Gary Marcus

using a robot hand to rotate a cube and spin the top in a stable configuration is a nice research project, but barely a "challenge." also it didn't work 8 out of the 10 times they tried it on the actual robot.

I don't know, others may, but is it inherently more impressive than the Baoding ball result from the week before? Factoring in the 80% failure rate? And the bluetooth sensors inside the highly souped-up cube?

it’s hard to see any obvious way in which the rubik’s demo is more impressive than the more flexible Google Brain Baoding demo that came out earlier in the month. what do you think @beenwrekt?

if they showed that in some general way that wasn't one (pre-rigged) object, I could maybe sorta squint and see that point. but sim to real with one overlearned prerigged object is a joke. no respectable academic journal would buy that argument.

They do have unseen modifications e.g. finger tying. While it's just one object, I really hope you appreciate how hard this must have been to pull off. While I'm with you that the PR is a bit silly, this is a significant advance.