CONCEPT
The Training Problem Under Automation
Bainbridge's diagnosis of the structural impossibility of training operators for exceptional situations by exposing them only to routine ones — a mismatch that renders conventional training inadequate for the very scenarios training is supposed to address.
The training problem is the fourth of Bainbridge's ironies. To handle exceptions, operators need exposure to exceptions. But exceptions are, by definition, rare — and automation makes them rarer. Training programs substitute simulated exceptions, but simulations are built from previously imagined failures, and the failures that actually matter are the ones no one imagined. The problem compounds across generations: senior operators built their pattern library during manual operation; their juniors, trained in automated environments, never develop the same depth. Bainbridge argued that no training curriculum could fully solve the problem, because the solution required something training programs cannot provide — actual experience with the unanticipated. The problem has migrated directly into AI-era knowledge work, where junior developers trained on AI-assisted codebases never encounter the debugging experiences that built their senior colleagues' judgment.
In The You On AI Field Guide
Conventional training assumes a model in which expertise is acquired through explicit instruction,