Mira Murati launches Thinking Machines Lab to make AI more affordable

Last September Mira Murati unexpectedly left his job As the Chief Technology Director of OpenAi, saying, “I want to create time and space to do my own study.” The Hearing at the silicone valley It was that she retired to start her own company. Today, she has announced that she is indeed the CEO of a new public benefit corporation called Think Machines Lab. Her mission is to develop the first -class AI in order to make it useful and accessible.

Murathi believes there is a serious gap between the rapid AI progress and the public understanding of technology. Even complex scientists have no firm understanding of AI’s capabilities and limitations. Thinking Machines Lab plans to fill this gap by building accessibility from the beginning. He also promises to share his work by posting technical notes, documents and actual code.

At the heart of this strategy is Murati’s conviction that we are still in the early stages of AI and competition is far from closed. Although it happened after Murati started planning his laboratory, The appearance of Deepseek– which claims that it builds model models for reasoning for some of the usual costs – it increases its thinking that newcomers can compete with more efficient models.

However, the Machine Thinking Lab will compete from the high end of large language models. “In the end blog post Tuesday. Although the term “Agi” is not used, Thinking Machines Lab believes that increasing the capabilities of its models to the highest level is important for filling the precipice that is identified. Building these models, even with the effectiveness of the Deepseek era, will be expensive. Although the Think Machines Lab has not yet shared its funding partners, it is confident that it will raise the necessary millions.

Murati’s terrain has attracted an impressive team of researchers and scientists, many of whom have Openai on their CVs. These include the former Vice President of Research Barret Zoph (who is already CTO at Mishing Machines Lab), a multimodal research leader Alexander Kirilov, John Lahman Special Project Head and top researcher Luke Metz, who left an open AI for a few months. The main scientist in the laboratory will be John Schulman, a key inventor of Chatgpt, who left Openai for Anthropic only last summer. Others come from competitors like Google and Mistral AI.

The team moved to an office in San Francisco late last year and has already started working on a number of projects. Although it is not clear what his products will look like, Thinking Machines Lab shows that they will not be copies of Chatgpt or Claude, but models of AI that optimize cooperation between people and AI – which Murati sees as the current obstacle in the field.

American inventor Danny Hilis dreams of this partnership between people and machines more than 30 years ago. AI Pioneer Protege Marvin Minski, Hilis built a super computer with powerful chips working in parallel – a precursor to the clusters that work today. He called it thinking machines. Before his time, thinking machines declared bankruptcy in 1994. Now a version of his name and perhaps his inheritance belongs to Murati.