(Bloomberg) — Meta Platforms Inc.’s lavish multimillion-dollar budget for recruiting top AI talent may not guarantee success, said Helen Toner, former OpenAI board member and director of strategy at Georgetown’s Center for Security and Emerging Technology.
The poaching of artificial intelligence researchers from the likes of OpenAI — with salaries in the tens of millions of dollars — and the debut of Meta’s new Superintelligence group comes after the Facebook operator developed a reputation for “having a dysfunctional team,” Toner said in an interview with Bloomberg TV. The practice of luring away high performers from each other’s AI labs has intensified among Silicon Valley companies since the launch of ChatGPT, she said.
“The question is, can it turn around Meta’s fortunes and turn it into a real juggernaut?” Toner said. “It’ll be difficult, there’s a lot of organizational politics at play.”
Meta’s troubles began compounding when China’s AI upstart DeepSeek came out of nowhere this year and put forward credible competition to Meta’s open source models. “The fact that DeepSeek was outshining them was really not a good look for the company,” according to Toner.
Chief Executive Officer Mark Zuckerberg is now plowing financial resources in, but whether he’ll be able to change organizational dynamics and make progress fast enough to retain top engineers is an open question. “Can Meta convince them that they are moving fast enough?” Toner said.
Toner, an influential voice in artificial intelligence, came into the limelight first as a board member of OpenAI and then for her vote to oust Sam Altman from the CEO post in late 2023. The Melbourne-educated academic departed from the board following Altman’s brief stepping down and restoration to the top job, and has since advanced her career in studying the AI race between the US and China.
That race is now spilling across borders as the two superpowers vie for the business and collaboration of other countries, Toner said. US companies like OpenAI and Chinese players like Alibaba Group Holding Ltd., DeepSeek and Zhipu AI are making plays for international partnerships with governments and businesses. South Korea’s Kakao Corp. is integrating ChatGPT and other AI services into the country’s most used social media platform, while Alibaba is adding new data centers in Southeast Asia.
China has a long history of working with other governments and is chipping away at the US tech monopoly globally, Toner said. “It’s certainly a strong showing they’re making,” she said. China’s models are widely available even if they are less technically sophisticated. They compete on the basis that they’re “cheaper, easier to use, and they help you adopt and customize.”
Toner hasn’t interacted with Altman since their clash in the November 2023 OpenAI boardroom battle. “At some point, we’ll wind up at the same event, the AI world is pretty small,” she said. “I’m sure we’ll both be happy to shake each other’s hand.”
More stories like this are available on bloomberg.com
