Meta CEOMark Zuckerberg’s aggressiveAI talent poachinghas been the talk of the tech circles lately. Now it appears that the Meta CEO is embracing a new and radical strategy by placing his biggest AI bet on the smallest teams at his company. As reported by Business Insider, during the recent earnings call Zuckerberg described Meta’s newSuperintelligence Labsas “a bit of a different setup,” which contrasts it with the company’s workforce which has more than 70,000 employees. At the heart of this new transformation is the TBD Lab. The lab is a secretive group which consists of elite AI researchers led by Alexandr Wang. This new lab has task of creating Meta’s most advanced AI models — what Zuckerberg calls “personal superintelligence.”

Small team with big goals

“I’ve just gotten a little bit more convinced around the ability for small, talent-dense teams to be the optimal configuration for driving frontier research,” Zuckerberg said. Meta CEO feels that breakthroughs in the filed of AI are best achieved by small and compact groups which can hold the complete problem space ‘in their head’ rather than depending on the massive engineering teams like the ones that manage Facebook’s newsfeed.This philosophy mirrors a growing trend across Silicon Valley, where lean teams are seen as faster, more agile, and more innovative. Startups like Hightouch, which has raised over $132 million with just 55 engineers, exemplify the model. On the other hand, Nat Friedman, former GitHub CEO and now part of Meta’s AI product integration team, has championed the idea that most tech companies are “two to ten times overstaffed.”

Internal tensions and structural challenges

Meta’s pivot towards the ‘startup mode’ has created a bit of a friction within the company. As reported by Business Insider, the creation of Superintelligence Labs has led to resentment and resignation threats among legacy researchers, who have been sidelined after coming of the new hires.On the other hand, experts also suggest that small teams inside large companies often struggle to deliver transformational change. “They can produce useful products and efficiency gains,” said Elliott Parker, CEO of Alloy Partners, “but rarely the kind of fundamental shift that reshapes the parent company.”Lately, Meta has reorganised its AI division. The company has dissolved its two major AI units and is now dealing with the risk of overlap between micro teams. The Superintelligence LabsDespite the risks, Zuckerberg remains bullish. He sees small, elite teams as the key to staying competitive in the AI race, especially as breakthroughs like the 2017 “Attention Is All You Need” paper—authored by just eight researchers—continue to shape the field.“For the leading research on superintelligence,” Zuckerberg said, “you really want the smallest group that can hold the whole thing in their head.”