What Is 'Recursive Self-Improvement' in AI?
Recursive self-improvement: AI helping engineers build a smarter AI, which helps build an even more capable one, and the process continues. Could we test those systems fast enough and keep meaningful human control? Serious questions, but a possible future risk is not something that has already happened.
California's Nov 18, 2026 AI safety report asks for power to monitor and, if needed, cap it.
Full episode: https://youtu.be/Hbj3FUtwEHE
#AI #AISafety #AIExplained
Daugiau trumpų vaizdo įrašų
Visi vaizdo įrašai
Will AI Take Your Job? The Chainsaw Lesson
Nov 20, 2026
3 peržiūros
Catholic Bishops Set New Rules on Investing in AI
Nov 20, 2026
2 peržiūros
2 Senators: AI Cyber Threats Are Outpacing U.S. Defenses
Nov 20, 2026
29 peržiūros
California Wants Power to Shut Down Dangerous AI
Nov 20, 2026
22 peržiūros
AI Won't Make the World Better. People Will.
Nov 18, 2026
17 peržiūros
Can AI Help Investigators Protect Children?
Nov 18, 2026
23 peržiūros
Pause AI Until It's Safe, or Keep Building?
Nov 18, 2026
26 peržiūros
What If We Stopped Building AI?
Nov 17, 2026
20 peržiūros