The A.I. researcher left OpenAI earlier this year after playing a key role in the attempted ousting of its CEO Sam Altman.
Ilya Sutskever, a pioneering A.I. researcher and OpenAI co-founder who recently left the high-flying A.I. company after a failed attempt to oust CEO Sam Altman, is launching a new A.I. company that aims to put safety first. Known as Safe Superintelligence, Inc. (SSI), the startup will focus singularly on safely creating an A.I. system able to surpass humans in intellect. “We will pursue safe superintelligence in a straight shot, with one focus, one goal, and one product,” said Sutskever of the new firm in a post on X.
Concerns regarding the existential threat posed by A.I. have been rife among those working with the technology and have led to divides in safety approaches across leading companies like OpenAI. Shortly after Sutskever stepped down from OpenAI in May after a nearly ten-year tenure, Jan Leike—with whom he co-ran a since-disbanded safety team overseeing A.I.’s long term risks—also resigned and said the company’s “culture and processes have taken a backseat to shiny products.”
https://x.com/ilyasut/status/1803472979873128498
SSI’s other co-founders include Daniel Gross, who formerly led A.I. efforts at Apple (AAPL) and served as a Y Combinator partner; and Daniel Levy, who previously worked alongside Sutskever at OpenAI as a member of its technical staff. While the trio has not revealed the specifics of how exactly they will pursue “safe” A.I., they described SSI’s mission as one that will lack “distraction by management overhead or product cycles” and remain “insulated from short-term commercial pressures.” The startup, which will have offices in Palo Alto, Calif. and Tel Aviv in Israel, has also yet to disclose its business model or investors.
Sutskever’s first major break in the world of A.I. took place in 2012 when he was studying at the University of Toronto under Geoffrey Hinton, an A.I. academic often referred to as the “Godfather of A.I.” An A.I. lab created by Sutskever, Hinton and another student was acquired by Google, which brought on Sutskever as a research scientist in 2013.
The computer scientist eventually joined OpenAI as a co-founder and chief scientist in 2015. Both OpenAI and Demis Hassabis’ Google DeepMind were interested in hiring Sutskever at the time, said OpenAI co-founder Elon Musk during a recent podcast. “That was one of the toughest recruiting battles I’ve ever had,” he said.
Musk, who has been vocal about the dangers of A.I., is seemingly not convinced by Sutskever’s newest venture. “Any given A.I. startup is doomed to become the opposite of its name,” the Tesla and SpaceX CEO wrote on X.
https://x.com/elonmusk/status/1803783427616821401
Sutskever made headlines last year as a member of the four-person OpenAI board that claimed to have lost trust in Altman and briefly removed him from the company in November before swiftly reinstalling the CEO amid pushback from employees and investors. Shortly after the failed ousting, Sutskever appeared to have a change of heart and said he “deeply regretted” his role in the firing. Sutskever himself stepped down as chief scientist of the A.I. company six months later. At the time, he hinted at SSI by stating that he was moving on to a project that is “personally meaningful to him.”
