Agents Will Toxicify the Open Internet; Private High-Trust Networks Win
sull · x · 2026-11-08
Quoting a take that agents will turn the open internet into a toxic wasteland — like humans retreating to private lobbies when cheaters overrun games — @sull adds that real social networks are already private chat groups escaping the cesspool. Enshittification will be fought by better systems, and P2P/private networks like Tailscale are well suited for this coming "new Internet" where the web contracts into high-trust private spaces.
More from AGI Musings
- Matei Zaharia to lecture on accelerating research with AI at Simons Institute — matei_zaharia · 2026-11-08
- Safety factors predate bridge disasters by a century — AI safety analogy corrected — davidmanheim · 2026-11-08
- 'Human tokens' math: human thinking costs 10,000x more than Astra at $500k per million tokens — flngr · 2026-11-08
- RLVR debate: 90% of human work isn't verifiable, so hill climbing won't generalize — yacineMTB · 2026-11-08
- Formal methods researcher decries secret proprietary math datasets for RL training — neurocy · 2026-11-08
- Defenders Can Match 1,000 Attacking AI Agents—and Hold an Information Edge — marktenenholtz · 2026-11-08