Comment by p1dda

Comment by p1dda 21 hours ago

It would be interesting to see which is faster/better in competitive coding, the human coder or the human using AI to assist in coding.

Snuggly73 18 hours ago

New benchmark for competitive coding dropped yesterday - https://livecodebenchpro.com/

Apparently models are not doing great for problems out of distribution.

p1dda 16 hours ago

It goes to show that the LLMs aren't intelligent in the way humans are. LLMs are a really great replacement for googling though

Reply View | 0 replies

asciimov 20 hours ago

It would only be interesting if the problem was truly novel. If the AI has already been trained on the problem it’ll just push out a solution.

wiseowise 18 hours ago

It already happened. Last year AI submissions completely destroyed AoC, as far as I remember.