When Meta CEO Mark Zuckerberg reached out to Mira Murati, the former CTO of OpenAI, earlier this year with an offer to buy her AI startup, Thinking Machines Lab, she declined. So he tried a more aggressive move—attempting to recruit her team. His top target was Andrew Tulloch, a former Meta engineer and now a key figure at Thinking Machines. To lure him back, Zuckerberg reportedly offered a package potentially worth $1.5 billion over six years.
Tulloch said no. And so did every one of Murati’s roughly 50 employees.
Meta’s deep pockets can’t always buy loyalty
Even in Silicon Valley, where massive compensation packages are common, walking away from a 10-figure offer is unusual. But this episode highlights a shift: in the world of generative AI, money isn’t the only magnet. Companies like OpenAI, Anthropic, and Murati’s Thinking Machines have built deep internal cultures and quasi-missionary zeal around their visions of artificial general intelligence (AGI). That ethos—combined with tight-knit teams and flat hierarchies—has created a kind of fortress against poaching, the Wall Street Journal reported.
OpenAI, Anthropic, and the culture of commitment
Meta has tried to recruit over 100 OpenAI employees and has landed only a handful—just 10 so far. Many researchers are staying put, believing OpenAI is closest to achieving AGI and wary of seeing their work used to power advertising platforms. Anthropic, led by Dario Amodei, has seen even fewer defections. All seven of its co-founders remain, bound by long-standing friendships and a shared belief in the principles of effective altruism, a movement that once united many in concern over AI’s existential risks.
Sutskever’s stealth approach at Safe Superintelligence
Meanwhile, Ilya Sutskever—OpenAI’s former chief scientist and co-founder—has built his new company, Safe Superintelligence (SSI), to be nearly unpoachable. The company hires largely unknown but promising researchers, discourages staff from listing SSI on LinkedIn, and keeps a low public profile. Zuckerberg reportedly tried to buy SSI outright. Sutskever refused.
Mira Murati’s quiet revolution
Murati, who helped lead OpenAI from a scrappy lab to the centre of the AI universe, started Thinking Machines in February and brought more than 20 OpenAI alumni with her—many from the team that built ChatGPT. The startup has attracted $2 billion in funding and is reportedly working on multimodal AI systems designed to interact with the world more naturally. Despite its low profile and modest office in San Francisco, Thinking Machines has become a prime target for poaching—and so far, no one has left.
Zuckerberg’s recruitment blitz
Zuckerberg’s push has extended beyond Thinking Machines. He recently appointed Shengjia Zhao, formerly of OpenAI, to lead Meta’s new superintelligence lab, and enlisted Alexandr Wang to help build out the team. But even with lavish offers, defections from startups like Anthropic, OpenAI, and Thinking Machines remain rare. Tulloch, for instance—long regarded as a machine learning prodigy—was once courted by OpenAI back in 2016 but only joined seven years later, after ChatGPT’s rise.
The value of mission over money
For Zuckerberg, the talent war is a reminder that even Silicon Valley’s biggest wallets have limits. In the AI race, trust, culture, and belief in the mission are proving more powerful than billion-dollar offers. While Meta may lead in capital and infrastructure, the heart of the AI movement still beats in the labs where researchers feel they’re shaping something bigger than the next app—something closer to artificial general intelligence itself.
Discover the latest Business News, Sensex, and Nifty updates. Obtain Personal Finance insights, tax queries, and expert opinions on Moneycontrol or download the Moneycontrol App to stay updated!
