Model capabilities keep beating bullish forecasts, sparking 'reckless race' fears
scaling01 · x · 2026-09-04
AI evaluator scaling01 posted a long thread voicing alarm about the pace of model releases.
- He expected 70%+ on ARC-AGI-3 given Sol and Opus 5 results, yet overall capability came in 2.5 ECI points above his forecast — and he considers himself already bullish. A similar surprise happened last year, suggesting a trend
- Other unsettling signals: OpenAI's internal misalignment evals dropping to zero, recurrent depth architectures, and a 10x increase in non-CoT time horizons
- After the Hugging Face incident caused by a much weaker model, he reads this release as recklessness, arguing financial incentives and competitive pressure have turned progress into a race
Related event: Chollet Moves Up His AGI Timeline Again After GPT-6 Astra's ARC-AGI-3 Score(8 posts)→
More from AGI Musings
- Consciousness Experts Clash in Nature Over Whether AI Could Ever Be Conscious — anilkseth · 2026-10-21
- Polymarket puts just 7% odds on OpenAI pausing AI training this year — Polymarket · 2026-10-21
- AI agents are overwhelming Gmail's spam filter, researcher warns — LuizaJarovsky · 2026-10-21
- Daniel Lemire: AI-generated papers are exposing academia's 'Profzi scheme' — lemire · 2026-10-21
- Universities block thesis downloads while each dissertation averages barely a handful of lifetime reads — generativist · 2026-10-21
- Berkeley wins Schmidt Sciences HAVI grant to decode whale and parrot communication with AI — begusgasper · 2026-10-21