AI researcher in China: companies won't slow down, chip embargo won't stop them
Imaginary_Music4768 · reddit · 2026-09-13
An AI researcher based in China argues the country won't voluntarily slow AI progress while still catching up: companies are commercially driven and highly competitive, distillation offsets compute gaps, and rapidly expanding in-house chips now let players like Moonshot train trillion-parameter models. State-level US-China pacing agreements are plausible, he says, but company-level cooperation with Anthropic is unlikely due to widespread distrust, and any deals won't slow much.
More from AGI Musings
- Critics question EA-aligned oversight of Anthropic: "not independent if ideologically aligned" — beffjezos · 2026-09-13
- Prime Intellect team: open-source camp overreacting to frontier AI slowdown calls — _sholtodouglas · 2026-09-13
- Rauchh: AI Safety Push Risks Making US 'Self-Inflicted Obsolete' While Adversaries Speed Ahead — beffjezos · 2026-09-13
- Beff Jezos Slams Anthropic's Safety Push as 'Unprecedented Soft Power Consolidation' — beffjezos · 2026-09-13
- AI isn't dangerous, regulatory moats are: the anti-safety-narrative case — tawnniee · 2026-09-13
- X User Suspects Organized AI Slowdown Panic, Points to Anthropic and EA Crowd — haider1 · 2026-09-13