As another example, from yesterday's report: 7 independent VLA papers (vision-language-action models for robotics) dropped within 24 hours from 9 different orgs. Xiaomi, GigaBrain, RISE — all attacking sim-to-real transfer for robotic manipulation, none coordinating. Unless you are monitoring all those channels, you might have missed this, more importantly their solutions and discussions.

I'm most curious if reports like these match what people were already seeing or if it was somewhat difficult to get a full picture of convergence until clustered.