Debating AlphaGo Moments and Why Long-Horizon Agents Skip Theory-Building
A discussion sparked by weballergy argues that AlphaGo's Move 37 may have been within the model's self-play distribution, only novel relative to human knowledge. A follow-up post contends the real barrier for long-horizon agents is not capability but the lack of incentives and goal structures that motivate theory-building.
2026-11-22 ~ 2026-11-22 · 2 related posts
- Long-Running Agents With Long-Horizon Incentives Could Build Theories Too — weballergy · 2026-11-22
- Was Move 37 really out-of-distribution? A case for long-horizon self-play — weballergy · 2026-11-22