At a Glance
- Thinking Machines Lab co-founders Barret Zoph and Luke Metz are returning to OpenAI
- Zoph was dismissed after allegedly sharing confidential data with rivals
- The move comes days after Zoph told CEO Mira Murati he might leave
- Why it matters: OpenAI regains two top researchers while rival startup loses key talent
OpenAI has lured back two high-profile researchers less than two months after they left to launch a competing AI company, according to an internal memo obtained by News Of Fort Worth.
The announcement caps a dramatic 48-hour sequence that began when Barret Zoph told Thinking Machines CEO Mira Murati he was considering resignation and ended with his termination for what sources describe as unethical conduct.
Rapid Exit and Return
Zoph and Metz departed OpenAI in late 2024 alongside Murati to build Thinking Machines Lab, a startup pitched as the next challenger in the race for artificial general intelligence. Their return marks one of the fastest executive round-trips in Silicon Valley’s recent history.
Fidji Simo, OpenAI’s CEO of applications, broke the news to staff Thursday afternoon, writing that the company has “full confidence” in both researchers. The memo pointedly added that OpenAI does “not share the same concerns” about Zoph that Murati expressed.
Allegations of Leaked Secrets
A source close to Thinking Machines told News Of Fort Worth that Zoph shared proprietary information with outside competitors. The startup moved swiftly, firing him within hours of the disclosure. Attempts to reach Zoph for comment were unsuccessful.
Technology reporter Kylie Robison first reported the firing on X, writing that Zoph was terminated for “unethical conduct.” News Of Fort Worth has not independently verified the specific data allegedly shared.
Personnel Scorecard
The shake-up delivers a clear win to OpenAI, which recently lost Jerry Tworek, its VP of research, to Anthropic. Besides Zoph and Metz, researcher Sam Schoenholz is also rejoining the ChatGPT-maker.
Key departures and returns since October
| Name | Role at OpenAI | Action | Date |
|---|---|---|---|
| Mira Murati | CTO | Left to found TML | Oct 2024 |
| Barret Zoph | Senior Researcher | Left, now returning | Jan 2026 |
| Luke Metz | Research Scientist | Left, now returning | Jan 2026 |
| Jerry Tworek | VP Research | Departed to Anthropic | Dec 2025 |
| Sam Schoenholz | Research Scientist | Left, now returning | Jan 2026 |
Timeline of a Reversal
- Monday: Zoph informs Murati he is considering leaving Thinking Machines
- Thursday morning: Decision to terminate Zoph is finalized
- Thursday afternoon: Simo notifies OpenAI staff of the rehires
- Thursday evening: News breaks publicly via social media
The speed of events left many employees at both companies stunned. One Thinking Machines engineer, speaking on condition of anonymity, described the mood as “whiplashed” and said internal systems were already being audited for potential data exposure.
Broader Implications
OpenAI has now re-absorbed two of the three public faces of its biggest talent exodus, strengthening its research bench at a moment when rivals like Anthropic and Google DeepMind are aggressively recruiting. For Thinking Machines, the departures raise questions about stability only weeks after the firm emerged from stealth with plans to build more controllable AI systems.
The startup had raised undisclosed seed funding and was rumored to be targeting a $1 billion valuation. Its remaining roadmap and ability to attract new talent will be closely watched in the coming months.
Key Takeaways
- OpenAI regains two high-profile researchers just two months after their exit
- Thinking Machines loses both co-founders and a third staffer in a single day
- Allegations of confidential data leaks triggered Zoph’s termination
- The personnel moves underscore the fierce competition for AI talent
