Lab insider explains the internal-external gap on AI progress and Astra scare
MajmudarAdam · x · 2026-09-13
MajmudarAdam argues the past two weeks look like orchestrated regulatory capture from outside, but the real story is a perception gap: capability leaps only come from scaling further on an existing law or discovering a new one (GPT2-4 pre-training scale; o1/o3's test-time compute axis).
Only a few releases felt like real leaps—GPT3, GPT4, o1/o3, DeepSeek R1, Fable/Mythos, Kimi K3, and now Astra. Because marginal releases dominate, "scaling has hit a wall" feels comfortable from outside. Inside labs, researchers read scaling-law plots directly: the question is how many unsaturated scaling axes remain—and Astra abruptly changed that internal picture, explaining why lab employees suddenly got scared.
Related event: OpenAI researcher: labs' internal progress far exceeds outside perception(5 posts)→
More from AGI Musings
- Security researcher: vibe coding will force software to consolidate around mature products — jd_pressman · 2026-10-31
- AI pushes circuit lower bound from 3.1n to 3.5n as UCLA-backed challenge aims at P vs NP — soumitrashukla9 · 2026-10-31
- What is work for? Cross-national survey of 8,234 adults on meaning in the AI age — soumitrashukla9 · 2026-10-31
- Which universities win top students may hinge on healthy AI-use boundaries, argues researcher — brianryhuang · 2026-10-31
- Researchers debate whether more inference compute can substitute for training compute — teortaxesTex · 2026-10-31
- Five agents can agree because they made the same mistake, krishnan warns with satire — krishnan · 2026-10-31