Andrew Tulloch has once again reminded Silicon Valley that purpose can trump pay. According to the Wall Street Journal, Meta dangled a package worth up to $1.5 billion over six years. Tulloch, a former Facebook engineer who now co-leads Mira Murati’s startup Thinking Machines Lab, politely declined.
Rival Labs Push Back
Tulloch’s rebuff comes amid an aggressive, and only partially successful, hiring blitz. Since early June, Meta has messaged well over a hundred OpenAI employees, luring just ten. The boldest catch so far is Shengjia Zhao, a Chinese researcher who spent three years at OpenAI and was appointed to run Meta’s Superintelligence Team.
Yet most of the people Zuckerberg has courted remain unconvinced. They see OpenAI as closest to cracking artificial general intelligence and prefer the nimbleness of smaller labs. They balk at the idea of funneling their breakthroughs into another ad-funded product suite.
Rival Labs Hold Firm
Anthropic has proved even harder to raid. Despite Meta’s sky-high offers, only two former colleagues – Joel Pobar and Anton Bakhtin have crossed over. The rest of Anthropic’s staff, including all seven founders, remain loyal to CEO Dario Amodei.
Ilya Sutskever, OpenAI’s co-founder turned CEO of Safe Superintelligence, has gone on the defensive as well. SSI operates in near-silence: staff avoid listing the company on LinkedIn, and Sutskever hires lesser-known prodigies he can mentor. Earlier this year he reportedly rejected an acquisition offer from Meta, effectively shutting a backdoor to his talent pool before it opened.
Thinking Machines’ Magnetic Pull
Then there is Murati whose emotional IQ and flat-hierarchy ethos kept engineers fiercely loyal. When she launched Thinking Machines in February, more than twenty OpenAI colleagues immediately followed – including co-founder John Schulman, one of the original architects of ChatGPT.
Backed by $2 billion in fresh capital, the company has leased a discreet Mission District office a stone’s throw from OpenAI headquarters. Its mission – building multimodal systems that feel as natural as human conversation remains largely under wraps, even to early investors.
Tulloch’s Decision and Meta’s Dilemma
Tulloch now finds himself at the center of the most lucrative courtship to date. The Australian mathematician first became a Meta wunderkind in the early 2010s, left to pursue machine-learning research at Cambridge, and re-emerged at Facebook’s AI Research Group in California.
His reputation for solving impossible optimization problems earned him a spot among the social network’s elite engineers. In 2016 Greg Brockman tried to recruit him into OpenAI’s founding team, noting in an email to Elon Musk that Tulloch was earning $800,000 a year and would likely push for more. Tulloch hesitated over the pay cut and stayed put. He finally joined OpenAI seven years later, just after ChatGPT’s viral breakout.
Australian man Andrew Tulloch turned down a $1.5 billion USD offer from Mark Zuckerberg to work at Meta on AI
He graduated from the University of Sydney as the top honours student in the Faculty of Science
AUSTRALIAN EXCELLENCE 🇦🇺🇦🇺🇦🇺🇦🇺 pic.twitter.com/Wzvgikbr4N
— Drew Pavlou 🇦🇺🇺🇸🇺🇦🇹🇼 (@DrewPavlou) August 2, 2025
Zuckerberg and Alexandr Wang, the newly appointed head of Meta’s superintelligence lab, bombarded Tulloch with messages this spring. Their pitch: come home, pocket a life-changing fortune, and build the future of AI with virtually unlimited resources.
Tulloch considered the offer, then declined. People close to him cite three reasons: loyalty to Murati’s vision, discomfort with Meta’s ad-driven roadmap, and doubts that any amount of funding can buy the creative freedom he currently enjoys.
Meta disputes the premise that it is striking out. Spokesperson Andy Stone maintains that only a “handful” of Thinking Machines employees received offers, adding that “the numbers floating around are wildly exaggerated.”
For Zuckerberg, the challenge is no longer writing a bigger cheque; it is convincing the Andrew Tullochs of the world that Meta’s future is the one worth building.
