Mira Murati Resurfaces with New Company and AI Governance Warning Mira Murati has re-entered the AI spotlight after 18 months of relative silence, using a tightly choreographed media appearance to launch a new product vision and issue a warning about how the industry is being governed.

A careful return after 18 months

Murati, formerly CTO of OpenAI and now CEO of Thinking Machines Lab, had largely avoided public stages despite having helped ship ChatGPT, DALL·E, and Codex. On Thursday in San Francisco, she sat down with Bloomberg’s Emily Chang in what was described as her first major media appearance in roughly a year and a half, a “carefully managed re-entry” into an AI conversation that has moved at “breakneck speed” without her.

During that quiet period, Thinking Machines raised around $2 billion, secured a gigawatt of Nvidia Vera Rubin compute, hired—and then lost—a notable number of researchers, and shipped a single product: Tinker, an API for fine‑tuning open‑source models.

The new bet: “interaction models”

Murati used the appearance to preview what the company calls “interaction models,” pitched as “a fundamentally different kind of AI interface.” Instead of today’s turn‑based prompt‑and‑response, the models are designed to ingest continuous streams of audio, text, and video in 200‑millisecond intervals, capturing interruptions, mid‑thought corrections, and pauses “in something closer to real time.”

Technically described as “full duplex,” the flagship TML‑Interaction‑Small model reportedly responds in about 0.40 seconds, “roughly the speed of natural conversation,” aligning with Murati’s thesis that powerful AI should enable “closer human collaboration, not less of it.” She stressed this is only a first step and declined to offer a release date, keeping Tinker as the company’s only shipping product for now.

Governance warning and OpenAI reflections

Beyond product talk, Murati argued that the AI sector “lacks structural governance checks,” warning that decisions about increasingly powerful systems are too concentrated in a small number of hands. She framed the current moment as one that will determine AI’s long‑term trajectory.

She also reflected on the chaotic November 2023 firing of OpenAI CEO Sam Altman, when she briefly served as interim chief. In hindsight, she acknowledged a need for more transparency and planning around that crisis, even as she defended the clarity of her decision‑making at the time.

Taken together, Murati’s reappearance positions Thinking Machines as both a technical contender in next‑generation interfaces and a critic of how power over those technologies is currently structured.

Continue reading https://foxvector.com/stories/019e9a10-a5c4-2039-71cd-1bd8ddac3afc