There is something satisfying, and deeply weird, when an LLM reports that it has fixed 37 of 41 failing tests (a 90% improvement) and then asks what to do next, and correctly interprets "Onwards!"
Posters: patio11. patio11
2 likes recorded by the source corpus.
Topics: AI & LLMs