Training AIs on millions of verifiable tasks in diverse RL environments will lead to AGI.

Added to korrents 24 Oct 2026 · How quotes work · Something wrong? Tell us

Shared subject first, then similar wording. They may agree or disagree.

• Scaling up GPT will never produce AGI, because a model trained on cross-entropy loss cannot get there; reinforcement learning in rich environments is required. GHGeorge Hotz29 Jun 2023AGIreinforcement learning
• A verifiable task can be optimised by reinforcement learning until a neural network performs it extremely well. AKAndrej Karpathy17 Nov 2025reinforcement learning
• Aggressive data curation may be harmful for frontier models that are significantly overtrained relative to Chinchilla optimality. DPDwarkesh Patel8 Sept 2026reinforcement learning
• Anything precise enough to benchmark is also precise enough to optimize for, and modern RL training increasingly resembles the shape of the benchmarks themselves. SKSayash Kapoor16 Apr 2026reinforcement learning
• Capability does not generalise for free: a model that will move mountains on an agentic task still tells the same bad joke it told five years ago. AKAndrej Karpathy20 Mar 2026reinforcement learning

See this claim on the map