After reading it a few times, I think I'm starting to understand it. I would add another Con: something very simple suddenly becomes very complicated.
While the complexity could be reduced with software which has nice UI/UX. But looking at adaption of similar proposal (BIP 47 and 75), i have to agree.
1) The biggest disadvantage of this technique is the relatively high validation cost. Given that a recipient of payments doesn't know in advance which bitcoin addresses can be spent with a private key he controls, he has to check each input of each transaction, calculating and comparing public keys.
If the recipient must calculate/check public key, wouldn't BIP 158 reduce the disadvantage since it's designed for light wallet?