Urgent.News

What's breaking now, across thousands of outlets.

AI

AI didn't make me a worse coder. It made me a worse reviewer.

I built a lot of apps last year. I approved most of the code and commits that went into them, and a good portion of it I could not explain. Not "couldn't explain to a junior." Couldn't explain to myself . The tests were green, the diff looked like something I would have written, and I hit approve. If you had stopped me forty minutes later and asked why that function needed a lock, I would have…

The headline reads: "AI didn't make me a worse coder. It made me a worse reviewer."

The author built numerous applications last year but struggled to explain certain parts of the code, even to themselves. They relied on AI tools to approve most of the commits and tests were green, making it difficult for them to justify their approval. The author's own projects served as the sole quality gate without a second reviewer.

This issue is not unique; many developers have experienced a decline in self-confidence and cognitive skills after using AI tools. A 2025 Stack Overflow survey found that one in five developers reported becoming less confident in their problem-solving abilities due to AI.

The author found an explanation in aviation research, where pilots trained on manual flight but then flew highly automated aircraft experienced a decline in cognitive skills, such as maintaining awareness and recognizing failures. The cognitive skills that decay faster than physical ones are more affected by AI, as AI mimics cognitive work rather than mechanical work.

The author proposes that the review process, which has become a habit, produces false confidence in the model's work, leading to a lack of self-assessment. The illusion of explanatory depth, a cognitive science concept, highlights how people overestimate their understanding of complex systems.

To address this issue, the author suggests a two-minute exercise: reviewing the last commit or PR written by AI, writing down its purpose, and identifying where the code would break if deleted. This exercise aims to reveal the illusion of understanding and highlight the cognitive decay caused by AI.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

More from Friday 21 August →