AI Leader Re‑Emerges With New Vision for Real‑Time Interaction

Mira Murati, once a behind‑the‑scenes figure at OpenAI and founder of Thinking Machines Lab, recently stepped into the spotlight for the first time in over a year. Her appearance on Bloomberg in San Francisco offered a glimpse into her new project and a candid look at past challenges. Thinking Machines has quietly built a team of researchers, raised funds, and released an API called Tinker that lets developers fine‑tune open‑source AI models. While competitors like Anthropic and Elon Musk’s xAI keep grabbing headlines, Murati decided it was time to make a noise about her own work. She described the lab’s “interaction models” as a fresh way to engage with AI. Instead of waiting for a prompt and receiving a single reply, these models process streams of audio, text, and video every 200 milliseconds. The goal is to capture the natural rhythm of human conversation—interruptions, pauses, and corrections—in near real time. Murati stressed that this is an early prototype and she did not give a launch date.

The interview also touched on her brief stint as interim CEO of OpenAI after Sam Altman was fired. Murati said she made decisions with the company’s mission and team in mind, but later admitted she would have wanted more information and a clearer transition plan. She avoided commenting on whether the outcome was ultimately positive. When asked about her trust in former leaders, Murati redirected to a broader issue: too many critical choices are made by a handful of people. She argued that good intentions do not guarantee sound decisions and called for stronger governance structures across the industry. Murati also addressed recent departures of key researchers from her lab. She explained that building a cutting‑edge AI center compresses typical organizational growth into months, and that while high salaries attract talent, they are not the sole factor. She added a light‑hearted remark that her daily thoughts do not revolve around outcompeting rivals. The conversation closed with a discussion of AI’s future impact on jobs and safety. Murati warned against complacency, noting that the present era will shape whether AI becomes a force for good or harm. She emphasized that humans must stay actively involved to steer the technology toward beneficial outcomes.

Actions