That's just multisig, plain and simple .They seem to be using some kind of stealth address scheme to generate the underlying addresses, but the redeemScript itself is just vanilla multisig. I presume it would work just as well with bare multisig.
Well, segwit is a whole different ball game, and requires a change to consensus rules. But again, surely the problem is the existence of witness scripts, not the existence of P2SH. Once you allow witness scripts in scriptPubKey, it's natural to allow them in redeemScript too.
P2SH makes the scripting capabilities of bitcoin far more useable. It allows you to have an address secured by an arbitrary script, without us having to invent an address format that can encode arbitrary scripts, and without someone sending to that address having to pay for a potentially very large trainsaction.
It also has the side effect of allowing you to keep the script secret until the first time you spend from the address - which I think is probably beneficial in some cases, but don't feel strongly about.
That's it. That's all P2SH does. It doesn't change the functionality of the bitcoin scripting language at all - with the minor exception of the secrecy issue. It just makes it rather more convenient to use complex scripts in the real world .
EDIT TO ADD: Deprecating P2SH would mean that the vast majority of institutional cold storage schemes would have to be completely redesigned from the ground up to be useable with future BCH (or more likely BSV). How is that going to help adoption?