Jeff Dean, one of Google's longest-serving and most influential executives, is leaving the company to launch Discovery Loop, an AI startup built to accelerate scientific research. The public benefit corporation announced Wednesday that Alphabet, Google's parent company, joined an initial funding round co-led by Radical Ventures and Khosla Ventures.
Dean will serve as CEO. His co-founders are three senior Google researchers: Sanjay Ghemawat, a senior fellow and top engineer; Quoc Le, a founding member of Google Brain; and Oriol Vinyals, a senior research scientist at Google DeepMind.
What Discovery Loop plans to do
The company wants to use large-scale computing to run thousands of experiments at once, partially automating research so more iterations happen in less time. It also plans to use AI to build better AI, a process called recursive self-improvement, which would cut human iteration out of the loop entirely.
The founding team argues this slow, sequential approach is the main barrier to faster progress. "While science and engineering have tremendously advanced society over past centuries, progress has traditionally relied on slow, sequential human iterations, creating a significant bottleneck," the company said in a press release.
"We think there is opportunity for AI to more fully automate what has traditionally been a very human-intensive experimental loop," Dean told the New York Times. "You will get both a higher quantity and a higher quality of experiments, and that will lead to scientific breakthroughs and advances."
Funding and Dean's Google years
Radical Ventures and Khosla Ventures co-led the initial round, with participation from Kleiner Perkins, Lightspeed, and Doerr Capital.
Dean joined Google in 1999 as its 30th employee. He helped build core infrastructure for Google Search, including its crawling, indexing, and query-serving systems, and later took a leading role in the company's early AI research and in Gemini's multimodal models.
Why this matters for science and research professionals
Discovery Loop is trying to automate the full experimental loop: designing experiments, running them, and interpreting results. If it works, researchers will spend less time on repetitive work and more time deciding which questions matter.
The startup has not published technical details or a timeline, so its claims need evidence. Still, the caliber of the founding team suggests this direction will draw serious funding and competition.
Researchers who want to prepare for AI-assisted discovery can work through an AI Learning Path for Research Scientists, which covers experimental design and lab automation. Broader coverage is available under AI for Science & Research.
Your membership also unlocks:
