Ask HN: What is one simple thing LLMs are insanely bad at?

A post on the technology news aggregator Hacker News has sparked a discussion among developers and AI enthusiasts about the practical limits of large language models. The entry, titled “I am looking for ideas on what to train a specialized model for! What is one simple thing you repeatedly ask ChatGPT, Claude, or another model to do that it still somehow messes up?” was published with a URL linking to the discussion thread (https://news.ycombinator.com/item?id=49443820) and has accumulated 20 points and 30 comments at the time of reporting.

Participants in the thread are sharing recurring tasks that current general‑purpose models struggle with, ranging from precise code generation and consistent formatting to accurate extraction of structured data from unstructured text. The conversation highlights a growing interest in developing niche or domain‑specific models that can address these persistent shortcomings, suggesting that the community sees value in targeted fine‑tuning to improve reliability for routine but error‑prone queries.