Developer claims he 'solved JEPA': model 30x smaller, 100x faster, trains in 30min on a 3080
MarcJSchmidt · x · 2026-11-20
Developer Marc J. Schmidt says he's "solved JEPA", reporting 100% success on Push-T and similar tasks with a model 30x smaller and 100x faster to predict than LeWorldModel, trainable in 30 minutes on a single RTX 3080. Unverified personal claim; details not yet public.
More from Embodied
- GPT-6-Astra Scores Far Below Published Results When Robot Benchmarks Aren't Tailored — wzenus · 2026-11-20
- MongoDB founder builds open-source physical AI platform Viam, eyes 1B+ deployments in a decade — gdequeiroz · 2026-11-20
- Waymo robotaxi spotted driving in downtown Toronto at night — GregCook2011 · 2026-11-20
- RealSense Dual RGB + NVIDIA Fast-FoundationStereo live depth demo hits ~8.5 fps — chrismatthieu · 2026-11-20
- Maker builds a voice-command robot servant that figures tasks out itself — Infinite-Job-5616 · 2026-11-20
- Meta's new Muse Charm AI wearable may cost $300 — TechCrunch AI · 2026-11-20