LLM solutions claimed to be better than task authors’ own
A user shares a list of tasks where they say LLMs—mostly GPT-6 Pro—found solutions significantly better and non-trivially different from the authors’ solutions.
TLDR
A user says LLMs, mostly GPT-6 Pro, found significantly better and non-trivially different solutions than task authors did. They share a list that is still being updated and say they’ll mark solutions they find particularly interesting.
Combined views
144.9K
1 Source, first seen ago