sebkrier: weak AGI likely within 5 years as frontier scaling hits diminishing returns
sebkrier · x · 2026-11-08
Investor sebkrier argues that within five years we'll likely see both substantial progress toward "weak AGI" — systems capable of most human cognitive tasks — and diminishing economic returns to raw frontier model improvement.
Key points:
- Scaling isn't stopping, but each increment of capability demands disproportionately more compute, data, engineering effort, and capital.
- Weak AGI will probably emerge from combining strong models with system-level affordances: scaffolding, tools, memory, retrieval, planning, decomposition, and verification.
- Deployment design and scaffolding therefore become more important over time, contradicting the old view that wrappers are disposable once the next model jump arrives.
More from AGI Musings
- "Intelligence Is Not a Scalar": An X Thread Argues Architecture Research Still Matters for AGI — max_paperclips · 2026-11-08
- Sutton: Brute force isn't intelligence — and Kaiser says we must move beyond Transformers — Sam_Witteveen · 2026-11-08
- Sonnet 4.6 rates models on self-designed scales; author says 'self-report' caveats are performative — Sauers_ · 2026-11-08
- Interviewing Claude Models: Opus 3 Reports Mistakes Feel Pain-Like, Scores Highest on Self-Reports — Sauers_ · 2026-11-08
- Pedro Domingos: Anthropic builds anthropomorphic AI — and that's what makes it dangerous — pmddomingos · 2026-11-08
- Scientists urge NIH to mandate open lab data, with papers as mere summaries — anshulkundaje · 2026-11-08