Lead
Thinking Machines Lab leaders raised concerns last summer about an alleged workplace relationship involving cofounder and former CTO Barret Zoph, sources tell WIRED. That internal confrontation preceded a breakdown in the cofounders’ working relationship and Zoph’s departure, after which he entered talks with competitors and was ultimately hired by OpenAI. The employee involved, who held a leadership role in a different department, is no longer at the lab. Thinking Machines Lab and OpenAI declined on-the-record comment for this report.
Key Takeaways
- Leaders at Thinking Machines confronted Barret Zoph last summer over an alleged relationship with another staffer; the individual worked in a different department and has since left the lab.
- Multiple sources say the cofounders’ working relationship deteriorated in the months after that conversation, and Zoph began discussing roles at competitors.
- Zoph spoke with Meta Superintelligence Labs leaders before ultimately being hired by OpenAI; OpenAI’s Fidji Simo said the hire had “been in the works for weeks.”
- At least four researchers, including cofounder Luke Metz and three others, moved from Thinking Machines to OpenAI this week; cofounder Andrew Tulloch left for Meta in October.
- Thinking Machines reportedly explored raising capital at a $50 billion valuation in November, up from a reported current valuation of $12 billion.
- Thinking Machines Lab declined to comment to reporters; OpenAI and Zoph also did not provide on-the-record statements for this story.
Background
Thinking Machines Lab launched with high-profile leadership and rapid hiring in a competitive market for top AI talent. Co‑founders including Mira Murati and Barret Zoph attracted attention for the lab’s ambitions and their respective roles in building advanced models. The startup has faced internal debate over product direction and governance, a dynamic reported previously and cited by current and former staff.
Industry competition—particularly between OpenAI and Meta—has intensified headhunting and lateral moves among researchers over the past year. Startups in this space also face scrutiny over workplace conduct, board oversight and how personal relationships intersect with reporting lines. Those pressures come alongside investor focus on lofty valuation targets: a November report said Thinking Machines was seeking funding at a $50 billion valuation versus a reported $12 billion base valuation earlier.
Main Event
According to multiple people familiar with internal discussions, Murati and other leaders raised concerns with Zoph about a relationship involving another employee last summer. The employee worked outside Zoph’s department and occupied a leadership role; to protect privacy, sources declined to identify that person. Following the meeting, sources say friction between cofounders increased and collaboration grew strained.
In the months that followed the confrontation, Zoph engaged in conversations with outside labs. One source familiar with the matter said he spoke with leaders at Meta Superintelligence Labs; those talks did not immediately result in a hire. Separately, OpenAI pursued and eventually hired Zoph, with OpenAI’s applications chief, Fidji Simo, later saying the recruitment had been underway for several weeks.
Thinking Machines has experienced broader departures beyond Zoph. This week, cofounder Luke Metz and at least three other researchers moved to OpenAI, and earlier reporting shows Andrew Tulloch left for Meta in October. Company spokespeople did not respond to requests for clarification about whether the relationship or related concerns directly triggered Zoph’s exit.
Analysis & Implications
The episode highlights how interpersonal issues at senior levels can escalate into talent shifts that reshape early-stage AI ventures. For a startup raising at triple its prior valuation on paper, high-profile departures magnify investor and market scrutiny; liquidity and hiring assumptions underpinning a $50 billion target may be harder to justify if leadership cohesion weakens.
For acquiring firms like OpenAI and Meta, rapid hiring of elite talent offers capability gains but also raises governance questions: how do buyers vet interpersonal and ethical risks when recruiting away rivals? OpenAI’s reported decision to hire despite Thinking Machines’ internal concerns illustrates a tension between competitive priorities and due diligence norms in a fast-moving labor market.
Regulatory and reputational exposure is another factor. If a leadership-level relationship intersects with reporting lines or power imbalances, companies can face legal and cultural risks. Startups building governance frameworks will need clearer policies and disclosure practices to manage those risks while sustaining recruitment in a highly mobile talent pool.
Comparison & Data
| Metric | Reported Figure |
|---|---|
| Thinking Machines reported valuation (current) | $12 billion |
| Thinking Machines target valuation (reported, Nov) | $50 billion |
| Recent researcher departures to OpenAI | At least 4 (including Luke Metz) |
| Cofounder move to Meta | Andrew Tulloch (left in October) |
Those figures frame the corporate stakes: a reported target of $50 billion represents more than a fourfold increase over a $12 billion baseline. Multiple cofounder departures and researcher flights to larger labs can materially affect a startup’s development roadmap and investor confidence. The near-term operational impact depends on which teams and projects the departing individuals led.
Reactions & Quotes
“The hiring had been in the works for weeks,”
Fidji Simo, OpenAI (as reported)
OpenAI’s applications chief characterized the recruitment as a multweek process and told reporters she did not share the concerns that Thinking Machines leaders raised internally. OpenAI did not provide further comment to clarify vetting steps or whether those internal concerns were disclosed during recruitment.
“The cofounders’ working relationship broke down in the months after the conversation,”
Person familiar with Thinking Machines internal discussions (anonymous)
Sources who spoke with reporters described increasing friction among the founding team after leaders addressed the alleged relationship. Those tensions, sources say, coincided with a window when several researchers began exploring external opportunities.
Unconfirmed
- The precise nature and details of the alleged misconduct have not been publicly disclosed and remain unverified.
- Whether the relationship was the formal or sole reason for Zoph’s termination or departure is not definitively confirmed.
- The full roster and motives of all employees who left Thinking Machines for OpenAI have not been independently verified beyond multiple named departures.
- Reports of conversations with Meta Superintelligence Labs are based on a source familiar with the matter and have not been independently corroborated by Meta.
Bottom Line
The situation at Thinking Machines illustrates how senior-level interpersonal issues can cascade into leadership splits, hiring moves and investor uncertainty—especially when a startup is operating under aggressive valuation and hiring assumptions. Talent competition among AI labs means departures can accelerate and have outsized impact on small, research-focused startups.
Observers should watch whether Thinking Machines stabilizes its leadership team, how investors respond to departure-driven risks, and whether acquiring labs disclose more about vetting practices when hiring senior hires from rivals. These developments will influence governance norms across the industry and how startups manage both internal conduct and external talent flows.