My Midjourney Experience

Midjourney is an AI project environment where typing in certain keywords or sentences will generate images based on your text. As someone that is fascinated with technology and architecture, I looked past the photorealistic images of people and fantasy scenes and saw potential... perhaps a glimpse of the future of our architectural practices?

None of what you are about to see is real, it is all computer generated based on the text I used to describe the scene and the massive database of reference material that Midjourney uses to understand what you might mean.

Starting simple, I decided to give it an easy one. Provide me a detailed view of Prairie Style and Frank Lloyd Wright might design. Seeing the above image I was immediately intrigued by what the future may hold in my trial and error.

Asking it to provide me a Design Study of Falling Water I had to smile. It wasn't what I had expected to get, but what it provided seemed familiar and captured the general aesthetic that I might expect.

Asking it to create for me "a Norman Rockwell style image of a country house with peeling paint at sunset" generated a variety of interesting combinations, including the one above. From here, I decided to go in a slightly different direction.

Suddenly I'm laughing out loud, as my creation words were "Show me the interior of an office created in Revit". I had expected something rather raw and clunky, instead, it generated something much better than I had any right to expect. To be fair I checked it off my mental checklist as something it did wrong, but really... did it? It was just more than I asked for.

After creating a series of marble statues designed in the style of Bernini (I can post those if you have an interest) I started working on various words related to stained glass. My oh my, can it generate beautiful stained glass. But I'll skip that as well and go directly to my glass sculptures that came next.

This request was "Create a statue of an American Eagle designed with Tiffany Glass. Back lit. " By the way, using Tiffany Glass as an example creates some of the most dynamic glass objects you'll see, which makes sense if you know Tiffany Glass.

Feeling I was getting off track, even if these experiments did teach me a lot, I asked it to design a home on a specific size lot and make it be green, black, and white. Also, make the house exist in Oregon, USA. While there is a drainage issue or two, the results I received were really impressive.

So what next? Well, I decided to go upscale. I asked it to design for me the most beautiful, ornate, detailed, and expensive room possible. While this might not 100% qualify, it certainly did its job.

While the previous command was great, it didn't quite meet my aesthetic. So I added the words at the end - "color: forest green, white, black, red, brown". Needless to say, I was pleased with the outcome.

Finally, I decided to go back down to something realistic. Asking it to design a 3D warehouse with metal panels for storage and manufacturing. As a result I had generated many beautiful designs, including the above.

Ultimately I have now had it generate at minimum a couple of hundred renderings based on a variety of criteria. It's rather easy to see where this is all going. Provide a database of your design criteria and then refine, refine and refine. I'm rather certain software like Conix is going in the right direction, but I can see these options working in a variety of ways.

I highly recommend trying your hand at Midjourney, ChatGPT, Conix, and other AI offerings as you see them become available. You might not use them today or tomorrow, but building on those skills and connecting the mental dots of "what if I could do this...." will take you a long way. Ultimately these will be more than just renderings, instead tied to the way we design, estimate, construct, and manage facilities and the future is closer than you might expect.

I will be writing more on the various solutions and how / why they should be implemented in your own practice or skillset in the near future. Until then, enjoy exploring a very small, elementary aspect of the future of work.