npub1ehyhh99y4ct4pr9xplnpffst6czak79pgzswa2uaeemckdp4ee3q63ndnc (npub1ehy…ndnc) npub1rk0a902jawxn6huj9h57z04s0ftf6n98sej4sy338pmur9xz99uqw8fkhq (npub1rk0…fkhq) We're still in the very early beginning of image generation from text prompts. There are already systems that can create texturised 3D models from 2D photo series or video, and there are systems that can create 3D models from text prompts. Those weird errors in image generators mostly stem from the fact that the model doesn't know anything about physics or 3D space, since it has only been trained on static 2d images.
While it is still far too early for bosses to fire their artists and writers, I expect generative systems to become a lot better in the coming years.
I think what really needs to happen is for lawmakers to make absolutely clear that AI output cannot be considered anybody's intellectual property, that any such output cannot be owned, copyrighted, or patented, that everything generated by an automated system is in the public domain from the second it gets generated.
Lord Caramac the Clueless, KSC /
npub19u…g6hsy
2023-12-10 00:41:57
in reply to nevent1q…vmz2
Author Public Key
npub19u4ha9q5zm5p0c6lq7055ftsxnn90nkc2pn4slewyzf5nzzf98nsug6hsyPublished at
2023-12-10 00:41:57Event JSON
{
"id": "39b3879c8ef79417284ba7b847e07dc15cb2bf58995a9c0a0de63acde5f3c8bf",
"pubkey": "2f2b7e941416e817e35f079f4a257034e657ced85067587f2e209349884929e7",
"created_at": 1702168917,
"kind": 1,
"tags": [
[
"p",
"cdc97b94a4ae17508ca60fe614a60bd605db78a140a0eeab9dce778b3435ce62",
"wss://relay.mostr.pub"
],
[
"p",
"1d9fd2bd52eb8d3d5f922de9e13eb07a569d4ca786655812313877c194c22978",
"wss://relay.mostr.pub"
],
[
"p",
"00a7ab8280576a40b4817882c0ec0e0d2407b777296e784dc6616d41fdfb374c",
"wss://relay.mostr.pub"
],
[
"p",
"0a99d175e964f4d1db5dd3151dc6d66ef07cb95863b01f2db4589835090e3228",
"wss://relay.mostr.pub"
],
[
"e",
"57b812629e7999133853dedd33893b2aaf4984563f817156f2349f64a5eb45d8",
"wss://relay.mostr.pub",
"reply"
],
[
"proxy",
"https://discordian.social/users/LordCaramac/statuses/111553342165789213",
"activitypub"
]
],
"content": "nostr:npub1ehyhh99y4ct4pr9xplnpffst6czak79pgzswa2uaeemckdp4ee3q63ndnc nostr:npub1rk0a902jawxn6huj9h57z04s0ftf6n98sej4sy338pmur9xz99uqw8fkhq We're still in the very early beginning of image generation from text prompts. There are already systems that can create texturised 3D models from 2D photo series or video, and there are systems that can create 3D models from text prompts. Those weird errors in image generators mostly stem from the fact that the model doesn't know anything about physics or 3D space, since it has only been trained on static 2d images. \nWhile it is still far too early for bosses to fire their artists and writers, I expect generative systems to become a lot better in the coming years. \nI think what really needs to happen is for lawmakers to make absolutely clear that AI output cannot be considered anybody's intellectual property, that any such output cannot be owned, copyrighted, or patented, that everything generated by an automated system is in the public domain from the second it gets generated.",
"sig": "0f76cd15d6d705cbb8eb3eeff79edf56b78eca562b3d56a4d470e60687e93672760cb035ce6c39e3a215a98fcc41e5193c53fb60352b166485cb815e9e9952df"
}