Latest update
"Too poor to use GPT-5 all the time": Stanford's Yejin Choi and Carlos Guestrin raise $110M for AI lab Noeri
First seen on X 25 hours ago@YejinChoinka ♥ 1,6456 views
Stanford professors Yejin Choi and Carlos Guestrin have launched Noeri, an AI lab backed by $110 million, betting on many specialized AIs that companies grow from their own data rather than one giant model.
Menlo Ventures led the round with Madrona, Kindred Ventures, Wing Venture Capital and NEA, Bloomberg and TNW reported Nov 16. Cisco, Nvidia and Snowflake joined as strategic investors. The valuation wasn't disclosed. The two are co-CEOs and are hiring in Palo Alto and Zurich.
Noeri's pitch is AI that keeps improving on the job. It tunes the agents a company already runs using real work: changing instructions, tools and the model used for each step, and training custom models when needed. A person approves each change, like a code review. It works with closed and open-weight models, and customers keep ownership of their data and workflows. No customers have been named.
The idea grew out of a Stanford microbiology and immunology team that wanted cheaper access to OpenAI models; Choi says a small AI built for them beat GPT-5 on their research needs, a claim not independently verified. "We are too poor to use GPT-5 all the time for research," she told Bloomberg. Guestrin said both founders worry "about centralization, about one company owning everything."
Choi, a MacArthur Fellow, previously taught at the University of Washington and worked at the Allen Institute for AI and Nvidia. Guestrin co-founded Turi, acquired by Apple in 2016, led Apple's central ML organization and co-created XGBoost. Choi's launch post drew 750,000 views and 1,600 likes.
Comments 0
Select a sentence in the story (long-press on phones) to comment on it.
“Through my wonderful time at NVIDIA and Stanford, one question has kept me up at night: Should the future of AI be monolithic or plural?”
