Mira Murati's Thinking Machines Lab is betting that the future of AI isn't faster prompts — it's conversations that feel human, a bet that pits her startup against OpenAI, Anthropic, and xAI in a market where the top players have raised more than $30 billion each.
Mira Murati's Thinking Machines Lab is betting that the future of AI isn't faster prompts — it's conversations that feel human, a bet that pits her startup against OpenAI, Anthropic, and xAI in a market where the top players have raised more than $30 billion each.
The former OpenAI chief technology officer made her first major public appearance in roughly 18 months Thursday, sitting down with Bloomberg's Emily Chang in San Francisco to preview what her company calls "interaction models." Rather than the turn-based prompt-and-response format that defines products like ChatGPT and Claude, Thinking Machines' models are designed to process continuous streams of audio, text, and video in 200-millisecond intervals — fast enough to capture interruptions, mid-thought corrections, and natural conversational pauses.
"It's about building systems that collaborate with humans rather than just responding to them," Murati said during the interview, according to TechCrunch's report. She framed the work as an early step and declined to provide a specific release date.
The technical challenge is significant. Low-latency multimodal streaming requires tightly integrated audio capture, vision pipelines, and stateful memory layers — architectural choices that increase compute and engineering costs compared with batch prompt flows. Thinking Machines' TML-Interaction-Small model responds in 0.40 seconds, according to the company, roughly the speed of natural conversation. The company has not disclosed test conditions or benchmark comparisons against competitors' real-time systems.
The competitive landscape Murati re-entered
Thinking Machines Lab has spent roughly 18 months raising capital, hiring researchers, and shipping one product: Tinker, an API for fine-tuning open-source models that launched in October 2025. The company secured $2 billion in funding and a gigawatt of Nvidia Vera Rubin compute capacity, according to The Next Web.
But the AI landscape has shifted dramatically since Murati left OpenAI in September 2024. Anthropic has raised $30 billion and reportedly drawn investor interest at an $800 billion valuation. Elon Musk's xAI has been folded into SpaceX ahead of what is expected to be a record public offering. OpenAI remains the dominant force in consumer AI, with ChatGPT maintaining the largest user base in the industry.
Meanwhile, Thinking Machines has lost several high-profile researchers. Co-founder and CTO Barret Zoph, co-founder Luke Metz, and founding team member Sam Schoenholz returned to OpenAI in January, according to reports. Five founding members have joined Meta, lured by compensation packages that reach into nine figures.
Murati downplayed the departures during the interview. Building a frontier AI lab from scratch compresses years of normal organizational volatility into months, she said. "When I wake up in the morning, I am not thinking about how to kill the competitor," she told Chang, drawing audience laughter.
Governance, the Altman firing, and the concentration of power
Chang pressed Murati on the November 2023 episode when OpenAI's board fired Sam Altman and she became interim CEO — a stretch employees inside OpenAI called "the blip." Murati said she felt clear about her decisions in each moment and that the company would have "imploded" without her involvement. But she acknowledged that clarity of intent is not the same as clarity about consequences, saying she would have pushed harder for more information and a better transition plan.
Asked whether she still trusts Altman, Murati sidestepped the question. Instead, she offered a broader critique: the AI industry has concentrated too many consequential decisions in too few hands, and too much attention has been paid to individual leaders' character and not enough to structural governance. Good people make bad calls, she said. Well-intentioned organizations drift.
On the broader trajectory of AI, Murati pushed back against both dystopian and utopian framings. Neither outcome is predetermined, she argued, and the current period is the one that will determine which direction the industry takes. She returned to a theme that connects her governance critique to her product philosophy: if humans take their hands off the wheel too soon, the future will look very different, and not better.
What it means for investors
Thinking Machines Lab remains a private company with no disclosed valuation, making direct investment impact difficult to assess. But the competitive implications are clear. OpenAI, Anthropic, and xAI are all racing toward real-time, multimodal interaction — the same direction Murati is heading. If Thinking Machines' interaction models deliver on their promise of sub-second conversational latency, the technology could pressure incumbents to accelerate their own roadmaps or acquire the startup outright.
For now, the company has one shipping product, a shrinking founding team, and a vision that remains unproven at scale. The window for quiet conviction, as Murati herself acknowledged, is closing.
This article is for informational purposes only and does not constitute investment advice.
