Earlier quoted context omitted.
Note that he wasn’t saying that it was identical to a classic database but rather that the applications being built were treating it like one. I’ve seen a number of hypotheticals batted around and it was clear that most of them could have been a database, perhaps with signatures on certain records, because they didn’t need any of the anonymous aspects. That could be a sign that developers don’t know how to architect…
Imho blockchain == distributed consensus in a low trust environment. When you are talking about a database and managing it you are by default outside of the low trust zone (or there exists an authority that everyone trusts)
This distinction matters because almost no real world situation is like that: people need to know that what they’re buying is what they were promised, sellers want to avoid doing business with anyone who fraudulently claims damages, etc. If you need to add reputations and/or trusted third parties, you get no value from a blockchain to make up for all of the negatives unavoidably added by using one.