Via The Rundown AI
| “Ilya Sutskever breaks silence on AI’s futureThe Rundown:
Safe Superintelligence founder Ilya Sutskever just appeared on the Dwarkesh Podcast, giving his take on scaling, ASI, his secretive startup, and more — arguing that research breakthroughs, not compute, will drive the next wave of progress. |
| The details: |
- Sutskever said that 2020-2025 was the “age of scaling”, but we’ve reached the point where research becomes the differentiating factor for AI breakthroughs.
- He forecasts 5-20 years until superhuman-like learning AI emerges, adding that the first ASI systems should be built to care about sentient life.
- Sutskever said that his startup, SSI, is taking a “different technical approach” to superintelligence, and called it an “age of research” company.
- He also revealed that SSI was raising at a $32B valuation and declined an acquisition offer from Meta, with his cofounder marking the only departure.
|
| Why it matters: Sutskever has been out of the spotlight since his exit from OpenAI, with SSI quietly working in the shadows — but his words carry massive weight in the AI world. His take on a “return to research” over compute comes at an awkward time, as the majority of the industry continues to pour massive money into scaling infrastructure.” |
Via The Neuron
| “Well, here are the top 10 things he had to say: |
- The “Jaggedness“ Problem: Ilya points out that current models are “jagged“—they crush PhD-level benchmarks but fail at basic tasks (like fixing a bug without breaking something else). He compares them to a student who studied 10,000 hours just to pass a test, but doesn’t actually understand how to learn.
- The “Age of Research“ is back: For the last 5 years, AI progress came from “scaling” (making models bigger). Ilya argues that pre-training data is running out, and we are returning to an era where ideas matter more than compute.
- The “Secret“ Principle: Humans learn faster than AI. A teenager learns to drive in 10 hours; an AI needs millions of simulations. Ilya claims to know the “missing machine learning principle“ that explains this gap but refused to share it—hinting this is exactly what SSI is building.
- “There are more companies than ideas“: Ilya’s bluntest observation. Everyone’s doing the same thing. The Silicon Valley mantra that “ideas are cheap, execution is everything“ breaks down when nobody’s having ideas.
- Emotions are the value function, not decoration: Ilya tells the story of a patient who lost emotional processing—he could still solve puzzles but couldn’t decide anything. Took hours to pick socks. Emotions tell you when to stop thinking and act.
- RL (reinforcement learning) now consumes more compute than pre-training: The balance has flipped. Long reasoning rollouts eat massive compute, and you get relatively little learning per rollout.
- The goal isn’t “finished AGI”—it’s a superintelligent learner: Think of a brilliant 15-year-old, not an omniscient oracle. Deploy it, let it learn on the job, and merge knowledge across instances. That’s the path.
- You can’t communicate AI power through essays—you have to show it: Ilya’s evolving view: gradual deployment matters because seeing AI do something is fundamentally different from reading about it. The world needs to feel the AGI capability.
- Ilya’s Timeline = 5-20 years to systems that learn as efficiently as humans and subsequently become superhuman. Wide range, but Ilya’s not hedging—he thinks it’s possible within that window.
- Research taste = beauty + simplicity + brain inspiration: When asked what makes great AI research, Ilya’s answer is almost aesthetic: “There’s no room for ugliness.“ The best ideas feel right from multiple angles simultaneously.”
|
0 Responses
Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.