AI-generated images open up a whole new world. To cover it all, and even explain the basic technology, would require a much longer article than is reasonable, but we can certainly share some of the fun and help you get started (for free).
Chess is a tricky subject for AI-generated images, and even if you tell it you want a ♛ and it knows it’s chess, it’s quite capable of producing a human monarch with a crown on his head:
Not quite what I asked for…
First I tried a simple “Chess Olympiad in Budapest“But the results were off the mark, so I rejected this solution almost immediately:
In a way, this isn’t quite how I imagined a Chess Olympiad… (the less said, the better).
I then asked Dall-E 3 “Illustration of the Budapest Chess Olympiad, Olympic rings“and it began to produce fun and creative results:
The fun thing is, if you liked the results, but would like to see more samples, just ask the AI to try again, and it will produce more:
When you send a message to Dall-E 3gives you a choice of four results.
I then fed Feeda brand-new player on the market, released barely a month ago at the time of writing. It’s a powerful AI produced by a German company called Black Forest Labsand is behind the new AI running on Twitter/Xcalled Grok 2. However, while Grok 2 requires a paid subscription, Flux is open source, so you can install it on your PC and use it.
Flux is accessible from a variety of sites, but one of them is the simple Fluxpro.art. With exactly the same prompt mentioned above, we get this:
which is nice, but not really inspiring. So what can you do if you feel you lack the overflowing creativity to get better results? There are many ways of doing this, such as looking at other people’s images to see how they did it, which is all very well, but suppose you don’t want to go through all that and just want to get the image in question. The answer is ….: Ask the AI! Specifically, ask an AI like ChatGPT or Claude, and tell it that you want an image about the 2024 Chess Olympics in Budapest and that you’d like something fun and creative. Then copy this wall of text (it can be very wordy) and send it to the AI, and suddenly that same AI produces:
The possibilities are endless, and you can ask for cartoon versions of parts in athletes’ outfits ready to compete for an Olympic medal:
And much, much more!
Use AI yourself
There are a wide variety of amazing resources out there, almost all of which come with some form of subscription allowing you to use them in large quantities, but there are a few things to bear in mind:
- Every AI image is different, with different strengths and weaknesses. These differences can be quite pronounced, so I encourage you to experiment with them before taking the plunge.
- If you don’t intend to produce a lot of images, many of them allow you to generate a limited number of images for free each day, without watermarks or other quirks.
- If you have reasonable technical knowledge and a computer equipped with a modern GPU (video card), you might consider installing a resource such as Flux and remove these restrictions.
The following resources can be used free of charge for a limited number of times per day, although they all require an account with an e-mail address.
Slab-E 3
This veteran AI is still an amazing resource and particularly good at imitating artistic styles and illustration-type images.
Ideogram 2.0
With its new version 2.0, it can produce stunning works that rival those of a graphic designer when it comes to symbolic imagery. It is generally considered the queen of text.
Click here to access Ideogram 2.0
Imagen 3
This is the new AI produced by Deep Mind, the creators of Alpha Zero, and it’s available for free to US users (or anyone with a VPN or VPS). I found its ability to produce cartoon and comic book images nothing short of brilliant.
Flux
As mentioned above, Flux is available in several versions, for example via the web with a free account at Fluxpro.art (and no doubt others) or you can opt for the technical route and install it locally. One of the virtues of the latter choice is that the community behind it is huge and there are many specialized addons, known as LoRAs, that you can create or download to teach it to produce images in a particular style or on a particular subject.
Mind you, this is a fun and fascinating world, not without its addictive aspects, but it can also be both rewarding and productive. For example, I know more than one independent game developer who can’t afford to hire a professional artist and who has used AIs to develop elements for his work and make his dreams come true.
Links
