🌀

Ilya Sutskever

Original nameIlya Sutskever

Israeli-Canadian Computer Scientist, Co-Founder and Former Chief Scientist of OpenAI

The Laws Behind the Technology · Theoretical Founders

AlexNetThe "Scaling Law" Idea of the GPT SeriesOpenAI

Who they are

Ilya Sutskever (1986– ) is an Israeli-Canadian computer scientist who studied under Geoffrey Hinton during his doctorate at the University of Toronto, and in 2012, with Hinton and Alex Krizhevsky, designed "AlexNet," winning overwhelmingly in the ImageNet image-recognition competition, the first major breakthrough of his academic career. In 2015 he became one of the co-founders of OpenAI and long served as chief scientist, one of the core figures driving the institution to hold to the "scaling law" idea of "raising AI capability by continually enlarging model scale and training-data volume," an idea thereafter proven the key methodological basis for the continual leap in the capability of the GPT series of models. In November 2023 Sutskever was caught up in the governance storm of the OpenAI board’s sudden removal of then-CEO Sam Altman, being one of the board members who drove the decision, after which the situation swiftly reversed and Altman was reinstated within days, while Sutskever gradually withdrew and formally left OpenAI in 2024 to found a new company called "Safe Superintelligence Inc.," focusing on advancing general AI research on the premise of ensuring safety.

Primary sourcesKrizhevsky, Sutskever, Hinton, "ImageNet Classification with Deep Convolutional Neural Networks" (2012, the AlexNet paper)

Key stories

The Fame-Making Work of His Doctoral Years

While pursuing his doctorate under Hinton, Sutskever, with the fellow graduate student Krizhevsky, designed the deep convolutional neural network "AlexNet," which won the 2012 ImageNet competition by a margin far above the competitors, a result that directly triggered the whole AI field’s turn to deep learning and established Sutskever’s core standing in the field early in his academic career.

Taking Part in a Removal and Then Bowing Out

In November 2023 Sutskever, as one of the OpenAI board members, took part in driving the decision to remove Altman as CEO, but this decision provoked a strong backlash within and outside the company, and Altman was reinstated a mere five days later, while Sutskever thereafter gradually withdrew from the company’s core decision-making and formally announced his departure from OpenAI in May 2024 to found a new company focused on AI-safety research — this experience of "taking part in a removal and then being ousted" is a case often mentioned in examining the internal divergences of ideas in the AI industry.

Relationships

Mentor

🧠Geoffrey Hinton 🔮The "Transformer Eight"　The Transformer architecture was the core technical basis for the later continual scaling and capability improvement of OpenAI’s GPT series of models

Nemesis

🤖Sam Altman

Echoes today

Below are how modern works borrow or reinterpret this name or story — not the original material. The two differ, so keep them apart.

The "Safe Superintelligence" companyThe new company Sutskever founded after leaving OpenAI deliberately chose not to pursue short-term commercial products but to focus on long-term AI-safety research, a positioning that is itself his stance on the industry controversy of "how to balance speed of development and safety."

Appears in

The Explosion of Deep Learning and Generative AIPerseverance in the "Neural Network Winter" and the Shock of AlexNet The Explosion of Deep Learning and Generative AIChatGPT’s Hundred Million Users in Two Months, a CEO Removed and Reinstated in Five Days

← Back to Figures