I think this is inherently a bad idea, especially coming from one of the core developers of bitcoin.
Throwing away part of the network just to appeal to random nerdery is a horrible idea to implement when in the end, the core developers should really worry about using the resources of the system as efficiently as possible, after all the infrastructural components of the system are the basis of everything.
People thought ipv4 was sufficient once, people thought 512kb ram were sufficient, etc. Being wasteful with a core resource in IT, given its short but rich history, is just wrong on so many fronts, i wouldnt even know where to start.
it's hard, intuitively, to reason about numbers as large as the keyspace of the implicated private keys. but resource inefficiency isn't even plausibly a concern in this space. it's not at all like ip addresses.
maybe a comparison will help: if the world population grew by a factor of a thousand and then every person lived for a thousand years and generated a thousand keys per second while they lived, there would still be no issue with 'wasting' keys.
if that's still counterintuitive, you can think about it like this: if this idea were a problem merely because a developer of the bitcoin client encourages it, then bitcoin would not be secure, as the idea does not need the support of any bitcoin developers. i can generate and throw away thousands (i still cannot do billions like bytecoin because my maths are rusty) of keys a second, and nobody can stop me from doing that. if i
(or even ten thousand people acting in the same way) could compromise addresses that way, then bitcoin would be worthless from the perspective of financial security.