Scaling01 on Astra: no AGI vibes, lacks taste, doubts 10T params, guesses 6-8T
scaling01 · x · 2026-09-07
- KOL scaling01, after burning through his Pro limits on Astra, says it doesn't feel like AGI: the model lacks conviction and taste, wasting time and tokens.
- He discounts rumors of 10T+ parameters — if true, scaling laws would be "cooked" — and guesses 6-8T instead.
- He also mocks the idea of a 100-day pre-training run, arguing it would waste 100 days of algorithmic progress, and suggests short iterations may be the base case for GPT-6. His verdict: Astra is just a slightly larger code monkey — the best available, but not AGI.
Related event: KOL Tests Astra, Sees No AGI; Speculates GPT-6 Uses Up to 5x Compute(3 posts)→
More from AGI Musings
- AI accusations shake literary prizes from Caine to Goncourt: who decides what's human? — TuhinChakr · 2026-10-26
- Tech discovers wargaming: applying to be Anthropic's tabletop exercise lead — round · 2026-10-26
- RSI is the culmination of the industrial revolution: a factory building a faster factory — tobias_rees · 2026-10-26
- NYT roundtable asks five experts what the rules for AI companies should actually be — ShakeelHashim · 2026-10-26
- Anti-AI content is algorithm gold and very profitable, quips commentator — annetgriffin · 2026-10-26
- Silicon Valley breakdown: five Stanford-area giants worth ~$18T — FinanceYF5 · 2026-10-26