Former OpenAI CTO Mira Murati has announced the launch of a new AI startup, Thinking Machines Lab. The company boasts a strong team of about 30 top researchers and engineers from OpenAI, Meta, and Mistral. The startup aims to build AI systems that encode human values and target a broader range of applications than its competitors.
Approximately two-thirds of the company's employees are former OpenAI staff, including notable researcher Barret Zoph, who left ChatGPT's creator to become the technical director at the startup. Additionally, OpenAI co-founder John Schulman has joined as the chief scientist after a brief stint at competitor Anthropic. Schulman expressed a desire to focus on AI alignment, which involves encoding human values into AI models to make them safer and more reliable—a key focus for Murati's new venture.
Sources indicate that more former OpenAI employees are expected to join the company. Thinking Machines Lab is reportedly in talks with investors to raise venture capital. Murati will serve as the CEO, joining a growing list of former OpenAI executives launching AI startups. Other companies like Anthropic and Safe Superintelligence have also attracted former OpenAI researchers and secured substantial funding.
The startup differentiates itself by having research and product teams collaboratively design its approach. It plans to contribute to AI alignment research by sharing code, datasets, and model specifications. Murati, who joined OpenAI in 2018 and led the development of ChatGPT, left during a period of significant organizational changes at the company.
