Andrew Tulloch, who reportedly turned down Meta's $1.5 billion offer before joining, is now leaving for AnthropicAndrew Tulloch, the AI researcher Meta once chased with a pay package reportedly worth up to $1.5 billion, is leaving the company after less than a year. He is joining Anthropic, the maker of Claude. The Wall Street Journal reported the move, and an Anthropic spokesperson later confirmed to Business Insider that Tulloch starts next week on its inference and performance team. He will work on how the company's AI models are trained and how they respond to users.Meta loses one of its highest-profile hires. Tulloch rejoined the company in October 2025 after leaving Thinking Machines Lab, the startup he co-founded with former OpenAI CTO Mira Murati. He worked in TBD Lab, the small team central to Mark Zuckerberg's AI strategy, within Meta Superintelligence Labs led by Alexandr Wang. His personal website described his role as "working on superintelligence."Andrew Tulloch waited for Meta's Muse AI agent to launch before quittingTulloch told colleagues he was leaving on Wednesday, a day after Meta launched Muse. He had reportedly held off on quitting until the product shipped. Muse is a personal AI agent built for mainstream users rather than developers. If given access, it can send emails, shop online and complete other tasks on a user's behalf.Why he is leaving remains unclear. Meta declined to comment on his exit, according to The Wall Street Journal, and Tulloch did not respond to requests for comment.Meta's reported $1.5 billion offer made Tulloch the face of the AI talent warTulloch's name became tied to the AI hiring frenzy last year. According to The Wall Street Journal, Zuckerberg first tried to buy Thinking Machines. When Murati refused, he approached more than a dozen of her employees, with Tulloch as his main target. The package could have been worth up to $1.5 billion over at least six years, assuming top bonuses and exceptional stock performance.Tulloch said no. Meta spokesperson Andy Stone called the reported figure "inaccurate and ridiculous." Months later, Tulloch joined Meta anyway. The terms he finally accepted were never disclosed.Tulloch helped train OpenAI's GPT-4o, GPT-4.5 and o3 modelsTulloch spent 11 years at Facebook and rose to distinguished engineer, one of its highest technical ranks. In 2023, he moved to OpenAI, where he helped train GPT-4o, GPT-4.5 and o3, before co-founding Thinking Machines in early 2025.OpenAI had tried to hire him years earlier. In 2016, OpenAI president Greg Brockman wrote to Elon Musk that Tulloch earned $800,000 a year at Facebook, while OpenAI offered recruits a $175,000 salary plus a $125,000 bonus, per the Journal. Tulloch balked at the pay cut.Anthropic lands Tulloch as AI labs keep trading top researchersTulloch is the third Thinking Machines co-founder to change jobs in recent weeks. Barret Zoph and Luke Metz, who had moved from the startup to OpenAI, left last month for Google and Meta respectively. Google's chief scientist Jeff Dean also quit last month with three senior researchers to launch an AI startup.For Anthropic, the hire comes at a busy time. Business Insider reported that the company is preparing for its initial public offering. This week, it also lost researcher Jacob Coxon, who resigned with a public warning that people building AI believe it could wipe out humanity by the end of the decade.Meta spent last year showing it would pay extraordinary sums for AI talent. Eleven months after Tulloch finally said yes, its most famous recruit is heading to a rival.