Who is Mira Murati?
When one of the most powerful figures at the company that started the modern AI boom walked away to build something of her own, the industry paid attention, and Mira Murati has held that attention ever since. She is the founder and chief executive of Thinking Machines Lab and the former chief technology officer of OpenAI. Albanian-born and trained as an engineer, she rose through OpenAI to lead much of its product and research work, overseeing the development and deployment of some of its best-known systems, before departing in 2024. In early 2025 she launched Thinking Machines Lab, which drew intense attention and one of the largest funding rounds ever raised by a young AI company. In 2026 the lab released Inkling, a large mixture-of-experts model, alongside the Tinker fine-tuning interface. Her team includes senior figures such as chief technology officer Soumith Chintala and chief scientist John Schulman, names that carry weight in the research community and that signalled the lab intent to compete at the very top of the field rather than at its edges.
What does Mira Murati think about AI?
Murati thesis is that AI should be customisable, understandable, and collaborative rather than a single one-size-fits-all product handed down from a lab. Thinking Machines emphasises tools that let people adapt models to their own needs and that make model behaviour more legible, which reflects her long-standing focus on responsible deployment and human oversight. She frames the goal as building AI that works with people and can be shaped by them, and she treats safety and customisation as connected rather than opposed, on the view that a system users can understand and steer is easier to trust and to correct than an opaque one. Having spent years at the centre of a lab that shipped some of the most widely used AI products, she speaks from the experience of watching how these systems behave once they reach millions of people, and that experience shows in her emphasis on careful rollout and on keeping humans meaningfully in control.
What is Mira Murati’s role in the AI race?
Murati is one of the most closely watched founders of the current wave, both because of her OpenAI pedigree and because Thinking Machines has attracted extraordinary resources. Her lab competes at the frontier while staking out a distinct position around customisation and openness, and its early releases, an open model and a fine-tuning platform, put that position into practice rather than leaving it as rhetoric. That makes her a central figure in the argument over whether the future belongs to a few closed megamodels or to adaptable, more widely controllable systems. In a race often framed as a contest of raw scale, she represents a bet that how a model can be shaped and understood matters as much as how large it is, and her prominence gives that bet unusual visibility.
Where does Mira Murati work?
She is the founder and chief executive of Thinking Machines Lab, the AI research company she started after leaving OpenAI. The lab was among the most heavily funded new entrants in the field, which gave it the means to hire senior researchers and to train large models from the outset. Murati leads both its research direction and the company as a whole, and she has assembled a team drawn substantially from the ranks of experienced researchers who had worked at the leading labs, positioning Thinking Machines as a serious frontier competitor rather than a niche player.
What are Mira Murati’s key projects?
Her current key projects are Inkling, a mixture-of-experts model that activates only a fraction of its parameters for any given task, which is a design aimed at capability without proportional cost, and Tinker, an interface for fine-tuning and customising models that puts her customisation thesis into a usable product. Earlier, as OpenAI chief technology officer, she was closely associated with the development and release of its flagship products, work that placed her at the centre of some of the most consequential deployments in the field. The line from that role to Thinking Machines is direct, since the lessons of shipping widely used systems inform the tools she is now building to let others adapt models responsibly.
What has Mira Murati written about AI?
Murati communicates mainly through interviews, public appearances, and the research and product releases of her lab rather than through a body of essays or opinion writing. Her public commentary emphasises responsible deployment, customisation, and the importance of making AI broadly useful and understandable, and she tends to speak in the measured register of someone who has had to make real decisions about releasing powerful systems to the public. That practical framing, drawn from experience rather than from theory, is the throughline of how she talks about the technology.
Does Mira Murati think humanity will survive AI?
Murati has generally taken a measured, safety-conscious stance rather than a doom-laden or dismissive one. She has emphasised human oversight, careful deployment, and understanding model behaviour, framing risk as something to be managed through good engineering and governance rather than predicting either certain catastrophe or guaranteed safety. Her public position treats the survival question less as a prophecy to be made and more as an outcome that depends on the choices labs and societies make now, which is consistent with a leader whose career has been about the practical work of putting powerful systems into the world without losing control of them.
Is Mira Murati a transhumanist?
There is no public evidence that Murati identifies as a transhumanist. She presents as a pragmatic AI leader focused on building and responsibly deploying capable, customisable systems, not as an advocate of transcending or replacing human biology. Her public emphasis on human oversight and on making AI understandable points toward keeping people in control of the technology rather than toward merging with it, and any attempt to attach the transhumanist label would go beyond what she has said or built.