Andrej Karpathy joins Anthropic

Andrej Karpathy is joining Anthropic, he posted on X while Google was holding its I/O keynote. In AI circles he needs little introduction: he was among the 11 people who co-founded OpenAI in 2015, led computer vision at Tesla for years, briefly came back to OpenAI, and more recently ran Eureka Labs, a startup built around AI-assisted education. When someone with that résumé moves, the whole field pays attention.

His path runs through most of the big shifts in modern AI. He earned a PhD at Stanford under Fei-Fei Li, whose ImageNet work helped kick off the deep-learning wave. At OpenAI he worked on early model research before Elon Musk recruited him to Tesla in 2017. There he served as Director of AI and ran the vision teams behind Autopilot and Full Self-Driving. He left Tesla in 2022, spent roughly a year back at OpenAI, then founded Eureka Labs in 2024.

Outside the labs he is almost as well known as inside them. Millions have followed his online courses on neural networks, and he coined the term «vibe coding», which caught on fast among developers. He is one of the few researchers who can move frontier-model work and still explain it in plain language to a wide audience.

At Anthropic he joins the pretraining team and will use Claude to speed up development of the next generation. «The next few years at the LLM frontier will be especially formative,» he wrote. He also said he plans to return to education work in time.

For Anthropic, landing him is a clear win. That an OpenAI co-founder chose Anthropic over going back to the lab he helped start sends a clear signal in the race toward the next generation of models. Karpathy is moving from explaining AI to millions of people back to building it. The rest of the industry will be watching to see what that means for Claude.