Interesting, I think have seen him argue that 50% earlier (a year back or so). It is a fair point, but I think he ignores the other effects in the ecosystem (such as miners eventually rationally agreeing to an adaptive blocksize).
I wonder what happens in this scenario in the case
@Peter R.'s subchains catch on: The 50% orphaning might not be too damaging in that scenario any more, as the sub-blocks come fast enough.
The most important takeaway from his post should maybe be: Vitalik also expects there to be an equilibrium, even without block reward
