Report comment

You read one thing you have no idea the context of, and your response is a petty denunciation with a posing attempt to anchor the conversation to what you and your friends like, and in doing so hopefully invoke the vestiges of some abject insecurity that the effigy for your posing, who is in fact a real person, might have had abused into them.

You and your friends can stay the fuck away from me.

Report comment

Added: Actually, your suggestion is equivalent, if the strength of the baseline connection is strong enough so that when you subtract off the negative connection, it remains positive. And it reorders the top eigenvalues. But if you want more than 2 clusters, you have to take several eigenvalues, anyhow.

Report comment

Report comment

I don’t know anything about neural nets, so maybe this system isn’t actually a neural net, but whatever it is I’m thinking of, it’s a structure where eventually the three nodes reach some kind of equilibrium. If we start with someone liking Israel and Chomsky, but not Palestine, then either that’s going to shift a little bit towards liking Palestine, or shift a little bit towards disliking Chomsky.

Use a Markov chain!

Basically, imagine a “traveler” on each node. Whenever the traveler is on a node, it randomly travels to another node. The probability of it going from Noam Chomsky to Palestine is Noam Chomsky’s support for Palestine.

Then, your own support for a node is the percentage of time that the traveler spends on it. Or rather, relative support is the ratio of time. If you have a lot of nodes supporting Israel and each other, the traveler will sort of get caught i them and spend a lot of time on Israel.

However there’s no notion of initial support–the time the traveler spends is completely determined by the links themselves, not your initial biases. And there’s no notion of “inhibition” comparable to what you see in biological neural networks, really.

But there is this concept of an equilibrium emerging from a lot of support-links. The equilibrium is that the average time it spends on a node approaches a defined limit as time approaches infinity. Basically the Law of Large Numbers, despite the lack of probabilistic independence. Those time-averages that it converges too take into account all the conflicting support links but are still static, unchanging numbers.

My use of a Markov chain for this is inspired by Google’s PageRank algorithm, where the traveler moves across links between websites, and the fraction of time it spends on a website is its importance, which is balanced with relevance in ranking results. Maybe we can get a concept of inhibition from there, because in PageRank the traveler also has a probability of going to a random, not necessarily linked other page. So there’s always this baseline probability of going to an unsupported node. Maybe “inhibition” can be a probability below baseline of going from Noam Chomsky to Israel?

Report comment

Did the Republicans make substantial shifts in policy that made this no longer operable?

Sadly, I don’t think people wised up to the fact that the Republicans aren’t emotively anti-woman, but simply have religious beliefs that as policy disproportionately affect women. There were high profile gaffs regarding women by Republicans in 2012, mostly involving bad word choice but also some ignorance. There just don’t seem to have been any gaffs nearly as juicy this time around as, say, Todd Akin’s. Either way, Yes, there’s a shelf-life. Outrage has to be served fresh for maximum tribal enjoyment.

Report comment

Report comment

There’s some demonization (although that’s perhaps too strong a word) of people who are perceived as inherently opposed to Gray ideals, especially religious people.

Report comment

Report comment

The thing *I* like most about the greys is that they tend to be more honest and sensible in what they say. That, too, is a result of being small and powerless.

(If grey politicians and pundits had much chance of actually WINNING ELECTIONS or influencing policy, I’m sure they’d start falsifying preferences to the degree the other tribes do.)

Report comment