Mira Murati made her first major media appearance in about 18 months on Thursday, offering a glimpse into the work underway at her startup, Thinking Machines Lab, while avoiding detailed timelines or broader announcements.
Speaking with Bloomberg’s Emily Chang in San Francisco, the former OpenAI chief technology officer described a new class of systems that Thinking Machines calls “interaction models.”
The appearance comes after months in which the company largely stayed out of public view while raising capital, recruiting researchers and launching Tinker, an API for fine-tuning open-source AI models.
New AI interface under development
Murati said the company’s models are designed differently from today’s prompt-and-response systems.
According to her, the technology processes continuous streams of:
- Audio
- Text
- Video
The system handles information in 200-millisecond intervals, allowing it to respond to interruptions, corrections and pauses in a way that more closely resembles human conversation.
Murati described the technology as an early step rather than a finished product and did not provide a release date.
Reflecting on OpenAI’s leadership crisis
Murati also revisited the turbulent period in November 2023, when OpenAI’s board removed Sam Altman and she briefly served as interim chief executive.
She said she felt certain about the decisions she made at the time because her focus was on protecting the company’s mission and employees.
Murati said OpenAI would have “imploded” without her involvement during the five-day period and the immediate aftermath.
Looking back, she said she would have sought:
- More information about the situation.
- A clearer transition plan.
She stopped short of saying whether she believes the episode ultimately had a positive outcome.
Governance concerns
When asked whether she still trusts Altman, Murati avoided giving a direct answer.
Instead, she spoke about what she sees as a broader problem across the AI industry: too much power concentrated in too few hands.
She argued that strong governance matters because even well-intentioned organizations can make mistakes.
According to Murati, discussions around AI have focused too heavily on individual leaders and not enough on institutional safeguards.
Talent departures and competition
Bloomberg’s Chang also asked about several high-profile departures from Thinking Machines in recent months.
Murati downplayed the issue, saying the process of building a frontier AI company compresses years of organizational change into a much shorter period.
She acknowledged that the race for talent has pushed compensation packages into nine-figure territory but suggested money is not always the deciding factor.
“When I wake up in the morning, I am not thinking about how to kill the competitor,” she said, drawing laughter from the audience.
AI’s future remains uncertain
Murati rejected the idea that either a utopian or dystopian future for AI is inevitable.
She said the current period will help determine how the technology develops and warned against removing humans from decision-making too early.
“If humans take their hands off the wheel too soon, the future will look very different, and not better,” she said.
Correction: The original story incorrectly stated the length of Murati’s tenure as OpenAI’s CTO.
TL;DR:
Former OpenAI CTO Mira Murati used her first major public appearance in 18 months to discuss Thinking Machines’ new AI models, revisit the 2023 OpenAI leadership crisis and call for stronger governance across the AI industry.
AI summary:
- Mira Murati made her first major media appearance in roughly 18 months.
- Thinking Machines is developing “interaction models” that process audio, text and video continuously.
- Murati said OpenAI would have “imploded” without her role during the 2023 crisis.
- She raised concerns about power being concentrated among a few AI leaders.
- Murati said AI’s future will depend on decisions made during the current period.
