OpenAI Cofounder Durk Kingma Joins Anthropic

Yet another prominent ex-OpenAI researcher is joining Anthropic. Today, Diederik “Durk” Kingma, one of the co-founders of OpenAI, announced that he’s joining rival Anthropic. Kingma, who was most recently at Google, is renowned for co-authoring the highly influential paper “Adam: A Method for Stochastic Optimization.” He brings a wealth of expertise to Anthropic’s growing team of AI luminaries.

Kingma’s impressive career spans academia and industry. He earned his Ph.D. cum laude from the University of Amsterdam in 2017, focusing on deep learning and generative models. His contributions to the field are significant, including the development of Variational Autoencoders (VAE), the Adam optimizer, Glow, and Variational Diffusion Models.

Kingma’s move follows a pattern of high-profile departures from OpenAI to Anthropic. Earlier this year, Anthropic recruited OpenAI’s former safety lead Jan Leike and co-founder John Schulman.

“Anthropic’s approach to AI development resonates significantly with my own beliefs,” Kingma stated on social media. “I’m looking forward to contributing to Anthropic’s mission of developing powerful AI systems responsibly.”

Anthropic’s founders, Dario and Daniela Amodei, themselves left OpenAI in 2019 due to disagreements over the company’s direction, particularly its deepening ties with Microsoft. Under their leadership, Anthropic has successfully branded itself as a more safety-conscious alternative.

Kingma’s extensive background includes key roles at major tech players. After co-founding OpenAI in 2015, where he led the algorithms team, he joined Google Brain (later merged with DeepMind) in 2018 . At Google, he led various research projects, primarily focusing on generative models for text, image, and video.

Kingma did note that he will be working remotely from the Netherland, but will visit the Bay Area occasionally.