LessWrong 2.0 Reader

The median AGI timeline of more than half of METR employees is before the end of 2030.

(AGI is defined as 95% of fully remote jobs from 2023 being automatable.)

If I were to think about it a little, I'd suspect the big difference that LLMs and humans have is state/memory, where humans do have state/memory, but LLMs are currently more or less stateless today, and RNN training has not been solved to the extent transformers were.

One thing I will also say is that AI winters will be shorter than previous AI winters, because AI products can now be sort of made profitable, and this gives an independent base of money for AI research in ways that weren't possible pre-2016.

It seems quite plausible for someone to falsely belive they were asexual in this situation.

I understand that if you are starving or nutrient deficient (zinc, vitamin D, vitamin B12, and iron) your sex drive can be at zero. If it's like that for long enough you may think that it is because it's inherent to who you are. You are wrong, but have no way of knowing that.

I agree with you on your assessment of GPQA. The questions themselves appear to be low quality as well. Take this one example, although it's not from GPQA Diamond:

In UV/Vis spectroscopy, a chromophore which absorbs red colour light, emits _____ colour light.

The correct answer is stated as yellow and blue. However, the question should read transmits, not emits; molecules cannot trivially absorb and re-emit light of a shorter wavelength without resorting to trickery (nonlinear effects, two-photon absorption).
 

This is, of course, a cherry-picked example, but is exactly characteristic of the sort of low-quality science questions I saw in school (e.g with a teacher or professor who didn't understand the material very well). Scrolling through the rest of the GPQA questions, they did not seem like questions that would require deep reflection or thinking, but rather the sort of trivia things that I would expect LLMs to perform extremely well on.

I'd also expect "popular" benchmarks to be easier/worse/optimized for looking good while actually being relatively easy. OAI et. al probably have the mother of all publication biases with respect to benchmarks, and are selecting very heavily for items within this collection. 

Whether or not to get insurance should have nothing to do with what makes one sleep – again, it is a mathematical decision with a correct answer.

Don't be overly naive consequentialist about this. "Nothing" is an overstatement.

Peace of mind can absolutely be one of the things you are purchasing with an insurance contract. If your Kelly calculation says that motorcycle insurance is worth $899 a month, and costs $900 a month, but you'll spend time worrying about not being insured if you don't buy it, and won't if you do, I fully expect that is worth more than $1 a month.

But do be actual consequentialist about it. If the value of the insurance is more like $10, but the cost is $900, I doubt peace of mind about this one thing is worth $890 a month.

Actually, I've changed my mind, in that the reliability issue probably does need at least non-trivial theoretical insights to make AIs work.

Your formula is only valid if utility = log($).

This is a synonym for "if money compounds and you want more of it at lower risk". So in a sense, yes, but it seems confusing to phrase it in terms of utility as if the choice was arbitrary and not determined by other constraints.

I do think that something like dumb scaling can mostly just work

The exact degree of "mostly" is load-bearing here. You'd mentioned [LW(p) · GW(p)] provisions for error-correction before. But are the necessary provisions something simple, such that the most blatantly obvious wrappers/prompt-engineering works, or do we need to derive some additional nontrivial theoretical insights to correctly implement them?

Last I checked, AutoGPT-like stuff has mostly failed, so I'm inclined to think it's closer to the latter.

I think the history of things being predicted Real Soon Now is one of the main counterarguments to short timelines. It just seemed Obvious that we were getting flying cars, or fusion power, or self-driving cars, or video-phones, for years, before in some cases we eventually did get those things, and in other cases maybe we'll never get those things because technology just followed a different path than we expected.

Like, maybe the "we'll just merge with the machines" people will turn out to actually be right. I don't believe it. But it could happen, and there are plenty of similar things that "could happen" that eventually add up to a nontrivial chunk of probability.

In the strongest sense, neither the human brain analogy nor the evolution analogy really apply to AI. They only apply in a weaker sense where you are aware you're working with analogy, and should hopefully be tracking some more detailed model behind the scenes.

The best argument to consider human development a stronger analogy than evolutionary history is that present-day AIs work more like human brains than they do like evolution. See e.g. papers finding that you can use a linear function to translate some concepts between brain scans and internal layers in a LLM, or the extremely close correspondence between ConvNet feature and neurons in the visual cortex. In contrast, I predict it's extremely unlikely that you'll be able to find a nontrivial correspondence between the internals of AI and evolutionary history or the trajectory of ecosystems or similar.

Of course, just because they work more like human brains after training doesn't necessarily mean they learn similarly - and they don't learn similarly! In some ways AI's better (backpropagation is great, but it's basically impossible to implement in a brain), in other ways AI's worse (biological neurons are way smarter than artificial 'neurons'). Don't take the analogy too literally. But most of the human brain (the neocortex) already learns its 'weights' from experience over a human lifetime, in a way that's not all that different from self-supervised learning if you squint.