So paintings just “lucked out” and already had a convenient scarce item that interested parties could agree represents the “original act of creation of the work of art,” because the technology to create an exact copy doesn’t exist, and it’s extremely expensive and difficult to create a convincing forgery.
In a similar way an NFT has value based on particular details about the NFT such as the author and database it is in, not just the bytes themselves. You can certainly create the same NFT with a new author, or on a new database, or etc. but it's no longer the same NFT at that point, it's a replica and can be distinguished from the original. And while an NFT published by the author of the art might be a interesting thing to own/control, the same NFT published by a random John Doe doesn't carry the same relevance. And fundamentally you still can't change the properties of the original NFT in the original database.
I think you missed the point, someone has issued the certificate and it only matters if you value that person's or entity's opinion. If you don't then it's just a meaningless piece of paper.
For example, PSA grades playing cards. A rare card that is graded highly by PSA will generally make it much more valuable than an ungraded version of the same card. However, if someone send me a card in the mail and I grade it a 10 and send it back, that means nothing because nobody cares about my opinion on cards.
Similarity, if PSA were to issue NFTs, they would have value specifically because it was PSA who authored the NFTs. I could take the exact same data and author an NFT myself, but nobody would care because they specifically want the one that was created by PSA, and that is something that is verifiable and that I cannot duplicate. They don't want it for any real tangible reason, they want it so they can say "I own an NFT from PSA".
Well it has something written on it, but yes pretty much. If the "somebody" is someone you care about though, then that can make it more than just a piece of paper, in the same way that if you care about the issuer of a certificate then it is more than just a piece of paper. I get the distinction you're trying to make, and I don't think it really matters, but perhaps it was a bad example and autographs would have been a better one.
To be clear, I don't own NFTs, I get the idea but they don't appeal to me. But I also don't own any expensive art either for the same reason, so to each his own. I do own a bunch of trading cards, and I'm sure plenty of people would just view them as pieces of paper. Point being, people value plenty of silly things, NFTs are just one of them.
Generally the content is hosted in IPFS and identified by a content hash which was recorded as part of the transaction which minted the NFT in a blockchain, which makes it somewhere between "difficult" and "impossible" for any third party (or first or second party, for that matter) to replace the content.
An NFT is basically like a 'certificate of authenticity' - whether you think it's BS or not depends entirely on who issued it. A item that includes such a certificate from a known reputable source may be extremely valuable to a collector, where-as the regular item may be practically worthless, and that can be true even if nobody can tell the items apart. At the same time, John Doe down the street signing 'certificates of authenticity' doesn't provide any value because nobody cares who John Doe is, in the same way that some random person signing an NFT doesn't automatically make it worth anything since anybody else can just come along an do the same. But if the author of an artwork came by and produced an NFT of it, that's unique in a way that people may care about.