Andrej Karpathy, OpenAI co-founder and creator of vibe coding, joins Anthropic

Andrej Karpathy, who was a co-founder of OpenAI, has joined Anthropic. He announced his move on X, claiming that he would be working in research and development for the AI firm. Karpathy is known as one of the best AI researchers in the world, making it a big coup for Anthropic.

In Short

• AI guru Andrej Karpathy joins Anthropic
• He co-founded OpenAI, coined the term vibe coding
• Karpathy also served as Tesla’s director of AI for five years

Anthropic has just bagged arguably one of the biggest coups in AI talent wars. Andrej Karpathy, who co-founded OpenAI, has joined the Dario Amodei-led startup. Karpathy is widely regarded as one of the best AI researchers in the world, and was behind the term “vibe coding” that has become widely used today.

Karpathy announced the move on X. He wrote, “Personal update: I’ve joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative.” He claimed that he will be working on research and development for Anthropic.

Who is Andrej Karpathy?

Andrej Karpathy co-founded OpenAI a decade ago before moving to Elon Musk-led Tesla in 2017. At Tesla, Andrej led the automaker’s director of AI team for just over 5 years.

In 2025, Karpathy coined the term vibe coding – the use of AI to write code. Since then, vibe coding has become a widely accepted term that is used by many when they use coding agents to work.

After leaving Tesla in 2022, he briefly returned to OpenAI before later starting Eureka Labs, an AI-focused education company. Andrej also holds a PhD in computer science from Stanford. In recent years, Karpathy has become a prominent public educator building a large audience on YouTube and X through technical explainers and posts about AI.

Recently he claimed that he had been in a “state of AI psychosis” since December last year, embracing “tokenmaxxing” and aggressively stress-testing frontier models.

Andrej Karpathy to start a new Claude team at Anthropic

Anthropic’s Nicholas Joseph reposted Andrej’s post on X. Nicholas leads Claude’s pretraining team – the team responsible for the large-scale training work that gives Claude its core knowledge and capabilities.

He wrote that Andrej will be working in the pretraining division, but will build a new team that will focus on using Claude itself to speed up pretraining research. This will essentially allow Anthropic to automate this key part of AI development. Anthropic has previously stated that Claude writes 90 per cent of the company’s code.

Nicholas wrote, “Excited to welcome Andrej to the Pretraining team! He’ll be building a team focused on using Claude to accelerate pretraining research itself. I can’t think of anyone better suited to do it — looking forward to what we build together!”

Anthropic’s hiring of Karpathy is another notable addition as competition for senior AI talent intensifies. Earlier this month, Ross Nordeen, a founding member of xAI and former Tesla employee, said he was joining Anthropic. The same day, Anthropic announced a deal with Elon Musk’s SpaceX to rent compute capacity at xAI’s Colossus 1 supercomputer.

Karpathy’s work at OpenAI and Tesla also surfaced during Musk’s recent trial against OpenAI, which the tech billionaire ultimately lost. In one email shown in court, Elon Musk described Karpathy as “arguably the #2 guy in the world in computer vision.” In another message about hiring him, Musk wrote, “The OpenAI guys are gonna want to kill me, but it had to be done.”

- Ends