Sign In

Close
Forgot your password? No account yet?

Profile

Digital artist and tinkerer originally from Fayetteville North Carolina USA. I currently reside in Spring Lake North Carolina. I'd be lying if I said I didn't miss the Cascades. Single, no children. Interests include science, art, reading, music, food, culture, the outdoors, etc. I enjoy making and building things. Art software used include Truespace, Poser, Photoshop, Bryce, and FLStudio. Primarily a Poser artist these days.

Love to roleplay. If anyone is interested in discussing roleplaying please contact me via my inbox to discsuss it. I also love to talk about art and making things and am looking to talk to and make friends with many kinds of local artists and craft people. Especially in the immediate area. If you want to discuss any art or craft especially construction techniques or you have questions about one of my submissions or want advice on how I did something please do not hesitate to ask. I'm quite approachable and love to help people when I can.

Thanks!

Latest Journal

My Thoughts on A.I. Art and a brief Review of A.I. Art Apps ***Long Read***

Recently I began dabbling in A.I. Art. For those who (have been living in a cave and) haven't heard, A.I. or Artificial Intelligence Art is a means of using complicated programming to interpret a prompt given and then creates an image based on it. It's far more complicated than that at the nuts and bolts level but that's the gist. Below are some various thoughts on the topic followed by a brief review of the ones I've tried.

My First Impression

A.I. Art and its associates A.I. Music and A.I. Writing are nothing new. Lifeway was (or is? I don't know.) A web based service which allowed a person to create a unique piece of music by selecting sound samples to use as triggers. But the phenomenon of computers creating things that used to be the sole - or soul- province of Human Imagination has recently become advanced enough to seize public attention. I had known of A.I. Art and had not bothered with it until I watched a Youtube video by The CrafsMan entitled something like 'This Action Figure Does Not Exist'. I was intrigued because the thumbnail showed what clearly appeared to be a solid object. The video then goes into detail concerning the A.I. Art powerhouse Midjourney and what it can do. This program, with just a few words of descriptive prompting, created an incredibly realistic looking image of a 1980s style toy which you'd be forgiven for thinking was a photograph of an actual object. I was instantly blown away and after hemming and hawing a couple days I decided to try it. It took me a bit longer than it should have to figure out how it works but once I got it going I was awed. I entered my first prompt and got a stunning image of a Samurai walking through a bamboo forest lit by japanese lanterns. A quartet of beatiful looking steampunk rayguns. Several others. I was made a believer let's say. (If you want to dip your toe into a good brief but very explanatory look at A.I Art for somebody with no prior knowledge I highly recommend CrafsMan's video.)

Initially as I stated I had little interest in A.I. Art because I had assumed it was some some niche experimental thing for nerds and programmers that probably didn't do a good job. Allow me to state how wrong I was. It turns out it's an incredible and frankly rather addictive process. The A.I. Art folder on my phone currently weighs in at over half a gigabyte and is growing.

What Is It and How Does it Work

A.I. Art is a complicated process that uses user prompts to interpret an image. It's nothing new; the experimental history of it dates back to the 1960s. It's only in the last decade or so as far as I know become advanced enough to be noteworthy enough to create an interest in public consumption. It's internally a complicated process but on the end-user level it goes something like this: You enter a prompt based on a few words, for example 'Cthulhu in R'lyeh with Glowing Crystals Stormy Sky' and wait. The software parses the net to find out what those things are then creates you the image you want - at least ostensibly. In reality you're just as likely to get an image of a dark story sea with a rock sticking out of it with disembodied green tentacles trailing down it and red patches of light all over. Or it may give you a mindblowing high quality unique image. The results are usually middle of the road. Like asking any artist to create an image of something from a description it's highly open to individual interpretation. In the 5 weeks or so I've been using it (as of the time of this writing) I've generated reams of unsatisfactory images but I've also retained around 600 gigabytes of stunning
imagery. The process is usually fairly quick taking under a minute but most A.I. Art engines caution that the process can take up to two minutes. The output of the programs can be a single image or a batch of images. All the A.I. Art apps I've used give you the option to save images and some allow you to publish them to a website if you make an account. A few have some cool features indeed which I'll go into in detail in the review section.

On the end use level the apps have a field to enter in your prompt which is typically limited to a about 120 characters. Most of the apps ask you to select an art style such as 'steampunk', 'oil painting', 'realistic' etc. Then the app asks you wait as it works. Most of the Apps will then generate either a single image or a small group of them. You can then choose to download what you like. upload to the Some of the apps also allow you to upload an image to use as a reference. This gives you an image based on thst reference or even a modified copy of it. What keywords your prompt contains as well as the order they're in and the art style you use all have an effect on the findal product. Also typing in prompts that contain better known material or more common material will get you better results as the app must interpret your request. Since it isn't watching Hulu it has to turn to Google for input and just as with a human artist more obscure requests are going to have less to go on. For example you are going to get far better results for 'Peter Quill' or 'Howard the Duck' than you are for 'Spartakus and Arcana'. The results are then typically relatively small in image size and file size though they look like they'd scale up to the size of a small poster. They usually save in a lightweight format like .jpeg or .png.

Reactions to A.I Art and Questions

This section contains various reactions I've seen on the topic plus some questions I've been asked and my answers.

Isn't this stuff all just a bunch of hype?
No it isn't. My initial reaction was that nothing created by a computer could possibly be any good. I was wrong and having seen the results and the interest it's generated I doubt it's going to go anywhere anytime soon which is likely a good thing.

But is the art any good?
Yes and no. Typically you get mixed results. You're asking a computer to create an image based on keywords. All art is highly interpretive and A.I. Art is nothing but interpretive. Suppose for example you go to a cake decorator and say 'gimme a pic of Thanos wearing the Infinity Gauntlet on a cake for my son's 12th birthday.' The cake artist might know exactly what you want and then she creates an awesome picture of Thanos. Or she might have no idea who he is in which case she turns to that great repository of conflicting answers we call Google. Then the results are hit or miss. He may not be in the pose you wanted or he may not be scary enough or he may make your kids cry. The A.I. algorithms doing your art have no idea what ANYTHING is so they parse the net for answers and then try to figure out what you want. Then there's the fact that the apps use procedural generators which tend to give some things a random look. Most of the apps I've used have trouble with people and living things and in particular faces. I've got a lot of results that looked like somebody took a Picassoand threw it in a Vitamix. But I've Gotten stuff that made me say 'yeah that's gonna get printed and framed' and a couple that seriously made me consider getting a tattoo!

This Sh😜t is going to take traditional artists' jobs from them!
No it isn't. Calm down. If you'll notice everytime a new innovation makes somebody's life easier somebody complains that it's going to put somebody out of business. Assembly lines didn't put factory workers on the breadline and Doctor Hawking still had a job as an iminent mathematician long after the advent of supercomputers. Nor did animated movies mean we don't still see live action. People
like both so both exist. 3D Printing hasn't caused the dinosaur-like
extinction of traditional crafters and propmakers and traditional artists are not going to die out either Art is a primary form of Human expression and people value the art they create as much for what it tells them about the artist as for the art itself. When I look at a painting done by Michael Boehme I see that he is inspired by the same things as me: Love and concern for nature, a passion for Planet Earth and a deeply awed sense of wonder concerning the mysteries of the universe and the future. When I look at an A.I.
generated piece it tells me some guys are outstanding programners
and a bit about the person who prompted it and that's it. I can have no emotional and intellectual kinship with a piece of silicon. A computer is no more aware that it's doing art than the sun is aware it comes up on time each morning. The computer doing the art could not be doing it apart from the fact a group of clever humans enabled it to. Computers are STILL not artistic or imaginative. Then there's the reality that it's a hit or miss process. You can still go to a human artist and get a much closer representation of what you want in exactly the style you want as opposed to making do with a piece of A.I. Art that's meh...good enough. Additionally dealing with human artists always gives us the potential to make friends, expand our social groups, learn, share ideas and such as can't be had with machines. So relax artists. You aren't going to come in and find GlaDOS in your office monday morning. What I DO think A.I. art will become is another powerful tool in our overall ability to get things done.

This stuff's not really art because it doesn't take any skill.
I highly beg to differ. As somebody who's been doing digital art for almost a quarter of a century NOTHING is more irritating to me than hearing somebody say 'you didn't do it, a computer did it for you.' Nothing is further from the truth. Art is art. A.I. Art is still art because not only did it take the creators of the software years of research and developmemt to make the computer ABLE to do it, it also relies on the imagination of the end user to create the art. As I stated earlier the computer imagines nothing. It has no idea what art is, what constitutes good or bad art, whether a topic is in good or poor taste. All that depends on the brain of the user. Not only that but there is certain skill in how you approach getting what you want out of the software. How you word the prompt, which keywords you use, which art style you choose and even which app to use all factor in. Then you have to ask yourself what art is? Does it give you a sense of satisfaction? Is it inspiring? Is it beautiful? Does it fit the need you created it for? If you can check off most or all of these I'd say it is art. Again, and I cannot overstate this- the computer did nothing on its own. The art was STILL created by humans even if indirectly by using a tool. I like beautiful imagery which inspires me regardless of where it comes from. If I want to frame it and put it on my wall.....it's art.

Yep..another example of how computers are going to take over.
Ok. So this one is really SQUIRREL! and doesn't have much directly to do with art but I'll address it anyway. No, Mr. Worrywart computers are not scheming to conquer us all. No matter what some Google engineer said about A.I. becoming sentient it hasn't. It has become a scarily accurate MIMIC of sentient as of late I confess . But in order for computers to take over they'd first have to have a REASON to take over. Lacking any creative or emotional capacity even if it had a reason it wouldn't have an impulse to act on it firstly. You could imagine any scenario and the concept predates Arnold Schwarzenegger dressed in biker leathers saying 'I'll be back'. Suppose a computer does come up with a notion, somehow, to conquer humanity. In order to do thst it first has to PLAN how. It can't just randomly set off all the aromic bombs; it'll destroy itself in the process. It can't control the weather as that technology doesn't exist so that's out. It could easily shut down the power grid and communications systems but even that we could combat. Since it is a COMPUTER and lacks imagination and critical thinking it will have to consult current and historical information on military science. No problem; it has access to every search engine on Earth, can bypass any firewall, hack any secure database (we're essentially imagining a Sci Fi scenario here) etc. Now it has the history of every military genius and every battle from Sun Tsu on. Now it has to distill that info into a workable plan and THAT is where its trouble starts. It has no ability to factor in things like skill, experience, natural cunning, judgement, snap decision making, et cetera. It can only calculate its plan based on staristical data alone. A.I. Art is less of a dire portent and more of an example of why computers WON'T conquer us all. Say you prompt 'Ghostbusters'. You may get a great image of the guys all standing around Ecto-1 in their Proton Packs or you may get something that looks like a handicapped Shoggoth trying to do a Bill Murray impression. The computer doesn't know if it did a good job or not and it doesn't have the ability to second guess itself. And I've gotten Thousands of results that from a human standpoint were failures. All the evil computer has to do is put one flawed plan into action and it's defeated. So in other words....Shut up Buckley.

What can I do with it?
What do you want to do with it? Pretty much anything you could with any other art. Print it out and hang it on your wall. Get it put on a T shirt. Get a tattoo of it. Use it as part of a craft project. Get it printed on Water Slide film and put it on your computer. It's a digital image so it can be transferred to pretty much any medium. And it's totally one of a kind. You'll never get the same result twice even if you use the exact same app and parameters. I think it'd be cool to turn some into lithophanes. I will caution this though: make sure the content is being used legally. You may think that since the art was created electronically that no one has any rights to it but not so fast. The software that created it constitutes months or years of research and development and proprietary technology created by people who then own the rights to that software and so either technically or explicitly own the right to the art by extension. Most of the app devs don't seem to be too anal about what you do with it for private use but if, say, you want to use some as the cover art for a novel you will very likely need to negotiate a license with the company who made the app and owns its rights and run into all the complexity that implies.

Is it worth it to get into A.I. Art?
YES! It absolutely is worth it. I wish I'd gotten into it sooner. Not only do I now have reams of high quality unique art that nobody else has but it's just fun and easy and habit forming. It's a great way to while away idle time and you can do it in the background while doing something like chatting or watching youtube. It's fun to see what results you're going to get and always keeps you guessing. The apps are typically small and run just fine on tablets or smartphones. The results can be both amusing and useful. There a million things you could do with them.

Who would benefit most from A.I. Art?
Lots of people. IMO the people who would benefit most are people who love art Ars Gratia Artis (or for its own sake) and want to have good quality original artwork to enjoy without spending a lot or having to buy a lot of art materials. It would most benefit people like me who have zero skill at traditional art mediums like painting and drawing. If you're building a diorama or a roleplaying terrain and need some art for billboards or posters or what ever it has you covered. I've used it extensively to make good quality Illustrations for some cards for a roleplaying game I'm working on. As I have no skill at drawing I was ecstatic when I stumbled onto Dall-E's ability to make illustrations in a style similar to card game art. If you're an incorrigible nerd like me and refuse to give up roleplaying and you need an authentic character pic but don't want to spend weeks in Poser you can have a decent one in ten or 15 minutes. Say you want to make an original prop or costume for a cosplay or LARP etc and can't figure out how you want it to look. A.I. Art has you covered once again. I plan to build physical props of several of the results I've gotten already. So basically anybody who needs good quality art on the cheap and done quickly and especially people who lack traditional art skills. As a side note say you're a highly gifted traditionalist artist who can already create anything you want any time. Should you use it anyway? Yes, because it's fun. And the results might inspire you.

What are some Pros and Cons to A.I. Art?

There are pros and cons. I'll list some of the major ones I've run into.

First the Pros:
Fun. You can easily spend hours doing it and is a good way to kill time.
Easy Access to high quality art without paying a lot.
Free to inexpensive to use - many apps don't cost anything and you don't have to buy piles of expensive art supplies. Subscriptions are typically not large fees if you are so inclined.
Accessible- most of the apps run just fine on smart devices so you don't need expensive high end hardware. This makes it valuable and attractive to folks without lots of money.
Wide variety: There are several options to choose from which all give different results for flexibility.
Easy to get into. You have good results within your first hour without complicated and time consuming training so you know right away if it's for you or not. If not you simply uninstall the app. No suite of supples you will now have no use for and take a loss on at ebay.

Inspiring: gives you plenty of food for thought within minutes. Get a quick sketch to base another project on etc.

Portable: Since most of the software runs fine on tablets and phones you can take them with you anywhere without having to drag around a laptop.
Doesn't require lots of hands on effort. You can run the app and let it work in the background as you do other things.

Wide variety of possibilities let you vary your results from app to app or style to style. You can keep running the prompt over and over till you're happy with your result.

Publish your designs: Many of the apps have an internal command to publish your designs in a personal gallery online after making an account.

And then the Cons:
Subscriptions and Paywalls. This was the first major drawback I ran
into. Many of the apps have some Limitation on the number of
features or images you can make without subscribing. SAAS is absolutely persona non grata to me so any software that requires a subscription or it won't work is instantly and permanently blacklisted.
Time to generate images: A.I. requires some patience as it takes time to generate the image and even though it's usually not long it can add up.
Inaccuracies: As with anything being interpreted by someone or something else it's impossible to get exactly what you want. You have to lower your expectations a little.

Images getting cut off: This happens in most of the software I've tried. You may get an otherwise perfect image only to see an important part of it is cut off because it doesn't fit in the image. Most output images in portait format.

Inability of the procedural process to accurately render certain subjects. Some topics are difficult to portray accurately; particularly people and animals. Especially faces. I've gotten plenty of results that looked like something torn apart and glued back together by monkeys, plenty of human centipedes, images on top of images, et cetera.

Bizarre image artifacts: All the software I tried have problems with strange artifacting being fairly common. Things like characters with extra limbs, 'human centipede' type chains of body parts, warped and distorted things like off round ballobs and planets. Objects stacked on objects- For expample getting a car stacked on a car or two scenes in the same image one atop another.
Easy to end up with several different apps. If you like a relatively clean device with as little installed on it as possible take note. I currently have four of these on my phone. I'm likely to end up with more if I find ones thst even better meet certain requirements.
Space Consuming: while individual images typically take up less than a couple megabytes I have generated thousands of them and it's exacerbated by the fact that it's easy to make several using the same parameters and get equally desirable results so that I have have ten iterations of the same concept. Get prepared to devote a separate folder(s) on your device to store them in and have plenty of storage space or have a flash drive handy.

Multiplicity: As stated above you may be tempted to run the same prompt many times with the same or different parameters.
Frustration: The apps can be frustrating for various reasons, not the least of which is not getting results close to your expectations right away. Some apps create a number of potential outcomes but select only one final one which you have no control over meaning you may not get the result you really want.

The list of pros and cons are pretty close with the cons maybe just barely edging out the pros. Don't let that stop you if you're so inclined to try it.

Review of The Apps I've Used

The following will be a brief review of several of the apps I've used and their pros and cons. These are my own private thoughts and none of these were requested, paid or sponsored in any way and I make all efforts to be accurate and objective.

Midjourney

This celebrated A.I. Art software hss been making a buzz of late and it's easy to see why. It creates images that are nothing short of just stunning both in the accuracy it interprets prompts and the quality of the art it creates. It currently occupies the top slot in my ranking of most capable A.I. Art Software. My initial impression was that it was slightly difficult to access. In order to use it you need to have a discord account and join one of several threads. Entering the prompts was difficult at first until I figured out how and then I was blown away by what I was seeing. It definitely deserves its current state of celebrity. The software outputs images in quartets and you can upscale and save them. Images can be saved individually or as groups of four. The app will give you a certain number of 'credits' which are required to generate images, upscale and save. On a trial basis you can create about fifty depending on how many upscales and saves you do. Subscribers can make about 200 a week. The software is currently in open beta development.

Pros:
Stunning performance. Midjourney currently occupies the top slot for the best of both interpretation of prompts and quality of artwork generated IMO.
Images are made in batches. You get four each generation so you
can take your pick.
Upscale images: You can scale up an individual image as you like.
Other apps I've used lack this feature.

Prompt administrative support. If you have any questions you join the support threaad and a representative answers your questions in real time.

Cons:
Paywall limited: The software is in current beta development and beyond a relatively limited trial period requires an ongoing subscription to use. Any action requires credits which run out fairly quickly. The company currently has no plan to allow users to buy credits as you go. They indicated they did not plan against such a feature in the future as well but no speculation was available.

Relatively difficult to access. Using the software requires users to create a Discord account and join its thread. This could be a bar to users who don't want to go to that trouble just to decide they don't want to keep using it.
Slightly confusing. It took me several minutes to figure out the correct syntax in which to enter prompts. Entering them wrong gets
an error message or simply does not do anything.

Non discreet: You enter prompts in a chat environment and others
can see your prompts and results and they can see yours. I don't
know if you can download or upscale other peoples' work or not.

Multiple users at the same time: Images are generated for dozens of users at the same time. Your request is qued and then you have to scroll up or down to find your result. It is easy to lose track of your place.

Software is in Beta mode. This -may- be a potential bar to some users.

As you can see the software currently has more cons than pros in my estimate. In order to avoid sounding unfair I would in no way refuse to use it again later if certain changes and improvements were made such as publishing an installable app and offering a perpetual license or the ability to buy credits as needed.

Would I recomment Midjourney?
The short answer is 'maybe'. If you're new to A.I Art and just want to try it out or you don't want to pay a monthly subscription you may want to try another program. If you don't mind its idiosyncrasies and specifically want to try it then try it. On the other hand if you don't mind subscription based services and want access to what I think is some of the highest quality A.I. Art out there then yes.

Nighcafe Creator

This is another A.I. Art service that's received a lot of press. My description will be very brief as I tried it only a few minutes. It makes very good art and is not limited to just one algorithm so you can choose a number of ways to generate your image. It also has a number of art styles as per other apps. It's only accessible on the web and is paywall limited to allowing only a few images a day. Fortunately you can earn or buy more without committing to a subscription. It has some interesting features like a running list of your work and the ability to go back and create edited versions later.

Pros:
High image quality: Is good at interpreting prompts and makes good quality images.
Ability to save as a basic feature. You don't need to subscribe to save your few daily images.

Easy: The site is easy to find and has a user friendly interface making it easy to use.

Buy or earn Credits: You can buy credit packs without subscription or earn them in numerous ways.
Daily Credits: Earn daily credits.
Numerous 'Diffusion' Choices: Choose from a number of algorithms to determine how the art is created.
'Evolve' an image: You have access to your past images and if you have credits you can 'evolve' or modify them. This is a unique feature among the software I've tried.

Cons:
Paywall limited access. This was my primary objection to this service. The number of daily credits you receive are limited. Credit packs are not cheap with plans starting at forty for around seven dollars or you can go 'pro' for a regular subscription.
Net Only: The site doesn't have an alp as far as I can tell which means having to open a browser to use it.
Slow: The process is a bit slower than others I've tried.

Would I recommend Nightcafe Creator?
I do recommend giving it a try at the least. The only real objection I see so far is the limitation with credits but fortunately the company seems to be quite generous with options to earn or buy them without subscribing. They are a bit costly.

Dream by Wombo

This was the third A.I. Art app I tried and the one I have the most time in as well as the first one I decided to 'keep'. Dream is available for download as an app and has a simple interface asking you to enter a prompt and then select one of a fairly generous number of art styles. Dream does not limit the number of images you can create per day and allows you to make images, save them and / or publish them to a gallery after making an account. The app works fast and outputs a single image. You can choose show the prompt you used or not when you publish it so others can see what you did.
Dream seems to excel at creating labscapes and natural scenery but most of the art styles seem to have difficulty rendering living things as well as things like objects and vehicles and has a particularly hard
time at faces. This of course depends muchly on the art style you
use and Wombo seems to be constantly improving their software. The app is particularly adept at natural images and landscapes and
llustration type art. There is a pro version available which allows a
number of features such as unlimited access to prompt history,
removal of ads, and an option to download animated versions of
your images. Some particularly interesting features afforded by
Wombo are an ability to order prints of your art from their site with
various options on what medium your art is printed on and how you
want it framed and start at about 44 dollars. Another very
interesting feature is the option to turn your art into NFTs if you
desire. Admittedly this is not a feature most people will use but it is
interesting to have the option none the less.

Pros:
Unlimited images per day.
Simple, user friendly interface
Basic level has most of the features unlocked.
Images are generated quickly.
Wombo frequently adds new features and improvements.
Publish your art to a gallery online
Buy prints of your art with various options.
Option to create NFTs of your art.
App has built in access to your gallery embedded in it.
Save your images as a homescreen backdrop for your device from inside the app.

Cons:
Results can be hit or miss
Certain subjects can be very garbled up-particularly bad at faces.
Most images have a slightly abstract look with most art styles. This along with the problem of rendering some things clearly seem to be the biggest drawbacks.
Pop-up asking you to subscribe shows up at random. Minor
problem at worst. It dismisses easily enough.
Pro level requires a regular subscription and pro features don't seem to be worth it in my estimation.
Images are saved in 'Trading Card' format with a frame and overlaid text which may be unpalatable to some people. Limits what you can do with the image without post processing.

Would I recomment Dream by Wombo?
Yes. I do recommend Dream especially for beginners. The fact that it allows unlimited images and has a generous helping of features at basic level make it a good app to learn the ropes of A.I. Art on. Its options to buy art prints and create NFTs make it particularly interesting.

Dall-E Mini and Craiyon

Dall-E is another highly talked about app which creates high quality images with a simple interface. Its successor Craiyon will be discussed in the same review since they're essentially the same. I'm given the impression -although admittedly I may be wrong- that Dall-E was only recently made available to the public. Both apps are available in various forms on the web and as apps. The version I use which is Dall-E Mini has a very minimal interface. It has some example artwork and a place to type your prompt and that's it. No additional options for art styles or anything and you largely control that with how you word the prompt. After tapping the 'Generate' command you're taken to another screen where the software generates the results. You will see the prompt and a counter showing time elapsed. When finished it outputs a series of 9 images which are all slightly different variations. Tapping on an image allows you to see an enlarged version and download it. There's no option to buy prints or publish and no access to prior prompts or results. You can download any or all of the 9 images you choose. Craiyon is very similar to Dall-E in that it has a nearly identical interface, processes images the same and has mostly the same features. You can save the entire grid of 9 images or save seperate ones. Dall-E and craiyon both seem especially good at rendering things other A.I. Art software have trouble with such as faces, people, animals and vehicles and they seem especially good at creating high accuracy representations of pop culture characters. Both are good at creating images in the style of art such as you'd find in card games. As with most of the A.I. Art apps I've used they have a problem with cutting off parts of images which are outside the bounds of the image from time to time. There is no limit to how many images you can create daily, nothing to subscribe to or pay for, etc.

Pros:
Unlimited image generation
High quality images
Nothing to pay for or subscribe to
Accesible both from the web and as an app.

Bare minimum user interface
Particularly adept at rendering accurate portrayals of people and animals
Makes images in the style of game art especially well
Dall-E Mini saves images which have no frame or additional overlay meaning they're more versatile and don't need post processing.

Cons:
Portions of images tend to be cut off.
Lack of 'Bells and Whistles' may be offputting to some people.
No option to publish artwork.
Craiyon saves images in 'trading card' format meaning you'll have to cut off part of the image to use it on anything you don't want the additional info on.

Would I reccomend Dall-E Mini / Craiyon?
I highly recommend these two apps. The minimalistic interface and lack of paywall plus the quality of images produced make them a very attractive option for beginners. Because Dall-E Mini downloads images in plain style without additional frames or text on them I recommend it over Craiyon. These apps would be good for people who create memes or want to use the art in games although I would definitely check the legality first.

Wonder

Wonder is a very aptly named piece of software. It has among the highest image quality of any of the programs I've used and in particular it excels at producing beautifully accurate high quality images of people and animals. It renders the faces most accurately of any app I've tried bar none, flawlessly in many instances in fact. The interface is much like that of Dream's with a field to enter a prompt at the top and a list of art styles at the bottom. In fact the two are so similar I have lost track in which I was working in! It has a number of art styles to choose from and I particularly like the results given by the 'pen and ink' style. The software outputs two images at once. The app allows you to save images with or without a frame and publish to a gallery after making an account. When publishing images it publishes both at once although only the thumbnail of the first one shows up in your gallery. Clicking on it lets you scroll back and forth between both. I did not notice an option to buy prints. You are limited to a certain number of free images a day, about twenty, but you have the option to go pro which unlocks unlimited daily images, faster image generation, more art styles and remove ads. In lieu of a regular subscription you may choose to pay about 25 dollars for a perpetual license which is an excellent option for people who eschew SAAS like myself. I consider it well worth it. If you want to try the full version before you pull the trigger on going pro you have the option to try it free for three days which I recommend.

Pros:
Second Highest quality artwork I've seen after Midjourney.
Absolutely the best I've used at making stunning high accuracy human images. Renders the most accurate and realistic faces of any I've tried bar none.
Good amount of features on basic level.
Outputs two images at once.
Allows you to save art with or without a frame.
Publish images to your gallery.
Images can be zoomed in to large size without suffering loss of detail. This makes them good candidates for use as framed art.
Is fairly good at making images of 'anthro' characters.
High fidelity at reproducing human images means they would make excellent character art for roleplaying games.
Allows a one time payment for perpetual license in lieu of an ongoing subscription. This is one of the most attractive features to me.

Cons:
Limited images daily unless you go pro.
Cuts off parts of images fairly regularly.
Somewhat slow. Pro speeds it up considerably.
Images sometimes have strange artifacts. Seems to be a phenomen in all of the software tried.
Did not see an option to buy prints. Could be there but overlooked.
Seems to do scenes rather than images of individual objects meaning it isn't so good for game item images.

Would I recommend Wonder?

I absolutely recommend this software. It is near the top if not my top recommended one. It is fast becomming one of my favorite A.I. Art apps. In fact it is the only A.I. Art program I've tried for which I'd recommend buying a pro license. Of that I recommend buying the $24.99 lifetime license as I think it would be well worth it. The app truly lives up to its name.

Conclusion
My experiences with A.I. Art have overall been overwhelmingly positive and I highly recommend anyone who enjoys digital art to try it. I want to state that I am in no way an expert and my opinions are just that- opinions. I've made every attempt to be honest and unbiased in the subject matter as well as accurate. I reiterate that I was not prompted by or received any remuneration from any of these companies in creating these reviews. Also note that as with anything in our world things are subject to change and if you read this at a much later date features, prices et cetera may have changed. Getting into this fascinating and rewarding phenomenon has certainly changed my outlook on it and I recommend anyone inclined to try it. Who knows? you may create something great.

View This Journal and 1 Comments

Statistics

Joined

2304
Pageviews
15
Followers
447
Favorites Given
44
Favorites Received
50
Submissions
12
Journals
26
Following

Shouts