Vincent Sitzmann: traditional computer vision is about to go away
vincesitzmann · x · 2026-10-22
MIT professor Vincent Sitzmann publishes an essay arguing that computer vision as we know it is ending: mapping images to intermediate representations (segmentation, 3D, tracking) is a historical artifact, vision is only well-defined within perception-action loops, and the boundaries between vision, robot learning, and control will dissolve. He predicts 3D representations will fade from embodied intelligence, and adds that scaling up video models will carry these algorithms into the real world since it is fundamentally just generative modeling.
More from Research
- AI will multiply text 50x, making 'judge each paper on its merits' impossible — paulnovosad · 2026-10-22
- Veteran reviewer: 'safe' cockroach papers get accepted, reviewers should reward flawed insight — plopesresearch · 2026-10-22
- Hyperparameter Tuning Is Itself a Machine-Learning Problem: Grid vs Random vs Bayesian — bravo_abad · 2026-10-22
- Michael Levin polls the public: how do physical, living, and minds overlap? — drmichaellevin · 2026-10-22
- ICLR reviewer: 2 of 3 assigned papers are complete AI slop, worst batch ever seen — AkariAsai · 2026-10-22
- Training-free 20B model Viveki generates real-time video heatmaps under 150ms — not-ekalabya · 2026-10-22