Last September, Mira Murati suddenly Leaving her work As the technology director of Openai, I say that I want to create time and space to make my own discovery. The Rumors in Silicon Valley She resigned to start her own company. Today, she stated that she was actually the CEO of a new public interest group called Thinking Machines Lab. Its mission is to develop AI for the purpose of making it helpful and accessible.
Murati believes that there is a serious gap between AI progress quickly and public understanding of technology. Even sophisticated scientists do not grasp AI's ability and limitations. Lab thinking plans to fill that space by building accessible from the beginning. It also promises to share its work by publishing technical notes, papers and actual codes.
As the foundation for this strategy is Murati's belief that we are still in the first phase of whose stage, and the competition is not closed. Although it happened after Murati started planning his lab, but Deepseek's appearanceTietWhich announced the construction of advanced theoretical models for a part of the usual cost, related to her thoughts that newcomers could compete with more effective models.
However, Machines Machines Lab will compete on the high -end of the large language models. Finally, the most advanced models will unlock the most change applications and benefits, such as allowing new scientific discoveries and technical breakthroughs, the company wrote in A blog post On Tuesday. Although the term AG AGI is not used, the Machine Lab believes that upgrading its capabilities to its highest level is very important to fill the gap that it has determined. Building those models, even with the effectiveness of the Deepseek era, will be very expensive. Although the Machines Lab thought has not shared its sponsorship partners, it is confident that it will increase millions of people needed.
Murati's yard attracted an impressive group of researchers and scientists, many of them had Openai on their resume. Those who include former Vice President of Research Barret Zoph (currently CTO at Thinking Machines Lab), head of multimodal research Alexander Kirillov, head of special projects John Lachman, and a researcher. Luke Metz's head, who left AI a few months earlier. The chief scientist of the laboratory will be John Schulman, an important inventor, who left Openai for anthropology only last summer. Others come from competitors like Google and Mistral AI.
The team moved to an office in San Francisco at the end of last year and started working for some projects. Although it is unclear what its products will look like, thinking that Machines Lab shows that they will not be a copy of TATGPT or Claude, but AI models optimize the cooperation between people and AI. , which Murati is considered the current bottle in this field.
American inventor Daniel Hillis dreamed of this cooperation between humans and machines more than 30 years ago. A guard of AI Pioneer Marvin Minsky, Hillis has built a supercomputer with powerful chips running in parallel with a pioneer for AI operating clusters today. He called it thinking. Before the era, the thought machines declared bankruptcy in 1994. Now is a variant of its name, and perhaps its heritage, belongs to Murati.