F. provided PG with the link to the following video which explores ten different AI ad-on programs. You may wish to postpone your viewing if you’re at work or put on some headphones and turn your screen so no one can see it.
4 thoughts on “AI Apps”
Comments are closed.
One useful tool she doesn’t mention is BingCreate. Do try it. Free and you can download the images. It is still evolving. Fast.
Another promising but paid one to keep an eye on is PromeAI, not just for what it does now, but for what it claims is coming.
A perfect example of how the handwringers are worrying about the wrong things.
Things that came to mind from that video are: cover art (obvious); interior chapter header art; book trailers with short animations, voiceovers, and custom *original* music.
Chatbots? “Faux AI” stories? Small minds.
There are more productive u$e$ for that tech.
Mind-blowing video — hard to keep up. Thanks for the link — I’ll be chewing on that for days.
I still prefer my commissioned-artist covers (so I can stand out from the slick AI art world), but everything else…
I’ve only done one audiobook (read-by-author) but it looks like I may be able to go the route with mass production given the current state of affairs. (But when do I get any writing done?)
Just in the single image above I see the trends I already hate: the female on the left has her age reduced, her nose upturned to be more childlike, her eyes exaggerated in both color and size, and her fingers slimmed (otherwise they wouldn’t fit into those gloves, which are the same size as the original’s fingers.
I already find these trends in anime and cartoons – really, have you seen the waist-size of the daughter in The Incredibles? with no room for ribs or annoying things like stomachs?
No wonder teenage girls have image problems – and adults aren’t doing a whole lot better. The constant barrage of these images is ridiculous.
One more thing: “AI” is now officially a tech bubble like the late 90’s: Hollywood stars with more dollars than sense are pouring their money in itbthat didn’t go into NFTs.
Its going to get to where you shake a tree and three “AI” sites come crashing down.
So it may be time to dust off the old dot.com rules:
1- Free isn’t free.
In the dot.com era free sites burned investor cash to harvest eyeballs on ads and make a profit that way. Hundreds of startups, each aiming for 5% of the ad market. Barely a half dozen survived. In the new bubble most sites, even the reputable ones, are harvesting prompts. They’re a form of currency they’re buying with investor cash. At least one (LEONARDO.AI) has the “whiff of honey” and seems to want to harvest personal data, too. I don’t know for sure but I gave it a hard pass. I’d rather be late to their party if they turn out to be honest.
2- Welcome back to Internet time.
As in, last month’s bleeding edge is this month’s also ran. Worse, last money’s standalone product in next months is next month’s added feature to an existing product from a major software company. (As in, Adobe is adding many of the “AI” art features to Photo Shop, MS has been previewing a professional art program called DESIGNER since november, Corel Painter likely won’t be far behind.) If you find a useful tool that does what you need well enough to pay for, DO NOT go for the cheaper full year option unless you absolutely know you’ll get your money’s worth right away. They may not be around that long. And promised features do REAL.SOON.NOW. may never come. Let common sense guide you. CAVEAT EMPTORM
3- We are at the proof of concept phase of the new tech.
All those apps will either evolve into much more versatile and useful tools that make today’s wonders look dated. Doesn’t mean they’re not useful, but the eventual tools will be like EXCEL to original VISICALC. The reason all the free apps are “free” is because by harvesting prompts they can teach their models to better interpret user prompts and evolve to be faster, cheaper to run, and more effective. Today, generating exactly the image you visualize might take take a couple dozen tries as you refine the prompt. And the tool records the prompt, the output, and user reaction (zoom, download, etc). This info gets rolled back into the model so the more people use it succesfully the faster it can get better. Eventually it’ll get good enough people will pay cash for the “PRO” version.
Bottom line: it is very much worth keeping an eye on the new tools, play with them, see how they might be useful. There may be value adders to some businesses out there and as they say, early adopters get a headstart.
But these apps are just the first generation. Some will bloom (I’m seriously intrigued by the music generator–I’ll be playing with it in due time) some will flash off the pan (a copycat rap generator?) and some will muddle through. And some will come late but take over by addressing a negpected market. The key frame image to animation tool seems promising. We may be in for an age of indie animation.
Two areas where the image generators are guaranteed to evolve is iterative prompts (each output image most tools create are fresh creations without regard to what came before. An obvious improvement is to incrementally adjust an existing image. Some are pretending to do so but fall short. Often the new image bears no resemblance to the “guide” image.) and another saving created image elements for compositing. (One other direction that comes to mind is the ancient POSER graphics tool of decades past.) Both Corel and Adobe, among others already offer a limited form of this for pixel editing images. Lots of rooms to grow in all those spaces.
Funny thing: with the last boomers retiring and taking the bulk of their liquid investment cash into safer long term bets, the tech investment world was looking bleak for startups, especially after the fall of SVG, but along comes a new bubble to keep things lively for a while. I figure three years or so before the dust settles.
Should be fun to watch things evolve.