Former OpenAI CTO Mira Murati has taken the wraps off her new company, Thinking Machines Lab. The startup has come out of stealth with a clear focus on artificial intelligence: making AI systems easier for people to understand, adapt, and use for their own goals.
Murati is leading Thinking Machines Lab as CEO. The company’s early leadership also includes OpenAI co-founder John Schulman as chief scientist and Barret Zoph, who led model post-training at OpenAI, as CTO.
A New AI Startup Built Around Customization
Thinking Machines Lab says it wants to build tooling that can make AI work for people’s unique needs and goals. Its stated direction is not only to build more capable systems, but also to make those systems more widely understood and more customizable than what is currently available.
That emphasis matters because the company is framing the next phase of AI development as a usability and transparency problem as much as a raw capability problem. The startup’s blog post, shared with TechCrunch, argues that AI has advanced quickly while important gaps remain in how people understand and shape these systems.
“The scientific community’s understanding of frontier AI systems lags behind rapidly advancing capabilities,”
The same post says knowledge about training these systems is concentrated inside top research labs. Thinking Machines Lab argues that this limits public discussion about AI and makes it harder for people to use AI effectively.
The company also says today’s systems remain difficult to customize for specific needs and values. Its public positioning is therefore centered on a practical question: how can powerful AI become more adaptable to different users, domains, and priorities?
What Thinking Machines Lab Plans To Build
According to the company’s blog post, Thinking Machines Lab plans to focus on multimodal systems that work with people collaboratively. The goal is to build systems that can adapt across a broad range of human expertise and support a broader set of applications.
The startup also says it is building models at the frontier of capabilities in areas such as science and programming. In its view, the most advanced models could enable applications with major downstream benefits, including novel scientific discoveries and engineering breakthroughs.
Murati summarized the company’s direction publicly by saying she started Thinking Machines Lab with a team of scientists, engineers, and builders. The company is focused on three areas:
- Helping people adapt AI systems to work for their specific needs
- Developing strong foundations to build more capable AI systems
- Fostering a broader effort around AI development
Those priorities connect the company’s technical ambitions with its product philosophy. Thinking Machines Lab is not presenting itself only as a model-building organization. It is also describing a push to make advanced AI more usable and better aligned with the people applying it.
Safety Is Part Of The Pitch
AI safety is also described as a core part of Thinking Machines Lab’s work. The company says it plans to reduce misuse of the models it releases, share best practices and recipes for building safe AI systems, and support outside alignment research.
That external research support may include sharing code, datasets, and model specifications. In the company’s framing, safety work is not only an internal process but also something that can be strengthened by broader access to technical materials.
“We’ll focus on understanding how our systems create genuine value in the real world,”
The company also wrote that important breakthroughs can come from rethinking objectives, rather than only improving existing metrics. That statement signals a broader definition of progress: not simply whether AI systems score better on benchmarks, but whether they create meaningful value in practical use.
Based on the source material, Thinking Machines Lab is tying safety, customization, and capability together. Its stated bet is that more capable AI should also be easier to steer, easier to study, and better connected to real human needs.
Murati’s OpenAI Background Shapes The Story
Murati left OpenAI last October after six years at the company. At the time, she said she was stepping away to “do her own exploration.”
Her OpenAI role is central to why Thinking Machines Lab is drawing attention. Murati joined OpenAI in 2018 as VP of applied AI and partnerships. She was promoted to CTO in 2022 and led work on ChatGPT, the text-to-image AI DALL-E, and Codex, the code-generating system that powered early versions of GitHub’s Copilot programming assistant.
She was also briefly OpenAI’s interim CEO after CEO Sam Altman’s abrupt firing. Altman has described her as a close ally.
Before OpenAI, Murati spent three years at Tesla as a senior product manager of the Model X. During that period, Tesla released early versions of Autopilot, its AI-enabled driver-assistance software. She also served as VP of product and engineering at Leap Motion, a startup building hand- and finger-tracking motion sensors for PCs.
A Team Drawn From Major AI Labs
Thinking Machines Lab’s launch follows months of rumors that Murati was hiring high-profile AI researchers and staffers for a new AI venture. The company’s blog lists 29 employees from OpenAI, Character AI, and Google DeepMind, among other top firms.
The company is also actively hiring. Its post lists openings for machine learning scientists and engineers, as well as a research program manager.
At one point, Murati was said to be in talks to raise over $100 million from unnamed VC firms. The company’s blog did not confirm or deny that.
Murati is also part of a wider pattern of former OpenAI executives launching startups. The source names Ilya Sutskever’s Safe Superintelligence and Anthropic as examples of rivals in that growing group.
For now, Thinking Machines Lab’s public message is focused less on a specific product and more on its direction: frontier AI systems that are collaborative, customizable, better understood, and built with safety work in view.