In September, Mira Murati unexpectedly left her work as the Openai technology director, saying: “I want to create time and space for my own exploration.” The rumor in the Silicon Valley was that she gave way to start her own company. Today she announced that she is indeed the general director of the up-to-date public corporation of benefits called Thinking Machines Lab. His mission is to develop the highest quality artificial intelligence with a view to being useful and available.
Murati believes that there is a sedate gap between the rapidly developing AI and the understanding of technology by society. Even sophisticated scientists do not have a mighty understanding of AI’s abilities and restrictions. Lab thinking machines plan to fill this gap by building accessibility from the very beginning. He also promises to share his work by publishing technical notes, documents and actual code.
Lying at the basis of this strategy, Murati’s conviction that we are still in the early stages of AI, and the competition is far from closing. Although this took place after Murati began to plan her laboratory, the appearance of Deepseek-who claimed that he was building advanced reasoning models for a fraction of an ordinary cost-he says about his thinking that novices can compete with more effective models.
However, the machine thinking laboratory will compete at a high level of vast language models. “Ultimately, the most advanced models will unlock the most transformational applications and benefits, such as enabling innovative scientific discoveries and engineering breakthroughs,” writes the company Blog post On Tuesday. Although the term “Aga” is not used, the thinking of LAB machines believes that increasing the capabilities of its models to the highest level is crucial for filling the identified gap. Building these models, even with the era of the Deepseek era, will be costly. Although Lab, the thinking of machines has not yet shared their financial partners, it is convinced that millions are necessary.
The Murati pitch was attracted by an impressive team of scientists and scientists, many of whom have OpenAI on their CVs. These include the former Vice President of Research Barret Zoph (who is currently CTO in Machines Lab thinking), Multimodal Research Chief Alexander Kirillov, head of special projects John Lachman, and the best researcher Luke Metz, who left Ai a few months earlier. The main scientist of the laboratory will be John Schulman, the key inventor of Chatgpt, who left Opennai to the anthropic only last summer. Others come from competitors such as Google and Mistral Ai.
The team moved to the office in San Francisco at the end of last year and has already started work on many projects. Although it is not clear what his products will look like, the thinking of LAB machines indicates that they will not be imitators of ChatgPT or Claude, but AI models that optimize cooperation between people and AI – which Murati considers to be the current bottleneck in the field.
The American inventor Danny Hillis dreamed of this partnership between people and machines over 30 years ago. Hillis, a protected pioneer AI Marvin Minsky, built a great computer with powerful systems running in parallel – a precursor to clusters that today operate AI. He called it thinking machines. Before their time, thinking machines declared bankruptcy in 1994. Now the variety of its name and perhaps the legacy belongs to Murati.