131–135 of 135 posts
Earlier quoted context omitted.
Yes, and that’s what a coding agent is too. Technically an agent is not (just) an LLM, but the two have become synonymous is discussions.
> that’s what a coding agent is too. Not quite.
Earlier quoted context omitted.
An AI agent mutates state (the repository) in order to get some metric (e.g. test suites) closer to passing. Same thing, really.
An AI agent isn't an evolutionary optimization loop.
Earlier quoted context omitted.
Read https://www.argmin.net/p/lore-laundering-machines Given time, we may find out that the solutions in this paper were also in the literature, as was the case in the anecdotes from the linked article :)
Then its utility as a best search agent is even more. It proves the statement, LLM will find the needle in the haystack if the needle exists.
Except that, instead of telling you that the idea came from a particular source, it fools you into thinking it came up with the idea. In other words, it cannot be described as "search", but rather "laundering" as the article describes it.
Earlier quoted context omitted.
> that mimics how people do real math Real people do not do math like AlphaEvolve...
What do you feel is fundamentally different about the feedback loop in AlphaEvolve compared to, say, Einstein and Grossman repeatedly running calculations until they found the right tensor setup for General Relativity? Or Euler filling his notebooks with computations? Or Ramanujan? Or Newton working out infinite series? Or Kepler, etc etc. They are all doing iterative search with feedback from a function that tells t…