AI-Research and Articles - Proof of Concept - Can AI make a functional Fantasy-Architecture-Design?

Once I first saw, that a lot of this was working as intended, I fell into a manic-episode. I was researching and testing for a 100-hour-week. My brain was constantly flooded with new ideas and applications. My first intent was to see, if a newly trained model on my personal artwork could design something else, than it was trained on.

I was once again seriously surprised by the output and the intense production-output-increase.

I can say it with confidence now, 3 months later. Once a proper pipeline has been build, a specific output load increases by a factor 2000% relative to purely traditional work.

My next question was, can I get this new architecture design into Unreal Engine 5, and once again I was shocked. I managed to pull all of this off within 3 days.

Day 1: I experimented with my new custom-LoRA in comfyui.
Day 2: I started editing some good shots in Photoshop, by bashing and overpainting to create a logical design, that still holds up, once you zoom in.
Day 3: I played around with AI-Mesh-Generation, and actually managed to make it work within 2 days.
Day 4+5: I started assembling all the parts into a coherent scene within Unreal Engine 5.