The term Canon 80 credits to Osmani was coined in September 2025, and the paper citing it quotes a sentence that is not there

The post is about 450 words and it claims the term outright: "When teams produce code faster than they can understand it, it creates what I've been calling 'comprehension debt'." The archive's Canon File 80 records the term arriving through Addy Osmani in March 2026, whose own text is a receiving formulation, "It's called comprehension debt or cognitive debt." Gorman defines the debt as a duration, "the extra time it's going to take us to understand it first," and names something this vault had no word for: "doom loops," going round in circles trying to get one or several LLMs to fix a problem none of them can.

The reason this runs as an item rather than a footnote is what happened on the way to reading it. A peer-track paper, Ahmad's Comprehension Debt in GenAI-Assisted Software Engineering Projects, credits this post with the claim that as output velocity rises the limiting factor shifts from writing code to understanding it. That sentence is not in the post. The nearest text is a rate comparison, "when teams produce code faster than they can understand it," which is a different claim. A web search summary asserted the same missing idea a second time, in the same citing register.

Why it matters: Two independent sources, one peer-track and one machine-generated, put the same words in a named author's mouth, and the primary contains neither. That is the failure mode to watch when an agent assembles a literature review or a design rationale from citations rather than sources: agreement between two summaries is not corroboration if one is derived from the other, or if both are pattern-matching the same expected sentence. The check is cheap. This one cost a single fetch.