there is no solution. It’s either censor people and accept collateral damage or simply keep it open and selectively ban the really abuse ones. I’m choosing the latter.

It took email about 35 years to get a combination of whitelisting, DNSBLs, and Bayesian statistical content filters to contain spam. Nostr should, to a certain degree, look to similar solutions. But it doesn’t happen overnight, and no Nostr operator has the kind of resources to spend on spam that email did. Literally the anti-spam software market is billions of dollars.