CEO and co-founder of Thinking Machines Lab
Mira Murati
Biographies
Profile
Mira Murati is the engineer who turned OpenAI’s research into products that reached hundreds of millions of people. Born in Vlorë, Albania in 1988, she left at sixteen on a United World Colleges scholarship, then earned engineering degrees from Colby College and Dartmouth’s Thayer School. Before AI, she did the unglamorous product work — a summer at Goldman Sachs, then Tesla as a Model X product manager starting in 2013, then Leap Motion as VP of product and engineering. The résumé matters because it explains what came next: she wasn’t the person who discovered attention mechanisms, she was the person who shipped them.
She joined OpenAI in 2018 and was named CTO in 2022, just in time to oversee the launches of ChatGPT, DALL-E, GPT-4, Codex, and Sora. In practical terms, ChatGPT’s success was her execution — taking research that had been sitting in GPT-3.5 and turning it into a product the world actually used. For three strange days in November 2023, she was interim CEO of OpenAI after the board removed Sam Altman. Reporting later suggested she had provided material to Ilya Sutskever’s memo that prompted the firing; within the week she signed the employee letter demanding Altman’s return. Make of that what you will — it captures the compressed chaos of that moment better than any tidy narrative.
She left OpenAI in September 2024 and launched Thinking Machines Lab in February 2025, bringing a wave of senior OpenAI talent with her including Lilian Weng, John Schulman, Barrett Zoph, and Luke Metz. By mid-2025 the company had raised roughly $2 billion at a $12 billion valuation led by Andreessen Horowitz, with Nvidia, AMD, and Cisco joining in — and by late 2025 was reportedly in talks at a $50 billion valuation, before they had launched a single public model. Their first product, Tinker, shipped in October 2025 — an API for fine-tuning open-weight models, pitched at researchers and developers who want to train models without managing distributed clusters.
What’s interesting for anyone building with AI today: Thinking Machines is deliberately positioning itself differently from OpenAI and Anthropic. Fewer black-box chatbots, more developer infrastructure. The company has committed to a gigawatt of Nvidia compute and publishes a research blog called Connectionism. Early 2026 brought a round of attrition — three co-founders (Zoph, Metz, and Sam Schoenholz) returned to OpenAI, Andrew Tulloch went to Meta. Whether Murati can keep the band together long enough to ship something that justifies the valuation is the open question.
Controversies
- The Altman ouster: Murati’s role in providing material to Ilya Sutskever’s memo against Altman, followed by her public support for Altman’s reinstatement, drew scrutiny about judgment and loyalty during the November 2023 OpenAI crisis.
- Sora training data: In a Wall Street Journal interview about Sora in March 2024, she appeared visibly uncertain when asked whether YouTube videos were part of the training data — the clip became a reference point for how AI labs avoid straight answers on provenance.
- “Sky” voice / Scarlett Johansson: GPT-4o launched with a voice many felt mimicked Johansson’s Her performance, which she had declined to license. OpenAI pulled the voice. Murati, as CTO, was the public face of the product at launch.
- “Some creative jobs shouldn’t have existed”: A comment at a Dartmouth event in June 2024 about AI eliminating creative work — “some creative jobs maybe will go away, but maybe they shouldn’t have been there in the first place” — was widely criticized and became shorthand for tech-leader tone-deafness about creative labor.
YouTube
Spotify Podcasts
