The thing blockchains _can_ be better at than a classic SQL database is tamper-evidence, but that's also possible with less overhead using classic PKI and/or Merkle trees. Using trusted third-parties to sign things could be better if something came to a court or other public form because then you'd be able to have some institutional weight rather than random internet strangers: “My filing record was signed by the USPTO and their signature was notarized by the DPMA before the date when the other party claimed…” sounds better than having to explain to a court why it's unlikely that you were able to suborn miners to favor your transaction.
The problem is that a blockchain can only trustlessly and verifiably encode things that are wholly represented on-chain. As soon as you bridge it to anything in the real world you're relying on some degree of trust, and trustlessness is all-or-nothing.
That's why only proper un-backed-by-anything cryptocurrencies actually make any sense - they're wholly on-chain. On the other hand their value flails around wildly like a balloon you forgot to tie off and that limits their practical value to basically nothing. Their monetary policy is in the hands of some un-elected group of randos accountable only to their own enrichment.
Stablecoins: who knows whether they're redeemable? You just have to trust the issuer and the legal system. They also get frozen all the time - Tether freezes more tokens than anyone and there's zero process and probably zero backing. Even Circle/USDC lied about their backing.
Deeds on the blockchain: not your keys not your house? Ok, no thanks? If you have to rely on the court to have final say then the real world diverges from the chain and of course then why even have a chain?
The iron law of blockchain is "if you think the blockchain is a good solution to any given problem you either don't know enough about the blockchain or you don't know enough about the problem."
It's not clear they're anything more than a technological curiosity that's been coopted by anarchocapitalist libertarian grifters.
Except the legal system doesn't have to go after the blockchain per se, they go after the users who still have to have some way to access the value in it.
Maybe, with absolute perfect OpSec, you can use a "censorship resistant blockchain" to finance your ($politically_unpopular_activity), but it's going to be very difficult to get mainstream users involved in a platform that could become increasingly radioactive.
Eventually it becomes a legal death spiral-- as mainstream opportunities leave the network, so does the legal fig leaf for participation.
People bring this up a lot.
1) In practice, blockchains get censored and edited all the time when mistakes happen.
2) I don't see the advantage vs. a publicly readable database that a distributed set of users can archive in a cryptographically secure way. Again, blockchain's main failure point is that it does what centralized databases (which can be cryptographically secure, some even are!) do, but it takes 200x more effort to do it.