Ilya Sutskever

• 1. Introduction

  ›

  Ilya Sutskever, a computer scientist born in 1986, wrote a short message in 2022 that set off a public argument. It read: "it may be that today's large neural networks are slightly conscious." Few researchers could post a line like that and be taken seriously. Sutskever could, because his work sits underneath much of the machine learning people now use every day. He specializes in the field, and he is counted among the most highly cited computer scientists in history.

  How does a boy who changed countries twice before his sixteenth birthday end up shaping the systems behind ChatGPT? Why did he later vote to remove the chief executive of the company he helped found, and then regret it? And what is he trying to build now that he has left that company behind?

  33342710

• 2. Three Countries by Sixteen

  ›

  Gorky was the name of the city where Sutskever was born in 1986. Today it is called Nizhny Novgorod, and it lies in Russia, which was then part of the Soviet Union. His family was Jewish. When he was 5, they immigrated to Israel, and he grew up in Jerusalem.

  School came easily to him. In eighth grade he began taking classes at the Open University of Israel, well before most students his age would touch university material. At 16 his family moved again, this time to Canada. He sat in a Canadian high school for only a month. Then the University of Toronto in Ontario admitted him as a third-year undergraduate.

  At Toronto he earned a bachelor's degree in mathematics in 2005. A master's degree in computer science followed in 2007, and a PhD in computer science in 2013. His doctoral advisor was Geoffrey Hinton. Years later, that same advisor would publicly weigh in on one of Sutskever's most controversial decisions.

  1220156922

• 3. Networks That See and Translate

  ›

  In 2012, Sutskever worked with Hinton and a fellow student, Alex Krizhevsky, to build AlexNet. It is a convolutional neural network, and the three of them are credited as its co-creators.

  That same year he spent roughly two months at Stanford University as a postdoc under Andrew Ng. He then returned to Toronto and joined DNNResearch, a new company that Hinton had spun off from his research group. In 2013, Google acquired DNNResearch and hired Sutskever as a research scientist at Google Brain.

  Inside Google Brain, Sutskever teamed up with Oriol Vinyals and Quoc Viet Le. Together they created the sequence-to-sequence learning algorithm. He also worked on TensorFlow, and his name appears among the many co-authors of the AlphaGo paper. His broader list of contributions includes CLIP, DALL-E, and the GPT models.

  Google did not hold him for long. At the end of 2015 he left to become a cofounder of a newly founded organization called OpenAI.

  282293024

• 4. The Scaling Ethos

  ›

  Sutskever is credited with establishing the scaling ethos at OpenAI, where he served as chief scientist. In that role he oversaw the research breakthroughs that led to large language models. Those same breakthroughs led to the launch of ChatGPT. Later he led the research that produced reasoning models such as o1.

  In 2023, he announced that he would co-lead a new effort called Superalignment. Its goal was to solve the alignment of superintelligences, and it set a deadline of four years. He wrote that superintelligence might seem far off, yet it could still arrive this decade.

  At that moment he was also one of six members of the nonprofit board that controlled OpenAI.

  3827103137

• 5. The Board Vote

  ›

  In November 2023, that board fired Sam Altman. Its public statement said Altman "was not consistently candid in his communications with the board".

  Behind the statement stood a 52-page memo that Sutskever himself had written. It leaned heavily on information from Mira Murati. It accused Altman of lying, manipulating executives, and fostering division inside the company. Sutskever submitted it after months of tension and dissatisfaction with Altman's leadership style, and he then joined the board in voting to remove him.

  Shortly afterward, at an all-hands meeting, Sutskever described the firing as "the board doing its duty". Within the following week he expressed regret for having taken part. Altman was reinstated a week after being fired.

  The fallout spread. Co-founder Greg Brockman resigned, and three senior researchers followed him out the door. Sutskever stepped down from the board and stopped appearing at OpenAI's office. Reports disagreed about what that meant. Some sources suggested he was leading his team remotely, while others said he no longer had access to its work.

  In an interview in October 2024, after winning the Nobel Prize in Physics, Hinton said he supported Sutskever's decision to fire Altman. He stressed his own concerns about AI safety.

  4041434455

• 6. Safe Superintelligence

  ›

  In May 2024, Sutskever announced that he was leaving OpenAI to focus on a project he called "very personally meaningful". His exit followed a turbulent stretch of leadership crises and internal debates over the direction of AI development and alignment. Hours later, Jan Leike, the other leader of the superalignment project, announced his own departure. Leike cited an erosion of safety and trust in OpenAI's leadership.

  In June 2024, Sutskever announced Safe Superintelligence Inc., founded with Daniel Gross and Daniel Levy. The company has offices in Palo Alto and Tel Aviv. He drew a sharp line between it and OpenAI, which sells revenue-generating products. Sutskever said the new company's "first product will be the safe superintelligence, and it will not do anything else up until then".

  Money followed quickly. In September 2024 the company announced it had raised $1 billion from venture capital firms including Andreessen Horowitz, Sequoia Capital, DST Global, and SV Angel. In March 2025 it raised $2 billion more and reportedly reached a $32 billion valuation, notably due to Sutskever's reputation.

  In June 2025, the company turned down an offer from Meta Platforms to buy it. Soon afterward, Gross left for Meta, and Sutskever took over as chief executive of the company.

  3247495054

• 7. Honors and Holdings

  ›

  MIT Technology Review put Sutskever on its 35 Innovators Under 35 list in 2015. In 2018 he gave keynote talks at Nvidia Ntech 2018 and the AI Frontiers Conference 2018.

  In 2022, he was elected a Fellow of the Royal Society. Time included him among its 100 most influential people in AI in both 2023 and 2024. From 2022 through 2024, he won the Test of Time award from Neural Information Processing Systems three years running. That award recognizes papers that shaped the field over at least ten years.

  In 2025, the University of Toronto, his alma mater, gave him an honorary doctorate. In 2026 he received the National Academy of Sciences Award for the Industrial Application of Science, which was presented for the first time in artificial intelligence.

  The same year brought a very different kind of record. During the Musk v. Altman trial in 2026, Sutskever confirmed that he held a $7 billion stake in OpenAI.

  6063646870