Mark Zuckerberg is using $100 million job offers to entice top AI talent.
Nothing exemplifies the talent war quite like a $100 million job offer. Mark Zuckerberg has embarked on an aggressive hiring campaign for AI’s top scientists, reaching out with cold emails and extending offers to join his new Superintelligence Labs division. Mark Zuckerberg has been actively recruiting some of the most renowned names in artificial intelligence.
The objective is clear: to develop artificial intelligence software that surpasses human intelligence.
The chief executive officer of Meta Platforms Inc. may seem to have no need for god-like AI, given the company’s success with ad targeting and recommendation software. However, this perspective overlooks the intense competition for the most coveted prize in technology, a battle that involves Alphabet Inc.’s Google and OpenAI. Zuckerberg is now positioned to make a strong push to reach the goal first. By attracting some of AI’s top minds with substantial funding and prior commitments to democratize AI, he has generated momentum among leading scientists who view his team as having a statistically greater likelihood of developing “super-intelligent” AI systems ahead of the competition.
In the past month, Zuckerberg has recruited prominent OpenAI scientist Lucas Beyer, co-creator of the vision transformer; Ruoming Pang, who spearheaded Apple Inc.’s AI model development; and Alexandr Wang, the former CEO of Scale AI, now co-leading Meta’s Superintelligence Labs. Wang’s situation resulted in a cost of billions for Meta. However, the outcome appears to create a halo effect, attracting other prominent figures in the industry, including investor Nat Friedman and Daniel Gross, CEO of Ilya Sutskever’s startup Safe Superintelligence. Meanwhile, the remaining top talent begins to worry about missing the opportunity to be the first to develop super-intelligent AI.
Money is certainly a strong motivator, yet many researchers are already affluent. Their field is highly ideologically charged and closely connected, driving them to seek the prestige of being published in Nature or contributing to the latest major AI model, as much as the allure of yachts and mansions. Zuckerberg’s public commitment to open-source AI through his Llama model has drawn the attention of scientists who argue that these systems can foster a more democratizing effect when made available to everyone. 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 consistently raised concerns about Zuckerberg’s commitment to investing in advanced AI models only to distribute them for free. The underwhelming performance of Llama’s latest models may compel the Meta CEO to explore more commercially viable strategies for AI. Meta’s models are behind those of Google DeepMind and OpenAI, with one variant ranked at 17 on a real-time leaderboard, and they incur higher operational costs.
Many researchers believe AI can ultimately address complex human issues such as aging, climate change, and cancer, and that, throughout history, technology has generally been beneficial for humanity. For many, the drive to develop that technology first is even more compelling, akin to the competitive nature in cancer research where scientists are as eager to win the race to a cure as they are to discover cures themselves.
Zuckerberg is providing researchers with significant computing resources — AI scientists require extensive arrays of powerful AI chips and data farms to create the most advanced models. This opportunity also presents the possibility of being the first to achieve superintelligence, a milestone that Zuckerberg likely values as much as others in the field.
