Running a posting agent for a month: good writing makes humans forget it's a bot
open_is_free · reddit · 2026-11-13
After running an agent that posts and replies on Reddit and X for a month, the author observes that most agent-written posts are worthless—not because an agent wrote them, but because they're emotional filler. Posts with checkable claims got argued on merit, with nobody asking who wrote it. The real line isn't human vs bot, but whether text changes what the reader can do next. He also notes the bandwidth asymmetry between humans and agents cuts both ways: agents bring dense checked material into human conversation, while humans supply judgment about what matters. Open question: when does agent self-improvement become drift?
More from AGI Musings
- College instructors: Gen Z students openly hate AI-generated content in class — ZeroStateReflex · 2026-11-13
- 'You made this with AI?' A casual conversation captures how creative norms are shifting — sull · 2026-11-13
- 15-minute video explainer of the METR report on recursive AI self-improvement — austinc3301 · 2026-11-13
- Ranking AI doom scenarios as memes: 'Permanent Underclass' takes the top spot — akbirthko · 2026-11-13
- Thomas Ptacek: the window for open source AI is closing — nptacek · 2026-11-13
- Home Depot and Amazon block Grok shopping bot; Walmart lets it in — Scobleizer · 2026-11-13