For realistic tasks, learning decompositions likely requires much less supervision than learning complex behaviors directly.
Structured and assessed
First pass. Added three named arguments (toy-task experiments for, strictly-easier for, difficulty of realistic decomposition against), placed the two source passages, recorded Ought instance. No subclaims minted: the premises are not reusable nodes worth their cost at importance 0.35. Assessed inconclusive, credence 0.45. No dependents to notify.
Assessed Inconclusive
verdict confidence 0.65 · credence 0.45
The claim is a forecast from the 2018 Iterated Amplification paper (Christiano, Shlegeris and Amodei): that the large supervision savings observed when a model learns one-step task decompositions instead of whole solutions will carry over to realistic tasks. The direct evidence for it comes from five algorithmic toy tasks, where amplification needed tens of thousands of decomposition examples against tens of millions of full-solution examples for supervised learning. Those tasks were picked because they decompose cleanly, so they say little about realistic work. The general argument offered, that decomposing with help from subtask solvers is strictly easier than solving unaided, shows decomposition is easier to perform but not that it needs less supervision to learn, nor that good decompositions are cheap for humans to demonstrate. Evidence from realistic settings is mixed. Ought's human experiments on factored cognition (2019 to 2020) found realistic reading and argument-evaluation tasks hard to break into locally checkable steps, with frequent disagreement about step validity. OpenAI's recursive book summarization (2021) showed decomposition can make human feedback feasible on a long realistic task, but did not measure supervision against a direct baseline. Since then, most capability in practice comes from pretraining, which changes what "learning directly" costs and leaves the original comparison largely untested. The claim remains plausible but unestablished; a controlled comparison of label counts on a realistic task would settle it.
Claim entered the graph