18
Score
Karpathy
Karpathy
3 entries in AI Capability Tracker
17
Score
An AI agent autonomously improved a language model training pipeline over two weeks, making iterative code changes that reduced validation loss without increasing training time.
14
Score
A larger model is running in production on consumer-grade hardware for extended periods.
mail
Get notified about new Karpathy developments
Primary sources. No fluff. Straight to your inbox.