Are Dune’s Thinking Machines Here?
Artificial Intelligence
2-minute read
Another key member from OpenAI has ventured out to start a new company. This time, it's Mira Murati, the former CTO of OpenAI who briefly served as CEO during the internal shake-up in November 2023. She departed in September 2024 and announced today the launch of her new AI startup, Thinking Machines Lab.
The reason I say "another" is because Murati follows in the footsteps of other former members like Ilya Sutskever, Andrej Karpathy, and Elon Musk, who founded Safe Superintelligence Inc, Eureka Labs, and xAI, respectively. Even earlier, some founding members had left to establish Anthropic.
A Truly Open and Safer AI Company?
On February 19, Murati shared this news on X, highlighting three core directions:
Visiting the Thinking Machines Lab website, you'll find more former OpenAI employees. For instance, CTO Barret Zoph was previously OpenAI's VP of Research. Chief Scientist John Schulman, one of the six co-founders, was deeply involved in developing ChatGPT and led AI safety tasks after the Superalignment Project was dissolved. He even briefly joined Anthropic. According to external reports, Murati has recruited about twenty employees from OpenAI.
The website also elaborates on the company's vision, emphasizing open science by regularly publishing technical articles, papers, and code. They also set specific goals for security, such as establishing high standards to prevent misuse and sharing safety-related technologies with the industry to accelerate AI alignment research.
Is the Company Name from 'Dune' a Satire or a Nod to Safety?
Murati's new company seems to challenge OpenAI, aiming to outdo her former employer with an open approach. However, the name Thinking Machines, borrowed from the sci-fi classic 'Dune,' seems a bit at odds with AI safety.
We've previously discussed how the 'Dune' universe lacks any computer-like technology because, in the protagonist's ancient storyline, there was a war against robot domination. Humans, reliant on computers and robots, created Thinking Machines that mimicked human thought to win an inter-empire conflict, only to be enslaved by them for a thousand years until a human rebellion sparked a holy war.
Cover of a prequel by Brian Herbert, son of the 'Dune' author, showing two Thinking Machines enslaving humans. (Source: Neo Encyclopedia)
Using this name for an AI company might seem ironic, but perhaps it's meant to highlight their commitment to AI safety. Another interesting coincidence is the recent release of the book, 'The Thinking Machine: Jensen Huang, Nvidia, and the World’s Most Coveted Microchip,' which also uses the 'Dune' reference.
Left: English cover, Right: Taiwanese cover, showing the design language difference with the international version emphasizing 'Nvidia Green' and the Taiwanese version featuring the subject's portrait. (Source: Amazon, Books.com.tw)
Regarding this coincidence of names, one can only say that the influence and popularity of Dune is truly enormous. Frank Herbert’s use of the name to refer to “machines with human-like thinking” was so apt that it has inspired countless imitators. As for whether Thinking Machines Lab will, as it claims, be more open and secure, and create artificial intelligence that benefits humanity, we’ll just have to wait and see.
