Nemotron 3 Ultra's 550B params mogged by models 20x smaller, SemiAnalysis blasts NVIDIA's frontier training
teortaxesTex · x · 2026-09-03
SemiAnalysis argues that while NVIDIA Research produces excellent architecture work (LatentMoE used in Kimi K3, GatedDeltaNets in Qwen), NVIDIA's bureaucratic culture undermines end-to-end frontier training: Nemotron 3 Ultra, with 550B total params (55B active), is outperformed by all Chinese models, including Qwen3.8 27B with 20x fewer parameters. teortaxesTex adds that architecture R&D output tends to be anticorrelated with model quality and questions the dogged criticism of Nemotron 3.
More from Companies & People
- Technically Speaking podcast launches with Databricks, WorkOS, Datadog founders for Season 1 — jamesmurdza · 2026-10-21
- GitHub HQ Hosts Open Source AI + Policy Talk with Mozilla Co-founder Mitchell Baker — kevinsxu · 2026-10-21
- Affiniti Pivots Into an AI-Native Operating System for 3,300+ Main Street Businesses — jfiance · 2026-10-21
- CodeRabbit Ships 5,000 PRs a Week With a Tiny Team, Hires by Strategy-Game Hours — omojumiller · 2026-10-21
- AI safety researcher Sören Mindermann moves to Tsinghua, co-founds Beijing AI Safety Hub — dhadfieldmenell · 2026-10-21
- Runway joins Claude Marketplace, letting enterprises spend Anthropic budget on video generation — runwayml · 2026-10-21