I think I have LLM burnout
Points and comments are a snapshot, not live.
Daily LLM use has led a developer to dread reading AI-generated text due to repetitive patterns and errors.
For about a year, the author has used LLMs extensively for coding and queries, shifting from writing code to describing designs and reviewing AI output.
While feeling more productive and exposed to new approaches, the author has grown weary of false assumptions, hallucinations, and stylistic tics like emphatic fragments and excessive emojis. The repetition of the same writing style and mistake types is the root cause. The author does not plan to stop using LLMs but finds the constant aggravation wearing.
What commenters are saying
Commenters share the burnout but offer counterpoints: one finds LLMs enabling pet projects previously out of reach due to research barriers.
Two camps emerge: those who see coding joy shifting from craft to higher abstraction, versus those lamenting that LLMs replace interesting debugging with hack-stacking against unpredictable output. Some suggest stylistic tweaks (e.g., instructing LLMs to answer like cavemen) to reduce reading fatigue. A few commenters are considering leaving programming because the problems have changed from interesting puzzles to managing LLM quirks.