4 of Google’s Top AI Brains Are Leaving—and Launching Their Own AI Startup

Jeff Dean, Google's chief scientist, alongside three other prominent artificial intelligence researchers, is departing the tech giant to establish a new start-up named Discovery Loop. This move highlights the shifting dynamics within Google and the continued availability of venture capital for innovative A.I. projects.

Dr. Dean, who was Google’s 30th employee, has been a pivotal figure in the company's history, significantly contributing to the development of its search infrastructure and A.I. research. Although his influence at Google has diminished in recent years, he has remained integral to the development of the company’s key technologies.

Discovery Loop joins a growing roster of start-ups pursuing a long-held ambition in Silicon Valley: creating A.I. that can enhance itself with minimal human intervention. “We think there is opportunity for A.I. to more fully automate what has traditionally been a very human-intensive experimental loop,” Dr. Dean stated in a recent interview. “You will get both a higher quantity and a higher quality of experiments, and that will lead to scientific breakthroughs and advances.”

As Google strives to compete with smaller, agile A.I. firms like OpenAI and Anthropic, many leading researchers are leaving the company, drawn by substantial venture funding and the opportunity to spearhead their own ventures. Over the past year, Google has seen the departure of several top figures, including Peter Norvig, its former head of research, who transitioned to the A.I. start-up Recursive Superintelligence.

“It is just a very different environment inside a start-up. It is much more fast-paced and focused,” remarked Tim Rocktäschel, who also left Google to co-found Recursive Superintelligence. “There is much less bureaucracy and politics.”

Dr. Dean, 58, is set to take on the role of chief executive at Discovery Loop, alongside longtime collaborators Sanjay Ghemawat, Quoc Le, and Oriol Vinyals. Their start-up has garnered seed funding from notable investors such as Radical Ventures and Khosla Ventures, with backing from Google’s parent company, Alphabet. Dr. Dean mentioned that Alphabet has agreed to provide computing resources for at least the first year to support their A.I. initiatives.

“Over 27 years, Jeff and Sanjay helped to drive some of the most significant technology transitions, from our early search infrastructure to the neural networks that helped create the modern A.I. era,” said Google’s CEO Sundar Pichai. He expressed personal gratitude for the opportunity to work alongside Dean and Ghemawat and confirmed Google's collaboration with the new venture.

Dr. Vinyals emphasized the significance of establishing their own company, stating, “Having extreme focus is very powerful. The ultimate form of extreme focus is when it is the only thing that matters — when it aligns with the mission of the company.”

Joining Google in 1999, Dr. Dean and Dr. Ghemawat designed complex software systems that enabled Google’s search engine to efficiently handle queries from billions of users. These innovations, including MapReduce and BigTable, became industry standards, influencing other tech giants like Amazon and Meta. Dr. Dean also founded Google Brain, a lab that catalyzed the industry's shift toward A.I.

Despite the initial successes, Google was slow to capitalize on advancements in chatbot technology, eventually merging Google Brain with DeepMind. The new entity, Google DeepMind, is now led by DeepMind founder Demis Hassabis, while Dr. Dean and Vinyals focus on developing Gemini, Google's response to ChatGPT.

While Google accelerated its A.I. efforts upon Sergey Brin's return, its technologies still lag behind competitors like OpenAI and Anthropic. Dr. Dean has also voiced dissent against the company's alignment with government policies, notably condemning federal actions against protesters in Minneapolis.

In a departure from Google’s structure, Dean announced that Discovery Loop would operate as a public benefit corporation, emphasizing a commitment to societal good. “There is a lot of good that can be done in the world if we can distribute scientific discoveries broadly,” he remarked. “We might make decisions that are not necessarily in the company’s purist financial interests.”