Using real robots is more convincing. It's easy to hide a simplifying assumption in a simulation. We say simulations are doomed to succeed. If these guys used

Using real robots is more convincing. It's easy to hide a simplifying assumption in a simulation. We say simulations are doomed to succeed. If these guys used 20 or 100 robots instead of 1000, it would not have made the news. Plenty of similar work has been done with small numbers of robots. However, This is still novel and a decent contribution. Pushing for scale has its own problems. Also, I have no problem with a bit of glamour and flash: it gets people interested and makes the sponsors happy. Bona fides: This was my area for a while.