Mira Murati, once the chief technology officer of OpenAI during the development of ChatGPT, has unveiled a new venture — Thinking Machines Lab — with an initial team of about 30 researchers and engineers.
Notably, the group includes several former OpenAI colleagues and hires from Character.AI and Google DeepMind.
Among the recruits are John Schulman, co-founder of OpenAI and now head of research at Thinking Machines Lab, Barret Zoph, another former OpenAI researcher who steps in as chief technology officer, and Jonathan Lachman, the ex-leader of OpenAI’s special projects unit. Murati, 35, has described the aim of the new venture as building AI capabilities that users can customise to address their own needs more effectively.
Murati joined OpenAI as vice-president of applied AI and partnerships in 2018, assuming the role of CTO in 2022. During a tumultuous period at OpenAI, she briefly served as interim chief executive when founder Sam Altman was abruptly dismissed and later reinstated in 2023. She ultimately resigned last September to “create the time and space” to pursue a fresh project.
In a statement posted on the new company’s website, Thinking Machines Lab said it is working on “helping people adapt AI systems to work for their specific needs”, emphasising open publishing and code transparency. Its mission is to bridge current gaps in AI understanding by granting broader access to advanced machine learning tools and fostering collaboration between humans and AI.
Murati’s former colleague at OpenAI, Sam Altman, has recently spoken about the transformative, if sometimes unpredictable, power of AI systems. He estimates that the latest models may already be capable of handling 5 per cent of “economically valuable” tasks, yet acknowledges these technologies will produce both “a lot of good and a lot of bad” in society.
