Yeah, I might, but here I was just surprised by the down-voting for contrary opinion. It seems like the thing we ought to foster not hide.
I'm interested in general purpose optimizers, but I bet that they will be evolved from AIs that were more special purpose to begin with. E.g., IBM Watson moving from Jeopardy!-playing machine to medical diagnostic assistant with a lot of the upfront work being on rapid NLP for the J! "questions".
Also, there's no reason that I've seen here to believe that Newcomb-like problems give insights into how to develop to decision theories that allow us to solve real-world problems. It seems like arguing about corner cases. Can anyone establish a practical problem that TDT fails to solve because it fails to solve these other problems?
Beyond this, my belief is that without formalization and programming of these decision frameworks, we learn very little. Asking what does xDT do in some abstract situation, so far, seems very handy-wavy. Furthermore, it seems to me that the community is drawn to these problems because they are deceptively easy to state and talk about online, but minds are inherently complex, opaque, and hard to reason about.
I'm having a hard time understanding how correctly solving Newcomb-like problems is expected to advance the field of general optimizers. It seems out of proportion to the problems at hand to expect a decision theory to solve problems of this level of sophistication when the current theories don't seem to obviously "solve" questions like "what should we have for lunch?". I get the feeling that supporters of research on these theories assume that, of course, xDT can solve the easy problems so let's do the hard ones. And, I think evidence for this assumption is very lacking.
Harsh crowd.
It might be nice to be able to see the voting history (not the voters' names, but the number of up and down votes) on a comment. I can't tell if my comments are controversial or just down-voted by two people. Perhaps even just the number of votes would be sufficient (e.g. -2/100 vs. -2/2).
Seems unlikely to work out to me. Humans evolved intelligence without Newcomb-like problems. As the only example of intelligence that we know of, it's clearly possible to develop intelligence without Newcomb-like problems. Furthermore, the general theory seems to be that AIs will start dumber than humans and iteratively improve until they're smarter. Given that, why are we so interested in problems like these (which humans don't universally agree about the answers to)?
I'd rather AIs be able to help us with problems like "what should we do about the economy?" or even "what should I have for dinner?" instead of worrying about what we should do in the face of something godlike.
Additionally, human minds aren't universal (assuming that universal means that they give the "right" solutions to all problems), so why should we expect AIs to be? We certainly shouldn't expect this if we plan on iteratively improving our AIs.
i don't see how your example is apt or salient. My thesis is that Newcomb-like problems are the wrong place to be testing decision theories because they do not represent realistic or relevant problems. We should focus on formalizing and implementing decision theories and throw real-world problems at them rather than testing them on arcane logic puzzles.
Given the week+ delay in this response, it's probably not going to see much traffic, but I'm not convinced "reading" source code is all that helpful. Omega is posited to have nearly god-like abilities in this regard, but since this is a rationalist discussion, we probably have to rule out actual omnipotence.
If Omega intends to simply run the AI on spare hardware it has, then it has to be prepared to validate (in finite time and memory) that the AI hasn't so obfuscated its source as to be unintelligible to rational minds. It's also possible that the source to an AI is rather simple but it is dependent a large amount of input data in the form of a vast sea of numbers. I.e., the AI in question could be encoded as an ODE system integrator that's reliant on a massive array of parameters to get from one state to the next. I don't see why we should expect Omega to be better at picking out the relevant, predictive parts of these numbers than we are.
If the AI can hide things in its code or data, then it can hide functionality that tests to determine if it is being run by Omega or on its own protected hardware. In such a case it can lie to Omega just as easily as Omega can lie to the "simulated" version of the AI.
I think it's time we stopped positing an omniscient Omega in these complications to Newcomb's problem. They're like epicycles on Ptolemaic orbital theory in that they continue a dead end line of reasoning. It's better to recognize that Newcomb's problem is a red herring. Newcomb's problem doesn't demonstrate problems that we should expect AI's to solve in the real world. It doesn't tease out meaningful differences between decision theories.
That is, what decisions on real-world problems do we expect to be different between two AIs that come to different conclusions about Newcomb-like problems?
If LW would update the page template to have the script in the html header, I think we'd be set. Isn't there a site admin for this?
I think this is critical, because rationality in the end needs mathematical support, and MathJax is really the de facto way of putting math in web posts at this point.
Wouldn't the right solution be to use MathJax?
As one of the folks who made this argument in the other job thread, I'm going to disagree with you. Paying an assistant $36k/yr seems low to me for the Bay Area, but $100k/yr is probably out of line. These all seem like assistanty things that draw more modest salaries. Indeed.com puts the average for administrative assistants in SF at $43k/yr, so given that it's non-profit, it's certainly in range. Do SIAI jobs come with health insurance?
Sorry. It didn't seem rude to me. I'm just frustrated with where I see folks spending their time.
My apologies to anyone who was offended.