Evolving AI: OpenAI & Google in Deep Trouble, Nvidia Just Acquired SUPER INTELLIGENCE!

Ilya Sutskever disappeared from the AI spotlight after leaving OpenAI and founding Safe Superintelligence (SSI). Now, after two years of extraordinary secrecy, we may finally be getting our first glimpse at what his new AI lab has been building. A reported comment from investor Gavin Baker suggested that SSI could reveal its first model in August 2026. SSI itself has not confirmed that timeline, but the possibility has attracted enormous attention, especially after NVIDIA entered a major strategic partnership with the company and reportedly gained access to SSI's closely guarded research. In this video, we break down Ilya Sutskever's vision for superintelligence, why he believes today's AI models still generalize dramatically worse than humans, and why simply scaling models with more data and compute may no longer be enough.
We explore Safe Superintelligence's approach to sample efficiency, continual learning, catastrophic forgetting, internal feedback systems, and Sutskever's fascinating idea of a “superintelligent 15-year-old” — an AI that doesn't begin knowing everything, but can continuously learn new skills from experience after deployment.

Could SSI be developing something fundamentally different from ChatGPT, Claude and Gemini? Or are expectations surrounding Sutskever's secretive AI company getting ahead of reality? Whatever SSI eventually reveals, its first model could tell us whether the next major AI breakthrough comes from building bigger models or from teaching machines to learn more like humans.

#IlyaSutskever #SafeSuperintelligence #SSI #Superintelligence #AGI #AI #OpenAI #NVIDIA #ArtificialIntelligence