Articles from dwarkesh.com
5 kept
- The Winning Essays for the Big Questions About AI
Three winning essays propose using AI foundations for pandemic eradication, advocating for light-touch AI policy for non-supply-chain countries, and adapting Hong Kong MTR's rail-plus-property model for AI lab profitability.
- Alex Imas and Phil Trammell – What remains scarce after AGI?
Economists argue that after AGI, scarcity may shift to human-relational services, but the labor share's fate depends on whether new capital varieties prevent satiation.
- What I learned this week - Pretraining parallelisms, Can distillation be stopped, Mythos and the cybersecurity equilibrium, Pipeline RL, On why pretraining runs fails
A set of rough notes exploring challenges in AI model distillation, cybersecurity offense/defense dynamics, and pipeline RL for training.
- Thoughts on AI progress (Dec 2025) - by Dwarkesh Patel
Dwarkesh Patel argues that current AI models lack the continual learning and on-the-job adaptability of humans, making AGI distant, but he expects explosive progress once true AGI arrives within a decade or two.
- Thinking through how pretraining vs RL learn
Reinforcement learning provides far fewer bits per FLOP than pretraining until models achieve high pass rates, limiting RLVR's ability to learn new capabilities.