Who is Andrej Karpathy and why he matters

Andrej Karpathy was born on October 23, 1986 in Bratislava, from where he moved with his family to Toronto, Canada at the age of fifteen. He studied computer science and physics at the University of Toronto, earned his master's degree at the University of British Columbia, and defended his PhD in computer science at Stanford University under the legendary Fei-Fei Li.

Even during his PhD, he designed and taught the course CS 231n: Convolutional Neural Networks for Visual Recognition, which became one of the most popular courses at Stanford. In 2015, he became one of the founding members of OpenAI, where he worked at the intersection of computer vision and natural language processing.

In 2017, Elon Musk personally recruited him to Tesla for the position of director of artificial intelligence. According to an email that was submitted as evidence during the recent Musk v. Altman trial, Musk described Karpathy as "probably the number two in the world in computer vision" — right after Ilya Sutskever, another OpenAI co-founder. At Tesla, Karpathy led the team responsible for the computer vision of the Autopilot system.

After leaving Tesla in 2022, he briefly returned to OpenAI, only to leave again in February 2024 and found his own startup Eureka Labs — an AI-focused educational platform. In February 2025, he also popularized the term "vibe coding", which describes how AI tools enable laypeople to create applications simply by entering text prompts.

Move to Anthropic: a new chapter in the talent war

Karpathy announced his move on Tuesday, May 19 on X. "I think the next few years at the frontier of large language models will be particularly formative. I'm excited to join the team and get back to research and development," he wrote.

Anthropic confirmed that Karpathy is starting this week and will lead a team focused on using Claude to accelerate pretraining research — the phase where a language model acquires its foundational knowledge and capabilities. In other words: Karpathy will work on making the next versions of Claude even smarter, more accurate, and more capable.

It is not the only significant move. In early May, Ross Nordeen, a founding member of xAI and former Tesla employee, also announced his transition to Anthropic. Talents from Elon Musk's circle are thus increasingly appearing in the ranks of the competition.

Musk, Altman, and Karpathy as a witness to history

Karpathy's name resonated in the courtroom literally hours before he announced his move. On Monday, May 18, the long-watched trial Musk v. Altman ended, in which Elon Musk sued OpenAI and Sam Altman for allegedly violating the company's original mission. The jury and judge ruled in favor of Altman.

During the trial, Karpathy's role was repeatedly discussed. In one of the submitted emails, Musk wrote that he had "borrowed" Karpathy from OpenAI for work at Tesla — along with other OpenAI employees who worked for free for months on autonomous driving development. "The people at OpenAI are going to want to kill me, but it had to be done," Musk wrote.

This context adds another dimension to Karpathy's current move to Anthropic. While the courtroom drama between Musk and Altman has ended, Karpathy decided to bet on a third player — a company that in recent months has become the most serious competitor to OpenAI.

Anthropic on the rise: Claude as a challenger to GPT

Anthropic, founded in 2021 by former OpenAI employees including Daniel and Dario Amodei, has transformed over the past year from a smaller player into a giant. According to available information, its valuation is approaching — or even surpassing — OpenAI's private market valuation.

The Claude model achieves comparable or better results than OpenAI's GPT-4o in many benchmarks. Claude particularly excels in the areas of safety, factual accuracy of responses, and the ability to work with long documents. Karpathy's expertise in pretraining could push these capabilities even further — and potentially move Claude ahead of models from both OpenAI and Google.

What it means for users and businesses

For regular users and businesses — including those in the Czech Republic and Slovakia — Karpathy's move could mean several concrete things:

Faster innovation. The war for AI talent between OpenAI, Anthropic, Google, and xAI means one thing: models will improve faster than they would without competitive pressure. We are already seeing that each new version of Claude and GPT brings leapfrog improvements.

Availability in the EU. Claude has been available across the entire European Union since May 2024. It works in Czech — it understands it and responds at a solid level, albeit occasionally with minor errors. With Karpathy on the team, the quality of Claude's multilingual capabilities could further improve, which is good news for Czech users who don't want to communicate exclusively in English.

Lower prices? Maybe. Claude offers a free plan with certain limits and a paid Pro plan for approximately $20 per month (roughly 460 CZK). Competitive pressure could lead to better pricing — similar to when OpenAI introduced free access to GPT-4o.

Czech and Slovak footprint in global AI

Karpathy is not the only Central European name moving the world of artificial intelligence. But his story is exceptional precisely for its breadth of scope — from education through research to industrial deployment. For young Czech and Slovak students, programmers, and researchers, he is proof that even a person from a small Central European country can stand at the very center of the technological revolution.

Several companies and research groups applying AI operate in the Czech Republic, from startups like Rossum through research teams at CTU and MUNI to AI departments of large corporations. With the growing demand for AI experts in Europe, Karpathy's story is inspiring — and his current move to Anthropic shows that career paths in AI are not linear and that a change of environment can bring new breakthroughs.