right now its exactly the same as mid-journey in beta.. you just add a prompt and it spits out an image. Looking forward to them adding these features to real software.
Majority of AI art stuff is legal The moment you hit "Agree/Accept" when it comes to any type of program that involves data, you've sold the rights of that data to be used by any company, that includes art data All legal, all fair, all profitable :)
@@NumaPuma42 This has nothing to do with the grey area of the dataset used in those AI image generators such as Midjourney. They collect the images of artist all around the internet that didn't consent the use of their work. The data that you mentioned is the one of the people that are using the app
Seems like there is "big news" on a daily basis. Thank goodness there are people like you to keep us informed. Thank you. BTW, I tried to take a photo of a wheat field. It turned out grainy.
That lighthouse was a little more than just inpainting. It not only generated an object in the scene, but made it available on its own layer that could be independently manipulated in the full scene. This is a huge advancement!
Spitting facts right here. I wouldn't say it's the best though, if you've seen some of the stuff that people manage to get out of Stable Diffusion. This does seem very versatile though in comparison.
@@tantecosenelmondo2350 Stable diffusion is a image generation method. Midjourney is just stable diffusion tweaked. Actually, all these AI services are just stable diffusion with their own flavours. You are wasting money buying these services since the dataset is opensource and you can run an implementaion of stable diffusion locally anyway.
OMG, Firefly is doing what i was thinking about. For example. The separation of elements in the image and being able to extract them from the Generated image is as layers. and once we can do this and extract 3D Geometry as layers from a Generated Image. It's going to be a wrap!
Absolutely Dystopian, will prob incorporate it in my work regardless. Nobody that doesn't will stay relevant going forward. Delivers competent results in a fraction of the time but it's no longer our work, it was handed to us and we just signed off on it. We're no longer artists we're just some kind of directors. Creation has been outsourced to an algorithm, and it's fast becoming better than all of us.
Taste will eclipse skill in the coming days. It sucks but the wise thing to do is augment your workflow. A creative with AI can do much more than an average person.
@@Kritmus I don't think so. The main problem is time. Everyone wants the work to be done as fast as possible. Not to be done as good as possible. And yet there are still things the AI can't do. Like it's rather seldom that you need a generic photo of some shoes, or other products. Most of the time there will be a product that doesn't yet exist. The Ai cant help there, maybe only by giving compositional ideas, which would be fine since people always used reference and that's not much different. Yet for other creative work that doesn't need to be rushed out. I'm sure people can do better than those generic results. But only people with years of experience and a good working strategy. And well thats where AI can become dangerous. Cause many people will think.. "Heey? Why should I learn that stuff when AI can do it for me?" And as people stop learning they'll get useless and dump and then, hell yeah, AI will become better than humans
Not to mention all the future humans who will never know what it felt like to actually create something for themselves from scratch when we are already being told that it’s a waste of time when something else can do it so much faster, why even bother doing it yourself? Well even if it’s slower, I’d still rather do it the “old fashioned way” because at least it will truly be *mine* and not some machines’ design 😢
That's what they say, but in reality they trained it on all pictures on their customers Adobe Cloud accounts. According to their TOR that is legal, but according to pretty much anyone else is far worse than scraping off copyrighted published works on the internet.
I was thinking just yesterday that people wouldn’t need Photoshop anymore if our AI tool of choice could do everything so quickly. Plus it’s going to better at superimposing something into an existing picture. I am glad Adobe have brought this to market although I highly doubt it works as smoothly as the promo video suggests.
Firefly is going to be a lot of fun. As for those saying that Adobe CC subscription is “too expensive”, I assume that you’re too young to remember that Adobe programs use to cost between $600 - $1000 EACH. Then, every few years they’d upgrade with more great features and you’d buy those for $150-$200 EACH. There are 28 apps in Adobe CC. 12 of those are big design apps. Do the math. $60 per month isn’t “too expensive”.
The problem I've seen people complain about is that most folks who get the Suite don't use all the software, if you're a pro and use all of it, it's a pretty good deal especially if you're making bank from your craft. Personally, I only need 3 apps and honestly, I'd rather make a onetime payment for a license that I using those other apps. I've been looking at alternatives but there's definitely a reason why Adobe is the best, just not financially the best for casuals.
What attracts me of Firefly is that is the model. If it is all license no fear for professionals having issues with someone down the line accusing us of ripping an image or art style from another artist.
@@creativebubblecat People need to let go of the idea that their employment is dependent on a third party, whether AI is involved or not. Nowadays, it is dangerous to rely on companies for employment. Also, it's interesting how AI bros are getting so high in their horses today. They will lose badly to competitors who use AI but can also draw. They wouldn't be able to get employment in the professional industry as well. They need to be very naive to think anyone would hire them just type a bunch of random words in a textbox and call it a day.
@@apache937 yeah you can dream about UBI, it will never happen. Most country doesn’t even have free healthcare when they can afford it, because of the capitalist mentality, and you fool dreams about UBI. This is just fantasy in your mind, nothing even show a sign of the slightest advancement toward UBI in politics
As a major MJ user and paid subscriber, I've got to say this is FAR more superior! And as a user of all of Adobe's tools, I cannot WAIT until they release this.
The 3d modelling part is what really interests me, if it can create models that could then be 3d printed to play on tabletop that'd be huge... and probably the death of games workshop haha. Truly an industry shakeup moment.
This is pretty cool, and I would love to work with it. Two observations: 1. It's taken adobe years to get to this point. They don't move especially quickly. Yes, it's state of the art today, this week, maybe even this month. But most of this has already been done by stable diffusion. And nearly all of it can be done with creative applications of Controlnet and depth mapping. This isn't going to be at all new and shiny by the time it comes out. It'll look old, used, obsolete. Open source has a much bigger workforce on this. Changes happen on the scale of days and weeks rather than years. I think the big thing adobe's bringing to the table here is ui, which is something SD lacks. 2. The training situation here is questionable. Adobe has no opt out system for artists that contribute to adobe stock. This model is built on adobe stock images. They're introducing a do not train tag for artists to use... but it hasn't been implemented yet, and the people involved had no say on this first model. The do not train tag will be on a per image basis, and it's opt-in by default. Now, when all is said and done, I don't think that's a moral problem. But huge numbers of their commercial customers will take issue with this. We know how it's going to go down, we've seen companies try to do this before.
Valid points! What Adobe does bring, is a userbase already inside their world. If they can just "slap on" AI on top of that, it won't matter if it's cutting edge for many of the people working as long as it's good enough. But time will surely tell.
When you 'sell' your art to Adobe stock you're giving them a perpetual royalty-free license to reproduce and modify your work. That's a pretty blanket agreement which would undoubtedly cover the use-case of training an AI with your images. Even if AI could reproduce your work exactly (it almost assuredly cannot), that would be covered under the license you agreed to by selling your stock images to them. About the only use-case that would be protected under the license would be using it for illegal purposes or 'naughty' pictures, so you can be certain that the built-in Adobe AI has safeguards built in to avoid it being used for those purposes, much like Stable Diffusion 2.0 did.
I’m a traditional artist. I draw, write, paint, photo, video & some music. I also am an Ai Designer through Midjourney. An artist can & will create art with Ai. A non artist will just write prompts & now they use chat gpt to do that for them as well. Artists will never disappear.
you know, people who checked the stock options for this said they were never explicitly told abt it being used in AI and their is no option to op out of it at the moment? The company hasn't said anything about I either :)
no day passes without some news in AI art. things are moving crazy fast! Btw Sebastian did you see that you can use defusion controlnet to generate batch pictures and then create a clip out of that? early stages, but blows my mind.
I’m still quite torn about this. On one hand it’s great to be able to visualize a rough concept quickly, and I think ai tools are great if they help inspire artists work. However, this looks like it could replace the jobs of a lot of artists (including photoshop artists who are actually doing it by hand). Regardless of its legality, as an aspiring professional artist, this is a bit concerning.
Adobe thanks you, designers. Your $56/mo. has made it possible for them to build products that result in fewer jobs for you. Take the threat of AI seriously. It may initially free you up to do other task but it will eventually replace you. Ask any blue collar employee who was automated out of a job.
so so very excited to see these tools offered with this level of polish. I have 50k+ image generations under my belt in the generative AI world and these more sophisticated implementations designed for WORK are actually quite exciting to see. =)
The promising thing about Firefly is definitely integration with other Creative Cloud apps. In-painting with Photoshops selection tools and background replication stuff that they've been doing for years... Yeah this promises to be a huge leap in terms of control and actual real use cases for image geneartion.
Forget it. Their decades old software don't integrate well even to this day. The only "integration" they ever do is the ability to import PSD and that's it (you can't save, btw).
Well, if it's like a one time payment, it'll be interesting to see the price tag. However, i fear it to be a subscription service. One can hope though, really good ideas and implementations.
A subscription service works out cheaper than buying the product outright. The subscription also means you always have the latest version. Given how rapidly AI is advancing at the moment, you would want to make sure you have the latest version! I don’t know why everyone is complaining about the subscription cost.
@@travelwell6049 Because it's highly exploitative, you are basically just renting the programs and no matter even if you pay thousands in subscription over years you will never own the programs, the moment Adobe decide to pull the plug on a program/feature/...Etc there is nothing you could do about it, the moment you are unable to pay the sub (whose price they can hike as much as they want) you lose access to your entire program suite, it's disgusting, I'd take a one-time purchase program with less features over this exploitative crap any day.
We all knew something like that would happen. And doing it with "Ethical trained models" will help in leaving no more excuses to "real artists" to complain and AI will be seen more for what it is, a powerful and much needed creative tool. That said, with Adobe being Adobe, the real question is if they gonna clip the wings to SD plugins for photoshop that are already quite good, and assuming the AI runs in the cloud only, who will keep them in check with influencing future art by steering the models whatever direction they want.
All this talk of "A.I ethics" is nothing but vapid corporate talk to dissuade people from gravitating towards open source free A.I and keep them slaves to corporate subscription based services and their whims, how else will they make their billions.
Ai art is a fancy techs and all but calling it as powerful or better than human is a real insult to true creation itself as well as human and life. All that inspired us in characters, environemnt, design, catoons and everything all done by talented human craft. Real powerful art tool is all about complex control, knowledge of how things work and all that is human's intelligent strength that a simple text typing isn't powerful enough ti replicate. Thank you for understanding and respect that all the works done for us all is by real human.
Saw the video it's very interesting Adobe Firefly has amazing features that no other A.i has but I don't think is gonna be free but I hope other companies look at this as an example for future A.i applications
I knew it was coming. But to see it... I have so many conflicting feelings. Keep up the good videos. (PS: there's no "link below" to sign up for the beta unless I really missed it)
They may have a "100% legal dataset" but they retroactively added all images from Adobe Stock without any option to opt out for the creators and haven't paid those creators a dime yet for the (mostly unwelcome) addition to their licencing. There still currently is no option to opt out if you're an Adobe Stock contributor and don't want to participate on their ML training.
@@nodnoh-2174 to change something retroactively like this with no opt out option is actually a pretty significant change in policy and how their platform fundamentally functions. Btw, as of right now, there still is no opt out option, only Adobe vaguely saying that they're looking into a possibility of one, and by then, it'll be way too late for existing contributors. They've now launched what they call their beta of Firefly, but they've yet to even talk about compensation with it's contributors, other than "We're working on it". I can literally not find any mention of concrete info on compensation - not even in their latest FAQ on that specific subject. They promise nothing but the fact that your images are being, have been, and will continue to be used for this project. Adobe’s president of digital media, David Wadhwani, literally called Adobe Stock images "fuel for growth" in a recent interview with CNBC.
Looks cool, it's great that you can work with specific elements to tweak what you want. But yeah, there's no stylization and painterly stuff which artists need. This is mainly for UI/print etc designers -- which the other AI can't do. An interesting niche.
This is the one I'm talking about. This was from adobe max, somehow I can't find the adobe max original video. Anyway, this is what I'm waiting for :) czcams.com/video/4ooLcfLXGLA/video.html
that’s a good point. I was surprised to see that Adobe were happy to let AI generated Art to be sold on their stock site. It’s possible however that the AI was trained on a data set that excludes those particular images.
any graphic designers here? or artist? I'm not one..aspiring to be...but as GD or Artist how do you feel about the advancement of AI technology? Are you guys fearful or do you think moving forward your skills are still valuable and irreplaceable?
if ai were done right from start, meaning without taking copyrighted images, the mayhem would have been avoided and we would have welcomed it (just speaking to myself here). but it wasn't. there is a mix of fear, and from what i see, its less about jobs (well a little, but if ai can do shitty jobs for shitty companies that doesnt care, maybe its for the best. still shitty jobs were good for giving fresh illustrators experience, now they will have hard time getting it) but more on legitimacy on someone's pictures, and our respective expertise. there is something, ai can"t give, its experience in making images. in that regard, most ai users, you know the one who discovered visual art 6 month ago, are up to a great wall crash. because us regular artist will be forced to adapt, and when we will have even better product with better ui and such, our knowledge of composition, color theory, anatomy, perspective. and more important, our own style (something all ai users seems to be lacking, for a good reason, it takes time to know yourself enough to decide this is what your style is. for now there is only style decided on just prompt like "art by X",and not by having thoughts about what you are drawing and how you are drawing it) thats the real skill of a drawer here, you have to draw to know it and know why image a is better than image b. i dont have doubt some ai users will get past that wall, but they will be a minority. once regular artists will get "updated", our knowledge will still be our edge over early ai users. and the only thing taht will make the cut will still be the ability to spend 5 bucks on a pencil and printing paper. most affordable material in the world to begin drawing. with ai you need at least a decent computer, which is what? 500 dollars if you build it yourself?
@@keranao thank you very much for your lengthy response. To be honest I'm still at a beginner and started watching classes but somehow got demotivated because all this AI started popping out.
It's Adobe, so expect this to have an excellent UX but be prohibitively expensive. I do think it's overall a good thing because nobody was ever going to beat Adobe at UX design, but a lot of what they ultimately do will be able to be imitated in A1111 or elsewhere.
What’s with this idea that Adobe is prohibitively expensive? I’ve been using their full creative suite for years, and have never been able to not afford it. The cost of it is a drop in the ocean in terms of my overall income from the creative work I do. Their software is also remarkably cheap for what it does, you can expect to pay approx 300 dollars per month for a subscription to 3DS Max, for example…
Not sure how soon this will be released. Based on my experience with Adobe's neural filters, I think their solution will be limited in how you can use it and customize it.
I'm feeling the same way. It's as if the filters have become really neutered since I first started using Photoshop 6. I haven't found many/any uses for their neural filters yet, and really don't like the browser method. What I think Adobe will be a champion with is workflow integration and smoothness of use in regards to the toolsets they give you. Where they're going to prove inferior on is stability, connection speed and limited functionality unless they allow third party apps seamless inclusion. Not like the current SD 3PA.
@@MattLathrum All these image-generation products and services are just a reskin of each other. All the tech is the same and publicly available one, they just put an UI on it (that's why they are appearing so quickly). These AI services are just quick cash grab until people discover that almost everything about stable diffusion and even the newest and most specific features are all open source and you can run it locally.
its extremely limited, if it thinks you are generating too much skin it will not let you make an image. if adobe thinks they can limit the scope of your art based on some false moral standing they are going to fail.
I'm interested with the workflow when we provide the base theme and we can keep the concept around that theme without breaking the composition, for now I cant control the output unless we train our own stuff like what people do with the LORA etc. full control the output is always great addition in the AI tools. The brush is neat, ngnl.
I think the chances are high that training ML models on existing art is going to end up legally kosher. Not that you can't get similar results with just public stuff, as in Adobe's example.
all other models datasets are also legal, there is no grey area, it is legal to scrape data to train ML, just because someone doesn't like that doesn't make it illegal.
Yeah, and online identity theft wasn't illegal before we had internet... Laws have to and do adapt to emerging ways of people being dicks to each other.
@@alexandercoll1045 nah. Say you make skyscapers. It looks good, realistic. Whatever. Do you own skyscrapers now or do you own that angle now? Hell no. Get over it. Ai will do and take over many of these creative processes overtime. No one will do it manually if they dont have to. Its the future and it aint stoppin. Stop crying and get on the train or get left behind. Or get ran over by it..
@@JakeshandleYT AI is the future, I use various forms of it all the time, but we can't just ignore the rights of artists who put their heart and soul into their work. Laws have always evolved to keep up with new tech, so don't be surprised if things change to protect artists too. We gotta find a balance, you know? Let's respect everyone's hard work while still embracing the future. It's about making sure we're moving forward in a way that's maybe, hear me out, _ethical_... Also, you might not want to argue against consent, lol
Adobe...i really hope they will not plague AI with patents like they are ruining other domains. Also i have high doubt the actual use will product results like in this video so easily. I want AI art to be more accessible (the need for powerfull hardware is a problem) but adobe is plague that we should avoid at all cost.
You don't need powerful hardware. You don't need any of those AI services as well. Almost everything about stable diffusion is open source and you can easily run it locally with an OK computer. Those AI services are just a reskin of each other. That's why they are appearing so fast, they just put an UI on top of already public available tech.
As a small indie shop. This is great for me. But at same time is bad to. Some services i am selling to my clients they could now do themselves with this. Product designers and artists, because are some people specialized in that will be soon out of work. Wait until it can do proper car generation, integration and relight and all. I know many that will remain out of a job. Whole agencies. But we know that everything we see here is In The Ideal Conditions, and same as a selection brush, there will rarely be as easy as a one click fix. AI is here to stay and w have to adapt and learn to use it or we go obsolete, as it was with first generation that started to use PC's for 3d work and made practical FX dead (close to), or old school matte painters.
Does anyone know any models like Firefly's with entirely stock or public domain training? I'm trying to show my artist friends the benefits of AI but won't touch it due to ethical worries with the training
There are no ethical worries, it's nonsensical corporate talk to dissuade people from using open source software and keep them as slaves to corporate whims and subscription based services.
@@caterpillar4153 There are no copyrighted images used, that's not how training neural networks works, maybe read about it a bit before going with the dumb "ethical A.I" narrative being promoted by highly unethical corporations, think for a damn second why they are doing this and you will understand you have been fooled by utter nonsense.
Oh that's why my Adobe products are slow lately. They are using my cpu to test their Firefly...you know...when it's telling you that it is processing in the cloud...That's probably my computer 😂 That's OK, this is very promising. Thanks!
They accepted AI-generated images to upload in Adobe Stock in January. Accordingly, a bunch of Ai-generated images are already in the training data. They even collab with Kris Kashtanova. I don't think the "legal" is in Firefly.😒
Interesting if they did include AI generated images in their dataset. I would venture to guess that someone over at Adobe did take measures to prevent that.
@@sebastiankamph Your optimistic thinking of Adobe is admirable, but it's important to note that Adobe is not an AI company, so they probably do not have that kind of model to identify and may have measures in place to prevent copyright infringement, it's difficult to determine which generated images are copyright-free from what we know about the tech. Additionally, when Adobe accepts AI-generated images in their stock, it's unclear what kind of AI technology the uploader used to generate them. As I said, Kris Kashtanova, one of Adobe Firefly influencers, used Midjourney and Stable Diffusion to generate her work with prompts, and her work got into the stock. Unfortunately, both companies' training data are questionable, making Adobe want to promote Firefly with a "LEGAL" slogan. Nonetheless, it's an interesting topic, but I don't think their training data is "clean." I am not sure whether you got paid by Adobe to share this video, but other people do, so if you don't, you might write a letter to them.☺
@@huangyu-mingart I wasn't paid for this. Whenever I am paid for anything, I am super clear about that. And while I can show a sponsor and their messaging in my videos I never sell my opinions. Thank you for the elaborate comment.
I think the advantage that Firefly has is that it seems non-destructive. Even more so when it's implemented into Phototshop
right now its exactly the same as mid-journey in beta.. you just add a prompt and it spits out an image. Looking forward to them adding these features to real software.
Good to know that they start to move onto a more legal ground. Hope this will become the trend for all image generating AIs
Majority of AI art stuff is legal
The moment you hit "Agree/Accept" when it comes to any type of program that involves data, you've sold the rights of that data to be used by any company, that includes art data
All legal, all fair, all profitable :)
@@NumaPuma42 This has nothing to do with the grey area of the dataset used in those AI image generators such as Midjourney. They collect the images of artist all around the internet that didn't consent the use of their work. The data that you mentioned is the one of the people that are using the app
Seems like there is "big news" on a daily basis. Thank goodness there are people like you to keep us informed. Thank you. BTW, I tried to take a photo of a wheat field. It turned out grainy.
There's too much news. Want to take over for a week? You got the dadjokes up and running already so you're halfway there.
That lighthouse was a little more than just inpainting. It not only generated an object in the scene, but made it available on its own layer that could be independently manipulated in the full scene. This is a huge advancement!
Good catch Matt! Adding layers to the generative AI process will add so much depth to the workflow.
Other AI models lacked layers and transparency. Adobe is onto something big here.
@@KillFrenzy96 That's a common and old feature. It's just that all those AI services didn't bother giving access to it.
It pains me to say but this is by far the best AI gen. Still, they’ll never hook me into their subscription scheme.
Spitting facts right here. I wouldn't say it's the best though, if you've seen some of the stuff that people manage to get out of Stable Diffusion. This does seem very versatile though in comparison.
@Maurice Steiner with adobe nah
@@giusepperana6354 Stable diffusion is better than Midjourney and this? Really?
Stable diffusion is best,if u use it locally in ur computer, there is so many models and thousands of lora,on civitai .
@@tantecosenelmondo2350 Stable diffusion is a image generation method. Midjourney is just stable diffusion tweaked. Actually, all these AI services are just stable diffusion with their own flavours. You are wasting money buying these services since the dataset is opensource and you can run an implementaion of stable diffusion locally anyway.
OMG, Firefly is doing what i was thinking about.
For example.
The separation of elements in the image and being able to extract them from the Generated image is as layers.
and once we can do this and extract 3D Geometry as layers from a Generated Image. It's going to be a wrap!
The images didnt look great. Think i'll be sticking with MJ5
@@armondtanz Sure, their model is not that great at the moment. However Look at the potential
No, once the 3d geometry has optimized, clean topology is perfectly unwrapped, THEN it will be a.... ahem... wrap.
Absolutely Dystopian, will prob incorporate it in my work regardless. Nobody that doesn't will stay relevant going forward. Delivers competent results in a fraction of the time but it's no longer our work, it was handed to us and we just signed off on it. We're no longer artists we're just some kind of directors. Creation has been outsourced to an algorithm, and it's fast becoming better than all of us.
Yes, its terrifying
True. What about the fun of learning something new. Learning how to create something. The art of solving problems?! The art of craftsmanship?
Taste will eclipse skill in the coming days. It sucks but the wise thing to do is augment your workflow. A creative with AI can do much more than an average person.
@@Kritmus I don't think so. The main problem is time. Everyone wants the work to be done as fast as possible. Not to be done as good as possible. And yet there are still things the AI can't do. Like it's rather seldom that you need a generic photo of some shoes, or other products. Most of the time there will be a product that doesn't yet exist. The Ai cant help there, maybe only by giving compositional ideas, which would be fine since people always used reference and that's not much different. Yet for other creative work that doesn't need to be rushed out. I'm sure people can do better than those generic results. But only people with years of experience and a good working strategy. And well thats where AI can become dangerous. Cause many people will think.. "Heey? Why should I learn that stuff when AI can do it for me?"
And as people stop learning they'll get useless and dump and then, hell yeah, AI will become better than humans
Not to mention all the future humans who will never know what it felt like to actually create something for themselves from scratch when we are already being told that it’s a waste of time when something else can do it so much faster, why even bother doing it yourself? Well even if it’s slower, I’d still rather do it the “old fashioned way” because at least it will truly be *mine* and not some machines’ design 😢
as an artist im actually here for this, AI aiding professionals instead of directly competing with them
I despise Adobe but this is a great step forward for their customers and the field as a whole, looking forward to free alternatives.
11:42 simple model training. This is amazing. I hope many adopt this ❤
Agreed one of the best features.
@@dmarsub this was my first wish when i started playing with mj. Can i just name this something and keep using it.
I love how much control this tool provides, and I hope it's as easy to use as it looks!
Yeah, I like the controls and customizations. Could be useful in client work.
I hope this doesn't come at an extra cost for paid subscribers. This is the type of product I look forward to as a CC subscriber.
That's what they say, but in reality they trained it on all pictures on their customers Adobe Cloud accounts. According to their TOR that is legal, but according to pretty much anyone else is far worse than scraping off copyrighted published works on the internet.
I was thinking just yesterday that people wouldn’t need Photoshop anymore if our AI tool of choice could do everything so quickly. Plus it’s going to better at superimposing something into an existing picture.
I am glad Adobe have brought this to market although I highly doubt it works as smoothly as the promo video suggests.
A mix of marketing, promises and features surely.
Firefly is going to be a lot of fun.
As for those saying that Adobe CC subscription is “too expensive”, I assume that you’re too young to remember that Adobe programs use to cost between $600 - $1000 EACH.
Then, every few years they’d upgrade with more great features and you’d buy those for $150-$200 EACH.
There are 28 apps in Adobe CC. 12 of those are big design apps. Do the math. $60 per month isn’t “too expensive”.
The problem I've seen people complain about is that most folks who get the Suite don't use all the software, if you're a pro and use all of it, it's a pretty good deal especially if you're making bank from your craft. Personally, I only need 3 apps and honestly, I'd rather make a onetime payment for a license that I using those other apps.
I've been looking at alternatives but there's definitely a reason why Adobe is the best, just not financially the best for casuals.
What attracts me of Firefly is that is the model. If it is all license no fear for professionals having issues with someone down the line accusing us of ripping an image or art style from another artist.
Yeah, this removes the 1. argument the outdated artists keep spamming on twitter. not that I cared about licenses ever
@@creativebubblecat yes. good that no one needs to work . we will get ubi
@@creativebubblecat People need to let go of the idea that their employment is dependent on a third party, whether AI is involved or not. Nowadays, it is dangerous to rely on companies for employment. Also, it's interesting how AI bros are getting so high in their horses today. They will lose badly to competitors who use AI but can also draw. They wouldn't be able to get employment in the professional industry as well. They need to be very naive to think anyone would hire them just type a bunch of random words in a textbox and call it a day.
@@apache937 yeah you can dream about UBI, it will never happen. Most country doesn’t even have free healthcare when they can afford it, because of the capitalist mentality, and you fool dreams about UBI. This is just fantasy in your mind, nothing even show a sign of the slightest advancement toward UBI in politics
@@apache937 there will be no ubi and if there is it will be the most insidious form of slavery ever developed
As a major MJ user and paid subscriber, I've got to say this is FAR more superior! And as a user of all of Adobe's tools, I cannot WAIT until they release this.
I am a retired newspaper cartoonist… thanks Adobe for your programs that I used.
The 3d modelling part is what really interests me, if it can create models that could then be 3d printed to play on tabletop that'd be huge... and probably the death of games workshop haha.
Truly an industry shakeup moment.
IF it can do that i will have to just give up on my decision of not wanting to pay a monthly subscriptions and buy an adobe plan
This is pretty cool, and I would love to work with it.
Two observations:
1. It's taken adobe years to get to this point. They don't move especially quickly. Yes, it's state of the art today, this week, maybe even this month. But most of this has already been done by stable diffusion. And nearly all of it can be done with creative applications of Controlnet and depth mapping. This isn't going to be at all new and shiny by the time it comes out. It'll look old, used, obsolete. Open source has a much bigger workforce on this. Changes happen on the scale of days and weeks rather than years. I think the big thing adobe's bringing to the table here is ui, which is something SD lacks.
2. The training situation here is questionable. Adobe has no opt out system for artists that contribute to adobe stock. This model is built on adobe stock images. They're introducing a do not train tag for artists to use... but it hasn't been implemented yet, and the people involved had no say on this first model. The do not train tag will be on a per image basis, and it's opt-in by default. Now, when all is said and done, I don't think that's a moral problem. But huge numbers of their commercial customers will take issue with this. We know how it's going to go down, we've seen companies try to do this before.
Valid points! What Adobe does bring, is a userbase already inside their world. If they can just "slap on" AI on top of that, it won't matter if it's cutting edge for many of the people working as long as it's good enough. But time will surely tell.
@@sebastiankamph I want vector and 3d ai. I also want text to mesh, but I get why that's a big ask
When you 'sell' your art to Adobe stock you're giving them a perpetual royalty-free license to reproduce and modify your work. That's a pretty blanket agreement which would undoubtedly cover the use-case of training an AI with your images. Even if AI could reproduce your work exactly (it almost assuredly cannot), that would be covered under the license you agreed to by selling your stock images to them.
About the only use-case that would be protected under the license would be using it for illegal purposes or 'naughty' pictures, so you can be certain that the built-in Adobe AI has safeguards built in to avoid it being used for those purposes, much like Stable Diffusion 2.0 did.
I’m a traditional artist. I draw, write, paint, photo, video & some music. I also am an Ai Designer through Midjourney.
An artist can & will create art with Ai. A non artist will just write prompts & now they use chat gpt to do that for them as well.
Artists will never disappear.
you know, people who checked the stock options for this said they were never explicitly told abt it being used in AI and their is no option to op out of it at the moment? The company hasn't said anything about I either :)
A ai generator made from their own stock of images with a streamlined easy-to-use GUI? I'm sold.
Love that chill talking tone and such informative. and of course the dad joke too
Thank you for the kind words 😊
no day passes without some news in AI art. things are moving crazy fast! Btw Sebastian did you see that you can use defusion controlnet to generate batch pictures and then create a clip out of that? early stages, but blows my mind.
Would be nice with a couple days break from the news 😊😅
“Art”
AI "art"
I’m still quite torn about this. On one hand it’s great to be able to visualize a rough concept quickly, and I think ai tools are great if they help inspire artists work. However, this looks like it could replace the jobs of a lot of artists (including photoshop artists who are actually doing it by hand). Regardless of its legality, as an aspiring professional artist, this is a bit concerning.
you could say the same thing about millions of factory workers ran out of job after the automation chain was put into use
Adobe thanks you, designers. Your $56/mo. has made it possible for them to build products that result in fewer jobs for you. Take the threat of AI seriously. It may initially free you up to do other task but it will eventually replace you. Ask any blue collar employee who was automated out of a job.
wowww new tech information thanks for this I think adobe combine all generative art
so so very excited to see these tools offered with this level of polish. I have 50k+ image generations under my belt in the generative AI world and these more sophisticated implementations designed for WORK are actually quite exciting to see. =)
Great video thanks for letting us know. LIked!
Thanks for watching and your support!
The promising thing about Firefly is definitely integration with other Creative Cloud apps. In-painting with Photoshops selection tools and background replication stuff that they've been doing for years... Yeah this promises to be a huge leap in terms of control and actual real use cases for image geneartion.
Exactly! I'd love to get this into my regular day to day workflow.
Adobe is too big.
Their business model SUCKS!!!!!!
Adobe has TOO MUCH POWER IN THE INDUSTRY!!!!!!!!
Forget it. Their decades old software don't integrate well even to this day. The only "integration" they ever do is the ability to import PSD and that's it (you can't save, btw).
Long live GIMP lol
Dear Adobe, please give us artists the manual CC apps free if you make enough money with marketing agencies buying this AI.
Well, if it's like a one time payment, it'll be interesting to see the price tag. However, i fear it to be a subscription service.
One can hope though, really good ideas and implementations.
I can't imagine a thing that would be less likely to be a one time payment
OH, IT WILL BE!!!!
Adobe does nothing else.
IT WILL COST!!!!!
FOR RICH KIDS ONLY!!!!
A subscription service works out cheaper than buying the product outright. The subscription also means you always have the latest version. Given how rapidly AI is advancing at the moment, you would want to make sure you have the latest version!
I don’t know why everyone is complaining about the subscription cost.
@@travelwell6049 Because it's highly exploitative, you are basically just renting the programs and no matter even if you pay thousands in subscription over years you will never own the programs, the moment Adobe decide to pull the plug on a program/feature/...Etc there is nothing you could do about it, the moment you are unable to pay the sub (whose price they can hike as much as they want) you lose access to your entire program suite, it's disgusting, I'd take a one-time purchase program with less features over this exploitative crap any day.
We all knew something like that would happen. And doing it with "Ethical trained models" will help in leaving no more excuses to "real artists" to complain and AI will be seen more for what it is, a powerful and much needed creative tool. That said, with Adobe being Adobe, the real question is if they gonna clip the wings to SD plugins for photoshop that are already quite good, and assuming the AI runs in the cloud only, who will keep them in check with influencing future art by steering the models whatever direction they want.
All this talk of "A.I ethics" is nothing but vapid corporate talk to dissuade people from gravitating towards open source free A.I and keep them slaves to corporate subscription based services and their whims, how else will they make their billions.
Ai art is a fancy techs and all but calling it as powerful or better than human is a real insult to true creation itself as well as human and life.
All that inspired us in characters, environemnt, design, catoons and everything all done by talented human craft.
Real powerful art tool is all about complex control, knowledge of how things work and all that is human's intelligent strength that a simple text typing isn't powerful enough ti replicate. Thank you for understanding and respect that all the works done for us all is by real human.
I came for the AI, but stayed for the dad jokes 😂
Great vid as always Seb. Good to see more competitors throwing their hat into the arena. More competition the better.
Absolutely! And thank you for the encouraging words 😊
Adobe is NOT welcome. They are too greedy.
Great video , and hope to see more on the Firefly AI and all it's capabilities, thank you for the update :
7:28 I see that AI has mastered the standard SF/Fantasy trope of phallic towers.
😅
Saw the video it's very interesting Adobe Firefly has amazing features that no other A.i has but I don't think is gonna be free but I hope other companies look at this as an example for future A.i applications
Did they pay adobe stock sellers? Interesting how it works
Powerful 🎈🎈thank you for the info a insights
So neat with the UI and the workflow.
OMG FINALLY ARTIST REJOIS WE MADE SOME PROGRESS .... now we have to get royalties :[
“My headphones in a studio” is giving very “Emily in Paris” vibes lol
This is wild. Can you do a Q1 retrospective of advancements on generative AI?
If they stop releasing new tools! Which it seems they aren't doing 😅
i hope this program not becoming just a new excuse to inflate even more their subscription scheme price 💀
it looks so good to be real
Is this for offline use or do artists have to share their unfinished work with Adobe in order to apply AI effects?
I knew it was coming. But to see it... I have so many conflicting feelings. Keep up the good videos. (PS: there's no "link below" to sign up for the beta unless I really missed it)
My bad, I forgot! Thank you for reminding me, I'll fix it
They may have a "100% legal dataset" but they retroactively added all images from Adobe Stock without any option to opt out for the creators and haven't paid those creators a dime yet for the (mostly unwelcome) addition to their licencing. There still currently is no option to opt out if you're an Adobe Stock contributor and don't want to participate on their ML training.
That's too bad. That's what work for hire type work is.
@@nodnoh-2174 not when you're not told of, or able to provide consent for the full scope of use for your work - which is very much a part of it.
@@pipkin5287 That's already covered in most terms of agreement. AI is just a new implementation of transformative work.
@@nodnoh-2174 to change something retroactively like this with no opt out option is actually a pretty significant change in policy and how their platform fundamentally functions. Btw, as of right now, there still is no opt out option, only Adobe vaguely saying that they're looking into a possibility of one, and by then, it'll be way too late for existing contributors.
They've now launched what they call their beta of Firefly, but they've yet to even talk about compensation with it's contributors, other than "We're working on it". I can literally not find any mention of concrete info on compensation - not even in their latest FAQ on that specific subject. They promise nothing but the fact that your images are being, have been, and will continue to be used for this project. Adobe’s president of digital media, David Wadhwani, literally called Adobe Stock images "fuel for growth" in a recent interview with CNBC.
Couple months into ai and we already have the "green" option.
2015 was more than a couple months ago
@@WillakaPlaeground i was thinking midjourney
I have the Adobe CC. Can't wait to try this
thanks
Looks cool, it's great that you can work with specific elements to tweak what you want. But yeah, there's no stylization and painterly stuff which artists need. This is mainly for UI/print etc designers -- which the other AI can't do. An interesting niche.
“For professionals” it’s just the new way of saying: unnecessarily expensive
Creative Cloud needs to END! SaaS needs to end. Adobe is EVIL.
This is good. But what I am actually waiting for is that Photoshop + AI Generation feature they've announced back in AdobeMax.
There are actually plugins for that in the works. I'm alpha-testing one of them
This is the one I'm talking about. This was from adobe max, somehow I can't find the adobe max original video. Anyway, this is what I'm waiting for :) czcams.com/video/4ooLcfLXGLA/video.html
Aren't a lot of Adobe Stock images generated by AI now? Community submissions
that’s a good point. I was surprised to see that Adobe were happy to let AI generated Art to be sold on their stock site.
It’s possible however that the AI was trained on a data set that excludes those particular images.
homie cracked a joke in the middle that was unexpected LMFAO
I hope those in the data set were properly compensated.
I would totally combine this with Midjourney.
0:28 There might be a grey area actually depending on certain factors...
When such tools use animated video sequences : a new industry forms to parallel the Los Angeles film industry?
any graphic designers here? or artist? I'm not one..aspiring to be...but as GD or Artist how do you feel about the advancement of AI technology? Are you guys fearful or do you think moving forward your skills are still valuable and irreplaceable?
Don't do it. making art for money will ruin your creativity.
if ai were done right from start, meaning without taking copyrighted images, the mayhem would have been avoided and we would have welcomed it (just speaking to myself here). but it wasn't. there is a mix of fear, and from what i see, its less about jobs (well a little, but if ai can do shitty jobs for shitty companies that doesnt care, maybe its for the best. still shitty jobs were good for giving fresh illustrators experience, now they will have hard time getting it) but more on legitimacy on someone's pictures, and our respective expertise.
there is something, ai can"t give, its experience in making images. in that regard, most ai users, you know the one who discovered visual art 6 month ago, are up to a great wall crash. because us regular artist will be forced to adapt, and when we will have even better product with better ui and such, our knowledge of composition, color theory, anatomy, perspective.
and more important, our own style (something all ai users seems to be lacking, for a good reason, it takes time to know yourself enough to decide this is what your style is. for now there is only style decided on just prompt like "art by X",and not by having thoughts about what you are drawing and how you are drawing it) thats the real skill of a drawer here, you have to draw to know it and know why image a is better than image b.
i dont have doubt some ai users will get past that wall, but they will be a minority. once regular artists will get "updated", our knowledge will still be our edge over early ai users. and the only thing taht will make the cut will still be the ability to spend 5 bucks on a pencil and printing paper. most affordable material in the world to begin drawing. with ai you need at least a decent computer, which is what? 500 dollars if you build it yourself?
@@keranao thank you very much for your lengthy response. To be honest I'm still at a beginner and started watching classes but somehow got demotivated because all this AI started popping out.
Pretty amazing. Look forward to giving this a go.
this is very nice
It's Adobe, so expect this to have an excellent UX but be prohibitively expensive. I do think it's overall a good thing because nobody was ever going to beat Adobe at UX design, but a lot of what they ultimately do will be able to be imitated in A1111 or elsewhere.
What’s with this idea that Adobe is prohibitively expensive?
I’ve been using their full creative suite for years, and have never been able to not afford it. The cost of it is a drop in the ocean in terms of my overall income from the creative work I do.
Their software is also remarkably cheap for what it does, you can expect to pay approx 300 dollars per month for a subscription to 3DS Max, for example…
Not sure how soon this will be released. Based on my experience with Adobe's neural filters, I think their solution will be limited in how you can use it and customize it.
Possibly, but they've been looking at the competitors and probably realized that they must go all in on this.
I'm feeling the same way. It's as if the filters have become really neutered since I first started using Photoshop 6. I haven't found many/any uses for their neural filters yet, and really don't like the browser method.
What I think Adobe will be a champion with is workflow integration and smoothness of use in regards to the toolsets they give you. Where they're going to prove inferior on is stability, connection speed and limited functionality unless they allow third party apps seamless inclusion. Not like the current SD 3PA.
It's all about the market pressure and "land grab" mentality that major corporations are showing for AI. Then expect limitations later 😁
@@MattLathrum All these image-generation products and services are just a reskin of each other. All the tech is the same and publicly available one, they just put an UI on it (that's why they are appearing so quickly). These AI services are just quick cash grab until people discover that almost everything about stable diffusion and even the newest and most specific features are all open source and you can run it locally.
its extremely limited, if it thinks you are generating too much skin it will not let you make an image. if adobe thinks they can limit the scope of your art based on some false moral standing they are going to fail.
I'm interested with the workflow when we provide the base theme and we can keep the concept around that theme without breaking the composition, for now I cant control the output unless we train our own stuff like what people do with the LORA etc. full control the output is always great addition in the AI tools.
The brush is neat, ngnl.
Regarding creating vector art for banners, etc. Microsoft Designer is already doing this (Canva competitor, but with AI).
Sweet, good tip!
I think the chances are high that training ML models on existing art is going to end up legally kosher. Not that you can't get similar results with just public stuff, as in Adobe's example.
There's nothing illegal about using public images that I can right click save anyway.
well i dont give a f
all other models datasets are also legal, there is no grey area, it is legal to scrape data to train ML, just because someone doesn't like that doesn't make it illegal.
Yeah, and online identity theft wasn't illegal before we had internet... Laws have to and do adapt to emerging ways of people being dicks to each other.
True..it is like you can look at my art but cannot try to understand it.. like whaat??
@@alexandercoll1045 nah. Say you make skyscapers. It looks good, realistic. Whatever. Do you own skyscrapers now or do you own that angle now? Hell no. Get over it. Ai will do and take over many of these creative processes overtime. No one will do it manually if they dont have to. Its the future and it aint stoppin. Stop crying and get on the train or get left behind. Or get ran over by it..
scrape while you can
@@JakeshandleYT AI is the future, I use various forms of it all the time, but we can't just ignore the rights of artists who put their heart and soul into their work. Laws have always evolved to keep up with new tech, so don't be surprised if things change to protect artists too. We gotta find a balance, you know? Let's respect everyone's hard work while still embracing the future. It's about making sure we're moving forward in a way that's maybe, hear me out, _ethical_... Also, you might not want to argue against consent, lol
Adobe...i really hope they will not plague AI with patents like they are ruining other domains. Also i have high doubt the actual use will product results like in this video so easily. I want AI art to be more accessible (the need for powerfull hardware is a problem) but adobe is plague that we should avoid at all cost.
You don't need powerful hardware. You don't need any of those AI services as well. Almost everything about stable diffusion is open source and you can easily run it locally with an OK computer. Those AI services are just a reskin of each other. That's why they are appearing so fast, they just put an UI on top of already public available tech.
So you basically say that you want easy reproduction of Art to sell?
My job as a book cover designer keeps getting easier.
This is awesome! 🙌
The one that is trained on everything.....from Deviant Art to Iamag to Flickr to personal photos will be better......?
Time will tell, but MJ and SD who trained on everything surely have fantastic models.
haha I love that you pointed at Adobe and said "send me beta". How could they refuse that? lol.
Best implementation I’ve seen. Too bad it’s a live service for sure… expensive
we can finally zoom and enhance!!!!
The magic of movies is here
I love that adobe is being forced to innovate, or at least pretend to innovate...
They need to be forced out of business.
Wow artist 👏
hi I have purchased adobe bundle family package, if this firefly is is released in the future, we can download it and use it?
As a small indie shop. This is great for me. But at same time is bad to. Some services i am selling to my clients they could now do themselves with this. Product designers and artists, because are some people specialized in that will be soon out of work. Wait until it can do proper car generation, integration and relight and all. I know many that will remain out of a job. Whole agencies. But we know that everything we see here is In The Ideal Conditions, and same as a selection brush, there will rarely be as easy as a one click fix. AI is here to stay and w have to adapt and learn to use it or we go obsolete, as it was with first generation that started to use PC's for 3d work and made practical FX dead (close to), or old school matte painters.
Dad jokes 2.0: topical and relevant 🤩 (edit omg a second one) (edit third one). 😮😮
Excited but afraid. Will there be such skill as Photoshopping in future?
So me honing my photoshop skills for half a decade and being all proud of it... I can now just unlearn all that?
Does this mean anything this tool touches cant be copywritten?
😲😲😲Wow! This looks great! Thank you for the video.
Thanks for watching! 😊
@@sebastiankamph I'm loving Adobe Firefly! I can't wait to see what more they add to it.
Hey Sebastian, do you think you can make a weekly 20 min video highlighting AI news and stuff that happend each week?
Does anyone know any models like Firefly's with entirely stock or public domain training? I'm trying to show my artist friends the benefits of AI but won't touch it due to ethical worries with the training
ethics? get new friends
There are no ethical worries, it's nonsensical corporate talk to dissuade people from using open source software and keep them as slaves to corporate whims and subscription based services.
@@AscendantStoic There are ethical worries! The use of copyrighted images will have consequences in the future.
@@caterpillar4153 There are no copyrighted images used, that's not how training neural networks works, maybe read about it a bit before going with the dumb "ethical A.I" narrative being promoted by highly unethical corporations, think for a damn second why they are doing this and you will understand you have been fooled by utter nonsense.
Hello there :D
where s the link
Looks good. Does it also now include AI? Midjourney incorporated AI into it
what is art?
I came for the title and stayed because of the firefly joke
😘
hahah this will make me buy adobe cloud again. Looks really good!!
Adobe is betting on it! 😅
This is awesome to make stories for children ad an AI voice and pictures
Is it free to use?
We’ve finally arrived to sci fi future: “enhance image…enhance image.”
can you import your own 3D models for it to turn into an image?
Sure
Oh that's why my Adobe products are slow lately. They are using my cpu to test their Firefly...you know...when it's telling you that it is processing in the cloud...That's probably my computer 😂
That's OK, this is very promising. Thanks!
Haha, you might be on to something!
If that's what they are doing, then there needs to be a CLASS ACTION lawsuit against Adobe.
@@sebastiankamph not funny.
@@peterbelanger4094 I don't believe it to be true at all. I do believe blackvx joked about it though.
They accepted AI-generated images to upload in Adobe Stock in January. Accordingly, a bunch of Ai-generated images are already in the training data. They even collab with Kris Kashtanova. I don't think the "legal" is in Firefly.😒
Interesting if they did include AI generated images in their dataset. I would venture to guess that someone over at Adobe did take measures to prevent that.
@@sebastiankamph
Your optimistic thinking of Adobe is admirable, but it's important to note that Adobe is not an AI company, so they probably do not have that kind of model to identify and may have measures in place to prevent copyright infringement, it's difficult to determine which generated images are copyright-free from what we know about the tech.
Additionally, when Adobe accepts AI-generated images in their stock, it's unclear what kind of AI technology the uploader used to generate them. As I said, Kris Kashtanova, one of Adobe Firefly influencers, used Midjourney and Stable Diffusion to generate her work with prompts, and her work got into the stock. Unfortunately, both companies' training data are questionable, making Adobe want to promote Firefly with a "LEGAL" slogan. Nonetheless, it's an interesting topic, but I don't think their training data is "clean."
I am not sure whether you got paid by Adobe to share this video, but other people do, so if you don't, you might write a letter to them.☺
@@huangyu-mingart I wasn't paid for this. Whenever I am paid for anything, I am super clear about that. And while I can show a sponsor and their messaging in my videos I never sell my opinions. Thank you for the elaborate comment.
@@sebastiankamph No worries. I just think you are doing a great video, but if you are not interested in getting paid by them, then never mind.
Cheers!
AI is an “original sins”now?