News

OpenAI Co-Founder Karpathy Joins Anthropic to Lead Pre-Training

· Rebecca Bellan, Lorenzo Franceschi-Bicchierai

OpenAI Co-Founder Karpathy Joins Anthropic to Lead Pre-Training

Andrej Karpathy is moving to Anthropic. The AI researcher, best known for co-founding OpenAI and leading computer vision at Tesla, will focus on pre training models at the company…

Andrej Karpathy is moving to Anthropic. The AI researcher, best known for co-founding OpenAI and leading computer vision at Tesla, will focus on pre training models at the company behind Claude. He announced the news on X, where he also shared a photo of himself holding a coffee mug with a personal inside joke. This is one of the few times he has switched jobs since leaving OpenAI in 2017.

Pre training is the foundational stage where a model learns from massive amounts of data. It is expensive and compute intensive, which makes Karpathy an obvious hire. He has deep expertise in both the engineering and research sides of building large scale neural networks. At Tesla he pushed hard on self driving technology, and at OpenAI he was instrumental in early GPT work. His move to Anthropic signals that the company is pouring resources into the next generation of its models.

The timing is interesting. Anthropic has been positioning itself as the safety conscious alternative to OpenAI. But if you are hiring one of the most prominent figures from that world, you are signaling that you want to compete on raw capability too. Karpathy is not a researcher who avoids the spotlight. He teaches, writes, and explains complex ideas to the public. That fits with Anthropic’s broader strategy of being transparent, or at least appearing more transparent than its rivals.

What happens next is the real question. Anthropic has raised billions and needs to show results. Karpathy’s focus on pre training suggests the company is investing in the model itself, not just alignment or safety layers. If they nail both, they could become a serious number two behind OpenAI. If not, this is still a big talent grab that shuffles the deck in AI research. Expect more movement soon.

Original source