For me it’s Text-to-Vector.
just got accepted into the beta late last night. going to play with it some today.
Can it be used standalone, or do you need a subscription to abobe creative suite?
How do I access Adobe Firefly?
Adobe Firefly is available as a standalone beta at firefly.adobe.com. The intent of this service is to get feedback. We invite you to request access to the beta and come play with this new model.2 days ago
nope, i was wrong. it’s completely in the browser and then you can download. it’s interesting that t hey described it as “in adobe apps” when you’re (at least for now) just generating JPGs.
i’m definitely fascinated by text to vector, extend image, sketch to image, and text to pattern. but none of those are available yet.
I won’t get excited. It’ll probably be amazing to only be cancelled after half a year.
A) trained on Adobe Stock and public domain (much more palatable to the lawyers I work with for clearance purposes)
B) text to brush
adobe rarely gets to beta without a full release. and their AI experience with sensei prepositions them as a strong player in this segment. i was just surprised it took as long as it did for them to release something.
I appreciate them waiting to announce more useful AI tools, than just releasing an image generating feature.
so, i played a tiny bit during a conference call.
this one was “Jesus takes a selfie with the disciples during the last supper” with settings at “photo” and “hyperrealistic.”
changing photo to “art” and “hyperrealistic” to “digital art” gave me this.
then art to “graphic” and digital art to “cartoon” gave me this.
getting more specific, i switched to photo, concept art, lighting as “golden hour” and composition as “narrow depth of field” and this was the result.
one interesting take by adobe, one of the four options for all of those was jesus as a women.
I was amused that the finger problem persisted - but this broke me!
this one wasn’t much better.
remember, firefly is less mature than midjourney, so it may take them a little time to fix weird stuff.
One pundit posits this release is a direct response to Canva taking some of their market share. The AI generation tools will supposedly be heavily featured in Adobe Express when they come out of beta.
Call me crazy, but isn’t that the “last knight” from Indiana Jones III photobombing in the background?
Speaking of Canva, The Verge had an article on them with an unfortunate clipping choice in the social media post:
Where do I sign up to be a
Here’s the full graphic:
And yet none of them showed Jesus as dark skinned.
Inherent bias in image sampling groups.
It’s one of the hidden dangers of AI, all biases in the data get carried into the output.
John Oliver recently did an episode about AI, it’s pretty good stuff.
I’m looking forward to some of the less obvious tools. The art pieces I’ve seen are interesting, but don’t hold a lot of appeal for me beyond just basic entertainment. I have my eye on Text to Brush, Sketch to Image, and Text to Template. That said, the Text effects look fun, too.
Interesting comparison to Midjourney.