Generative AI Fundamentals Map
Map/Loops/reinforcing

Domains make the data

reinforcing loop · 4 steps · Trace on the map →

The cycle

Agents and robots doreal workTraces and physicaldataTraining data growsStronger modelsR
  1. 01Agents and robots do real workLong-horizon agents
  2. 02Traces and physical dataTrace capture
  3. 03Training data growsTraining data supply
  4. 04Stronger modelsReasoning models

What drives it

Every domain lane produces data the data lane lacks: traces of agents at work, robot experience, lab results. The domains are suppliers to the fundamentals, not only users.

What breaks it

Model collapse if generated data is not checked, and data rights if the work belongs to someone else.

What to watch

Share of training data that is experience or synthetic rather than scraped text.

Also feeds it

Sources

  1. 01Silver and Sutton’s Era of Experience (The Decoder)
  2. 02TechXplore: AI models collapse when trained on recursively generated data (Nature)