OpenAI Founding Member Who Coined “Vibe Coding” Joins Anthropic

Andrej Karpathy has joined Anthropic, the AI researcher announced on Tuesday. The move returns him to hands-on frontier lab work after more than a year of independent projects.

The hire places one of the most visible figures in AI research at a direct competitor to OpenAI. Karpathy was a founding member there before two separate stints at the company.

From OpenAI Founding Member to Anthropic Researcher

Karpathy joined the original OpenAI team in 2015. He left in 2017 for Tesla and rejoined OpenAI in 2023 for about a year. He departed again in February 2024.

Karpathy framed that second exit as a personal choice rather than the result of any internal dispute. He moved into independent work soon after.

He launched Eureka Labs, an AI education startup. He also built a large following through his Zero to Hero video series on neural networks and language models.

The Anthropic announcement marks his first return to a major frontier lab since that departure.

I’ve joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative… I remain deeply passionate about education and plan to resume my work on it in time,” he shared in a post.

Follow us on X to get the latest news as it happens

Coining Vibe Coding Between Stints

Karpathy created a phrase that reshaped how developers talk about AI-assisted programming. In a February 2025 post, he introduced the term vibe coding.

He defined it as a workflow where the model writes code. The user accepts changes without reading the diff. The term entered mainstream developer vocabulary within weeks.

BeInCrypto coverage has documented its use in building a crypto trading bot in a single weekend. The same workflow helped separate winning crypto exchanges from laggards.

It also lowered the entry barrier for Web3 builders with no programming background.

Karpathy has since refined the idea into what he calls agentic engineering. Humans focus on specifications and oversight while autonomous agents handle execution.

Frontier LLM Research the Next Chapter

Anthropic was founded in 2021 by former OpenAI researchers. The company has positioned itself as the safety-focused alternative to its larger rival.

Its Claude model family competes directly with GPT. Anthropic shipped Opus 4.7 in April with stronger long-form reasoning and vision capabilities.

The Karpathy hire follows a steady stream of senior moves between Anthropic and OpenAI in recent quarters. OpenAI countered days after the Opus 4.7 release with GPT-5.5.

That model was pitched as OpenAI’s most capable system for autonomous, multi-step work. The talent war between the two labs has steadily intensified.

Neither Karpathy nor Anthropic has disclosed the specific team or projects he will work on. His public statement points to LLM training and agentic systems.

His prior research on neural network design, computer vision, and synthetic data has direct application in those areas.

Karpathy also said he plans to resume his education work in time. Eureka Labs and related output will likely continue alongside his Anthropic role.

The coming months will show which areas of Anthropic’s research Karpathy gravitates toward. His arrival may also shift how the company presents its frontier work to developers.

Subscribe to our YouTube channel to watch leaders and journalists provide expert insights