Long before he became one of the most recognisable names in artificial intelligence, Ilya Sutskever was just another shareholder standing in line at a microphone, waiting to ask Elon Musk a question about cars.
The moment came at Tesla’s annual shareholder meeting on June 9, 2015, held at the Computer History Museum in Mountain View. A young, curly-haired Sutskever, at the time a researcher at Google Brain, introduced himself simply as a shareholder and asked Musk to weigh in on the future of Autopilot — specifically, when the feature would evolve into a fully autonomous system, one reliable enough that a driver could safely fall asleep at the wheel.
“I’d like you to say a few words about the Autopilot, specifically when it will become a full go-to-sleep-in-the-car solution, given the rate at which you see it being developed, and given what you know about how quickly regulation and laws can change to accommodate that,” Sutskever asked.
Musk’s answer was, in hindsight, a preview of a pattern that would come to define much of his public commentary on self-driving cars in the decade that followed. “From a technical feasibility standpoint,” he said, “I think it’ll be technically possible in approximately three years.” He also flagged that regulatory approval could take longer still, since laws around autonomous vehicles would need time to catch up with the technology.
Three years came and went in 2018, and full autonomy did not arrive. Musk has continued to push out similar timelines for “full self-driving” nearly every year since, a habit that has been well documented and mocked in equal measure.
A Recruiting Battle Already Underway
What makes the exchange notable in hindsight isn’t really the Autopilot forecast — it’s who was asking the question. At the time, Sutskever was a rising star in deep learning circles at Google, and unbeknownst to most people in that room, he was likely already being courted for a new AI venture that Musk was helping put together.
Musk has since described the tug-of-war over Sutskever’s talent in fairly dramatic terms. In a conversation on the Lex Fridman podcast, Musk said his once-close friendship with Google co-founder Larry Page effectively ended over the recruitment of Sutskever, calling it “a tough recruiting battle” fought mainly between himself and DeepMind’s Demis Hassabis, with Sutskever going back and forth before finally deciding to jump ship.
That jump happened just months after the Tesla meeting. By December 2015, Sutskever had left Google to become a co-founder and the first Chief Scientist of OpenAI, the research lab Musk was backing alongside Sam Altman, Greg Brockman, and others, with a stated mission of ensuring artificial general intelligence would benefit humanity rather than being controlled by a handful of corporations.
From Shareholder Mic To Chief Scientist
Sutskever would go on to spend nearly a decade at OpenAI, becoming one of the most cited researchers in the field and playing a central role in the technical work that eventually led to ChatGPT. He later found himself in the middle of one of Silicon Valley’s most dramatic corporate episodes, voting in favour of ousting Sam Altman as CEO in November 2023 before Altman’s swift reinstatement days later. Sutskever departed OpenAI the following year and went on to found his own venture, Safe Superintelligence Inc., alongside Daniel Gross and Daniel Levy, dedicating the company entirely to research rather than commercial products.
Musk’s own relationship with OpenAI, meanwhile, has taken a very different path. He stepped away from the company’s board in 2018 and has since become one of its most vocal critics, even suing Altman and OpenAI over allegations that the lab abandoned its founding nonprofit mission — a legal battle that reached trial in 2026.
It’s a strange bit of tech history trivia: a full head of hair, a modest question about when cars might drive themselves, and a young researcher who, within months, would help build one of the most consequential companies of the AI era.