Most of those attacks, as the paper points out, are fixed. The one that is still relevant is the issues around sampling decoys vs real spends. At the time, with high probability you could assume the most recent tx in the set of decoys was the real source of funds because money is spent frequently. Monero improved their sampling algorithm, but it's still an arms race. The existence of the problem at all, whether or not its reasonably dealt with, should tell you monero isn't cryptographicly strong privacy, its obfuscation. This isn't a Zcash vs Monero thing, people should do their own comparisons. But that has to start with facts.

How much non-default "churning" should a Monero user manually choose for maximum privacy?