Economy USA

Anthropic Lands OpenAI Vo-Founder Andrej Karpathy in a Big AI Talent Grab

Anthropic Lands OpenAI Vo-Founder Andrej Karpathy in a Big AI Talent Grab
Andrej Karpathy Director of AI Tesla a keynote speaker at the Train AI conference at Pier 27 in San Francisco, Ca. on Thurs. May 10, 2018 (Michael Macor / San Francisco Chronicle / Getty Images)
  • Published May 20, 2026

With input from CNBC, Forbes, Axios, Fortune, Reuters, Business Insider, and Bloomberg.

Anthropic just scooped up one of the best-known names in AI.

Andrej Karpathy, who helped found OpenAI before heading to Tesla to run AI work there, said Tuesday that he is joining Anthropic. The company said he will be part of its pretraining team, the group that helps Claude build the core knowledge and capabilities behind the model.

Karpathy said he is excited to get back to research and development, calling the next few years at the frontier of large language models especially important. That tracks with his reputation: he is one of those rare AI figures who has credibility in research, industry and education all at once.

This is a big win for Anthropic, which has been on a hiring spree as it battles OpenAI for top technical talent. The company is trying to scale Claude fast, and names like Karpathy matter not just because they are famous, but because they know how to build the thing from the ground up.

Karpathy’s path through the AI world has been a winding one. He was part of OpenAI early on, then left in 2017 for Tesla, where he led computer vision work for Autopilot. Musk later pulled him over while still involved with OpenAI, and Karpathy’s time at both companies came up repeatedly in the Musk v. Altman trial that wrapped up Monday.

He eventually returned to OpenAI for a brief stretch, left again, and then started Eureka Labs, an AI-focused education startup. Along the way, he became a widely followed voice in the AI world, known for detailed posts, sharp opinions and, more recently, coining the phrase “vibe coding” for using AI in a looser, more intuitive way to write code.

Now he is back in the thick of it.

Anthropic said Karpathy will help build a new effort aimed at using Claude itself to speed up pretraining research, which is the grunt work that gives AI models their basic intelligence before they are fine-tuned into useful products. It is the kind of role that fits him well: technical, foundational and very much at the frontier.

The move also lands at a tense moment in the AI race. Anthropic has been rising fast, with its valuation reportedly blowing past OpenAI’s in private markets. At the same time, it is fighting on several fronts, from talent wars to legal and political battles over how far AI companies should go.

Karpathy said he still cares deeply about education and plans to return to that work later. For now, though, he is back where the industry’s biggest fights are happening: inside the lab, at the edge of the next model.

Wyoming Star Staff

Wyoming Star publishes letters, opinions, and tips submissions as a public service. The content does not necessarily reflect the opinions of Wyoming Star or its employees. Letters to the editor and tips can be submitted via email at our Contact Us section.