OpenAI cofounder Ilya Sutskever announced a new venture yesterday, Safe Superintelligence Inc. The project will seek to launch a “safe” artificial general intelligence, hearkening to OpenAI’s founding goals but without its current pressures from investors and consumers.
The project—cofounded by OpenAI colleague Daniel Levy and Apple’s former AI lead Daniel Gross—comes after Sutskever announced his departure last month from OpenAI. Sutskever was part of the OpenAI board that ousted CEO Sam Altman in the fall, though he soon backtracked and worked to bring Altman back. Sutskever did not provide much detail on Safe Superintelligence, including its investment structure. He says safety would be engineered into the AI system from the beginning rather than adding fixes in response to discovered lapses.
Artificial general intelligence describes a general purpose data system with human-like capabilities, with the term “superintelligence” referring to a system whose capabilities extend beyond humans. Listen to an explainer of AGI here.
Comments