Meta plans to cut approximately 600 jobs in its artificial intelligence division as chief executive Mark Zuckerberg reshapes the company around a vision of “personal superintelligence” that empowers individuals rather than automates work, reports the New York Times.
The layoffs will hit Meta’s Superintelligence Labs, which has several thousand employees across AI research, products and infrastructure. The cuts aim to eliminate organisational bloat from three years of rapid hiring and help the company develop AI products more quickly. Alexandr Wang, Meta’s chief AI officer, says reducing team size means fewer conversations will be required to make decisions, with each person taking on more responsibility and having a greater scope and impact.
The cuts do not affect Meta’s newest AI hires, some of whom receive pay packages numbering well into the hundreds of millions of dollars. Zuckerberg has spent billions recruiting top researchers from OpenAI, Google and Microsoft since restarting AI efforts in June with a $14.3 billion investment in ScaleAI, the start-up co-founded by Wang.
“Superintelligence is now in sight”
Zuckerberg articulated his vision for AI in a July letter declaring that “developing superintelligence is now in sight” as Meta’s systems began showing signs of self-improvement. He positioned personal superintelligence as distinct from competitors’ approaches, rejecting the view that AI should “be directed centrally towards automating all valuable work, and then humanity will live on a dole of its output”.
“Meta’s vision is to bring personal superintelligence to everyone,” Zuckerberg wrote. “We believe in putting this power in people’s hands to direct it towards what they value in their own lives.” He described personal superintelligence as systems that know users deeply, understand their goals and help achieve them, integrated into devices like glasses that see and hear users’ context throughout the day.
The philosophical shift comes as Meta struggled to keep pace with OpenAI, Google and Microsoft after ChatGPT’s 2022 launch. After early success with its open-source Llama model, progress stagnated. Strategic errors led to product development issues over the past 18 months before Zuckerberg’s June reset.
In August, Zuckerberg split Meta Superintelligence into four groups: FAIR for AI research, a team focused on superintelligence called TBD, a products division and an infrastructure group. The planned cuts affect employees at FAIR, products and infrastructure but spare TBD, which manages Meta’s large language models and continues hiring researchers under Wang’s management.
Zuckerberg framed the superintelligence race as defining the decade ahead. “The rest of this decade seems likely to be the decisive period for determining the path this technology will take, and whether superintelligence will be a tool for personal empowerment or a force focused on replacing large swaths of society,” he wrote. Meta executives emphasised the cuts do not signal retrenchment on AI efforts, with superintelligence remaining among Zuckerberg’s top priorities.
