Actually we have the opposite problem most protocols shy away from solving DNS, and even the ones who did solve it with centralized solution are trying to remove it, and most are just using capability system.
For example Willow identifies the namespace as a pubkey, but don't bother with key rotation for that namespace, there is delegation of course, but if the master key is lost it is over.
Peergos has centralized PKI and they are planning to remove it and rely entirely on capability tokens for the only part that use them ; establishing a friendship connection and grant friends access to some stuff.
So really Mlkut name system, if it works, it will be like a sovereign DNS that you use, possibly along side good old ICANN to rotate the Namespace keys of Willow if lost, and point to a file in Peergos with a short URL.
Put another way, most protocols use identity systems that are cryptographic by nature, but they don't exclude a layer of indirection through DNS to deal with the needs that can't be done with cryptographic primitives... That is exactly where Mlkut Name System would fit.
For example, NIP05 could have used DNS TXT records directly instead of relying on HTTPS, but you would need the clients to treat that as the root of identity.
I don't think it is hard to create clients that do that in Nostr, or in Willow or in Peergos. In fact it is much easier in Willow because you don't have to worry about interop with all other Willow clients and communities, since you don't actually care about interop outside the community.
