DeepSeek

Raised as a usability problem by 13 people on Hacker News over 3 years (2024–2026), in 12 threads.

People raised many different problems with DeepSeek, 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 12 complaints

Confusing model version names make it hard to tell releases apart2026-10-19
Webpage is stuttery and nearly unusable on Android phone2026-10-02
Cannot make it stay concise; ignores preferences despite many workarounds2026-08-02
Responses take 2-3x longer than comparable OpenAI or Anthropic models2026-06-08
Mobile page is very wide, with content squeezed into a tiny column on the left2026-05-24
Animation keeps pushing page content down while the user is reading2026-05-24
Configuration is arcane, poorly documented, and the sandboxing permissions are confusing2026-05-23
LLM transcription errors are hard to spot because text gets hallucinated2025-10-20
API is very slow, sometimes unusable for the user2025-05-12
Losing useful past conversations; endless scrolling to find old chats2025-04-28
Responses are too slow to make it usable2025-01-29
Quantized model runs at 4 tokens/sec, too slow for interactive use2025-01-28