Schmidhuber on RSI and AI Moats
max_paperclips · x · 2026-07-15
This is a long interview revolving around Recursive Self-Improvement (RSI), with key points including:
- He believes AI labs will find it hard to build a durable moat around RSI.
- Researchers should focus on directions that truly push model capabilities, rather than only on labs' internal closed advantages.
- He predicts that data centers may significantly depreciate in the future.
- He has relatively lower concerns about AI safety, emphasizing that models still need several key improvements.
The interview timeline also covers: how far superintelligence is from humans, whether ChatGPT surprised him, whether AI takeoff will happen suddenly, and the idea that "intelligence means efficiency."
Related event: Schmidhuber Argues Recursive Self-Improvement Offers No Lasting Moat(3 posts)→
More from AGI Musings
- jjvincent invokes Terence Tao: ceding exploration to AI means ceding human agency — jjvincent · 2026-09-11
- Better languages emerged from struggle: AI shortcuts may cost the commons — jjvincent · 2026-09-11
- OpenRouter agents now out-consume humans as AI usage arrives in three waves — AccBalanced · 2026-09-11
- If AI teleports us to solutions, how do underlying fields develop? — jjvincent · 2026-09-11
- Economist argues safe AGI comes from engineers inside big labs, not regulation — paulnovosad · 2026-09-11
- Op-ed: the ">10% extinction" narrative is liability evasion — AI is just software, and the vendor is the defendant — gerardsans · 2026-09-11