TL;DR
Andrej Karpathy, one of the 11 original co-founders of OpenAI, has joined Anthropic’s pre-training team. He will form a new group that will use Claude himself to accelerate the most expensive phase of the development of the frontier model.
Andrej Karpathy, one of the original co-founders of OpenAI and one of the world’s most renowned AI researchers, announced on monday who has joined Anthropic. The move is a major coup of talent for the creator of Claude in his race to stay on the frontier of developing great language models.
Update from Andrej Karpathy
Karpathy joins anthropic pre-training team, led by Nick Joseph, where you will create an entirely new group focused on a surprisingly recursive goal: using Claude himself to speed up pre-training research. Pre-training, the intensive massive computing phase that gives a frontier model its core knowledge and capabilities, is the most expensive part of building systems like Claude. Finding ways to make that process faster and more efficient could reshape the economics of the entire AI industry.
In an He added that he remains “deeply passionate about education” and plans to return to that work in time.
The hire caps a career arc that has touched nearly every major turning point in modern AI. Karpathy earned his PhD at Stanford under Fei-Fei Li, the computer scientist behind ImageNet, focusing on deep learning and computer vision. He was among the 11 people who co-founder of OpenAI in 2015, where he worked on deep learning research before leaving in 2017 to join Tesla as head of AI.
At Tesla, Karpathy led the computer vision teams behind Full Self-Driving and Autopilot, the programs underpinning the electric car maker’s ambitions for autonomous vehicles. He left in July 2022, returned to OpenAI for about a year, and then left again in 2024 to found Eureka Labs, a startup that applies AI assistants for education. Eureka Labs’ work is now on pause while Karpathy supports Anthropic.
The moment is remarkable. Anthropic has become a magnet for top technical talent at a time when its main rival, OpenAI, has experienced a series of high-profile departures. Over the past two years, OpenAI has lost more than a dozen senior executives and researchers, including CTO Mira Murati, reinforcement learning pioneer John Schulman, and, most recently, three executives who left in a single day in April 2026.
For Anthropic, landing in Karpathy indicates that the company can attract talent of the highest caliber as it scales both its research and commercial operations. The company, led by CEO Dario Amodei, has attracted investor interest with a valuation of about $800 billion and is reportedly exploring an initial public offering that could happen as early as late 2026.
Karpathy’s new role also underscores a broader trend in cutting-edge AI: using existing models to improve the next generation. If Claude can significantly speed up his own pre-training process, it would be a practical demonstration of recursive self-improvement, one of the capabilities the AI security community has long followed closely. Whether that prospect excites or unnerves observers may depend on how much trust they place in the anthropic security-minded culture has cultivated since its foundation.
For now, Karpathy appears to be exactly where he wants to be: back in the lab, building models on the frontier.
