Why do Chinese AI labs keep up at a fraction of the cost? Reddit debates the efficiency gap
budfischer · reddit · 2026-09-30
A Reddit thread asks why Chinese AI labs keep improving while spending only a fraction of what US labs do, floating explanations like better training efficiency, stronger post-training, better use of open research, and superior data strategy.
The poster highlights a recent trend: Chinese labs buying large amounts of handcrafted expert data from US firms such as SurgeAI and Mercor, and argues such data exports should be restricted like GPUs if the US wants to keep its lead.
More from Companies & People
- Modal's Runtime conference lands Oct 1 in SF with Cognition, Anthropic, NVIDIA speakers — charles_irl · 2026-09-30
- UChicago Booth hires Applied AI faculty: IC track, department-funded research — ethayarajh · 2026-09-30
- Yacine recalls interviewing Rajan Nagarwal amid 30-second WiFi drops and a fire alarm — yacinelearning · 2026-09-30
- A masterclass on the one-person AI creative team: two marketers push AI to its limits — alexgoughcooper · 2026-09-30
- GPT-6 naming chaos: Terra became Sol, Sol became Astra, out came GPT-6.1 Sol — Balance- · 2026-09-30
- X User Likens FTX to EA/Rat Orgs' 'Hiring Your Friends' Problem — EigenGender · 2026-09-30