The man who helped assemble the most expensive research team in AI history wants you to know it's not about the paychecks.
The Summary
- Meta AI chief Alexandr Wang pushed back on claims his SuperIntelligence Lab researchers are mercenaries, despite reports of $100 million offers to poach talent from rivals.
- Wang called industry feuds between AI leaders "unfortunate" and hopes animosities will "subside over time", even as he reconciled with former colleague Yann LeCun who previously called him "inexperienced."
- The defense comes as Meta assembles what may be the most capital-intensive AI research operation in history, testing whether throwing money at the problem actually builds culture or just buys resumes.
The Signal
Meta reportedly made $100 million compensation packages to lure top AI researchers to its SuperIntelligence Lab. That's not a typo. Nine figures for individual contributors. When those numbers leaked, the narrative wrote itself: Zuckerberg buying a championship team with Facebook's money printer.
Wang isn't having it. "I think it's like an incorrect assumption to think that the researchers are just money motivated," he told the Core Memory podcast. "For most of them actually, the financial prospects of them staying wherever they were looked very, very, very strong."
"This is one of the larger narrative violations between external perception and what the day-to-day inside is like actually."
The timing matters. Meta is making this talent play while the rest of the AI industry is fragmenting into warring camps:
- Sam Altman versus Dario Amodei
- Altman versus Elon Musk (now in court)
- Wang versus his former Meta colleague Yann LeCun
LeCun called Wang "inexperienced" and said he didn't fully understand AI researchers. That's the kind of public swipe that usually calcifies into permanent enmity in this industry. But Wang says he ran into LeCun in India weeks later, and LeCun congratulated the team on launching the Muse Spark AI model.
Wang called the industry tensions "unfortunate" and expressed hope that "all these animosities subside over time" so people can focus on building "this incredibly important technology." It's a remarkably diplomatic stance for someone who just got publicly dressed down by one of the field's founding figures.
Here's what Wang is really saying: Meta didn't just buy talent, they bought focus. The researchers who took those packages could have stayed at OpenAI, Anthropic, Google, wherever. They had equity, they had options. What they didn't have was Mark Zuckerberg's willingness to pour Meta's entire infrastructure budget into compute and tell them to build without worrying about quarterly earnings calls.
The Implication
Watch whether Meta's culture bet pays off. $100 million packages buy you 12-18 months of someone's attention. Building a research culture that produces breakthroughs instead of papers requires something money can't directly purchase: shared conviction about what you're building and why.
If Wang is right that his team came for the mission, not just the money, we'll see it in the product velocity and research output. If he's wrong, we'll see it in the attrition rate when those packages vest and competitive offers arrive. The AI talent war isn't over. It's just entering a new phase where the teams that built cohesion under pressure separate from the ones that were just highly compensated free agents all along.