Some years ago, i bought a 'lifetime membership' to some media editing software. It was a new company, relatively unknown, and was therefore offering the lifetime membership for $15.I opened it up recently, which i rarely do now, and noticed that it is offering A.I. image generation.
I started screwing around with it, and as I did, I noticed a strange repeating artifact in some of the images. Odd lines that seemed out of place.
I figured out that is was the A.I. simulating all of the artists signatures which has been present in all of the scraped art used to train it. There was another strange warping artifact sometimes in the corners, and I realized that was the A.I. replicating the natural warping of a page in a book. Which means it was also trained on scanned images from art books.
I determined something from all of this...
Corporate resale A.I. services are based in Piracy.
Piracy as defined by them. The unautorized use of their images or software, is what they call piracy.
And yet, the entire foundation of their new goldrush product is built on Piracy. By their own standards, they stole all of the data they are using.
So I am going to predict that in the not too distant future, some sort of landmark lawsuit will happen which determines if these large corporations have any right to claim their data is proprietary, as it was built by them scraping text, images, video, code, and data, from us, without our permission. And therefore, belongs to everyone.
atyh /
npub1p3…57qpc
2023-05-16 00:42:06
Author Public Key
npub1p3xpg98qzz7qn6sstcr49qcuwedmhqnt792z5rqqq92wkhn2etuqj57qpcPublished at
2023-05-16 00:42:06Event JSON
{
"id": "951939c4396c55c50cb38f81935ab312823ccb5dc907f102fa5884303410ba12",
"pubkey": "0c4c1414e010bc09ea105e0752831c765bbb826bf1542a0c000154eb5e6acaf8",
"created_at": 1684197726,
"kind": 1,
"tags": [
[
"mostr",
"https://h.atyh.net/u/atyh/h/Mw3k6yfQg4KL4GbYST"
]
],
"content": "Some years ago, i bought a 'lifetime membership' to some media editing software. It was a new company, relatively unknown, and was therefore offering the lifetime membership for $15.I opened it up recently, which i rarely do now, and noticed that it is offering A.I. image generation.\nI started screwing around with it, and as I did, I noticed a strange repeating artifact in some of the images. Odd lines that seemed out of place.\nI figured out that is was the A.I. simulating all of the artists signatures which has been present in all of the scraped art used to train it. There was another strange warping artifact sometimes in the corners, and I realized that was the A.I. replicating the natural warping of a page in a book. Which means it was also trained on scanned images from art books.\nI determined something from all of this...\nCorporate resale A.I. services are based in Piracy.\nPiracy as defined by them. The unautorized use of their images or software, is what they call piracy.\nAnd yet, the entire foundation of their new goldrush product is built on Piracy. By their own standards, they stole all of the data they are using.\nSo I am going to predict that in the not too distant future, some sort of landmark lawsuit will happen which determines if these large corporations have any right to claim their data is proprietary, as it was built by them scraping text, images, video, code, and data, from us, without our permission. And therefore, belongs to everyone.\n",
"sig": "d6286d3fe3fc0066107039b9f75bb0a86c345b90dd9872b2aa7f28288de23db660cc532481d96ff3ca415889016f023a1b61391a5b19544e3a481e24889125c6"
}