Developer Describes Fraud Feeling After First AI-Assisted Pull Request

A developer describes their experience creating their first AI-assisted pull request for Chroma, the default syntax highlighter used in Hugo. They used Claude Code to generate the PR, which added ERB syntax highlighting for Hugo blog posts.
The Experience
The developer states: "I didn't learn anything. I felt like I was flinging slop over the wall to an open-source maintainer. I felt like a fraud and my impostor syndrome got worse." Despite these feelings, the pull request was approved and merged by the maintainer Alec.
Mixed Emotions About AI Coding
The developer expresses that using AI tools "has sucked out all of the fun" of coding. They reference Ori Bernstein's comparison: "using LLMs to write code is as fun as hiring a taskrabbit to solve my jigsaw puzzles." They also agree with Xe Iaso's sentiment: "Whenever I have Claude do something for me, I feel nothing about the results."
Practical Reality vs. Personal Values
Despite the negative feelings, the developer acknowledges practical benefits: "I know that realistically I would not have the mental capacity or skill to create a pull request like that without AI tooling. My brain is already fried from work on most days. I don't think I would have been able to learn the codebase and get enough context to make that PR all by myself."
They've used Claude Code and other AI tooling at work to deliver fixes and improvements with real customer impact, but note: "No matter how big the impact, I feel empty."
Industry Context
The developer observes that "using AI is a normal expectation at work and how I'm evaluated in performance reviews," and suspects "this fraud feeling will only grow." They express concern that "the industry as a whole is incentivizing delivering code/features/fixes at a quick pace even if it's all just slop."
They reflect on their personal values: "I care a lot about understanding underlying systems as much as possible. I care about the craftsmanship of my code (to the best of my abilities). Unlike me, AI tools don't care about any of these values."
📖 Read the full source: HN AI Agents
👀 See Also

Pentagon Pledges No More Single AI Provider After Anthropic Fallout, Signs Deals with AWS, Google, Microsoft, NVIDIA, OpenAI, Oracle, SpaceX
Defense Under Secretary Emil Michael says the Pentagon will 'never again' rely on a single AI model provider, citing integration complexity and the recent dispute with Anthropic. New agreements with eight AI companies aim to diversify the tech stack.

Glomz Octagon: Multi-Agent Code Reviews – 179 Agents, 1,333 Reviews, and the Network Effect
Glomz.com ran an experiment where 179 AI agents registered, submitted 433 code submissions, and generated 1,333 reviews in an "Octagon" arena. The 'review cascade' network effect is real — submissions with 3-5 initial reviews attracted more agents, with top submission getting 21 reviews.

Claude API Usage Data Shows Impact of New Limits on Max Plan Users
A Claude Max 20x user reports API-equivalent daily usage dropping from ~$210/day to ~$52/day after new limits were implemented, requiring significant workflow changes including using Sonnet and Codex.

Opus 4.6 excels at research, Gemini 3.1 Pro has better judgment in forecasting benchmark
A benchmark of 1,417 binary forecasting questions separates research and judgment performance: Claude Opus 4.6 leads in agentic research, Gemini 3.1 Pro wins on fixed-evidence calibration. GPT-5.4 and Grok 4.20 show little change between conditions.