05/11/2023
Okay, so for the first initial observation, I won't overload this one with images.
MJ vs. Leonardo.AI.
It's fascinating to think that I started this as a text-to-image AI generator comparison. I would love to say that I have a clear winner, but I simply don't. Comparing the two is like comparing apples to oranges.
In the traditional text-to-image sense, I would have to say this one goes to Midjourney. Midjourney seems more fleshed out, training model-wise. It's really easy to pick up and play. Midjourney is easy to use and can create really decent images with almost no input. That being said, in my opinion, MJ has focused on one thing and one thing only: interpreting human language into desirable images. This hyperfocus on one aspect of training is both a pro and a con.
That leads me to Leonardo.ai. Oh boy, where do I start? First off, this is the "create a character" of prompt building, slider bars and all. There are so many UI options and settings to learn, even in the base text-to-image. It certainly isn't rocket science, but it's not as easy as the "type-to-picture" experience you would have with MJ. That being said, it may not be as easy to use, but it makes up for that in control. You have so much control over the direction the image will go and what direction you want it to take.
So, to sum it up, Midjourney is the guided tutorial of text-to-image. It does 90% of the work for you. Leonardo, on the other hand, is the full game past the guide. You better know the ropes or you will be lost.
I will post more as I learn more; this is the rough comparison thus far. It's important to note that Leonardo.ai is free to try, while MJ used to be free. Interpret that as you will.
PS: images are of me trying to image to image within leo, it is possible to add text doing this so of course i had to try it pic 1 , crude image of a sign , pic 2 image to image with prompt