When Mark Zuckerberg revamped Meta’s artificial intelligence operations this year, he recruited a new leader — Alexandr Wang, a 28-year-old entrepreneur — to build a team of top researchers from rivals like OpenAI and Google.
That team, called TBD Lab (for “to be determined”), was placed in a siloed space next to Mr. Zuckerberg’s office at the center of Meta’s Silicon Valley headquarters, surrounded by glass panels and sequoia trees. Mr. Zuckerberg wanted to separate the new A.I. group from the bureaucracy of the company, which owns Facebook, Instagram and WhatsApp, two people with knowledge of the matter said.
Five months later, that divide has become more than physical.
In meetings this fall, Mr. Wang has privately told people that he disagreed with some of Mr. Zuckerberg’s longtime lieutenants, including Chris Cox, the chief product officer, and Andrew Bosworth, the chief technology officer, the people briefed on the conversations said.
In one case, Mr. Cox and Mr. Bosworth wanted Mr. Wang’s team to concentrate on using Instagram and Facebook data to help train Meta’s new foundational A.I. model — known as a “frontier” model — to improve the company’s social media feeds and advertising business, they said. But Mr. Wang, who is developing the model, pushed back. He argued that the goal should be to catch up to rival A.I. models from OpenAI and Google before focusing on products, the people said.
The debate was emblematic of an us-versus-them mentality that has emerged between Meta’s new A.I. team and other executives, according to interviews with half a dozen current and former employees of the A.I. business. TBD Lab’s researchers have come to view many Meta executives as interested only in improving the social media business, while the lab’s ambition is to create a godlike A.I. superintelligence, three of them said.
To pursue TBD Lab’s aim, Meta is pushing other departments to the sidelines. Mr. Bosworth was recently asked to slash $2 billion off next year’s proposed budget for Reality Labs, the virtual reality and augmented reality division that he oversees, two people said. The money went to the budget for Mr. Wang’s team, they said. (A Meta spokesman disputed the claim and said next year’s budget was not yet final.)
Some Meta employees have also disagreed over which division gets more computing power. Those working on the social media ranking algorithm have argued that more of the company’s new computing power should be used to improve their business rather than training A.I. models, people with knowledge of the matter said.
Juggling the tensions falls to Mr. Zuckerberg. The 41-year-old has long surrounded himself with an inner circle of loyalists, but has been open about how much he wants to win the A.I. race. He has invested billions — including $14.3 billion in Mr. Wang’s A.I. start-up — and declared that superintelligence will lead users into “a new era of individual empowerment.”
“Meta spent a huge amount of money to bring in Alexandr Wang, and that must have changed both the internal topology and the feeling of the company,” said Tomasz Tunguz, a general partner at the venture capital firm Theory Ventures.
“You have an outsider who now has carte blanche,” Mr. Tunguz added. “That probably means that a lot of previous ways of working are disregarded, which is hard.”
In a statement, a Meta spokesman, Dave Arnold, disputed that Mr. Wang and TBD Lab have had conflicts with Mr. Cox and Mr. Bosworth. He declined to make the executives available for interviews.
“Not only is Meta’s leadership aligned on building superintelligence while continuing to grow the core business, but our company is an outlier in how much our A.I. investment is improving the ads and recommendations systems at the heart of our business,” Mr. Arnold said.
Mr. Zuckerberg began retooling Meta’s A.I. efforts this spring after its newest A.I. models underperformed those of competitors. To catch up, he invested in Mr. Wang’s start-up in June and hired the entrepreneur to be the new chief A.I. officer. Mr. Wang is not an engineer but is seen as one of the best-connected A.I. entrepreneurs.
Mr. Zuckerberg and Mr. Wang then went on a recruiting blitz for top A.I. researchers, offering pay packages worth hundreds of millions. Mark Chen, OpenAI’s chief research officer, said Mr. Zuckerberg had gone as far as delivering homemade soup to the doorsteps of some OpenAI employees to persuade them to join his company.
(The New York Times has sued OpenAI, claiming copyright infringement of news content related to A.I. systems. OpenAI has denied those claims.)
The new roster of all-star researchers arrived on Meta’s campus in Menlo Park, Calif., in July. Their mandate was clear: Right the wrongs of the generative A.I. team, which was no longer in charge of developing the company’s next batch of chatbots.
A month later, Meta restructured its A.I. division into four groups. One was focused on A.I. research, one on products and one on infrastructure such as data centers and other hardware, while TBD Lab was focused on building superintelligence. Collectively, the groups worked under an organization called Meta Superintelligence Labs. Mr. Wang would lead the entire effort.
Amid the restructuring, Meta lost dozens of senior A.I. researchers, many jumping to the very rivals that the company had just recruited from, like OpenAI and Google. Some executives also left, including Yann LeCun, Meta’s chief A.I. scientist; John Hegeman, the chief revenue officer; and Clara Shih, the head of A.I. business. In October, Meta laid off 600 employees in the broader A.I. division but left TBD Lab untouched.
To prepare for the arrival of superintelligence, Mr. Zuckerberg has pledged $600 billion to build data centers, which are the computing facilities that power A.I. and support nearly every aspect of Meta’s business.
This fall, some employees who work on Meta’s social media algorithms, which determine what content people see on Instagram and Facebook, argued that the computing power from those data centers should be used to improve their algorithms instead of training A.I. models, two people with knowledge of the matter said. They said the algorithms made money, while the A.I. models didn’t, the people said.
Meta, which gives away its A.I. models in what is known as “open sourcing,” plans to make money from the technology by adding A.I. to its popular products. That includes incorporating A.I. characters that people can talk to into Instagram or adding a personal A.I. assistant to the company’s smart glasses. The more time users spend on its apps, the thinking goes, the more money Meta can make from advertising.
That plan has led to philosophical differences. Mr. Cox, who has worked on products like the Facebook news feed, and Mr. Bosworth, who leads hardware development, have said they envision Meta’s superintelligence as useful for the social networking business.
In one recent meeting, Mr. Cox asked Mr. Wang if his A.I. could be trained on Instagram data similar to the way Google trains its A.I. models on YouTube data to improve its recommendations algorithm, two people said.
But Mr. Wang said complicating the training process for A.I. models with specific business tasks could slow progress toward superintelligence, they said. He later complained that Mr. Cox was more focused on improving his products than on developing a frontier A.I. model, they said. It’s unclear if Mr. Wang, Mr. Cox and Mr. Bosworth have resolved their debate.
On a recent call with investors, Susan Li, Meta’s chief financial officer, said a major focus next year would be using A.I. models to improve the company’s social media algorithm.
TBD Lab had a brief stream of departures over the summer, but things have appeared to stabilize. In a recent test of loyalty, researchers on the team reached their first “vesting” period on Nov. 15, when they got to own some of their company shares outright and could sell them, two people said. Only two researchers left the lab, which has around 100 people.
On Thursday, Mr. Wang plans to host his annual A.I. holiday party in San Francisco with Elad Gil, a start-up investor. (The invitation featured an A.I.-generated image of humanoid robots drinking cocktails next to a Christmas tree, according to a copy seen by The Times.) It’s unclear if any top Meta executives were invited.
Eli Tan covers the technology industry for The Times from San Francisco.
The post Meta’s New A.I. Superstars Are Chafing Against the Rest of the Company appeared first on New York Times.
