What the Multicoiners are claiming is the equivalent of there being, in their own words, literally millions of different incompatible internets all coexisting, with their own browsers, dev teams ISPs, domains and everything the Internet has. Sorry. That is impossible.
This analogy does not hold completely in a context with standardized lightning cross chain swaps and decentralized exchanges. I mean, I doubt there will be millions, but there it's definitely possible that more than one coin prevails.
If there's added utility (which I doubt) and engineering viability in creating another binding on top, then yes, it could happen. The relevant question is if users will see utility in using coins with different attributes (such as stronger privacy) in a LN-XCAT-DEX context.
I'd argue it would happen up to the point where the marginal comparative utility justifies it. This could be the one-coin-to-rule-them-all point, or lightning to bind them all, or further up. Not to worry, cause the free market will always drive itself towards maximum utility.
This is right; in the end there will be a single layer that binds everything, just like the internet is bound by a single layer. What that layer is isn't the point, that there is one layer, is the point. Bear in mind that all coins are not identical, and all are inferior to BTC.
It is possible to join all coins together, but this will not conjure up developers to make sub-Bitcoin coins viable in and of themselves. This is the error of Multicoinism, thinking that all coins are interchangeable, when they are absolutely not.
Bitcoin is superior to all coins because it has so many people working on it. It will be technically superior going forward, and it will be impossible for any alt-Coin to catch it. This is completely separate from the companies writing software to make BTC's features accessible.