Andrej Karpathy is one of the most respected voices in AI, and he’s now joined forces with the fastest growing startup of all time.
The Slovak-Canadian computer scientist announced on X that he has joined Anthropic, calling the next few years at the frontier of large language models “especially formative.” The move ends what has been an extended break from frontier lab work — a period Karpathy spent building educational content on YouTube and launching an edtech startup — and places one of AI’s most recognizable figures squarely inside the company that has emerged as OpenAI’s most serious rival.
A Career Defined by Firsts
Few people in AI carry a résumé quite like Karpathy’s. He earned his PhD from Stanford under the supervision of Fei-Fei Li, before becoming one of the founding members of OpenAI in 2015. Two years later, Elon Musk secretly recruited him — while still on OpenAI’s board — to lead Tesla’s self-driving AI effort. Court testimony from the ongoing Musk v. Altman trial revealed that Musk emailed a Tesla VP at the time: “The OpenAI guys are gonna want to kill me. But it had to be done.”
At Tesla, Karpathy served as Senior Director of AI, leading the computer vision team behind Autopilot and briefly overseeing the Optimus humanoid robot program. He left in July 2022, and after a brief return to OpenAI in 2023 — where he worked on a team related to ChatGPT — he stepped down again in February 2024, insisting both times that there was no drama behind his exits.
The Education Detour
What followed was perhaps the most publicly beloved chapter of Karpathy’s career. His YouTube lectures on LLMs and neural networks attracted millions of views, with his one-hour deep learning lecture alone crossing 3.1 million views. His “Neural Networks: Zero to Hero” series became required watching for a generation of AI practitioners.
In July 2024, he took the education mission further by launching Eureka Labs, an AI-native edtech startup. The company’s vision was ambitious: pair high-quality course materials with AI-powered teaching assistants to make expert-level education accessible to anyone, anywhere, in any language. Its first product, LLM101n, was an undergraduate-level course designed to walk students through training their own AI model. “It’s the culmination of my passion in both AI and education over about two decades,” Karpathy said at the time.
The education chapter hasn’t been closed — in his Anthropic announcement, Karpathy said he “remains deeply passionate about education” and plans to resume that work in time.
Musk Tried, Anthropic Won
The hiring is a notable win for Anthropic, not least because Karpathy had been openly courted in the other direction. When Karpathy posted a warm remark about the Tesla Diner in Los Angeles last year, Musk quote-posted it with a pointed plea: “Andrej, my long lost brother, let us work together again!” It was the kind of public recruitment attempt Musk rarely makes — a signal of just how much he valued bringing Karpathy into the xAI orbit.
It didn’t work. Karpathy chose Anthropic instead.
The decision carries symbolic weight. Karpathy had previously noted wryly that fewer researchers leave Anthropic to start companies than leave other frontier labs — partly because Anthropic’s equity is less liquid. Now he’s on the other side of that observation.
What It Means for Anthropic
Karpathy’s hire is significant beyond the name recognition. He is, above all, a hands-on research scientist who has consistently said his preference is to be close to the technical work — not in a corporate executive role. His stated reason for leaving Tesla was that his job had drifted away from the engineering. His announcement at Anthropic specifically flagged getting “back to R&D” as a draw.
Anthropic is currently locked in a tight race with OpenAI and Google DeepMind at the frontier of LLM development. The company has also attracted attention for its Claude Code tool and its work on agentic AI systems — exactly the kind of territory where Karpathy’s technical instincts and broad influence could matter.
For the wider industry, the move is a reminder that the talent war at the top of AI is far from settled. The next few years, as Karpathy put it, will be “especially formative” — and he’s now betting on Anthropic to be where that future gets built.