March 23, 2026
AI or DIY? Feelings.exe crashed
I Created My First AI-Assisted Pull Request and I Feel Like a Fraud
Dev uses AI to ship tiny feature, feels fake — internet splits between “tool” and “cheat”
TLDR: A dev used an AI assistant to add ERB code highlighting to Hugo’s Chroma and felt like a fraud, even after the change was accepted. Comments split between “it’s just a tool, ship it” pragmatists and craftsmanship defenders, with jokes about manager vibes and warnings this AI-fueled future is here to stay.
A developer used an AI assistant to add ERB code highlighting to the Chroma highlighter in Hugo and got the change merged—then confessed they felt like a total fraud. Cue the comment-section fireworks. Some readers said the quiet part loud: is AI coding just cheating? Others fired back: if users get value, who cares how the sausage is made.
The hottest clash came when one commenter pushed back on the doom: “They absolutely do care how software is built,” arguing companies value quality—just not always the same way craftspeople do. Pragmatists piled on with the reality check: being a “great programmer” means shipping results, not writing maximal code. Another voice reframed the whole thing: you knew the problem for years, the tool finally made it possible—that’s not fraud, that’s progress.
Meanwhile, the jokes flew. One quipped the dev has discovered the manager life: talking all day (now with chatbots) and still wondering if anything happened. Another warned, darkly comic: wait until your 1,000th AI-assisted commit—it’s coming. The thread turned into a referendum on modern work: fun vs. output, craftsmanship vs. speed, identity vs. KPI. The post may be about one tiny feature, but the vibe is big: welcome to coding in the feelings era.
Key Points
- •The author used Claude Code to create a pull request to Chroma adding ERB syntax highlighting for Hugo.
- •The PR was approved and merged by the project maintainer.
- •The author reports feeling like a fraud and not learning from the AI-assisted effort.
- •AI tooling enabled the contribution despite limited time and capacity to learn the codebase.
- •At work, AI use is expected and tied to performance reviews; the author perceives incentives favoring rapid delivery over craftsmanship.