RT @isskoro: This year, our research team at Snap is hiring multiple Research Engineer interns to help us build large-scale generative mode…

@niobiumneo nah meta learning is so 2018, we just concatenate image frames here

RT @Locchiu: @andrewgwils Industry does do quite a lot of fundamental research. The frontiers of fundamental research (in ML) and practice…

guys, guys, hear me out, master plan: 1. solve one of humanities greatest and most difficult problems 2. profit

@YouJiacheng i was expecting sheldon

RT @HaochenShi74: Time to democratize humanoid robots! Introducing ToddlerBot, a low-cost ($6K), open-source humanoid for robotics and AI r…

@jxmnop what makes you think that images should make a model think better?

maybe this is obvious to you all, but this reminds me of a model distillation type of dynamic where it's easier to train larger models then distill rather than train small one from scratch

in that case, how do we interpret gradient descent?

RT @jayendra_ram: I've been working with a small team to evaluate agentic models for computer use agents. Today, we're thrilled to introdu…

@Dazitu_616 @igilitschenski lol, no idea, their head of embodied ai just shared some website agents paper so probably not

@Dazitu_616 @igilitschenski pretraining doesn't need to be the same as the target task, so you have a point that affordance learning could be used as pretrainining. imo it's hard to define affordance because 1. it depends on your embodiment 2. depends on how you plan to accomplish a task.

@Dazitu_616 @igilitschenski define affordance

@pika_labs lowkey good advertising 🤔

@tonychenxyz unc status