After OpenAI co-founder Ilya Sutskever stepped down as chief scientist, he co-founded Safe Superintelligence Inc. (SSI) with his colleague Daniel Levy and Apple's former artificial intelligence director Daniel Gross, dedicated to building safe superintelligence. The move attracted widespread attention, especially after Sutskever previously led the OpenAI Super Alignment team and was involved in the brief ouster of OpenAI CEO Sam Altman in November 2023. The establishment of SSI marks Sutskever's shift in its focus on artificial intelligence security from OpenAI to a more focused and independent platform.
OpenAI co-founder Ilya Sutskever recently announced that he will devote himself to new entrepreneurial projects after resigning as OpenAI's chief scientist in May this year. He and OpenAI colleague Daniel Levy and former Apple artificial intelligence director and Cue co-founder Daniel Gross announced that they are co-founding Safe Superintelligence Inc., a start-up whose mission is to build safe superintelligence.
In information posted on the SSI website, the founders said that building secure superintelligence is "the most important technical problem of our time." Furthermore, "We view safety and capabilities as one, as technology problems solved through revolutionary engineering and scientific breakthroughs. We plan to advance capabilities as quickly as possible while ensuring our safety is always ahead of the curve."
So, what is superintelligence? It is a hypothetical concept that represents a theoretical entity with intelligence that far exceeds that of the smartest humans.
The move is a continuation of Sutskever's work at OpenAI. He was a member of the company's Super Alignment team, responsible for designing ways to control powerful new artificial intelligence systems. But with Sutskever's departure, the team was disbanded, a move that was harshly criticized by one of its former leaders, Jean Leike.
SSI claims it will “advance the development of secure superintelligence in a straight line, focused on one goal and one product.”
Of course, OpenAI's co-founder was instrumental in the brief ousting of CEO Sam Altman in November 2023. Sutskever later said he regretted his role in the incident.
The establishment of Safe Superintelligence Inc. will enable Sutskever and others to focus more on solving the problem of safe superintelligence and achieve product development with a single goal.
SSI aims to solve the extremely challenging problem of building secure superintelligence, and its future development deserves continued attention. Whether Sutskever's new company succeeds in achieving its ambitious goals will have far-reaching consequences for the field of artificial intelligence.