But quadratically with block size meaning at 16MB blocks or so a 30% miner might still be able to block all nodes permanently.
No. As I explained (did you even read?), parallel validation routes around the quadratic hash time issue.
More importantly, as miners who create blocks exhibiting this quadratic hash time issue have their blocks orphaned, they will be bankrupted. Accordingly, the creation of these blocks will be disincentivized to the point where they just plain won't be built.
For an attacker disrupting the network for a while might pay of via puts or rising altcoins or just by hurting Bitcoin.
No. As I explained (did you even read?), parallel validation routes around the quadratic hash time issue. A Bitcoin network with a significant proportion of miners implementing parallel validation can not be stalled by quadratic hash time.
Also let me remind you of the resource discussion further up. Of course it is relevant to this debate. Why do you oppose the technically sound and sustainable solution? Particularly as it happens to also bring other important benefits?
There was no resource 'discussion' upthread, an inadequate strawman of resource consumption was used to cast aspersions upon parallel validation.
I oppose The SegWit Omnibus Changeset mostly due to considerations other than segwit itself. Namely: the SF nature; the backdoor of versionbit changes; and the centrally-planned magic 4:1 ratio. But mostly because the 1.7x or 2x or whatever capacity increase it ends up being is too little, too late, and we'll just be back at this same argument before the year is up.