$100 million job offers: How Mark Zuckerberg is luring the best AI minds

By Parmy Olson

 

Nothing says talent war like a $100 million job offer. Mark Zuckerberg has been on a hiring blitz for AI’s most revered scientists, sending them cold emails and offering them roles in his new Superintelligence Labs division whose goal is nothing less than to build artificial-intelligence software that’s smarter than humans.

 

You might wonder why the Meta Platforms Inc. chief executive officer, whose company already prints money from clever ad targeting and recommendation software, needs to build god-like AI, but you’d be underestimating the hottest prize in tech, which Alphabet Inc.’s Google and OpenAI have been vying to win. Zuckerberg is now coming from behind with a viable shot at getting there first. Having attracted some of AI’s top brains with huge sums and previous pledges to make AI free for all and potentially more impactful, he’s now created momentum among other leading scientists who see his team as having a statistically higher chance of building “super-intelligent” AI systems before anyone else.

 

 

In just the last month, Zuckerberg has poached leading OpenAI scientist Lucas Beyer, who co-created the vision transformer; Ruoming Pang, who led Apple Inc.’s efforts at building AI models; and Alexandr Wang, the former CEO of Scale AI who now co-leads Meta’s Superintelligence Labs. In Wang’s case, the cost to Meta was billions. But the result seems to be a halo effect as other big names in the field join, such as investor Nat Friedman and Daniel Gross, the CEO of Ilya Sutskever's startup Safe Superintelligence, and the remaining top talent starts to fear missing out on being the first to build super-intelligent AI. 

 

Of course, money is a great motivator, but many of these researchers are already wealthy, and their field is so ideologically charged and so close-knit that they’re motivated by the glory of being published in Nature or having a hand in the biggest new AI model, just as much they are by the prospect of yachts and mansions.

 

Zuckerberg’s public commitment to open-source AI with his Llama model has already attracted scientists who believe such systems can have a more democratizing impact if they’re free for all. OpenAI made a similar bet early on, sharing much of its research freely “for recruitment purposes,” according to its then-chief scientist Sutskever, before taking that work behind closed doors.

 

Investors have long questioned Zuckerberg’s willingness to invest in advanced AI models and then give them away, and the lackluster performance of Llama’s most recent models may put pressure on the Meta CEO to consider more commercial approaches to AI. Meta’s models lag those of Google DeepMind and OpenAI — a variant is ranked at 17 in one real-time leaderboard — and they’re more expensive to run. 

 

Many researchers reckon AI can eventually solve intractable human problems like ageing, climate change and cancer, and that, overwhelmingly through history, technology has been a net good for humanity. But for many, the desire to build that technology first is even more powerful, a dynamic not so different to the field of cancer research where scientists want to win the race to a cure as much as they want to find cures at all. 

 

The key catnip Zuckerberg is offering researchers is vast amounts of compute — AI scientists need huge arrays of powerful AI chips and data farms to develop the most cutting-edge models —  and with it the chance of being first to cross the superintelligence threshold, which Zuckerberg likely cares about as much as anyone else in the field.  (Disclaimer: This is a Bloomberg Opinion piece, and these are the personal opinions of the writer. They do not reflect the views of www.business-standard.com or the Business Standard newspaper)