Karpathy Joins Anthropic From OpenAI

Andrej Karpathy was there at the beginning. A founding member of OpenAI, he left to run computer vision and autonomous driving at Tesla, then returned to OpenAI before departing again in 2023. He has spent the years since building educational content — his neural network tutorials became something close to required reading for a generation of AI practitioners. That independent chapter now appears to be closing. As of May 19th, TechCrunch is reporting that Karpathy has joined Anthropic, specifically to work on pre-training. The move lands him at OpenAI's most credible rival, working on the part of the stack that determines, more than anything else, what a model can ultimately become.

If confirmed, here is what this means. Pre-training is where the foundational bets are placed — it is expensive, slow, and the results are irreversible in the sense that every subsequent capability is built on top of what gets learned here. Karpathy joining Anthropic's pre-training team is not a communications hire or a research adjacency; it is a signal that Anthropic is investing in the deepest layer of the process. For OpenAI, losing a founding voice to a direct competitor carries a specific sting beyond any single technical contribution. It feeds a narrative — already circulating — that the culture which made early OpenAI magnetic has dispersed. For the broader field, it concentrates more of the world's serious pre-training thinking inside the small cluster of frontier labs, and Anthropic's cluster just got meaningfully denser.

Watch for an official Anthropic announcement or a statement from Karpathy himself — his previous departures have been accompanied by candid, personal explanations, and if that pattern holds, it will tell us far more than the job title does.

NewsHive monitors these sources continuously. All signal titles above link to the original reporting.

Intelligence by NewsHive. Need help navigating what this means for your business? Contact GeekyBee →