Andrej Karpathy, a co-founder of OpenAI and an AI researcher who previously worked at OpenAI and previously led AI at Tesla, has joined Anthropic.
“I joined Anthropic,” Karpathy posted on X Tuesday. “I think the next few years at the LLM front will be a particularly formative time, and I’m very excited to join the team here and return to research and development.”
Karpathy will start this week at Anthropic, working in pre-training with team leader Nick Joseph. According to the company, Pre-Training will be responsible for conducting extensive training that will equip Claude with core knowledge and competencies. This is also one of the most costly and compute-intensive phases in building a frontier model.
An Anthropic spokesperson told TechCrunch that Karpathy will be building a team focused on using Claude to accelerate pre-training research.
Karpathy is one of the few researchers who can bridge the gap between LLM theory and large-scale training practice. Hiring him to build a team like this is a clear sign that Anthropic believes AI-powered research, rather than pure computing, is the way to stay competitive with OpenAI and Google.
During his time at OpenAI, Karpathy focused on deep learning and computer vision before leaving to join Tesla in 2017. He led Tesla’s Full Self-Driving (FSD) and Autopilot programs until his retirement in 2022.
He then returned to OpenAI for a year and left again in 2024 to start Eureka Labs, a startup focused on applying AI assistants to education.
Karpathy hasn’t shared many updates about Eureka Labs since its launch, and it’s unclear whether the renowned researcher will continue with the startup. I also teach an online course called Neural Networks: Zero to Hero, which helps students learn how to build neural networks from scratch in code. I also have a YouTube channel where I post lectures on LLM and AI semi-regularly.
“I still have a deep passion for education and plan to resume my education efforts in time,” Karpathy said.
TechCrunch has reached out to Karpathy for comment.
Separately, Anthropic has added Chris Rohlf to its Frontier Red team, which stress tests advanced AI models against serious threats. Rohlf is a cybersecurity industry veteran with over 20 years of experience. He previously worked on Yahoo’s acclaimed cybersecurity team known as “The Paranoids,” and most recently joined Anthropic after six years at Meta. Rohlf is also a research fellow at Georgetown’s Center for Security and Emerging Technologies, where he worked on CyberAI projects.
“There is a real opportunity before us to leverage AI to dramatically improve cybersecurity, and I can’t think of a better company or team to join us at this important moment,” Rohlf said in a post on X.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
