Emile

Karma: 10,769

Emile 21 Jul 2008 12:56 UTC
4 points
on: Should We Ban Physics?
Heh, somedbody’s been reading Unicorn Jelly …

Emile 8 Oct 2008 12:03 UTC
3 points
on: Make an Extraordinary Effort
Hmm, it would be interesting to try to develop a video game aimed at teaching rationality; I’m sure it would be possible to make a game world with the kind of baffling chaotic problem Eliezer talks about. It could be presented as a world with weird physical laws, in which you perform experiements … but it could also be something completely different, as long as the baffling element is there.

There are some aspects that are hard to reproduce in a game, such as the fact that you don’t even know whether a solution exists and/or how much effort is required, or the social pressure to do things a particular way … but then there are plenty of ways game have of teaching you things that are subtly wrong...

Anyway, even if it’s not possible to make a game that makes you really challenge a baffling chaotic problem, it’s probably still possible to make one that can teach you quite a bit about rationality.

(by the way, I’m a game programmer / game designer, so I’ll start taking notes on what could be possible, what’s already been done etc.) (I already know of a game that gave you “incorrect” feedback on the rules—“don’t do this or you’ll die” where in fact, you don’t always die if you don’t)

Emile 13 Oct 2008 17:15 UTC
4 points
on: The Ritual
I enjoy the beisutsukai posts too, it’s nice to see similar ideas presented under different lights; plus it’s a nice change of tone, it gives a bit of variety.

Emile 15 Oct 2008 10:42 UTC
2 points
on: Ends Don’t Justify Means (Among Humans)
I wonder where this is leading … 1) Morality is a complex computation, that seems to involve a bunch of somewhat independent concerns 2) Some concerns of human morality may not need to apply to AI

So it seems that building friendly AI involves not only correctly building (human) morality, but figuring out which parts don’t need to apply to an AI that doesn’t have the same flaws.

Emile 17 Oct 2008 8:11 UTC
5 points
on: Traditional Capitalist Values
“It smells like Eliezer is trying to refute a strawman.”

PK: Elizier isn’t really refuting anything, he’s just saying that personally he’s not sure this is the best system, without going into too much detail as to why he thinks that.

Emile 25 Oct 2008 18:02 UTC
1 point
on: Belief in Intelligence
I wouldn’t count gravity (or the pinball machine) as an optimization process, and I think Elizier would say the same, only better. Once you start saying that gravity, gas expansion etc. are optimization process, it kinda sounds like everything is an optimization process, which isn’t very useful.

The pinball machine question is still usefuln because it helps refine the concept. I’d say the difference between the pinball machine and Kasparov is that once you know the probability distributions for every individual “choice” of the machine (whether a ball will bounce left or right, etc.), you know all about the machine and can use those distributions to prove the ball will reach the bottom. (If it’s a bit hard to imagine for a pinball machine, so imagine a simplified model with a bunch of tubes going around, and some “random” nodes where the ball may fall left or right—like a pinball or patchinko machine, but with a countable number of paths)

Unlike the pinball machine, Having a probability distribution for each of Kasparov’s choices isn’t enough to predict the result of the Kasparov-Elizier game.

You need to know that Kasparov is trying to win to predict the result.

There’s no equivalent knowledge for the Pinball machine—having a model of each of it’s possible “choices” is enough.

Emile 25 Oct 2008 18:06 UTC
1 point
on: Belief in Intelligence
(damn, it’s Pachinko, not patchinko)

You still need to know about gravity to make local predictions about the path of the ball in the pinball machine, but unlike for Kasparov, you don’t need to think at all about future or final states.

Emile 25 Oct 2008 18:28 UTC
1 point
on: Belief in Intelligence
(Damn, it’s Eliezer, not Elizier. Sorry!)

Emile 27 Oct 2008 9:45 UTC
−1 points
on: Aiming at the Target
Tim Tyler: Stars are optimisation processes. They maximise entropy production.

No, at least, not by the definition of “optimisation processes” that Eliezer uses. Stars don’t plan to maximize entropy productions, gravity doesn’t plot a course down a hill so that a ball reaches the lowest point. Both can get stuck in a local maxima. Reread the post about Kasparov, and try to come up with a corresponding story about stars. Elizier is not talking about the same kind of thing as you are.

Emile 27 Oct 2008 15:34 UTC
0 points
on: Aiming at the Target
Tim: I’m not offering a definition of “optimization process” or “intelligence”, I’m just pointing out that when you say “gravity is an optimization process”, or “stars are optimization processes”, you’re missing the whole point of the Kasparov post.

There’s a difference between the unpredictability of a ball rolling down a hill, and the unpredictability of Kasparov’s chess moves. This difference is not obvious.

Emile 28 Oct 2008 10:59 UTC
1 point
on: Measuring Optimization Power
I agree with Tim that the presence of a deterministic solution shouldn’t be enough to say whether there’s an optimization process going on. But then, Eliezer didn’t say “Since there’s a deterministic solution, no optimization is going on”: it’s more like “it’s possible that no optimization is going on”. Optimization power isn’t required, but might be useful.

Emile 28 Oct 2008 14:48 UTC
0 points
on: Measuring Optimization Power
Ben Jones: I think a process can be deterministic and (relatively) simple, yet still count as an optimization process. An AI that implements an A* algorithm to find the best path across a maze might quality as a (specialized) Optimization process. You can make more accurate predictions about it’s final state than about which way it will turn at a particular intersection—and you can’t always do so for a stone rolling down a hill.

But I’m not sure about this, because you could say something similar about the deterministic checkers player—maybe it uses A* too!

Emile 18 Nov 2008 13:10 UTC
0 points
on: Whither OB?
I wouldn’t mind to the posting becoming weekly instead of daily, that’d be better than a drop in quality.

I’m a bit skeptical about a community site, but you never know. I wouldn’t mind a forum / wiki, but I don’t think it could replace the blog or furnish enough material …

Emile 29 Nov 2008 13:50 UTC
0 points
on: Thanksgiving Prayer
What’s the point of discussing whether atheism is a religion? There must be something about that back in the discussion of Rube and Bleggs.

Same for discussion of whether someone is an atheist or a theist by using different definitions for “god” or “universe” or whatever … once you know Elizier’s beliefs, addind a label isn’t useful information.

Emile 4 Dec 2008 12:03 UTC
0 points
on: Permitted Possibilities, & Locality
douglas:

What the heck does the theory of relativity have to do with AI? The theory of relativity’s only worth is that it gives an accurate description of how the universe works.

Your system might have some common mathematical grounds with relativity, but I don’t think that makes it “based on the Theory of Relativity”. Or at least, the validity of the theory of relativity doesn’t say anything about how applicable the mathematical models are applicable to your domain. Sounds like marketese hype to me.

(Sorry for the somewhat agressive diverging issue; I don’t have much to say about the main post beyond “I want to see more about that”, or “I’m curious as to what Robin will answer”)

Emile 20 Dec 2008 15:04 UTC
4 points
on: Complex Novelty
RDV: Fun comes from overcoming adversity.

Got anything to back that up ? I expect most game designers (I’m one) will tell you that fun is about learning new skills. You may not have noticed, but most games don’t involve any actual threat to the player. (simulated) “adversity” is something you see often in games, but that’s because it’s a “rich” kind of challenge that can give rise to a lot of complex decisions. And also because before video games, the best way to make complex challenges in games was to have another player.

Emile 20 Dec 2008 15:50 UTC
2 points
on: Complex Novelty
Looks like everybody is having comment problems. Maybe someone should post a big sign saying

DON’T REPOST IF YOU GET “The requested URL could not be retrieved” OR YOU’LL POST A DUPLICATE!!!

(I’m using Firefox 3.0.5 on Windows XP, if anyone’s trying to troubleshoot this)

Emile 20 Dec 2008 16:59 UTC
2 points
on: Complex Novelty
Roland: theoreticians of game design (Well—some guys, Raph Coster, Dan Cook …) often reduce “fun” in games to learning—and those aren’t just games that would appeal to this audience. The space of mathematical problems to solve seems like a close enough approximation of the space of complex systems in which one can learn, and, by a happy coincidence, there’s plenty of theoretical work on the complexity of mathematical problems and how large that space might be.

Emile 20 Dec 2008 17:09 UTC
1 point
on: Complex Novelty
Roland: to put it differently, I don’t think that those who hate mathematics don’t enjoy the “aha” of insight (and might greatly enjoy it in video games) - but mathematics in school (especially in the early grades) doesn’t have much to do with insight.

Games are challenges presented in a really inviting and “casual” context. They’re naturally inviting.

Maths is challenges presented in a somewhat dry and “serious” context, being able to enjoy them mostly has to do with being able to understand the language.

(I wonder if there’s any formal study of the kind of games enjoyed by people who like or don’t like maths … it’d be interesting)

Emile 20 Dec 2008 17:19 UTC
11 points
on: High Challenge
Kaj: But if they were made into orgasmium, even for a short while, then they wouldn’t want it to ever stop...

Well, yeah, but most people, if they had their brain removed and replaced by a slab of jelly, wouldn’t raise any objections either.

I don’t think wireheading is that different. A wireheaded brain is basically a hacked brain that doesn’t work any more.

(you could also replace “made into orgasmium” by “given heroin”)