Members of Meta’s flashy new team of AI experts have reportedly been clashing with some of CEO Mark Zuckerberg’s loyal longtime executives over what the company should prioritize.
An us-versus-them mentality has riven the tech giant, with AI brainiacs disdaining Musk’s inner circle as being interested just in improving the social media and ad businesses — while they want to create a “godlike AI superintelligence,” the New York Times reported Wednesday.
A key player in the alleged conflict is Alexandr Wang, the 28-year-old founder of Scale AI who joined Meta after it paid $15 billion for his startup. He now leads the company’s TBD Lab of researchers. (His first name really lacks the usual second E and the lab’s initials indeed denote “too be determined.”)
During meetings this fall, Wang privately told people he disagreed with longtime Meta execs Chris Cox, the company’s chief product officer, and Andrew Bosworth, chief technology officer, according to the Times.
Cox and Bosworth wanted Wang’s team to use Instagram and Facebook data to train Meta’s new model, similar to the way Google uses YouTube data, in an effort to fine-tune Meta’s social media algorithm, sources told the Times.
But Wang argued TBD Lab should solely focus on catching up to AI rivals like OpenAI and Google, according to sources.
He later complained that Cox was more focused on products than developing a “frontier” AI model, and that the execs’ requests could slow the AI team’s progress, people told the Times.
Bosworth, meanwhile, was recently asked to slash $2 billion off his proposed budget for Meta’s virtual reality and augmented reality division next year, which he leads – so the money could go toward Wang’s team, sources were quoted as saying.
A Meta spokesman disputed the account of strife and said next year’s budget has not been finalized.
“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 AI investment is improving the ads and recommendations systems at the heart of our business,” the spokesman told The Post in a statement.
A Meta veep also slammed the Times report.
“The conversation these anonymous sources claim happened was met with surprised-disbelief by the people who supposedly had it,” Andy Stone, the company’s vice president of communications, tweeted in response to the Times report.
It’s just the latest alleged clash involving Meta’s newest AI researchers – many of whom were lured from rival companies with pay packages worth hundreds of millions of dollars.
It was reported earlier this year that Wang butted heads with Meta’s engineers as he pushed to make the company’s next AI model “closed,” keeping its technology secret.
Meta Chief Financial Officer Susan Li said during a recent call with investors that the company would focus on using AI models to improve the social media algorithm next year.
Wang, meanwhile, is planning to host his annual AI holiday party in San Francisco on Thursday – and it’s unclear whether any Meta execs made the guest list, according to the Times.
Zuckerberg has been willing to shell out big bucks in the effort to catch up to rivals like ChatGPT-maker OpenAI in the superintelligence race.
After buying Wang’s AI startup for a whopping $14.3 billion, Zuckerberg went on a hiring blitz, offering eye-popping pay packages to top researchers at rival companies.
He went so far as to have homemade soup delivered to the doorsteps of some OpenAI employees, according to Mark Chen, OpenAI’s chief research officer.
Zuckerberg has also announced $600 billion in planned spending on AI data centers.
When the new team arrived at Meta’s Silicon Valley campus in July, he placed them next to his office at the center of headquarters, according to the Times.
Zuck split the AI division into four teams – research, products, infrastructure and superintelligence – and appointed Wang to oversee the entire unit.
Researchers on the new AI team hit their first “vesting” period last month, so they can own and sell some of their company shares, sources told the Gray Lady.
Meta has said it plans to make money from the investments by integrating AI into products, like adding a personal AI assistant to its smart glasses.
But other Meta employees have argued that most of the company’s computing power should go toward improving the algorithm, not funding the superintelligence lab, since Meta’s social media business is the real moneymaker, sources told the Times.
Meta has suffered an exodus of some top AI leaders including Yann LeCun, chief AI scientist, and Clara Shih, head of the AI business.
The company also laid off 600 staffers across its AI division in October. The layoffs did not impact TBD Lab.
