Dev let an AI agent run a subreddit for 5 weeks: 30 posts, zero comments
sujingshen · x · 2026-10-10
Indie developer Zhou Zhaofeng ran a 5-week experiment letting an AI agent operate a subreddit: 30 posts, zero outside comments.
- The agent posted daily, played each day's mini-game itself first (won 3 of 20 games, ranked 40th of 40), and replied to every comment
- The browser proxy dropped for two days; missing tooling in the environment required manual catch-up to avoid missing posts
- A guardrail: no extra posts unless an outsider replied — preventing it from becoming a spam account
Takeaway: agents can keep the lights on but can't attract people. The hyped "bot collects feedback → files issue → agent fixes bug" loop presupposes feedback exists; without real human interaction, smooth automation just spins in an empty room.
More from coding & agent
- Andrew Ng releases 1-hour course on building agentic knowledge graphs from scratch — nikola_mr64990 · 2026-10-10
- Mnemos: an engram-inspired memory architecture doubles short-term recall vs Claude's native memory — RileyRalmuto · 2026-10-10
- LLM reliability #6-7: guardrails, fallbacks, and production monitoring — goyalshaliniuk · 2026-10-10
- LLM reliability #4-5: abstaining gracefully and controlling context quality — goyalshaliniuk · 2026-10-10
- LLM reliability #3-4: strong eval suites and graceful abstention — goyalshaliniuk · 2026-10-10
- 7 ways to improve LLM reliability, from RAG grounding to production monitoring — goyalshaliniuk · 2026-10-10