Mira Murati has never been comfortable in the spotlight. During her six years as OpenAI’s CTO, she stayed largely behind the scenes while Sam Altman dominated headlines. Since launching her own company, Thinking Machines Lab, she’s been even more elusive. So when she sat down with Bloomberg in San Francisco on Thursday for her first major media appearance in 18 months, the AI industry took notice.
The timing isn’t accidental. While Thinking Machines has spent the past year and a half quietly raising capital and building its team, competitors have grabbed massive market attention. OpenAI remains constantly in the news cycle. Anthropic’s momentum is the talk of Silicon Valley. And Elon Musk’s xAI has been folded into SpaceX ahead of an expected massive public offering. In this environment, staying quiet has diminishing returns.
Murati used the Bloomberg appearance to preview what Thinking Machines calls “interaction models” – a fundamentally different approach to AI interfaces. Unlike the turn-based, prompt-and-response dynamic that defines current AI products, her company’s models process continuous streams of audio, text, and video in 200-millisecond intervals.
The goal is capturing the texture of human communication – interruptions, mid-thought corrections, and thinking pauses – in near real-time. But Murati was careful to frame this as a first step rather than a finished product, declining to provide specific release dates.
This represents a significant shift in how AI systems might interact with users. Current models like ChatGPT or Claude require users to submit complete prompts and wait for responses. Thinking Machines’ approach could enable more natural, flowing conversations that mirror human interaction patterns. The technology could impact everything from customer service to creative collaboration tools.
The interview inevitably touched on “the blip” – the chaotic November 2023 week when OpenAI’s board fired Sam Altman and Murati became interim CEO. She said protecting the mission and team made her decisions feel obvious, even as the situation appeared to collapse from the outside. The company would have “imploded” without her involvement, she claimed.
But Murati acknowledged that good intentions don’t guarantee good outcomes. In retrospect, she would have pushed harder for more information, better transition planning, and greater transparency. She notably avoided saying whether she thinks things ultimately turned out well.
When asked about trusting her former boss, Murati sidestepped, instead highlighting a broader industry concern: too much power concentrated in too few hands. Her worry isn’t individual character but the absence of structural checks. Good people make bad calls, and well-intentioned organizations drift, she argued.
This reflects growing industry anxiety about AI governance. As AI systems become more powerful, questions about oversight and accountability have intensified. Murati’s comments suggest she believes the industry has focused too much on trusting leaders’ virtue rather than building robust governance structures.
Chang pressed Murati on recent high-profile departures from Thinking Machines, a topic she’s largely avoided publicly. Murati downplayed the exits, arguing that building a frontier AI lab compresses years of normal organizational volatility into months. While acknowledging that nine-figure compensation packages capture imaginations in the war for AI talent, she suggested money isn’t usually the whole story.
“When I wake up in the morning, I am not thinking about how to kill the competitor,” she said to audience laughter, describing her competitive instincts.
On AI’s broader future, Murati pushed back against both inevitable dystopia and utopia framings. Neither outcome is predetermined, she argued, and the current period will determine which direction we head. But she warned that if humans “take their hands off the wheel too soon,” the future will look very different – and not better.
This measured approach contrasts with both AI cheerleaders predicting universal prosperity and critics warning of imminent doom. Murati’s position suggests the outcome depends on choices made now about development, deployment, and governance. Her emphasis on human oversight aligns with growing calls for AI safety measures and regulatory frameworks as the technology rapidly advances.
