Tech News, Tech & Coding · Free Online Tools

IT Question Answer
Tools
Back to AI
Prominent AI researcher Andrej Karpathy joins Anthropic

Prominent AI researcher Andrej Karpathy joins Anthropic

AIBy Muhammad FareedVentureBeat; Axios; TechRadar; The Decoder
aianthropicresearchtalent

The announcement

On May 19, 2026, Andrej Karpathy said on X that he had joined Anthropic, in a post he called a personal update. The post appeared on the day Google's I/O developer conference began. He wrote that the next few years at the frontier of large language models will be formative and that he was excited to get back to research and development. He added that he remains committed to education and plans to return to it in time. Anthropic confirmed the hire.

His role

Anthropic's head of pretraining, Nick Joseph, said Karpathy will join the pretraining team. A company spokesperson said he will start a team that uses Claude to accelerate pretraining research. The spokesperson tied that work to what is known as recursive self-improvement, meaning AI helping to train or improve its successors with less human involvement. Outlets differ on whether he leads a new group or joins an existing one, and some said he reports to Joseph.

Who he is

  • One of OpenAI's original co-founders.
  • Led computer vision for Tesla's Autopilot from 2017 to 2022 and served as Tesla's director of AI.
  • Returned to OpenAI in 2023 and 2024, where he built a team on mid-training and synthetic data generation.
  • Earned a PhD at Stanford under Fei-Fei Li and helped create the CS231n course.
  • Since 2024 has worked on education through Eureka Labs and on open-source projects, which will likely slow while he works at Anthropic.

Why pretraining matters

Pretraining is the stage where a model learns broad abilities from large amounts of data, and its quality shapes everything built on top. A respected researcher can also help recruiting, since leading researchers can choose where to work and their choice is read as a signal about where important work is happening. Karpathy's choice of Anthropic over his former employer was widely framed as a talent win.

The talent race

Experienced researchers who can train and scale large models are scarce, and expertise is concentrating at a few well-funded labs. Anthropic was growing quickly at the time, as covered in our reports on its operating profit outlook and its funding round and IPO filing.

What to watch

Look for any public research or tools from his team, how the self-improvement work is governed, and whether other senior researchers follow.

Advertisement