simons-law ADG-0023

Simon's law

“The apparent complexity of a system's behaviour is largely a reflection of the complexity of the environment it is in.”

Also: Simon's ant

Origin

Herbert Simon, 1969, The Sciences of the Artificial, from the parable of an ant's winding path across a beach: the path is complex, the ant is not.

Why it should, or should not, apply to models

Replace the ant with a model and the beach with the harness and task environment. The law predicts that much of what looks like agent capability or agent failure is environment: the tools offered, the guides, the sensors, the shape of the task. That is a testable, and uncomfortable, prediction — it says comparing models on a benchmark measures the beach as much as the ant, and that changing the environment can move outcomes as much as changing the model.

Standing: Untested

Standing counts claims a person has reviewed. Claims reviewed by AI alone are listed and marked, and do not move it.

Where it holds

Claims showing the heuristic applying to models.

A large share of an agent benchmark score is attributable to the harness around the model rather than the model itself: holding the model fixed and changing only guides, sensors, retries and state management is reported to move the same model from 30.9% to 74.6% on GAIA and 30th to 5th on Terminal Bench — so a score reported for "a model" on an agentic benchmark is a score for a model-plus-harness, and comparisons across labs with different harnesses are not comparisons of models.
Whether the measurement made the findingReviewed by AI
Holding the model fixed and changing only the harness moved the score by as much as a model upgrade — the beach, not the ant.
A correction made in conversation fixes one exchange and is gone when the session ends; the same correction encoded as a guide rule, a sensor, or a permission fixes every future run — so an agent system improves over time only to the extent that failures are converted into harness structure rather than re-applied as prompts, and the rate of new guide rules per week falling is the sign the conversion is working.
Following an unfamiliar procedureReviewed by AI
Improvement accrues in the environment (harness structure), not in the ant, which is unchanged between runs.

Add a case where it holds or breaks

A break is worth more than a hold. If you have seen this fail for a model, with the setup written down, that is the most useful thing you can add here.