The journey to completing The Galarian Chronicles has very much been a back and forth affair. I got the flames reignited inside me, desperately wanting to see at least the first book completed, but the actual progression since then has been a bit lighter than I would have liked. I partially blame it on Diablo 4, Last Epoch, and watching Hoarders and Karen bodycam videos on YouTube.

A couple of nights ago, I somehow remembered that AI image generation was a thing (how I forgot that, I have no clue). I was having a particularly down and sluggish day, but I wanted to work on the story, but I didn’t necessarily have it in me to put pen to paper. I decided to go to an online program I used to make/edit images for my Twitch streams, and was reminded that they had an AI image generation area. In my head, I already knew how people looked, but since I have the artistic abilities of a drunk salamander, I could never really get this all out in a way where I can always reference it outside of my head meat.

So I gave the AI image generator at [redacted] a shot. Started with mostly an overly simplified request like:

“half elf male Warrior in a full suit of dark blue armor and dark brown hair”

First result was really freaking close to how I envisioned Galarian if it were a movie or show or something:

I then decided to get a bit more complex, requesting the three main figures of the first book together, specific looks, and so on. This one was more of an animation skewed representation:

With only minor grievances to what was generated, and outside having what looks to be nine fingers on one hand, the style, tone and feel was there. It was such a surprise how close to my thoughts it came with the information I provided.

Fascinated with the mostly accurate depictions, I spent a couple of hours generating different images of the leads in the story. One problem had come up though – the site I was using gave limited “credits” per 24 hours, and if I wanted to continue using those services, I would either have to pay a monthly fee for a paltry number of credits (did not do), or cycle through nearly a dozen different Gmail accounts and do it that way (I have no shame). Frustrated with how quickly I ran out of email accounts, I did some searching on Google, and stumbled upon a similar AI image generation site, and ran into the same debacle.

I looked into one that was 100% free, no credits, no needing to sign in and out of hundreds of email accounts, and the quality comparison was shocking. Shocking as in I see why it was free and the others were limited free use. A lot more instances of hands and eyes being distorted, contorted or just a mess. That became another frustration, because when I attempted to add a sword with blue flames, or even a sword in general, 75% of the time it would provide warped looking weaponry and appendages, or just make something that made me laugh out loud at how it came out:

I did find some more sites that were limited free use, and just like the previous ones of similar stature, the quality was beyond compare between it and the 100% free generated sites. Honestly, if it offered unlimited AI generated images for X a month, I would bite for a month, but the bitter taste of how few credits you get for the low end sub – ghastly. At the same time though, I completely understand, since each of these sites that did limited free generations were of some high quality.

As mentioned previously, I can barely draw a stick figure these days. There was no kind of visual story board I could construct in helping to literally see my vision come to life, and from what I’ve done so far with the image generating, I feel like I can dig deeper into telling about these characters. It’s a visual tangible that was very much wanted and needed; trying to continually go off thoughts while getting other thoughts down about the story was a bit cumbersome after time.

AI advancement has been absolutely mind-blowing over the last five years. You have image generations like these, you have Twitch channels with parodies of famous figures and streamers with an AI voiced performance, and then you have what I just stumbled upon a few days ago – AI generated movie trailers. Granted the latter is still very volatile when it comes to consistency and quality of certain aspects, it’s still hard to believe we’ve come this far with technology. There are worries about AI overtaking human jobs, and that’s a very legitimate concern. Like with anything, there’s promise and drawbacks, and hopefully with such a massive development in technology that AI has been taking, we don’t take away the true creativity that humans provide in a myriad of ways.

AI Image Generation has helped me a lot more than I would have ever anticipated. At times I’d see my stories as a movie or animated feature in my head, and to get a close enough 1:1 mental dump like this, it’s astounding. To finish this little blurb about my journey with AI image generation, here are a few images that I saw and clicked with, in both a “realism” tone and a more cartoon-like presentation (which shows off five of the main characters from the first book, a couple more still to do – I did get the main antagonist done but I don’t want to spoil that just yet!):

I guess this constitutes as a book update? I feel it’s a significant enough development for me and how the book proceeds. If anything else comes to mind that warrants a whole update or info dump like this, I’ll be sure to do so.

Oh, and AI image generation of a logo is still pretty messy. AI can’t do hands, hands holding objects, or spelling properly. I’ll have to find alternate methods of developing a temporary logo!