HOME OF THE THE WORD DUPLICATION ILLUSION

Links 12/17: Silent Site, Holy Site

The world’s largest hotel is in Saudi Arabia, hosts 10,000 guests, and looks pretty much how you would expect the world’s largest hotel in Saudi Arabia to look.

Legends of Chinese immigrants in California, unsourced, sometimes a bit implausible. “John the Chinese laundry man was the laughingstock of Weaverville, California. For months he washed the Anglo miners’ clothes and never charged a penny for his services. But a year later one of the miners came across John wearing fine clothes in Sacramento. He had washed enough gold dust out of pants cuffs and shirttails to set himself up for life.”

Common vs. Specific Factors In Psychotherapy – Opening The Black Box. Key quote: “Neither variability in competence nor adherence [to the principles of the therapy involved] was related to patient outcome…extent of training might also not be relevant to outcome.”

Magic cards with @dril quotes as text. EG 1, 2

The size of a nation’s legislature tends to be about the cube root of its population. Also, the US House of Representatives is “one of the world’s most undersized” legislatures.

Sam Altman on the increasingly repressive climate in the Bay Area; makes some of the same points as my article about Kolmogorov complicity, but better, and with more personal experience. Tyler Cowen’s response. Related: GSS survey data shows high IQ predicts “free speech absolutism”.

Related: Heterodox Academy offers OpenMind, “a free, online platform designed to depolarize communities and foster mutual understanding across differences”.

More on the link between autism and transgender, with a few more studies than I’d seen before. Although only 5-10% of autistics have gender dysphoria, up to 25%-50% of transgender people may be autistic.

Late Christmas shopping idea: gravitational distortion placemats.

Contra Turkheimer and others, a new team finds no tendency for environmental influence on intelligence to be stronger in the poor, not even in the United States. [EDIT: Or maybe it doesn’t contradict Turkheimer, just show his results don’t extend to adults]

Also, even though the obvious evo psych explanation for bitter taste is that it’s supposed to warn us of potentially toxic molecules, there’s no real relationship between bitterness and toxicity.

The New I-66 Tolls Offer Great Insight Into Commuter Psychology. Commuters okay with a road being illegal to use (except for certain groups), but angry when it was legal to use but with a very high toll.

People Learned To Survive Disease; We Can Handle Twitter. Interesting take on cultural evolution including a micro-review of new James Scott book.

Some rare good news: the grad student waiver tax will not be in the final tax bill.

This month in the FDA: liberalization of rules on genetic tests like 23andMe (official statement, media summary). Related: probably legal for police to get your DNA from a genetic testing company if you’re a suspect; some good discussion of the exact warrant requirements buried in the Reddit comments. 23andMe has announced they will fight any such requests; unclear what other companies will do.

No, it’s not just your imagination: recent mystery interstellar asteroid Oumuamuamuamuamuamuamua does look kind of like a spaceship. Some good discussion in the comments here. And Robin Hanson on what it might teach us about interstellar space.

Venus only has one earthquake every hundred million years or so, but it’s a doozy.

Was James K. Polk the greatest US president? And Garrett Jones on which US president made the largest positive contribution to global income (hint: it’s James K. Polk)

The 100 most-discussed scientific papers of the year. A combination of health-relevant, politics-relevant, clickbaity, and groundbreaking new science. My girlfriend is lead author of #16.

The latest thing AI is outperforming humans at, this time very close to my heart: fantasy cartography. Example here. H/T Gwern.

Related, though you’ve probably seen it already: DeepMind has made an AI that can learn to play at superhuman level in various games including chess, Japanese chess, and Go – after just a few hours of practice.

Related: MIRI’s 2017 fundraiser. For those of you who don’t know, they’re a research institute that looks into the possibility of future AI superintelligence and how to make it safe for humans. I can vouch for them as good people; see also Zvi Mowshowitz’s I Vouch For MIRI.

Percent of people in different countries on who think life is better vs. worse than fifty years ago. More vs. less market freedom seems to be pretty big explanatory variable; being in Latin America doesn’t help.

This article purports to rank all generals and prove that Napoleon was the best. It’s gotten a lot of coverage, but it seems trivially wrong to me – as far as I can tell, it gives each general credit for their win vs. loss record, but doesn’t adjust for number of battles. So a general who fought 30 battles and won 50% would be “better” than a general who fought 10 battles and won 100%. As such, I can’t endorse it – but it’s a cool way of looking at things and I hope someone tries something similar and does it right – which would probably involve starting with a prior that each general is average and treating each battle as a new piece of Bayesian evidence.

We often hear that the amount parents talk to their baby is vital in explaining their development and life outcomes, so Scientific American profiles a South American tribe where parents practically never talk to their babies. But how many people from that tribe get into Ivy League colleges, HUH SCIENTIFIC AMERICAN?

In the IGM poll of economists, which I’ve cited a few times here as a good measure of expert opinion, top economists generally favor repealing Net Neutrality. H/T Buck, who writes that “if you think that repealing net neutrality is clearly bad, I’d love to bet you about it. Betting is a tax on bullshit and I feel like the internet is particularly full of bullshit at the moment; I’d like to do my part to clean it up a little while also hopefully making a little money. I’d love to hear your concrete predictions about how the world will be worse as a result of the repeal of net neutrality. I’m willing to spend at least a thousand dollars betting on this topic.”

Bay Area politicians die as they live: causing delays for local commuters.

This entry was posted in Uncategorized and tagged . Bookmark the permalink.

433 Responses to Links 12/17: Silent Site, Holy Site

  1. Sniffnoy says:

    Scott, is the “Reddit comments” link regarding genetic testing supposed to link to a particular comment? Right now it just links to the comment thread as a whole with some do-nothing anchor attached.

    Edit: Same with your other Reddit comments link, about the asteroid.

    Also the James Polk link could probably use some context (or maybe even more context).

  2. AnonYEmous says:

    Sam Altman on the increasingly repressive climate in the Bay Area; makes some of the same points as my article about Kolmogorov complicity, but better, and with more personal experience.

    I’m all for his goals, but his argument is weak. Viewed through the lens of making your argument about Kolmogorov complicity, it’s a much weaker version of your argument.

    To put it simply, you argue that Satoshi, in this example, wouldn’t be able to invent Bitcoin because he would be witch-hunted before he did it. He argues that Satoshi would’ve kept his mind closed on race issues and therefore on other issues, so he never would’ve tried to make Bitcoin. But I think you can easily just check yourself on an issue despite having some opinions on it – even I can, and I’m a Kolmogorov if you ever saw one. That doesn’t stop you from being open to new ideas, as far as I can tell. The entire thing just comes off to me as “meh”.

    Oh, and also:

    Considering both distributional effects and changes in efficiency, it is a good idea to let companies that send video or other content to consumers pay more to Internet service providers for the right to send that traffic using faster or higher quality service.

    I really don’t see how this doesn’t end up enforcing a monopoly on the market. Having to wait for your video to load is a bitch and a half, and combined with the already-existing natural monopoly structure that exists, there’s almost no way for the leader of the market to be dethroned. It would be better if consumers were simply charged extra, but then they might opt out of that and Netflix / Youtube and co. would prefer that this didn’t happen. Not to mention that it can easily extend to more than just video streaming. Top economists? I’m not impressed.

    • Why does it enforce a monopoly on the market?

      Suppose Comcast is both ISP and distributor of movies. They could give their movie firm an advantage with their ISP customers by charging high prices to their competitors, but why should they? They can get the same rents by charging everyone, themselves included, the price that maximizes their profits. If one of the competitors can do a better job of providing movies, Comcast makes more money, from their charges to that competitor, than it could make as a monopoly provider.

      As far as I can tell, all of the arguments really come down to worries about price discrimination. For some reason, people who think income inequality is the world’s biggest problem have never noticed that price discrimination, trying to charge different people different prices for the same good based on how much each customer is willing to pay, tends to reduce inequality.

      So far as the standard economic analysis is concerned, price discrimination is ambiguous–it can make the world, on net, either better or worse off. But since one effect is to make being an ISP more profitable, one likely result is more ISPs, hence more competition. And allowing ISPs to charge more for services that are more costly to produce is a pretty unambiguous plus. We don’t want a world where VWs and Teslas have to sell at the same price.

      • AnonYEmous says:

        Suppose Comcast is both ISP and distributor of movies. They could give their movie firm an advantage with their ISP customers by charging high prices to their competitors, but why should they?

        Because they could easily dominate the online movie market and make way more money that way?

        I don’t really know how online movie distribution works, but let’s imagine it’s similar to a Youtube video, which I bet it is. If you can deliver the movie 10% faster than the competition – i.e. ensure that buffering never has to happen or happens far less frequently – then any other advantage is not going to be sufficient to overcome that advantage, more or less. Now you dominate all movie sales, everywhere. Maybe selling some internet bandwidth is more lucrative, but I doubt it.

        Besides, this isn’t a good example of what I’m talking about anyways. Instead, I’m talking about all ISPs charging extra for fast service, or even just “regular service” which is now fast compared to other services (of video streaming, or whatever data-heavy thing we’re talking about). Under this system, all online movie distributors or video streamers or whatever are now forced to pay a bundle of cash, which they’ll gladly do, unless they can’t because they are too dang small.

        As far as I can tell, all of the arguments really come down to worries about price discrimination.

        They come down to worries about small companies not having the money they need to choose the price they’d want, and as a result those companies getting shafted. I also don’t care much about income inequality and I’m not planning to start, so don’t give me that.

        Anyways, here’s the bottom line: if ISPs are willing to charge differently for costlier services, that’s one thing. But if they charge the websites directly and if they charge them not by, say, amount of visits or something like that, then large companies will have a large advantage. That means less competition and less effective competition. It seems like you like competition, so hopefully you don’t like that potential outcome – and that outcome is what concerns me.

        • stucchio says:

          I don’t really know how online movie distribution works, but let’s imagine it’s similar to a Youtube video, which I bet it is. If you can deliver the movie 10% faster than the competition – i.e. ensure that buffering never has to happen or happens far less frequently – then any other advantage is not going to be sufficient to overcome that advantage, more or less.

          This doesn’t really make sense. I’m not going to watch e.g. Wonder Woman (which might be in the Comcast library) over the Walking Dead (in the Netflix library) simply because Wonder Woman buffers less.

          • hello says:

            What if Walking Dead was in both libraries?

          • Evan Þ says:

            At the margins, I absolutely would. Say I was indifferent as to which of two movies I’d watch. If I knew (or reasonably suspected) one would stop several times for a minute of buffering, I’m probably going to pick the other one.

            And as we’ve learned before here, the margins matter.

        • But if they charge the websites directly and if they charge them not by, say, amount of visits or something like that, then large companies will have a large advantage.

          Why? They aren’t going to try to charge the same price to a small company as to a large one, since they know the large one can pay much more.

          As best I can tell, the large companies have been pushing for the net neutrality rule, which suggests the opposite of your view.

          What the ISP is selling to their customers is access to the net. Anything they do that reduces the quality of that reduces the amount they can get their customers to pay for it.

          The ISP has two reasons to charge differently for different services. One is that different services cost them different amounts to carry–if a service results in very large volumes being downloaded, that slows things for their customers unless the ISP expands its facilities, which is costly. So they would like to charge a low price to services that don’t take much bandwidth, a high price to services that do.

          One way to do that would be to charge the customers different rates depending on what they use, but it might be easier to do it at the provider end instead.

          The other thing the ISP would like to do is to price discriminate, to charge a high price to people (customers or services) that are willing to pay a high price because the connection is very valuable to them, a low price to people who are only willing to pay a low price.

          Neither of those puts a particular burden on small or new companies. If anything the ISP has an interest in price discriminating in favor of new services in order to let them get established, increasing the value of an internet connection and hence what customers will be willing to pay for it.

          • Gobbobobble says:

            The ISP has two reasons to charge differently for different services. One is that different services cost them different amounts to carry–if a service results in very large volumes being downloaded, that slows things for their customers unless the ISP expands its facilities, which is costly. So they would like to charge a low price to services that don’t take much bandwidth, a high price to services that do.

            Or just charge based on bandwidth usage, like NN already allows them to do. Or, as mentioned downthread, put a data cap on the most egregious uploaders. Like they already do for (much more frequently downloading) residential users.

            The whole point of NN is that a byte is a byte is a byte. You don’t need to take that away to price discriminate based on network usage.

          • Mark Atwood says:

            the large companies have been pushing for the net neutrality rule

            This is one of *my* bailiwicks, much like economics of law, medieval cooking, and SCA fighting are yours.

            Everyone who’s lifeblood is their internet endpoint wants NN, from the very largest tech companies, down the the webcartoonist living on t-shirt sales. The small deviations from that base fact are some large companies who have decided that if NN will fall, they need to get ready to be cooks instead of dinner.

            You are suffering from a variant of the lamppost effect, where you think only the largest companies want this, because they are the ones who have the lobby footprint and “get out the word” muscle to be heard.

            To a good first and second approximation, the only companies that have heavily invested to kill NN are the debt-leveraged corrupt-local-politics incumbent last mile ISPs, and to a good first approximation, that means “Comcast” and “AT&T”.

            Know this, when I use terms like “corrupt politics”, “dirty lawfare”, and “they lie”, about those companies, I am not being even the slightest hyperbolic.

          • actinide meta says:

            On the whole I’ve been skeptical that eliminating NN regulation will be a disaster, and certainly most of the arguments to that effect are pretty weak.

            But your post actually suggests a better argument to me: transaction costs. ISPs have a relatively standardized, low cost way to bill their own customers. If they want to charge businesses to prioritize their traffic (with the ISP’s customers), as things stand they have to negotiate those deals individually, with lots of bargaining costs, risks of negotiation failure, etc. Maybe it’s worth it for every ISP to negotiate with Netflix (because they use so much bandwidth) but not worth it to negotiate with some new startup (because they are *currently* insignificant). If the negotiations take the form “pay us, or we will deprioritize your traffic”, then this actually helps startups. But if ISPs offer to give incumbents’ traffic an advantage over their “default” priority rules – and this is a natural ask for the other party in the negotiation – then a new startup is in the position of having to negotiate complex deals with a large number of ISPs worldwide before it can compete on a level playing field, and this is going to be difficult and expensive. Similarly, if an ISP wants to play in some business itself, convincing them to let you pay to compete with them is going to be very difficult even if in the long run that is the optimal solution for all parties. Almost by definition, when a new idea first comes along almost everyone thinks it is worthless.

            I still don’t think this is a fatal problem. ISPs can’t be too aggressive in throttling “default” flows without harming the value of their service to their customers, and it seems like if transaction costs of this sort become a serious problem, someone will come along to intermediate this kind of deal. The latter could even lead to a better world than the one we have now, where the internet actually prioritizes traffic in accordance with its economic value. (In theory, ISPs could do this more easily by negotiating variable cost plans with their own customers, and letting customers shape their own traffic. But the reality is that customers hate variable pricing.)

            But this is the clearest vision I’ve had so far of how we could wind up in a significantly worse place.

          • actinide meta says:

            An empirical argument that lack of NN regulation won’t lead to crushing transaction costs for new entrants is that content delivery networks, which are legal, essential to delivery of things like video with competitive economics, and require just this sort of negotiation (plus worldwide hardware presence!), are available on reasonably nondiscriminatory terms to everyone.

        • Conrad Honcho says:

          When something bad like that happens, I’m completely open to discussing regulation to correct the bad behavior. I’m not ideologically opposed to regulation, but I do have a healthy fear of the Law of Unintended Consequences. Pre-regulation of hypothetical problems leaves you open to the unintended consequences with no way of knowing if you actually prevented the hypothetical bad behavior.

          If Comcast starts doing evil stuff to Netflix, I’m pretty sure we’ll all know about it. But let them do the evil first before we go about adding new regulations.

          • Gobbobobble says:

            It’s not new. As has been mentioned, Net Neutrality goes back to the 90s, the Title II reclassification in 2015 was just the latest method.

            The rules were adapted from those developed for real telegram and telephone abuses, it’s not a what-if fever dream.

          • pontifex says:

            What makes you think you would ever know about the evil stuff Comcast does to technology companies? The deals are made behind closed doors.

            It’s not clear why Netflix (or any other random internet company) would want to publicize the details of these deals, beyond the extent required by SEC disclosure laws. It would scare investors to know how dependent the internet company was on the tiny oligopoly of last-mile providers. Comcast can offer companies who don’t publicize the deals better terms, and punish the ones who do make a stink.

          • pontifex says:

            By the way, this is also true of patent law. Companies pay patent trolls to go away all the time. But you almost never hear about it. Usually one of the conditions of the patent deal is that the company being extorted can’t make a stink about the extortion.

          • Aapje says:

            Yet the big companies don’t seek to weaken patent law, because they benefit from having a bunch of patents as weapons that increase the barrier to entry for new competitors.

            Those who actually end up hurt are not very visible, as the damage is what they cannot do, so we don’t see a loss (but merely not a gain), so the real damage is not very visible, as we cannot see what the level of innovation would be otherwise.

            Similarly, removing Net Neutrality runs the risk of damaging mainly upstarts in a similar way.

          • MoebiusStreet says:

            Net Neutrality goes back to the 90s, the Title II reclassification in 2015 was just the latest method. … The rules were adapted from those developed for real telegram and telephone abuses

            The Title II regulations go back to 1934, and were revisited as recently as the 90s. As you mention, they go back as far as telegraph.

            Why would anyone think that a set of regulations that was engineered for a circuit-switched voice network would be a good approach to the regulation of a packet-switched high-bandwidth data network? There are huge fundamental differences in how those networks operate, and what their goals are.

            I think everyone directly involved in the regulation agrees that this is a problem. While NN was in effect, the FCC had agreed to just overlook the aspects of the regulation that make no sense.

            But a law that can be exercised at the pleasure of the government is just a tool waiting for a really crappy administration – say, Donald Trump’s – to misuse against its enemies.

          • Gobbobobble says:

            Why would anyone think that a set of regulations that was engineered for a circuit-switched voice network would be a good approach to the regulation of a packet-switched high-bandwidth data network? There are huge fundamental differences in how those networks operate, and what their goals are.

            Why wouldn’t they? Voice is a subset of data and packet-switching is essentially just an automated form of circuit-switching. As demonstrated by VoIP, the internet can be considered a bigger and better iteration on phone networks. Do you not consider internet a telecommunications technology?

            If you want to argue for fundamental differences you’ll have to actually make your case instead of assuming it as a given. NN can be distilled to a simple principle:

            Networks that carry data from A to B are not allowed to discriminate or interfere with the data being carried

            What am I missing that makes this not apply to both phone lines and internet connections?

          • MoebiusStreet says:

            NN can be distilled to a simple principle

            No. The argument was that a particular implementation of NN was absolutely necessary. Thus, we’re debating whether that implementation makes sense, not whether some nutshell summary of a philosophy is a good one.

            Voice is a subset of data and packet-switching is essentially just an automated form of circuit-switching

            You seem to be arguing that because something works well for a subset, we ought to be able to apply the same principle to the whole superset. That’s clearly a fallacy.

            And in practice, the FCC even acknowledged that when they were enforcing NN. They recognized that portions of the regulations were clearly silly, and promised not to enforce those portions.

            So the second half of my response was regarding the political dangers of trusting the government with “don’t worry, we won’t enforce this silly law, even if we think it could gain us significant political advantage”.

            If we want to craft a NEW NN framework, it would make a whole lot more sense. But that doesn’t seem to be what most people are arguing for – they’re arguing for the re-imposition of the legacy regs.

          • Gobbobobble says:

            No. The argument was that a particular implementation of NN was absolutely necessary. Thus, we’re debating whether that implementation makes sense, not whether some nutshell summary of a philosophy is a good one.

            Disagree. Unless “trust Comcast et al in the (supposedly) free market to take care of it” counts as a different implementation instead of a philosophical difference.

            You seem to be arguing that because something works well for a subset, we ought to be able to apply the same principle to the whole superset. That’s clearly a fallacy

            You appear to be arguing that just becomes something works well for a subset, we ought to not apply the same principle to the whole superset. Which is also a fallacy. Some make sense, some don’t. The point of that excerpt was that your “Why would anyone think” bluster was unwarrantedly dismissive.

            You completely ignored my actual example of what voice and data have in common as support for why this falls into the “makes sense” bucket.

          • David Speyer says:

            “When something bad like that happens, I’m completely open to discussing regulation to correct the bad behavior.” This seems like a good time to mention that the proposed method of ending NN is for the FCC to switch from regulating ISPs under Title II to regulating them under Section 706. Under Section 706, the FCC would not have the authority to reimpose these regulations. If the FCC wanted to relax the regulations and remain able to rapidly reimpose them, it would keep the Title II authority and just relax the regulations.

          • bean says:

            If the FCC wanted to relax the regulations and remain able to rapidly reimpose them, it would keep the Title II authority and just relax the regulations.

            That’s not really an improvement. Giving the government power to do whatever it wants, and trusting that it won’t use it unless it has to rarely ends well. That’s one of my main objections to Title II regulation. If we need to keep the net neutral, let’s come up with a law specifically designed to do so, one that doesn’t give them power to set rates.

      • Murphy says:

        Trying to challenge a big ISP who already has their own infrastructure built out is spectacularly expensive and there isn’t even much reward if you successfully drive prices down and there’s a non-trivial chance they can just temporarily drop their prices until you go bankrupt then buy up any infrastructure you built out. After all they have all those other regions where they have effective monopolies to bankroll them.

        Trying to muscle in is high cost, high risk and low reward.

        Throwing more fistfuls of everyone elses money at them is not going to help that. If anything it’ll just entrench them more.

        NN is pretty similar in many ways to monopoly rules in other markets: you don’t get to use a monopoly in one market to take an unfair advantage in another.

        Online video services are currently in a good state, low prices, stiff competition etc.

        ISP’s that offer other services have already demonstrated a willingness to blackhole, slow or cripple protocols and services that compete with their own with ISP’s slowing competitors video services or trying to block VOIP protocols that compete with their phone services.

        It’s not a hypothetical or “why would they”, it’s “they are actually already doing this” where they’ve been allowed. Your hypothetical seems to be heavily contaminated by optimism and you may be missing the optimal strategy of “eating little girls” because you want the nice option to be true.

        Which isn’t great for competition. being allowed drop network traffic from their main competitor into a black hole or just degrade service for a large fraction of their customers and being able to do so transparently so their customers aren’t even informed it’s happening and can be mislead to beleive that the competitors service is just buggy or crap…. it’s not terribly appealing.

        there’s also an overhead cost and the death of future competition.

        Today very little is stopping me from setting up my own streaming service or other high-bandwidth service.

        I need a box connected to the net and something to send.

        Even as an average guy I could buy or rent enough kit to service a reasonable number of customers with my meager savings and start my own outfit.

        that all changes if I need a legal team to negotiate rates with a couple hundred ISP’s. I can no longer treat the network as just, the network, now I need to be part of the chumocracy to get anywhere.

        Now there is the possibility that my network provider might simply negotiate peering arrangements for traffic with their trunk provider and once they’ve done that I don’t need to face and negotiation costs! wouldn’t that be ideal! perhaps such a system will grow up! oh wait, that’s exactly the system the big monopolies want to burn down.

        once the cancer has been allowed to grow it will be almost impossible to cut out.

        • MrBubu says:

          I think DavidFriedman’s Argument is fundamentally solid. However I am afraid that the environment in which e.g. video streaming is happening is “too difficult” for humans.
          1. There is no way to know how much you will like a movie
          2. Humans overvalue trivial inconveniences. Sometimes waiting another 10s for a vastly better movie is too much
          3. If you make the wrong choice and watch a worse movie because of a 10s delay YOU WILL NEVER KNOW. There is no chance for learning.

          • Murphy says:

            It sounds like a lovely argument. And it’s very hopeful and optimistic.

            but when hopeful and optimistic arguments yield predictions counter to real observations in the field it implies that the people making nice sounding predictions are missing something.

            We all want nice things to be true but sometimes the optimal strategy that yields the most advantage for an actor isn’t the nice strategy.

            Sometimes the optimal strategy isn’t to be a nice competitive actor in the marketplace, sometimes the optimal stratagy is to play unfair, destroy your competitors through means other than fair competition and then jack up the prices to the maximum the market can bear.

        • poignardazur says:

          I hear the “it’s impossible to make money competing with ISPs, they’d just crush the competitor” argument all the time.

          It’s really confusing me, because this exact impossible thing happened in France: there was an oligopoly, Free happened, drove the prices down, France is now one the developed countries with the cheapest internet in the world (I checked, the average price in France is around 25€/month for 1GB/s, against 150$/month in major US cities unless you have Google Fiber).

          How does this not happen in the US? Was there something special about the French market? Did the scenario you theorize (new ISP arrives, gets crushed by major ISPs) actually happen?

          I know Google tried to wedge in on the market with Google Fibre, but apparently they stopped expanding. But they sure do undercut prices where they set up. Did they make money on balance?

          • A1987dM says:

            France is now one the developed countries with the cheapest internet in the world (I checked, the average price in France is around 25€/month for 1GB/s, against 150$/month in major US cities unless you have Google Fiber).

            FWIW all other places in Europe I’m familiar with are more similar to France than to the US (both in terms of prices and of number of different ISPs you can choose from), so I guess if anything it’s the US market that there’s something special about.

          • Murphy says:

            Simple answer:
            brutally enforced government regulation.

            An incumbent provider and providers with significant market power have universal service obligations whereby they can be required by law to allow other suppliers to use their network infrastructure with non-discriminatory conditions.

            Sorry, not the libertarian answer you wanted but the genuine real answer to how france succeeded so hard is Heavy Handed Government Regulation.

            Also: mobile wireless infrastructure being vastly easier to enter compared to landline infrastrcture where you would need to gain access to vast amounts of dark fiber in order to compete effectively.

          • Evan Þ says:

            To follow up on @Murphy, “brutally-enforced government regulation” is also why Google Fiber didn’t get built here in Seattle: the existing infrastructure down to leases on the poles was owned by Comcast, and setting up new fiber would’ve involved negotiating permits from a dozen different city departments. Faced with such bureaucracy, Google threw up its hands and went elsewhere.

          • 1soru1 says:

            A regulation that says ‘no competitors’ can reasonably be expected to have a different effect from one that says ‘no monopolies’.

          • poignardazur says:

            @Evan Þ: Yeah, I’ve been hearing a lot of stuff like that. I’d appreciate sources if you can find them, because the subject drives me mad. Doesn’t the US have systematic local loop unbundling? If not… first off, what the fuck? And second, why isn’t this what everyone is campaigning for?

            I mean, it seems everyone agrees the major problem the US have is a lack of competition. But… nobody since to be talking about it at all, except to say “You silly libertarians, your ideas have no powers in this land, for there is no competition!”… but, like, is nobody worried about fixing it?

            Why are thousands of websites and youtubers and Mozilla Firefox making huge-ass petitions to maintain an overall minor, legally dubious bit of legislation (don’t get me wrong, I don’t think it should have been repealed, but it’s not the end of the world either) when they should obviously lobby for better pro-competition legislation?

            That’s a rhetorical question, by the way. I can already guess most “people are more interested in scandals that in solutions” types of answer.

          • Evan Þ says:

            @poignardazur, I’m afraid I’ve only got one source off the top of my head, but here’s this article about Google Fiber and Seattle.

          • poignardazur says:

            Thanks!

          • Murphy says:

            @Evan

            I think it may be partly down to a different view on regulation and competition.

            There’s a view that seems to be popular with a lot of americans where “regulation” is some kind of synonym for “evil”.

            If you see it as some kind of us vs them, regulation vs market competition thing then it’ll look like some kind of contradiction.

            But there’s a lot of people who are all for “yay capitalist market competition” who also aren’t anti-regulation and that mix seems to be more a little more common in Europe.

            So you regulate hard but in such a manner as to create a structure within which market competition can work best.

            Meanwhile the regulations=bad view leads people towards hoping that if they strip away all regulations then that will somehow immanentize the world spirit and bring about optimal market competition.

            There are plenty of shoddy regulations out there that make things worse, often enforcing a poor equilibrium, unfortunately what some people get from that is just “regulation=bad”

          • cassander says:

            @murphy

            So you regulate hard but in such a manner as to create a structure within which market competition can work best.

            this is something that is easy to say but very hard to do. See, for example, the ACA exchanges. It’s also very hard to sustain over time, because once the regulatory framework exists, there will always be a natural tendency pushing the regulators achieving desired outcomes instead of simply ensuring a level playing field.

          • 1soru1 says:

            > this is something that is easy to say but very hard to do.

            Lots of things are hard. Given that it is demonstrably possible, any country that wants to stay competitive with its peers is either going to have to put in the effort required to learn how to do it, or find some economic niche that doesn’t require widespread internet.

          • cassander says:

            @1soru1 says:

            Lots of things are hard. Given that it is demonstrably possible, any country that wants to stay competitive with its peers is either going to have to put in the effort required to learn how to do it, or find some economic niche that doesn’t require widespread internet.

            The correct response to “X is hard” is not “Someone, somewhere, once succeeded at X, therefore everyone should try to do X”. It’s “have appropriate humility about the likelihood of achieving X and try to find solutions that minimize reliance on X.”

          • shenanigans24 says:

            I know the 1934 regulation didn’t prevent monopolies it created one. Why would it be different today.

            Airline regulation was the same. Not one single new airline was approved prior to deregulation.

      • The Nybbler says:

        Suppose Comcast is both ISP and distributor of movies. They could give their movie firm an advantage with their ISP customers by charging high prices to their competitors, but why should they?

        Why does Amazon refuse to allow Chromecasts to show Amazon Video? Sometimes, the would-be monopolist is just an asshole.

      • Andrew Cady says:

        ISPs can (and do) already sell more costly service at a higher price TO THE USER. Allowing them to double-bill the other end of the connection for that same cost will allow them to capture a higher portion of value that they don’t create. That doesn’t seem like a good thing because there is less money to encourage the creation of that value.

        “More ISPs” doesn’t make sense. ISP is a natural monopoly, not perfect competition market. The ISP competitive market is fantasy.

        • A1987dM says:

          “More ISPs” doesn’t make sense. ISP is a natural monopoly, not perfect competition market. The ISP competitive market is fantasy.

          Huh? AFAIK most cities in Europe have around half a dozen major ISPs, and that’s not counting mobile ones. Why is the US that different? (Not a rethorical question — I would find your claim unlikely if I didn’t know about the current status of the US market, and it’s very possible that whatever it’s keeping it from becoming more like the European one is unlikely to be eliminated in easy ways such as just restoring net neutrality regulations.)

      • sunnydestroy says:

        I posted this on an open thread but maybe I’ll get more responses here.

        Looking for any critiques, criticisms, or agreement on these economics papers about the effect of Net Neutrality regulations on economic efficiency and on rural access to broadband. They’ve been shared with me, but I don’t feel qualified to judge their quality.

        I found this July 2017 review of the economic literature that reaches a mixed conclusion. More specifically, that the regulations can be harmful or helpful depending on the conditions in the marketplace. In general, restrictions on what advanced features ISPs can offer lower economic efficiency. However, the blocking of content users value was found also found to be harmful. The literature suggests ISPs have characteristics differing from the telecom companies that Title II was written for. I don’t know if I’m interpreting this right, but it looks like Net Neutrality is more likely to be economically detrimental?

        One grain of salt for this research is that Connolly is a former chief economist at the FCC and anti-neutrality. I’ve heard that a lot of the economists in this paper are anti-neutrality as well, so maybe the sampling of studies is selective?

        The other paper is also by Connolly. The abstract says the Open Internet Order might widen the digital divide (Paywall so I haven’t seen the beyond the abstract):

        In its 2016 Broadband Report, the Federal Communications Commission (FCC) recognizes that a rural/urban digital divide remains prevalent—especially with respect to broadband adoption. It also highlights several policies that the FCC has undertaken purportedly to reduce the divide, including the 2015 Open Internet Order (OIO)—in which the stated intent is to enforce “network neutrality.” However, long before the OIO, studies have raised concerns that network neutrality policies will discourage investment by internet service providers (ISPs) in broadband infrastructure, to the detriment of broadband accessibility, and may increase average consumer costs—both of which would only further exacerbate the digital divide.

        Commenter pontifex offered one criticism:

        So, the crux of the argument seems to be this:

        “…with current net neutrality regulation as specified in the 2015 OIO, ISPs are not allowed to charge any fees to CSPs. Hence, ISPs are forced to generate revenue solely from last-mile fees, which is likely to result in an increase in the average price that is paid by end consumers.”

        This doesn’t really make sense, though. If the ISP charges Netflix $20 to connect to each customer, Netflix has to increase its prices by $20 per customer, in the long run. The content companies are not magical unlimited pots of money. Price increases eventually have to be passed on to consumers.

        Someone also told me that the EU’s net neutrality regulations have stifled investment in rural areas when compared to the US. A counterpoint to this is what commenter pontifex has told me in another thread:

        The US had government programs and policies specifically designed to incentivize internet service providers to provide service to rural and semi-rural areas. Some of them were quite expensive and controversial. What is considered “broadband” in one place might not be in another. And the way that internet service is provided is drastically different in some countries. The US model of private entities owning the last mile is not universal.

        Stats-wise, there does seem to be a gap between the EU and US rural broadband access for networks with speeds equal to or greater than 25 Mbps.

        The 2016 European Commission report on broadband access says:

        However, only 39.2% (12.0 million rural households) could benefit from NGA broadband.

        With NGA defined as equal to or greater than 30Mbps speed.

        For the US, this FCC report says 39% of rural Americans lack access to 25Mbps broadband, which would mean 61% of rural Americans have access to broadband of at least 25Mbps speed.

        So according to 2016 data, the difference is 39.2% (EU) vs 61% (US) for rural high speed (>25Mbps) net access.

  3. ilikekittycat says:

    Net neutrality predictions contingent on net neutrality stays repealed for at least several years and becomes the new regime, so to speak (no states overriding it, no democrats taking over congress and taking it back)

    1. At least one new realm of promising innovative content production (twitch streaming, podcasts, something currently associated with the little guy bootstrapping something rather than needing expensive corporate production) will wither away compared to its current form, because people will be nudged away from “getting addicted” to strange new things that have huge megabyte demands.

    2. US ranking will continue showing up in the bottom quartile for internet connection speed average in the developed world and no blitz of infrastructure will make things better all of a sudden. All these libertarian dream plans about “well what about the guy WHO ONLY NEEDS 4 mbps to check email twice a day and wants to get it for $8/month” will amount to naught. People will continue being dissatisfied with payment plans and bundling deals they only grudgingly take (i.e., I don’t think there are market forces wanting to give us a vastly better internet waiting to be released, the best possible world for this repeal is “things stay mostly the same.”)

    3. New limits and bounds that don’t make good common sense for the medium of bandwidth will be implemented for the average internet consumer: i.e., data limits making the jump from phone plans to internet wired connections

    4. Internet service that isn’t part of what the average 50 year old needs will suffer or degrade: I’m actually not worried about Netflix at all, I think Netflix will be accommodated at the cost of a thousand other things that were once “default” parts of the internet, the same way that once upon a time a little image hosting space, usenet access, etc. was the default for having a broadband ISP account but isn’t anymore

    • actinide meta says:

      As far as I know, data limits and 4mbps plans were already legal under NN regulations.

      • Gobbobobble says:

        Then what do they need NN repealed for? If Netflix using too much bandwidth is their True Objection, slap an upload limit on them.

        • Mark Atwood says:

          Because they are lying.

          They want NN repealed so they can demand “revenue sharing” from profitable companies.

          We may yet snatch a victory from this setback, there is now a fast rising pressure on local and state politicians about the cozy corrupt local monopoly arrangements that the last mile ISPs have lied, cheated, and defrauded their way into.

          • Gobbobobble says:

            Precisely. And I hope you are right about the last part.

          • Alex Zavoluk says:

            ” is now a fast rising pressure on local and state politicians about the cozy corrupt local monopoly arrangements that the last mile ISPs have lied, cheated, and defrauded their way into.”

            This is the actual problem. Even if NN repeal allows ISPs to engage in all of the site-specific throttling and what-not that everyone is so afraid of, putting it back into place won’t stop the ISPs from continuing to screw everyone who lives in a local monopoly.

      • ilikekittycat says:

        You are correct that both of those things are legal, and I don’t think net neutrality will have much of a legal effect honestly, it’s mostly gonna be changes in the culture of how the ISPs get to treat customers

        1. In a lot of situations, like my present one with Cox cable, you’re just helpfully being “counted.” I get “1 TB” of data cap for each month in the last year, but nothing happens when i repeatedly go over. My sense is that that, and new sorts of limitations, will become the norm, and become enforced, if net neutrality repeal persists

        2. 4mbps are already legal, and some people do have them (people in the boonies, say, who only get sattellite connections.) I’m trying to find something to argue against the proponents of the repeal, and one of the only things I can get from them is that too many people are in plans that are more than they need, and after this, you’ll see far less people getting (say) a 20 mbps plan with bundled cable Tv when all they need is the 4mbps. I apologize if this was just what I heard from who I was arguing with specifically and doesn’t actually represent a widespread prediction from net neutrality repeal people, but that’s one of the only things I’ve gotten from arguing the point so far

        This whole scheme seems backwards, honestly, it should be the net neutrality people that defend the status quo ante and pro-repeal people that have to make predictions and risk money about how things are gonna improve

        • GNUNotUnix says:

          This whole scheme seems backwards, honestly, it should be the net neutrality people that defend the status quo ante and pro-repeal people that have to make predictions and risk money about how things are gonna improve

          I think regulators should have the burden to show that their regulations are improving the world. Private individuals’ freedom should be the default stance, and it’s the people with guns (the govt) who would like to impose their views who should have the higher burden.

          Your view especially doesn’t make sense in the NN case because NN was implemented just a few years ago by the Obama Administration. I never noticed any difference before and after NN, so I think the burden is on the regulators to really show me that I was being harmed before NN and I was helped dramatically after NN. I don’t think anyone did that. I just got a bunch of emotional pleas.

          • Gobbobobble says:

            Your view especially doesn’t make sense in the NN case because NN was implemented just a few years ago by the Obama Administration.

            It has been mentioned several times already that this is flagrant misinformation.

          • GNUN: Your view especially doesn’t make sense in the NN case because NN was implemented just a few years ago by the Obama Administration.

            Gobbo: It has been mentioned several times already that this is flagrant misinformation.

            Yeah, mentioned, but never explained. Someone above said that things were getting worse somehow, and so Obama’s regulations were keeping the status quo? I admit that I don’t know too much about how the Internet works, but reading all the comments gives me the impression that no one else does either. If someone could explain this better, maybe Obama’s regs would make sense. At this point, I am with GNUN.

          • Montfort says:

            It has been explained several times, but I wouldn’t necessarily have expected you, personally, to have read it. Prior to 2014, the FCC had attempted to enforce Net Neutrality though the legal authority to do this was not totally clear. This effort produced the Open Internet Order, adopted 2010 (though, again, the FCC was investigating unfair practices even before this, e.g. Madison River for allegedly blocking VOIP traffic in ’05, Comcast for throttling torrenters in ’08).

            Then, in 2014, a court ruling (Verizon v. FCC) stated that most of the Open Internet Order was invalid unless ISPs were designated “common carriers.” Then came the Net Neutrality hullabaloo of 2014-15, until the FCC under Obama decided to designate the ISPs common carriers and enforce Net Neutrality once more.

            Now, at the end of 2017, the FCC has voted to repeal the earlier decision, dropping the designation of “common carrier.”

            So the argument is that although yes, technically, the recent vote reversed a decision from just a few years ago, in fact the FCC has been attempting to regulate ISPs according to Net Neutrality since the mid-2000s.

    • Evan Þ says:

      There’re enough twenty-something-year-olds that I’m not really concerned about #4. What I’m concerned about is the next big idea getting choked in its cradle by needing to negotiate deals with all the ISP’s… and we’ll probably never notice, because there’re always a hundred other possible reasons for a Big Idea to not grow to maturity.

      • MrApophenia says:

        I do wonder if we’re about to see a huge resurgence in peer to peer networking (of both piracy and non-piracy varieties). If I can’t get a decent bandwidth connection to my content provider of choice anymore, I guess I’ll just need to go back to BitTorrent, or some new equivalent.

        • Andrew Cady says:

          But the ISPs are now allowed to throttle bittorrent too. They can throttle by protocol. For any centralized service they have reason to throttle, they have the same reason to throttle its P2P alternative.

    • shenanigans24 says:

      My internet was more expensive and slower in Germany. I hear the same about cell service which was also garbage when I was there. This in a country vastly smaller where the infrastructure should be cheaper. I don’t know why there’s a gap between my, and all my friends experiences there vs what is supposedly happening.

      I get my internet from wifi now in the US by AT&T. This seems like the most likely future and one with much lower entry costs.

  4. Incurian says:

    The general ranking guy doesn’t seem to take attack and defense into account, just raw numbers. From what I remember, the historical force ratio for an attacker to overcome a defender is 1.5-6:1, depending how well prepared the defenses are (those numbers are both from memory and possibly not too reliable to begin with, but I still think it’s an important factor). There are probably a lot of nitpicks about his method, which I understand wasn’t meant to be especially rigorous anyway, but this seems like a major one that could have been overcome easily.

    The other issue that springs to mind is the naming. I think even if this were a rigorous assessment, the most you could say is “most winningest” not “best.” Imagine a general with crummy politicians who is forced into a war he can’t win, but does his best with brilliant delaying actions (each of which probably classified a defeat), punching well above his weight. Or Vietnam where the US “won” all the major engagements but didn’t accomplish their putative strategic objective.

    • Alex Zavoluk says:

      There are certainly differing ways in which one could define a “great general.” Maybe some generals put themselves into a place where they didn’t even have to fight. General Sherman did this late in the Civil War, just marching his troops all over the place to push the Confederates around without fighting nearly as much he would if he attacked them directly. That would be hard to capture in data, though.

      As Scott points out, raw number of battles fought is highly emphasized in this methodology. One conception of “best general” would be more like “highest probability of winning a battle, all else held equal.” Alexander has a 100% win rate, though 9 battles is kind of a small sample size. On the other hand, he captured a massive amount of territory. Along a similar line to the “avoid having to fight at all” idea I mentioned above, maybe certain generals can arrange their victories to accomplish more, or win by a larger margin, needing one decisive victory rather than a war of attrition through several muddy victories.

      I wonder if the model looks weird because generals are not randomly assigned to battles. That is, if you’re in a hopeless situation, you send your best general, and maybe the other general sends their worst, and the battle is a lot closer than you expect.

      The author mentions technology, but that may be one of the primary reasons Lee is ranked so low. The North had a reasonable manpower advantage in most cases, but it had a really big technology advantage, for example in ships, trains, and guns. Lee took risks because a war of attrition would inevitably favor the North; they didn’t pan out, but probably had a higher chance of winning than fighting more conservatively.

      • powerfuller says:

        I disagree that a war of attrition would have favored the North, and that Lee’s risky battle of annihilation strategy was thereby justified. It’s true that the North had almost limitless resources compared to the South, but it didn’t have limitless morale. If Lee had adopted a defensive strategy more along Longstreet’s ideas (who seemed to foresee the trenches of WWI surpassing Napoleonic tactics), it’s very possible the North would have exhausted its will to fight on more Fredericksbergs and Cold Harbors.

        But Lee disliked being called “The King of Spades” by his troops and the press when he reinforced Richmond early in the war, and like much of the Southern leadership was stuck in the old mindset of glory and valor in battle. So he turned to big battles. Had he swallowed his pride and kept digging, he might have won. Washington was a better general the Lee because he knew he didn’t have to win his war so much as not lose it. Lee’s prowess as a tactician is probably overrated anyway; he main strengths were his audacity, which Union generals mostly lacked, and his nigh godlike levels of charisma. But the latter probably worked to his detriment in the end, as he became overconfident in his army at Gettysburg, which was in several ways a tactical disaster.

        • Alex Zavoluk says:

          It’s true that an infinitely long war would have seen the North eventually give up, but I think they would under almost any circumstances win a war of attrition long before that point, particularly with the blockade. Moreover, Sherman proved that the Union army could do quite a lot without having to shove themselves into entrenched positions, and that some of the commanders at least had learned by the end of the war that such tactics were futile.

          I’ve mostly seen Gettysburg described along the lines I described, as a gamble designed under the conclusion (even back in early 1863) that the North would win a long war. Vicksburg was captured the day after the end of the Battle of Gettysburg, and made waging war after that point quite difficult, regardless of Gettysburg.

          You make an interesting point about Napoleonic vs. trench warfare. It’s something I’ve wondered myself, why there seemed to be so many more defensive fortifications in the Civil War than previous wars. I don’t know of any good answer. The most I’ve been able to come up with is improvements in firearms technology, which might have made “digging in against a numerically superior force” a viable tactic, when before you would have simply been overrun by a bayonet charge or cavalry. I’m not actually sure, though.

          • Chevalier Mal Fet says:

            I get nervous posting stuff like this with so many bona-fide experts around, particularly in military matters, but here’s my layman’s understanding of the increased importance of fortification in the Civil War: The Civil War was the first war in which rifled guns were a major factor.

            Now, rifling had been around for ages before the war – the British even experimented with an all-rifle unit in the Peninsula. But early rifles couldn’t really compare with standard muskets as anything other than a specialist weapon. The reason lies in the way a rifle works. I’m going to run through the basics just in case anyone’s unfamiliar, but rifling a barrel involves a series of spiral grooves carved into the sides of the barrel of a gun. This imparts spin on a projectile, so that it’s more accurate – like the rotation of a football.

            Now, in order for rifling to be effective, the projectile has to grip the sides of the barrel, otherwise it won’t spin, right? But if your bullet is literally touching the sides of the barrel as you load it, loading is a very difficult and tedious process as you force your ball down. Your rate of fire is slow, at a time when 3 rounds per minute was considered respectable. On the other hand, if you use a smaller ball for quicker loading, the ball rattles around the barrel as you fire and you lose all the gains in accuracy that rifling provides. Might as well have a more cheaply-produced musket in that case.

            What changed this equation was the Minie ball. The Minie had a chamber at the base of the bullet which expanded when weapon was fired. When you loaded, the small ball was easily forced down the barrel, so you had a high ROF. When you fired, the bullet expanded and gripped the rifling, so you had great accuracy. Suddenly armies could put out accurate, aimed fire at 800+ yards instead of ~150. A man with a bayonet can cross 100 yards and stick you before you get off more than a round or two (what cavalry can accomplish is left as an exercise to the reader). Close quarters fighting and cavalry charges are viable in that scenario (although I suppose it’s worth noting that even in Napoleon’s day cavalry charging head-on into formed bodies of infantry and breaking them was a relatively rare occurrence). At 800 yards? Not so much. And of course none of this touches on the massive improvements artillery underwent in the 50 years between Waterloo and Appomattox, but I don’t know much about the specific innovations there so I must pass over them in silence. I DO know that the other guy became far more capable of killing you at distances never imagined before.

            So, the only solution is to dig. Early on, generals tried wars of maneuver, and suffered appalling casualties as most battles devolved into bludgeoning matches while the defenders rapidly dug in (soldiers learn quick what keeps them alive). In the wide-open spaces of the sparsely settled western theater, this maneuvering continued until the end of the war (Grant’s Vicksburg campaign and Sherman’s Atlanta campaign are good examples). In the confined East, though (especially northern Virginia, carved up by so many rivers), you couldn’t really avoid the bludgeoning match, and so the war there quickly came to resemble 1914 (compare Grant’s Overland campaign and siege of Petersburg with the Battle of the Marne and the Race to Sea, for example).

          • Evan Þ says:

            Vicksburg was captured the day after the end of the Battle of Gettysburg, and made waging war after that point quite difficult, regardless of Gettysburg.

            I’ve heard this often myself, but why? I see how capturing the whole Mississippi was a huge gain for the North, but why was it a huge loss for the South? Were they previously getting a lot of supplies from the Trans-Mississippi, was it just the loss of communications, or was it something else?

          • Alex Zavoluk says:

            Yes, the Minie ball and the spread of rifling was primarily what I meant when I referred to improved technology (also smokeless powder, which is more powerful and reliable than black powder, and rapid-fire technology such as the breech-loading rifle and gatling gun). What I wasn’t sure about is to what extent these more accurate and longer distance rifles changed tactics.

            Also, I didn’t think early rifles had that much range? I thought it was more like an improvement from 50-100 yards to 200-300 yards.

            “I’ve heard this often myself, but why? I see how capturing the whole Mississippi was a huge gain for the North, but why was it a huge loss for the South? Were they previously getting a lot of supplies from the Trans-Mississippi, was it just the loss of communications, or was it something else?”

            It separated the South into 2 regions, making it much more difficult to get supplies between Arkansas/Oklahoma/Texas/Louisiana and the rest of the South, and prevented them from moving troops and supplies up and down the river. It was also part of a larger plan to generally cut the South off from foreign trade. That wiki article claims the consensus is that capturing the river was as important as the major Eastern battles.

          • Nornagest says:

            Wikipedia gives an effective range for Revolutionary War-era rifles of 200 yards. I have been told by Revolutionary War enthusiasts that 300 was achievable in skilled hands.

            The Minié rifle was good out to about 600 yards, so we’re talking roughly a doubling of range over a hundred years. Aside from the better gas seal described above, most of the difference might be down to the ballistics of the Minié “ball” — while the two weapons used similar powder charges and fired projectiles of similar weight, Minié projectiles are roughly bullet-shaped, while early rifles actually fired balls, which are very aerodynamically inefficient. Better manufacturing practices would also have contributed to accuracy and therefore effective range — the Civil War was just about when it became possible to reliably manufacture stuff to set tolerances.

          • Chevalier Mal Fet says:

            I’ve heard this often myself, but why? I see how capturing the whole Mississippi was a huge gain for the North, but why was it a huge loss for the South? Were they previously getting a lot of supplies from the Trans-Mississippi, was it just the loss of communications, or was it something else?

            Archer Davis, in his book How the North Won, argues that the Mississippi had been effectively closed by the middle of 1862, and that the real significance of Vicksburg was the loss of Pemberton’s 30,000 veterans. Without that army, the Confederates in the West lacked the strength to stop the real hammer blows of Chattanooga and Atlanta.

            Also, I didn’t think early rifles had that much range? I thought it was more like an improvement from 50-100 yards to 200-300 yards.

            Early ones did not, but by 1861 the Springfield had an effective range of 400 yards. The 1863 model developed during the war could reach 800 – 1000 yards at maximum range, 500 yards effective. Most soldiers didn’t have the marksmanship to exploit all that range, of course, but it was more than enough for your common grunt to reach out and touch people with – ranges were quintiple what they’d been 50 years earlier.

            I think that’s really all you need to explain the increased importance of field fortifications during the Civil War. Soldiers adapt to environments of increased lethality remarkably quickly, I believe.

          • Evan Þ says:

            It separated the South into 2 regions, making it much more difficult to get supplies between Arkansas/Oklahoma/Texas/Louisiana and the rest of the South, and prevented them from moving troops and supplies up and down the river.

            How frequently were the Confederates doing that, though? Given the state of Southern railways, I’d be surprised if any major amount of supplies were coming across the Mississippi at any point in the war.

            Perhaps naively, Chevalier Mal Fet’s argument (that the main point was the capture of Pemberton’s troops) makes more sense to me.

          • baconbacon says:

            I’m not sure of it’s relevance to the Civil War, but the rail lines were pretty important to the effectiveness of digging in during WW1. Without them a portion of any dug in line could be attacked for a prolonged period before reinforcements and extra supplies could arrive. This is one reason, especially before the telegraph, that digging in is very risky. You could find yourself in an awkward position suddenly, and be forced to hold out for an indefinite period of time.

          • Doctor Mist says:

            @Chevalier Mal Fet-

            I love this place. I’ve known the term “Minie ball” for decades, and understood the importance of rifling for even longer, but had never actually heard what made the minie ball so important. Thanks!

          • The original Mr. X says:

            The Civil War was the first war in which rifled guns were a major factor.

            Ahem.

            Also, I think the importance of rifles in the USCW is exaggerated. Both sides gave their troops marksmanship training which ranged in quality from “abysmal” to “functionally non-existent”, so whatever the theoretical range of their rifles (up to 800 or 1,000 yards, as the Crimean War shows), in practice troops engaged at ranges not much greater than those of the Napoleonic Wars.

          • Aapje says:

            In general, armies eventually started retreating from the idea that longer range accuracy was important for the regular infantryman, focusing more and more on the short and medium range.

          • John Schilling says:

            In general, armies eventually started retreating from the idea that longer range accuracy was important for the regular infantryman, focusing more and more on the short and medium range.

            More that they started realizing that crew-served weapons were a more efficient way of delivering fire at long ranges and that personal weapons were important mostly in the sort of confused fighting that doesn’t allow three men to coordinate on setting up a machine gun or mortar. Which is predominantly (but not always) going to be close-quarter battle.

          • John Schilling says:

            I’ve known the term “Minie ball” for decades, and understood the importance of rifling for even longer, but had never actually heard what made the minie ball so important.

            Note to potential time travellers and/or writers of alternate-history fiction: There is nothing about the Minié ball that couldn’t have been invented and produced in quantity as soon as the rifle itself was invented, sometime in the sixteenth century. It’s just that nobody came up with the (in hindsight simple and obvious) idea for another three hundred years.

            If you have reason for some group of plucky underdogs to, say, conquer the world in the seventeenth or eighteenth centuries, this might come in handy.

          • Eric Rall says:

            I’ve heard this often myself, but why? I see how capturing the whole Mississippi was a huge gain for the North, but why was it a huge loss for the South? Were they previously getting a lot of supplies from the Trans-Mississippi, was it just the loss of communications, or was it something else?

            The big strategic value of the Mississippi was the ability to ship things up and down the river, not across it. The South had relatively little in the way of roads and rails and depended heavily on river traffic for travel and freight, and the Mississippi and its tributaries made up a huge portion of this. The Union had already denied the Confederacy most of the value of this by 1863, first by securing the Ohio and Tennessee rivers (which flow into the Mississippi and connect big chunks of territory to the Mississippi river network) early in the war, and later by gradually taking control of most of the Mississippi proper.

            What the surrender of Vicksburg did on top of this was to take an entire Confederate army off the board (about 30k soldiers) and free up the Union armies that had been facing it (about 75k soldiers) to attack the heart of the Deep South, and to give the Union the full use of the Mississippi river network to move armies around and bring up supplies and reinforcements.

          • engleberg says:

            @Lee disliked being called ‘The King of Spades’ when he reinforced Vicksburg early in the war-

            Lee was the best engineering soldier in the US Army, and had a pet engineering battalion for rapidly strengthening weak points in his line, but digging exhausts troops (especially if they die of smelly diseases before antibiotics) and exhausted troops don’t charge well. Meanwhile the North left field fortifications to the enlisted swine; there’s a mention in Sherman’s memoirs.

            @There is nothing about the Minie ball that could not have been invented and produced in quantity as soon as the rifle was invented, about the middle of the sixteenth century-

            Hmm. Light plate body armor was still pretty common through the late seventeenth century, so you’d want a heavy slug that smashed through it. I’d give alternate histories a percussion cap. Simpler than flints, much less wheel locks, and you could still just stick a match in the hole when it failed. And maybe cap and ball revolvers with replaceable cylinders, not long after.
            In the Civil War, revolvers, Minie balls, and lever action carbines among them improved small arms while the improved cannon were mostly too big to be moved fast on land. Parrot guns, Columbiads, great at sea. On land the generals wanted all the Napoleonic-era light cannon they could get and a minimum of multi-ton tubes. Dupuy’s old encyclopedia showed units that started with ten battalions of line troops and one company of skirmishers ending the war with ten battalions of skirmishers and five or so line.

          • cassander says:

            @John Schilling

            Note to potential time travellers and/or writers of alternate-history fiction: There is nothing about the Minié ball that couldn’t have been invented and produced in quantity as soon as the rifle itself was invented, sometime in the sixteenth century.

            I’m not sure about this, for two reasons. One, minie balls, at least sime designs, had an iron plug in the base. That would have made them considerably more expensive to make in quantity, especially since my understanding is that a fair bit of bullet casting in early modern armies was done in the field.

            Two, the 19th century the quality of rifling was much better and more consistent. I’m not sure how much of a difference this would have made performance wise (surely someone has fired minie balls out of a 16th century musket) but I’d bet a fair bit that rifling was, relatively at least, a lot cheaper by the 19th htan in the 16th

          • My understand was that the plug in the bottom of the minie ball turned out to be a mistake. If it was left out the design still worked, presumably because the pressure on the hollow expanded the base into the rifling, and there was no longer the risk of the plug being forced through the ball, leaving a ring of lead blocking the barrel.

            But that’s from a discussion in a book aimed at a popular audience, so I don’t know how reliable.

          • bean says:

            @cassander
            My understanding of the plug issue agrees with David’s. But I also share your thoughts on rifling. It’s relatively difficult to cut, and smoothbores are easier to make. Before the industrial revolution, it may not have been worth it for most soldiers.

          • John Schilling says:

            Hmm. Light plate body armor was still pretty common through the late seventeenth century, so you’d want a heavy slug that smashed through it.

            A .58 caliber Minié ball has almost the same weight as a .75 caliber musket shot (500 vs 600 grains), 30% greater sectional density, and because of the better gas seal about 10% greater muzzle velocity. Better still in terms of retained velocity And if you really need it, the original version’s wrought-iron base plug could be stretched to provide a hard iron tip.

            If you don’t need it, as noted, you leave the iron base plug out altogether. I think even a pure-lead Minié renders any plausible armor obsolete.

          • Nornagest says:

            Hmm. Light plate body armor was still pretty common through the late seventeenth century, so you’d want a heavy slug that smashed through it.

            If it worked that way, shotgun slugs would penetrate armor better than rifle rounds. They certainly don’t for modern body armor; I don’t know if the same’s true for steel but I don’t have any particular reason to think it wouldn’t be.

          • Barry Steakfries says:

            Early ones did not, but by 1861 the Springfield had an effective range of 400 yards. The 1863 model developed during the war could reach 800 – 1000 yards at maximum range, 500 yards effective. Most soldiers didn’t have the marksmanship to exploit all that range, of course,

            Truth, especially since civil war troops generally fired standing or kneeling – I’d be happy to drop a man-sized target at 250m from those positions using a modern rifle w/ iron sights and I was a decent (70th percentile?) marksman (regular US Army infantry).

          • Aapje says:

            @Nornagest

            The general rule is that basic* armor piercing ammunition has to have a strong penetrator core** that doesn’t shatter or flatten easily (usually made of hardened steel, tungsten, or tungsten carbide), so it puts all its energy into a small region of the armor. The typical way that armor works is by spreading the energy over a large area.

            To defeat armor you want to not play into that, which means a relatively small, heavy bullet that keeps it shape and that hits the target at high speed.

            * Shaped charges work differently, but are for bigger projectiles and bigger targets than humans
            ** This has to be jacketed with softer material, because that strong core won’t engage the rifling properly.

          • Protagoras says:

            Alex Zavoluk mentions smokeless powder, which confuses me. Was smokeless powder used in significant quantities in the American Civil War? Wikipedia doesn’t seem to think it was particularly practical before the 1880s (though the first attempts were admittedly in the late 1840s).

          • engleberg says:

            I like alt-Renaissance Minie balls, though I like alt-Renaissance percussion caps better. A useful link between brass and plopping a lead ball onto your powder. Sixteenth-century black powder varied wildly in quality, and so did the tubes -why I don’t expect high-velocity US Civil War rifle performance from anything much earlier, so let’s assume someone from Vasari would be needed. Da Vinci? Too obvious, and his designs were too symmetrical. Michael Angelo: artist of giant buttocks and Minie balls. Cellini: defending Rome with Minie ball artillery. Raphael: too cartoony.

        • cassander says:

          Washington was a better general the Lee because he knew he didn’t have to win his war so much as not lose it

          Washington’s opponents lacked the ability to physically occupy every important part of the US. Lee’s opponents could, and eventually pretty much did, occupy the entire south. Lee had to do more than just not lose, he had to scare the north away, a problem more similar to yamamoto’s than Washington’s.

      • fahertym says:

        You’re alluding to the difference between “tactics” and “strategy,” one of many reasons why the list makes no sense. Tactics is what generals do on the battlefield, strategy is what generals do before they get to the battlefield.

        Alexander the Great was a master tactician. He knew how to line up his infantry, when to reinforce weakened lines, and when to order cavalry charges on the flank. Sherman was a great strategist. He knew how to organize armies to push enemy forces into corners and force surrenders. The two can’t really be compared.

        • cassander says:

          > Tactics is what generals do on the battlefield, strategy is what generals do before they get to the battlefield.

          I like to phrase it, strategy is where you march, tactics is what you do when you get there.

    • jchrieture says:

      Any metric that ranks generals George Catlett Marshall and William Joseph Slim both at “null” will receive respect chiefly from SSC readers who imagine that winning battles is the chief measure of a general, or have never listened to (the audio recording of) Truman’s citation of Marshall’s distinguished service (1945), or have never read Field Marshall Slim’s much-admired strategic monograph Defeat Into Victory (1958), from which the following passages are excerpted:

      Morale is a state of mind. It is that intangible force which will move a whole group of men to give their last ounce to achieve something, without counting the cost to themselves; that makes them feel they are part of something greater than themselves.

      If they are to feel that, their morale must, if it is to endure — and the essence of morale is that it should endure — have certain foundations. These foundations are spiritual, intellectual, and material, and that is the order of their importance.

      Spiritual first, because only spiritual foundations can stand real strain. Next intellectual, because men are swayed by reason as well as feeling. Material last — important, but last — because the very highest kinds of morale are often met when material conditions are lowest.

      … I learnt, too, that one did not need to be an orator to be effective. Two things only were necessary: first to know what you are talking about, and second and most important, to believe it yourself.

      It is characteristic of generals Marshall and Slim, that although both were eminently skilled at tactics, strategy, and logistics, their uniquely outstanding genius was affective … and to this listing of affective military geniuses might credibly be added the name “Abraham Lincoln.”

  5. philosophicguy says:

    Re Sam Altman and Bay Area intellectual oppression, Tim Ferriss cited similar reasons for his recent move from San Francisco to Austin. Here’s part of his explanation when asked in a reddit AMA.

    “5) Silicon Valley also has an insidious infection that is spreading — a peculiar form of McCarthyism (https://en.wikipedia.org/wiki/McCarthyism) masquerading as liberal open-mindedness. I’m as socially liberal as you get, and I find it nauseating how many topics or dissenting opinions are simply out-of-bounds in Silicon Valley. These days, people with real jobs (unlike me) are risking their careers to even challenge collective delusions in SF. Isn’t this supposed to be where people change the world by challenging the consensus reality? By seeing the hidden realities behind the facades? That’s the whole reason I traveled west and started over in the Bay Area. Now, more and more, I feel like it’s a Russian nesting doll of facades — Washington DC with fewer neck ties, where people openly lie to one another out of fear of losing their jobs or being publicly crucified. It’s weird, unsettling, and, frankly, really dangerous. There’s way too much power here for politeness to be sustainable. If no one feels they can say “Hey, I know it makes everyone uncomfortable, but I think there’s a leak in the fuel rods in this nuclear submarine…” we’re headed for big trouble.”

    • Scott Alexander says:

      Austin seems to be becoming the Bay Area Alternative Schelling Point. I wonder if I should start looking into it more, just in case.

      • Scott says:

        If you move here, I promise to organize a “whalecome party” to meet you at the airport with signs bearing terrible puns.

        Incidentally, did you know that on that 100 most-discussed papers page, Katja Grace et al. are listed as the authors of “A Solution of the P vs NP Problem”?

        (The actual author was Norbert Blum, and no, it wasn’t.)

        –Scott Aa.

      • TheZvi says:

        Serious question, what would it take to move the Bay Area Alternative Schelling Point to New York? Willing to apply non-trivial optimization pressure, perhaps capital.

        • Incurian says:

          Reduce the housing costs.

          • Wrong Species says:

            +1

            Moving from the most unaffordable housing to the second most unaffordable housing in the country is hardly an improvement.

        • Scott Alexander says:

          I think the main reason Austin seems better than New York is that if you don’t like the Bay Area, it’s probably because either it’s too crowded, too expensive, or too liberal – and New York is the only other place in the US that might be equally bad on all those axes.

          I’m not saying Austin is better than New York, I’m saying Austin is better than New York if you assume people already hate San Francisco.

          • mupetblast says:

            SF is worse than NYC on this front because it’s NYC politics as if NYC were made up solely of Manhattan.

        • drethelin says:

          Just consider how much someone would have to pay you to move to San Francisco and then multiply that by about fifty.

          That should give you a rough ballpark of how expensive it would be to move the core of the bay area community to a new city, after which I think the rest would follow over time

        • Reasoner says:

          Why New York instead of Austin? New York is supposed to have lots of single women, right? (Good because geeks tend to be male.) Anything else?

      • Incurian says:

        It couldn’t hurt! Also I know this guy there who is always having get togethers with fun activities and good food for the local LWers.

      • vaniver says:

        Excellent! Tho more excellent if this had happened before I had left Austin 😉

      • Alex Zavoluk says:

        Yes, definitely! We’d love to have you. (Conflict of interest note: I’m the organizer for the Austin LW meetup group).

    • pontifex says:

      Is it really “Silicon Valley” that has the infection, or Blue Tribe in general?

      I guess SV concentrates a lot of Blue Tribe people. But so do blue enclaves in red states, like Portland, Seattle, or Austin. From what I can tell, the stuff they’re saying and doing in SF today, they’ll be saying and doing in the other blue enclaves in a year or two. Portlandia is not about Portland.

      Also, I really dislike the equivocation between Silicon Valley and San Francisco. San Francisco is not part of Silicon Valley. SF tends to have the most toxic kind of blue tribe culture, because it’s all 20-somethings fresh out of college. Unless they strike it rich, they move to the south bay or east bay when they want to start a family.

      • Reasoner says:

        In my very limited experience, SF’s toxic culture flows from the colored hair artist contingent, not young people per se.

      • thingdreams says:

        I find that the increasingly common use of the word “toxic” to refer to discussions, attitudes, people etc. basically exemplifies the problem you’re ranting about.

        It’s one of those lazy words that substitutes for actual reasoning and description. I guess it conveys some sense of Extra Importance. And kind of sounds cool.

        It means poisonous. I guess “poisonous” is too blatant somehow. But using “toxic” conveys the same kind of absolute rejection and more importantly the idea of becoming “sick through exposure”, without being as reminiscent of Cold War or other government propaganda, or various literary dystopias.

        • Hyzenthlay says:

          I find that the increasingly common use of the word “toxic” to refer to discussions, attitudes, people etc. basically exemplifies the problem you’re ranting about. It’s one of those lazy words that substitutes for actual reasoning and description.

          This is a good point, and probably a good case for trying not to use the word. I wonder if there’s another way of saying more clearly what people usually mean by this?

          When I think of the word “toxic” in relation to cultures or communities, what I usually mean is:

          1. High levels of anger and negativity
          2. High levels of moral judgment toward those outside the community, or those who don’t share the community’s values
          3. A tendency to react to relatively slight infractions or violations of community norms with shunning, shaming, name-calling, etc.
          4. A tendency to use emotional pressure or emotional manipulation to coerce agreement, rather than trying to persuade people through debate
          5. A pervasive climate of fear and paranoia within the community; tendency for even ingroup members to express anxiety that they will accidentally violate the community’s norms and be punished for it.

          • thingdreams says:

            Well I’ve noticed that word used a lot by people matching most of your list, 2-4 at least, about people who trigger their sensitivities. I guess you can argue with them about it and be mutually toxic together.

          • Hyzenthlay says:

            People tend to respond to negativity with negativity, which is why the cycle is so difficult to escape. The justification is usually something along the lines of, “we can’t play nice with these people because they don’t respond to reason and politeness, we have to use their own dirty tactics against them.” (And people on the other side of whatever the debate is will be using the exact same rationalization.)

            So toxic communities breed equally toxic reactionary communities, etc. etc., with both sides pointing to the other and saying, “Well, they did it first! We’re just acting in self-defense!”

      • Brad says:

        It’s telling that three out of four examples you choose are in the Pacific Northwest. None of them are especially large metro areas. Probably a majority of the “blue tribe” lives in Bos-Wash and SoCal. If not, close to it.

        • Nornagest says:

          Bos-Wash is about 50 million people; the LA metro area is about 12. Adding San Diego gets you to 15, but San Diego is relatively reddish for a large city (maybe because of all the military presence there). Of all of those, I expect no more than 50% are Blue Tribe in Scott’s original sense, and even 50% is probably being generous — Northeastern union workers aren’t Blue, most urban blacks aren’t, most first-language Spanish speakers aren’t. I’m not sure that gets you to a majority, although I don’t have a really solid idea of how big Blue Tribe is.

          • Brad says:

            My logic is this:
            Rural America is around 60 million people out of 320 million and contains essentially no members of the blue tribe. That leaves 260 million Americans in non-rural America. Of that 75-80 million are in BosWash + SoCal. That’s only around 30%. While I’m sure that even Oklahoma City and Colorado Springs have some blue tribe members I think it is reasonable to say that they are more concentrated in larger cities on the coasts than smaller cities in the interior. Metropolitan statistical areas with fewer than 250,000 people (largest of which are places like Yakima, WA; Binghamton, NY; and Tuscaloosa, AL) total around 30 million non-rural people. In terms of the largest ex-BW+SoCal MSAs there’s Chicagoland which is reasonably blue tribe, but beyond that there’s Dallas, Houston, and Miami (total 20 million) none of which is especially heavily blue tribe.

            Maybe 50% + 1 is a bit of a stretch, but I don’t think “or close to it” is so very far off.

          • Gobbobobble says:

            Yeah, when I ran some fast’n’loose numbers I came to mid30s-low40s % – higher than I expected to be in such a small geographic area (Midwest born’n’raised), but didn’t think far off enough from the original claim to be worth mentioning.

            Similarly, I was surprised that Rural America only constituted 20% until I looked it up and saw that places like Topeka and Little Rock count as Urban. Which “huh, yeah, okay, that doesn’t match the usual connotations but is technically correct”

      • bean says:

        But so do blue enclaves in red states, like Portland, Seattle, or Austin.

        How are you getting Portland and the People’s Republic of Seattle as “blue enclaves in red states”? Oregon and Washington are only red states if we look at the county voting map and ignore the fact that some counties have a lot more votes than others.

        • Mark Atwood says:

          I live in Seattle.

          Your objection is meaningless, because it describes *all* the blue city enclaves.

          • bean says:

            Washington and Oregon are blue states. Clinton beat Trump by >10% in both, which I think is probably a fairly good metric. Yes, the hinterland is red, but so is the hinterland in California.

          • Mark Atwood says:

            Washington is not a blue state.

            Seattle is an idiotically blue city, that when it notices there is the rest of Washington State, it’s with annoyance.

          • bean says:

            Washington is not a blue state.

            Seattle is an idiotically blue city, that when it notices there is the rest of Washington State, it’s with annoyance.

            Speaking as someone who spent 5 years as a resident of the dry side of the state, the feeling is mutual, only they don’t give us much opportunity to avoid noticing them. But rural California is pretty red, too. I suspect they feel the same way about LA that I did about Seattle. There’s maybe a bit better balance, but the last time the GOP won Washington was 1984.

          • Mark Atwood says:

            The people on the wet side, but not in King County, likewise have a similar view of Seattle.

            It’s pretty… special. The best two things said about most of the residents are 1) at least they are not San Franciscans, and 2) I find the “Seattle Freeze” to be comfortable and relaxing, not upsetting.

          • Mark Atwood says:

            the GOP won Washington was 1984.

            There would have been a GOP governor here in 2002, but the King County Elections office fell back to the D party playbook of a generation past and from the Chicago machines, and kept “discovering” boxes of “overlooked” ballots in back office filing cabinets, until they had faked enough votes to squeak out a win.

            D partisans either change the subject when they are reminded, or they consider it “payback” for Bush v Gore. I just consider it to be the illumination of everything I need to know about the local D party.

          • Alex Zavoluk says:

            Almost every city is much bluer than the surrounding suburbs and countryside. Maybe not Boston (for some reason the mountainous part of Massachusetts, in the west, is also full of Democrats).

            But the rural Illinois counties outside Chicago are just as red as those across the border in Missouri. Upstate NY is just as red as Kansas. It’s just that NYC and Chicago make the whole state look blue on a presidential election map.

  6. Ashley Yakeley says:

    The Abraj Kudai is still under construction, so it’s not yet the world’s largest hotel.

    • Nornagest says:

      Why does all the new supertall contruction in Saudi Arabia look like it belongs in Disneyland? The stuff in Dubai doesn’t look like that, so it’s not an Arab thing.

  7. Brandon Berg says:

    IMO that IQ study is less contra Turkheimer than a clarification and complement. Turkheimer studied seven-year-old children, whereas the new study is on older adolescents and young adults. The heritability of IQ is much higher in adults than in children, a phenomenon known as the Wilson Effect.

    The reasons for this aren’t entirely clear; my primary hypotheses are that child and adult IQ tests are testing something different, with the former being more trainable, and/or that public school gives students what they need to reach their genetic potential with respect to IQ, such that any head start high-SES children get from more stimulating home environments fades out after 13 years of formal education.

    I’ve long suspected that Turkheimer’s findings were, due to the Wilson effect, not applicable to adult IQ. One study doesn’t prove anything, but it’s nice to see some evidence in support of my hypothesis.

    • bbartlog says:

      There are a couple of other hypotheses I like:

      1) measuring IQ at a particular age reduces the apparent heritability because people grow up at different rates. Maybe everyone ends up with IQ pretty similar to their parents’ (plus randomness) once fully grown, but if we look at 7 year olds, the variation in how far along they are towards this goal state makes the heritability lower. This could be tested if we had full age-dependent data on the parents as well, but generally our data limitations are such that we have to use measured adult IQ for the parents.

      2) motivation and interest/curiosity are heritable and end up being an important contributor to IQ development as people move in to young adulthood. At earlier ages, parental influence dominates innate interest to a greater degree, so that young children can be pushed in the direction of intellectual pursuits that they aren’t really that in to. Once they’re teenagers, this is less effective, and inborn curiosity or other inherited motivation becomes a prerequisite for developing high IQ.

      Mind you, *any* heritable developmental mechanisms that were important but didn’t kick in until the teen years would also work for 2). Maybe some people end up with an excess of hormones and lose all interest in intellectual pursuits for a few years; maybe there’s some signal to the brain that it’s done developing and it’s time to start doing more stuff, and that kicks in earlier for some people than others.

    • Larvatus Prodeo says:

      While the results accord with my priors, I think the study itself is a bit of a mess, for several reasons:

      1) They don’t have zygosity data on twins, so they have to rely on more assumptions than in a typical twin study.

      2) They use test scores averaged across several years, which may increase the prominence of genetic variance which is more stable than environmental variance.

      3) The appendix suggests that there may be moderation of shared environmental effects, with stronger family effects in high-SES families. The method they use is not particularly well-suited for studying that kind of moderation though.

  8. mikewp says:

    I’m so surprised at the difference between Canada and the US, in the Pew survey.
    For two countries that are not all that different (I’ve lived in both), it’s shocking to me that Canada is 55%/24% positive/negative, while the US is 41%/37%.
    Is there more analysis of that survey somewhere?

    • mobile says:

      It’s shocking to me that Canada and the U.S. are not 80%/10% or that the average worldwide result is not 90%/5%. What the hell do you people think the world was like 50 years ago?

    • arabaga says:

      Here is the link to the report: http://www.pewglobal.org/2017/12/05/worldwide-people-divided-on-whether-life-today-is-better-than-in-the-past/

      FYI it is actually 37%/41% for the US, not vice versa.

      I was going to say the election of Trump could have made up the 15-20% difference between Canada and the US, but that doesn’t seem to be true: http://www.people-press.org/2017/08/04/partisan-shifts-in-views-of-the-nation-but-overall-opinions-remain-negative/

      Between August 2016 and July 2017, opinion did indeed shift in the negative direction for about 15-20% of Democrats, but it shifted in the positive direction for 20-25% of Republicans. So the overall change is from 36%/44% to 37%/41%.

      So I don’t know what can explain the difference between Canada and the US. There doesn’t seem to be an explanation in the report.

      • Nornagest says:

        Trump’s election probably didn’t hurt, but it’s more a symptom than a cause. He basically ran on “fuck everything, burn it all down” (this is not meant as an indictment of Republicans; Bernie’s message was kinder and gentler but fundamentally driven by the same motivations), and that’s not a pitch that gets much traction when people feel like they’re doing well.

      • Slocum says:

        My thought is that it’s political tribalism — at this point, in the U.S., few partisans on either side are going to answer that the world is better than the past at a time when their party is out of power. Republican voters used to think the world was going to hell because Obama whereas now Democrats think so because Trump.

    • dndnrsn says:

      Canada and the US are not as similar as one would think. We’ve been diverging in terms of many social attitudes – for example, while Alberta has the reputation of being “Canada’s Texas” there is less support for social conservatism there than in the NE US.

    • notpeerreviewed says:

      I’m speaking from ignorance here, but my sense is that there is far less ideological rancor and partisan dysfunction in Canadian society/politics than in the United States. A disturbingly large number of Americans have come to believe it’s necessarily to literally fight each other in the streets.

      • dndnrsn says:

        I would say that, on the whole, this is true. Politics are less polarized here, a lot more voters are swing voters, the parties are less demographically polarized than in the US. A major problem, I think, is Canadians thinking that things are like they are in the US – for example, it’s not unusual to see left-of-centre Canadians casually assume that the Conservatives are a party reliant on white voters and unable to succeed as far as %s go with other voters; this is true of the Republicans but it is not true of the Conservatives.

  9. Montfort says:

    Japanese chess, and shogi

    are the same thing.

  10. MattH says:

    The amount of words kids hear stuff was always really obvious BS. Kids in big families should here many more words than kids in smaller families. Thus it implies birth-order effects. A second child now hears 3 people speaking instead of 2. There are no birth order affects on achievement, or family size affects.

    • BlindKungFuMaster says:

      I think there is actually a birth order effect on achievement, but it’s in the wrong direction. And there used to be a family size effect on IQ in some draft data, also in the wrong direction (tentatively explained by more infectious diseases).

    • Meister says:

      An even better natural experiment in whether the amount of words kids hear matters: non-deaf children raised by deaf parents. Despite never hearing their parents speak, they show no deficiency in language ability as adults and often have the benefits associated with being bilingual due to using sign-language at home

      • Evan Þ says:

        Confounding factor: they grow up seeing lots of sign language, even if they don’t hear oral language at home. Sign language very plausibly hooks into the same parts of the brain.

      • Ozy Frantz says:

        To add to Evan’s point, Deaf children raised by non-Deaf parents do experience serious language delays, while Deaf children raised by Deaf parents do not, presumably because of the early exposure to sign.

    • Chalid says:

      A parent is likely to talk way less *with* their second and successive babies. Perhaps more words are being spoken nearby, but no one thinks that’s all that matters – if it was you could raise your baby by podcast.

  11. daneelssoul says:

    About the net neutrality survey, I feel like a lot of the reason that it was skewed was just that the pro-net neutrality arguments were more complicated than the anti-net neutrality arguments meaning that the proponents of such arguments were less confident in them. If you read lots of the comments by people who listed themselves as “uncertain”, they were along the lines of here’s a potential bad effect of removing net neutrality, but I’m not sure what it will mean, while the agree people were just saying something along the lines of removing it allows for more accurate pricing, and usually make no attempt to even address the issues brought up by the other side.

    • Maggie says:

      I came here to basically say this. Reading the comments, I was surprised all the people raising concerns about vertical price discrimination were “uncertain”. Vertical price discrimination is exactly what “the masses” are worried about.

    • poignardazur says:

      Isn’t that an Isolated Demand for Rigor?

      I mean, seriously, how many people shouting “No fast lanes!!!” on forums and in youtube videos know what Google Global Cache is and why it’s important?

  12. Campion says:

    Re: Polk, in addition to the argument that the Mexican War was a sketchily-justified land grab for more slave states, it was also Grant’s opinion that his administration was actively hindering the war effort at times because it didn’t want Scott or Taylor to get too high a reputation and run for president as a Whig (which Taylor ended up doing successfully anyway).
    With regards to the ranking of generals, there’s more problems with the methodology than just what you pointed out. He’s determining the odds of one side winning the battle based purely on number of soldiers (to his credit, he admits in the article that further research would need to account for other factors). Based on this, his rating of Lee at Fredericksburg gives him a WAR of over .50, since Lee was outnumbered, even though any adequate model would expect a Confederate victory there to be more likely than a Union one. Looking at his data for Lee also shows some other problems with his model, particularly the use of a naive win/loss/tie model and the way he treats all battles equally. The most ridiculous display of this is how Lee gets awarded point for the Seven Days Campaign: he is counted as winning Gaines’ Mill, tying Glendale, and losing Malvern Hill, Oak Grove, and Beaver Dam Creek, giving him a 1-1-3 record and a substantial negative contribution to his WAR. Looking at this, one would think the Seven Days were a disaster for Lee’s army, instead of a decisive success that drove back and demoralized the Union Army and saved Richmond from capture. Aside from the question of how to assign win/loss/ties for some of the more unclear or indecisive battles, he has made multiple unambiguous errors: he has Gergovia listed as a victory for Caesar, when it was an unambiguous victory for Vercingetorix, and Cold Harbor listed as a Union victory when it should be a Confederate one. The fact that I could find two egregious errors in by only looking briefly at 5 or 6 of his data sets suggests that there are likely at least several more in there somewhere.
    More broadly, the entire method is limited by the fact that it’s only looking at tactical skill, which is only part of a general’s skill set. Hannibal gets one of the highest ratings from the model, as he usually does when you’re looking purely at tactical skill, but this misses his major weakness, that he was never really able to capitalize on his victories: he smashed Roman armies on the battlefield repeatedly, but it never really got him much closer to capturing Rome or winning the war. The failure to consider strategic significance is related to the obvious problem you noted of not weighting for numbers of battles: a general who wins a war quickly and decisively in a few battles would be given a lower rating than a general who wins battles repeatedly but never gets anywhere because of it. Lysander, who I’d rate as one of the greatest generals of history, only won two battles, because that was all he needed to win to win a war that had been going on for a generation. As Sun Tzu said, “To win one hundred victories in one hundred battles is not the acme of skill. To subdue the enemy without fighting is the acme of skill.”

    • Vamair says:

      Even as it’s counted there, Alexander Suvorov won 63 battles and lost none, so his count should be a bit higher than Napoleon’s.

    • The original Mr. X says:

      To be fair to Hannibal, Rome’s commitment to winning the war was really quite extraordinary. (I remember my old history lecturer telling me that, at the height of the Second Punic War, Rome had a higher percentage of its male population under arms than any of the western powers did in WW1.) Whilst in retrospect we know that Hannibal’s strategy didn’t work, I’m not sure we can reasonably blame him for failing to realise just how crazy the Romans were, when basically any other ancient state would have folded after Cannae at the latest.

      • Campion says:

        Yeah, I am being a little hard on Hannibal, given how tough the Romans were and how little support he was getting from his government, and I can see the argument that he did about as good a job as he could given his situation. But I still think it’s fair to say he was a greater tactical genius than he was a strategist, and would therefore be overrated by a purely tactical rating, even though he’d still rate pretty high on a rating that looked at a wider range of abilities.

        • DeWitt says:

          I do think you’re being rather harsh on Hannibal, yes. If he were to ever try capturing Rome, there’d be two problems he’d run into, things he was very well-aware of: an army not very well-suited to storming the city(Rome couldn’t well be besieged, given that it had a navy where Carthage did not), and Rome having an immense amount of allies in the surrounding area that could very well come in with a relief force.

          Hannibal, then, went on to try and have southern Italy defect against Rome, which he was actually rather succesful at. And really, what else was he to do? What would a greater strategic genius have done?

          Finally, it’s not clear to me he’s quite so much of a tactical genius either, given that he lost Zama, which seems to have been preventable in hindsight. Even so, Hannibal is a highly impressive example of a general, and made do much, much better than he had any right to do.

          • cassander says:

            I think it’s also fair to say that hannibal’s strategy would have worked on basically any ancient empire besides rome. Most ancient empires had networks of “allies” that tended to bolt whenever they got half a chance, only the roman allies were actually loyal enough to stay loyal when the chips were down, and would eventually even fight a war to become more closely integrated with rome itself. No one else managed to build such loyalty.

          • The original Mr. X says:

            Finally, it’s not clear to me he’s quite so much of a tactical genius either, given that he lost Zama, which seems to have been preventable in hindsight.

            To be fair, that is IIRC the only major battle he lost, and Scipio was a tactical genius as well.

          • Protagoras says:

            Also, as I understand his situation at Zama, he had inferior cavalry (for the first time ever in his career) and inexperienced troops. He may have correctly judged that a lot of ambitious tactics he might have tried would have had no hope of success with the forces available to him, and so employed tactics within the capabilities of his available men (tactics which predictably but perhaps unavoidably proved insufficient against Scipio’s veterans and superior cavalry).

          • DeWitt says:

            Hannibal did, basically, get screwed out of the cavalry he should’ve received at Zama. Even so, the elephants he’d received were soundly defeated by some highly clever maneuvering of Scipio’s, something which could well have been avoided by someone deploying them in a different manner.

  13. ManyCookies says:

    @Sam Altman

    Someone on the subreddit brought up Tyler Cowen’s reply:

    Nerd that I am, I am immediately reminded of the theory of price indices. If you go to a new country with the same goods and prices as your home town, you won’t buy very much. Alternatively, if your port of call has radically remixed relative prices, you will do lots of shopping and go home pretty happy.

    And so it runs with shadow prices for speech, including rights to say things and to ask questions. Whatever you are free to say in America, you have said many times already, and the marginal value of exercising that freedom yet again doesn’t seem so high. But you show up in China, and wow, your pent-up urges are not forbidden topics any more. Just do be careful with your mentions of Uncle Xi, Taiwan, Tibet, Uighur terrorists, and disappearing generals. That said, in downtown Berkeley you can speculate rather freely on whether China will someday end up as a Christian nation, and hardly anybody will be offended.

    For this reason, where we live typically seems especially unfree when it comes to speech. And when I am in China, I usually have so, so many new dishes I want to sample, including chestnuts and pumpkin.

    The litmus test for a group’s free speech isn’t how their members treat outside controversies, which might not even be controversies for them (“and thus you gain no merit by your tolerance!”). The litmus test is how they treat their own controversies, be it political or tribal or whatever, and China’s responses to out-of-line speech can be a tad more extreme than the Bay Area’s.

    • Anatoly says:

      On “ẗheir own controversies”, see also Steve Hsu’s opinion:

      “Actually, Sam understates the case. In Beijing no one would hesitate, in private conversation, to question the communist party or some action by Xi Jinping. There are basically no restrictions on what you can discuss in China — people are “reality based”! You only get in trouble if you *widely broadcast* anti-government views through social media or other platform.

      The US is a different story entirely. You can easily get an angry emotional response from someone who is supposedly highly educated and rational by asking simple questions like “How reliable are IPCC climate projections? Can we really model such complex phenomena?” or “Do students admitted through preference with SAT scores hundreds of points below the rest of their class have a good shot at succeeding in challenging majors?” :-(“

      • Wrong Species says:

        “Get in trouble” is one way of putting it. In China, you go to jail. In America, you get booted off twitter.

        • Calvin says:

          You miss the point – OP’s argument is that if you raise a controversy in private conversation in China nothing happens, but if you raise a controversy in private conversation in the Bay Area you lose significant social status.

          • Wrong Species says:

            I understand that. My point is that is glosses over can happen in China. China may in some sense allow opinions that are socially unacceptable in the US but that doesnt makes it better.

          • Conrad Honcho says:

            Except the average person is not broadcasting their opinion on controversial topics to the masses, in China or the US, but they are holding casual conversations on a regular basis.

          • Wrong Species says:

            What exactly do you think Facebook and Twitter are?

          • Conrad Honcho says:

            Places where average people, if they use them at all, share pictures of their breakfasts.

      • ManyCookies says:

        In Beijing no one would hesitate, in private conversation, to question the communist party or some action by Xi Jinping.

        But is there meaningful disagreement going on in these conversations, or is everyone on the same page about whether the party screwed up or not? Would they be willing to have the same conversation in front of a strong party supporter or a party official (the closest equivalent to a SJW in this context)?

        And why the heck are we limiting the scope to private personal conversations, online and open freedom of speech are pretty freaking important – we’re on an online forum right now! Sure the Bay Area’s epistemic standards are just like China’s… when you conveniently hedge around the parts where China’s is way way worse. I’m not trying to excuse any behavior, but this is seriously “Trump = Hitler” levels of tenuous comparison.

      • Enkidum says:

        “In Beijing no one would hesitate, in private conversation, to question the communist party or some action by Xi Jinping.”

        This is… not really true? How many mainland Chinese do you know who would, for example, question the validity of their current South China Sea claims? You get just as hostile a response as you would in America, despite the deficiency of the Chinese position being orders of magnitude more obvious than the analogous examples you give for the US.

        • powerfuller says:

          @Enkidum

          Anecdata, but most of the younger people I talked to in Mainland China were vaguely critical of the government but fairly pessimistic about things changing, much like Americans in conversation. A lot expressed a general dislike of the Communist party/leadership. They were most critical of the environmental degradation, somewhat critical of internet censorship, and less critical of foreign policy, Tibet, etc. I didn’t talk to many older folks about politics, but the biggest criticism I got from some was that China wasn’t Maoist enough anymore.

        • MoebiusStreet says:

          How many mainland Chinese do you know who would, for example, question the validity of their current South China Sea claims? You get just as hostile a response as you would in America

          I’ve had exactly this conversation with a few different PRC natives. I think that they’re very likely to side with their government here. But I don’t have the feeling that it’s cost me any social status to broach the topic. And that is the key difference.

      • John Schilling says:

        Actually, Sam understates the case. In Beijing no one would hesitate, in private conversation, to question the communist party or some action by Xi Jinping.

        Is it safe for a Beijing native to question the CCP’s decision to gun down two thousand or so Chinese citizens in and about Tiananmen Square on 4 June 1989? Is it, for that matter, safe to acknowledge that the date of “4 June” exists and might be significant, without using deniable circumlocutions?

        I must have missed that.

        • MoebiusStreet says:

          Yes, absolutely, these areas of discussion are acceptable in private conversation. I’ve been part of a number of such conversations, including Party members and at least one with a current employee of the equivalent of PRC’s FBI. In a personal situation there are absolutely no legal problems, and on this topic there’s not even any argument.

          I think that some of the folks responding on this topic are conflating official legal response (which only occurs when publicly broadcasting a statement, on the air or on social media) with the far more easy-going face-to-face conversation, where there seems to be much less of sacred cow problem in China.

      • Zakharov says:

        It seems more likely that we just don’t know what you can’t bring up in polite conversation in China – and of course, that would vary depending on who you were talking to.

  14. Quadratic says:

    Clearly the ISPs think that Net Neutrality is going to make them a lot more money, or they wouldn’t have lobbied for it so hard. The main concerns from my understanding is that

    1. ISPs are basically monopolies for lots of the country. Net Neutrality doesn’t really help competition.

    2. Startups won’t be able to compete. Something like youtube or netflix wouldn’t be able to afford these fast lanes and couldn’t be made, and since this is the future, these startups would be things that are even better than youtube and netflix.

    3. It sets a nasty precedent that the internet could become more closed and monitored like what’s happening in europe and china.

    • Scott Alexander says:

      What you say makes sense, but I like Buck’s point – predict what’s going to happen, in decideable terms. And then preferably offer to bet on it.

      • anonymousskimmer says:

        I doubt any nation will nuke any other, no matter how much nuclear weapons proliferate. I’m hypothetically willing to make a bet on that.

        I also doubt that any nation would willingly give (non-state actor) radicals nuclear weapons. I’m likewise hypothetically willing to bet on that.

        I doubt the chances are high that (non-state actor) radicals could get their hands on nuclear warheads. But these chances increase the more nations have more nuclear weapons, and this increased, albeit low, probability is not something I like.

    • Squirrel of Doom says:

      Both YouTube and Netflix were started long before Net Neutrality was enacted in 2015.

      • PSJ says:

        We have had net neutrality regulation in one form or another continuously since 1996. The “2015” talking point is a lie told in order to make it seem like we didn’t need it before so why should we need it now. Looking at the IGM poll, the Agrees seem to have much less idea on what net neutrality entails than the Uncertains or Disagrees. It’s hard to make specific predictions on a short timescale since the ISPs will be acting in a way as to specifically minimize the risk of being regulated in the future. If someone wants to propose a measure of skew of the Alexa top 100 in terms of traffic distribution, and can show a reasonable choice of trend curve through that data, I am willing to bet that the 2019 data is more skewed towards the top few websites than predicted by a curve fit from data from 2000-2018. But it’s a complicated bet and I have already spent wayyyy too much time talking about this topic on the internet, so I’m not that pressed to do it.

        Sources:
        Original Act of Congress
        Covers the period from 1996 to 2005
        Covers the period between 2005 and 2015

    • Startups won’t be able to compete.

      I think you have it backwards. It’s in the interest of the ISP to price discriminate in favor of startups so as not to block them, because the successful startup ends up increasing the value of internet connection to the ISP’s customers and that’s what the ISP is selling.

      • Gobbobobble says:

        It is not the only thing the ISP is selling. The big players all bundle telephone/cable/internet together. Which is why Netflix and YouTube are the main scapegoats now and VoIP services have been in the past.

    • poignardazur says:

      2. Startups won’t be able to compete. Something like youtube or netflix wouldn’t be able to afford these fast lanes and couldn’t be made, and since this is the future, these startups would be things that are even better than youtube and netflix.

      Even assuming that ISPs drive prices higher for fast lanes, startups will still be able to compete. They don’t have to buy directly from ISPs, they can buy from CDNs who buy fast lane access in bulk from ISPs, and redistribute it in detail (eg Cloudflare and co).

      Actually, I think this is already happening, or will happen soon. CDNs want their data delivered as fast as possible, and that means paying to store cache proxies directly in ISP distribution networks.

      And that’s for big download or streaming startups. Most startups don’t need that much broadband.

  15. Ketil says:

    The bitter/toxic research seems silly to me – what poisons are known today have little bearing on what types of poisons would affect the evolution of humankind. This is the trap of averages, where little to no observed effect in the average is used to rule out any effect at all.

    • Aapje says:

      We also don’t eat random compounds, but plants and such. So a more interesting study would take a sample of plants that are present in nature and see if their bitterness correlates to their toxicity.

      • wintermute92 says:

        My first thought: nightshades are common, with appealing traits like “tomatoes”. They’re also poisonous from high alkaloid content, which we perceive as intensely bitter.

        It would only take one or two families of bitter-poison plants to justify a response like this.

        • Aapje says:

          The upside of eating bitter plants may also be less if they are less nutritious on average. If a MealSquare bar and a equally sized puffed rice cake have the same chance of being somewhat poisonous, I’d much prefer to eat the MealSquare. I’d have to eat way more rice cakes to ingest the same calories, so I would then ingest way more poison in total.

      • Loris says:

        That’s kind of what I was thinking – particularly in relation to fruit.

        It’s called Müllerian mimicry

        Plants want animals to eat their fruit and spread the seeds – but not until the seeds are ready. Bitterness would be a way of telling the animal to leave the unripe fruit alone. A prediction of this model is that ripe fruit would have maximally fertile seeds and would be significantly nutritionally better for the animal. (If there’s a delay between fertility and ripening that wouldn’t be a problem.)

  16. Le Maistre Chat says:

    On the ASD/gender dysphoria link: “between 60% and 90% of children diagnosed are no longer gender dysphoric 10 years after diagnosis.”
    That’s… really interesting. It undermines the Narrative and matches my own psychiatric history.

    • Lillian says:

      This is absolutely true, but at the same time, i understand that the older the person, the less likely the dysphoria is to go away. So an 8 year old who insists she is a boy is most likely to develop into a fairly normal young woman by the time she’s 18. On the other hand an 18 year old who insists she’s a guy is very likely a guy, and a 28 year old who insist she’s a man will continue insisting it for the rest of her life. That’s going off memory, though.

      From the abstract of one of the linked papers: “Both boys and girls in the persistence group were more extremely cross-gendered in behavior and feelings and were more likely to fulfill gender identity disorder (GID) criteria in childhood than the children in the other two groups. At follow-up, nearly all male and female participants in the persistence group reported having a homosexual or bisexual sexual orientation.”

      And from the other: ” The rates of GID persistence and bisexual/homosexual sexual orientation were substantially higher than base rates in the general female population derived from epidemiological or survey studies. There was some evidence of a “dosage” effect, with girls who were more cross-sex typed in their childhood behavior more likely to be gender dysphoric at follow-up and more likely to have been classified as bisexual/homosexual in behavior (but not in fantasy).”

      This illustrates the importance of psychological screening before starting hormone replacement therapy, and also the need to be very reluctant to start any irreversible treatment before the individual is an adult.

      • Le Maistre Chat says:

        This is absolutely true, but at the same time, i understand that the older the person, the less likely the dysphoria is to go away. So an 8 year old who insists she is a boy is most likely to develop into a fairly normal young woman by the time she’s 18. On the other hand an 18 year old who insists she’s a guy is very likely a guy, and a 28 year old who insist she’s a man will continue insisting it for the rest of her life.

        Plausible. I first experienced gender dysphoria at 14 and got over it.
        I was fairly well-adjusted through elementary age, able to make a typical number of friends even when my family moved. Though in retrospect there was evidence of something wrong, I didn’t get psychiatric evaluation until a germ phobia was noticed during puberty, which was diagnosed as OCD. The SSRI Celexa/Citalopram made the anxiety manageable, but I wasn’t able to fully function until having a religious experience at age 14. As a freshman, I was now finding it hard to make friends. Being >6 feet tall with A cups, unpopular, with nerd interests and getting the stereotypical “you’re not a real girl” message from guy nerds, dysphoria set in…
        I don’t remember how long after this I got the Aspergers diagnosis.

        This illustrates the importance of psychological screening before starting hormone replacement therapy, and also the need to be very reluctant to start any irreversible treatment before the individual is an adult.

        YES. It creeps me out how parents whose child displays even mild gender dysphoria are being pressured to accept their permanent transgender essence to remain Blue people in good standing.

        • Ozy Frantz says:

          In experiencing gender dysphoria as a teenager and getting over it, you are very unusual (although we don’t have the studies to say precisely how unusual). Even medical professionals who accept the persistence results at face value believe that gender dysphoria in adolescence is generally permanent. In part for this reason, the standard of care is not to give puberty blockers until the minor is well into Tanner stage 2. Of course, hormones should not be begun until the minor is mature and certain of their decision, usually after the age of 18.

          The persistence results are somewhat complicated, because at the time gender dysphoria in children was a particularly broad diagnosis, particularly in children assigned male at birth. A boy who played with dolls and had only female friends could be diagnosed with gender dysphoria in children, even if he insisted repeatedly and unambiguously that he was male. From a gender-identity-based model (as opposed to the Blanchardian model), there is no reason to expect that a boy with a clearly articulated male gender identity would grow up to be trans, regardless of his taste in toys or friends. It is true that children who insist they are a different gender are far, far less likely to desist than children who do not. Unfortunately, we do not have the data for persistence rates for a narrower definition of gender dysphoria. In the meantime, I think parents should be aware that desistance is common and make it clear to their children that desistance is a fine and acceptable outcome, but should also make the right decisions for their children’s mental health, which may include social transition.

          That said, I’m glad you resolved your gender issues in a way that worked for you.

          • Le Maistre Chat says:

            In experiencing gender dysphoria as a teenager and getting over it, you are very unusual (although we don’t have the studies to say precisely how unusual).

            Maybe. I’m people-oriented yet have what’s now an ASD diagnosis, apparently in the 99.9th percentile for height, and was told after taking an adult IQ test that my verbal score (but not visual-spatial) was beyond their ability to measure.
            I’m not trying to be an extra WEIRD person, honest.

          • chridd says:

            “The persistence results are somewhat complicated, because at the time gender dysphoria in children was a particularly broad diagnosis, particularly in children assigned male at birth.”

            My own suspicion is that the gender dysphoria criteria for children isn’t too broad or too narrow, but rather measuring the wrong thing—that is, whether a child likes dolls or what gender their friends are or what clothes they wear (which are the sort of things the criteria I’ve seen talk about) might not be related at all to whether they’re gender dysphoric, or only weakly correlated, and so there are both a lot of false positives and a lot of false negatives.

            It may be that actual gender dysphoria in children looks a lot like autism, and not (as people expect) like gender non-conformance.

          • Le Maistre Chat says:

            @chridd: Yeah, and those factors really need to be disentangled before making parents think they’re only good people if they help a diagnosed child become transsexual with puberty blockers and more extreme medical intervention.
            I’m not even saying that’s accepted practice in 2017, but the DSM is sensitive to Leftist pressure.

            Tangent: I thought it was funny when I got into weightlifting in college that female bodybuilders are on male hormone treatments, so there’s no objective difference between them and a serious transman.
            So I do natural powerlifting.

          • a reader says:

            In experiencing gender dysphoria as a teenager and getting over it, you are very unusual (although we don’t have the studies to say precisely how unusual). Even medical professionals who accept the persistence results at face value believe that gender dysphoria in adolescence is generally permanent.

            Actually there is a survey of FTM detransitioners and according to it, the average age of coming out/transition was 17 years and the average age of detransition was 21 years:

            https://www.surveymonkey.com/results/SM-RCWYCNGM/ (3rd graph)

          • Ozy Frantz says:

            a reader: That is a study of detransitioners and thus does not say anything one way or the other about whether adolescent detransitioners are very unusual. However, even in this survey, 91% of detransitioners state they still experience gender dysphoria.

        • Lillian says:

          Note that like Ozy i believe dysphoric youth should be allowed to socially transition, while also making it abundantly clear that the route back is always open. If a seven year old insists on being a she, then she’s a she, and if at nine years old wants to try being a boy again, then he’s a boy, then if a year later wants to go back to being a girl, then well, want the old girl name or want to try a new one? Social transition is reversible, so kids should feel free to have at it as many times as they think they need.

          Once puberty starts though, it gets more complicated since the possibility of puberty blockers enter the picture. Generally i would propose to only give those to adolescents who both have persisted in insisting they are a gender opposite their sex, and have gotten a taste of their biologically determined puberty and it’s making their dysphoria worse instead of better. The blockers are fairly safe side-effect wise and are reversible by simply going off them, so they can be on them until they hit the late teens, and then if they’re still dysphoric then they can start discussing cross-sex hormones. As i understand it, that is in fact the current standards of care, and it seems to be me like the best approach in terms of minimizing harm.

        • Hyzenthlay says:

          I’m another person who found my gender dysphoria getting milder as I got older. It probably peaked in my teens and early twenties and now it feels like it’s barely an issue for me. I still don’t really identify as female that strongly (and don’t even fully know what it means to “feel” male or female) but my desire to have male anatomy has mostly gone away.

          One thing that helped me, weirdly, was skipping periods for several years using hormonal BC. Having a sense of control over my own biology made that biology feel less like a scary alien thing that had me under its power.

          • Le Maistre Chat says:

            I’m glad to hear you grew out of it.
            I have a hypothesis that, for most of history, most gender dysphoria was females wanting to live as men. I’m agnostic on whether that was typically accompanied by a desire to have male anatomy and whether they tended towards being straight, hitting on women, or celibacy.
            You have an interesting point with hormonal BC. I think periods, pregnancy and the much-reduced role of the latter in many modern women’s lives could interact with a brain chemistry issue to cause confusion.

          • Aapje says:

            The female body has a more fluctuating hormonal balance, which might cause more unhappiness with the body than what men experience.

        • a reader says:

          I think giving puberty blockers to 11 years old children (“Tanner stage 2”) is a very bad idea. Afaik, if very young boys (before school) manifest gender nonconformity (they want dolls, dresses etc.), it’s because their brains are quite feminine, because they weren’t masculinised enough in utero (probably not enough hormones). Natural puberty with its hormones boost is probably just what many of them need to naturally solve their problem – and it produces that massive desistance because it somewhat masculinises the brain, so that now many of them feel male enough and can live as cis (homosexuals mostly), only a minority of them remain feminine and need transition.

          • chridd says:

            What I said above applies here, too: I think that the large amounts of “desistance” is actually because being trans is a completely separate thing from being masculine or feminine. Trans women don’t transition because they’re super-feminine, they transition because part of their brain expects a female body. I suspect the people who “desist” were never trans in the first place, and only thought to be trans because there’s a misconception that AMAB people who like dolls, wear dresses, and have female friends are trans (a misconception that’s codified in the DSM). Meanwhile, people who realize they’re trans in adulthood were always trans, but no one realized sooner, because there’s a misconception that trans girls like dolls, wear dresses, and have female friends.

            If I’m right, then going through puberty doesn’t really affect gender identity, and puberty “masculinizing” the brain isn’t necessary to explain the change in gender identity, because there was no change in the first place, only a misconception. And gender nonconformity isn’t really a problem, so there isn’t really a problem for puberty to solve here. That said, feminine cis boys shouldn’t be given puberty blockers, because they’re cis and their brains expect them to go through male puberty.

            Really what should be done is figuring out what the actual signs of children being trans are.

          • a reader says:

            Afaik there are 2 kinds of MTF trans:

            Type 1: Early onset transsexuals
            Called “homosexual transsexuals” by Blanchard & Bailey.
            Feminine from early childhood, attracted to men in adulthood.

            Type 2: Late onset transsexuals
            Called “autogynephile transsexuals” by Blanchard & Bailey.
            Not particularly feminine but frequently nerdy in childhood.
            Attracted to women or to both sexes as adults; sometimes asexual.

            My former comment applied only to the first type, the ones with early onset. The second type, due to their late onset, don’t receive puberty blockers, because they come out only after puberty is over.

            I suppose autism is linked to Type 2 (the nerdy ones).

            More about the two types in this book (very interesting, although controversial):

            The Man Who Would Be Queen by J. Michael Bailey

            I suppose an address like faculty.wcas.northwestern.edu means the PDF is legal, not pirated.

          • chridd says:

            I think that type 1 is what people think trans people look like, and type 2 is what trans people are actually like (typically), so I kind of suspect that the (apparent) existence of type 1 trans people is something self-fulfilling prophecy-ish: either some of the people who are mistakenly thought to be trans (due to the misconceptions I mentioned above) transition anyways, or there’s some selection bias where trans people who are coincidentally stereotypically feminine realize and transition early, or type 1 trans people are type 2 people who interpret their feelings through a type 1 lens. (If not that, I still suspect that type 2 is the default manifestation of gender dysphoria, and type 1 is something else.)

            I think type 2 trans people aren’t actually late-onset, it’s just that people don’t notice the signs because they’re looking for the wrong thing.

            (Also, if you’re talking about the same distinction as tailcalled, then type 2’s are sometimes straight (attracted to their assumed gender at birth)[1]. Also, I’m not really convinced there are two types; it’s not something I’ve personally seen, and here and Ozy’s blog are the only places I frequent where I’ve seen people actually make that claim.)

          • Nornagest says:

            I think that type 1 is what people think trans people look like, and type 2 is what trans people are actually like (typically)

            I think there might be some availability bias here. All the transwomen I’m aware of knowing more closely resemble Blanchard & Bailey’s “autogynephile” type, but if B&B are right about there being a dichotomy, then I’d never notice the “homosexual” type — they apparently tend to transition early, so I’d only have known one before her transition if I’d happened to grow up with her (which is unlikely), and by all accounts they tend to pass very well.

            B&B would have enjoyed a larger sample size, and they wouldn’t have been as reliant on luck. So they’d be in a position to actually notice. They could still be wrong, of course — I just don’t think my perceptions are good evidence against the idea.

          • Le Maistre Chat says:

            The autogynophile “lesbian” type is extra WEIRD. I know that wanting to pass as the other sex is a weird/rare behavior, but it’s not hard to model why it pops up across cultures, either as a brain disorder that makes the bearer passionate about passing or female freedom-seeking behavior in a patriarchal society.
            This type, though, makes minimal effort to pass and leverages SJ to impose the burden on everyone else they see. It’s a blatantly domineering behavior, and because it’s linked to politics, it feels much worse than just, like, having a domineering boss and more like 2+2=5 in 1984.
            I think I was only suspected by one girl when in menswear, when I forgot to cut my nails before meeting people I knew from the internet. It never would have occurred to me that I WASN’T the one at fault.

          • chridd says:

            > I think there might be some availability bias here […]

            I don’t see how that relates to what I said? I’m not saying that feminine AMAB people who transition don’t exist; I’m saying that their femininity isn’t solely caused by them being trans. I’m saying that being trans by itself looks more like type 2, and type 1 is caused by either a combination of being trans and something else, or is not caused by being trans at all.

            In any case, since the test “Are they feminine?” as an attempt to find out whether they’ll be a trans woman when they grow up seems to have both a high false positive rate (large desistance rates) and a high false negative rate (plenty of trans women aren’t feminine during childhood), it seems likely that this test is simply wrong (and that’s a simpler explanation than that people who pass the test are a different type of trans than those who don’t). And just because a trans woman isn’t feminine in childhood doesn’t mean that she wasn’t trans then; there may be I’m fairly certain there are signs of being trans that don’t have to do with being masculine or feminine.

    • anonymousskimmer says:

      This seems quite similar to the fact of remission likelihoods in childhood tic disorders, childhood OCD, and presumably many other psychological disorders.

      https://www.ncbi.nlm.nih.gov/pubmed/23428237

  17. mondsemmel says:

    Re: EA donations in general, and MIRI’s fundraiser in particular: There’s a donation matching challenge until December 31 for 10 EA charities (including MIRI, and several GiveWell-recommended charities). See here: https://2017charitydrive.com/

  18. suntzuanime says:

    Some rare good news: the grad student waiver tax will not be in the final tax bill.

    Some rare good news: the hated tyrant has promised to leave our beloved tulip subsidies intact.

    You’ve read the same articles as I have about the PhD overproduction crisis. When you have an overproduction crisis it seems like maybe you should stop giving preferential tax treatment to the thing being overproduced? It would be a start, anyhow.

    • There are many policies which, if rolled in slowly or set to take effect with enough warning, could be very beneficial but at the same time if imposed by surprise would be disastrous.

    • peacetreefrog says:

      UChicago economist John Cochrane on this:

      Moreover, nobody stopped to ask, why do universities pretend to charge tuition, and then waive it?Just how hard would it be for universities to adapt to the tax by not charging tuition in the first place? Why is tuition like medical bills, with a phoney-baloney list price and then everyone gets a huge discount of one sort or another?

      Carlos Carvalho and Richard Lowery figured out the answer to this question: Many graduate students, especially in the sciences, get funding from the federal government, and to a lesser extent from private sources. The university charges “tuition” to the grant. So “tuition” is just a way for universities to tax federal grants, and to transfer money that would otherwise flow to students, departments, and research instead to central budgets and general university operations.

      https://johnhcochrane.blogspot.in/2017/12/universities-and-taxes.html#more

    • Slocum says:

      There was such an easy workaround for that one (offer a job and a scholarship separately rather than tying them together) that I never worried about it. I suspect that may be why it was dropped — bad publicity but no actual effect.

    • Gobbobobble says:

      20k + free tuition

    • phil says:

      Who’s holding whom hostage?

      ——–

      Since the language is interestingly similar, from the above Cochrane post:

      “though it clearly is income — if, in return for work, your employer waives rent in a company-provided apartment, that is taxable income — and though it seems to me that it would have taken only minor tax finesse to avoid it and keep the federal money flowing, the university lobby seems to have preserved our exemption from the tax bill. Having a sympathetic hostage is a great lobbying tactic. “Look at the secretaries and administrators we’ll have to fire” would have played less well.”

    • suntzuanime says:

      I’m literally, *literally* a grad student with a stipend and a tuition waver who would be paying the tax. Don’t try your ad hominems on me you smarmy piece of shit.

      • Sniffnoy says:

        Dude, suntzuanime had a particular reason for that insult, piling on afterward like that is… really not helpful.

      • Kevin C. says:

        @Sniffnoy

        Except this is far from the only time he has displayed this attitude, like when he anointed himself arbiter of the Overton Window and began pronouncing certain views and the commenters who held them as needing to be banished completely from our society. I’ve only held back from responding because my comments would have likely been intemperate enough to see me banned.

      • Montfort says:

        Knock it off, Kevin. Commenters breaking the no-ad-hominem norm isn’t a free pass for third parties to drop in and get their licks in.

    • quanta413 says:

      I’m in the same boat as suntzuanime. Ignore any personal pain it would have caused for the moment. The tax would’ve been a sudden shock to a large sector and it would’ve blown for a few years for a some people who weren’t at fault, but I find it highly unlikely a movement towards many fewer Ph.D.s being produced will happen in any way besides “over the screaming objections of universities and professors at the cost of some collateral damage”. There are also good economic reasons to believe the tax exemption is poor policy as can be seen in the link to cochrane’s blog (see peacetreefrog’s post). And most grad students are young people who can still change their life relatively easily and who in the future will be upper middle class or higher anyways, so as far as some group of people getting whacked by a swift adjustment for the greater good of a less stupid tax system, it’s one of the better possibilities.

      And since I’m also a grad student, I also know that when you claim it’s the norm for grad students to work 80 hours of week, you’re full of it. I mean, maybe you force your graduate students to work 80 hours a week (do you?), but that’s not the norm. And I also know that a lot of grad students spend most of their work time doing things a less educated chunk of the population could do despite grad students being trained far past that. And then grad students finish grad school and ~2/3 of them end up in jobs similar to the sort of work they could have done with their bachelor’s and a few year’s work experience.

    • anonymousskimmer says:

      This is fine except you aren’t talking about “things”, you’re talking about life experiences with consequent formative impact on people.

      I believe that development of a person’s exploration and justification abilities is a positive, regardless.

      The main problem is that people aren’t having their (international) job-search and job-creation abilities developed.

  19. suntzuanime says:

    Net Neutrality strikes me as one of those issues like climate change where great uninformed masses have been whipped into a frenzy by dishonest demagogues, but there’s a real point buried underneath all the bullshit. I don’t expect losing NN to shut down the internet, but creating more points for negotiation creates more potential frictional loss from transaction costs and creates more attack surface for monopolists and scammers, and it’s not clear that there’s much value to be gained. The internet is a key part of our infrastructure, ISPs should be utilities, not engines of economic dynamism.

  20. Candide III says:

    Scott, why are you using Audacious Epigone’s IQ vs free speech absolutism chart without attribution? I can understand your not wanting to link to him and send traffic his way, but attribution is basic etiquette.

    • Scott Alexander says:

      Because the post is a bunch of alt-right cliches strung together, I don’t think I can get away with linking it, and I assume Epigone prefers that their points be spread without attribution than not spread at all. If I’m wrong and they tell me, I’ll take it down.

      • ManyCookies says:

        Did you not want to link the article because it’d (really) distract/detract from otherwise interesting data, or because you’re wary of potential backlash?

      • Candide III says:

        I’ll omit all citations in my next paper — on the assumption that other papers’ authors would prefer that their points be spread without attribution than not spread at all — and see how that goes over. /sarc

        PS: I might not have been clear enough above and there might be a misunderstanding. One doesn’t have to link to show attribution, just mentioning the name/handle in plain non-hypertext letters should be enough. That’s why I said I understood your not wanting to link to him.

      • Svejk says:

        One of the reasons for proper citation etiquette is to allow readers to judge the quality of the source. Earlier this week I was part of a discussion of a dataset that turned out to have been improperly compiled by its original source; this was not discovered because the data had been “laundered” through a trusted intermediary (with no ill intent, neither party had noticed the error).

      • a reader says:

        Now I feel less guilty for reading the Audacious Epigone, if even Scott Alexander does it…

        I would love to see a rational, objective, thorough and sincere rebuttal of the alt-right, that concedes them all the points where they are right but debunks with solid arguments all the points where they are wrong. But I understand that such a thing is very risky, that even the tiniest concession to the alt-right narative may mark someone as “public enemy”.

        • 1soru1 says:

          The points where the alt-right are ‘correct’ are pretty much entirely where they claim ownership of mainstream (or occasionally leftist) ideas. They just have a knack for presenting them in such a way that it seems like they are the only ones daring to say that commonplace.

          For example, ‘GOP Senators are heavily influenced by corporate donors’, ‘libertarianism is unworkable’, ‘low-IQ people tend to be kind of stupid’, ‘woke college students are annoying’, …

      • Douglas Knight says:

        You could link to Razib Khan instead. He doesn’t an composite “absolutist” metric, thus neither a simple graph; and he defines “free speech absolutism” by combining just two questions, rather than AE’s all 5.

  21. Lillian says:

    The size of a nation’s legislature tends to be about the cube root of its population. Also, the US House of Representatives is “one of the world’s most undersized” legislatures.

    This is thanks to the Reapportionment Act of 1929, which established an effective permanent ceiling of 435 Representatives. Prior to that, there had been an expectation that the size of the House would be raised after every couple of censuses. It is precisely because our House of Representatives is undersized that we have gotten two splits between the electoral and popular vote in the last couple of decades. Increasing the size of the House would fix the problem (if you think it is one) without requiring any Constitutional amendment. Much more importantly though, it would make Representatives more responsive to their constituents, which i think is to the benefit of everyone.

    • Incurian says:

      Not if your constituents are idiots.

      • Lillian says:

        Huh, weird i could have worn i saw an analysis showing a larger House would eliminate the split. Then again it might have been that crazy 10 000 Representatives proposal. Finer grade representation does make such splits less likely, but the point at which it eliminates it for 2016 may be unrealistically high. Regardless, i do think we could stand to increase the size of the legislature to make it more responsive to the public

    • On the other hand legislatures that are too large have a very hard time getting anything done as their size gets too far above Dunbar’s Number. Look at Catherine the Great’s attempt to create a Russian legislature to see how badly a too-large body can be at getting anything done.

      • Evan Þ says:

        So the real problem is the size of the country, and letting California and Texas secede would go a ways toward fixing things?

        Alternatively, how well is the Indian legislature working?

        • Lillian says:

          Alternate solution: Decrease power of Federal government and devolve more to the states. The national legislature not being very representative is less if an issue if they have limited power. It also less of an existential crisis when the other side takes control of the Federal government. This would mean no more same sex marriage in the more conservative states, but the most socialist ones would be free to institute pseudo People’s Republics if they want.

    • hopaulius says:

      The USA has myriad legislatures, at the national, state, county, city, and smaller levels. It would be interesting to count that total of legislators and see how it compares to the cube root of the total population.

      • hopaulius says:

        There are 7,383 state legislators in the US. http://www.ncsl.org/research/about-state-legislatures/number-of-legislators-and-length-of-terms.aspx. The cube root of the US population is ~689. I would say we have too many legislators.

        • Lillian says:

          State legislators should be numbered according to the cube root of each state’s population, not the cube root of the entire population. That said, off handedly it looks like a few states have Houses that are too big, most egregiously New Hampshire, but none seem to have Houses that are too small. It’s an interesting contrast to the state of affairs at the Federal level.

          • Rob K says:

            And New Hampshire does in fact have problems with the size of its legislature. While you might hypothesize that a legislature that size would make the legislators closer to and more responsive to their constituents, in practice since so many of the legislators are part-timers without support staff, decision making power is much more centralized with the legislative leadership than in most states.

    • cassander says:

      There’s a literature on the optimal size of legislatures. Do we have any evidence that more representation leads to better outcomes?

  22. Sfoil says:

    Trying to calculate “Wins Above Replacement” for generals isn’t sufficient, because warfare is highly contingent and chaotic, whereas baseball is a controlled contest. I wonder if this method isn’t best for sorting out who’s who among pre-Napoleonic European limited warfare generals and their counterparts in other parts of the world. There’s probably something to it, though; the far right consists of .Napoleon, Julius Caesar, al-Whalid, Shingen, Nobunaga, Frederick the Great, and U.S. Grant. All of these (well, I am actually not very familiar with Shingen) are unquestionably giants of military history. MacArthur, Rommel, and Lee all have large but divisive reputations, although I’m a little surprised to see Saladin on the left.

    European, American, Japanese, and Islamic generals seem to be well represented. I’m pretty suspicious of the relative lack of historical Chinese and Mongol generals though; maybe I’m missing them. Admirals seem to be getting shortchanged as well; both Horatio Nelson and Yi Sun Shin are unremarkably above average.

    • bean says:

      Admirals seem to be getting shortchanged as well; both Horatio Nelson and Yi Sun Shin are unremarkably above average.

      I suspect this is due to low numbers of battles. Nelson is credited with four battles, and his WAR/battle is over .5, which is really high. Poking at other Admirals, Nimitz is given credit for Midway, Coral Sea (loss), Iwo Jima and Okinawa. That leaves out two years of the war, and gives him a pretty low WAR. Spruance gets 5 battles, all wins, while Halsey gets 8 (6-0-2). But Halsey gets credit for several battles in the Solomons that he wasn’t actually at. Something’s wonky with his count of battles, if nothing else. And the bit where Norman Scott’s WAR/battle rivals Nelson’s is particularly amusing. I suspect that a better number will need to consider total number of battles, too.

      • gbdub says:

        Plus Coral Sea wasn’t even really a loss. The US repelled an invasion, and set themselves up well for Midway by putting two carriers (temporarily) out of action for that battle.

        Naval battles in particular I think are tough to score, since the strategic winners are a little less obvious and “tonnage sunk” is such a tempting score to judge.

        • bean says:

          Plus Coral Sea wasn’t even really a loss. The US repelled an invasion, and set themselves up well for Midway by putting two carriers (temporarily) out of action for that battle.

          I just assumed that was obvious.
          (Actually, I was writing quickly and that was not the point I was trying to make, but it was notable enough to call out.)

          Naval battles in particular I think are tough to score, since the strategic winners are a little less obvious and “tonnage sunk” is such a tempting score to judge.

          True. As a test, I pulled up our old friends Beatty and Jellicoe. His algorithm is good enough to count Jutland as a British victory (so why’d it mess up Coral Sea?), but that’s Jellicoe’s only entry. Beatty gets that, Dogger Bank, and Heligoland Bight. This is impossible to take seriously.

    • DeWitt says:

      There’s probably something to it, though; the far right consists of .Napoleon, Julius Caesar, al-Whalid, Shingen, Nobunaga, Frederick the Great, and U.S. Grant. All of these (well, I am actually not very familiar with Shingen) are unquestionably giants of military history. MacArthur, Rommel, and Lee all have large but divisive reputations, although I’m a little surprised to see Saladin on the left.

      I’m not so easily convinced. Number crunching is neat for data analysts and the like, but history is a chaotic and unpredictable field where it will do you little good. The generals that rank high on his list had many things more going for them than sheer skill at arms. From the top of my head:

      Napoleon, the man to be so immensely far ahead of any other general on this whole list, was the first general to also have access to a nation-state’s worth of citizens to draft. This is something of a recurring theme in history, where certain innovations in warfare cause major early successes until people catch on and an equilibrium is reached; Napoleon wasn’t nearly as succesful at Waterloo, when his opponents had caught on. You could (erroneously) credit Napoleon with inventing the Napoleonic warfare that was named after him, but much of the groundwork had already been laid for him because of that whole revolution thing you may have heard about.

      Alexander the Great has a similar thing going on: his father’s military reforms were revolutionary enough that Macedon managed to unify nearly all of Greece, which might well be called more impressive than the latter conquest of Persia. Defeating the Persians in pitched battles was something the Greeks had been doing for the past two centuries, and the one thing Alexander managed to do was doing so offensively, rather than defensively. This is giving Alexander rather much negative credit, and the man was evidently a skilled commander, but he had many things going very very right for him, and attributing his wins to tactical and strategic genius alone leaves out many issues.

      Julius Caesar doesn’t belong on this list. This isn’t me saying he was a poor general; I’d have argued he belonged on the list’s other end then. It’s me saying that the data on Julius Caesar is horribly off, that the one account of his campaigns are literally propaganda that he wrote by himself, and that conquering Gaul is very easily done when most of the legwork is performed by.. Other Gauls. It’s unclear that Caesar was uncommonly strategically or tactically skilled, his opponents were stubborn people beset on many sides by their own people also being bought off, and there’s such a scarcity of reliable data that any respectable historian would have to be so cautious with his assessment that the end result would be awfully vague.

      Frederick the Great was indeed an excellent general, but we have to remember that it’s easy to win battles when you obsessively ensure your army is much more well-trained and disciplined than others of the time. Having such an elite fighting force means cutting into the size of your army, and for all the victories Frederick won, he suffered losses on the other side of his country: a large reason the Seven Years’ War ended up a shaky win for Prussia at all is because a lucky Russian inheritance saw the new czar switch sides. A clever commander, to be sure, but not one without his flaws.

      al-Whalid is a general who had very much going for him, in the sense that he commanded a great many people from warrior cultures who had never before been properly unified, and then also had the fortune of leading them against two empires that were immensely exhausted from having fought one another so much. As with Greece and Mongolia, history shows that when areas plagued with endemic warfare unify, they tend to draw upon immense forces of highly trained men. As with a few iterations of China, history teaches that empires in disarray are rather soft targets. Both of these meant that much of what the Arabs conquered was done with armies that had been fighting themselves not half a century ago, against opponents that were very much weakened from mutual warfare in turn.

      This isn’t a place where I’ll need to list the ways in which the Union held the advantage over the Confederacy, but even so, Grant is another general who had very much going for him. The man was a skilled general, and he certainly has a place on a list like this, but he didn’t fight under such duress and hardship that he performed extremely much above expectations.

      This leaves the two Japanese generals set to the right that I’m not about to criticise very much: Takeda Shingen and Oda Nobunaga. Both were feared and skilled generals; neither, however, were poised at such advantageous positions to wage war as were people like Alexander or al-Whalid. I suspect that fractured areas like Sengoku Japan are where such highly skilled generals truly begin to stand out: at the beginning of the Sengoku Jidai, it’s difficult to speak of any one out of fifty or so clans to have a great advantage over the other, and so the individual skill of each commander counts for much more. Something similar would go on in Greece, where I’d want to name Epaminondas as someone whose genius shook the city-states present there, or China, where the Qin government’s reforms were instrumental in causing it to conquer what would become China’s first empire.

      As for the generals you mentioned that seem to be shunted to the left a bunch more..

      Lee and Rommel fought doomed wars. Lee fought against a numerically, technologically superior army that had better supplies to boot; Rommel was stranded in unfamiliar lands while under the command of highly incompetent Italians and given much fewer supplies than an army his size would need. As with Hannibal: what could they reasonably have done? An earlier commenter posts about Lee possibly being able to get better results out of starting WW1 half a century early, but it’s unclear to me that he wouldn’t have been relieved from command for doing such a thing. It’s very easy to blame a war’s loss on the generals involved, but I’m not really sure that these two men could have done much better than they did.

      MacArthur is contentious, but I don’t know enough of him to really make a proper judgement here. Regardless, ‘not absolutely brilliant’ is probably correct in history’s assessment of the man, and I hesitate to say more.

      I’m a little surprised to see Saladin on the left.

      Why?

      I’m pretty suspicious of the relative lack of historical Chinese and Mongol generals though; maybe I’m missing them.

      Subutai should probably have been included here somewhere, and his being absent is suspicious.

      China is a different issue, because it really does have a relative lack of people that I’d argue should belong on a list like this. The gist of the matter is that Imperial Chinese generals have very, very sucky jobs: if they’re succesful, and manage to win victories, chances are that envy and fear of the general gaining too much power will get them exiled or killed. If they’re unsuccesful, they tend to get executed more often than not anyhow. Yue Fei is an example of a general who could have been (even) greater than he already was, had he not been born in the wrong land at the wrong time.

      Still!

      Trying to calculate “Wins Above Replacement” for generals isn’t sufficient, because warfare is highly contingent and chaotic, whereas baseball is a controlled contest.

      This is probably correct, and the one takeaway I’d like my post to have: history doesn’t allow for experiments and charts like these should be taken with the biggest ever grain of salt.

  23. US says:

    One aspect of the gender dysphoria/autism link that may be worth considering: Which denominator is best? What’s the proportion of autistics (with/-out?) actual sexual (/partnership) experience that have gender dysphoria? A substantial proportion of in particular male autistics never have sexual/romantic experiences throughout their lives (estimates vary, but it’s not a small subset), and presumably such experiences is an important component of sexual identity. A few related data:

    “Engström, Ekström, and Emilsson (2003) recruited previous patients with an ASD diagnosis from four psychiatric clinics in Sweden. They reported that 5 (31%) of 16 adults with ASD had ”some form of relation with a partner.” Hofvander et al. (2009) analyzed data from 122 participants who had been referred to outpatient clinics for autism diagnosis. They found that 19 (16%) of all participants had lived in a long-term relationship.
    Renty and Roeyers (2006) […] reported that at the time of the[ir] study 19% of 58 ASD adults had a romantic relationship and 8.6% were married or living with a partner. Cederlund, Hagberg, Billstedt, Gillberg, and Gillberg (2008) conducted a follow-up study of male individuals (aged 16–36 years) who had been diagnosed with Asperger syndrome at least 5 years before. […] at the time of the study, three (4%) [out of 76 male ASD individuals] of them were living in a long-term romantic relationship and 10 (13%) had had romantic relationships in the past. […] A total of 166 (73%) of the 229 participants endorsed currently being in a romantic relationship or having a history of being in a relationship; 100 (44%) reported current involvement in a romantic relationship; 66 (29%) endorsed that they were currently single but have a history of involvement in a romantic relationship; and 63 (27%) participants did not have any experience with romantic relationships. […] Participants without any romantic relationship experience were significantly more likely to be male […] [P]revious studies that exclusively examined adults with ASD without intellectual impairment reported lower levels of romantic relationship experience than the current study, with numbers varying between 16% and 31% […] The results of our study can be best compared with the results of Hofvander et al. (2009) and Renty and Roeyers (2006): They selected their samples […] using methods that are comparable to ours. Hofvander et al. (2009) found that 16% of their participants have had romantic relationship experience in the past, compared to 29% in our sample; and Renty and Roeyers (2006) report that 28% of their participants were either married or engaged in a romantic relationship at the time of their study, compared to 44% in our study.” (link)

    • chridd says:

      I don’t see how this is relevant? Gender dysphoria is about being comfortable with one’s body and with how one is seen, not about romantic or sexual relationships. In some cases, being trans can cause one to be uncomfortable with relationships (since sex involves parts of the body they might not be comfortable with, and romantic relationships often involve gender roles, and probably other reasons), but I don’t think being in a relationship has any effect on whether one is trans. (From what I remember, I think there have been trans people who thought their problem was not being in a relationship and found that it didn’t actually help.)

      • Hyzenthlay says:

        I don’t see how this is relevant? Gender dysphoria is about being comfortable with one’s body and with how one is seen, not about romantic or sexual relationships.

        In theory, yes, but gender identity doesn’t exist in a vacuum, and romantic/sexual feelings often are a big part of someone’s sense of their own gender. In fact, I’d say it’s rare for those things to be totally disconnected from each other…though they can interact in complex and often unpredictable, individualized ways that often won’t conform to society’s expectations.

  24. bean says:

    So a general who fought 30 battles and won 50% would be “better” than a general who fought 10 battles and won 100%.

    That’s not quite true. AIUI, a general who fought 30 battles and won 20 would rate about the same as a general who fought 10 and won all, his estimates of relative strength being equal.
    His WAR/battle statistic is a lot more interesting, and rates quite a few people higher than Napoleon.

  25. hanmeng says:

    Bitter melon (which can indeed be quite bitter) is used in many different Asian and Caribbean cuisines, but it’s also assumed to have various medicinal properties.

  26. wintermute92 says:

    There’s a story about Feynman going to meet someone who he was told was “a great general”. As I heard it, Feynman asked “well what makes a great general?” And the person who had used the phrase said “well, perhaps a general who has won five battles in a row?” And Feynman said “well if every battle is decided at random, then wouldn’t about 3% of generals be great?”

    I never believed the story, because why would anyone who knew military history set ‘wins’ or ‘consecutive wins’ as the standard for greatness? But now I’m doubting myself; I’ve got a bad habit of assuming people instinctively reason like Bayesians.

    Surely the “greatest general” is whoever consistently did better than an average general in their circumstances, like troop count or position on the field. (Amended, perhaps, for whatever else we value: performance across many types of fighting? Development of novel strategies?) Napoleon makes a good show here, but in the West the title perhaps goes to Caesar or Hannibal for a string of utterly improbable victories clearly won by the decisions of the general.

    (I don’t know enough military history from elsewhere to cast a vote.)

    • bbartlog says:

      Your comment about decisions makes me think of another aspect of generalship which is remarkably important, at least historically – charisma. Although we give generals credit for brilliant decisions and maneuvers, many of the greats on the list are also known to have had some nearly preternatural ability when it came to commanding the loyalty of their troops. For example, Napoleon, on his return from Elba, as he went along took command of all the French soldiers and units that were sent to oppose and capture him.

      • Protagoras says:

        It can be hard to fully disentangle cause and effect, though. Being perceived as a winner is in itself something that tends to inspire loyalty.

    • sohois says:

      I can’t recall where I heard it from but I’m sure there is some old bit of wisdom about the best generals showing their worth not in the battles that they fought but in those that they didn’t fight, i.e. the best generals will manoeuver in such a way that their enemies are beaten before they even step on the field, or they can end a war with a single battle or such things. I’m probably mangling the idea a bit but I’m sure my point is easy enough to grasp.

      I suppose it comes down to a debate about the value of strategy versus tactics. Winning in individual battles is measuring the tactical ability of a commander. Actually winning the war or taking great spoils is presumably a demonstration of strategic ability, though of course not every general will have that kind of strategic control; I remember reading of Rommel that he continually pushed for certain actions to be taken by the German forces but was frustrated at Hitler’s ignorance and poor grasp of strategy.

  27. Maggie says:

    Re: I-66. I think the main reason people were angry is they were led to believe the tolls would be $3-5 and they ended up being $40. Also, I’d be worried about sufficient signage (and radio announcements) warning people of the toll to give them the opportunity to seek an alternate route.

    • Guy in TN says:

      Agree. There’s also the issue that HOV/hybrid lane are seen as serving a larger social good (rewarding people who reduce pollution), while restricting access to the poor does not serve that purpose.

      Scott’s phrasing is interesting. He says:

      Commuters okay with a road being illegal to use (except for certain groups), but angry when it was legal to use but with a very high toll.

      Under the second scenario, the road is still illegal to use for certain groups- people who can’t afford to pay the toll. Likewise, you could phrase the first scenario as “legal to use for everyone, but requiring the purchase of a hybrid vehicle”.

      In either case, the road will be legal for some people and illegal for another. And we are choosing whether that factor will be based on money, or based on something else.

    • Janet says:

      Another factor is that the high toll then diverts a very large amount of traffic from the Interstate onto the (also highly congested) secondary roads and bridges en masse. So, even if you never had any intention of using the Interstate and weren’t heading into the city at all, your commute can suddenly and unexpectedly become dramatically worse. That’s true even if you’re doing the “responsible” thing and riding the bus or carpooling. Nor can you effectively modify your behavior, if you weren’t ever going on the toll road, nor even saw a sign for the toll road, and the traffic surges are unpredictable to you. Of course, nobody from VDOT is measuring the slowdowns on alternate roads when they declare the tolls a “success”.

      • Evan Þ says:

        What’s the aggregate delay there, versus the aggregate delay for cars stuck on a jammed freeway without the tolls, or the aggregate delay when SOV’s aren’t allowed on the freeway at all?

        It seems this’s making the pain more visible, or differently-distributed, but I’m not convinced it’s increasing delay overall.

        • Janet says:

          Well, as I said… nobody’s measuring anything except the speed on the Interstate, so it’s not measured, i.e. it’s assessed a value of zero by fiat. I don’t know how you would “aggregate” delay totals, when now the delays are affecting people who weren’t going to go on the Interstate at all. They have instead been given the gift of a chaotic and stressful trip to their doctor’s appointment, or dropping the kids off at school, because the surface roads are now completely stopped.

          I can see the rationale for giving commuters heading into DC the option of a paid “fast lane” or a free “slow lane”. I can see that people who really, really, really need to get to the city might find even a $40 toll better than the alternative of being late or missing something. I’m mildly in favor of toll roads in general, as I think it’s fairer for roads to be paid for by the people who use them, more use = more payment. But toll roads shouldn’t cause bystanders these kinds of headaches, and VDOT should be measuring all of the costs before declaring their project a success.

          • Evan Þ says:

            Yes, VDOT should be measuring all the costs, even though that’s difficult.

            My argument is twofold: Suppose that before this change, ten people were stuck on the freeway and delayed half an hour each, while ten other people didn’t get any delay on city streets. Now, suppose that after the change, eight people stayed on the freeway and were only delayed 15 minutes, while the other twelve were delayed ten minutes on city streets. That means 12*10=120 min delay introduced, with 8*15=120 minutes delay removed… the change actually breaks even! Change the numbers just a bit, and it’d be an improvement!

            But, remember that previously SOV’s weren’t allowed on the freeway at all. So, they’d already be clogging up city streets, meaning that people were already delayed somewhat there, which makes those numbers very likely an improvement.

          • anonymousskimmer says:

            @Evan Þ

            Time is not fungible. Whether within one person or between two people.

            Assuming the total number of vehicles on the roads don’t change (just their distribution), 8 voters get a better deal while 12 voters get a worse deal.

            And from anecdotal experience a given amount of traffic causes greater delays on roads with stop signs and lights than on highways.

          • Evan Þ says:

            @anonymousskimmer, on a societal level, I’m perfectly happy to give twelve voters a slightly worse deal to give eight a much better deal.

            Though, politically speaking, it can indeed be more difficult.

          • anonymousskimmer says:

            I guess the only real problems with that morally is that the people themselves aren’t necessarily making the decision (even if in the form of a gamble), and that given enough such decisions some few people will be relatively very well off while another few people will be relatively very poorly off.

    • Michael Crone says:

      The article here says that the earlier ban on solo communters was widely violated. Highly likely in my view that the expected payment under the ban (“ban”?) was an order of magnitude below the actual tolls turned out to be. https://www.google.com/amp/s/wtop.com/dc-transit/2015/10/i-66-hov-violators-will-soon-have-to-pay-up/amp/

  28. Deiseach says:

    Dear Sam,

    I was inclined to be sympathetic to your plaint about being horribly censored in San Francisco, right up until you started dropping the Galileo comparisons.

    Every kook likes to claim “They persecuted Galileo, and now they’re persecuting me!”

    I’m taking a wild guess here, Sam, but I’d say your smart people with radical ideas for intelligence augmentation and life extension being hounded out of San Francisco are not Galileos (Galilei?)

    Okay. So I gritted my teeth and read on, despite the increasing “stupid dumb wrong pop culture version of history” and decreasing “making a point here” tone of the piece, until the end where alas, you have lost me forever.

    Now I want to tie you in a sack, give that sack a good beating with a heavy stick, and then toss it into the famous Bay of the Bay Area. But why? you gurgle, as the weighted sack with you, bruised and bleeding within, sinks beneath the Pacific waters.

    Because this, I murmur softly as the depths enfold you.

    I can picture Galileo looking up at the sky and whispering “E pur si muove” here today.

    Yeah, that invented piece of myth-making propaganda that never happened, and you put your own unique twist on it. You may be a shit-hot entrepreneur, I have no idea, but I do know if I were judging “Should I invest with this person? Is he someone with breadth and depth?” that question would have been answered: nah, he’s another kid who uncritically swallowed the same kind of pap as he is (ironically) arguing against vis-à-vis the suffocating atmosphere of conventional orthodoxy in what is permissible to think and “everyone knows that…” pseudo-knowledge.

    Oh, how you wannabe scientists or science-adjacent or “I fucking LOVE science!” types love you some Galileo persecution similes.

    Yours,

    Disgruntled “for feck’s sake pick up a decent history book and stop rotting your brain with The World War II History Channel Bigfoot Specials” Old Curmudgeon.

    • poignardazur says:

      That’s a little extreme, but yeah. Fuck this shit.

    • BlindKungFuMaster says:

      I have actually no idea where you are coming from, could you elaborate? What kind of pap did Sam swallow? Can’t just be the pop culture version of history, can it? Is he not persecuted enough to make Galileo similes and therefore he has to beaten and drowned? Help me out, I can’t do allusions and I’m not American.

      • rlms says:

        See here. TL;DR actually Galileo deserved it.

        • quaelegit says:

          Edit: First off highly recommend the link, it’s long but very entertaining and discusses the history of heliocentrism and astronomy in Europe as well as Galileo’s tale specifically.

          To elaborate on rlms’ comment a bit more: Galileo spent a couple decades insulting and attacking his powerful friends (Jesuits, the Pope), so when his enemies got him in political trouble his former friends didn’t help him out. I don’t know if that’s “deserved it” (his enemies did cheat to get him in trouble) but he has no one to blame but himself that he lost the friends who could have helped him avoid punishment (which ended up being pretty minor anyways).

          Both Galileo and Giordano Bruno pattern match more closely to “smartasses who had it coming” than “martyrs for science” (and calling Bruno a scientist is a bit like calling Christopher Nolan a scientist because he made “Interstellar”.)

        • poignardazur says:

          Well, “deserved” it is a little extreme.

          More like, tl;dr Galileo isn’t a free “get out of everyone calling you an idiot” card.

          • Deiseach says:

            Also, if Altman wants an example of “expressing politically incorrect thinking gets techies in trouble” he does not have to go all the way back to Galileo for an example.

            I seem to dimly recall there was some chap in the news a while back, now what was his name again – Bert? Peter? Jim – I think his initials were J.D.?

            I note Altman isn’t touching that one with a ten-foot bargepole, safer and easier to rely on the lazy thinking of “recognised pop culture example of Good Guy accused of Wrongthink who was Right After All (and makes me and my mates look burnished by the glow of reflected glory by comparing ourselves to Good Guy Hero Boldly Standing Up To Theocratic Tyranny)”.

      • Deiseach says:

        I have to agree with poignardazur, that was perhaps somewhat extreme, but when I see someone dispose themselves upon their fauteuil and then commence to emulate Byron’s Manfred (Oh woe is me, I am so damned, so gloriously damned due to my genius, my genius which is far above these common dolts who persecute me), it instead makes them look like Mr Mantalini as far as I am concerned:

        Mr Mantalini, who had doubtless well considered his part, no sooner heard these words pronounced in a tone of grief and severity, than he recoiled several paces, assumed an expression of consuming mental agony, rushed headlong from the room, and was, soon afterwards, heard to slam the door of an upstairs dressing-room with great violence.

        …The dressing-room door being hastily flung open, Mr Mantalini was disclosed to view, with his shirt-collar symmetrically thrown back: putting a fine edge to a breakfast knife by means of his razor strop.

        ‘Ah!’ cried Mr Mantalini, ‘interrupted!’ and whisk went the breakfast knife into Mr Mantalini’s dressing-gown pocket, while Mr Mantalini’s eyes rolled wildly, and his hair floating in wild disorder, mingled with his whiskers.

        …At this crisis of his ravings Mr Mantalini made a pluck at the breakfast knife, and being restrained by his wife’s grasp, attempted to dash his head against the wall — taking very good care to be at least six feet from it.

        First, the examples he gives of unheralded geniuses being run out of the Bay Area on a rail due to the drone-like groupthink that is now ruling the roost there, it’s “people peddling the elixir of immortality”. It’s California, Sam; there’s a rich tradition of kooks promising that their patented diet of kale and liver extract will allow you to live to be two hundred. Getting the bum’s rush does not mean that minds are closed by liberal shibboleths to the glories of “intelligence augmentation, genetic engineering, and radical life extension”, it’s because even in the climate of “Hello, I am a venture capitalist, here take this shed-load of money and set fire to it fund your start-up”, some ideas are still regarded as too wacky to take seriously.

        Second, he had to go and title it E Pur Si Muove, didn’t he? Juuuuuusst in case anyone missed the teeniest nuance of “Halp, I iz bein’ oppressed!” he’s got going on. The anecdote is dubious (even the Wikipedia article pitches it too strong, I think) but it’s a great propaganda image, in the vein of And When Did You Last See Your Father? – blond, blue-eyed innocence betrayed by its natural candour into too much frankness before a seemingly-friendly (but in actuality hostile) interlocutor, who will take its words and use them against it, and condemn an entire family to ruin in doing so.

        As I said, every kook and hobby-horse rider loves the comparison with Galileo to glorify their self-identified “persecution” in a Crackpot Syllogism – the great minds of the past had to struggle with opposition and persecution, I have to struggle with opposition and persecution, therefore I too am a great mind!

        Sam Altman is not the only guy who can fling quotes about, I’ll quote back at him Jimmy Durante from I’m The Guy Who Found The Lost Chord:

        People said I was mad, but that didn’t discourage me
        They said Mozart was mad
        They said Puccini was mad
        They said Louis was mad
        Who’s Louis? My uncle, he was mad!

        Thirdly, this pearl of false equivalence:

        In addition to the work Newton is best known for, he also studied alchemy (the British authorities banned work on this because they feared the devaluation of gold) and considered himself to be someone specially chosen by the almighty for the task of decoding Biblical scripture

        Newton’s Biblical interpretation had feck-all to do with his mathematical work, except insofar as Bible Code types love the kind of logical patterning they can tease out of the material. Once again, this is the “Great wits are sure to madness near alli’d;
        And thin partitions do their bounds divide” view of History, that Art only proceeds from pain and the best rock musicians all die young of drug overdoses.

        Poppycock. Newton’s crankery had nothing to do with his talent, and had he been devoid of it then it wouldn’t have made a straw’s worth of difference to his achievements. A crank Newton without the mathematical talent would have been a crank who lived and died in obscurity, not a bold pioneer of daring thought that pushed the boundary of what could not be said in polite society.

        Fourthly, you ain’t Galileo or Newton, Sam. And neither are your persecuted pals. And a guy who is sitting pretty as a lauded entrepreneur who is probably not living on rice and beans (though granted, if I believe the stories about rents in the Bay Area, you have to be at least a millionaire to be able to afford a cardboard box on the side of the road) looks rather absurd donning the tatters of the scorned, mocked and socially derided by the powerful makers and breakers of reputations. (If he genuinely is at risk of being kicked off the board of the two nuclear energy companies of which he is chair for not being sufficiently Politically Correct, I will be very surprised, but then again the one in California may possibly be that extreme, I don’t know).

        Fifthly and finally, this is the age-old complaint of a businessman decrying regulation, bureaucracy, and “why are these paper-pushers stopping me from making money in the free market of competition of the fittest?”; a robber-baron of the Gilded Age could happily sign his name to:

        I don’t know who Satoshi is, but I’m skeptical that he, she, or they would have been able to come up with the idea for bitcoin immersed in the current culture of San Francisco — it would have seemed too crazy and too dangerous, with too many ways to go wrong. If SpaceX started in San Francisco in 2017, I assume they would have been attacked for focusing on problems of the 1%, or for doing something the government had already decided was too hard.

        Yes, drat that interfering government telling us what we can and can’t do with our hard-headed horse sense, gumption and go-getterness! Under the august robes of a historical monument lie the gaiters of “money-making opportunities may be lost because the kind of guys who come up with good ideas that can be developed into profitable business concerns aren’t mealy-mouthed enough for the cliques that control public opinion”.

        Well, perhaps you do have something in common with Galileo after all, Sam. He too was a pragmatic opportunist who bigged himself up, wasn’t backwards about coming forwards, used the ideas of others without being too concerned over sharing the credit*, and wanted to get a nice profitable sinecure out of his scientific work (branding the satellites of Jupiter as the Medicean Stars was the kind of marketing stroke that would be applauded today, never mind back in the early 17th century).

        *Not accusing Altman of this or any other chicanery, just if you want to hitch your wagon to a star, be very aware of the nature of that star before you start going “Me and my buddies – Galileo Redivivus, us!” Galileo wasn’t an innocent dreamer, is all I’m saying.

    • quaelegit says:

      >Galileos (Galilei?)

      I think it would be third declension — so “Galileones”? 😛

    • Nick says:

      Deiseach, check out the Business Insider article on Sam’s post:

      The essay, titled “E Pur Si Muove” — Italian for “And yet it moves” referencing a quip the heretic scientist Galileo Galilei said on his deathbed — drew swift and strong reactions from both supporters and detractors.

      In Altman’s view, people who have criticized those ideas for businesses have cast the entrepreneurs behind them as “heretics,” like how the Catholic Church sentenced Galileo to house arrest for (correctly) saying the Earth revolves around the sun.

      “[H]eretic scientist Galileo Galilei”!

      • Deiseach says:

        Better yet, now the anecdote has shifted from Galileo before the Inquisition to Galileo on his deathbed! Next it’ll be Galileo’s ghostly voice from beyond the grave 🙂

  29. Freddie deBoer says:

    Contra Turkheimer and others, a new team finds no tendency for environmental influence on intelligence to be stronger in the poor, not even in the United States

    can someone hook me up with the full text, my attempts at using my institution’s library site have left me paralyzed with frustration once again

  30. maldusiecle says:

    It’s baffling the way you and Cowen are interpreting that net neutrality study. If you look at the confidence-weighted numbers, those economists are more in the camp of uncertainty than approval. But even if they were 100% in favor of its repeal, what makes you think that their opinion is the “expert” opinion? The implications of net neutrality are broader than just the economic. If you took a poll among people who have studied the internet and its social implications, I might be a little more convinced.

    The implications of net neutrality are for our civil society. The internet, which was created by public research rather than private companies alone, is one of the most powerful communication tools in history. The question of who gets to access it, and how, is no neutral economic optimization. Of course there’s a huge economic aspect to it, but that’s not the whole question. As with many of these kinds of questions, it’s primarily an issue of power: whether the internet’s communicative power should be available to everyone, or whether it should be meted out purely by the grace of large, monopolistic ISPs which are already among the most hated corporations in the US.

    I realize this sounds a bit abstract, and that’s deliberate. I doubt the ISPs will immediately enact political censorship, or any of the other scary scenarios people have played up. I think we’ll see a series of subtle, complex systems of control whose net effect is to crush certain kinds of political dissent and minimize the free expression of certain marginalized groups. It’ll be a bureaucracy, most of all, and most of the damage it does will be more stupid than malevolent.

    • Murphy says:

      It also seems an odd choice of “experts” choosing economists vs network engineers.

      There’s a physical cost to setting up your infrastructure to be able to prioritize all packets from [list of 1000 companies who’ve given us money to get top priority].

      You could build a simple router which passes every packet in the direction it needs to go as fast as it can…. or you can build a complex router which does a lot of extra work inspecting each packet to decide how fast it should process it. That extra complexity is not free.

      By the time you’ve received the packet, looked up the table and decided where in the queue it should go and sorted it into the queue you could have simply processed every packet faster.

      It’s better described as “throttling all packets that are not in their priority paid list”

      The non-networking-experts against NN like to think of it as if it’s some kind of extra tier being added on top such that it hurts nobody when the technical reality is that it’s a monopoly provider making everyone service crappier so that they can charge more for non-crappified service.

      • poignardazur says:

        That’s BS. All ISPs already have structure for prioritizing packets in place (to differentiate between things like downloads and voice chats). Besides, the way these things work, all calculations are done once when opening the connection and then cached. The overhead you’re talking about is negligible.

        More importantly, the technology ISPs use for giving higher priority aren’t as simple as “put some packets on faster optic fibres”. The bulk of it is ISPs loaning cache space to major content providers.

        • Gobbobobble says:

          That’s BS. All ISPs already have structure for prioritizing packets in place (to differentiate between things like downloads and voice chats).

          I’m pretty sure that’s exactly what NN is supposed to prevent, at least one ISP got in hot water for throttling VoIP traffic.

          Besides, the way these things work, all calculations are done once when opening the connection and then cached. The overhead you’re talking about is negligible.

          This… sounds odd, but we’ve left the limits of my networking knowledge. My understanding was that data going from A to B has to pass through several routers and the exact path taken varies from packet to packet. Not quite “try all paths at once and see what works” but something approaching that.

          • Brad says:

            Would you consider prioritization of VoIP protocol packets (regardless of provider) to low latency channels to be a violation of NN? Because that seems like a pretty reasonable policy to me.

          • John Schilling says:

            Neutrality by definition doesn’t have a sign. If you can prioritize VoIP packets to a low latency channel, then you can “prioritize” VoIP packets to a high latency channel. That’s almost exactly what NN was designed to prevent, and the people who thought it imperative to prevent that by regulation called for Net Neutrality, not Net Implement the Obviously Reasonable Prioritization Scheme Rather Than the Obviously Evil One.

            So I’m guessing that the obviously reasonable prioritization to low-latency channels would be just as verboten as the obviously evil prioritization to high-latency channels. And somebody will find reason to make the complaint official.

          • rlms says:

            Various influential people (Eric Schmidt, Vint Cerf) propose allowing discrimination between different types of traffic, but not between traffic of the same kind from different sources.

          • Gobbobobble says:

            That’s actually pretty reasonable, but I worry it doesn’t make as good a Schelling point as “a byte is a byte”. It opens the door to legalistic wrangling over “what type of traffic is this, *really*?”. And what happens next time someone comes up with a new flavor of traffic?

            Google Chairman Eric Schmidt aligns Google’s views on data discrimination with Verizon’s: “I want to be clear what we mean by Net neutrality: What we mean is if you have one data type like video, you don’t discriminate against one person’s video in favor of another. But it’s okay to discriminate across different types. So you could prioritize voice over video.

            What about video + voice, a la Skype or Twitch? Live news? The uncharitable & pessimistic answer is “depends on if we own a competitor in that space”

          • John Schilling says:

            Various influential people (Eric Schmidt, Vint Cerf) propose allowing discrimination between different types of traffic, but not between traffic of the same kind from different sources

            How important is Vint Cerf’s opinion in the 2010s, really? Because if I google “net neutrality violations”, virtually everyone that Google thinks I need to pay attention to puts Madison River at or near the top of the list, and Madison River was all about deprecating VoIP regardless of the source.

            Also, pretty sure that if such a rule were put in place, it would be gamed by inventing new types of traffic that get top priority “regardless of source” but are impractical for anyone to actually implement without extraordinary access to / support from the providersin question.

          • MoebiusStreet says:

            I’d very much like to live in a world where a surgeon doing an operation by telepresence can be bumped up to top priority with no latency and jitter.

            I get that there’s some amount of wiggle room for this in NN, but I question whether the protocols involved can be handled in that way (or if it needs to be shunted into a different protocol just to enable this, increasing costs), particularly in an area where packet inspection is out due to privacy demanding encryption. It seems very much to me like we’d want a way to give traffic to/from County General Hospital priority, or at least the portion of that traffic that’s on port #X.

          • Gobbobobble says:

            I’d very much like to live in a world where a surgeon doing an operation by telepresence can be bumped up to top priority with no latency and jitter.

            It seems very much to me like we’d want a way to give traffic to/from County General Hospital priority, or at least the portion of that traffic that’s on port #X.

            They can set up or join a private network if it’s so important (which I agree it likely is). It’s not like this is a whole new thing: you don’t ship kidney transplants via UPS, you use a specialized medical service.

            Hospitals are big boys as it is, I’d bet most are much closer to the backbone than residential ISP customers already.

            Or, assuming a public interest case, carve out a special exception: explicitly medical traffic gets an “information superhighway” analogue of ambulance sirens and lights.

            Importantly, note that it’s illegal for anyone other than emergency services to purchase and operate in this meatspace fast-lane.

          • Another Throw says:

            This… sounds odd, but we’ve left the limits of my networking knowledge. My understanding was that data going from A to B has to pass through several routers and the exact path taken varies from packet to packet. Not quite “try all paths at once and see what works” but something approaching that.

            I think you are confusing design principles with implementation particulars, and for different systems, too. Your end-to-end transportation protocol should treat the network as a black box were all kinds of wild and crazy things can go wrong in order to be robust against them. But TCP being robust against (for example) multi-path routing does not imply that networks actually do it.

            Routing loops are a terrible failure mode, and having your routers willy nilly send packets around redundant routes are a really great way to get there. My impression is that all the routers in a network generally agree on a non-cyclic, spanning subgraph of the network to route traffic on which is periodically updated as the status of various links change (in order to keep it both non-cyclic and spanning). In those cases where someone really smart decides load-balancing between multiple routes is important, the balancing is accomplished at the stream level: all the packets in the same stream are routed on the same route. Because, even though your stream should be robust against it, multi-path routing is inefficient. The route packets in the same stream take rarely changes in practice.

            Furthermore, just because your end to end protocol should not depend on intermediate routers maintaining consistent state information does not imply that intermediate routers do not do their damnedest to maintain consistent state information anyway. While network layering models look really cool on the classroom blackboard, as an engineering practice, layering is considered harmful. Any router that isn’t inspecting packets—already in its memory!—in order to glean useful information is leaving money on the table:

            “This packet appears to be a session initiation, maybe I should expect more packets between these endpoints and cache some information about it?”

            “This packet appears to be fragmented and the link I am routing it onto (and every other link all the way to the destination) supports the reassembled length, maybe I should hang on to it for a new ms and see if I can reassemble it to save on the fragmentation overhead?”

            “This packet seems to be hopelessly malformed, why am I bothering to route it again?”

            “This packet appears to be part of a DDOS attack, maaaaybe I should do something about that?”

            “This packet appears to belong to a real-time application originating from the CEO’s office, maybe I shouldn’t get my engineer fired?”

            Blindly forwarding packets hasn’t been the norm for decades. If you haven’t noticed, computation has gotten insanely cheap, and basically every installed router is capable of doing a lot of it on all the packets it sees at line speed. Turning on Not-Net-Neutrality is as simple [ha ha] as rolling out a new configuration to all your routers. And the limiting factor is line speed, not computation. Doing so is not going to reduce the throughput of the router.

            Because, look, every network engineer that has built a more complicated network than plugging two computers together in Networking 101 has rapidly come to understand that not all bytes are created equal, and every nontrivial non-ISP network is already using the exact same equipment that ISPs have in order to implement Not-Net-Neutrality. Which they have helpfully named “Quality of Service.”

            What’s in a name?

            Imagine for a moment that instead of debating whether we should be “repealing Net Neutrality” we were instead debating whether we should be “allowing ISPs to implement industry standard Quality of Service features.”

          • Gobbobobble says:

            Thanks for the networking info. However…

            “This packet appears to belong to a real-time application originating from the CEO’s office, maybe I shouldn’t get my engineer fired?”

            This is a sort of “Quality” of Service anti-feature that NN is supposed to prevent. Mr Bigshot should not get to cut the line.

            Imagine for a moment that instead of debating whether we should be “repealing Net Neutrality” we were instead debating whether we should be “allowing ISPs to implement industry standard Quality of Service features.”

            Okay, cool, so the ISPs are going to be considered accessories to any illegal activity that they carry? They’re getting so much information from the packets, they can’t claim common carrier ignorance any more.

          • Another Throw says:

            This is a sort of “Quality” of Service anti-feature that NN is supposed to prevent. Mr Bigshot should not get to cut the line.

            When that realtime application is the CEO holding a video conference about that 300 million dollar deal, which your company lost when you looked like incompetent assholes from the bloody thing not working because everyone in accounting was watching cat videos instead of working?

            Yes, he absolutely should. Trust me.

            Okay, cool, so the ISPs are going to be considered accessories to any illegal activity that they carry? They’re getting so much information from the packets, they can’t claim common carrier ignorance any more.

            There might be just a few orders of magnitude of difference in computation between “reading the packet header and the header of the encapsulated packet to set a couple bits in your internal cache” and “intercepting then reassembling a 6 GB stream, breaking AES, decrypting it, running it through Google Deep Mind, and determining that it looks an awful lot like child pornography video.”

            Is Maersk an accessory to my smuggling because they read the shipping manifest I supply, and weight the container so they could more efficiently load the vessel?

          • Gobbobobble says:

            Oh right, I forgot we should decide public policy based on anecdotes about perfectly-legal kludges on one’s private network used to compensate for a failure to manage personnel. Clearly Mr Bigshot’s super duper important meeting also means he should get to cut in front of traffic from Cathy’s Cat-video Corp next door going out to her paying customers. That’s just quality service.

          • Another Throw says:

            I’m sorry, I forgot we live in a world where setting engineering standards by how the words “torsion member” makes you feel is a perfectly valid approach.

            Edit: To elaborate, suppose we’re considering repealing the regulation that says “no torsion members on public bridges” (which may or may not have been the result of lobbying by the Torsion Member-Free Builder’s Association). (Also, do bridges even have torsion members? Not an bridge guy.) And on the one hand we have a very vocal contingent that is radically opposed to the rule change because “torsion member” sounds like a dirty sex thing and don’t want them on their public bridges. Public bridges are platonic ideals that should be free of dirty sex things, after all. And on the other hand, we notice that practically every private bridge built since torsion members were invented contains torsion members, because decades of experience indicates that they Just Work Better. To such an extend that the manufactures of bridge components no longer even MAKE products that are not designed to implement torsion members. And it turns out that all public bridges have all the parts needed to implement torsion members, too. And all the operators of public bridges are asking to implement the torsion members, because they think (/claim to think) it will make their bridges work better (and, because of price stickiness, will temporarily increase their operating margin).

            The correct question to ask is whether public bridges are materially different in structure that there are solid engineering reasons to believe that torsion members will not make them work better—in fact, whether these engineering reasons indicate that even trying to implement torsion members on public bridges will be counter productive?

            Doubling down on “but platonic ideal of dirty sex thing free” with an extra helping of snark is not helpful.

            Sometimes, dirty sex things turn out to be excellent bridge components.

          • Another Throw says:

            Oh, and incidentally, priorities have to be set on the side of the link the traffic is coming from. A corporation (or resident) can set the priorities of traffic that leaves their premises, but their ISP has to prioritize the traffic that enters it. So in our example, the ISP would have to prioritize the teleconference over the cat videos on behalf of the corporation.

            Basically everyone, everywhere would LOVE LOVE LOVE to prioritize their inbound traffic over the low bandwidth last mile. But can’t. Because NN.

            The only solution available is to BUY BIGGER PIPES MAN! Because of how the market is currently structured, buying bigger pipes is exactly what the ISP wants you to do. Being able to use the pipe you already have more efficiently doesn’t make them any more money because they charge by bandwidth.

            I would, personally, be willing to pay a moderate premium for the ability to do inbound traffic shaping. Even at my house even. Wait, especially at my house.

          • poignardazur says:

            @Another Throw: I think you had everyone until you mentioned the CEO thing. That part’s more controversial.

            @Gobbobobble: The problem is, “a byte is a byte” is the mother of all leaky abstractions. First, different types of data (HTTP requests, buffered streams, unbuffered video calls) need to be treated differently.

            But second, even identical types of data need to be treated differently. If you have a small website and you send a 10min video every two weeks, you don’t need any special treatment. If you’re Youtube and you send like 2.000 10min videos every minute, then you start to need either special infrastructure, or else overload both your networks and the ISPs, and then everybody looses.

            So the big content providers need to set up proxy servers directly in ISP’s distribution networks: Google Global Cache, Netflix Open Connect, and probably others that don’t have names.

            Note that these servers were perfectly allowed by the now defunct 2015 FCC rules. My point is about net neutrality in general: the fact is, it’s more complicated than “a byte is a byte”.

            My ideal, I don’t know if you’d call it net neutrality, would be something like “A dollar is anonymous”. That is, everyone can buy the same services from ISPs for the same price, without them playing favorite. The obvious problem is that “don’t be a jerk just to make money” is really difficult to enforce, which is why I think net neutrality efforts should focus way less on legislation and more on increasing competition.

          • Gobbobobble says:

            If you’re Youtube and you send like 2.000 10min videos every minute, then you start to need either special infrastructure, or else overload both your networks and the ISPs, and then everybody looses.

            So get special infrastructure. What’s the big deal? Why does everyone else on the network have to suffer (or pay protection money) because a few services got big?

            My ideal, I don’t know if you’d call it net neutrality, would be something like “A dollar is anonymous”. That is, everyone can buy the same services from ISPs for the same price, without them playing favorite. The obvious problem is that “don’t be a jerk just to make money” is really difficult to enforce, which is why I think net neutrality efforts should focus way less on legislation and more on increasing competition.

            These are good points. We absolutely do need to work on increasing competition. I do argue for that. Better competition, like the you-must-share-the-wires system mentioned upthread, would do wonders toward making this less of a concern. But when you want to go from Safeguard A to Safeguard B, the responsible path is via having A+B, not “tear down A and wait for B to materialize eventually”. So the hot issue I spend my time arguing is “don’t tear down A with no plan”

    • Anon. says:

      I doubt the ISPs will immediately enact political censorship, or any of the other scary scenarios people have played up.

      Twitter, facebook, and google, companies which control access to information on the internet to a far greater degree than any ISP, do enact political censorship. Instead of worrying about future hypotheticals, worry about what is actually happening right now.

      • maldusiecle says:

        You have that completely backwards. If Facebook or Twitter censor your content, you’re free to post it elsewhere. If an ISP censors it, the content may become completely inaccessible.

        • Anon. says:

          You are “free to post it elsewhere” in the same sense that a homeless man is free to buy a Ferrari.

          • maldusiecle says:

            Your argument that sharing content is impossibly hard would be more convincing if it weren’t a comment on a blog with its domain and (small, sure) network of advertisers. It’s not that hard to post elsewhere.

          • Evan Þ says:

            @maldusiecle, where is it hosted? Does Scott have his own server under his desk (which wouldn’t really scale), or is he renting space from some larger corporation that can be pressured? And how long did it take him to build up a large enough audience to make his own domain successful, while blogging at livejournal and LessWrong?

          • Brad says:

            @Evan Thorn
            That extends the argument well beyond the one Anon. had made. He compared Twitter, facebook, and google to ISPs. Both of those groups are made up of very large corporations.

            Personally I lean towards maldusiecle’s side of the argument. Google is the closest case, but I have very little regard for the argument that being kicked off facebook and twitter means you can’t get your message out. There’s still more than plenty of man-hours spent outside those walled gardens.

          • maldusiecle says:

            Yes, there are difficulties associated with creating and hosting your own website. Obviously. And just as obviously, those difficulties are trivial compared to the difficulty of getting your content out to the world if an ISP has blacklisted the content. Paying an IT person to set up a server under your desk is annoying and somewhat pricy, but less annoying and pricy than fighting an ISP would be. Mentioning LessWrong if anything weakens your argument: it’s an independent platform not affiliated (to my knowledge) with any major corporation.

          • Evan Þ says:

            @maldusiecle, sure, censorship by Facebook/Twitter/Google/etc isn’t as bad as censorship by ISP’s. What I’m arguing is it’s still pretty bad.

        • The Nybbler says:

          If an ISP censors it, the content may become completely inaccessible.

          Infrastructure providers were already doing political censorship as well; ask gab.ai or The Daily Stormer. Net Neutrality rules weren’t doing a damn thing for preventing censorship of one side of the political spectrum, so cries about removing those rules resulting in political censorship rang hollow for much of the population.

          • maldusiecle says:

            You’re allowed to feel however you want, but as a substantive argument this comment has no merits. Political censorship was already present, and therefore we should…make it even easier to enact censorship?

            But anyway, as far as I’m concerned the Daily Stormer is a terrorist propaganda website. I have no more problem with its being censored than I would with ISIS being censored.

          • The Nybbler says:

            Your reply undermines your original argument

            As with many of these kinds of questions, it’s primarily an issue of power: whether the internet’s communicative power should be available to everyone, or whether it should be meted out purely by the grace of large, monopolistic ISPs which are already among the most hated corporations in the US.

            My point about political censorship is that these weren’t really the stakes. It’s whether the internet’s communicative power should be available only to those with views acceptable to one side of the political spectrum, or meted out purely by the grace of large, monopolistic ISPs (etc).

            But when you say it’s OK for the Daily Stormer to be censored you’re saying “the internet’s communicative power should be available to everyone” isn’t your true position in the first place.

          • maldusiecle says:

            I would have thought it was obvious when I said “the internet’s communicative power should be available to everyone” I didn’t mean for the “everyone” to include violent criminals promoting and organizing acts of terrorism and violence. Guess I was wrong -_-

          • Gobbobobble says:

            Was the Daily Stormer censored by ISPs? I thought I remembered it was their hosting provider…

            ETA:

            I didn’t mean for the “everyone” to include violent criminals promoting and organizing acts of terrorism and violence

            I do. Those should be taken care of by law enforcement, not ISPs.

            That’s the deal with common carriers: you (the carrier) treat all data equally and we (the law) don’t hold you accountable for the content of that data.

            This category of censorship is incidental at best to NN, the only thing making it not utterly orthogonal is the ISP snooping tech being in place makes it cheaper for LEOs.

          • Brad says:

            Neither an ISP nor a hosting provider, the trouble they had was with a DDoS mitigation services provider (Cloudflare) and with many different domain registrars.

            I fail to see what that has to do with NN, but who am I to tell “much of the population” from “one side of the political spectrum” that their all consuming resentment is causing them to hit the wrong target again.

          • The Nybbler says:

            I would have thought it was obvious when I said “the internet’s communicative power should be available to everyone” I didn’t mean for the “everyone” to include violent criminals promoting and organizing acts of terrorism and violence. Guess I was wrong -_-

            Indeed you were; the plain meaning of “everyone” is “everyone”. Not “everyone except those the infrastructure providers think is bad enough” (Note that the Daily Stormer, while a neo-Nazi board, is legal in the US, and that in the US even criminals have freedom of speech once released.). The latter both lacks moral authority and fails to be a Schelling point.

          • rlms says:

            @The Nybbler
            What is your proposal to deal with a monopoly of large companies behaving badly? If it is to involve the largest, monopolyest organisation in the world and hope that somehow makes things better, I think there are some libertarians here who would like a word with you. Since AFAIK there is no government interference that would cause underproduction of censorship-free infrastructure providers, if there are fewer of them than Daily Stormer readers would like that just indicates the strength of their preferences.

          • Gobbobobble says:

            Indeed you were; the plain meaning of “everyone” is “everyone”. Not “everyone except those the infrastructure providers think is bad enough” (Note that the Daily Stormer, while a neo-Nazi board, is legal in the US, and that in the US even criminals have freedom of speech once released.). The latter both lacks moral authority and fails to be a Schelling point.

            True, but still has nothing to do with Net Neutrality. Muddling ISPs and hosting/domain services under “infrastructure” is a bunch of FUD. NN only regulates the tubes, stop acting like a failure to prevent censorship before it even reaches the tubes is a partisan bias in NN.

          • The Nybbler says:

            True, but still has nothing to do with Net Neutrality. Muddling ISPs and hosting/domain services under “infrastructure” is a bunch of FUD. NN only regulates the tubes, stop acting like a failure to prevent censorship before it even reaches the tubes is a partisan bias in NN.

            I’m not claiming that there’s a partisan bias in NN. I’m saying that the argument for NN that NN will prevent partisan bias on the Internet rings hollow when it has failed to do so — particularly when some of those arguing for NN are the ones who have been implementing the partisan bias.

          • Gobbobobble says:

            Explain how removing NN will reduce partisan bias on the internet.

          • honhonhonhon says:

            “If you eat pizza you will get fat.” -> “But I already am fat!” -> “Explain how eating pizza will make you thinner.”

            For the other side:
            “If you spray with pesticides this year, we can’t use the land next year” -> “But you overfarmed so we won’t be able to anyway.” -> “Explain how spraying will make it usable next year.”

          • Gobbobobble says:

            “We shouldn’t get rid of our sprinkler system: it drastically reduces fire damage” -> “I am concerned about fire damage that sprinklers are not designed to prevent. We should get rid of the sprinklers.” -> “Explain how removing the the sprinklers will reduce fire damage”

            (edited for better mapping. It’s still not great, but few metaphors are)

      • Nornagest says:

        If Twitter, Facebook, and to some extent Google decide to pick up the role of information commissars, I’d be less concerned about that blocking information than about compartmentalizing it. If you’re pro-Blue and you can’t get your message out on Facebook and Twitter, sites specializing in Blue news will emerge to pick up the slack. But that means that you’ve just removed most of the opportunity for Blues and Greens to interact with each other, and further segregated the population into Blue and Green clusters all vigorously agreeing with each other. We already have quite enough trouble with people doing that voluntarily; we don’t need to add a structural layer to it.

        • Bugmaster says:

          I’m not convinced this would be such a bad idea. If Blues and Greens have few — or, ideally, none — opportunities to interact, then there’s no chance of Blue-vs-Green violence. It might also lead to more relaxed Blue and Green philosophies. It’s easy to hate the evil Other when it’s right there across the road from you; it’s much harder to do so when the Other is some sort of a mythological creature whom no one had seen for the past 20 years.

          • John Schilling says:

            I’m not convinced this would be such a bad idea. If Blues and Greens have few — or, ideally, none — opportunities to interact, then there’s no chance of Blue-vs-Green violence.

            Blues and Greens always have the opportunity to interact. They can elect separate Blue and Green governments which regard each other as evil incarnate, amass arsenals of nuclear missiles to defend themselves against each other, and then decide that the best defense is a good offense. Or, if sufficiently intermingled, they can elect a Blue government to stomp a demographically-destined boot in Green faces forever or until the Greens decide to readjust Blue demographics, whichever comes first.

            There’s always a chance of Blue-vs-Green violence. The question is whether there’s a chance for Blue-vs-Green peace.

      • birdboy2000 says:

        We can and should oppose both.

  31. dndnrsn says:

    I’m dubious about the suitability of baseball-style statistics to battles. Baseball produces a lot of data – 162 games a season. The sample size is going to be much bigger. This isn’t just true comparing sports to other things; it’s true between sports – looking at an MLB team’s raw record is going to tell you more about how good they are than looking at a boxer’s. Additionally, everything relevant to a baseball team is within that baseball team, more or less – how good a general is might not have an impact on the mission they are assigned or the resources they have available.

    • bean says:

      Now that I think this over more, what on Earth was he thinking using raw WAR? That only works when all of your data is fairly consistent coming in. That’s true of sports, not so much of war. Napoleon won this evaluation by a huge margin because he fought lots of battles and was good enough to win most. Bad generals usually get replaced before they get that many battles. But apparently he didn’t even think to normalize WAR until requested.
      And then there’s the bizarre, like Thomas Cocharane having one battle and -0.5 WAR. I admit that CacafuegoEl Gamo is going to be hard to find, but still…

      • dndnrsn says:

        And the way he counts battles is strange – just going with a Wikipedia list, ignoring that the Wikipedia list doesn’t do a good job of differentiating between more old-fashioned battles over in a few days at most, and modern “battles” that are really campaigns. Apparently Patton only fought in two battles, Operation Torch (which, apparently, was also a single battle) and the Battle of the Bulge. Or, Zhukov won Operation Barbarossa, which was a battle, and Operation Mars never happened. Or, apparently Haig and Foch are near the top of his “greatest generals” list, as far as I can tell.

        • bean says:

          Exactly. I think that this is why his 20th century numbers don’t look that good. In the old days, you’d have battles of a few days, with the commanding general physically present. In WWII, Halsey was listed against several battles that happened while he was fleet commander, but that he wasn’t present at, and which have no reflection of him as a tactical commander. Nimitz has a rating, even though he was in tactical command at no battles at all.
          And then I looked at MacArthur, and it got worse. Bataan is separate from the Philippines (and both have WARs above -.5, which is absurd at least for Bataan), he’s credited with a loss for Coral Sea (two problems here), Manila is separated from Luzon, and there’s no mention of anything in New Guinea. Because none of them have proper battle navboxes set up, apparently. Seriously, do not use this for anything after 1870.
          Edit: The more I look, the worse it gets. Haig is ranked quite highly. One of his two defeats is the First Day of the Somme. Several other actions from the Somme are credited to him, too. But he’s only got 3 credits after 1916. 7 of Attaturk’s 11 battles were part of Gallipoli. Well, 6, because one of them was Gallipoli itself. Did he run this by anyone familiar with 20th century military history?
          Edit 2: Just hit peak stupid. Montgomery is 6-0-0. Market Garden didn’t get scraped for some reason. This is failing basic sanity checks left, right, and center.
          Edit 3: Actually, that wasn’t peak stupid. Lloyd Fredendall‘s WAR being only barely negative was.

          • dndnrsn says:

            It also doesn’t consider battles anything but “win/loss” which… Well, he has Haig listed as “winning” at Passchendaele, but the Wikipedia article’s sidebar has “see analysis” for the outcome, which is usually a sign that it was not an overwhelming victory. Meanwhile, Ludendorff comes out with a loss for that one – and Ludendorff’s fairly overwhelming victory at Tannenberg is counted the same as Haig’s I-guess-he-won-maybe at Passchendaele.

            I don’t think even the most diehard Haig partisans claim that he was one of history’s greatest generals.

          • bean says:

            I think Haig’s popular reputation is undeserved, but he’s definitely a lot higher than he should be. Seriously, who is this guy and has he had any exposure to military history since high school?

          • dndnrsn says:

            Haig’s popular reputation is certainly undeserved – the whole “WWI generals were idiots bahaha” thing is unfair. But the argument for them is more “the conditions were not ripe for much other than ghastly attritional warfare and they were doing the best they could” not “well, the British won the war, ergo, Douglas Haig was a military genius not far away from Alexander.”

          • bean says:

            @dndnrsn
            Agreed, although I’d say that even if Haig was an Alexandrian-level genius, we couldn’t tell due to the circumstances.
            (The stupid keeps coming. Richmond Kelly Turner is credited with Savo Island, which makes a weird sort of sense, but only Kwajalein from among the Marshalls, and all three of the Marianas separately. And no credit for Iwo Jima or Okinawa. Oh, and Yamamoto? Coral Sea (v), Midway, Guadalcanal, Naval Battle of Guadalcanal, Rennel Islands (v). What on Earth is going on here?)

          • Paul Zrimsek says:

            On the other hand, it’s always bothered me that sabermetrics didn’t involve any actual sabers, so at least they fixed that. But I’ll bet my prize Manstein rookie card that this guy’s coming from the baseball side, not the military history side.

          • Civilis says:

            On the other hand, it’s always bothered me that sabermetrics didn’t involve any actual sabers, so at least they fixed that. But I’ll bet my prize Manstein rookie card that this guy’s coming from the baseball side, not the military history side.

            Purely at a tongue in cheek level, wouldn’t Football (American) be the obvious analog to use at deriving a statistical analogy between sports and warfare, given how often American Football is compared to war?

            To try to stretch this analogy, the different team positions can be compared to the different roles of generals. The quarterback is your strategic commander; what’s important is, did he win the war or campaign (ie, score a touchdown)? If so, how long and how many casualties did it take (attempts and completion percentage)? Were there any spectacularly failed offensives (interceptions)?

            Your receivers and running backs are the more field-level commanders. Some situations call for bold deep strikes to gain lots of territory (passes), some for overcoming heavy defenses for shorter advances (rushes). The things you can compare are territory gained (total yardage), successful attempts, and how many decisive battles they’ve won (touchdowns) or lost (fumbles).

            Your defensive generals are going to be harder to rate, but we can differentiate between ‘merely’ stopping the other side’s offensives (tackles) and spectacular counteroffensives (sacks and interceptions).

          • Paul Zrimsek says:

            I watched a few of ESPN’s Spanish-language MNF broadcasts this season; the announcers usually call the quarterback “mariscal de campo”.

      • The original Mr. X says:

        Now that I think this over more, what on Earth was he thinking using raw WAR? That only works when all of your data is fairly consistent coming in. That’s true of sports, not so much of war.

        Or, to put it another way: WAR. What is it good for? Absolutely nothing.

    • dndnrsn says:

      Looking at it, this seems like the stats that they throw on TV when MMA fights are happening. It might be interesting that fighter so-and-so gets 2.676 takedowns per fight, but absent knowing who those takedowns were against and under what conditions, not extremely useful for saying how good so-and-so is at taking people down. They’re not there because they’re useful statistics, but because people used to baseball or football or whatever are used to associating statistics with “this is a Legitimate Sport.”

      The things that this turns up were already areas of controversy, anyway.

    • baconbacon says:

      The issue is much deeper than that. One of the things that makes WAR reasonable for sports is that they are turn based. You can swap out players and they will take (roughly) the same number of at bats, and there is a strict opportunity cost for having player X on the field/court instead of player Y.

      Without that assumption you cannot have a meaningful WAR, and since wars aren’t fought around “total wins given a set number of engagements” makes the exercise (beyond the pun of WAR and war) pretty meaningless.

      Finally, the comment “its a normal distribution, but with this one anomaly 23.5 SDs away, that makes it legit” made me die inside.

    • bean says:

      Just realized something else on this. Isn’t WAR inherently per-season? I’d guess it’s something so deeply ingrained that he just didn’t notice it when putting this together. (The bit where Napoleon was so far out should have been a clue, though.)

  32. Athreeren says:

    Katja Grace apparently also wrote Norbert Blum’s erroneous proof of P=NP, with the same co-authors. I haven’t checked whether other similar mistakes appear in the list of most discussed papers.

  33. MT says:

    Re: Napoleon – Win-loss in battles is the wrong metric. It’s a battle-centric viewpoint on winning war, which purports that the best general is the one who forces a climatic battle of annihilation and then wins it. War is usually not won this way, one major exception being the Germans crushing the French at the onset of WWII.

    This book argues that thesis for those interested:

    https://www.amazon.com/Allure-Battle-History-Wars-Have-ebook/dp/B01MR0SLZ3/ref=sr_1_1?s=digital-text&ie=UTF8&qid=1513706837&sr=1-1&keywords=the+allure+of+battle

  34. Buckyballas says:

    If anyone liked that article on Chinese immigrant legends, I recommend the Ken Liu short story “All the Flavors”. Full text can be found here. Short story collection can be purchased here.

  35. P. George Stewart says:

    Isn’t bitter more a sign of medicinal rather than toxic?

    • Nornagest says:

      Same thing — or rather the former is a subset of the latter. Toxic means it’s messing with your biochemistry, medicinal means it’s messing with your biochemistry the way you planned for it to. We use the active ingredient in nightshade to counter nerve gas, and large doses of Tylenol are the most popular (though not the most effective) suicide drug.

      “The dose makes the poison” has been a medical adage since Paracelsus.

  36. skybrian says:

    It’s worth noting that AlphaZero’s performance is considerably less impressive if you measure learning per game played, instead of learning per hour. In other words, it does well because a cluster of machines in a data center can play billions of games extremely cheaply and in parallel.

    If each game played cost a dollar then we would probably still prefer training some humans from scratch rather than AlphaZero from scratch because a fully-trained AlphaZero would cost over two billion dollars.

    This doesn’t matter for domains that are cheap to simulate. In domains where experience actually costs time and money, machine learning from scratch needs to be much more efficient to be practical.

    • Ilya Shpitser says:

      More to the point, the domain needs to be sufficiently formalized where reinforcement learning + a big cluster can give you a solution. Most problems aren’t like that.

      Reinforcement learning works on things that can be phrased as (big versions of) “Wumpus world” from “AI, the Modern Approach.”

    • Douglas Knight says:

      Where do you get 2 billion games from? Multiplying 700k steps by 4k minibatch size? On page 15, it claims 20-40 million training games. I don’t understand the discrepancy.

      Whether 10⁷ or 10⁹, how should we think about this number? Should we compare it to the cost of training an individual? Or should we compare it to the collective of all humans trying to learn the game? I think 10⁷ is impressive by that benchmark. If we magically identified the 1k potentially best chess players with no exposure to the game, brought them together and had them play 40k games each, would that recapitulate the history of chess? go? Or would we need 10k players? 100k? So I think 10⁷ is impressive by that standard. 10⁹ is less impressive, but since it is more like the total number of games played, skipping the step of identifying innate talent, that’s pretty good, too.

      • skybrian says:

        Yeah, I skimmed and did the simple multiplication, but apparently that’s wrong?

        If it cost $1 per game, maybe we wouldn’t use training from scratch at all; we’d have it learn from human games or from playing a smarter game engine.

      • Controls Freak says:

        I think there’s still a bit of a question about what “counts”. Is it games? Moves? Some combination? The fundamental problem is that we can learn from thinking about moves as much as we can learn from making moves – so can AZ. When we think about moves, we’re ‘making’ them in our mind, thinking about how they might play out, and updating our ‘policy network’ in the process. Similarly, AZ was using some of the MCTS results beyond just ‘moves played’ for updating during self-play.

        In general, you can think of the problem as such. If you start an algorithm on the search space of chess, but just let it run and run and run, training all the while, but not being “allowed” to make any moves, it will get smarter (at least, if your training is somewhat stable and such)… but it won’t have “played” any games. On the other end, we could use the barest of resources in the MCTS in order to decide what move it’s going to play, and just blitz out quadrillions of “games”. Neither strategy is as optimal as something in the middle. Allow it to think about more possible moves between “actual moves” (and use that thinking to update the policy network), but make it actually play through a bunch of games, too.

        In the end, I don’t know if there’s much of a way to measure this… for either contestant. We don’t really have good information for how many moves AZ “considered” (and used for training), and it’s clearly impossible to make that same calculation for humans (especially when you consider training outside of “games”, computer-aided training, etc.).

        I think the verdict is still out on how much brute-forcing these types of algorithms are actually doing. We clearly understand how humans create conceptual categories to help avoid having to brute force all the time, but we’re still in the infancy of explainable DL. There’s some reason to believe that these types of NNs are actually creating categories for patterns (they were explicitly built to try do that). One data point is that when the trained AZ policy net plays with the MCTS, it’s still very good, but it’s not superhuman. Of course, this isn’t exactly apples-to-apples, either, because we can’t just strip out “human MCTS” and test ourselves without it.

        • skybrian says:

          I guess the question is whether self-play is more like thinking or more like learning from experience. If you think of it as experience, it seems like learning efficiency should be measured using the amount of experience needed in the denominator. Using wall-clock time for a highly parallel algorithm seems like a rather silly way to measure experience.

          If you think of self-play as more like thinking hard, I’m not sure what goes in the denominator. Maybe joules?

    • Squirrel of Doom says:

      Perhaps a fair point but it almost doesn’t matter if it took four hours or four years.

      The achievement is still absolutely breathtaking, and I don’t know if people thought it was possible before it was done.

    • Calvin says:

      But how is the statistic of learning per games played more relevant to the question of “impact (of similar/future AI) on humanity”? It’s not like future AI will have lower processing capabilities than present. The relevant stat for humans is not learning per game it’s learning per unit of time.

      • DocKaon says:

        I think it’s highly relevant because there are only relatively small number of cases where you can produce nearly an infinite number of perfectly accurate “games” in which to learn from. The acquisition of data appears to be one of the main limiting factors on AI. It has succeeded in cases where there is large amounts of pre-existing data or it’s easy to generate perfect data through an automated process (i.e. a game of chess played against itself is a perfectly accurate model of a game of chess). In the real world, data is limited and frequently incorrect and models are slow, incomplete, and require calibration with real empirical data. It’s important to understand whether data is a fundamental bottleneck which would significantly reduce AI risk. If an AI would require 22 million actual attempts at talking itself out of a box to learn how to manipulate its’ jailers, that is a lot less scary than if it can just sit in it’s box computing for 4 hours to acquire the same skills.

        As some with a background in experimental physics, I’m biased to think even an AI god will have to do experiments to understand the world around them.

    • jchrieture says:

      Quantum simulation is yet another discipline in which machine learning is demonstrating transformational power.

      As evidence of which the Portland-based quantum simulation corporation Schrödinger LLC — continuing its history over the past 27 years of doubling in size every 44 months — is presently advertising more than forty technical positions.

      To continue a theme from two other job-related SSC comments (here and here), these new technologies, markets, investments, and careers (and paychecks) are composing an ongoing transformational shift in humanity’s cognitive “cleaving of reality at the joints” — a shift in which the fundamental joints are appreciated to be affective as contrasted with ratiocinative.

      In essence, AlphaZero has a “good feeling” about the moves that it recommends — “feelings” for which it can offer no coarse-grained / humanly-understandable explanation. Similarly, human brains and quantum density mechanisms alike can be regarded (computationally) as tensor network structures, whose geometric and algebraic structures again elude coarse-grained / humanly-ratiocinative understanding, even as these same structures become ever-more-amendable to computational instantiation (via Google’s TensorFlow language / cloud-hardware environment, for example).

      Michael Nielsen’s free-as-in-freedom on-line textbook Neural Networks and Deep Learning (2017) is a good place to learn about the shared mathematical underpinnings of these new disciplines, enterprises, and careers.

      It was interesting (to me anyway) to read that Nielsen is both a Research Fellow at Y Combinator (which was a founding investor in “Relationship Hero”) and a member of the Steering Committee for the journal Distill; the latter journal being heartily recommended (by me) to the general SSC readership … job-seekers and career-pursuers especially.

  37. romeostevens says:

    The link on commuters objecting to a toll road. This is part of the same class as ‘consumers hate price discrimination’, ‘people resist efficient auctions’, ‘fair price folklore’ etc.

    I think the generative seed for these things is an anti-molochian instinct. People recognize and (rightly!) resist things that collapse values into being uni-dimensional. That this happens below the level of explicit awareness is what causes it to be coopted by political zero-sum narratives of free markets vs equality or whatever other nonsense reigns at the time. (THIS SUNDAY! [value] vs [value] in a no holds barred cage match and only one can win!)

    To generalize: Why does something seem ‘interesting’ over other things? My guess is that it typically is because the person finding it interesting hasn’t yet discovered the seed generating the category of [thing]. So every time another instance pops up, their system 1 says ‘ooh another data point’ and continues crunching on what might an efficient compression for this class.

    • dank says:

      I think you’re on to something here. From the narrow perspective of a driver’s choice on whether to use the HOV lane, they now have strictly better options. In the larger meta-view, now that there is a new revenue stream, the incentives for the agents running the highway have changed. Those changes might mean bad things for the driver down the road.

      For example, if they want to maximize the number of people paying the high fee, they can try to make driving in the other lanes as painful as possible.

  38. jchrieture says:

    Research PhD Opportunities/University College London: UCL Associate Professor Patrick Luyten is accepting PhD students in the next academic year (2017) in the following areas:

    (a) Research on the role of epistemic hypervigilance in borderline personality disorder (BPD).

    This topic is part of a larger series of studies concentrating on epistemic hypervigilance in BPD, and involves testing some key hypotheses concerning epistemic hypervigilance in relation to attachment and mentalizing in a series of experimental studies.

    (b) Embodied mentalizing in patients with functional somatic disorders.

    As part of a broader research programme, this topic will investigate the role and correlates of impairments in embodied mentalizing in patients with chronic somatic complaints.

    Professor Luyten belongs to an active & growing community of cognitive researchers who pursue research topics of broad interest to SSC readers, by evidence-based methods that are broadly compatible with rationalist traditions.

    For the right SSC person(s), here is a concrete opportunity to pursue a medical research career, in one of the world’s most culturally vibrant cities.

    Background readings:

    @article{Kalisch:2015aa, Author = {Kalisch, Raffael
    and M\"uller, Marianne B. and T\"uscher, Oliver},
    Journal = {Behavioral and Brain Sciences}, Title =
    {A conceptual framework for the neurobiological
    study of resilience}, Volume = {38}, Year = {2015}}
     
    @article{Fonagy:2017aa, Author = {Fonagy, Peter and
    Luyten, Patrick and Allison, Elizabeth and Campbell,
    Chloe}, Journal = {Borderline Personal Disord Emot
    Dysregul}, Pages = {11}, Title = {What we have
    changed our minds about: Part 1. Borderline
    personality disorder as a limitation of
    resilience.}, Volume = {4}, Year = {2017}}
     
    @article{Fonagy:2017ab, Author = {Fonagy, Peter and
    Luyten, Patrick and Allison, Elizabeth and Campbell,
    Chloe}, Journal = {Borderline Personal Disord Emot
    Dysregul}, Pages = {9}, Title = {What we have
    changed our minds about: Part 2. Borderline
    personality disorder, epistemic trust and the
    developmental significance of social
    communication.}, Volume = {4}, Year = {2017}}

    • Nornagest says:

      Go away, John.

      • quanta413 says:

        I think this is his most devious post yet. Trying to lure poor unsuspecting undergraduates into more academia with promises of glory.

        • jchrieture says:

          As it turns out, job opportunities abound that relate to cognitive attachment.

          E.g., Patrick Luyten’s articles (above) provide a a neuroscientifically grounded and evidence-tested rational for attachment therapies that differ only inessentially from the attachment therapies that are marketed by the perennial SSC sponsor “Relationship Hero”. whose ad appears at right (on my browser anyway).

          That Relationship Hero is hiring affirms that a need exists — the market is speaking! 🙂

  39. Peter Shenkin says:

    Joseph Altonji agrees that requiring net neutrality is bad, saying “High bandwidth traffic imposes externalities on other users.”. But high-bandwidth traffic is already charged for, because (AFAIK), the content providers pay per byte of traffic they produce and consumers pay more for the higher bandwidth needed to consume high-byte traffic. Why is this mechanism insufficient to control the externalities?

    Please correct me if I am wrong, but what loss of net neutrality is about is allowing backbone providers and ISPs to accept extra money from content providers to favor their traffic over others based on willingness to pay, not on byte count. And to charge end-users extra to receive specific channels, in addition to what they pay for bandwidth.

    It’s not about charging high-bandwidth producers and consumers more per byte produced or consumed. That already takes place in the net-neutral world we have had up to now.

    Again, please correct me if I am wrong.

    • Sniffnoy says:

      This is basically my understanding.

    • Murphy says:

      Don’t forget ISP’s blocking specific protocols like VOIP where they compete with other services they offer like traditional phone services.

      Also IP4 already had a QoS byte specified in RFC’s that could be used to specify type of service needed.

      https://en.wikipedia.org/wiki/Type_of_service

      Unfortunately the makers of an unnamed highly popular proprietary OS thought they were being real smart by having their OS mark everything as top priority, thereby making the whole thing pointless. And now it’s deprecated.

      • Gobbobobble says:

        Unfortunately the makers of an unnamed highly popular proprietary OS thought they were being real smart by having their OS mark everything as top priority, thereby making the whole thing pointless. And now it’s deprecated.

        Sounds like they did us a favor. If you need special priority, build a dedicated private network.

        • Mark Atwood says:

          If I was building an ISP from scratch, I would:

          * charge by the packet, and also have a congestion charge (guarantee that no more than 2% of the time buckets will be congested)
          * honor priority QoS tags, but charge extra for those packets
          * charge more for IPv4 packets, less for IPv6 packets
          * charge less for packets correctly using flow tags
          * charge less for packets from heavy sources (media streams, CDNs, commercial VPNs, cloud service providers) that will direct connect to one of my core switches. Absolutely require that packets switched over those ports be IPv6 and flowtagged.
          * Other than that port-connection option, no paid “fast lanes”, no “revenue sharing”.
          * provide a on demand xkcd/806-compliant tech support tier, but there will be at $1000 charge for misusing it
          * unless the customer fills out a “I know what I am doing, really” form, I will:
          ** block SSL/TLS/VPN with broken encryption
          ** block port 25
          ** block ads and ad beacons
          ** force https redirect
          ** block various other network bogosity
          * market my latency and dwell numbers, not just raw bits per second

          I’m told that customers prefer a known monthly charge instead of variable cost recovery charge. Too bad. I’m sure that people would like to buy a “$100 a month take all you need” plan from their grocery store or their gasoline station too.

          • Brad says:

            If you can pull customers from across the world for this kind of niche product you can make a nice little business out of it (e.g. Colin Percival’s tarsnap). But an ISP is inherently geographically concentrated and needs to sign up a decent percentage of the customers in the covered area in order to pay the fixed costs. So this is not a good business model.

          • Mark Atwood says:

            There are multiple runs at building intelligently run high quality ISPs, picking off geographic clusters. They don’t fail because the customers don’t like them. They fail because Comcast and AT&T exercises lawfare against them.

            I know of one well run WISP that failed because the local incumbent ISP kept “accidentally” bumping their backhaul antennas, kept “accidentally” damaging the cables into their cabinets, and then finally just cowed their hardware suppliers into no longer shipping them parts.

            Comcast well earned their “worst company in America” awards.

          • The Nybbler says:

            Unless you’re a monopolist, you’re never going to get anyone to sign onto that. The Bastard Operator From Hell keeps his job because the people he serves aren’t the same as the people he services; an independent business doesn’t have that option.

          • Mark Atwood says:

            I have *been* the BOFH. I would rather be a luser under the BOFH, than a supplicant to Comcast.

            In this hypothetical ISP I would run, I would offer Simon Travaglia right of first refusal to be the Director of Operations.

          • actinide meta says:

            Even in a world where people are happy with unpredictable variable pricing, I think there’s a problem with this economic model. You can only bill your customers, not your upstream. So if I understand correctly you are charging people variable prices for what they receive as well as what they send. But they don’t really control what they receive! This means your scheme is not actually incentive compatible:

            * Some jerk can run up a big bill for you by sending you lots of high priority packets you don’t want
            * It’s in the selfish interest of websites and advertisers to send everything out with QoS tags; it costs the user lots of money, but they don’t have the tools to detect this. Even if you tried to give the customer a billing report breaking their costs down by domain name or something, I doubt you could make it correspond to what the user sees as the website they were visiting (e.g. you will just see that all the traffic comes from Cloudflare or something).

            I think it would be great to get to a world where the Internet supports “economic” QoS. It would be hidden from most consumers behind some simpler pricing model, but people with serious bandwidth needs would deal with it directly. But I’m pretty sure it needs new technology beyond packet level QoS bits, and/or for the variable costs to be paid at the “sophisticated” (i.e. the business) end of the connection rather than the consumer end.

          • Jiro says:

            provide a on demand xkcd/806-compliant tech support tier, but there will be at $1000 charge for misusing it

            This basically turns any service call into a gamble where the consumer gambles that you don’t screw up the “determine who’s responsible” process, and he loses an extra $1000 if you do screw up.

          • Gobbobobble says:

            You can only bill your customers, not your upstream.

            Why not?

          • Mark Atwood says:

            This basically turns any service call into a gamble where the consumer gambles that you don’t screw up the “determine who’s responsible” process, and he loses an extra $1000 if you do screw up.

            Only if you choose xkcd/806-compatible callflow. “Misusing” it is not “root cause analysis determined that the misconfiguration was in the CPE”, misusing it is “you didnt already know what CPE means, and you were not able to talk to level-three support as a technical peer”.

            There are businesses and industries were I would take that $1000 risk, and take the xkcd/806 call flow, and there are many many businesses and industries where I would not, and would instead choose the standard CSR callflow.

            That particular wishlist item may be becoming moot anyway, my last half dozen times I’ve had to call an 800 customer support line, my experience was almost xkcd/806 compliant anyway. More industries may be wising up to the actual bottomline cost of bad CSR, or maybe I’ve just gotten really good at getting to a L3 CSR without noticing.

          • Murphy says:

            The problem is the number of highly competent people who still skip the “turn it off and on again” step.

            hanging out with an acquaintance who’s hobbies include making his own system-d and assembly programming to tune devices to tune performance to integrated devices. Far more technically competent than myself. The other day he was knocking his head against trying to get a system to connect to the printer and he’d diagnosed the problem as a DHCP problem with the printer losing it’s registration etc etc wireshark etc etc

            So I turn the printer off and on again and away it goes because he’d skipped that step.

            Lots of our kind do. It takes mental effort to force yourself to when you can see a more complex problem that you’re fully capable of comprehending.

          • Mark Atwood says:

            “Turning it off and back on again” working just means that there is stupid bug in the software of that device, or in the remote services that that device is connecting to, probably either a setup race condition, a mistake in the implementation of a protocol state machine, or a resource leak, or a smashed stack or corrupted heap.

            Power cycling is not a fix. It’s a temporary workaround, until someone competent can figure out what the actual problem is. Sometimes, unfortunately, that time never comes.

            Going back to the original xkcd/806, the end user probably could have just kept power cycling until they randomly got assigned to another DSLAM port, and that’s what a normal CSR would have made them do. The 806-compliant CSR, on the other hand, actually looked at the problem and the systems involved, and fixed the real underlying issue.

          • poignardazur says:

            The reason most ISPs don’t charge you by packet sent is the same reason most restaurant don’t charge you for condiments or for using their toilets: it’s not the expensive part of their service, they don’t want the extra accounting, and the consumers don’t want the hassle.

            Usually ISPs just focus on people who are obviously trying to game the system (in the metaphor, people who try to bring 30kg of condiments home).

          • Evan Þ says:

            @Mark, very good point, but power-cycling can get the user back to an acceptable state much faster. Plus, if there isn’t adequate logging, you might never be able to find what was the actual root cause.

        • poignardazur says:

          That’s ridiculous. Do you really think Blizzard is going to spend billions of dollars making their own worldwide private network just so you can play Overwatch with no lag? I’m sure people living in Madagascar will be happy not being able to use Skype until Microsoft gets around to extend their private network to the island.

          In practice, different services need different treatment. Video chat and online games need high-priority service because a 0.1s delay is noticeable. Streaming and file downloading can get away with some buffering, as long as it doesn’t impact the total download time.

          • Gobbobobble says:

            If zero lag is absolutely critical, then that’s on them, yes. Overwatch seems to do fine without special privileges currently. Not much actually needs a data motorcade, they just want it if they can get it at someone else’s expense.

            Besides, the original example was hospital data. Bringing in online games and video/voice systems is further evidence to me that dismantling NN is just going to lead to constant strife over who belongs where on the totem pole.

            Fuck that, make it an even playing field and if that’s not good enough they can make their own.

          • poignardazur says:

            Again, no they can’t.

            Forgive me for being presumptuous, but I’m guessing you’re visualizing the Internet as something that exists in the USA and maybe Europe too, but Internet covers the Earth, and the Earth is HUGE. I can’t begin to estimate how expensive it is, but is has to be trillions of dollars at least. Transoceanic cables aren’t cheap.

            Right now, I’m in Korea, yet I’m able to have stable-ish visioconferences every week during rush hours with my colleagues in China, Europe, and Canada, using a startup-developed service called appear.in. This wouldn’t be possible if this startup had to build / rent a network completely separate from the internet yet somehow connecting all of us with optic fiber across two continents one ocean and one sea.

            And our reception would be way less smooth if our respective ISPs didn’t do some QoS sorting to prioritize our packets over torrent/download/http packets. I’m not talking theoretically, this is stuff every ISPs does, and needs to do to provide the best quality-per-dollar-spent.

            When I’m downloading a movie, I don’t care if my download takes 30 seconds longer because other services are prioritized. But when I’m having a professional conference across the world, I very much do care that I’m able to understand more than 1/10th of the words my teammates are saying.

  40. Lillian says:

    You know, thinking about it, the link between transsexuality and autism fits very well with my experience interacting, since most of the transgirls i’ve known are huge dorks. Incredibly adorable* dorks, which is why i kept falling into relationships with them, but dorks nonetheless. Like just a few days ago i was commenting on this to my Boyfriend, because the son of one of mom’s friends came out as trans, and apparently dating a few transwomen makes me qualified to deal with it. So fine, i ask for some contact info, get a Skype name, and just from that i already know: giant dork. My sample of transmen is smaller, but one out of three was in fact diagnosed with Asperger’s. The other two were a very stereotypical non-effeminate gay man, and a charismatic pretty-boy who left a trail of broken hearts in his wake. So yeah, i’m overall rather unsurprised.

    *(The ones who pass that is. The ones who don’t seem to be consistently annoying and moderately creepy. No idea which way the cauality arrows points though, maybe i’m just a close-minded gender essentialist or something.)

    • Le Maistre Chat says:

      *(The ones who pass that is. The ones who don’t seem to be consistently annoying and moderately creepy. No idea which way the cauality arrows points though, maybe i’m just a close-minded gender essentialist or something.)

      I see this type in the same places I encounter people who identify as poly. They always hit on women, so I think “There goes a yellow lizardman.”

      (Blue lizardmen forever. <3)

      • powerfuller says:

        I would have guessed a blue lizardman was somebody who is sad because nobody ever believes his sincerely answered survey responses… but yeah, go Blues!

      • Lillian says:

        Blue lizardmen forever. <3

        Heh, we may share a body type, a nerdy hang-out, and an absurdly high verbal intelligence, but we definitely do note share a taste in men. I like me some orange lizardmen, especially if the kind who doesn’t mind me seducing the rest of the harem. Perhaps my being enthusiastically bisexual has something to do with it.

    • Deiseach says:

      Using Skype makes you a giant dork? There must be an awful lot of dorks out there, so!

      (Note: not a Skype user) 🙂

    • Ratte says:

      The ones who pass that is. The ones who don’t seem to be consistently annoying and moderately creepy. No idea which way the causality arrows points though, maybe i’m just a close-minded gender essentialist or something.

      Obvious explanation: passing is generally pretty difficult and requires both substantial effort and sensitivity to how others perceive you; people who can’t manage any of that are going to evince other social defects as well. Also there’s the confounder that people tend to view the negative quirks of attractive people in a more positive light.

      FWIW, I have a lot of transfolk (dozen+) in my friend circle and every single one is a nerd of some stripe, most are into kink (also a stereotype), and about half of the ones I know well enough to be able to comment are into diapers at some level (another really bizarre stereotype I’ve heard from several people independently). I realize that I can’t draw inferences from my immediate circle about the community at large, but stuff like this feeds the nasty little imp that whispers things that would get me banned from progressive spaces.

      • Le Maistre Chat says:

        FWIW, I have a lot of transfolk (dozen+) in my friend circle… and about half of the ones I know well enough to be able to comment are into diapers at some level (another really bizarre stereotype I’ve heard from several people independently).

        … (retches)

        stuff like this feeds the nasty little imp that whispers things that would get me banned from progressive spaces.

        My nasty little imp whispers “gender dysphoria is a social construct whose prevalence varies with which sex it’s more inconvenient to be. It being harder to be a man is why estimates of people living as the other sex have increased from 1 in 50,000 to 1 in 500 in the past 25 years.”

        • Ratte says:

          Pretty much, yeah. I can’t deny that many of them seem better off, however.

        • ManyCookies says:

          It being harder to be a man is why estimates of people living as the other sex have increased from 1 in 50,000 to 1 in 500 in the past 25 years.”

          Object level “It’s much harder being a man” aside, the trans-woman:trans-men ratio seems relatively steady from a Wikipedia eyeball; the 1994 study shows a 3:1 ratio, and more recent estimates are in the same ballpark. If the increase in trans prevalence was because of male hardship, we’d expect a huge ballooning of that ratio. And I guess you can make a social construct explanation for the female side, but at that point “Being trans is (relatively) way less stigmatized” seems like the much simpler explanation.

          • Le Maistre Chat says:

            OK, that’s interesting. If the ratio hasn’t become increasingly skewed as the phenomenon has become more visible, I need another explanation.
            Olyslager and Conway’s methodology looks sketchy. Prevalence may still be 1:10,000 or less like it was 25 years ago.

      • Lillian says:

        While i’ve definitely seen the higher than average interest in kink, this is the first time i’ve heard of any transfolks being into diapers specifically, let alone that it’s a stereotype. Like, i would expect some would be into it given the higher interest in kink, but not that it would be common enough to be a specific stereotype. Though there was this transgirl i knew who was into urine and bestiality , and rather distressingly she would not shut up about it. Remember how i said the ones who don’t pass tend to be off-putting? She had a man’s head on a woman’s body, which made her very presence somewhat unpleasant. At least she had the decency and self-awareness to wear loose clothes in public and not insist that strangers gender her as female, nor make use of the ladies’ room. Otherwise she would have been intolerable, instead of just kind of annoying.

        And yeah, while it’s easy to think that passing means more effort, because much of the time it does, fact is some people are just ugly. A transwoman who has very masculine features is SOL unless she can afford some seriously expensive surgery. Given the human tendency to be less forgiving of unattractive people, i feel like i cannot be completely confident that non-passable transwomen are particularly obnoxious, when it could very easily be my own biases at play. Nonetheless, a couple of the transgirls i dated pointed out this phenomenon themselves, and would derisively refer to the non-passable ones as “lumberjacks”, which is meaner than i’m willing to be. So whether reality or bias, it’s not just cis-gendered individuals who see this pattern.

  41. eucalculia says:

    Slightly confused by the economists who are 10/10 confident that they are uncertain on the outcome of net neutrality… is that more or less certain than being 1/10 confident that you’re uncertain.

    • suntzuanime says:

      My guess is that 10/10 confidence in uncertain means that you’re pretty sure we don’t have the necessary facts to answer the question, and 1/10 means you might well have the facts but just not know the right model because this isn’t your area of expertise.

  42. JPNunez says:

    Machines with really high IQ support converting the whole earth into paperclips.

    Why would normies with low IQ oppose this glorious plan?

  43. Prof. Quincy Adams Wagstaff says:

    “Some rare good news:…”

    Oh try harder.

  44. sty_silver says:

    So to just throw this question out there…

    is there any reason why we should care what the majority of economists think? Does majority view of economists have a good track record of correlating with truth?

    • John Schilling says:

      We can’t know what the majority of economists think, because “economists” is an ill-defined, fuzzy-bounded and inhomogeneous subset of humanity that doesn’t even come with a Directory of Economists from which we can poll.

      Which is why this sort of thing always uses a poll or survey of the world’s top economists, where “top” can be defined in whatever way best suits the interest of whoever is paying for the poll. Occasionally that will be someone interested in an accurate forecast and willing to share it with you even if you aren’t sharing the cost, but what sort of economist would participate in something like that?

  45. The Nybbler says:

    Altman’s “clarification” is backpedaling to wishy-washy uselesnes.

    I didn’t mean that we need to tolerate brilliant homophobic jerks in the lab so that we can have scientific progress.

    Sorry, Sam. But Silicon Valley itself is in large part based on the work, in a lab, of a brilliant racist jerk (he was probably anti-gay as well, but I don’t know). So that’s what you should have meant.

    • Nornagest says:

      Yeah, I was distinctly unimpressed by that. What does it even mean to tolerate contrarian ideas as a society if tarring and feathering their carriers at a personal level is still totally okay?

    • Nick says:

      I don’t mind clarifying the argument “we need to allow free speech because sometimes society is wrong,”* since I think that’s a fair argument for free speech, and a value-neutral one at that—I don’t think anyone today is perfectly happy with exactly how the world is, since conservatives want it more conservative in various ways and progressives want it more progressive in various ways. That’s a line of reasoning which can appeal to anyone who has something potentially unpopular to say.

      What I do mind is framing this as “But just think of all the important progressive causes which aren’t currently popular!” Sorry, Sam, but I’m unpopular enough already; I’d rather not help accelerate this process. Do we care about protecting free speech generally here, or do we just care about protecting the most extreme ideas coming out of American progressives? Is the Overton Window only allowed to move left?

      *To clarify, neither of the things I put in quotation marks here are actual quotes of Sam or anyone else.

      • Le Maistre Chat says:

        Do we care about protecting free speech generally here, or do we just care about protecting the most extreme ideas coming out of American progressives? Is the Overton Window only allowed to move left?

        That’s no window; it’s a R’yleh Door. You know how Cthulhu only wants to swim left from his house.

      • Brad says:

        I’ll say the same thing here I said when similar responses were made to Scott’s post. You are the intended beneficiary but the not the intended audience. And when you attack them for saying the things they need to say in order for the intended audience to listen you are only hurting yourself.

        • Nick says:

          I considered this objection before posting, but decided to post anyway because a) Sam Altman’s critics are unlikely to ever read this, and b) I don’t think making his message more palatable requires retreating and reframing to the extent that he did, so my disagreement is a matter of degree here. That aside, I really didn’t intend for this to be an attack—I’m criticizing it, to be sure, but I tried here as any other time to be respectful to Sam. If I could have been kinder then I’m sorry and I’ll try to be in the future, but I’m outside the edit window now.

          ETA: apology.

        • quanta413 says:

          I know not exactly what you are talking about Brad, but a tangent if you will.

          I think that arguing how free speech will help “X Object level thing” to a group of people who think they are winning (or think they will win soon) won’t convince anyone of anything because they’ll say (WARNING: CARICATURE INCOMING) “I already don’t silence good people; I only silence bad people who want to do bad things. Why are you saying I shouldn’t keep silencing the bad people? The obvious solution is to allow ‘good debate’ and ban ‘bad debate’.” Notably, some of the most strident SJWs (“hate speech isn’t free speech”) and people as far right as the dread Jim (“free speech for us but not for them”) agree on this point, they just disagree on who is bad and who is good. And in a certain sense, their argument is correct. You don’t have legal rights to engage in violent threats like telling someone “tomorrow I will drive to your house and murder your whole family”. And I’m willing to do the totally not brave thing and agree with the consensus that this judgement is correct.

          Sam himself even acknowledges this, but doesn’t really give any coherent argument for why the line should be drawn somewhere between “Genetic Engineering to enhance whatever” being ok and “tomorrow I will drive to your house and murder your family” not being ok instead of both statements not being ok. And that’s easy mode with his audience. He’s not trying to openly argue that it’d be acceptable to say “Build the wall.”

          On the flip side, when a group is losing or downtrodden enough (but aren’t in immediate existential danger like a war) that the only plausible path to their goal involves persuading people they are right, it can be pretty easy to get them to endorse the abstract principle of free speech.

  46. Jaskologist says:

    For those of you who have been watching, US life expectancy fell for the second year in a row in 2016. Looks like they’re blaming opioids for it; the previous year they didn’t seem to know why, and mortality had increased for a number of caused.

    This continues a change from 2012, when life expectancy stopped rising, and sat still for a while before it began falling. I think at this point we can say that we’re in a new (bad) trend.

  47. Careless says:

    So I’m seeing 4 helipads on the hotel with presumably two more on the other side, but I see no possible way of getting off them short of parachuting

  48. apollocarmb says:

    I gave a wrong answer to a question and only now just realised. Frustrating. I guess I shouldn’t have rushed.

  49. quintopia1 says:

    Do people seeing transportation policy never pay attention to the news?

    It hasn’t been but a few years since the massive outrage over the tolls set by GA DOT on the I-85 HOT lanes… and they were reaching much less than 50 dollars at peak. Afterwards, the Washington Post said “Virginia transport policy makers should pay attention to what happened in GA.”

    https://www.washingtonpost.com/news/dr-gridlock/wp/2015/04/02/atlanta-hot-lanes-hold-lessons-for-d-c-region/?utm_term=.e6f99e4c5d75

    Looks like they were pretty much ignored.

  50. JPNunez says:

    The whole Sam Altman post is ridiculous. He wants the ability to say anything without losing social status, and yet the moment someone tells him the wacky idea that prolonging human life may not be a good idea, he:

    -writes a long blog post about how he was told such a wacky idea
    -shames the whole place where it was told such a wacky idea by comparing it with an undemocratic state
    -threatens with leaving the place

    I think Sam Altman is the one who cannot tolerate free speech and is an hypocrite.

  51. anonymousskimmer says:

    Why would those of higher IQ be more open to “free speech” (however they mean that) than those of lower IQ?

    Would it be because those of lower IQ see a bunch of arguments which, to the extent they understand them, seem like sophistry and objectifying (which, to be honest, is how I see a lot of so-called high-IQ arguments from the likes of poly-sci ‘policy wonks’). Arguments which the lower-IQ cannot argue against because they’ll be strawmanned or ad-hominened as not even understanding the argument and thus unable to object to the repercussions which they of the lower IQs see as likely to happen.

    Free speech doesn’t come with the power to implement or influence policy based on that speech, however those with higher IQs are more likely (as individuals) to have access to this power than those of lower IQ. And the lower-IQ know this.

    The closedness toward free speech by the lower-IQ may really be aversion to the actions taken based on that speech than a dislike of free speech itself. And they have quite real reasons for this wariness.

    We do have evidence that those who are better off have more difficulty empathizing with those who are worse off than the reverse.

  52. Morgan says:

    Altman is using “heresy” in a way that really confuses the issue, especially with the Galileo reference.

    Galileo was condemned by the Catholic Church in its capacity as body with genuine temporal legal power. Heresy in that sense comes under the same class of “government intervention” that he’s referring to in referencing the restrictions placed on Newton’s alchemical studies (he also fails to mention that the ‘devaluation of gold’ was only part of the issue, and alchemy was also restricted to try and minimize charlatans defrauding people, which…seems like a perfectly reasonable thing for the legal system to try and restrict, but that’s a side issue).

    “Heresy” in the sense he’s using it in the rest of the post, and as refers to the actual Bay Area system, is not in the same legal sense. The consequences are informal, social, and not part of the legal system. There are no legal consequences for saying “meat eating is ethically wrong”, “we can and should develop techniques for radically extending the human lifespan”, “the earth is flat”, “evolution is a thing”, “evolution isn’t a thing”, “there is no God”, “God exists and requires that we worship him”, “all your problems can be solved with kale”, or whatever else. Depending on where you say those things, there may well be social consequences, and that’s what he’s describing.

    The conflation is easily made because “heresy” and “blasphemy” even in the strict religious sense no longer carry the same legal weight for most of us as they did for Galileo. Religious law isn’t temporal law, and the consequences for committing a religious offence in a secular country are…social stigma, essentially. Committing literal heresy today has no legal effect. That was really not true in the Renaissance.

    There being a powerful social stigma attached to the expression of controversial opinions might be a problem, but it certainly isn’t the same problem as legal limitation on speech, and equating the two does no-one any services.

    Echo chambers stifling innovation via soft social pressure is a VERY DIFFERENT problem from Trump banning the CDC from using the word “fetus”, and the fact that Altman’s “Clarification” conflates the two suggests he doesn’t fully appreciate that.

    And the lack of clarity around that distinction may well be making the former problem worse. The example about vegetarianism cited in the body of the “Clarification” is probably illustrative of a lot of what’s going on there. Responding “banning meat consumption would infringe my rights” to “meat-eating is unethical” implies that “this is unethical” == “this should be illegal”, and I think that reflects a broader tendency to confuse discouraging something with advocating banning it.

    The tolerant free-speech society Altman envisions is only possible if we’re able to clearly discuss our ethical concerns without leaping to “and so it shouldn’t be allowed”, and equally, without hearing “I have some ethical concerns about this” as meaning “and so it shouldn’t be allowed”. We need to be able to distinguish “things you shouldn’t do” from “things you shouldn’t be allowed to do”, partly because free speech is all about that distinction (You shouldn’t say the earth is flat, because that’s stupid. You should be allowed to say the earth is flat, because you have a right to free speech), and partly because we need to be able to discuss complex ethical questions without advocating banning things except when we actually mean it.

    For that to be possible, the first step is to draw that line really sharply in any discussion of free speech, and reiterate it as often as possible.

    My main concern with Altman is that he’s not doing that, and that’s actively counterproductive, because it reinforces the idea that any means of disagreeing with someone’s ideas (telling them it’s stupid, telling them it sounds unethical, telling them you don’t personally want to discuss it with them, telling them you think less of them for holding that position) is equivalent to banning specific forms of speech.
    It’s easier both to debate the nuances of free speech rights, and to use those rights to discuss controversial issues, if that distinction is plain.