Jeff Dean exits Google to co-found Discovery Loop startup

Jeff Dean, Google's chief scientist, is walking out after 27 years, and he is not leaving alone. As Wired reported alongside other outlets, three of his longest-running collaborators are going with him: Sanjay Ghemawat, Oriol Vinyals (a VP at DeepMind and a technical lead on Gemini), and Quoc Le (a Google Brain co-founder). Together they are launching Discovery Loop, a Delaware Public Benefit Corporation whose stated mission is to automate the full experimental loop of the scientific method, starting with machine learning research and expanding into hardware design, drug discovery, and clean energy.

The departure is being cushioned unusually well. Google itself is a founding investor and the cloud partner, and the seed round is being co-led by Radical Ventures and Khosla Ventures, with Lightspeed, Kleiner Perkins, Doerr Capital, and Alphabet also participating. Four of the people who helped build Google's AI infrastructure are getting to spin out a company with Google's blessing, Google's compute, and Google's money on the cap table.

Investors read it less charitably than the press release. Alphabet shares reportedly fell about 5% on the news, which arrived alongside a broader reshuffle at DeepMind: Demis Hassabis is stepping back from day-to-day leadership to become chair and Alphabet's chief scientist, and Koray Kavukcuoglu, the unit's CTO, is taking over as SVP running the group. All of this lands while the flagship version of Gemini remains unreleased against a planned June launch.

The honest caveat is that 'automate the scientific method' is a mission statement, not a product. Discovery Loop has not shown a demo or a benchmark, the seed round is not closed, and the startup declined to disclose its valuation. What the reporting does not tell you is how Discovery Loop's IP is walled off from Gemini and the research the founders were running last month, or what Vinyals's exit specifically does to the Gemini timeline. The forward-looking read is that if Google seeding its own competitors becomes the pattern, the winners are small labs that can now recruit a Google-scale co-investment without giving up Google-scale compute.