Post by Michael Koker

Andrej Karpathy joined Anthropic today. Karpathy was a founding member of OpenAI in 2015. He left to run Autopilot at Tesla for five years, came back to OpenAI in 2023, then left again to do his own thing. His own thing turned out to be Eureka Labs, an AI-native education company. Before any of that he did his PhD at Stanford under Fei-Fei Li and taught CS231n, the convolutional neural networks course that trained roughly half the people now building AI systems. He is also the single best explainer of how this stuff actually works. His YouTube series building GPT from scratch in a Jupyter notebook is the closest thing the field has to a Feynman lecture. nanoGPT, micrograd, and the 'Software 2.0' essay are all his. He coined 'vibe coding' last year. He could have done anything today. Gone back to OpenAI. Joined Meta or xAI for a check with a lot of zeros. Stayed at Eureka full time. Started his own lab. He picked Anthropic. He is not the first. Jan Leike walked out of OpenAI's superalignment team in 2024 and went to Anthropic. Mike Krieger, the Instagram co-founder, took the CPO role there. Now Karpathy. Karpathy's craft is stripping neural networks down to their essentials and explaining what is actually happening inside them. Anthropic's interpretability work, the strongest in the field, is doing the same thing from the research side. Putting him next to that team is more interesting than putting him next to any other lab's training infrastructure. The race is not over.