Update 5/26: Adobe Firefly is now part of Photoshop Beta, which I tried out yesterday. The image generation quality still isn’t quite there, and Photoshop itself seems to be a little buggy, but the inpainting functionality that they now have is just awesome.
I just got my invite to the Adobe Firefly beta.
We don’t know what the pricing is going to be like yet and, during the beta, you’re not allowed to use the images commercially, so don’t cancel your Midjourney subscription just yet.
But, if you’re using Midjourney to create book covers and other commercial images, either for yourself or as a professional cover designer, here are four reasons you’re eventually going to want to switch.
- Adobe’s AI content is fully licensed. The AI is trained on images that Adobe has rights to and it will pay artists for the use of their styles. This means that you’ll be fully covered on both the ethical and the legal fronts.
- Midjourney’s best plan is the $30 per month deal, where you can get unlimited images once you switch into “relax” mode. It’s not clear what Adobe’s Firefly is going to cost, but Adobe Express costs $9.99 a month, so it will probably be somewhere in the same ballpark.
- Adobe Firefly has an excellent interface. Very easy to use, completely intuitive, and fully interactive — you switch styles and the images automatically change. No need to regenerate and wait. Oh, and it’s fast. Very fast.
- It doesn’t just generate images. It also does text effects, and, according to the preview video, will soon have inpainting, outpainting, AI-generated vector graphics, 3D images, AI-generated templates, and a ton of other artist-friendly features.
Watch a demo of Adobe Firefly in the video below:
Now, in the video above, the images are updated instantaneously. In real life, it takes a few seconds. But it’s still dramatically faster than most other platforms I’ve tried, including Midjourney.
So how good are the images? Pretty good. I signed up for the beta right after it was announced, and got my invite this morning, and I’ve been playing around with it throughout the day.
The quality of the images is not quite Midjourney level, but definitely high-end Stable Diffusion quality. And you can rate the images, so I expect that they’ll continue to improve as more people use the system.
Here’s another image I just created:
I can see using that on a book cover.
Plus, Adobe Firefly will create custom fonts for you, using AI.
Here are some that I just created:
You type in the text, then give it an AI prompt — such as “made of chocolate” or “deep space” — and it will create a custom font for you. What a fantastic way to create custom lettering for a book cover.
Of course, you can’t use Firefly for real yet, because the free, beta-release images aren’t licensed for commercial use — and they have the Adobe sticker on them.
My favorite part about Adobe Firefly is the community gallery.
Hover your mouse over any of the images and a “try prompt” button pops up. Press it, and it will take you to the image generation screen where you’ll see several generations of that same prompt — instantly. No need to mess around with the settings and configurations to figure out how the original artist was able to get that image.
For example, I liked the gallery image of a cute little fuzzy monster, and clicked on it and it immediately produced four similar results:
The prompt was simply “cute monster with a long tongue” — no special settings, filters, or prompt engineering required. I really like platforms that give you good images without you having to twist yourself inside out trying to craft the perfect prompt.
Again, this isn’t ready for prime time yet. It’s still in beta. But from what I’ve seen so far, it’s an excellent platform. Yes, all the features are already available in Stable Diffusion — but they’re extremely difficult to use, and there’s no mechanism to pay artists for their training data. And Midjourney might have better images, but the difference in quality isn’t that dramatic — but the difference in interface is huge. Midjourney’s interface — well, Midjourney doesn’t have an interface. Midjourney has a Discord server, which is just about the worst way to interact with an image generator imaginable. I can’t see how Midjourney will ever be able to add inpainting or outpainting, for example, with that interface. And, again, they don’t pay artists.
I’m definitely keeping an eye on Adobe Firefly and waiting for it to come out of beta. Meanwhile, check it out, play with the settings, and rate a few images.
MetaStellar editor and publisher Maria Korolov is a science fiction novelist, writing stories set in a future virtual world. And, during the day, she is an award-winning freelance technology journalist who covers artificial intelligence, cybersecurity and enterprise virtual reality. See her Amazon author page here and follow her on Twitter, Facebook, or LinkedIn, and check out her latest videos on the Maria Korolov YouTube channel. Email her at maria@metastellar.com. She is also the editor and publisher of Hypergrid Business, one of the top global sites covering virtual reality.
Thanks for this article—I am so glad to hear that more options are on the horizon that deal with the ethical issues. I couldn’t invest in MidJourney because of the annoyance of Discord. So I’ve been puttering around with Wombat’s Dream, which adds new features and filters constantly and is cheap. Then, I go in and tweak with my faves quite a bit—keeping in mind what I learned in the previous article about copyright of AI, and needing proof that you significantly changed the images. I appreciate this entire topic being covered so well here. I think, for me, it’s the universe noticing that I have ten kid’s books to publish and no longer have ability to spend hours stooped over a drawing. Now I can lean back and click. Make popcorn and click some more.