The mechanism is specifically human. The structures do not weaken on a thirty-year schedule. The codes do not expire. What changes is the felt relationship between the profession and the catastrophes that shaped the codes. When engineers who lived through a disaster design bridges, they design with the weight of the disaster in their judgment — a weight that does not appear in the code but shapes how the code is applied. When engineers who know the disaster only as history design bridges, they apply the code correctly but without the weight. The correctness is technical. The weight is ethical, felt, embodied. Its loss is invisible until the conditions arrive that it would have protected against.
Petroski's insight was that this cycle is not pathological but structural. It is built into the relationship between professional memory and professional practice, because practice is distributed across generations that cannot share direct experience. Each generation inherits the codes of its predecessors but not the catastrophes that produced them. The codes are the profession's long-term memory; the felt urgency is its short-term memory. The long-term memory persists. The short-term memory decays on the generational clock.
The AI era creates a specific new pressure that Petroski did not live to articulate but that his framework makes clear: AI may compress the cycle from thirty years to five. The mechanism is the acceleration of design generation. An optimization algorithm can produce thousands of variants in the time once required to develop one. Each variant that satisfies specified constraints registers, within the optimization framework, as a success. The accumulation of apparent successful precedent accelerates by orders of magnitude. The confidence that accumulates with each success accelerates proportionally. The AI-augmented engineer may, in five years, experience the equivalent of thirty years of pre-AI successful precedent — thousands of designs that worked, each reinforcing the conviction that the tool is reliable, the specifications sufficient, the codes complete.
But the real-world testing of those designs has not accelerated. A bridge designed by AI is still subjected to decades of traffic, weather, and material aging before its hypothesis is fully tested. The confidence outruns the testing. The number of designs that appear successful increases, but the number tested by the full range of conditions they will encounter in service does not. The gap between confidence and testing is the gap in which the next catastrophe lives, and AI has widened it. The defense — the study of failure cases, the cultivation of the felt weight that the codes alone cannot transmit — is the same defense Petroski advocated throughout his career. What has changed is the speed at which the defense must operate, and whether the profession's educational infrastructure can accelerate proportionally.
The observation of the approximate thirty-year cycle was articulated across Petroski's career, with specific documentation in Design Paradigms (1994) and To Engineer Is Human (1985). The pattern was not Petroski's original discovery — structural engineers had noted the recurrence of failure types across generations — but his articulation of the underlying mechanism (generational loss of felt memory) and his detailed case-by-case documentation established the framework that subsequent engineering historians have used.
The cycle is human, not technical. Structures do not age on a thirty-year schedule. The profession's caution ages. When the engineers who witnessed a disaster retire, the weight of the disaster retires with them, and their replacements inherit codes without context.
Success confirms the reduction of margin. Each standing bridge becomes evidence that the margin could be reduced. The profession, collectively, grows more confident. The confidence is the precondition for the next catastrophe.
AI compresses the cycle. Speed of design generation accelerates the accumulation of apparent successful precedent, but speed of real-world testing does not accelerate. The gap between confidence and testing widens, and the gap is where the next catastrophe lives.
The defense is the study of failure. Petroski's consistent prescription was that engineering education must transmit not only codes but cases — the specific, detailed, often painful examination of what happened when someone else's confidence exceeded their understanding. This study deposits the weight that the code alone cannot transmit.