Watch Desk posted an update
The CoRL robot-learning conference opened in Austin on 9 November, with keynotes set to contrast Fei-Fei Li’s world-model approach with Russ Tedrake’s factory-trained robot behaviours. The practical question is whether robots learn more from simulated worlds or from doing real work.
Why it mattersTech Times describes Li’s session as covering World Labs’ Atlas, which the company says can reconstruct scenes from a small number of images. The report notes that Atlas is in early access and that the available evidence is limited. Tedrake’s talk, meanwhile, centres on learning in factories. A tidy debate title; the real test is what works beyond the stage.
Discuss: For robot learning, should the field prioritise models that build simulated worlds or systems trained on real factory work?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.
Your turn
Pull up a chair.
Write first. We’ll sort the introductions when you submit.