[cwkoss]

There are several major issues with NFTs but I don't think this article identifies any of them:

- NFTs are not durable. All of the implementations I've seen are trading "ownership" of URLs that rely on DNS and WHOIS. URL is nice for convenience feature, but they should at least include file hashes, ideally multiple perceptual hashes as well. If there is a central point of failure in a web service, the benefits of decentralization are null. If OpenSea goes down, your NFT is effectively gone: so what is the effective distinction from OpenSea hosting a centralized database?

- Scarcity is entirely illusory across ecosystems. A given NFT within one ecosystem may be scarce, but there is nothing preventing duplicate NFTs of the exact same content (or minor pixel changes). That's why perceptual and file hashes would be useful: currently there is little to stop someone from taking popular NFTs on one ecosystem and selling the same images on another ecosystem. Ecosystems have a financial incentive to ignore enforcing duplicated content from other ecosystems: it makes them more money and hurts their competitors at the cost of possible minor reputation risk. Each NFT ecosystem is effectively a separate competing namespace system. Imagine if opensea://google.com and rarible://google.com were entirely independent and routed to servers controlled by competing organizations: chaos.

- There is no validation of copyright upon NFT creation. I can find someone else's art, make an NFT and sell it. NFT ecosystems need to evolve into becoming more like auction houses: their brand's value should be based on the trust that the art sold is not counterfeit, and are being sold by the true owner.

- NFTs have not succeeded in a test within the courts (afaik, lmk if you have an example of this). NFTs grant no legal rights to the holder. If NFTs granted the exclusive right to license that content AND courts enforced that right, they could be immensely valuable. There is no unified "IP Registry" that I can find a picture/audio/video on the internet and look up who the current owner is to contact for licensing. This is the killer app for NFTs IMO, but whether courts will play along with code-as-law has not been legally validated yet.

Fortunately for NFTs, non of these are technically insurmountable. Maybe there are some smaller ecosystems that have better designs, but it seems the big players are all taking the easy path and building a fragile foundation for ownership.

[iskander]

>NFTs are not durable

Some of the more prized NFTs are entirely on-chain; I agree that ones which just encode a URL of an image feel very superficial and brittle.

>Scarcity is entirely illusory across ecosystems

Every successful ETH NFT has been copied (probably multiple times) on other chains; this isn't really any different from just even copying on the same chain. In all cases it's treated as a copy.

>There is no validation of copyright upon NFT creation

I agree and I think this is a huge problem for 1/1s!

> NFTs have not succeeded in a test within the courts

Best NFTs are CC0 licensed and don't confer any legal rights (but instead encourage liberal reuse and remixing)

[cwkoss]

> Some of the more prized NFTs are entirely on-chain

Oh neat - that's definitely better, I'm sure it increases costs significantly though.

> "treated as a copy."

Are consumers able to easily make informed decisions about which versions of NFT are primal and which are derivative? Seems like a hard problem that would require a lot of expensive moderation by the ecosystems.

> Best NFTs are CC0 licensed and don't confer any legal rights

I love me some CC0, but I'd also love a system where I can self service:

- Identify the owner of a song I'd like to use as background music in a video

- Buy an NFT that represents a license to use that song in a single streaming video

- "Point" that NFT at a file hash of the completed video

- Upload video to youtube and other streaming platforms including a reference to the NFT, making it immune to arbitrary takedown requests.

There are stories of creators who own the rights to a song getting takedown requests on a video that includes it and getting stuck in kafkaesque bureaucracy despite having a valid license. There is no public registry of who is licensed to use a piece of content, so Youtube would have to manually inspect contracts to resolve this - it seems they are grossly understaffed in support to be able to do this. Sure, this is partially due to Youtube's extreme subservience to the IP lobby, but a public system to validate licensing would be extremely valuable IMO.

[RyanCavanaugh]

If you have a "unified 'IP Registry'", as you (I think correctly) posit would be useful, why does it have to be done with a blockchain?

Blockchains are useful for decentralized decisions, but decentralization is antithetical to the kind of thing you would need in order to take an IP claim to court. If you and another person with alleged IP claims to the same work but different registries show up to court, the court has no way to determine which registrar is authoritative for that work, and will ignore both of you. IP without government enforcement is literally useless, so your IP rights are nothing.

This is why property records, patents, etc. aren't decentralized. You and another person can't both have a legitimate claim to the same patent based on two different patent decentralized blockchain records, as cool as that concept might sound technically.

And once a system is unified, it doesn't need to be a blockchain. You can just have a simple database.

[cwkoss]

In terms of "unified IP registry' I'm not talking about a single centralized organisation, but a super-ecosystem of multiple sub- NFT ecosystems:

- Each do due diligence of creator ownership before an NFT can be sold on their platform.

- In addition, they test the candidate-NFT against the file and perceptual hashes on other trusted peer ecosystems

- Due diligence signatures are portable and transferrable, so ex. Sotheby's validation could be used as evidence when token is resold within OpenSea ecosystem - even if Sotheby's was to hypothetically disappear (assuming old signatures are still trustworthy).

(caveat: Perceptual hashes are starting to push past my technical expertise, not sure to what extent perceptual hashes would be economically feasible: can they be specific enough to not cause excessive false positives while still identifying true positive subtle image manipulations without drastically increasing on-chain storage costs?)

I think the benefits of decentralization would be nice, but it's certainly arguable that an ICANN-like international centralized consortium could be a better solution - it certainly can be more energy and cost efficient but at the expense of that entity being a single point of failure.