I'm a game developer, professionally!

You may know me from things like: Where the Water Tastes Like Wine, The Museum of Mechanics: Lockpicking, Gone Home, Bioshock 2, or maybe something else.

Right now I work as a Technical Narrative Designer at Remedy Entertainment in Stockholm, Sweden

Perhaps there are other aspects of my personality that may also be revealed here on this website


Email
johnnemann@dimbulbgames.com
Discord
Johnnemann

Is that you can see, very precisely, what the person generating it was hoping to accomplish through the prompt. I wonder if seeing the image and seeing below it like "attractive white woman mother sympathetic Rockwell holding [product]" would do anything to people's awareness of the attempts there in, like, advertising or editorial illustrations. Would requiring people to publish the prompt change how they generated images? I dunno. Interesting.

Also interesting to think about how the process of writing a prompt might make explicit some things that illustrators do subconsciously. Although I imagine a lot of that is already baked into the models through the trained bias... Things like linking "sympathetic" and "white" are probably already encoded in there.


You must log in to comment.

in reply to @johnnemann's post:

Only yesterday I saw a discord buddy commenting that the single biggest benefit about generative AI is that it's training its users to communicate their desires clearly and in plain (if stilted due to the nature of prompts) language. It's an interesting notion to chew on, and it'll be worth watching to see if those lessons stick outside of writing prompts.