There seems to be some friction within Meta’s newly formed AI team, particularly between these tech-savvy newcomers and some of CEO Mark Zuckerberg’s long-standing executives. It appears the frustration is growing over differing priorities within the company.
On one side, the AI specialists reportedly view the existing leadership, particularly in relation to Musk’s circle, as too focused on enhancing social media platforms and advertising. They’re aiming for something more ambitious: the development of a “godlike AI superintelligence,” according to a report from a major news outlet.
At the heart of the disagreement is Alexander Wang, the 28-year-old founder of Scale AI, who joined Meta following a hefty acquisition of his startup for about $15 billion. He’s currently at the helm of a research lab with a somewhat ambiguous scope. Interestingly, the lab’s name—TBD Lab—plays into this lack of definition.
During a meeting this fall, Wang apparently expressed his dissent toward senior Meta figures like Chris Cox and Andrew Bosworth. They were advocating for the use of Instagram and Facebook data to enhance new AI models, which is a strategy reminiscent of how Google leverages YouTube data.
Wang, however, maintained that the focus should solely be on catching up to competitors like OpenAI and Google. He voiced concerns that Cox’s focus on products was detracting from the essential work of pushing forward with advanced AI models, fearing that such demands might delay progress for his team.
In the midst of this, Bosworth found himself tasked with slashing $2 billion from the 2024 budget for Meta’s virtual and augmented reality projects, aiming to reallocate those funds to Wang’s team. A Meta spokesperson, though, dismissed reports of any conflict, stating that next year’s budget was still under discussion. They emphasized a unified leadership approach toward developing superintelligence while also bolstering their core social media business.
The communications vice president for Meta even challenged the narrative from the news report, suggesting that the alleged private conversations were surprising to those supposedly involved.
This isn’t the first instance of tension related to Meta’s AI initiatives. Earlier in the year, Wang reportedly clashed with engineers over a push to keep the next AI model private.
Moreover, during a recent call with investors, CFO Susan Lee reiterated that the focus would be on using AI to enhance social media algorithms moving forward. Meanwhile, Wang is scheduled to host an AI Holiday Party in San Francisco soon, though it’s uncertain if Meta’s executives will be attending.
Zuckerberg has shown a willingness to invest heavily in catching up with competitors like OpenAI in the race for superintelligence, recently acquiring Wang’s company and aggressively recruiting top talent from rival firms, even offering extravagant perks like delivering homemade soup to certain OpenAI staff members.
In his bid to get ahead, Zuckerberg has also announced plans for a significant investment in AI infrastructure, amounting to $600 billion. When the new AI team was established at Meta’s Silicon Valley office, it was strategically placed near Zuckerberg’s office, underlining the importance of this initiative.
The AI division has been organized into four distinct teams, with Wang overseeing them all. According to a report, these researchers recently reached their first “vesting” milestone, allowing them to claim some shares.
Meta intends to leverage its AI investments within its product lineup, including features like a personal AI assistant for smart glasses. Still, there are voices within the company arguing that their main revenue stream lies in social media, suggesting that greater resources should be directed towards enhancing existing algorithms instead of backing ambitious superintelligence projects.
The company has faced losses of several high-profile AI leaders recently and also laid off around 600 employees within its AI division back in October, although TBD Lab was not impacted by these layoffs.
