Are you that sure about that number that you will have all the dangers of a hard fork with no replay protection when it is years, maybe even decades from potential necessity?

I can very well imagine that everyone might use one big Lightning Network in the future and that there will be no need of ever closing it. There is also a cost to closing a channel, so there is an incentive of not paying mining fees that people might not want to ever close it. And people might not worry about the benefits of closing a channel as much as you think, since the only benefit is that there is no risk of your funds being lost. It makes sense to assume that people will not open a channel with a friend, but with a big well known node that has open channels with many other users. That way if that well known node betrays the well established trust starts closing channels, everyone would know about it, since it would be a big deal and maybe even be in the news and that could cost that node a great deal. This node won't be able to close all the channels at one either, since it will be literally impossible to put so many transactions in blocks. Here a bigger block size is actually a problem. You see, it isn't all that simple.

But I can tell you this, the benefits of S2X are none for many years to come and costs and dangers are huge and there is no cost in postponing the fork and there are certainly benefits to do so. This is obvious to anyone who thought about it for a reasonable amount of time, which tells you something about that NYC agreement, whatever you think that something is.