[Former OpenAI Chief Scientist Ilya's Secret Research Unveiled: Making AI Learn While Working Like Humans] According to monitoring by Beating, SSI has yet to release a model, but an overseas blogger pieced together its possible research direction from public clues: enabling AI to continue learning after deployment, rather than stopping progress once training is complete. The clues mainly come from Ilya Sutskever himself. He has repeatedly stated that the biggest flaw of existing models is their generalization ability, which is far weaker than that of humans. His ideal AI would be like a 'super-intelligent 15-year-old teenager,' capable of entering any industry and quickly mastering tasks through minimal experience. The Wall Street Journal also reported that SSI is studying 'overlooked aspects of human brain function.' After reviewing its confidential research, NVIDIA decided to invest and helped SSI expand its computing power tenfold within a year. Based on this, the developer speculated that SSI might be training an AI capable of identifying errors, absorbing experiences, and retaining old skills without forgetting them. However, this remains speculative. SSI has not disclosed its architecture, models, results, or demonstrations. [Original Link]