the people that were shitting on the idea that video models learn accurate world models look kinda stupid rn

cypherpunk wannabe, engineer working on @frames /open-frames /opencast
https://stephancill.co.za create stories at https://stories.steer.fun

The video from a few days ago comparing Google and Zora (IIRC) videos of a person cutting a tomato are the only comparison I need. 

Sure, there's room for progress. Especially visible in the tennis/basketball (release glitch or not enough "snap") and running (weird arm mechanics). But still. This is not only "usable", but can be more than "good enough" for most purposes.