Friday, December 19, 2008


Super awesome day Wednessday. I spoke with professor Bickle (Chair of UC's Philosophy department) about doing some serious study into cognitive science, and he was very enthusiastic and encouraging. he gave me the names of some people on campus who I could contact that are more deeply involved with the things I'm looking to study (neural networks, reinforcement learning, modeling biological processes, bioinformatics...). He also mentioned that he'd like to get me connected with the head of the Budapest program because he is visiting Indiana University for six months; now theres some serendipity.

I was all charged and energized after the meeting, and I went to Babba Budans to continue studying. Had a good conversation with the barista about the roles humans set themselves up to play and what it means to be free of them.

Then I finished up reading about Eric Baum's Hayek program, it turns out he's already implemented much of what I thought was an awesome and revolutionary idea in terms of using economic forces to act on "societies of mind." He even created "investor agents" to attempt meta-learning by adjusting the populations of agents and mutating them. However, there was no discussion of the "efficient frontier" of agents; the groups of agents that are maximally effective in the face of uncertainty for a given level of risk. The concept may have been overlooked, or it might be implicit in what Baum actually did.

An interesting question is what "risk" means in this context. It might be tied to the amount of computer time it takes to process, it might have to do with whether or not the program finishes at all. I'm also wondering if the populations of agents that Baum's evolutionary process represent such an efficient frontier, without having been explicitly applying that concept. It may be that his approach is another way of finding efficient frontiers in finance; that would be worth researching in itself.

So after reading all that ecstatically, my energy was spent and I fell asleep in my chair at the coffee shop for a half hour. Woke up in time to run home, play guitar for an hour, change, and go to the Southgate house for the Faint show.

The Faint was awesome, I'm really glad Jude got me into them a while back. They were the perfect band to see live and at the Southgate house; they had an awesome light show, their stage presence was exciting, and the music was perfect. Exactly the right mix of electronic and... rock? However you'd like to describe it, it blows your eardrums in a pleaseing way.

I find their website hilarious, by the way. Its a techie band, and they choose to make their home page nothing but a string of their tour date; I appreciate the irony.

Yesterday I went up to dads and put my clubman handlebars on my bike, check out the photo. Actually maybe don't, its a really badly shot and blurry photo. Anyway for posterity's sake, I'll include it.

It went smoothly, except for the fact that the wires on the headlight switch are supposed to run through the handlebars, and the new ones have no such holes. Dad and I ground the casing to make space for the wires to poke through; I felt a little bad about doing that since I believe they're original equipment (and therefore relatively rare, being from '65), but the damage is small and hardly noticable. When I got it all back together and started it (after some difficulty), the engine ran at about 2/3rds full throttle (eg: really fast and loud) at idle, so I was afraid I screwed up the throttle cable and was gonna have to redo it all. Turns out that adjusting the throttle cable is super easy, and I fixed it in a jiffy. It was hard to resist wanting to ride it immediately, but the thought of ice was enough to give me pause. Next thing to do is hack off the back of the seat, try to fix the tank, and give it some absurd paint job.

Regarding Fiction, computers, and politics; I just made the connection between Michael Crichton's remarks (echoing Feynman's) on current-day scientists and policy-makers reliance on computer models for data and an old Issac Assimov story called "The Machine that Won the War". The idea is that some big war's just been finished, thanks to this great computer's abilities. The three protagonists are sitting around discussing this, and each has some reservations about the succesfulness of the comptuer. The first guy says "I can't understand how the computer could have made good descisions at all! The reports I was getting from the feild were so spotty and contradictory that I had to choose what was valid enough to feed into the computer, and half the time I just ended up guessing!.

The second guy, the comptuer operator says "Ha! thats funny, because the data you were feeding into the computer was so bad that roughly half the time I had to interpolate it to fill in the gaps just so the computer would spit out a descision. I felt bad sending these to the feild, but I had no other choice."

They both regard the third guy guiltily, who was the admiral in the feild receving these garbage reccomendations. He grins knowingly shakes his head. He says "The computer's recommendations were so bad that I completely disregarded them. But the war was decided by a computer, in fact, the simplist possible kind..." and he flips a coin.

I'm also remined of some of the things that Eric Baum's posted on his website on the subject. I'm likewise convinced of its hoaxfulness, and it strains the credibility of any knowledgable who takes it seriously. I suppose that the general populace can be forgiven since its been presented to them by authority figures as fact for so long... maybe not forgiven, but understood.

Matt Peterson had some great thoughts on Ruroni Kenshin, and related to his review of Marcuse's "One Dimensional Man" (the review is also a great peice of thought). His thoughts on Kenshin and how it reminds him of both Marcuse and Atlas Shrugged:
"Being part of a movement means either following orders or being a manager. Kenshin doesn't want to give commands because he knows people should make their own decisions. And he doesn't want to senselessly kill any more. Movements to topple irrational and ineffective hierarchies become hierarchies themselves."

For some reason, I've always struggled to verbalize what I think is so great about Kenshin, but I think that captures it pretty well. Kenshin is a powerful individual, and he chooses to use his power for only his own ends. He refuses to be the tool of others. At the same time, he deeply respects other's lives and declines to use anyone as his tool. I think that describes the ideal ethical character.

No comments: