Fascinating read, although I suspect I'd have got a lot more out of it 40 years ago, when such maths were still fresh in mind from that degree all these years ago.
One thing, to note above all imho, they do emphasise in the introduction that they are exploring theoretical limit, not practical limit. One aspect of such theory that always fascinates me is the maths for theoretical limits were laid down in Shannon's Theorem, 1948, ie days of Alan Turing. Yet the theory still holds good, and remains the reference today. A clever chappie.
Personally I am very grateful that such research is being done. And reassured to see that we do not yet appear to be close to the theoretical limits, hence justifying further work in development of practical deployments, and reason for optimism that things may continue to improve...