Meta reportedly sat down with Thinking Machines Lab about an acquisition roughly a year ago. That deal never happened. What happened instead is more interesting — Meta started picking off TML’s founders one by one, and TML responded by going straight to Meta’s own talent pool to replace them. That’s not a failed acquisition story. That’s a rivalry.
I cover AI toolkits for a living, which means I spend a lot of time watching which organizations are actually building things versus which ones are positioning. The talent movement between Meta and Thinking Machines Lab tells me something useful about both.
Seven Founders and Counting
Meta has poached seven of TML’s founding members. Seven. That’s not a coincidence or a couple of opportunistic hires — that’s a deliberate strategy. Meta clearly sees something worth dismantling in what TML was building, and the most direct way to slow a competitor is to absorb the people who know the product best.
But here’s what makes this story worth paying attention to from a toolkit reviewer’s perspective: TML didn’t fold. They went recruiting. And they went recruiting at Meta.
That kind of counter-move takes confidence. You don’t go poaching from the company that just gutted your founding team unless you believe your roadmap is still worth executing. TML is signaling that the project continues, and that they can attract serious people to it even after taking significant hits.
What the Chintala Appointment Actually Means
In early 2026, Soumith Chintala was appointed CTO of TML. If you’ve spent any time in the ML space, that name means something. Chintala co-created PyTorch, which became one of the most widely used frameworks in the field. His move to TML isn’t just a headline — it’s a signal about where he thinks the interesting work is happening.
When someone with that kind of track record makes a move, it shifts how other researchers think about their own options. Talent follows talent. TML landing Chintala as CTO, especially in the middle of an active talent war with Meta, changes the calculus for anyone sitting on the fence about where to go next.
What This Means for the Tools Coming Out of Both Camps
From where I sit, reviewing what actually works in the AI toolkit space, organizational stability matters more than most people admit. The tools that hold up over time tend to come from teams with continuity — people who were there when the architecture decisions were made, who understand why certain tradeoffs exist.
When you lose seven founders, you lose institutional knowledge. Meta is a large enough organization that it can absorb those people and extract value from them, but TML’s original product vision takes a hit regardless. The question is whether the new team TML is assembling can pick up that thread or ends up building something adjacent to it.
That’s not a knock on TML — it’s just the honest reality of what talent churn does to a product. I’ve seen it happen with smaller toolkit companies where one or two key departures changed the entire direction of a product. Seven is a lot to absorb.
The Acquisition That Wasn’t
The failed acquisition attempt is the part of this story I keep coming back to. Meta wanted TML badly enough to sit down and talk about buying it. When that didn’t work out, they apparently decided the next best option was to hire away the people who built it.
That tells you Meta views TML as a genuine threat, not a curiosity. You don’t spend that kind of political and financial capital on something you’re not worried about.
For anyone evaluating AI tools right now, that context matters. TML is building something that one of the largest AI organizations in the world tried to buy and then tried to weaken. That’s a different kind of validation than a funding round or a benchmark score.
Where I’m Watching Next
- What Chintala’s technical priorities look like in practice once TML starts shipping under his direction
- Whether the researchers TML is pulling from Meta bring product sensibility or pure research focus
- How Meta uses the seven founders it acquired — are they building something new or reinforcing existing infrastructure
The talent war between these two organizations is still active. Neither side has made a decisive move yet. But the fact that TML is still standing, still hiring, and now has Chintala at the helm suggests this isn’t a story about a startup getting slowly absorbed. It looks more like a fight that’s just getting started.
🕒 Published: