DeepMind researcher: multimodal compute will be priced in world units, not tokens
prateekj · x · 2026-11-11
A Google DeepMind researcher argues tokens are the wrong unit for physical-world AI: a minute of video costs different token counts across Gemini, Qwen and other models, and 3D content like gaussian splats has no token count at all. Buyers of vision compute think in images, video-seconds, scenes and objects—so infra for this intelligence will likely be priced in units of the world rather than units of text.
Related event: Developer Argues Tokens Are the Wrong Unit for Pricing Multimodal Compute(3 posts)→
More from Infra
- Infrathon: 10,000 GPU-hours benchmark shows agents beat human infra baselines — UW · 2026-11-11
- MLPerf at SC25: 12 university teams to run MLPerf Training in 48-hour Student Cluster Competition — TheKanter · 2026-11-11
- MLPerf Training v6.1 results land Nov 16 as MLCommons heads to SC26 in Chicago — TheKanter · 2026-11-11
- The Best Test for Clinical AI: Pull the Network Cable and See if the Visit Keeps Going — MaziyarPanahi · 2026-11-11
- FT traces how a financier couple in a 1920s Shanghai villa built an AI cloud empire on Nvidia ties — zijing_wu · 2026-11-11
- AI Agent Fact-Checks Nvidia's $7.5B China Chip Sales Story in 20 Minutes, Flags One False Ownership Claim — peterjliu · 2026-11-11