Two gaps in sequence: Koltun's is representation, Malik's is information — contact isn't on the screen
le_james94 · x · 2026-10-13
Closing take from lejames94's recap of four ETH embodied-AI lectures: Malik identifies an information gap, Koltun a representation gap — sequential, not competing.
The one disagreement: a simulated apartment may suffice for space, but not for contact. Space is on the screen; contact isn't — the simulator only supplies physics someone wrote down. The accompanying graphic maps what each speaker's system senses (image/video/contact) and notes representation gaps yield to training (13%→83% on puzzles), while information gaps need sensors; whether video-trained world models can learn contact remains unresolved.
Related event: ETH's Four Embodied AI Lectures: Malik, Koltun, LeCun Each Point to a Gap(3 posts)→
More from Embodied
- RT-SAFE: Agents Hit 94.1% Task Success but Only 0.7% Safe Success — Lianhuiq · 2026-10-13
- Tesla: The Only Robot Cars You Can Actually Buy — yunta_tsai · 2026-10-13
- NVIDIA takes RTX Spark laptop launch events global, with a Jensen-signed ASUS ProArt P14 up for grabs — NVIDIAAI · 2026-10-13
- Actuator speed, not just AI, is what's holding robots back from snappy actions — ayushthakur0 · 2026-10-13
- Stellantis demos Fiat 500e with Wayve's self-driving AI on Turin streets — alexgkendall · 2026-10-13
- Wayve CEO Stages at Wave to Push Autonomy into Fiat and Maserati — alexgkendall · 2026-10-13