https://preview.redd.it/5qmk13wj7b3c1.jpg?width=1982&format=pjpg&auto=webp&s=bd43f39744651140e60378219562b7d9cd9cca7d

https://preview.redd.it/zm74vbwj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=9d7742363543db083462e6114ad72209505533b8

https://preview.redd.it/tb5w24wj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=0aad4eacbf12c7549ad1b5a44eceb7973470ef55

https://preview.redd.it/l5d8w2wj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=63ebab22550adb9644107443f2b502a6ac4090ca

https://preview.redd.it/rh8jhcwj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=c02df795d2b278d57d3ac0b80ee9e4355cab9c54

https://preview.redd.it/4bui54wj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=790916a1e24b686729a786deb503c3fcf88c26db

https://preview.redd.it/k8d773wj7b3c1.jpg?width=2048&format=pjpg&auto=webp&s=29af57b433eb301313d897ddae72f69dcfdf8a10

https://preview.redd.it/knbvecwj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=0fc513f33e3bea83c7814b768f3ff008ba9a03b8

https://preview.redd.it/t8d5bdwj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=19c9dce1d188d38135f1badeb2a6f52d8b61a960

https://preview.redd.it/gyxwf4wj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=50bd49d2c731475ece681e90557b976688200913

https://preview.redd.it/thyux6wj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=fe62db672e09fc8ee9519c37685a4dc73ccdb7b4

https://preview.redd.it/zwl8e4wj7b3c1.jpg?width=960&format=pjpg&auto=webp&s=174b1a85ddeec14852788747e6b131e500907042

  • mrfroggyman@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    10 months ago

    I’m not sure these are AI generated

    1. Image generation sucks really bad at text

    2. Image generation sucks at using IPs

    3. Image generation can’t really take an existing pic and generate new stuff on top of it (at least not the tools I know about)

    • NeuroPalooza@alien.topB
      link
      fedilink
      English
      arrow-up
      1
      ·
      10 months ago

      At least with Stable Diffusion you can train a LORA to be really good at using a specific IP, and can do img to img prompting where it draws from a specific reference image. That said the text was almost certainly done in Illustrator or something