Anthropic poaches another OpenAI co-founder

Another chapter in the great AI talent reshuffle just dropped, and Anthropic is once again reaping the rewards.

Durk Kingma, an OpenAI co-founder and fundamental machine learning researcher, announced on X that he is joining Anthropic. Kingma noted that he will be working mostly remotely from his home base in the Netherlands, though he did not share which exact division he will join or lead.

In his public post, Kingma stated that Anthropic’s approach to AI development resonates deeply with his own personal beliefs. He emphasized his excitement to build powerful AI systems responsibly alongside former colleagues from both OpenAI and Google.

The architect behind foundational generative algorithms

While Kingma kept a much lower public profile than figures like Sam Altman, his technical footprint across modern generative AI is massive.

After earning his machine learning doctorate at the University of Amsterdam and working as a doctoral fellow at Google, Kingma joined OpenAI as a founding research scientist. There, he led the algorithms team, helping lay the theoretical and algorithmic groundwork that eventually powered breakthrough systems like DALL-E 3 and ChatGPT.

In 2018, Kingma stepped back into angel investing and advising before returning to Google Brain later that summer. He remained with the organization through its high-profile merger with DeepMind in 2023, keeping him at the cutting edge of frontier model research.

Anthropic keeps assembling the original OpenAI crew

Kingma’s arrival is just the latest domino to fall in an extraordinary string of talent poaching by Anthropic.

Back in May, OpenAI’s former alignment and safety lead Jan Leike walked out the door and joined Anthropic, followed shortly by Instagram co-founder Mike Krieger stepping in as head of product. In August, fellow OpenAI co-founder John Schulman also announced his jump to Anthropic.

Anthropic itself was born out of this exact dynamic. CEO Dario Amodei was OpenAI’s former VP of research before leaving over disagreements regarding commercialization speed and direction, taking early leaders like former policy head Jack Clark with him.

Why it matters

We think this move shows that the talent war at the frontier of AI is no longer just about who has the biggest compute cluster; it is about corporate culture and mission alignment.

For enterprise customers and developers, Anthropic’s ability to hoard this level of algorithmic brilliance suggests Claude’s long-term research pipeline is in exceptionally good hands. Kingma’s pedigree in foundational generative algorithms means Anthropic is investing directly in core model architectures rather than just tweaking product features.

For the broader industry, OpenAI’s rapid pivot toward an aggressive, profit-driven corporate structure appears to be taking a genuine toll on its scientific morale. By positioning itself as the responsible, safety-first haven for top-tier researchers, Anthropic is successfully recreating the early intellectual powerhouse of OpenAI under its own roof, while OpenAI is left having to backfill some of its most foundational minds.