Andrew Tulloch, co-founder of Mira Murati’s venture Thinking Machines Lab, has confirmed his departure to rejoin Meta.
Tulloch helped construct the startup’s AI fine-tuning technologies and key infrastructure. His departure comes as Meta is aggressively recruiting AI experts and restructuring under a new subsidiary named Superintelligence Labs.
He reportedly turned down a $1.5 billion offer previously.
What this means for creators
As you decide where to plug in your work (platforms, APIs, frameworks), talent flows like this reveal which companies are doubling in on infrastructure – these could be good locations to align with or target integrations.
A high-stakes exit indicates that core tooling and agent building are still a high-leverage frontier. If you can create a solid foundation in model orchestration, tuning, or agent deployment, you could carve out a successful niche.
It’s a reminder that the AI space is still volatile. Even startups with serious ambitions lose important employees. Have contingency plans and adaptable structures.
What this means for entrepreneurs
Meta’s ability to entice elite personnel demonstrates its long-term strategy for controlling the AI stack. When pitching a business in tooling or infrastructure, it’s likely that you’ll be compared to or acquired by larger companies.
Founders should take precautions against key person risk. Ensure that knowledge is spread. If someone leaves, your roadmap should not come to a halt.
Talent is the battleground. Recruiting, retention, and contracts are as important as your product in AI.
