Frontier open-source models have been out for ages and nothing happened, doomers told
tszzl · x · 2026-09-21
In a thread echoing @thetreygoff, tszzl engages with the argument that frontier-class uncensored open-source models have been widely available for a while without the disasters doomers predicted — with the quip that doomers and EAs "only update in one direction." tszzl concedes the point but notes public models still aren't at the hypothetical "February Mythos" tier of capability.
More from AGI Musings
- Krauss and Sabine Hossenfelder to discuss AI wonders and nightmares on Sept 23 podcast — skdh · 2026-09-21
- How long until LG TVs influence your AI agents to sell you more LG gear? — tobowers · 2026-09-21
- Upcoming essay asks why humans are conscious but AI cannot be — AnnaCiaunica · 2026-09-21
- "AI psychosis": overindexing on LLM reach while ignoring context rot and human judgment — bendee983 · 2026-09-21
- Ray Dalio says AI could be both the next big bubble and a real technological revolution — Olivier__OG · 2026-09-21
- Oliver Habryka's team launches AGI.WTF, a Sept 2026 Berkeley conference on making sense of AI — RichardMCNgo · 2026-09-21