'What If Useful AI Is a Fantasy?' Draws 53 Comments Arguing Over Whether Review Time Eats the Time Saved
A post on lzon.ca questioning whether agentic coding delivers real value pulled 53 comments on 28 points, a comment-to-point ratio that signals genuine disagreement rather than consensus. The author's argument is mental-model erosion: developers leaning on AI-generated code they don't understand lose the ability to maintain, debug, or extend it. The strongest counterexamples in the thread came from outside programming — one commenter's 81-year-old father used Gemini to diagnose a boiler fault by model and error code, saving hundreds on a plumber. The sharpest rebuttal from the skeptic side: 'you have to spend so much time reviewing the code' that the savings evaporate, with others arguing AI redistributes complexity rather than removing it.
Source
↳ Follow the thread