My latest book, The Shadow Over Vinland and Other Tales of the Cthulhu Mythos, is available for pre-order on Amazon. The print volume will follow soon. Very horror! Much eldritch! Samples to come at the Cold Fusion Media website!
My latest book, The Shadow Over Vinland and Other Tales of the Cthulhu Mythos, is available for pre-order on Amazon. The print volume will follow soon. Very horror! Much eldritch! Samples to come at the Cold Fusion Media website!
I want the deluxe edition with the cover that is the actual woodcut.
The edition that actual Vikings stole from a monastery somewhere in Britain!
I was planning at one point to make the actual carved emblem, but A.I. proved surprisingly capable…
Been doing a little experimentation with A.I. text-to-imagery generation myself. It’s pretty good with making product logos and such (e.g. your carved wooden emblem; got that from NightCafe, I’m guessing?) and fair at posing living creatures and people in a given setting if you’re careful enough in describing the poses and setting (e.g. “sitting/standing/walking on [color] carpet/desk chair/wooden floor/tiled kitchen floor/throw rug/etc.”), but so far, I’ve not had much luck getting any such A.I. engine to portray things like body language and actions (e.g. “guy cupping gal’s chin in his right hand as they gaze into each other’s eyes”) or behavior in general. It’s like the engines have got a pretty good grasp of nouns and adjectives, but not so much of verbs as yet.
It should be interesting to see how A.I. cover art generation will affect future entries on this site. On the one hand, as I said of the pseudohumans tag recently, 3D renderings will stop looking so hideously plastic and start looking like something from a Pixar movie—which might get a little bland after the zillionth time someone does it, but will still be a marked improvement over what we’ve been getting up to now; and the quality of all other kinds of made-to-order art will likely improve as well. On the other hand, considering how many times a terribly composed/drawn/rendered cover has left us asking “Who could possibly have said ‘Yeah, that’s just what I wanted see on my cover!’ after looking at that?” on here, I think cover designers who turn to A.I. engines to do the artistic legwork for them will still be providing us with plenty of material to fill our days with laughter.
(Case in point: that Wall Pilates Workouts book cover that “won” a Lousy Award this year? I think that was probably made using one of the earlier DALL·E engines, which were often known for royally screwing up renderings of human faces and limbs and digits (possibly on purpose) in pretty much the way the human model is royally screwed up on that cover. Nonetheless, somebody evidently took one look at that horribly disfigured rendering of a human body in the A.I. engine’s display window and said “Works for me!”)
Shucks, we have a whole slew of covers with good pre-existing artwork, marred by bad layout of inappropriate fonts.
A.I. will be a tool, like having access to very specific stock photography. Still doesn’t mean that people will know what to do with it, any more than owning a ratchet set makes you a mechanic.
(I found out that no one A.I. engine will give me what I’m looking for. I can’t remember which one gave me the emblem; Adobe’s Firefly gave me the border and background; and then I still needed to edit and reproportion them in Photoshop, experiment with a dozen fonts, etc.)
From what we’ve already seen, it’s not just the layout cover designers will be screwing up, but the artwork itself. There’s apparently a significant portion of lousy cover designers—especially the ones who do the “cargo cult template” covers—who treat the cover design programs and templates on offer at Amazon and other major outlets as if using them guarantees the cover will be good; probably the same kind of people who think the cruise control feature for a vehicle is just like the autopilot on a plane. They’ll probably start thinking of A.I. the same way: “If I tell it to design a cover appropriate to my book’s genre and style, whatever it spits out must by definition be appropriate to my book’s genre and style.”
Judging by some of the stuff I’ve seen these text-to-imagery A.I. programs spit out in response to my instructions so far, however (e.g. one gave me a picture of a guy with a regular pair of arms at his shoulders, but two forearms and hands stemming from his left arm, one of which also had rather bizarrely deformed fingers), there are going to be a lot of hilariously grotesque glitches in some of these cover designs that these naive and trusting little lambs using the programs will suppose are probably just part of some new kind of avant-garde surrealist art style all the popular people are using these days. In other words, the advent of A.I. cover design may actually increase the number of hilariously bad covers we’ll see being produced in the near future.
Wouldn’t be surprised. Most of the problems with the AI art are obvious if you look at them for two seconds…but there are many covers on this blog that, if the author looked at them for two seconds, the problem would be obvious.
I foresee a number of romance covers with three-legged heroes who live in castles that could only exist in an Escher drawing…and the authors of those books saying, “But, what’s wrong with it? Why are people making fun of me?”
Looking forward to it and congrats! Not a lousy cover, however 😉