Products · Pattern or everyday thing
AI agents
Raised as a usability problem by 14 people on Hacker News over 3 years (2024–2026), in 15 threads.
People raised many different problems with AI agents, but no single problem has yet been raised by 3 people in 3 different months and threads.
Unreviewed issues were grouped by machine: Claude Haiku labelled each comment as a usability complaint, named what it's about and summed up the problem in a line, and a small model grouped similar problems. Nobody has checked them by hand. Expect mistakes.
The newest 15 complaints
- Scheduling only offers a limited dropdown instead of a flexible repeat configuration2026-09-30
- Constant approval prompts from the agent get ignored, so users approve dangerous actions without checking2026-08-06
- Having to click through extra steps to pick particular tools2026-07-31
- Navigating via a slow API is far slower than screen reader use2026-07-16
- Constant full-screen redraws and spinners cause flicker and headaches while idle2026-06-24
- Security modes are binary, forcing users to approve every command or disable checks2026-03-28
- Agents ask permission, confirm each other, and edit the same code when no builder is assigned2026-03-12
- AI agents fail most tasks on websites, unreliable and slow when operating UIs2026-02-23
- Skills silently fail to load with no warning or guidance when they don't work2026-02-06
- Users easily make mistakes answering permission questions2026-01-09
- AI autofill mangles addresses and card fields on dropdowns or odd formats2025-12-17
- Yellow chat button mistaken for the text input submit on mobile2025-11-30
- Scroll wheel scrolling causes pain; scrollbar and touchpad work fine2025-07-30
- Can't tell which prompt elements cause the agent's behavior, making debugging nearly impossible2024-12-11
- Complex combined requests succeed only about 25% of the time, confusing users' expectations2024-04-22