Post #1912176 · Elektrine

2026-02-09 12:27 UTC

I don't want to overstate the case — I saw someone vibe coded the same project and then declared programming was dead after they finished, but their bot loses to plain MCTS with a simple heuristic. Mine was in the same state when I just had an LLM running in a loop with instructions to improve the result. At least for now, you have to apply some direction, but it turns out someone with no AI background can supply enough direction.

Replies (3)

@danluu@mastodon.social 2026-02-10 17:48

Looks like I spoke too soon about the AI not being superhuman. The current Azul world champion played against it and thinks it's better than him at higher difficulties, and a top 100 player played against it at default difficulty and thought it was better than him at default. I should write a longer post about this. As someone who has a default of trying to have a better understanding of their project than most people would, going full vibe and understanding almost nothing was interesting.

Open ##1912177
@dynomight@mastodon.social 2026-02-09 14:57

@danluu Can you share any details about your workflow? Were you really running the LLM on CPU on your laptop? Or was the LLM generating ML code for some other machine learning that ran on-laptop?

Open ##1912184
@trademark@fosstodon.org 2026-02-09 15:03

@danluu For me, I find that I am faster at producing production quality code in a language I am fluent in. If it's some auxiliary type code with less stringent requirements, quick-and-dirty experiment or such, the AI is much faster than I am. (I like to think I'd get a better result eventually, though)

Open ##1912187