4 ms·
The thing is that the AI can read a puzzle faster than a human can. If someone put any effort towards an AI-based setup, it would easily beat human competitiors
by MattRix 2y ago
The thing is that the AI can read a puzzle faster than a human can. If someone put any effort towards an AI-based setup, it would easily beat human competitiors (well, up until the point the puzzles got too difficult for it to solve).
- petercooper 2y agoI've always done AoC "properly" but this year I've decided to actually use it as a learning experience for working with LLMs (and I don't get up early so will never sully the leaderboard) and trying some experiments along the way. I think the strategy for the harder puzzles is to still "do" them yourself (i.e. read the challenge and understand it) but write the solution in English pseudocode and then have an LLM take it from there. Doing this has yielded perfect results (but less than perfect implementations) in several languages for me so far and I've learnt a few interesting things about how they perform and the "tells" that an LLM was involved.