I uhh have no idea whats happening. I pressed space once and then - things.
10/10
Could be cool if the images were generated on the fly depending on the stuff happening, but maybe that'd be too hard to get working right.
Viewing post in ML Yume jam comments
It would take too long to generate images in real time and Midjourney still puts out messed up images too often. I also need to make sure the text kinda matches the image. GPT-4 can recognize what's in images but that ability isn't public yet. Plus I would have to pay for every person playing the game. So yeah doing this stuff real time isn't worth it at the moment.