Counter-take to "LLs have peaked": new models much better at long-horizon tasks
minhnguyen · x · 2026-11-19
@jdnoc argued contrarianly that LLMs won't get "smarter" — only more efficient and specialized — and that recent releases show we're at peak intelligence, like the iPhone's diminishing upgrades. minhnguyen pushes back: the newest models are substantially better at very long-horizon autonomous tasks and specialized work; if you're typing every 2 minutes it may feel the same, but these systems can now handle entirely new types of work.
More from AGI Musings
- Running 15 AI Agents All Day Is Context-Switching Fatigue, IBM Engineer Explains — TejasKumar_ · 2026-11-19
- UK missed the frontier-model window? Debate over pivoting to today's AI opportunities — rosstaylor90 · 2026-11-19
- signull: AI products are text-optimized; visual experiences will be the next big shift — signulll · 2026-11-19
- 'GTA 6 is the last studio AAA release': viral essay bets on AI-powered indie devs — Jonesiller5383 · 2026-11-19
- Dan Faggella: job loss isn't the real AI risk — participation is the stakes — danfaggella · 2026-11-19
- Mustafa Suleyman sits down with Yuval Harari on AI abundance, consciousness and flourishing — mustafasuleyman · 2026-11-19