I'd temper my views of the apparent recent Chinese robotics progress.
It's more than nothing, but there can still be a big gap between what is presented and the real timeline to follow; see e.g. the US example with Tesla's "Paint It Black" video, what was it, around a decade between the video and the AI getting to the level shown in that video?
My view doesn't need tempering. I talked about progress, which is indisputable, whatever its pace or current limitations might be. And that was the entire point: robotics is not a stagnant field.
What looks good on TV, what impresses a news show, these do not directly map to the hard stuff.
Likewise the other way around, this hand and wrist by itself is more impressive breakthrough than any of what I saw in those two videos, despite being much less flashy and 6 years ago: https://www.youtube.com/watch?v=x4O8pojMF0w
I'm also sure I've seen exactly the same two clips of an AI learning how to control a simulated avatar in Two Minute Papers a dozen times each over the years, before I got tired of the channel mostly becoming "Incremental LLM release! Also Unreal Engine sometimes!"
Note however that this does not mean I'm saying "there is definitely no progress in robotics"! Not at all! I'm saying that the stuff we find easily is the spectacle, not the substance, and that progress with the substance is much much harder to gauge.
It's more than nothing, but there can still be a big gap between what is presented and the real timeline to follow; see e.g. the US example with Tesla's "Paint It Black" video, what was it, around a decade between the video and the AI getting to the level shown in that video?