Ilya Sutskever

Sutskever was born in Russia and raised in Israel, became interested in computing at age five and emigrated to Canada with his family as a teenager. He was admitted to the University of Toronto’s mathematics program out of Grade 11 and graduated in 2005 with an honours BSc in mathematics.⁠[1]

He went on to earn a master’s and a PhD in computer science at Toronto as a student in the machine learning group under Geoffrey Hinton, co-authoring papers with Hinton on recurrent neural networks and sequence modelling, and held Ontario Graduate Scholarships (2006–2009) and a Google Graduate Fellowship (2010–2012). In 2013 he completed his doctoral thesis, Training Recurrent Neural Networks.⁠[1][2][3][4]

Hinton later recalled that Sutskever’s immediate reactions to problems were ones that had taken experts in the field a long time to arrive at.⁠[1]

In 2012, Sutskever, Alex Krizhevsky and Hinton published “ImageNet Classification with Deep Convolutional Neural Networks,” describing a large deep convolutional neural network with 60 million parameters, later known as AlexNet, trained to classify about 1.3 million high-resolution images; it took first place in the ILSVRC2012 visual object recognition contest. Its source code was later preserved by the Computer History Museum.⁠[5][3][1]

After his doctorate, Sutskever was a postdoc in Andrew Ng’s group at Stanford, then co-founded DNNresearch, Hinton’s spinoff company, which Google later acquired; he then spent about three years as a research scientist on the Google Brain team, contributing to the AlphaGo project.⁠[2][1]

In September 2014, Sutskever, Oriol VinyalsOriol VinyalsOriol Vinyals is an artificial intelligence researcher who joined Google around 2013 and worked at Google Brain and then DeepMind. He helped develop methods including sequence-to-sequence learning (seq2seq), knowledge distillation, and image caption generation, co-led the AlphaStar project that defeated professional StarCraft II players, and served with Jeff Dean as an overall technical lead of the Gemini models. In August 2026, he left Google with Jeff Dean, Sanjay Ghemawat, and Quoc Le to co-found the AI company Discovery Loop.Open the full entry and Quoc V. LeQuoc LeQuoc Le (Vietnamese name Le Viet Quoc) is an artificial intelligence researcher and early member of Google Brain who led the 2012 experiment in which a neural network running on 16,000 processor cores taught itself to recognize cats from YouTube videos, and who helped develop deep learning methods including sequence-to-sequence learning (seq2seq), paragraph vectors, and neural architecture search (AutoML). In August 2026, he left Google with Jeff Dean, Sanjay Ghemawat, and Oriol Vinyals to co-found the AI company Discovery Loop.Open the full entry published “Sequence to Sequence Learning with Neural Networks,” using a multilayered long short-term memory (LSTM) network to map an input sequence to a vector and decode it into an output sequence, achieving a BLEU score of 34.8 on the WMT’14 English-to-French task; sequence-to-sequence models were subsequently used in machine translation.⁠[6][1]

The same year he received the University of Toronto’s Innovator of the Year award.⁠[3]

In 2015, at age 29 and working on the Google Brain team, Sutskever was named to MIT Technology Review’s Innovators Under 35 list in the “Visionaries” category.⁠[7]

On December 11, 2015, Sutskever and Greg Brockman published the announcement of OpenAI, a non-profit AI research organization whose funders committed $1 billion; it named Sutskever as research director. He later became chief scientist and led the development of the large language models behind ChatGPT.⁠[8][1]

In The AGI Chronicles, published in 2026, Kevin RooseKevin RooseKevin Roose is a technology journalist and author who spent about nine years as a technology columnist at The New York Times and co-hosted the podcast Hard Fork, and is now a contributing writer at The Atlantic and co-host of Machine Gods, made with NPR. He has written four books, including The AGI Chronicles (2026), about the race to build artificial general intelligence.Open the full entry writes that in 2017 Sutskever and Brockman floated auctioning rights to future artificial general intelligence to governments including the United States, China and Russia, an idea that evolved from an international coalition to manage AGI risks into a fundraising strategy known internally as the “countries plan”; Dario Amodei objected in a December 2017 email. An OpenAI spokesperson said the company briefly considered turning itself into an international cooperative endeavor but did not consider selling AGI to China or Russia, and the plan was dropped after OpenAI secured MicrosoftMicrosoftMicrosoft is a technology company founded by Bill Gates and Paul Allen in 1975 and headquartered in Redmond, Washington, with businesses spanning Windows, Microsoft 365, Azure, enterprise software, artificial intelligence, gaming and devices.Open the full entry funding.⁠[9]

The AGI Chronicles also recounts that at an OpenAI retreat outside Yosemite in September 2022, Sutskever, Sam Altman and three other top executives appeared at a bonfire in white hotel bathrobes; Sutskever brought a wooden demonic statue he had bought, said it symbolized a misaligned AGI that had escaped human control while pretending to be aligned, placed it in the pit for a hotel worker to set alight, and shouted “For the glorious future of humanity!”⁠[10]

In Empire of AI, published in 2025, Karen Hao writes that in 2023 Sutskever told his team, “We’re definitely going to build a bunker before we release AGI,” adding that getting into the bunker would be optional.⁠[11]

Sutskever was named to TIMETIMETIME is an American newsmagazine and media brand founded by Henry Luce and Briton Hadden, which published its first issue on March 3, 1923; it is headquartered in New York and known for its Person of the Year selection and the TIME100 list. Since 2018 it has been personally owned by Marc and Lynne Benioff; its current Chief Executive Officer is Jessica Sibley, and its Editor-in-Chief is Sam Jacobs.Open the full entry’s list of the 100 most influential people in AI in 2023 and 2024.⁠[1]

On July 5, 2023, OpenAI announced its Superalignment team, co-led by Sutskever and head of alignment Jan Leike, with the goal of solving the core technical challenges of superintelligence alignment within four years; the company committed 20% of the compute it had secured to date to the effort over four years.⁠[12]

On November 17, 2023, OpenAI’s board announced that Sam Altman was leaving as CEO, saying he had not been “consistently candid in his communications with the board”; the board then comprised chief scientist Sutskever and independent directors Adam D’Angelo, Tasha McCauley and Helen Toner.⁠[13]

In her book The Optimist, Keach Hagey writes that Sutskever and chief technology officer Mira Murati gathered material on Altman’s conduct, including screenshots from Murati’s Slack channel, and that after Sutskever provided it to board members, they moved to oust Altman and appoint Murati as interim CEO. Karen Hao’s Empire of AI likewise assigns much of the responsibility for the decision to Sutskever and Murati.⁠[14][11]

Altman returned within a week. On November 29, 2023, OpenAI announced a new initial board of Bret Taylor (chair), Larry Summers and Adam D’Angelo, with Sutskever no longer serving as a director; Altman said “I love and respect Ilya” and that the company was discussing how he could continue his work at OpenAI.⁠[15][11]

On May 14, 2024, Sutskever announced that “after almost a decade” he had decided to leave OpenAI; Altman called him “easily one of the greatest minds of our generation,” and research director Jakub Pachocki succeeded him as chief scientist. Superalignment co-lead Jan Leike resigned around the same time.⁠[16]

On June 19, 2024, Sutskever, former Y Combinator partner Daniel Gross and former OpenAI engineer Daniel Levy announced Safe Superintelligence (SSI), with offices in Palo Alto and Tel Aviv; the company called safe superintelligence its “mission, our name, and our entire product roadmap” and said it planned to advance capabilities as fast as possible while keeping safety ahead.⁠[17][18]

On September 4, 2024, SSI announced it had raised $1 billion from NFDG, a16zAndreessen HorowitzAndreessen Horowitz (a16z) is a venture capital firm founded in 2009 by Marc Andreessen and Ben Horowitz. It invests from the seed stage through the growth stage and provides operational support to its portfolio companies.Open the full entry, Sequoia, DST GlobalDST GlobalDST Global is a technology investment firm founded by Yuri Milner in 2009 that focuses on growth-stage investments in internet companies worldwide; it has invested in companies including Facebook, Twitter, Alibaba, JD.com, Airbnb and Spotify, and in 2026 led the US$400 million Series B round of the AI visual content company Higgsfield.Open the full entry and SV AngelSV AngelSV Angel is a San Francisco venture capital firm cofounded in 2009 by angel investor Ron Conway and former Google executive David Lee. It typically writes small follow-on checks in seed rounds and has invested in companies such as Airbnb, Pinterest, and Dropbox; in 2018 it said it would no longer raise new funds from outside investors, and from 2022 it set up growth funds.Open the full entry, to be used for computing power and hiring; the round reportedly valued the company at $5 billion.⁠[19][20]

In October 2024, after winning the Nobel Prize in Physics, Geoffrey Hinton said he was “particularly proud of the fact that one of my students fired Sam Altman,” referring to Sutskever.⁠[21]

In December 2024, Sutskever accepted an award and spoke at the NeurIPS conference, saying superintelligence would eventually be achieved and would be “different, qualitatively” from current AI: such systems would be agentic “in a real way,” and their reasoning would make them more unpredictable.⁠[22]

In 2025, SSI raised $2 billion at a $32 billion valuation and partnered with Google Cloud to power its research; its backers also include Alphabet, Lightspeed Venture PartnersLightspeed Venture PartnersLightspeed Venture Partners backs companies from seed through late-stage financing. Its website distinguishes Lightspeed Management Company (LSVP) from Lightspeed India Partners (LSIP): the two are separately operated investment advisers, with the U.S., European, and Israeli offices belonging to LSVP and the India and Southeast Asia offices belonging to LSIP.Open the full entry and GV, among others.⁠[23]

On June 6, 2025, the University of Toronto conferred an honorary Doctor of Science degree on Sutskever. The university noted that he has been elected to the Royal Society in the United Kingdom.⁠[1]

On July 3, 2025, Sutskever wrote to the team and investors that co-founder and CEO Daniel Gross had officially left SSI as of June 29, that he himself was now formally CEO, and that Daniel Levy was president. Meta CEO Mark Zuckerberg had reportedly tried to acquire SSI and was in talks to hire Gross; Sutskever said the company was focused on its own work.⁠[19][24]

On July 26, 2026, SSI announced a long-term strategic partnership with NVIDIA that it said would let it scale its compute by 10 times; NVIDIA also made an undisclosed investment, and SSI is to use NVIDIA’s Vera Rubin GPU platform. NVIDIA said it signed the partnership after obtaining rare access to the company’s closely guarded research. SSI follows a “straight shot” research approach without commercial product releases or short-term revenue cycles.⁠[19][23]

Personal website: Public link: https://www.cs.toronto.edu/~ilya/ (opens in a new window)⁠[2]

Safe Superintelligence: Official website: https://ssi.inc/ (opens in a new window)⁠[18]

OpenAI: Official website: https://openai.com/ (opens in a new window)⁠[8]