Elon Musk poached a Google scientist and justified it by telling Larry Page he shouldn't have been 'so cavalier about AI safety'
- A newly released biography on Musk details how he justified poaching a Google scientist to then-CEO Larry Page.
- Larry felt 'betrayed' and was mad at Musk, the latter told his biographer.
In 2015, Elon Musk and Sam Altman convinced one Google scientist to jump ship — with a $1.9 million salary and starting bonus — and join them in cofounding OpenAI.
Walter Isaacson's biography on Musk, written after shadowing the tech billionaire for three years and published on Tuesday, detailed the aftermath of the incident and recounted how Musk justified poaching talent from Google to the company's then-CEO Larry Page.
At the heart of their conflict was a disagreement between Musk and Page over the future of AI, which resulted in Musk attempting and failing to thwart Google's acquisition of AI company DeepMind in 2014.
A year later, when Musk and Altman were looking to start up their own competing AI lab, OpenAI, they poached Google scientist Ilya Sutskever to become a cofounder and chief scientist at the new company. Sutskever joined Google's AI unit, Google Brain, in 2013 along with Geoffrey Hinton — also known as the "godfather of AI."
"Larry felt betrayed and was really mad at me for personally recruiting Ilya, and he refused to hang out with me anymore," Musk told Isaacson about this escalation of tensions between him and Page.
"And I was like, 'Larry, if you just hadn't been so cavalier about AI safety then it wouldn't really be necessary to have some countervailing force," Musk added.
In the early stages of OpenAI, Musk and Altman agreed that having a large number of competing AI systems was better for AI safety — AI's threat to humanity has been a long-running concern of Musk — and encouraged openly sharing the company's source code, per the biography.
By 2018, however, Musk left OpenAI's board. In an interview in April this year, Musk criticized the company for being "obviously for-profit" and "closed-source" — referring to the company's shift from its non-profit roots.
On his part, Sutskever justified OpenAI's opaque and closed-source GPT-4 in a March 2023 interview with the Verge, by saying that the company's past approach to openly sharing research was wrong.
Sutskever is now co-leading a team at OpenAI focused on addressing the threat of "superintelligent" AI going rogue and endangering humanity.
For context, competition over AI talent has intensified further in the past year. Even non-AI firms are offering up to $300,000 for candidates with AI skills.
When Musk started his own AI startup — xAI — in July, he again poached AI experts from Google and OpenAI. Seven of the 12 employees making up the new company previously hailed from the two tech giants.
Musk and Page did not immediately respond to requests for comment from Insider, sent outside regular business hours. Sutskever did not immediately respond to a request for comment sent via OpenAI outside regular business hours.