Hacker Next
The unreasonable effectiveness of an LLM agent loop with tool usesketch.dev
440 points by crawshaw 4 days ago | 318 comments