Naming the nameless
Untitled Science Fiction Project is a cheeky placeholder I’ve been using for a long time now. Various stabs at a better working title fell short. Now I’m deeper into the story and I found a working title that I’m happy with.
In Scrivener, I label scenes based on which faction is present. In the corkboard view, the label colors help me keep the focus moving from group-to-group so that none of the characters drop out of the narrative for too long.
Here are the labels:
Viewing this menu and the corkboard many, many, times, a pattern emerges. Dragons and monsters. It won’t spoil more than the opening chapter or two to reveal that there are two groups of heroes in play. Team Dragon gets its orders from someone or something whose avatar in cyberspace is a magnificent Chinese dragon. The Noble Monsters are soldiers-for-hire who strive to be, well, noble. Dragons and Monsters: FIGHT! Dragons and Monsters: UNITE! Dragons & Monsters has a heavy fantasy vibe, so it may not survive until publication. It’s far less cumbersome than its predecessor. (I still catch myself saying Monsters & Dragons. Camping out next to a major trademark may be hazardous to your health.)
More AI slop!
Facebook has a bias in favor of showing posts that are accompanied by pictures. To feed that algorithm, I am willing to turn to generative AI. I’m not excited by it and would not use generated images for publication. For a one-off social media post, I will yield to temptation.
Image generation has come a long way since I started working on Dragons & Monsters. Initial attempts to conjure settings and characters produced various amounts of visual confusion and abomination. The state of the anti-art is much, much better these days. It took me three steps in Copilot to get the D&M one-off.
Are the auroras where they should be? No. Is the ice coverage right for my planet? No. Nordlys is an ice age planet, but like Earth, it isn’t completely frozen-over. But for my purposes, this image is acceptable.
For the first years of AI image generation, text was a major point of failure. Letterforms were distorted or flat-out wrong. Spelling was approximate. The engineers have overcome this problem. I suspect that a separate text system has been added to make the text work. It may not even be generative in the manner of the background image. It knows something about fonts and layout.
That this worked at all was a pleasant surprise. Embossing effects have been around for a long time, but it’s a pretty niche style.
So, yeah, it’s AI slop, but it was very easy to get slop that met my short-term need. Yay?