Andrej Karpathy — AGI is still a decade away
dwarkesh.com · 27,216 words · saved by 7 readers
"The problems are tractable, but they're still difficult”
Playback speed × Share post Share post at current time Share from 0:00 0:00 / Transcript 370 17 66 Andrej Karpathy — AGI is still a decade away "The problems are tractable, but they're still difficult” Dwarkesh Patel Oct 17, 2025 370 17 66 Share Transcript The Andrej Karpathy episode. Andrej explains why reinforcement learning is terrible (but everything else is much worse), why model collapse prevents LLMs from learning the way humans do, why AGI will just blend into the previous ~2.5 centuries of 2% GDP growth, why self driving took so long to crack, and what he sees as the future of educati
saved by
related reading
- AI in 2025: gestalt — LessWronglesswrong.com
- Ilya Sutskever — We're moving from the age of scaling to the age of researchdwarkesh.com
- Dario Amodei — "We are near the end of the exponential"dwarkesh.com
- Animals vs Ghosts – karpathykarpathy.bearblog.dev
- As Rocks May Think | Eric Jangevjang.com
- Dario Amodei — "We are near the end of the exponential"dwarkesh.com
- The upcoming GPT-3 moment for RL | Mechanize, Inc.mechanize.work
- A Short History of Artificial Intelligenceevery.to
- The Extreme Inefficiency of RL for Frontier Models - Toby Ordtobyord.com
- GenAI Handbookgenai-handbook.github.io
- RLHF: Reinforcement Learning from Human Feedbackhuyenchip.com
- Thoughts on AI progress (Dec 2025)dwarkesh.com