Andrei Karpathy (ex-Director of AI at Tesla, founding member of OpenAI) released agenthub, a code implementation of his later autoresearch idea (it emulates a single PhD student doing research to improve LLM training) but I’m actually curious: how can it be used outside of LLM training? the first idea I had is to run hundreds or even thousands of A/B tests simultaneously to optimize key business metrics curious to see how it plays out https://github.com/karpathy/agenthub