Demos vs. Reality: Coding Agents Stuck in Loops Burning API Budgets
Particular-Swan2404 · reddit · 2026-09-18
A Reddit user calls out the gap between AI influencer demos ("agents build entire SaaS apps while you sleep") and reality: paid coding agents choke on basic tasks.
His own example: running Claude Code in the background to clean up legacy repo docs, it got paralyzed by a simple file check, burned half his API tokens asking itself for approval, then dumped gibberish into logs. Other common failure modes include silent rate limits triggering broken loops, git hook permission deadlocks, and inventing fake directory trees to refactor "ghost code" for 20 minutes.
He asks the community how they handle silent background crashes without blowing their budget on looped API calls.
More from coding & agent
- ChatGPT web can now open GitHub PRs directly, despite outdated docs saying it can't — dotey · 2026-09-18
- Sakana AI Launches Fugu Max, a Multi-Agent Orchestrator Routing Tasks to Leanest Capable Models — tkasasagi · 2026-09-18
- No Java SDK needed: one plain Java 25 file to call TypeSafe's Jev API — therealdanvega · 2026-09-18
- fast-jev-compaction: Claude Code plugin cuts a ~1M-token session to 86K in 1 second — viksit · 2026-09-18
- Claude Code 2.1.276 by the numbers: shipped in under 6 hours, +440 prompt tokens — ClaudeCodeLog · 2026-09-18
- Claude Code 2.1.276 fixes regression breaking all requests behind proxies — ClaudeCodeLog · 2026-09-18