Meta had already explored acquiring Thinking Machines Lab before those talks stopped short of a formal offer, while an August report identified Tulloch as the researcher who had turned down Zuckerberg's reported $1.5B-plus package. His eventual move gives that recruitment effort a different outcome without an acquisition.
The departure matters because it shifts a founding researcher from Mira Murati's independent lab to a large platform company. Later coverage of two more founding-team members working at Meta makes this an important early marker in the two organizations' overlapping talent story.
First-order effects
Meta gains Andrew Tulloch's research and company-building experience; Thinking Machines Lab loses a co-founder from its founding team.
The move resolves the immediate recruitment question raised by the reported August compensation approach, even though he had declined that earlier package.
Second-order effects
Thinking Machines Lab faces a harder retention and execution challenge as Meta becomes a destination for members of its founding circle, rather than only a prospective acquirer.
Meta's ability to recruit from a lab it had previously considered buying gives it another route to build AI capability without completing a corporate transaction.
Third-order effects
If leading researchers increasingly move individually from new frontier labs to incumbent platforms, talent acquisition can become an alternative to lab acquisitions, concentrating scarce leadership inside cash-rich companies.
The pattern could make independent labs more dependent on retaining cohesive founding teams; whether it persists will depend on their ability to offer researchers autonomy and durable institutional backing.
The trend: This is one data point in the concentration of frontier-AI talent as platform companies use recruitment alongside investment and acquisition discussions to assemble research capacity.
Few people realize that once you reach a certain level of wealth, money no longer drives your decisions. I really like Thinking Machines' openness about their research and hope they build cool products. Hopefully Andrew's departure won't slow them down.
Rumor has it Andrew Tulloch returned to Meta because Zuck offered him $3.5B. I don't buy the “everyone has a price” argument. For him, $3.5B vs. $1.5B barely matters. Building something from scratch as a cofounder of Thinky sounds way more fun. Anyway, Llama 5's gonna be good.
leaving as co-founder is kind of degenerate, after you've convinced tons of people to believe in you and invest their time to work with you and join you.
This is bearish all around... Andrew makes himself a valueless mercenary Thinking Machines clearly isn't high optionality enough for him to stay Meta keeps plowing billions on individuals that could not give two shits about metas success Once upon a time they bought IG for a
Andrew has a long history with Meta. He had already turned down a $1B offer from Meta earlier this year. But previous to cofounding Thinking Machines in January — and his work at OpenAI before that — he had served an 11-year stint at Meta as a distinguished engineer.
So either: Zuck said ok $1B wasn't enough, how does $5B sound? Or Tulloch read all the AI bubble tweets and thought hmmm actually fully liquid stock I can dump all of ASAP sounds pretty good after all Or He decided he just really wanted to work with Alexandr
Imagine being this guy, able to reject a $1B offer from Meta. You can literally walk around virtually anywhere and know that you are with no debate, better than every single person in a 10-mile radius of you. What could they have possibly offered him this time better than $1B?
mom how did we get so rich your daddy worked at meta for 11 years and then he fucked around and found out that he could boomerang after shopping at oai and thinky
Andrew Tulloch, the star AI researcher and Thinking Machines Lab co-founder who turned down a $1bn offer from Meta earlier this year, has decided to join Meta after all. Scoop w/ @MeghanBobrowsky https://www.wsj.com/... via @WSJ
Let's say the initial package was $1.5B over the equivalent of 3ys and then Mark had to double it to pry Andrew from #Thinking machines. [embedded post]