Andrej Karpathy joins Anthropic

Andrej Karpathy said on May 19 that he had joined Anthropic, returning to full-time research after leaving OpenAI last year and spending recent months on education-focused work. The move places one of OpenAI’s founding figures inside a direct rival as major AI labs compete for researchers who have built large models at scale. (axios.com) Axios and Forbes both reported the hire on Tuesday, and Anthropic confirmed details of his role to other outlets. (forbes.com) Karpathy wrote on X that “the next few years at the frontier of LLMs will be especially formative” and said he was excited to “get back to R&D.” Anthropic said he began this week on its pre-training team, the group responsible for the large-scale training runs that shape Claude’s core capabilities. (techcrunch.com) ### What job is Karpathy taking at Anthropic? Anthropic said Karpathy is working on pre-training under team lead Nick Joseph, an early Anthropic hire and former OpenAI researcher. Forbes reported that he will help launch a new team focused on using Claude itself to develop future models, rather than relying only on raw compute and hardware. (axios.com) Bloomberg reported that his work is focused on helping train new artificial intelligence models. TechCrunch separately described pre-training as the stage that gives Claude its base knowledge and capabilities, making it one of the most compute-intensive parts of building a frontier model. (techcrunch.com) ### Why does this move stand out inside the AI industry? Karpathy was one of OpenAI’s founding members and later led Tesla’s Autopilot and AI efforts before returning to OpenAI and then departing again in 2025. That background gives him a rare résumé across two of the best-known AI organizations and one of the most prominent real-world AI deployment efforts in the auto industry. (forbes.com) Axios described the hire as a major gain for Anthropic in the competition for elite AI talent. Forbes said the move comes as Anthropic, OpenAI, Google and Meta compete for a small pool of researchers with experience building advanced systems. (bloomberg.com) ### What did Karpathy say publicly about the decision? Karpathy said in his May 19 post that he had joined Anthropic and expected the coming years in large language models to be important. He also said he remained “deeply passionate about education” and planned to resume that work later, signaling that his public teaching and educational projects are not necessarily over. (forbes.com) CNBC reported that Karpathy framed the decision as a return to research and development. VentureBeat noted that his move raises questions about how his open-source and education efforts will fit alongside work at Anthropic, though the company has not publicly laid out those details. (axios.com) ### Where does this leave his work outside Anthropic? Bloomberg reported that Karpathy had launched a startup focused on AI and education after leaving OpenAI for a second time last year. Public reports on May 19 did not say Anthropic had acquired that effort or folded it into the company. (techcrunch.com) Tech and business outlets reporting the hire focused instead on his immediate assignment inside Anthropic’s model-development organization. Anthropic’s public description, as cited by Forbes and TechCrunch, centers on pre-training and a new team using Claude in the model-building process. (cnbc.com) ### What comes next to watch? Anthropic said Karpathy started this week, making his first visible milestone the build-out of the pre-training group under Nick Joseph. Any next public signal is likely to come through Anthropic research updates, Claude model announcements or future posts from Karpathy about his education work. (forbes.com) (bloomberg.com)