Update Oct. 4: Tim Brooks, one of the executives on the Sora project at OpenAI, is also leaving. On X, Brooks has already announced his new employer: Google DeepMind.
“I had an amazing two years at OpenAI making Sora,” Brooks said. He will also be working at DeepMind on a similar topic: video generation and world simulation. This makes him yet another departure at OpenAI, with the exodus being extensive this year.
OpenAI’s board seems to consist of different members every day in 2024. Since the company’s management turmoil late last year, the shifting of roles has been incessant. Now CTO Mira Murati and Chief Research Officer Bob McGrew, both veterans at the AI company, are leaving. Meanwhile, OpenAI will soon let go of its nonprofit nature and CEO Sam Altman will get equity.
The latter was reported by Reuters. It’s another step towards commercializing the company. External investment is needed to keep OpenAI afloat.
However, the departures of Murati and McGrew are a lot more public. They are the next episodes in a saga whose end just cannot be seen. First Sam Altman disappeared, then he was rehired. In the months that followed, administrators left at intervals in a move that suggests they were all simply serving out their contract terms.
Murati: interim CEO and face of GPT-4o
Mira Murati was briefly CEO at OpenAI, for four days to be exact. Her predecessor and successor turned out to be the same person: Sam Altman. Despite photos and press releases that suggested unanimity, peace was far from restored with Altman’s return. One of the departures was Ilya Sutskever, co-founder and Chief Scientist at the AI company.
Murati made a recent appearance on behalf of OpenAI when it introduced the new, voice-driven, GPT-4o, among other things. That appears to be the final chord of a six-and-a-half-year track record. McGrew worked at OpenAI for nearly eight years, as a researcher, VP of Research and one month as Chief Research Officer. He, like other prominent administrators, shared the ominous tweet“OpenAI is nothing without its people” sent out into the world late last year.
Time for technology?
The soap opera surrounding OpenAI knows no end. Yet there are signs that the upheaval is almost complete. Despite former backer Elon Musk’s fiery criticism on the matter, OpenAI’s mission really does seem to have changed, one from a nonprofit to a tech giant and AI kingmaker for others such as Apple, Microsoft and Salesforce.
That products are still being built within the walls of OpenAI despite this hassle was demonstrated again recently. The company launched o1, the smartest model to ever power ChatGPT, albeit without some capabilities such as being able to browse the Internet like GPT-4o and other LLMs can.
Read also: OpenAI launches o1, making ChatGPT smarter than ever