Moonlake AI: World Models Need Causality and Objects, Not Pretty Pixels
chrmanning · x · 2026-10-03
Chris Manning, Distinguished Member of Technical Staff at Moonlake AI, argues in his talk "World Models Need Causality, Not Pretty Pixels" that photorealistic video walkthroughs are not world models:
- Pretty pixels aren't a world model: he cites Genie 3 (demoed by Riley Goodside) and Marble — both produce convincing video, but neither encodes what objects are or how you can act on them.
- Objects must be separated from background: Moonlake's simulations split scenes into a background plus foreground objects rather than raw generated video.
- Why code and objects matter: building simulations this way yields causal, actionable structure, which is what you need to extract a robot policy from a simulator.
A clear counterpoint to the "generated video as world model" approach.
More from Embodied
- Robots Need Robot-Friendly Office WiFi, and It's a Bigger Problem Than You Think — broodsugar · 2026-10-03
- Alexandr Wang shows off a ~$50 DIY AI gadget that catalogs his wine cellar — alexandr_wang · 2026-10-03
- XPRIZE offers $10M prize to enable meaningful communication with wild animals — PeterDiamandis · 2026-10-03
- Mobile bin-dumping robot demo ran 12 hours at IROS without a single reset — chris_j_paxton · 2026-10-03
- Airbound's 6.6 lb carbon fiber drone carries 11 lb, outlifting Amazon's 78 lb MK30 — kyliebytes · 2026-10-03
- AI mods made one person 1000x faster; AR app explosion is next, glasses may go mainstream — Dr_Singularity · 2026-10-03