As a comics artist, one rap I've had against AI is its lack of consistency with regard to panel-by-panel continuity. Some of the examples you show at the very end would suggest that shortcoming may soon be addressed. Fortunately, I'm pretty old and will probably be dead before I lose any work over it.
Yeah, it's almost there in terms of continuity and consistency. That said, I think there will still be plenty of room for artist. Comics are such a unique medium and there is (as you know) a LOT of nuance in the visual storytelling. You could train a model on Wally Wood's 22 panels that always work-- but in terms of prompting? I don't think it would do a great job. At the end of the day, something like Krea will bridge a gap between thumbnailing and final pages. To be honest, I kind of see it as a real time saver for most comic artists...which, I mean your deadline margins are super tight to begin with!
Ha ha! I'm not old enough. The writing is SO on the wall. I just signed up at a technical college to get some real-world skills, hopefully a couple of years ahead of the pack. (Which is how I was successful in indie comics; I was basically 2 years earlier than all the competition, in both business development and skill refinement.) So now I'm gonna be an industrial electrical contractor! After making comics for more than half my life, I don't feel too bummed. I've seen and experienced all the best parts of comics, and right now, I derive more satisfaction from construction hobbies IRL than I do drawing. But man.., I feel bad for the young artists. When anyone can generate a great comic, (or a personalized Netflix), from their smartphone, art skills will be about as valuable as being good at lawn darts.
You're not going to lose any work because people are generally are disgusted when presented with ai generated images as "art". There is a demand of human skill and expirience behind a picture. They have to lie to sell us ai art, otherwise, noone cares. It's garbage
40 plus year artist, graphic designer. Currently CMO/CEO for a business owned by my wife and I am really excited about using this amazing tool to expand creative ideas, some of which lay dormant on archive files waiting to be repurposed!! Let's GOOOOO!
@@bellalawrencenye9793 Same thing was said when computers were able to use programs like AutoCad and Photoshop - did those that learned and used those programs ruin the industry? It's just another revolution in technology that will grab hold - you still need the creative outlet and creative chops to put it together and at this time, you still need to know programs to truly do something professional all the way. And artists will take hold of this tech and just become better at their workflow. Its called evolution in technology
StableDiffusion was doing more and better than that for half a year now. But most of it went under radar because free opensource projects like SD don't have great marketing or polished UIs
Perfect take. That’s exactly how I see it and I love Krea for this workflow. As I mentioned in the video, Martin is 40x the traditional artist I will ever be. And his outputs from Krea will reflect that. What’s great about Krea is that it inspires me to get better at traditional methods, so I can use it better!
@@TheoreticallyMedia this is a good thing, with how the usual AI Image generator software works there has been an increase of artist haters that think artists are/will be obsolete. Art creation is a skill like any other I wish these artist haters stop using untrue statements like "why would I learn art when the AI can do it 10x better than you can". Even picking prompts differs between photographers, artists and people that have no prior knowledge. To me such people don't value the craft and just want to play pretty image gambling.
Mind is officially blown. I have been struggling to finish my long form screenplay/graphic novel project for years since suffering a life threatening illness and injury. It looks like I'm gonna be able to do it now. Thanks man.
Depends on how you look at it. This does not devalue traditional art in any way. Zero effect on fine art-- as I can't see any of this printed on a canvas and being displayed at Moma. Commercial art will take a hit-- but I also argue that genAI can speed up an artist's workflow, allowing them to work on personal projects and pieces. From a film perspective: Storyboard artists will still be hired. Some will use GenAI, some won't...but someone still has to do that job. The Gen artists will just be able to do it faster. But-- again, someone will still need to be hired, because there will be notes to be addressed. There are always notes to be addressed.
@TheoreticallyMedia I get your point, but consider this. Why would a talented young art student invest in learning to paint well, when anyone can create a perfect oil painting in seconds. The perceived value of that talent hit the floor. Fine arts at the MoMA is not relevant. Only 0.001% of artists ever show there, and it has nothing to do with technical talent anyway. It's conceptual.
@@Reddkomet if there's a market that you can compete in, you don't have to be the best. But if an artist can make one painting per day, and an AI can make 1000 paintings at a better quality per day, the artist will not be able to compete.
@@nobody-u-know you could apply that to anything that’s manufactured. Should we ban factories as well? Ultimately it’s another artist using AI that is going to replace another artists, not a machine by itself.
Holy crap. Stuff like this gives me analysis paralysis. Limitations are what make me an artist. Give me unlimited power and I don't know what to do. I used to do animation by hand with cell drawing and a VHS camcorder, frame by frame. Once it because easy I lost interest. If just anyone can do it, I think it looses value, at least from the perspective of the artist.
So this is kind of funny: I spent a little time teaching high-school age kids Photoshop and video editing. A lot of them were into Digital Painting with Procreate. But they'd always want to (obviously) work on a tablet, instead of on a monitor. But what's funny is: When we went over to the (traditional) art area: They'd always be intimidated by large blank canvases. Like, they had no idea how to fill it up. It made me realize that working on those smaller screens, they could create an infinite image if they wanted to-- but only if the screen size was tablet sized. On a 5x5 canvas-- it was too much space!
I already do this using my own art ( watercolor and vector) as a base for AI ... LOVE the results I get as they are unique to my style and vision. I wish artists would experiment more with AI instead of hating on it so much!
I actually think there are a TON who are working and experimenting with AI-- they just aren't as vocal about it as those who are opposed to it. Watercolor and Vector sounds really cool!!
@@TheoreticallyMedia I'm new to vector, it's a mostly unconquered territory for me, but watercolor is my passion. And you are right, many artists are holding back and speaking up. more of us need to do so!
@@GringaCR Most of my paintings have been personal private commissions, but if you are new to watercolor I would seriously recommend splurging on the paper ( Arches ofc) and finding a good set of paints that suits your style. I have some Daniel Smith watercolors, but to be honest I always fall back to my White Nights St Petersburgh paints for 80% - 90 % of my works. I just love them, except for the blues , those I use mostly DS, as I find they have much better pigmentation. I have one set of squirrel paint brushes, which I mostly use for background blends and then WN brushes for everything else. I focus on floral and fantasy for the most part.
I agree!! Ironically, it's the short sighted who will miss out on how these tools can help innovate our craft. These same people would have protested the invention of the printing press back in the day.
To be honest, with this many images that it produces, doesn't this get tiring? Even just watching this stuff happen so fast makes my vision blur and my mind kind of zone out. It's not satisfactory in any way. One of the things I enjoy the most about art is the research and getting inspired by other artists I love, then applying it to my own designs. It's a careful trial and error that can't be replicated with AI where it constantly feeds you "amazing" stuff. Just some thoughts. It just feels like taking mediocre art to a "pro" level without all the knowlege that would come with actually getting to the AI image's level. Empty in a way.
I mean, one use case for you might be to do some quick perspective/layout images in a LCM like this, export that, and then working in a traditional method of top of it. If you play around with it, and give it some detailed prompts, it will often surprise you with visual connections you may not have thought of. Something like this could be more of a brainstorm session, than a finished product. Just something to consider.
So let's say you're painting in Krea and it's doing the realtime image generation thing... if it flashes an image that you really like, but then quickly changes to something else, is there a way to get back to a specific one of its previous iterations on your images? (If not I suppose you could record a screen capture video as you're going along)
So, that has happened a BUNCH of times to me...I'm like: "No wait!!!" Going to put that in as a suggestion to the devs. I'm not sure what the solve might be-- but they are also way smarter than I am.
Yeah, we‘ll have a capturing mechanism very soon. You should be able to grab any frame then. Consistency will also likely improve so that the whole thing feels a bit more like an animation, but that’s a bigger model architecture change -Titus
I see so many traditional industrial standard techniques being used here: digital photomanipulation, puppeteering, remixing, kitbashing, speed painting but on steroids! Except for detailing and refining. That's the work of the digital artist, for now.
Agreed! And glad you see it from this perspective! I think when you truly “get it” you see all the possibilities that AI opens up. I mean, even a traditional painter can scan an oil painting in, then animate it. It opens up a whole new market and area for them to explore, potentially even a whole other audience. All the meanwhile, it has not devalued the original painting.
@@TheoreticallyMedia Takes a lot of heart to convince these artists to take up AI tools. But I guess they will eventually. Many years ago, there was a traditional landscape artist who animated his Four Seasons art using VFX. It was such a rarity back then.
First of all great video. So well put together, but am I the only one that feels bad for artists? I am not an artist, but it seems like the equivalent of everyone going to a concerto where we watch a player piano instead of a pianist. Don't get me wrong. These tools are super cool and I love using them, but there is something bittersweet about looking at things now and not knowing if a human created it or not.
Nah, the savvy artists are quietly using this technology to their advantage. The naysayers are just a very loud minority. To use your example: I think this is like seeing a concert with a pianist who is accompanied by a player piano, allowing for a wider range of musical notes and tones. Executing things that a solo pianist could never do. In the early days of looping, I saw a Cellist do a set with a looping pedal and a number of guitar effects. Of course, purists hated it, while I went out and bought a looping pedal the next day!
Another superb video. What a tool.!! The faster latent Consistency models, are making this all so much more 'fun' as we get closer and closer to instant feedback. It's also great for learning tools like ComfyUI as you can now much more quickly understand what settings do what as you update your images and animation.
Hey, thank you so much! On the animation side, I can’t help but think (as the consistency improves) someone will add a keyframe/timeline feature to all this, and we’ll basically have an instant animation studio. Insane. It’s funny, the speed aspect: my musician brain turned on last night as I was futzing in Krea, and I realized this is a bit like “jamming” with another player. You play a few notes, they react, you react to that…and then things keep evolving from there. Also, I really need a cheap Wacom now!
@@TheoreticallyMedia . Yeah absolutely. I'm now regretting selling my drawing monitor. I really like the "jamming" comparison. I can 100% see this working with a timeline. Just testing my 'work in progress' Controlnet Rig After Effects plugin, and the potential for realtime feedback directly in After Effects as you animate the rig. Set keyframes for text prompts along a timeline) and build up the scene. It really feels like it's coming to the point where 2023 was the year the tools came out in early versions and 2024 will be the year this all starts to really bare fruit and have use cases in more commercial projects.
As much as it will level the playing field and quicken art generation process - it will homogenize art industry as modern DAWs did with electronic music industry: everybody and their mother can create electronic music now, so every electronic music producer sounds like any other. There are so much junk music rn that it's incredibly hard to find anything worth your attention. tldr most likely with AI uprising we will lose our last shreds of creativeness and most art will be some regurgitated parts of previously created art pieces bend by the angle of crudely made strokes in some random paint app
I mean, obviously with the guitars on the wall, this comment hits for me. I think home recording/DAWs have this double edged sword of democratizing and homogenizing. You are of course totally right: A lot of Music has gotten super bland, but there were a number of other factors there: like everyone using the same stupid Splice Loops, and the playlist culture that Spotify wrought. But, it has also given us hundreds of artist that I’ve discovered in all kinds of genres that I never would have run across. Granted, I’ve got some weird music tastes, so these are artists/groups that never would have had mainstream success to begin with. But I’ll say, the savviest of those artists have also figured out how to additionally support themselves around their art. Merch, Patreon, etc… It’s funny, and this is a conversation I can easily go for hours on, but the majority of musicians aren’t overly concerned about AI. Most of them see it and think “how can I use this to my benefit” To be honest, I think that’s likely a learned skill from a lifetime of getting screwed over. Musicians have always had to be a bit scrappy to make ends meet, and they’ve gotten pretty good over the centuries at rolling with the punches!
@@TheoreticallyMedia Whoa, didn't expect such an elaborate answer, well thank you for your attention. I feel the same regarding musicians and situation as a whole. Though to be fair advances AI has made recently in visual design and art is a bit stronger compared to music generation. And considering same sampling and DAW influence on music currently mentioned above it's already quite frequent that you can't really tell if it was AI generated or it war created by human hand. So I feel the same may happen with visual art, thus my previous reaction
As an artist, I'm stoked for this! It looks like a lot of fun. Also, you have solid in depth no-nonsense videos and that is a gem among youtube videos! Thanks again man. Did you ever see Tafi's AI and if you did, did you get into the beta?
How Does it look like lots of fun, that people Take screenshots of animations, or steal many drawings from a singe atrist, to copy them even more direcly...?
@@julle4083 Tafi is the company who makes Daz Studio, a 3d character creator program that's been around for idk like 2 decades or more. They basically took their lessons and mutability and turned it into an AI driven text prompt to 3d character generator.
Wow, these are some huge development in AI artwork. Thank you for sharing this, Tim. I am an academic and am always looking at different ways that AI can be used to help with education. These development have academic implications as well. Keep up the great work.
that's awesome! and thank you! You've got me thinking now-- it's a great way to teach various movements and periods of art. Have a student draw something and then prompt for: Renaissance, then Cubist, then Art Nouveau-- It'd be an interesting way of visualizing-- particularly to the student, as they'd feel a connection to each era, since it is their input image!
@@milionSTSocrates said the same about reading books - he didn't believe students should be taught to read and write. For the most part, he was right. But the ways he was wrong was well worth the loss of debate, discussion, and raw experience. So: you're right. Sort of. People may lose the ability to draw, but not the ability to make art.
@@milionSTalso, i love making clothes, but nobody makes their clothes anymore. So. Do mass produced clothes make us lazy? Not really. We spend our time on higher levels of work and creation. And those skills are preserved by the few of us who love work and process and creating.
This new feature in KREA , As a compositional tool, is SUPER DOPE, the image input feature to abstract elements to a new image is like WHOA. Great tool so far, if you screen record, or better still it has a Timelapse record like procreate, then this would be an experimental animation tool. If I create a short film with stock video, then I suppose I could create an animation from the output , I’ve always liked KREA for illustration ……External screen : MIND BL🤯WN… Great overview Tim
Thank you so much!! And yup: there’s another unthought of use case! LCMs in general are really going to change the way we generate from here on out! I’ve seen the animate tool in Procreate, but I haven’t played with it yet- but, 100% if it works the way I think it does: we’ve got an animation machine on our hands here!
@@TheoreticallyMedia I like it when a developer from the beginning creates tools that are open to 3rd party applications being an option for expanding one’s creativity and work flow. Applications like this even inspire other developers in their own lane to consider other use cases that can sharpen their tools pallets . I could definitely see this company developing a range of realtime plugins for post production image processing too. This is almost like the choices you get with using a modular synth in comparison to a regular synth, or signal processing sound… this is sparking so many ideas , like for exhibition space and interactive installations like the van gough / Dali immersive exhibitions… I’m really looking forward to this. Once again thanks for keeping us on point with top choice creative innovation.
I actually tried Ever Art AI . I was satisfied with the results I got . I was about to subscribe , but I found their subscription plan too pricey , for their unlimited training , and generative image . Their first two tiers on the other hand , was nice , but you're only allowed a certain number art to generate per month .
It occurs to me that we need one with an alpha background option and a no foreground option. These would allow layering in photoshop or gimp. It would also remove some of the "your stealing my art" arguments. I find I'm having to erase a lot of things in gimp to clean up the product.
I agree- there’s almost a threefold workflow here: generate, futz, then post-cleanup. The idea of an alpha background would be great. It would also be cool if you could lock certain areas of the output, since it does tend to generate pretty wildly with the smallest of changes. But again, this tech is like 2 weeks old, so we can give it some time to mature!
@@TheoreticallyMedia Yep its getting hard to keep up with the innovation and my cash flow is zero so I can't subscribe to most. I suspect an AI heal/ detail 'brush' will come to gimp and paint apps soon to further build up the innovations. I need a much bigger computer already.
Ha! To be honest, I’ve still got a crate of them in my basement. In those early years I couldn’t bring myself to throw them away. Of course, these days I’m yelling at them to stop stealing all my printer paper to draw on! (As a note: they HAVE sketchbooks i buy them all the time. But apparently stealing my printer paper is a source of joy to them…)
0:36 Krea looks very impressive, but it would be amazing if it could be thoroughly trained on custom characters. I wish I could use it to depict my custom characters from every angle possible. I need to pose my custom characters in different scenes. I also often need to combine multiple custom characters in a scene. It seems like Krea would be amazing for scene creation, and the only thing missing is the ability to accurately & consistently add custom characters to scenes.
Are you familiar with DAZ 3D? It's a figure customization tool used in animation, games, and concept work. I'm already using it to create and pose specific characters, with unique features and expressions. I use those for reference and consistency, placing them into Clip Studio to ink.
As mentioned in a previous post, a 40-plus-year artist, and graphic designer. Currently CMO/CEO for a business owned by my wife and I. I signed up for the "Creator" account. Getting a lot out of turning my creations into models. However, I'm finding it challenging to purchase more tokens, even though my subscription states I can "purchase more at any time". Has anyone run into the issue of not being able to purchase more tokens? Thank you.
@@TheoreticallyMedia great idea! I'll do just that. I sent an email. I'm working on a project and need to purchase more tokens! Appreciate your content! Gulf Shores Alabama!
they for sure have added stuff since this video was published-- but if anything, I'm seeing more options? Maybe try a different browser? That's weird...
I'm signed up for Krea beta. I cannot wait to get in and connect it to Blender for rapidly putting together scenes. I was thinking about some pretty interesting use cases. - For example, using primitives to create an output, then render the scene and use them both as image prompts to finalize the scene. So very exciting! Also looking forward to hopefully getting into EverArt and using my stuff to train it for consistency. Everyone who is interested in creating digital art just got their minimum quality leveled up in a huge way with these tools. Thanks for the video Tim!
I mean, to a degree it’s control net- but wayyyy faster. Stable Diffusion just release SD Turbo, taking steps from 50 to 1! Insane! I think 24 will be (among many other breakthroughs) the year of real time.
You don't have the Krea link in your bio - FYI in case it was an affiliate link or similar. This was AMAZING - I can't believe this video is a month old!
You should develop your "Henchmen" idea further. It's a solid idea. You should consider seeking funding to turn this into a TV series or a movie. I'll bet you could even convince Apple or Amazon to fund it.
So funny enough, the whole 12 issue script was written, and it did get optioned for a TV show at one point- but that whole game is basically a lottery ticket. We got shopped, but never picked up. But I totally agree, I think it would have made an awesome TV show. With IP, it’s always weird. I could get a call tomorrow about it and…well, it’ll still probably not get sold, haha.
@@TheoreticallyMedia Cool, I just signed up for the waiting list (hope I will get access soon). I want to try it with messy traditional materials such as charcoal and wax stick and see what I can tune it to and also geometrical City type shapes in Blender.
@@jamesabell9494 omg, that is such an amazing idea. I’ll say, talking with Krea yesterday, they’re really excited to see what insane things we’re all going to do with it. They’re as excited as we are!
I do a lot of expressive drawing so it would be great if I can someway tweak it so it resembles and evolves the lines etc will experiment with it asap when I get access!@@TheoreticallyMedia
Ooh, you could use this to get some custom details, then put it through Leonardo AI with the Img2Img sliders set at different rates to see what comes out.
I’m just sad to see the job market evaporate for trade artists and myself (20s) just trying to get into the market for graphic design and art only for it to literally be erased before my own eyes.
Nah, you'll be fine. The truth is, technology like this still requires artists to manipulate it, and...well, to be honest, address notes. Because there are always notes. Build relationships with other artists, clients, co-workers...basically everyone you run across professionally. You'll get jobs. Any company that is relying on AI Generated Images whole cloth is not a company that you'll want to work for anyhow.
Thanks for your content! Is it possible for you to maintain and regularly update a list of AI media tech? For your patron members? Would help a lot as the landscape changes quickly and would love to have a quick reference resource.
In limbo. Still hasn’t really been stress tested yet. Particularly with a tool like LCMs. But seriously, don’t stress about copyright. Internationally, many countries have already approved it. The US will hem and haw, but I think it’s a done deal at this point.
I can’t thumbs up this enough. Of course, as a guy who obsessively follows all of it, it also gets super overwhelming from time to time. But yeah, literally i wake up everyday thinking: how did the world change today? It’s amazing
@@TheoreticallyMedia I read the book 'Dark Matter' this week. It kept me thinking about how many possibilities AI development can create, and how each little advance in one area can result in a completely different world. Mind blowing haha. Thanks for your work! 🤝
@@aiforgemaster oh awesome! I love Blake Crouch! Recursion is really good too! And thanks for reminding me, he’s got a new one out that I forgot to read!
@@TheoreticallyMedia it used to be that ideas I had about the future would take 20 years to become realized by others. Now, ideas I have about the future become realized in less than a week!
Oh, that would be amazing. I love watching Z-Brush timelapses-- that is one piece of software my brain is not built for, but I am always amazed watching the wizards who do use it!
I don't quite understand the nuances of how training works. Without anyone explicitly telling it that a boxing ring is called a boxing ring, how is it able to draw one from a prompt asking for one? Is the training completed from the cartoon strip just an additional layer on top of a much denser training image set?
You had me laughing so hard at "scandalous Godzilla".. omg But wow - this tool, man I don't know what to say aout where all this is headed. A professional Artist here.. semi retired.. but truly bewildered.
i think ill ask the ai to create a browser addon which automatically filters and hides videos with clickbait and sensational titles, that would change everything for a change lmao
Thank you! I have a question: is it possible to upload a photo and turn it into a semi-impasto oil painting where the brushstrokes and the paint actually looks realistic? I still haven't found an AI tool that can do this in a satisfactory way, including photoshop's generative fill function.
Well, to be fair, with the Henchmen comic pages, I do own the copyright to those. But, to your overall point: There is sure to be a lot of saber rattling over the copyright issue, and a lot of lawyers are going to make a lot of money off of it. But, at the end of the day, it looks like because the material is being trained, and not reproducing 1:1 copies-- it'll land in favor of AI. The US copyright office recently decided you can't copyright AI generated material, without a human doing some kind of post work on it. But, that's a tricky rabbit hole-- as what constitutes post process work? If I put a filter on it? Or if I wholly repaint the image by hand? How do you quantify that? And hell-- how do you even prove it? If I send a Midjourney Image to the Copyright Office and don't declare that it is AI generated, how can anyone prove otherwise? It's going to be a mess...and ultimately, all of this only applies to US Law. Japan and many other countries have already OK'd AI generations...so, it's pretty much a done deal.
Good question. I’ll reach out to find out. I think it likely ends up feeding their training data. I mean, it’s the smart play on their side. I’m never too concerned about that, only because any image or set is just a blip in an ocean of data.
LCMs highlight a problem with services like Krea. This sort of real time feedback is a million times better running locally on your own machine, not only because of the low latency, but because of the ability to use your own models. Models that you might not want to share publicly.
Probably just going to have to try myself, but when you undo something in that Krea - that is e.g. erase something you draw - does the generated image also go back to the previous generation, or does it make something completely new?
Playing around with Krea I saw a few (artistic) nsfw generations pass by. Nothing too gratuitous, but it was there. And I mean, did you see the Godzilla image? Haha
You are my go-to man on the latest advances in AI art. Your demos are a revelation and inspiration. You said AI art just changed forever. Looks like I'm also changed forever. Can't wait to try them out.
So-- sort of. I'll have to try it out, but it generally alters the image somewhat. See the Pirate example to see what I mean. That said, you can then "trace" your own drawing after that to likely get something super close. But, it isn't 1:1.
Wow, this is really going to help out small businesses. I don't spend crazy amounts on freelancers but it adds up annually and any savings help. Just signed up, can't wait to experiment
Such an insane level of control. And once things like training models are applied to it? I mean- there isn’t going to be any image that can’t be endlessly micromanaged! Haha
I just paid a stupid amount for some radiator work in my house. The part was $7. Luckily, the guy was cool enough to show me how to do it, so next time around I can save some money.
To be honest, the legality on all this stuff is so in flux, pay it no mind. Just make stuff-- lawyers (on both sides of the argument) are going to get rich battling it all out. If you make something that is generating enough money that someone calls on you-- well, you've made enough money that you've probably got a whole set of more important problems anyhow!
Can’t disagree. And again, I think it comes to a matter of “the more you bring, the better you’ll get” You might see it also a a weird way for folks who have been promoting to get back into traditional art. The thing about AI art is: I don’t think anyone who spends a ton of time experimenting with it has no artistic background. They all come from somewhere.
Crypto bros /Nft bros and AI generative content prompters gross me out. Ai will never take away the ability for artists to draw but will steal their livelyhood after sucking them dry to be their own competition. We cannot stop the march of technology the same way no one could stop cars from replacing horses or robots replacing factory workers and leaving them to die. Society has been treating artists and artistic fields as afterthoughts or as "not true careers" for years while silmultaneously being unable to live in a artless environment. AI art is the nail in the coffin imo. Studios will use this to replace artists. Ai is a tool that could be amazing but if you believe for one second the AI bros will use it as intended you are dellusional. AI generated content is ground for abuse and when you will have destroyed the humanity out of art you'll look back in shame. If you lots even have that sort of empathy or self awareness to begin with. All these "40 years old in the art industry" blokes in this comment section remind me of naive boomers amazed by technology they didn't have in their childhood and not fully grasping the consequences of what they are fawning over.
I mean. You’ve pretty much got everything nailed down here. Everything about your post is correct. There will/is a gold rush of NFT/Web3 Chads making a run right now (I get emails asking me to sponsor their startups all the time. I turn them down) Studios will for sure, attempt to abuse the technology the moment it hits a technical level, despite their promises not to. That is exactly what the SAG/WGA strikes were about (well, that and streaming money) Artists have always been devalued through history. I’m pretty sure the first caveman to do a wall painting was asked: “you gonna do that all day, or maybe go help us hunt for food?” That said, I do have discussions all the time with people in the VFX and Post industries, and they aren’t the people you might think they are. They’re very thoughtful and aware of the double edged sword this technology is. I talk with those “40 year old” Boomers, who are excited, because they see the prospect of being about to create a movie or a comic that they’ve carried in a dream since they were a kid. I don’t know, the whole “AI has no Soul” argument reminds me of old “Drum Machines have no soul” bumper stickers.
I'll say that while it was generating, I had to carefully check my footage-- There for sure was some...scandalous generations that passed by. Even more scandalous than that Godzilla image!
@TheoreticallyMedia Interesting. That's one of my biggest gripes with AI generators......most tend to be hyper restrictive - overly censoring. Sure, in some cases with clever prompt engineering one can sort of circumvent some of that censoring, but still, I truly hate the overly abusive NSFW restrictions seen with platforms like MJ and many others.
@TheoreticallyMedia As a fan of Frazetta, for example, so much of his subject matter cannot be simulated via most AI art generators. Imagine how quickly your prompt would be rejected trying to describe a warrior standing over a pile of steaming bloody dead human bodies! 😉
@@LouisGedo haha. Frank also loved his scantily clad (or not clad) ladies as well! Love everything about his work- For sure Midjourney doesn’t allow it, and awhile back I tried name calling him and that was a no-go. It wasn’t even close. Weirdly, I’ve been trying Conan prompts as well (cinematic) and MJ steadfastly refuses those too. I mean, I get stuff- but it isn’t very good. That said, I’ll give it a shot in Everart (the second platform discussed) and see how it goes!
If Krea would give us the option to use our own gpu's it would let them scale quicker. i like when we are allowed to run ai apps on our own machines because of this issue with developers needing to spend $$'s to scale their services.
The underlying technology (LCMs) are open source, so technically you can run it yourself. Although, it’s only about a week old, and from what I’ve heard, configuring is a bit of a technical challenge, much less being able to screen share etc, If you’re up for the challenge, there should be a GitHub around to do it. You will be far braver than I!
I love this and I cant stop using it to be honest but, is it me or you all get japanese people? I mean, not always, but I would say.. 80% of the times I drew people. Not complaining thou, just curious
Well, I think there is a fairly narrow range of things this one has been trained on. I do notice that as well. Next week, I’ll have a new version of the tech that’ll turn some heads! Gotta test it, but it looks super promising!
What are the other guys you recommend following on RUclips concerning the exponential use of AI within digital art? I think you mentioned someone called Martin? And then there were some blender guys - do you have their names or channels somewhere so I can follow them too? And thanks for all your work and information its very eye opening
Super cool, I think for skilled artists it could be an interesting tool to experiment with. I wonder if these models are also trained on copyrighted works? Or if users will be able to copyright their outputs
What are you saying? Of course it's trained on copyrighted materials as any generative AI machine. And you will never have any copiright on something that is not done by human. And you also mentioned skiller artists? Are you aware that skilled artists have zero interest on that?
Haha, yup! Our old Dutch Footie Pirate gave us the answer! It’s funny, I can’t help but think one day someone is going to show Daniëlle van de Donk one of these videos, and she’s going to be all, WTH?!
I mean, look, art is never the safest career path in the best of times. I’ve pretty much made a living with a camera, music, or writing for the last 20 years. There have been years were it was LEAN. I was the definition of starving artist. There was a point where McDonalds had hamburgers that were 2 for a dollar. I bought 40 of them and put them in a freezer. (I’ll never eat one again) Making a living in art is never based in talent and skill. It’s about meeting and connecting with people. Networking with others in your field and rising as they grow, and then offering a hand to those coming up under you. If you’re truly passionate about making art; whatever the medium, and you’re willing to get punched in the nose, everyday for ten years, and never give up. Then yes. Do it. AI isn’t going to stop you. Might even make your path a little easier.
Currently starving from the sudden reduction of jobs and lost some oppurtunities in this industry over some AI. That's years of my life wasted and having my wrist damaged from doing studies and work at the same time to somehow survive. If you still got hopes and dreams in this career just forget it. it's an industry, in the end it's all about the money.
As a comics artist, one rap I've had against AI is its lack of consistency with regard to panel-by-panel continuity. Some of the examples you show at the very end would suggest that shortcoming may soon be addressed. Fortunately, I'm pretty old and will probably be dead before I lose any work over it.
Yeah, it's almost there in terms of continuity and consistency. That said, I think there will still be plenty of room for artist. Comics are such a unique medium and there is (as you know) a LOT of nuance in the visual storytelling. You could train a model on Wally Wood's 22 panels that always work-- but in terms of prompting? I don't think it would do a great job.
At the end of the day, something like Krea will bridge a gap between thumbnailing and final pages. To be honest, I kind of see it as a real time saver for most comic artists...which, I mean your deadline margins are super tight to begin with!
Ha ha! I'm not old enough. The writing is SO on the wall. I just signed up at a technical college to get some real-world skills, hopefully a couple of years ahead of the pack. (Which is how I was successful in indie comics; I was basically 2 years earlier than all the competition, in both business development and skill refinement.)
So now I'm gonna be an industrial electrical contractor! After making comics for more than half my life, I don't feel too bummed. I've seen and experienced all the best parts of comics, and right now, I derive more satisfaction from construction hobbies IRL than I do drawing.
But man.., I feel bad for the young artists. When anyone can generate a great comic, (or a personalized Netflix), from their smartphone, art skills will be about as valuable as being good at lawn darts.
@@solarydays any "limit to AI" is only a limit for NOW. Beware.
You're not going to lose any work because people are generally are disgusted when presented with ai generated images as "art". There is a demand of human skill and expirience behind a picture. They have to lie to sell us ai art, otherwise, noone cares. It's garbage
@@MarkOakleyComics Do you honestly think the career of a comicbook artists is gonna be dead soon because of AI?
40 plus year artist, graphic designer. Currently CMO/CEO for a business owned by my wife and I am really excited about using this amazing tool to expand creative ideas, some of which lay dormant on archive files waiting to be repurposed!! Let's GOOOOO!
I feel so sorry to all the families and artists that will never be employed in your company. Thank you for ruining the industry,
@@bellalawrencenye9793 Agreed
@@bellalawrencenye9793 Same thing was said when computers were able to use programs like AutoCad and Photoshop - did those that learned and used those programs ruin the industry? It's just another revolution in technology that will grab hold - you still need the creative outlet and creative chops to put it together and at this time, you still need to know programs to truly do something professional all the way. And artists will take hold of this tech and just become better at their workflow. Its called evolution in technology
@@CreativePunk5555 The old ways must never be forgotten for it is the back bone of all this. Plus when your computer is broke down you can sketch
@@bellalawrencenye9793 but at least AI will never be able to whine in public as well as you.
This is genuinely much more useful than midjourney or dall-e since it gives artists a lot more control than the text to image models.
Exacly.
Hope there is better tech than this that grants em even more control.
StableDiffusion was doing more and better than that for half a year now. But most of it went under radar because free opensource projects like SD don't have great marketing or polished UIs
Artists? Do you mean prompter? Or maybe AI intusiast or AI bro?
Trust me, artists have zero interest on this.
This is what I think AI assisted art should be like. If its like a drawing app you csn have allot more accuracy and get whst you actually want.
Perfect take. That’s exactly how I see it and I love Krea for this workflow. As I mentioned in the video, Martin is 40x the traditional artist I will ever be. And his outputs from Krea will reflect that.
What’s great about Krea is that it inspires me to get better at traditional methods, so I can use it better!
@@TheoreticallyMedia this is a good thing, with how the usual AI Image generator software works there has been an increase of artist haters that think artists are/will be obsolete. Art creation is a skill like any other I wish these artist haters stop using untrue statements like "why would I learn art when the AI can do it 10x better than you can". Even picking prompts differs between photographers, artists and people that have no prior knowledge. To me such people don't value the craft and just want to play pretty image gambling.
Mind is officially blown. I have been struggling to finish my long form screenplay/graphic novel project for years since suffering a life threatening illness and injury. It looks like I'm gonna be able to do it now. Thanks man.
Oh my God you are hilarious!! @terroristswin3461
@terroristswin3461I'd bet if wikipedia was a new tool you'd dismiss those using it to say they're uneducated swine lol
Who gona read it?
@@Bluejay-ri1yf Wikipedia is digital library. Ai do the work for you.
@@danieladamczyk4024when you cook food do you say you cooked the food or the oven? Ai is a tool just like anything else.
Wow, this was deeply depressing
Depends on how you look at it. This does not devalue traditional art in any way. Zero effect on fine art-- as I can't see any of this printed on a canvas and being displayed at Moma.
Commercial art will take a hit-- but I also argue that genAI can speed up an artist's workflow, allowing them to work on personal projects and pieces.
From a film perspective: Storyboard artists will still be hired. Some will use GenAI, some won't...but someone still has to do that job. The Gen artists will just be able to do it faster. But-- again, someone will still need to be hired, because there will be notes to be addressed. There are always notes to be addressed.
@TheoreticallyMedia I get your point, but consider this. Why would a talented young art student invest in learning to paint well, when anyone can create a perfect oil painting in seconds. The perceived value of that talent hit the floor. Fine arts at the MoMA is not relevant. Only 0.001% of artists ever show there, and it has nothing to do with technical talent anyway. It's conceptual.
@@nobody-u-knowwhy do anything if someone can do it better?
@@Reddkomet if there's a market that you can compete in, you don't have to be the best. But if an artist can make one painting per day, and an AI can make 1000 paintings at a better quality per day, the artist will not be able to compete.
@@nobody-u-know you could apply that to anything that’s manufactured. Should we ban factories as well? Ultimately it’s another artist using AI that is going to replace another artists, not a machine by itself.
Holy crap.
Stuff like this gives me analysis paralysis. Limitations are what make me an artist. Give me unlimited power and I don't know what to do. I used to do animation by hand with cell drawing and a VHS camcorder, frame by frame. Once it because easy I lost interest. If just anyone can do it, I think it looses value, at least from the perspective of the artist.
So this is kind of funny: I spent a little time teaching high-school age kids Photoshop and video editing. A lot of them were into Digital Painting with Procreate. But they'd always want to (obviously) work on a tablet, instead of on a monitor.
But what's funny is: When we went over to the (traditional) art area: They'd always be intimidated by large blank canvases. Like, they had no idea how to fill it up. It made me realize that working on those smaller screens, they could create an infinite image if they wanted to-- but only if the screen size was tablet sized. On a 5x5 canvas-- it was too much space!
I already do this using my own art ( watercolor and vector) as a base for AI ... LOVE the results I get as they are unique to my style and vision. I wish artists would experiment more with AI instead of hating on it so much!
I actually think there are a TON who are working and experimenting with AI-- they just aren't as vocal about it as those who are opposed to it.
Watercolor and Vector sounds really cool!!
I would love to see your art and chat. Am also new to watercolor artist who loves AI. @arraybabe8514
@@TheoreticallyMedia I'm new to vector, it's a mostly unconquered territory for me, but watercolor is my passion. And you are right, many artists are holding back and speaking up. more of us need to do so!
@@GringaCR Most of my paintings have been personal private commissions, but if you are new to watercolor I would seriously recommend splurging on the paper ( Arches ofc) and finding a good set of paints that suits your style. I have some Daniel Smith watercolors, but to be honest I always fall back to my White Nights St Petersburgh paints for 80% - 90 % of my works. I just love them, except for the blues , those I use mostly DS, as I find they have much better pigmentation. I have one set of squirrel paint brushes, which I mostly use for background blends and then WN brushes for everything else. I focus on floral and fantasy for the most part.
I agree!! Ironically, it's the short sighted who will miss out on how these tools can help innovate our craft. These same people would have protested the invention of the printing press back in the day.
To be honest, with this many images that it produces, doesn't this get tiring? Even just watching this stuff happen so fast makes my vision blur and my mind kind of zone out. It's not satisfactory in any way. One of the things I enjoy the most about art is the research and getting inspired by other artists I love, then applying it to my own designs. It's a careful trial and error that can't be replicated with AI where it constantly feeds you "amazing" stuff. Just some thoughts. It just feels like taking mediocre art to a "pro" level without all the knowlege that would come with actually getting to the AI image's level. Empty in a way.
I mean, one use case for you might be to do some quick perspective/layout images in a LCM like this, export that, and then working in a traditional method of top of it.
If you play around with it, and give it some detailed prompts, it will often surprise you with visual connections you may not have thought of.
Something like this could be more of a brainstorm session, than a finished product.
Just something to consider.
The creative industry doesn't care about artistic satisfaction, they care about results.
5:35 damn that Godzilla got me feelin all types of ways
That second image of its tail... is not a tail. I need a frontal view now.
Haha, that is a whole other video!
So let's say you're painting in Krea and it's doing the realtime image generation thing... if it flashes an image that you really like, but then quickly changes to something else, is there a way to get back to a specific one of its previous iterations on your images? (If not I suppose you could record a screen capture video as you're going along)
So, that has happened a BUNCH of times to me...I'm like: "No wait!!!"
Going to put that in as a suggestion to the devs. I'm not sure what the solve might be-- but they are also way smarter than I am.
Yeah, we‘ll have a capturing mechanism very soon. You should be able to grab any frame then. Consistency will also likely improve so that the whole thing feels a bit more like an animation, but that’s a bigger model architecture change -Titus
Yes please.
We would love to have a project cache of in between images.
Like ipad keeps recording everything.
And krita also records as we draw.
@@danrazART yeah we’re working on that
I see so many traditional industrial standard techniques being used here: digital photomanipulation, puppeteering, remixing, kitbashing, speed painting but on steroids! Except for detailing and refining. That's the work of the digital artist, for now.
Agreed! And glad you see it from this perspective! I think when you truly “get it” you see all the possibilities that AI opens up.
I mean, even a traditional painter can scan an oil painting in, then animate it. It opens up a whole new market and area for them to explore, potentially even a whole other audience. All the meanwhile, it has not devalued the original painting.
@@TheoreticallyMedia Takes a lot of heart to convince these artists to take up AI tools. But I guess they will eventually. Many years ago, there was a traditional landscape artist who animated his Four Seasons art using VFX. It was such a rarity back then.
This needs to arrive to my iPad asap!
So, obviously I'm not the greatest digital artist-- but I am totally picking up a cheap Wacom tablet soon!!
I prefer do Art the old fashion way instead
So glad to see Dreams still getting some love. Was the coolest project I’ve ever seen.
First of all great video. So well put together, but am I the only one that feels bad for artists? I am not an artist, but it seems like the equivalent of everyone going to a concerto where we watch a player piano instead of a pianist. Don't get me wrong. These tools are super cool and I love using them, but there is something bittersweet about looking at things now and not knowing if a human created it or not.
Nah, the savvy artists are quietly using this technology to their advantage. The naysayers are just a very loud minority.
To use your example: I think this is like seeing a concert with a pianist who is accompanied by a player piano, allowing for a wider range of musical notes and tones. Executing things that a solo pianist could never do.
In the early days of looping, I saw a Cellist do a set with a looping pedal and a number of guitar effects. Of course, purists hated it, while I went out and bought a looping pedal the next day!
Im an artist, and im very confident to say artists who does it as a career aren't very happy about it.
Another superb video. What a tool.!!
The faster latent Consistency models, are making this all so much more 'fun' as we get closer and closer to instant feedback. It's also great for learning tools like ComfyUI as you can now much more quickly understand what settings do what as you update your images and animation.
Hey, thank you so much! On the animation side, I can’t help but think (as the consistency improves) someone will add a keyframe/timeline feature to all this, and we’ll basically have an instant animation studio. Insane.
It’s funny, the speed aspect: my musician brain turned on last night as I was futzing in Krea, and I realized this is a bit like “jamming” with another player. You play a few notes, they react, you react to that…and then things keep evolving from there.
Also, I really need a cheap Wacom now!
@@TheoreticallyMedia . Yeah absolutely. I'm now regretting selling my drawing monitor. I really like the "jamming" comparison.
I can 100% see this working with a timeline.
Just testing my 'work in progress' Controlnet Rig After Effects plugin, and the potential for realtime feedback directly in After Effects as you animate the rig. Set keyframes for text prompts along a timeline) and build up the scene. It really feels like it's coming to the point where 2023 was the year the tools came out in early versions and 2024 will be the year this all starts to really bare fruit and have use cases in more commercial projects.
As much as it will level the playing field and quicken art generation process - it will homogenize art industry as modern DAWs did with electronic music industry: everybody and their mother can create electronic music now, so every electronic music producer sounds like any other. There are so much junk music rn that it's incredibly hard to find anything worth your attention.
tldr most likely with AI uprising we will lose our last shreds of creativeness and most art will be some regurgitated parts of previously created art pieces bend by the angle of crudely made strokes in some random paint app
I mean, obviously with the guitars on the wall, this comment hits for me.
I think home recording/DAWs have this double edged sword of democratizing and homogenizing. You are of course totally right: A lot of Music has gotten super bland, but there were a number of other factors there: like everyone using the same stupid Splice Loops, and the playlist culture that Spotify wrought.
But, it has also given us hundreds of artist that I’ve discovered in all kinds of genres that I never would have run across. Granted, I’ve got some weird music tastes, so these are artists/groups that never would have had mainstream success to begin with.
But I’ll say, the savviest of those artists have also figured out how to additionally support themselves around their art. Merch, Patreon, etc…
It’s funny, and this is a conversation I can easily go for hours on, but the majority of musicians aren’t overly concerned about AI. Most of them see it and think “how can I use this to my benefit”
To be honest, I think that’s likely a learned skill from a lifetime of getting screwed over. Musicians have always had to be a bit scrappy to make ends meet, and they’ve gotten pretty good over the centuries at rolling with the punches!
@@TheoreticallyMedia Whoa, didn't expect such an elaborate answer, well thank you for your attention.
I feel the same regarding musicians and situation as a whole. Though to be fair advances AI has made recently in visual design and art is a bit stronger compared to music generation.
And considering same sampling and DAW influence on music currently mentioned above it's already quite frequent that you can't really tell if it was AI generated or it war created by human hand.
So I feel the same may happen with visual art, thus my previous reaction
As an artist, I'm stoked for this! It looks like a lot of fun.
Also, you have solid in depth no-nonsense videos and that is a gem among youtube videos! Thanks again man.
Did you ever see Tafi's AI and if you did, did you get into the beta?
Oh man, thank you so much for that! Trying to keep it a no-fluff zone!
How Does it look like lots of fun, that people Take screenshots of animations, or steal many drawings from a singe atrist, to copy them even more direcly...?
What is Tafi AI exactly? 🤔
@@julle4083 Tafi is the company who makes Daz Studio, a 3d character creator program that's been around for idk like 2 decades or more. They basically took their lessons and mutability and turned it into an AI driven text prompt to 3d character generator.
Hater's gonna hate.@@m_lies
Wow, these are some huge development in AI artwork. Thank you for sharing this, Tim. I am an academic and am always looking at different ways that AI can be used to help with education. These development have academic implications as well. Keep up the great work.
that's awesome! and thank you! You've got me thinking now-- it's a great way to teach various movements and periods of art. Have a student draw something and then prompt for: Renaissance, then Cubist, then Art Nouveau--
It'd be an interesting way of visualizing-- particularly to the student, as they'd feel a connection to each era, since it is their input image!
That sounds like a great idea. So many great hands-on learning possibilities with AI, it is just unbelievable. @@TheoreticallyMedia
They will become lazy and without an ounce of self-realizaton trough hard work. Goodbye brain.
@@milionSTSocrates said the same about reading books - he didn't believe students should be taught to read and write.
For the most part, he was right.
But the ways he was wrong was well worth the loss of debate, discussion, and raw experience.
So: you're right. Sort of. People may lose the ability to draw, but not the ability to make art.
@@milionSTalso, i love making clothes, but nobody makes their clothes anymore.
So. Do mass produced clothes make us lazy? Not really. We spend our time on higher levels of work and creation.
And those skills are preserved by the few of us who love work and process and creating.
This new feature in KREA , As a compositional tool, is SUPER DOPE, the image input feature to abstract elements to a new image is like WHOA. Great tool so far, if you screen record, or better still it has a Timelapse record like procreate, then this would be an experimental animation tool. If I create a short film with stock video, then I suppose I could create an animation from the output , I’ve always liked KREA for illustration
……External screen : MIND BL🤯WN… Great overview Tim
Thank you so much!! And yup: there’s another unthought of use case! LCMs in general are really going to change the way we generate from here on out!
I’ve seen the animate tool in Procreate, but I haven’t played with it yet- but, 100% if it works the way I think it does: we’ve got an animation machine on our hands here!
@@TheoreticallyMedia I like it when a developer from the beginning creates tools that are open to 3rd party applications being an option for expanding one’s creativity and work flow. Applications like this even inspire other developers in their own lane to consider other use cases that can sharpen their tools pallets . I could definitely see this company developing a range of realtime plugins for post production image processing too. This is almost like the choices you get with using a modular synth in comparison to a regular synth, or signal processing sound… this is sparking so many ideas , like for exhibition space and interactive installations like the van gough / Dali immersive exhibitions…
I’m really looking forward to this. Once again thanks for keeping us on point with top choice creative innovation.
Totally amazing!!! At 12.00 minutes the AI model created an image of a boxer knocking the head off the other boxer. Did you notice that?! lols!!
Ha! Missed that!! I love when you guys catch the insanity! Usually by the time I’m publishing the video, my eyeballs have gone numb!
I actually tried Ever Art AI . I was satisfied with the results I got . I was about to subscribe , but I found their subscription plan too pricey , for their unlimited training , and generative image . Their first two tiers on the other hand , was nice , but you're only allowed a certain number art to generate per month .
you might want to look into Scenario. While not free, it does have a pretty good trial period: ruclips.net/video/v_FXC0iq1Sk/видео.html
It occurs to me that we need one with an alpha background option and a no foreground option. These would allow layering in photoshop or gimp. It would also remove some of the "your stealing my art" arguments. I find I'm having to erase a lot of things in gimp to clean up the product.
I agree- there’s almost a threefold workflow here: generate, futz, then post-cleanup. The idea of an alpha background would be great.
It would also be cool if you could lock certain areas of the output, since it does tend to generate pretty wildly with the smallest of changes. But again, this tech is like 2 weeks old, so we can give it some time to mature!
@@TheoreticallyMedia Yep its getting hard to keep up with the innovation and my cash flow is zero so I can't subscribe to most. I suspect an AI heal/ detail 'brush' will come to gimp and paint apps soon to further build up the innovations. I need a much bigger computer already.
My kids sketches from when they were in kindergarten are worth a lot more now.
Ha! To be honest, I’ve still got a crate of them in my basement. In those early years I couldn’t bring myself to throw them away.
Of course, these days I’m yelling at them to stop stealing all my printer paper to draw on!
(As a note: they HAVE sketchbooks i buy them all the time. But apparently stealing my printer paper is a source of joy to them…)
0:36 Krea looks very impressive, but it would be amazing if it could be thoroughly trained on custom characters. I wish I could use it to depict my custom characters from every angle possible. I need to pose my custom characters in different scenes. I also often need to combine multiple custom characters in a scene. It seems like Krea would be amazing for scene creation, and the only thing missing is the ability to accurately & consistently add custom characters to scenes.
Are you familiar with DAZ 3D? It's a figure customization tool used in animation, games, and concept work. I'm already using it to create and pose specific characters, with unique features and expressions. I use those for reference and consistency, placing them into Clip Studio to ink.
This really is a tremendous leap forward. 🚀
Congratulations to the developer 🎉
Thank you for this video
As mentioned in a previous post, a 40-plus-year artist, and graphic designer. Currently CMO/CEO for a business owned by my wife and I. I signed up for the "Creator" account. Getting a lot out of turning my creations into models. However, I'm finding it challenging to purchase more tokens, even though my subscription states I can "purchase more at any time". Has anyone run into the issue of not being able to purchase more tokens? Thank you.
Are you on Twitter/X? You might want to hit them up there with issues. They seem to be very responsive on that platform:
@@TheoreticallyMedia great idea! I'll do just that. I sent an email. I'm working on a project and need to purchase more tokens! Appreciate your content!
Gulf Shores Alabama!
They seem to have reduced the options or my browser is not displaying the page properly or something
they for sure have added stuff since this video was published-- but if anything, I'm seeing more options? Maybe try a different browser? That's weird...
I'm signed up for Krea beta. I cannot wait to get in and connect it to Blender for rapidly putting together scenes. I was thinking about some pretty interesting use cases. - For example, using primitives to create an output, then render the scene and use them both as image prompts to finalize the scene. So very exciting!
Also looking forward to hopefully getting into EverArt and using my stuff to train it for consistency.
Everyone who is interested in creating digital art just got their minimum quality leveled up in a huge way with these tools.
Thanks for the video Tim!
Did you get your invite code? I'm still waiting. My guess is that they have limited server time and need to parcel it out judiciously.
Scandalous Godzilla 😂😂😂
Never missed leg day and can deadlift and squat like a BEAST!
I mean, this method of using AI has been around for over a year now within the concept art field :p
Although this seems more optimized
I mean, to a degree it’s control net- but wayyyy faster.
Stable Diffusion just release SD Turbo, taking steps from 50 to 1!
Insane!
I think 24 will be (among many other breakthroughs) the year of real time.
5:34 Never had I realized that Godzilla is THICC XD
Dang...Godzilla's been working on his glutes
Never misses leg day and always focuses on deadlifts and squats! Literally…ahem, a beast!
Wow thanks! Your videos are the best!!❤
Oh bless you! Thank you so much!! I’ll keep your comment in mind as the hundreds of angry comments about the thumbnail start rolling in! Haha
You don't have the Krea link in your bio - FYI in case it was an affiliate link or similar. This was AMAZING - I can't believe this video is a month old!
That Godzilla killed me 😂 😂
ScandalZilla! I wanna take him to Gen-2 and animate some twerking!
@@TheoreticallyMediaplease do this. Seriously.
Nah, just kidding.
But not.
You should develop your "Henchmen" idea further. It's a solid idea. You should consider seeking funding to turn this into a TV series or a movie. I'll bet you could even convince Apple or Amazon to fund it.
So funny enough, the whole 12 issue script was written, and it did get optioned for a TV show at one point- but that whole game is basically a lottery ticket. We got shopped, but never picked up. But I totally agree, I think it would have made an awesome TV show.
With IP, it’s always weird. I could get a call tomorrow about it and…well, it’ll still probably not get sold, haha.
@@TheoreticallyMedia
That's really cool. Is it expensive to get shopped or pitched to the various decision-makers in the tv/film industry?
That's really interesting, with the other window capture in Krea, is it possible to rig it with a web cam that records real drawings?
It does have a webcam option as well! Sorry, forgot to mention that!
@@TheoreticallyMedia Cool, I just signed up for the waiting list (hope I will get access soon). I want to try it with messy traditional materials such as charcoal and wax stick and see what I can tune it to and also geometrical City type shapes in Blender.
@@jamesabell9494 omg, that is such an amazing idea. I’ll say, talking with Krea yesterday, they’re really excited to see what insane things we’re all going to do with it. They’re as excited as we are!
I do a lot of expressive drawing so it would be great if I can someway tweak it so it resembles and evolves the lines etc will experiment with it asap when I get access!@@TheoreticallyMedia
Neither of these are letting me sign up … when will they be live?
Both in beta right now. Krea, just sign up- for EverArt, there should be a signup on the main screen?
Ooh, you could use this to get some custom details, then put it through Leonardo AI with the Img2Img sliders set at different rates to see what comes out.
YESSSS! Toolbashing is one of my favorite things in the world! Hmmmm, off to Leo I go!
I’m just sad to see the job market evaporate for trade artists and myself (20s) just trying to get into the market for graphic design and art only for it to literally be erased before my own eyes.
Nah, you'll be fine. The truth is, technology like this still requires artists to manipulate it, and...well, to be honest, address notes. Because there are always notes.
Build relationships with other artists, clients, co-workers...basically everyone you run across professionally. You'll get jobs. Any company that is relying on AI Generated Images whole cloth is not a company that you'll want to work for anyhow.
this is unrelated...but try teaming up with a writer and put out indie comics
Thanks for your content! Is it possible for you to maintain and regularly update a list of AI media tech? For your patron members? Would help a lot as the landscape changes quickly and would love to have a quick reference resource.
That's a REALLY great idea! Let me think about how to best implement that!
Where is the copy right
In limbo. Still hasn’t really been stress tested yet. Particularly with a tool like LCMs.
But seriously, don’t stress about copyright. Internationally, many countries have already approved it. The US will hem and haw, but I think it’s a done deal at this point.
WHAT DO YOU RECOMMEND FOR PUTTING VIDEO FILTER ON A VIDEO SEGMENT WITH A PROMT, PREF WITHOUT A TON OF RESTRICTIONS LIKE VIOLENCE ETC
Few more details? What’s the use case? When you say filter, do you mean a stylize pass, like Gen-1 or Kaiber?
@@TheoreticallyMedia YOU HAVE A BANANA UP YOUR BUM
@@PuppetMasterdaath144 how’s so? I see it?
@@TheoreticallyMedia iN bum bum yes , see
AI is an interesting topic to study, every day we have something exciting.
I can’t thumbs up this enough. Of course, as a guy who obsessively follows all of it, it also gets super overwhelming from time to time.
But yeah, literally i wake up everyday thinking: how did the world change today?
It’s amazing
@@TheoreticallyMedia I read the book 'Dark Matter' this week. It kept me thinking about how many possibilities AI development can create, and how each little advance in one area can result in a completely different world. Mind blowing haha. Thanks for your work! 🤝
@@aiforgemaster oh awesome! I love Blake Crouch! Recursion is really good too! And thanks for reminding me, he’s got a new one out that I forgot to read!
"Futzing". Nice. Haven't heard the term in decades
Well of course it just changed forever. Every day, it changes forever.
Literally, every morning since March of 2023, I wake up and my first thought is: how did the world change today?
It’s awesome.
@@TheoreticallyMedia it used to be that ideas I had about the future would take 20 years to become realized by others. Now, ideas I have about the future become realized in less than a week!
Thanks Tim. Very interesting. Id love to see a really good Z-Brush artist working live in Krea.
Oh, that would be amazing. I love watching Z-Brush timelapses-- that is one piece of software my brain is not built for, but I am always amazed watching the wizards who do use it!
I cant see anything like that in KREA. no options to create or do something. I asked tor a ticket, but never got any invitation code or something.
It hasn’t been released wide yet, but it probably isn’t long. They’ve been doing a lot of open tests via their Discord, so I’d recommend joining that!
Ok, thank you for quick response. @@TheoreticallyMedia
I don't quite understand the nuances of how training works. Without anyone explicitly telling it that a boxing ring is called a boxing ring, how is it able to draw one from a prompt asking for one? Is the training completed from the cartoon strip just an additional layer on top of a much denser training image set?
🤯This is big! And 'Krea' sounds like a wonderful woman too! 'Scandalous Godzilla' 🤣
Great video! Fantastic 👍
I gotta say, I always have fun making these videos, but there was something extra fun in this one! I think it was because of Scandalous Godzilla!
You had me laughing so hard at "scandalous Godzilla".. omg But wow - this tool, man I don't know what to say aout where all this is headed. A professional Artist here.. semi retired.. but truly bewildered.
Wow! Looks like another fun rabbit hole to explore
It goes deep!!
Damn AI gave Godzilla some god level Gyaaat
I need my own JARVIS for art.
It was a real mistake having my JARVIS draw Ultron. Now he won't shut up about how he's going to destroy me. Sigh.
5:35 Godzilla has CAKE... lolz!!!
Our boy never misses leg day, and it shows!!
At some point in the relatively near future, we're going to be able to tell an AI to give us a Terminator film starring Bruce Lee.
Literally will be my first prompt!
i think ill ask the ai to create a browser addon which automatically filters and hides videos with clickbait and sensational titles, that would change everything for a change lmao
Thank you! I have a question: is it possible to upload a photo and turn it into a semi-impasto oil painting where the brushstrokes and the paint actually looks realistic? I still haven't found an AI tool that can do this in a satisfactory way, including photoshop's generative fill function.
This is my question as well! When will this be possible?
I think thats stable diffusion dreambooth but could be wrong, also just like in this video you can upload your own photos to train the model
I was wondering when we would get real time image generation
I only had to wait a few months.
Start wondering about Lottery Ticket Numbers!!!
Very interesting... Where do you think copyright fits in to all this. You used copyrighted material and who owns the copyright of the results?
Well, to be fair, with the Henchmen comic pages, I do own the copyright to those. But, to your overall point: There is sure to be a lot of saber rattling over the copyright issue, and a lot of lawyers are going to make a lot of money off of it. But, at the end of the day, it looks like because the material is being trained, and not reproducing 1:1 copies-- it'll land in favor of AI.
The US copyright office recently decided you can't copyright AI generated material, without a human doing some kind of post work on it. But, that's a tricky rabbit hole-- as what constitutes post process work? If I put a filter on it? Or if I wholly repaint the image by hand? How do you quantify that? And hell-- how do you even prove it? If I send a Midjourney Image to the Copyright Office and don't declare that it is AI generated, how can anyone prove otherwise?
It's going to be a mess...and ultimately, all of this only applies to US Law. Japan and many other countries have already OK'd AI generations...so, it's pretty much a done deal.
Thanks for the thorough reply @@TheoreticallyMedia
what happens to your own images that you upload in everart.Do they end up in their data bases free to use for everybody else?
I was wondering this.
Good question. I’ll reach out to find out. I think it likely ends up feeding their training data. I mean, it’s the smart play on their side.
I’m never too concerned about that, only because any image or set is just a blip in an ocean of data.
Oh, wait- I think I misunderstood! No, I can’t see anyone else’s trained sets. Just mine.
if thats the case then that is very cool. potential massive time saver
@@TheoreticallyMedia
LCMs highlight a problem with services like Krea. This sort of real time feedback is a million times better running locally on your own machine, not only because of the low latency, but because of the ability to use your own models. Models that you might not want to share publicly.
Probably just going to have to try myself, but when you undo something in that Krea - that is e.g. erase something you draw - does the generated image also go back to the previous generation, or does it make something completely new?
It seems to. I’ll double check that for you, but I think since the seed is locked, it would make sense that it does.
could i import my art stylized images and then manipulate them like this
heard jsut saw god zilla nvm
Haha, turns out, he’s been the real star of the video! Scandalous Zilla!
@@TheoreticallyMedia the PNG thing but yes this soft core baddie lizard could crush me like a soda can
Are these tools censored?
Playing around with Krea I saw a few (artistic) nsfw generations pass by. Nothing too gratuitous, but it was there.
And I mean, did you see the Godzilla image? Haha
Wow! Absolutely amazing! 🤯
It really is game changing
Krea take my f**kin money, i want this nowwwwww. Been looking to leverage my art with AI and DALLE just cant do it
Such a gamechanger. Honestly, Midjourney/Dall-E and the others are probably having a really bad morning.
1:19- so interesting the Krisa from kotor is now a programmer to redraw her home galaxy.
You are my go-to man on the latest advances in AI art. Your demos are a revelation and inspiration. You said AI art just changed forever. Looks like I'm also changed forever. Can't wait to try them out.
how to get access of real time generation
Check out the links in the description. Krea or you can try over at Hugging Face
What ai software allows me to upload my own drawings and the ai cleans up them and colors them?
So-- sort of. I'll have to try it out, but it generally alters the image somewhat. See the Pirate example to see what I mean. That said, you can then "trace" your own drawing after that to likely get something super close. But, it isn't 1:1.
Wow, this is really going to help out small businesses. I don't spend crazy amounts on freelancers but it adds up annually and any savings help. Just signed up, can't wait to experiment
thanks for the rundown, this is a big deal
Such an insane level of control. And once things like training models are applied to it? I mean- there isn’t going to be any image that can’t be endlessly micromanaged! Haha
welding trade school, here I come... Can't code any more and there goes art
I just paid a stupid amount for some radiator work in my house. The part was $7.
Luckily, the guy was cool enough to show me how to do it, so next time around I can save some money.
Great minds (me ;) and obviously you). Finally someone Who talks ab. Akira.
I really hope Krea will be available in a local install.
The underlying technology, LCMs are, but it also (literally) a week old, so it is a bit janky in terms of installing it.
I just registered on everartai but it says site is down 😢
Hmmm, I’ll check in with them. Might have been a weird issue?
@@TheoreticallyMedia it says "we will send you an email when the site is live "🤔
I think we could make dedicated plugin in PS for this to work seamlessly.
I'm sure Adobe is already looking to implement this into Gen-Fill. That said...I'm always 50/50 on the results I get from GenFill...
wait, how can you use ps when they state Very Clearly that anything AI you do with them is Theirs, All Theirs and all Iterations?
To be honest, the legality on all this stuff is so in flux, pay it no mind. Just make stuff-- lawyers (on both sides of the argument) are going to get rich battling it all out.
If you make something that is generating enough money that someone calls on you-- well, you've made enough money that you've probably got a whole set of more important problems anyhow!
NovelAI's Image Generator Version 3 is pretty insane imo. and I'd consider myself a AIGI critical "artist".
I’ll dig into it!!
hmm... this is a nice way for AI "prompt artists" to be actively involved in the creation of the image with tactile input
Can’t disagree. And again, I think it comes to a matter of “the more you bring, the better you’ll get”
You might see it also a a weird way for folks who have been promoting to get back into traditional art. The thing about AI art is: I don’t think anyone who spends a ton of time experimenting with it has no artistic background. They all come from somewhere.
Gaaaame chaaangeeeer ! What great tools ! I am going to have fun very soon.
You didn’t tell us HOW to get into it. I’m in Korea now
Still in beta. Best bet for right now: jump on their Discord. They open the door pretty often for testing.
I wonder, if this technology will be ever used for photorealistic graphics in video games.
Hi. They have both waiting lists, now. But it's nice to see tools for detailed adjustments.
Godzilla opened up a whole bakery 😂
Crypto bros /Nft bros and AI generative content prompters gross me out. Ai will never take away the ability for artists to draw but will steal their livelyhood after sucking them dry to be their own competition. We cannot stop the march of technology the same way no one could stop cars from replacing horses or robots replacing factory workers and leaving them to die. Society has been treating artists and artistic fields as afterthoughts or as "not true careers" for years while silmultaneously being unable to live in a artless environment. AI art is the nail in the coffin imo.
Studios will use this to replace artists. Ai is a tool that could be amazing but if you believe for one second the AI bros will use it as intended you are dellusional. AI generated content is ground for abuse and when you will have destroyed the humanity out of art you'll look back in shame.
If you lots even have that sort of empathy or self awareness to begin with. All these "40 years old in the art industry" blokes in this comment section remind me of naive boomers amazed by technology they didn't have in their childhood and not fully grasping the consequences of what they are fawning over.
I mean. You’ve pretty much got everything nailed down here. Everything about your post is correct.
There will/is a gold rush of NFT/Web3 Chads making a run right now (I get emails asking me to sponsor their startups all the time. I turn them down)
Studios will for sure, attempt to abuse the technology the moment it hits a technical level, despite their promises not to. That is exactly what the SAG/WGA strikes were about (well, that and streaming money)
Artists have always been devalued through history. I’m pretty sure the first caveman to do a wall painting was asked: “you gonna do that all day, or maybe go help us hunt for food?”
That said, I do have discussions all the time with people in the VFX and Post industries, and they aren’t the people you might think they are. They’re very thoughtful and aware of the double edged sword this technology is.
I talk with those “40 year old” Boomers, who are excited, because they see the prospect of being about to create a movie or a comic that they’ve carried in a dream since they were a kid.
I don’t know, the whole “AI has no Soul” argument reminds me of old “Drum Machines have no soul” bumper stickers.
Horses left the streets full of shit and were forced to carry heavy weights all day so if anything we should be glad cars replaced fhem
What advances on LCM and real-time generation have been made since this video? Is there anything available that’s free right now?
How restrictive is Krea in term of NSFW creations? Does it do blood, gore, famous people, nudity?
I'll say that while it was generating, I had to carefully check my footage-- There for sure was some...scandalous generations that passed by. Even more scandalous than that Godzilla image!
@TheoreticallyMedia
Interesting. That's one of my biggest gripes with AI generators......most tend to be hyper restrictive - overly censoring.
Sure, in some cases with clever prompt engineering one can sort of circumvent some of that censoring, but still, I truly hate the overly abusive NSFW restrictions seen with platforms like MJ and many others.
@TheoreticallyMedia
As a fan of Frazetta, for example, so much of his subject matter cannot be simulated via most AI art generators. Imagine how quickly your prompt would be rejected trying to describe a warrior standing over a pile of steaming bloody dead human bodies! 😉
@@LouisGedo haha. Frank also loved his scantily clad (or not clad) ladies as well! Love everything about his work-
For sure Midjourney doesn’t allow it, and awhile back I tried name calling him and that was a no-go. It wasn’t even close.
Weirdly, I’ve been trying Conan prompts as well (cinematic) and MJ steadfastly refuses those too. I mean, I get stuff- but it isn’t very good.
That said, I’ll give it a shot in Everart (the second platform discussed) and see how it goes!
@@TheoreticallyMedia just make sure you don't use the word "stepmom" in any of your prompts.
Tim, The Enchanter.
Haha, is that my new Myers-Briggs?
@@TheoreticallyMedia Sorry. For some reason as you were signing off, the Monty Python movie character came to mind.
@@Agg1E91 oh, duh- of course! I make fire without flint nor tinder!
Haha; sorry, I would have caught that if I’d had my coffee!
If Krea would give us the option to use our own gpu's it would let them scale quicker. i like when we are allowed to run ai apps on our own machines because of this issue with developers needing to spend $$'s to scale their services.
The underlying technology (LCMs) are open source, so technically you can run it yourself. Although, it’s only about a week old, and from what I’ve heard, configuring is a bit of a technical challenge, much less being able to screen share etc,
If you’re up for the challenge, there should be a GitHub around to do it. You will be far braver than I!
I love this and I cant stop using it to be honest but, is it me or you all get japanese people? I mean, not always, but I would say.. 80% of the times I drew people. Not complaining thou, just curious
Well, I think there is a fairly narrow range of things this one has been trained on. I do notice that as well.
Next week, I’ll have a new version of the tech that’ll turn some heads! Gotta test it, but it looks super promising!
What are the other guys you recommend following on RUclips concerning the exponential use of AI within digital art? I think you mentioned someone called Martin? And then there were some blender guys - do you have their names or channels somewhere so I can follow them too? And thanks for all your work and information its very eye opening
Super cool, I think for skilled artists it could be an interesting tool to experiment with. I wonder if these models are also trained on copyrighted works? Or if users will be able to copyright their outputs
What are you saying? Of course it's trained on copyrighted materials as any generative AI machine. And you will never have any copiright on something that is not done by human. And you also mentioned skiller artists? Are you aware that skilled artists have zero interest on that?
Are you able to upload an image and paint over it in KREA?
Nevermind, commented early 😅
Haha, yup! Our old Dutch Footie Pirate gave us the answer!
It’s funny, I can’t help but think one day someone is going to show Daniëlle van de Donk one of these videos, and she’s going to be all, WTH?!
I love how your video titles feel like click bate but are essentially true every time
So am I going to starve if I choose the art path as my career?😍😍 yipeey
I mean, look, art is never the safest career path in the best of times. I’ve pretty much made a living with a camera, music, or writing for the last 20 years.
There have been years were it was LEAN. I was the definition of starving artist. There was a point where McDonalds had hamburgers that were 2 for a dollar. I bought 40 of them and put them in a freezer. (I’ll never eat one again)
Making a living in art is never based in talent and skill. It’s about meeting and connecting with people. Networking with others in your field and rising as they grow, and then offering a hand to those coming up under you.
If you’re truly passionate about making art; whatever the medium, and you’re willing to get punched in the nose, everyday for ten years, and never give up.
Then yes. Do it. AI isn’t going to stop you. Might even make your path a little easier.
Currently starving from the sudden reduction of jobs and lost some oppurtunities in this industry over some AI. That's years of my life wasted and having my wrist damaged from doing studies and work at the same time to somehow survive. If you still got hopes and dreams in this career just forget it. it's an industry, in the end it's all about the money.
@@aldavedario4818
I'll try to forget your comment, art is all I will be able to work of😔
Awesome video! I can't wait to try this!
I can't wait for you to try it too! It is SO MUCH FUN. Seriously, book an afternoon when you get access! It's sort of addicting.
This is an dystopian nightmare and I cannot wake up...