1. 18
    1. 5

      Aw, I thought this was going to be about feeding AIs room descriptions from classic text adventures and seeing what they drew. I’d love to see DALL-E’s impression of Flood Control Dam #3.

      1. 2

        This idea is very compelling to me because I’m working on my own text based MUD-like experience. I chose it as a game project because it only requires writing and programming - the two things I can do in abundance. Eventually being able to have art automatically generated from room descriptions would be incredible and something I hope to see as possible someday.

      2. 2

        This idea briefly nerd-sniped me. Thank you.

        Here is DALL-E 2 given the prompt “You are standing in an open field west of a white house, with a boarded front door. There is a small mailbox here.” https://imgur.com/a/i4KzZ9f

        1. 2

          Nailed it! Please keep going!

          1. 2

            Alright you asked for it! I took an hour and made a small blog post: https://nathancraddock.com/blog/2022/ai-text-adventure/

            1. 1

              Thanks! Those are impressive images, but I still feel the first one from DALL-E captured the scene in my mind best. To me the White House in Zork is very plain, not a mansion.

              (And I just noticed that iOS autocorrect capitalized White House, above — which also explains why Midjourney desided to show it as a mansion 😂)

              1. 1

                Yeah that first one with DALL-E was really good! It was still fun to explore midjourney though

    2. 1

      I really enjoy these creative uses of AI image generation. Especially interesting was the idea to split an image in two halves to extend the image to the sides.

      Also enjoyed seeing a screenshot of Indiana Jones and the Fate of Atlantis. I love that game.