RL on checkable tasks
Reinforcement learning on tasks with answers that can be checked automatically.
capabilityBuilt today Open on the map →
Facts
- Lane
- Algorithms and learning
- Wave
- 4 of 12 (build order, not a date)
- Needs
- 8 boxes all the way back through hard links
- Unlocks
- 63 boxes downstream through hard links
- Leverage
- None: already built
- Cluster
- Agents at work
Evidence as of October 2026
Reasoning models came from training on maths and code, where answers can be checked. Verifier’s law predicts those domains fall first.