Hacker News Reader

Comment by anonporridge | Hacker News Reader

Moxie's recent post is a fantastic illustration of how NFTs as they are currently implemented are bullshit, https://twitter.com/moxie/status/1448066579611234305

I'm not sure I'd go so far to call them an outright scam, but they're highly fragile assets that likely won't stand the test of time.

I would be interested to see if someone is working on implementing NFTs in a way that I can feel confident that the digital art will remain provably attached to the NFT for decades and centuries in future.

View on HN

From this tweet, it sounds like NFT's are not much more reliable than html <a> tags? If the host is down, the NFT is nothing more than just a pointer to a 404?

Some NFTs point at images hosted in a traditional, centralized fashion. Others point to IPFS hashes which are basically permanent.

Pretty much.

And worse, the host of the image could theoretically change it on you and you'd have no way to prove (at least in a trustless way) what the original image was.

Yup. Here's a tool for checking most ETH based NFTs: https://checkmynft.com/?address=0x5c61afa47570ab2b562606fa57...

it depends on how it's implemented. its much harder to implement onchain so a lot of projects will do things like <a> tag which is not trustworthy long term like moxie said.

there are well respected projects that are fully onchain such as cryptopunks and autoglyphs

Yeah a huge portion of NFTs on Ethereum are like this. Hicetnunc does it right where by default the NFT is minted pointing at an IPFS hash.

IPFS would definitely be better, but there's no guarantee that the data will be stored on IPFS forever.

I'm thinking something more like a hash of the artwork embedded in the NFT directly rather than just a link to it. That way, even if the link breaks, you can still prove that a piece of digital art belongs to the NFT.

As long as you've saved a copy of the art (the exact copy) you can restore it to ipfs even if no one else pins it.

Ah, maybe I need to spend some time digging into IPFS.

Is it correct to say that an IPFS address can only point to a specific piece of data, and re-uploading that data will generate the same address?

Yup! IPFS address is a hash of the contents.

Nice! I've got to find time to setup an IPFS node.

It's very quick, and there's also docker images that you can use (handy for projects that use IPFS).

> Is it correct to say that an IPFS address can only point to a specific piece of data, and re-uploading that data will generate the same address?

Technically I believe the answers to those two questions are, respectively, “no (because hash collisions are possible)” and “yes (because the same content will always have the same hash)”.

Yup it's possible that someone some time in the future can manufacture hash collisions and poison any IPFS link they want.

Is this the guy who put crypto in Signal?

Reply on news.ycombinator.com