Mira Murati Aims to Bridge Human-AI Gap with New Firm

Former OpenAI CTO Mira Murati launched a new artificial intelligence startup with an aim to bridge the gap between rapidly advancing AI technologies and the public’s understanding of them.

See Also: From Silos to Synergy: Gen AI Aligns IT and Security Teams

Murati quit OpenAI in September after a six-year tenure that saw her contribute to the development of ChatGPT and DALL-E. She was briefly OpenAI interim CEO when Sam Altman was abruptly fired in November 2023 after being reinstated by a reconstituted board.

The startup, christened Thinking Machines Lab, is an attempt to create AI systems that can adapt to individual user needs. The company in a blog post outlined its approach, focusing on the importance of developing multimodal AI systems that can collaborate with people. Unlike the fully autonomous agents that have dominated conversations in the industry, Thinking Machines Lab’s tools are meant to be “more flexible, adaptable, and personalized AI systems.”

“Our goal is simple, advance AI by making it broadly useful and understandable through solid foundations, open science and practical applications,” Murati said in a social media post.

The name “Thinking Machines” harks back to the vision of American inventor Danny Hillis, a protégé to AI pioneer Marvin Minsky over three decades ago. Hillis developed a supercomputer with parallel processing chips, a precursor to the clusters that now power AI, in a company he ran called Thinking Machines. It declared bankruptcy in 1994.

The 2025-launched Thinking Machines Lab team includes former OpenAI luminaries such as John Schulman, co-founder and deep reinforcement learning pioneer, and Barret Zoph, the ex-chief research officer at Sam Altman’s company. The company’s staff also includes top engineers and researchers from firms such as Google DeepMind, Character AI and Mistral.

The San Francisco company’s strategy involves enabling users to adapt AI systems to their specific needs, laying technical foundations for more capable AI models and promoting a culture of open science. The company says that understanding AI’s potential requires sharing knowledge and collaborating with the broader research community, which is why it plans to publish technical papers, share code and engage with external researchers.

AI safety is said to be a key focus, with the lab taking an “empirical and iterative” approach to prevent misuse and enhance the technology’s societal benefits.

Central to its efforts is multimodality, or the ability for AI systems to understand and interact across multiple forms of data such as text, images and audio. These systems help with human-AI collaboration, making technology more intuitive and effective in real-world scenarios, the company said.

“We’ll focus on understanding how our systems create genuine value in the real world,” the blog post said. “The most important breakthroughs often come from rethinking our objectives, not just optimizing existing metrics.”

Thinking Machines Lab has not disclosed the specifics of its product offerings, but Murati’s team is working on models that could affect fields like scientific research and programming. The company has also not shared how much funding it has raised so far, or from who, amid reports of a potential $100 million investment.

Murati’s new venture comes at a time of rapid advancements in AI, with companies such as OpenAI, xAI and DeepSeek pushing the envelope on new capabilities. Other former OpenAI employees have also launched their own firms Safe Superintelligence and Anthropic.