Skip to content
Slow Variables

Ask the data

Answers come from the same store as the site; every number is checked against the record it cites.

Crosswalk

The same observations carry a diffusion address (bucket, valve) and a capture address (layer, sub-layer). Diffusion asks how fast value moves; capture asks who keeps it. The leak on the diffusion diagram is the capture lens.

Diffusion bucketRelationCapture layerNoteShared indicators
Methodssame valveModelCapability growth is the model layer's product; inference price at fixed capability ⇄ price-per-token pass-through; open-weights lag ⇄ open-vs-closed gap.METR 50% time horizonMETR 80% time horizon50%/80% horizon ratio
Methodsinput toCompute & physicalCompute growth and hardware perf-per-dollar feed the methods stock.
Productssame valveServing & orchestrationProductisation runs through serving, routing and agent infrastructure.
Productssame valveDeployment & applicationUplift RCTs, pilot→production and app-layer margin net of inference measure the same valve.
Early adoptionsame valveAdoptersBTOS, BBD, Ramp and the Anthropic Economic Index are adoption from both ends.
Adaptationsame valveLabour & consumersProductivity, wages, employment and consumer surplus.
Adaptationleak fromDeployment & application / Deployment / FDE / implementation servicesDeployment/FDE services capture part of the surplus that adaptation creates.
Return arrowreturn arrowTraining inputExpert data, RL environments and continual-learning infra are the deployment→invention channel; lab recoupment ⇄ capex-to-revenue.
Leakleak fromCompute & physicalMargin stacking, HHI, circular financing: surplus exiting the chain at semis and cloud. Interpretive spine: Teece, Nordhaus, Perez, Ding (docs/interpretation).