NFTs, Generative NFTs, Generative Art Coding

What Went Wrong with the “Isekai Meta” NFT Drop (and Other Generative NFT Drops in the Past)?

A fundamental of generative art coding, actually.

Photo by Jørgen Håland on Unsplash. And the intended metaphor here is that these are two different sheep, but they look pretty much the same. Good for sheep; bad for NFTs!

Today while scrolling through Twitter, I saw an interesting post by Twitter user #HashBastardsNFT mentioning an NFT drop that contained a number of duplicate NFTs. Here was his example photo:

Um, wow, that’s a LOT of dupes. For this many to have come up at all (regardless of the coding error they reportedly had), they must’ve had a bare minimum number of traits. Cute, for sure. But they should have had many more traits, I think.

I usually have to remind myself that not everyone codes generative art for a living. I truly have quite a strange (and enjoyable) job, to be honest. Having dupes in a generative set is an absolutely fundamental issue — especially as almost all generative NFT sets are advertised as having many thousands of “programmatically unique” (or some such similar wording) images.

Here’s how it works, for anyone interested:

1. First, you generate an image based on your rarity table. You select a background, a body, a shirt, a hat, etc.
2. As you select these things, you keep track of what you select. For most coders, that means you create a “DNA” string representing each NFT.