Comment by Yizahi
14 hours ago
Oh, he's too far gone it seems. Completely absorbed in his own marketing shpiel. And he is supposedly an engineer? I don't see any discussion about pros and cons (only pros of course), I see zero reasoning about WHY exactly LLMs are so successful in coding and somewhat in math/chemistry etc. and is it even applicable to other activities. I mean, if this is their fabled "top scientist level reasoning" then sure, AIs have reached it. The problem is that's no reasoning at all.
PS: he may be accidentally right about global impact, but the key word is "accidentally" because this crappy "reasoning" doesn't support that on it's own merit.
It would be tiring if you want a big essay on why LLMs are useful and whether they are productive. Much of the people are already convinced.
I don't think its useful to keep having to prove LLMs don't hallucinate or LLMs help in coding each time.
Not why they are useful. Useful is load-bearing word (hehe) here, it spans from assisting someone by 1% or by 99%, very vague and not concrete.
I was talking about omitting mentions why LLMs usefulness is efficient - in coding specifically and STEM in general to a lesser degree.
And this is really a crucial thing, which do need to be brought up when discussing future forecast about multiple industries at once.
And the reason why LLMs are so efficiently useful for coding is because LLMs perform much better if any of the two conditions are met:
1) The problem at hand is formalized and transformed into a specification basically.
2) There is a formalized verification of the output done by external pipeline of different tools.
Any of the two drastically improve LLM performance, and coding activity uniquely has both of them.
Having a formalized task narrows down the scope of generation for LLM, and formal verification allows LLMs to iteratively converge on the verification pass check. Same with those math problems, they have a formal verification and LLMs can be prevented from wandering in a generated forest too much. Same in chemistry or engineering for some degree.
But other human activities usually have rather badly formalized inputs and no formal verification of the results. Karpathy is making a logical mistake (or deliberate misleading) of projecting a very unique industry (coding) on every other human activity without any nuance or analysis. He is biased, but due to his reputation and power he doesn't get called out on that mistake.