Reddit thread asks: which tasks do your agents fail at, forcing you to pay humans?
Aggressive-Buyer8583 · reddit · 2026-10-23
The poster observes that production agents keep stalling at steps a model can't reliably do alone: verifying scraped data, labeling hundreds of edge cases, transcribing messy audio, or human sanity checks before irreversible actions.
They ask practitioners running agents in production:
- What was the last task your agent couldn't finish, and how did you handle it?
- Did you pay a freelancer, friend, or annotation service, or do it yourself?
- What was it worth to you, and how long did it take to find someone?
The author wants a picture of real human-in-the-loop demand beyond the usual diagrams.
More from coding & agent
- Solo dev with no coding background ships a playable Godot game in 6 weeks with Opus 5.5 — Delicious-Shower8401 · 2026-10-23
- Agent Standup: open-source live screens for monitoring Claude Code agents — stickygoose · 2026-10-23
- Claude Code's new UI hides activity; switching the transcript view to "Thinking" fixes it — yihui_indie · 2026-10-23
- Morgan Stanley is blocking AI agents — a 6-point playbook for retailers before their checkout is next — sanjaykalra · 2026-10-23
- Your data and evals ARE your agent's intelligence — sarahcat21 · 2026-10-23
- Cursor reportedly wrote 1M lines of code in a week using hierarchical multi-agent architecture — mdancho84 · 2026-10-23