Former OpenAI CTO Mira Murati launches AI startup, assembles top researchers

Mira Murati, former Chief Technology Officer at OpenAI, has launched a new artificial intelligence startup called Thinking Machines Lab, bringing together a team of around 30 leading researchers and engineers from companies like OpenAI, Meta, and Mistral.

Announced on Tuesday, the startup aims to develop AI systems that integrate human values while expanding their usability across a wider range of applications than its competitors, the company said in a blog post.

A significant portion of the startup’s team consists of former OpenAI employees, showcasing Murati’s influence in attracting top-tier talent. Among them is Barret Zoph, a well-regarded AI researcher who left OpenAI on the same day as Murati in late September. Zoph will serve as the Chief Technology Officer at Thinking Machines Lab.

The startup’s Chief Scientist, John Schulman, co-founded OpenAI before moving to rival AI firm Anthropic in August. He cited a desire to focus on AI alignment - the process of ensuring AI models adhere to human values and safety measures, which is also a key priority for Murati’s new venture.

According to sources, more former OpenAI employees are expected to join the company in the coming months.

Thinking Machines Lab seeks to differentiate itself from competitors by integrating research and product development from the outset. The company has pledged to contribute to AI alignment research by openly sharing code, datasets, and model specifications.

"While current AI systems excel in areas like programming and mathematics, we are working to develop models that can adapt to a broader spectrum of human expertise and enable a wider range of applications," the startup said in a statement.

Murati’s departure from OpenAI was part of a wave of high-profile exits following governance changes at the company. Prior to her tenure at OpenAI, she held leadership roles at Tesla and Leap Motion, an augmented reality startup.

Her new venture joins other startups founded by former OpenAI executives, including Anthropic and Safe Superintelligence, both of which have raised billions in funding and attracted former OpenAI researchers.

Thinking Machines Lab is already in discussions with investors to secure venture capital funding, reported Reuters.