Hey Tim, just a heads up, there is actually a great new feature on this platform, called Image BOMP; The info for which is rather buried deep in the image generation help section. Essentially it's a means for a quick blend of any two images together directly on the Canvas, to generate a new image taking attributes/aesthetics and subjects from both. It works pretty well too, imho. (With the Select tool activated, click on an image, and drag it over another image. Hold the image in place until the bottom image turns green. A new image will generate on the Canvas. The new image aspect ratio will match the bottom layer image.)
I've never been so confused and turned off by an interface concept. Next time you talk to them, can you ask if there will be an adult version and not just this one for children?
Thank you so much for making this video because there are many changes to the look of the Kaiber website and the video creation process. Your video has saved us so much time! Now we know what to do
@@TheoreticallyMedia I know! It's like evolution, but backwards. All joking aside, I wouldn't have guessed it's only been 18 months or so since that abomination of a spaghetti video.
🎯 Key points for quick navigation: 00:00:00 *🚀 Kyber's Return and New Features* - Kyber reintroduces itself with a significant platform overhaul, introducing new features, workflows, and a fresh design known as "super studio." - Offers a unique generative design aesthetic and integrates advanced tools like audio reactivity, video upscaler, and creative canvases. 00:01:48 *🎨 Creative Workflow in Super Studio* - New creative workflows called "flows" and tools like Luma's Dream Factory facilitate advanced AI-generated video and image creation. - Enhanced features for organizing and iterating creative outputs using infinite canvases and AI-driven modules. 00:05:26 *🎛️ Organization and Workflow Efficiency* - Introduction of node systems, collections, and buckets to efficiently manage and link creative assets, preventing disorganization. - Modular approach to creating diverse creative outputs while maintaining workflow simplicity. 00:07:05 *🧪 Experimentation with Image Lab* - Image Lab allows for custom artistic workflows and aesthetic alterations using various tools and prompts. - Encourages creative experimentation and remixes within Kyber, demonstrating flexible cross-tool functionality. 00:09:09 *💰 Pricing and Platform Integration* - Kyber's pricing model at $15/month provides access to various AI tools and platforms, emphasizing affordability. - Potential integration of additional AI video generators due to the opening of major AIs' APIs, positioning Kyber as a competitive all-in-one platform. Made with HARPA AI
Tim, this new Kaiber playgroung is really exciting actually ....... maybe the best thing yet in AI considering the potential to add additional tools and for a decent price for those like me who are on a very low, fixed disability income.
That's a great step in the right direction, I love the NOD system, although at the moment it's a bit useless. what they really want to create is using different input as different effects. for example : image in nod 1 is the styling, image in nod 2 is the character, video in nod 3 is the motion, nod 4 is the color correction and nod 5 is a reference. and all these nods are effecting one video output
At this point it feels like it's the step to right direction BUT it was also done by product managers and tech people and this is not yet the thing that is going to be used. The right thing would be to invite/record people who are producing/editing AI generated videos for their projects and make something that streamlines that process. For example when I was working on a music video, I was generating images with Midjourney and then cropped the images, pull them into a Google Colab for facial syncing, export the results, then pull it into Davinci Resolve, and then mix the animated face back into the cropped images, edited the pieces yet again Davinci Resolve and so on. And I think for a long time this is how it is going to be: people generate pieces and stitch them together in a professional editing software. What the platform should provide the easiest way to assist the generation including pulling everything up to the same aspect ratio/resolution/frame rate + upscaling, color schema so you can just drop those into your editor.
I've been sitting on 3 months of subscription credits with Kaiber hoping they would improve things. Saw their announcement email and I'm glad I waited. I've been wanting to make music videos for things I've been working on, and the previous version of Kaiber was a little nightmare for wasting credits on bad generations.
The one thing I really appreciated about “old” Kaiber is how they provided you with few preview frames of your output before you committed to a generation. That’s something I keep banging the drum on. I want that to become a standard across all the generators.
@@TheoreticallyMedia Absolutely agree. In my experience with those previews, unfortunately though, they would change drastically in the final render. The first frame would look fine but then, despite the evolve level being 0-1, it would still have a bizarre and unusable outcome. Ideally a low resolution preview of all 4 preview clips would be great, rather than the still image first frame. Old Kaiber seemed fine if you just wanted a lucid music visualizer, or something to spark a creative idea. I'm a big fan of Linkin Park, but feel like their claim to having used Kaiber to make their music video for "Lost" is a big stretch. Pretty sure they had proper animation done and overlayed with some Kaiber generation backgrounds. They used Kaiber for other unreleased tracks, and they worked as obscure visualizers but not anywhere near the level of the Lost video (song example: Healing Foot).
All the new AI companies (I'm not sure what "back" means) want to offer you a "canvas". The thing takes 30-seconds to load a canvas. Just let me use it already. Also, you can't delete a canvas so this thing is just polluting your assets area. Why do none of these companies hire a UX consultant?
Can we all just appreciate that we are reaching the first end goal of video generation? ❤ Now movie studios have a great alternative for expensive visual effects
I've been using Kaiber for a while now and I love it. Customer service is outstanding. I've created some great music videos with the product. Haven't tried the new version but I'm looking forward to it.
8 years ago I shifted from adobe photoshop to figma, because of endless canvas, I feel this canvas gives you an ease of creative freedom. And definitely taking this initiative kaiber will standout in the market, it is understood for their current modules, but this is going to change the game of generative ai tools. One suggestion, try to add some new examples , stop testing business man prompt :) more examples generates more ideas.
Woe, so Kaiber finally stepped their game up. Oh, you're the first I believe to upload a TH-cam video on Kaiber Ai. I was expecting MattVidPro or Matt Wolfe or even Al Samson. I like the direction that Kaiber Ai is going to. But like I had commented to you on a earlier video of yours, Runway Gen Alpha 3 and Kling Ai to me ( TO ME) And I say that because I know and understand that everything is relative and subjective. To me I still favor Kling Ai 1.5 to generate image to video. However I will play with Kaiber because I had created over 30 TH-cam videos using that program. But I'm heavily leaning towards Kling Ai along with some other programs to make everything professionL YOU can't JUST rely on Gen Alph 3 or Kling Ai. You have to use other programs such as Adobe Aftereffects and such for postproduction. Excellent video.
Well, I do mention in the video that we'll see some more generators in Kaiber soon-- and given Kling/Runway etc have APIs available now...welllll....I'd say keep an eye out!
Good to see Kaiber still about, and well done on keeping the pricing sensible (I do get though these things cost). Will be purchasing a month top play with very soon. Great video Tim :D
At this point, I don’t care much about new AI video generators. I only care about new features that give the user more control. AI video generators have gotten better at producing aesthetically pleasing videos, but they still fall short when it comes to giving the user more control over the details of the output. How about some inpainting, camera controls, motion brushes, ability to pose multiple consistent characters together, the option to add a start frame & end frame, etc.? These are important features to make AI video useful for storytelling projects (e.g., making an AI movie).
agreed. And all of that is (eventually) coming. I always do like to note, that OG Will Smith Eating Spaghetti was March of 2023. So-- we have come a long way in-- what, a year and change? And a lot of that stuff exists-- it's just connecting it all together in non-janky ways-- and obviously, it is currently not always reliable. I'd say by mid next year, we'll start seeing platforms that are integrating all of that and more.
Wow. I've been making AI videos and doing short films. I used Kaiber, but it was still a little off. I have a fun Space video on my TH-cam. You have provided me with some new ways to make my new one. Thanks!
That's awesome! I've got a deeper look at Kaiber coming up next week! Like I said in this video, I feel like I've barely scratched the surface with what this can do!
Hi Tim. Literally just got off Kaiber as I've been exploring it for the last week. I haven't even began to watch your new vid yet. Anyway, I'll shaddap a sec and watch your video LOL
@@TheoreticallyMedia I like Kaiber because of their unique style. Sometimes, those imperfections are great because you can create surrealistic imagery and effects. Anyway, just wanted to say I more or less came to the same conclusion as you. I'm even at a crossroads now on whether to be a paying member since that is indeed a bargain.
Got curious after your cello comment and came across the album Out There by Eric Dolphy. A very fitting album name, but had some pretty interesting stuff. Oh yeah, great overview of Kaiber too, haha. I hadn't played around with it yet
I'll have to check that one out! I listened to a lot of Dolphy playing with Coltrane, but I didn't dig too deep into his own catalog. And that album cover is kind of awesome! I'll give it.a listen today! New Kaiber is interesting for sure-- I think we're going to be seeing more of these node systems popping up with creative AI tools. I still have to play around with it more-- It's funny, this is one of those situations where if you have preexisting knowledge of Comfy, you have to untrain yourself. But, I think it's a great system to introduce the idea of nodes to new users. That said, I'm pretty sure I only scratched the surface with what can be done here! Looking forward to spending more time w/ it (and getting some dev tips) next week!
Appreciate it!! Yeah, I've always loved Kaiber for the commitment to exploring different artistic styles! I can't wait to really dig into the Image Lab area, because I think that's going to be a pretty powerful tool-- I just didn't have long enough to explore it for this video! But-- that's why I really want to circle back to a bigger deep dive!
I'm talking to them this week, so I'll bring that up. One of the things I really liked about "old" Kaiber is how it gave you preview frames before you'd generate, so at least you could see how things were looking before you committed. That is something that I've been pushing on a lot of the AI Video generators-- I think that would go a long way: Either A) a few preview frames to start, or a low-Res version (like 320p w/ watermark) just so you can tell if it is worth generating.
I just checked-- it looks like the Kaiber Mobile app is still the "old" Kaiber, but I'm sure they'll have a revamp soon enough. It'll be interesting to see how they port this idea over to mobile.
@@TheoreticallyMedia thanks for the speedy response! I checked as well. It looks like it might work on mobile, in "desktop mode" if one's device is up to it. I'm not sure mine is... Super cool looking tool. Thanks for sharing all the good stuff!
@TheoreticallyMedia Can you ask the Kaiber devs if they have plans for reaching out to platforms like Udio to see if they'd be willing to license access to Udio tools inside Kaiber? Because then I don't have to pay separately for audio, image, video AI tools if I can have them all in one playground space
It's still there-- I think that's under the Video Lab/Video Restyle flows. I'll dig a bit deeper into that when I get the chance to do my full walkthrough/tutorial on Kaiber. Hopefully within a week or two.
You can in the “old” Kaiber, which you can access via a dropdown. I know they’re working on a revamped version coming in a week or two for the new version,
Kling just announced that they were having some issues w/ bad actors abusing the API, causing false flags in their system. Check out their Discord-- they said to reset your password and they have a thread where you can put in your Username if you've been running into a generation issue.
Great overview Tim….i like this natural progression of platforms have a kind of intranet studio that you can tailor different laboratories of workflows in the same environment…. Ps could you do a feature on how to connect multiple platforms and a bit of ai automation of render tasks . Cheers
I was a huge fan of the Original Kaiber but I'm not impressed with what they have done especially for the price they are charging the video is just not up to current standards. Very dissapointly to be honest
Give it some time. Currently the credit pricing on Luma is obviously based on the API calls. I was looking around, and obviously the credits to generate are less for the old Kaiber model. So, it’ll be interesting to see the credit costs of the other models they add in. (I’m sure Runway/Kling/Minimax are on deck…)
there''s a chance to put the aesthetic tools to VIDEO RESTYLE like to old kaiber? i need that to my video to video project., and the legacy kaiber now is not working super slow cant even generate
It for sure takes a note from a Comfy styled interface-- but, I think the smart part on Kaiber's part is to have a "Comfy Light" UI that gets people used to the node workflow, and expands on that. You might be surprised how confused/overwhelmed people get when they see a Comfy workflow the first time. This is a bridge step, I feel.
I've been using Kaiber for about a month, and I'm not seeing where we choose the animations in the new Kaiber. It's like a completely different Universe now, and I'm confused.
Pretty cool, more for the potential than the current execution. Played around for a bit, video generation is still average. Not quite sure I get the praise for their pricing though, you get 25 videos, 5 seconds each for $15....not great, even less if you're using it for flux images as well.
Damn man i havent used kaiber in awhile, they were my go to, but kind of got burned out, kind of difficult to get the AI to adhere with your prompts, but if they added Luma and flux, thats gonna increase my bill😂. Only thing now is i have to buy a laptop
It's a start, but "flow" is not what I'll call it when you have to generate in Flux and manually add to Luma. (I instinctively tried to link the output of Flux to the first frame of Luma) Also, there are no controls for Flux. As Einstein would have it, make things as simple as possible, but not simpler. Give us some controls under an "advanced" tab, if you're worried about scaring newbies.
Had a sub with kaiber for about a year was hoping the update would be a upgrade on their own workflow and some improvements aimed at people making music videos and lipsync updates etc . I don't like the new UI are workflow. Thankfully you can still switch back to legacy mode
can you do a start to finish video on the new superstudio, on how to make a old style kaiber motion video how to prompt video stle etc, the new kaiber super stataion isnt user friendly at all i cant even find where to prompt, get previews, select severity of the render, select in the style of....... so confusing compared to the old format
Kaiber's new interface is awful, extremely confusing. And this is coming from someone who knows a little something about ComfyUI and node based AI. I always liked Kaiber but what the hell are they doing? They should have just kept their easy drag & drop front interface and added the new models & generator options. And utilizing Luma at this point is disappointing since Luma took a step back recently and is nowhere near as good as MiniMax or Kling
Yeah, I get that. And if you're someone who uses Comfy, I can see it being extra confusing as you'll be thinking about it as a workflow that it actually isn't. That said, the old Kaiber is still there. You can access it via a pull down, and that is super appreciated. I'll say, there was one guy on X who said a lot of what you're saying, but played around for an hour or two and started to actually like the new workflow...so, it might be a case where it is so jarring, it might take a bit to get used to. TBH, I felt much the same when I first saw it: Like...what IS this? (Oh, and MM/Kling is coming soon to Kaiber...just FYI)
@@missoats8731credits drive me insane. Could mean 10 credits per used or 35. 1000÷30= 33 vids. 33 vids is a quarter of a 2 min video. (With the re runs,).
Eh, “bad” is subjective. I like the canvas idea. Perfect? No. But give it time and let them build off it. But, not every redesign is going to work for everyone, I get that. I’m still surprised at the amount of people who prefer to generate with Midjourney on Discord. But, I also understand, as the MJ website still has these weird UI quirks, where not everything is easily accessible.
You can still get to it; the old Kaiber can be accessed via a pull down menu. Which is awesome; I wish everyone did that. Funny enough, I just generated some stuff in Midjourney v1- why? Because they still have it around!
wait, when you say you get "flux, luma" etc. You mean you can use it with your account or they're giving unlimited generations with that amount of money?
yup. And the credits are going to be interesting considering the API calls. So, when they add in the other generators, they'll all likely have different credit costs depending on what the API you are using. Interesting times, for sure. I'm HOPING for a "FastPass" type system at some point, allowing for cross generation on the API platforms and the "main" platform. Although, I can also see that as being a logistical nightmare for the various companies to work out. Still-- if someone does, that'll be huge.
@@TheoreticallyMedia yep that would be great. I saw luma is 30 credits I think (couldn't try it with the 22 free) I'm using freepik mostly for mystic (wait and see if the second version is coming too)
Agreed. I mean, that most things to be fair. But, there is still a large segment of users who either lack the hardware to run Comfy, or who find the learning curve of Comfy to be too steep. The important part is that the engines to create are available to everyone.
@@TheoreticallyMedia doesnt comfyui can be hosted online? or i've seen there's an online comfyui recently, is it good enought compared to the self-hosted ?
Weird and warped? Not my cup of tea. I've been fighting weird and warped since this all started. Did you pay for all those credits? Or did they give them to you? It's customary to tell people if you've been given anything of value from the company you're covering.
No, no-- I think it very much is meant to be "Baby's First Node Workflow"-- if you step outside the Comfy world, a LOT of people are overwhelmed and confused by it. I think this is a great way of easily introducing the idea. It was smart to limit the amount of in/out points as well. They'll come with time, but for launch: Keep It Simple Stupid. This might just make node systems mainstream!
![เป็นไปได้ไหม - WanMai [Performance Video]](http://i.ytimg.com/vi/Eh5hItD4Mr0/mqdefault.jpg)
Check out the NEW Kaiber here: kaiber.ai?via=tim
Hey Tim, just a heads up, there is actually a great new feature on this platform, called Image BOMP; The info for which is rather buried deep in the image generation help section.
Essentially it's a means for a quick blend of any two images together directly on the Canvas, to generate a new image taking attributes/aesthetics and subjects from both. It works pretty well too, imho.
(With the Select tool activated, click on an image, and drag it over another image.
Hold the image in place until the bottom image turns green.
A new image will generate on the Canvas. The new image aspect ratio will match the bottom layer image.)
For TH-cam algorithm "Your videos are inspiring, entertaining, and suitable for all ages".
APPRECIATE THAT!!
I've never been so confused and turned off by an interface concept. Next time you talk to them, can you ask if there will be an adult version and not just this one for children?
KAIBER AI CUSTOMER SERVICE IS FIRST CLASS TOO VERY APPRECIATED
Love that crew! They’re so nice!
Thank you so much for making this video because there are many changes to the look of the Kaiber website and the video creation process. Your video has saved us so much time! Now we know what to do
Thanks for the update. One good thing about kaiber is that there credits doesn’t end on month end. It’s been two years I have them ❤ shall use it now.
Yup! Rollover credits are the best! And most ethical, in my opinion!
Same. I have a bunch of credits I can play around with on the revised platform.
Same here. I got like 130k credits from friggin 2 years ago?..I forgot all about this app😂😂
It's incredible to think we probably have 4-5 eras of AI video in 2 years.
I’m doing a talk about that tomorrow. The OG Will Smith? March of 2023. Think about that,
That’s crazy!
@@TheoreticallyMedia WOOW. That's hard to wrap your mind around. Will that be on TH-cam or is that in person?
@@TheoreticallyMedia I know! It's like evolution, but backwards.
All joking aside, I wouldn't have guessed it's only been 18 months or so since that abomination of a spaghetti video.
🎯 Key points for quick navigation:
00:00:00 *🚀 Kyber's Return and New Features*
- Kyber reintroduces itself with a significant platform overhaul, introducing new features, workflows, and a fresh design known as "super studio."
- Offers a unique generative design aesthetic and integrates advanced tools like audio reactivity, video upscaler, and creative canvases.
00:01:48 *🎨 Creative Workflow in Super Studio*
- New creative workflows called "flows" and tools like Luma's Dream Factory facilitate advanced AI-generated video and image creation.
- Enhanced features for organizing and iterating creative outputs using infinite canvases and AI-driven modules.
00:05:26 *🎛️ Organization and Workflow Efficiency*
- Introduction of node systems, collections, and buckets to efficiently manage and link creative assets, preventing disorganization.
- Modular approach to creating diverse creative outputs while maintaining workflow simplicity.
00:07:05 *🧪 Experimentation with Image Lab*
- Image Lab allows for custom artistic workflows and aesthetic alterations using various tools and prompts.
- Encourages creative experimentation and remixes within Kyber, demonstrating flexible cross-tool functionality.
00:09:09 *💰 Pricing and Platform Integration*
- Kyber's pricing model at $15/month provides access to various AI tools and platforms, emphasizing affordability.
- Potential integration of additional AI video generators due to the opening of major AIs' APIs, positioning Kyber as a competitive all-in-one platform.
Made with HARPA AI
Amazing - thanks! You've got a great personal way for these videos - and very informative, just a joy to watch, thank you!
Tim, this new Kaiber playgroung is really exciting actually ....... maybe the best thing yet in AI considering the potential to add additional tools and for a decent price for those like me who are on a very low, fixed disability income.
Agreed. Love that they clearly went in with some idea of how overloaded we all are. Plus: a platform that is built to grow!
That's a great step in the right direction, I love the NOD system, although at the moment it's a bit useless. what they really want to create is using different input as different effects. for example : image in nod 1 is the styling, image in nod 2 is the character, video in nod 3 is the motion, nod 4 is the color correction and nod 5 is a reference. and all these nods are effecting one video output
At this point it feels like it's the step to right direction BUT it was also done by product managers and tech people and this is not yet the thing that is going to be used. The right thing would be to invite/record people who are producing/editing AI generated videos for their projects and make something that streamlines that process. For example when I was working on a music video, I was generating images with Midjourney and then cropped the images, pull them into a Google Colab for facial syncing, export the results, then pull it into Davinci Resolve, and then mix the animated face back into the cropped images, edited the pieces yet again Davinci Resolve and so on. And I think for a long time this is how it is going to be: people generate pieces and stitch them together in a professional editing software. What the platform should provide the easiest way to assist the generation including pulling everything up to the same aspect ratio/resolution/frame rate + upscaling, color schema so you can just drop those into your editor.
I've been sitting on 3 months of subscription credits with Kaiber hoping they would improve things. Saw their announcement email and I'm glad I waited. I've been wanting to make music videos for things I've been working on, and the previous version of Kaiber was a little nightmare for wasting credits on bad generations.
The one thing I really appreciated about “old” Kaiber is how they provided you with few preview frames of your output before you committed to a generation. That’s something I keep banging the drum on. I want that to become a standard across all the generators.
@@TheoreticallyMedia Absolutely agree. In my experience with those previews, unfortunately though, they would change drastically in the final render. The first frame would look fine but then, despite the evolve level being 0-1, it would still have a bizarre and unusable outcome. Ideally a low resolution preview of all 4 preview clips would be great, rather than the still image first frame.
Old Kaiber seemed fine if you just wanted a lucid music visualizer, or something to spark a creative idea. I'm a big fan of Linkin Park, but feel like their claim to having used Kaiber to make their music video for "Lost" is a big stretch. Pretty sure they had proper animation done and overlayed with some Kaiber generation backgrounds. They used Kaiber for other unreleased tracks, and they worked as obscure visualizers but not anywhere near the level of the Lost video (song example: Healing Foot).
The video came at the right time, I had already been thinking about canceling Kaiber because I no longer knew what to do with it.😅Thank you! 😊
Excellent to hear!!
Finally a workflow that looks more intuitive and I dare say professional. The best T2I and I2V in the same place is a real timesaver.
Really looking forward to checking this out. We're getting one step closer to an all-in-one platform.
This is actually really cool and clever. Basically a simplified comfyUI but using existing online services like Luma as its nodes. That's really neat
All the new AI companies (I'm not sure what "back" means) want to offer you a "canvas". The thing takes 30-seconds to load a canvas. Just let me use it already. Also, you can't delete a canvas so this thing is just polluting your assets area. Why do none of these companies hire a UX consultant?
Eh, give it time. It's launch day. I'm sure a lot of that will be figured out as time goes on.
Can we all just appreciate that we are reaching the first end goal of video generation? ❤
Now movie studios have a great alternative for expensive visual effects
Awesome update 👍🏽 Its a bit confusing to figure out the best approach for image to video generation given so many tools out there!
I've been using Kaiber for a while now and I love it. Customer service is outstanding. I've created some great music videos with the product. Haven't tried the new version but I'm looking forward to it.
Kaiber did spin out of action , good to see them back .
8 years ago I shifted from adobe photoshop to figma, because of endless canvas, I feel this canvas gives you an ease of creative freedom. And definitely taking this initiative kaiber will standout in the market, it is understood for their current modules, but this is going to change the game of generative ai tools.
One suggestion, try to add some new examples , stop testing business man prompt :) more examples generates more ideas.
Woe, so Kaiber finally stepped their game up. Oh, you're the first I believe to upload a TH-cam video on Kaiber Ai. I was expecting MattVidPro or Matt Wolfe or even Al Samson. I like the direction that Kaiber Ai is going to. But like I had commented to you on a earlier video of yours, Runway Gen Alpha 3 and Kling Ai to me ( TO ME) And I say that because I know and understand that everything is relative and subjective. To me I still favor Kling Ai 1.5 to generate image to video. However I will play with Kaiber because I had created over 30 TH-cam videos using that program. But I'm heavily leaning towards Kling Ai along with some other programs to make everything professionL YOU can't JUST rely on Gen Alph 3 or Kling Ai. You have to use other programs such as Adobe Aftereffects and such for postproduction. Excellent video.
Well, I do mention in the video that we'll see some more generators in Kaiber soon-- and given Kling/Runway etc have APIs available now...welllll....I'd say keep an eye out!
Thanks for sharing! This is just the start 🙂
Can't wait to see what's next! And next...and next...and next!
ideal for batch creation i guess
Good to see Kaiber still about, and well done on keeping the pricing sensible (I do get though these things cost). Will be purchasing a month top play with very soon.
Great video Tim :D
At this point, I don’t care much about new AI video generators. I only care about new features that give the user more control. AI video generators have gotten better at producing aesthetically pleasing videos, but they still fall short when it comes to giving the user more control over the details of the output. How about some inpainting, camera controls, motion brushes, ability to pose multiple consistent characters together, the option to add a start frame & end frame, etc.? These are important features to make AI video useful for storytelling projects (e.g., making an AI movie).
agreed. And all of that is (eventually) coming. I always do like to note, that OG Will Smith Eating Spaghetti was March of 2023.
So-- we have come a long way in-- what, a year and change?
And a lot of that stuff exists-- it's just connecting it all together in non-janky ways-- and obviously, it is currently not always reliable.
I'd say by mid next year, we'll start seeing platforms that are integrating all of that and more.
Wow. I've been making AI videos and doing short films. I used Kaiber, but it was still a little off. I have a fun Space video on my TH-cam. You have provided me with some new ways to make my new one. Thanks!
That's awesome! I've got a deeper look at Kaiber coming up next week! Like I said in this video, I feel like I've barely scratched the surface with what this can do!
Thanks so much Tim for your great videos 🙏
krea added video creation with like 4 sources including runway and minimax … you should do a vid on that …
Planning on it! Love Krea! I often shout them out as the only place that is doing creative video upscaling. Kind of like Magnific for video!
Thx for the info!
Yeh, I have fondness for Kaiber as well. Love the new price: CHEAP. Can’t wait to see more from them. Thanks again, Timmah!
1000%! Kaiber kept it weird and I love them for it!
Hi Tim. Literally just got off Kaiber as I've been exploring it for the last week. I haven't even began to watch your new vid yet. Anyway, I'll shaddap a sec and watch your video LOL
Haha! That’s awesome’ old school Kaiber fan here and glad to see them back in action!!
@@TheoreticallyMedia I like Kaiber because of their unique style. Sometimes, those imperfections are great because you can create surrealistic imagery and effects.
Anyway, just wanted to say I more or less came to the same conclusion as you. I'm even at a crossroads now on whether to be a paying member since that is indeed a bargain.
Got curious after your cello comment and came across the album Out There by Eric Dolphy. A very fitting album name, but had some pretty interesting stuff.
Oh yeah, great overview of Kaiber too, haha. I hadn't played around with it yet
I'll have to check that one out! I listened to a lot of Dolphy playing with Coltrane, but I didn't dig too deep into his own catalog. And that album cover is kind of awesome! I'll give it.a listen today!
New Kaiber is interesting for sure-- I think we're going to be seeing more of these node systems popping up with creative AI tools. I still have to play around with it more-- It's funny, this is one of those situations where if you have preexisting knowledge of Comfy, you have to untrain yourself.
But, I think it's a great system to introduce the idea of nodes to new users.
That said, I'm pretty sure I only scratched the surface with what can be done here! Looking forward to spending more time w/ it (and getting some dev tips) next week!
❤❤❤
Yay tim a+ always on top of the best AI updates ❤ A+ for kaiber revamp sure best yet to come n n w mininworkflow dope the price is a flex
Appreciate it!! Yeah, I've always loved Kaiber for the commitment to exploring different artistic styles! I can't wait to really dig into the Image Lab area, because I think that's going to be a pretty powerful tool-- I just didn't have long enough to explore it for this video!
But-- that's why I really want to circle back to a bigger deep dive!
So Kaiber is the new comfy basically.
I think of it as Comfy Lite. Basically, a little starter pack that (I hope) makes folks a little more…ahem, comfortable, with node workflows.
So does this mean I can stop paying for so many different subscriptions? It seems to have everything as models of the other AIs I use. That is insane!
The New Update is not good your credits are spent before you know if you like your result or can even use it just eats up credits
I'm talking to them this week, so I'll bring that up. One of the things I really liked about "old" Kaiber is how it gave you preview frames before you'd generate, so at least you could see how things were looking before you committed.
That is something that I've been pushing on a lot of the AI Video generators-- I think that would go a long way: Either A) a few preview frames to start, or a low-Res version (like 320p w/ watermark) just so you can tell if it is worth generating.
so kaiber was, in fact, sleeping. until they heard the APIs drop
Tell Kaiber I want My Crystal back, My Lightsaber won't work without it.
; )
Haha, number one note is on the chalkboard!
yayyy you!!! Excited for this thanks for sharing.
“Finger Wonk” will be added to the dictionary.
I think the new benchmark for text-to-video should be: Spaghetti eating Will Smith.
I need to generate this today! Haha, that might be my new “Man in a blue business suit!”
Does that sweet UI work on mobile? If not, is there a workaround to access the features?
I just checked-- it looks like the Kaiber Mobile app is still the "old" Kaiber, but I'm sure they'll have a revamp soon enough. It'll be interesting to see how they port this idea over to mobile.
@@TheoreticallyMedia thanks for the speedy response! I checked as well. It looks like it might work on mobile, in "desktop mode" if one's device is up to it. I'm not sure mine is... Super cool looking tool. Thanks for sharing all the good stuff!
Very interesting, tthank you :)!!
@TheoreticallyMedia
Can you ask the Kaiber devs if they have plans for reaching out to platforms like Udio to see if they'd be willing to license access to Udio tools inside Kaiber?
Because then I don't have to pay separately for audio, image, video AI tools if I can have them all in one playground space
NICE VIDEO VERY INFORMATIVE..But what happened to the good old video transform function ?
It's still there-- I think that's under the Video Lab/Video Restyle flows. I'll dig a bit deeper into that when I get the chance to do my full walkthrough/tutorial on Kaiber. Hopefully within a week or two.
@@TheoreticallyMedia thank you very much, looking forward to it
Thanks for the update. Good explanation. Can you explain in future videos how to work with the API to generate videos.
I don't find the non-linear workflow appealing. Things can get messy even without all this stuff on my screen :D
I love the idea of the canvas and having my work flow in one place.. I have way too many tabs and folders open when creating music videos 🤦🏿♂️
8:35 Slender Man playing the Sax
Now THAT'S an AI movie I want to see!
Oh, must have missed this one. Look good.
Oh, it JUST dropped today!
@@TheoreticallyMedia gotcha
How do we know what version we have in FLUX?
I believe it’s dev 1.1. I’m pretty sure I saw that somewhere in the documentation. (Ahem, pretty sure)
Cant you make videos effected by audio anymore? couldnt find it
You can in the “old” Kaiber, which you can access via a dropdown. I know they’re working on a revamped version coming in a week or two for the new version,
@ thank you for the answer
question I made three music videos and then they changed the format AHHHH can we still up load music tracks ?
Who else here is still waiting for a video generation from Pika or Kling that they prompted 2 weeks ago? 😁
Kling just announced that they were having some issues w/ bad actors abusing the API, causing false flags in their system.
Check out their Discord-- they said to reset your password and they have a thread where you can put in your Username if you've been running into a generation issue.
Great overview Tim….i like this natural progression of platforms have a kind of intranet studio that you can tailor different laboratories of workflows in the same environment…. Ps could you do a feature on how to connect multiple platforms and a bit of ai automation of render tasks . Cheers
I was a huge fan of the Original Kaiber but I'm not impressed with what they have done especially for the price they are charging the video is just not up to current standards. Very dissapointly to be honest
Give it some time. Currently the credit pricing on Luma is obviously based on the API calls. I was looking around, and obviously the credits to generate are less for the old Kaiber model.
So, it’ll be interesting to see the credit costs of the other models they add in. (I’m sure Runway/Kling/Minimax are on deck…)
there''s a chance to put the aesthetic tools to VIDEO RESTYLE like to old kaiber? i need that to my video to video project., and the legacy kaiber now is not working super slow cant even generate
Btw Flow is available for comfyui
This seems similar to a Stable Diffusion + Comfy UI workflow
It for sure takes a note from a Comfy styled interface-- but, I think the smart part on Kaiber's part is to have a "Comfy Light" UI that gets people used to the node workflow, and expands on that.
You might be surprised how confused/overwhelmed people get when they see a Comfy workflow the first time.
This is a bridge step, I feel.
what flux model is?
I believe Dev 1.1
I've been using Kaiber for about a month, and I'm not seeing where we choose the animations in the new Kaiber. It's like a completely different Universe now, and I'm confused.
Gotta love that free plan. 10 seconds per month.😂
Pretty cool, more for the potential than the current execution. Played around for a bit, video generation is still average. Not quite sure I get the praise for their pricing though, you get 25 videos, 5 seconds each for $15....not great, even less if you're using it for flux images as well.
I’ve been looking for that style of website method for a website I’m wanting to make. The interface type.
Damn man i havent used kaiber in awhile, they were my go to, but kind of got burned out, kind of difficult to get the AI to adhere with your prompts, but if they added Luma and flux, thats gonna increase my bill😂. Only thing now is i have to buy a laptop
So, no more music clip then ?
Returning in a week or so? You can still access the old Kaiber via a pull down and use it
💀R.I.P. Kaiber AI🪦
It's a start, but "flow" is not what I'll call it when you have to generate in Flux and manually add to Luma. (I instinctively tried to link the output of Flux to the first frame of Luma) Also, there are no controls for Flux. As Einstein would have it, make things as simple as possible, but not simpler. Give us some controls under an "advanced" tab, if you're worried about scaring newbies.
Had a sub with kaiber for about a year was hoping the update would be a upgrade on their own workflow and some improvements aimed at people making music videos and lipsync updates etc . I don't like the new UI are workflow. Thankfully you can still switch back to legacy mode
can you do a start to finish video on the new superstudio, on how to make a old style kaiber motion video how to prompt video stle etc, the new kaiber super stataion isnt user friendly at all i cant even find where to prompt, get previews, select severity of the render, select in the style of....... so confusing compared to the old format
“Kaiberspace,” haha
Kaiberspace is a full on gift to the Kaiber Marketing team! I'm over here just handing them gold!! haha
how many credits costs a Luma video generation inside Kaiber?
Kaiber's new interface is awful, extremely confusing. And this is coming from someone who knows a little something about ComfyUI and node based AI. I always liked Kaiber but what the hell are they doing? They should have just kept their easy drag & drop front interface and added the new models & generator options. And utilizing Luma at this point is disappointing since Luma took a step back recently and is nowhere near as good as MiniMax or Kling
Yeah, I get that. And if you're someone who uses Comfy, I can see it being extra confusing as you'll be thinking about it as a workflow that it actually isn't. That said, the old Kaiber is still there. You can access it via a pull down, and that is super appreciated.
I'll say, there was one guy on X who said a lot of what you're saying, but played around for an hour or two and started to actually like the new workflow...so, it might be a case where it is so jarring, it might take a bit to get used to. TBH, I felt much the same when I first saw it: Like...what IS this?
(Oh, and MM/Kling is coming soon to Kaiber...just FYI)
so, for 15 bucks, do u get UNLIMITED flux and dream machine generations?
It says you get a 1000 credits. Whatever that means. But it's not unlimited. That would be insane.
@@missoats8731 ahhhh... thx. that sounded way too good to be true.
@@missoats8731credits drive me insane. Could mean 10 credits per used or 35.
1000÷30= 33 vids.
33 vids is a quarter of a 2 min video. (With the re runs,).
@@missoats8731 it's 40 credits to generate a video. So pricing is...not great.
It's taking 5 credits to generate one photo so you will get 200 pics if you only generate pictures
"groundbreaking"... Reselling api calls and adding a bad UX. So innovative
Eh, “bad” is subjective. I like the canvas idea. Perfect? No. But give it time and let them build off it.
But, not every redesign is going to work for everyone, I get that. I’m still surprised at the amount of people who prefer to generate with Midjourney on Discord. But, I also understand, as the MJ website still has these weird UI quirks, where not everything is easily accessible.
I thought this new update was garbage
3:01 HAHAHAHAH!!
Oh no, not the spaghetti strings.
Do you have a discord?
Am i the only one who thinks the update is terrible? The other way was way easier
You can still get to it; the old Kaiber can be accessed via a pull down menu. Which is awesome; I wish everyone did that. Funny enough, I just generated some stuff in Midjourney v1- why? Because they still have it around!
How about you make 1 min movie secne using only kaiber and shows us its merits and demerits
Probably do something soon. I'm prepping one right now, but likely using Kling for that. But I might sneak in some Luma/Kaiber.
Next one for sure--
40 credits for a video is not cheap at all. I don't understand why you say it's cheap.
even more expensive than Kling on professional mode, it's very expensive for a single video generation.
@@WanderlustWithT Yes, and not the same quality at all.
super node based 🤔
reminds me alllot of morph Studio. they actually had an awesome canvas style interface, though.. vid quality was eh
wait, when you say you get "flux, luma" etc. You mean you can use it with your account or they're giving unlimited generations with that amount of money?
Ok, got it. Credits system. Those remains or reset monthly?
It could be a good way to have Luma, not having luma : )
yup. And the credits are going to be interesting considering the API calls. So, when they add in the other generators, they'll all likely have different credit costs depending on what the API you are using.
Interesting times, for sure. I'm HOPING for a "FastPass" type system at some point, allowing for cross generation on the API platforms and the "main" platform. Although, I can also see that as being a logistical nightmare for the various companies to work out. Still-- if someone does, that'll be huge.
@@TheoreticallyMedia yep that would be great. I saw luma is 30 credits I think (couldn't try it with the 22 free) I'm using freepik mostly for mystic (wait and see if the second version is coming too)
Can't generate flipbook style video with this new superstudio playground.
Just tried it. What a mess!!!! I have been doing AI images for 2 years. This is the worst product to jump into.
Can anyone recommend an ONLINE AI video enhancer?
Krea has a solid one. Creative upscaler. Pretty cool
@@TheoreticallyMedia thanks, Tim
I don't see anything here you couldn't do with Flux in ComfyUI and an IP Adapter. It seems like they are charging you for nothing.
so, would you rather recommend comfyui over this ai?
Agreed. I mean, that most things to be fair. But, there is still a large segment of users who either lack the hardware to run Comfy, or who find the learning curve of Comfy to be too steep.
The important part is that the engines to create are available to everyone.
@@TheoreticallyMedia doesnt comfyui can be hosted online? or i've seen there's an online comfyui recently, is it good enought compared to the self-hosted ?
I'm trying to lip sync a non human character. What is the best AI program?
Weird and warped? Not my cup of tea. I've been fighting weird and warped since this all started.
Did you pay for all those credits? Or did they give them to you? It's customary to tell people if you've been given anything of value from the company you're covering.
Looks like comfyui for kids 😄 Sorry Kaiber #nohate
No, no-- I think it very much is meant to be "Baby's First Node Workflow"-- if you step outside the Comfy world, a LOT of people are overwhelmed and confused by it. I think this is a great way of easily introducing the idea.
It was smart to limit the amount of in/out points as well. They'll come with time, but for launch: Keep It Simple Stupid.
This might just make node systems mainstream!
@@TheoreticallyMedia Ah didnt know that. (Im looking now an old video about notes😱)
Thank you for the info💪
1st... 52 seconds after upload.... i do still have a Kaiber login, I'll try it again......#
You beat me by 3 seconds. Congratulations!
You win the chicken dinner!!!
First?
hmmm...second
Chicken dinner?
and it's not even dec. 2024 yet - much less dec. 2025...
It's crazy. I always like to point out that the OG Will Smith Eating Spaghetti was March of 2023. NOT that long ago!
You didn’t show any interesting kaiber style new videos? Only images
I'll cover that in the deep dive later! This is just the first of many videos, I'm sure. They've got a lot more planned!
1000 credits per month sounds like nothing. For me it's either unlimited or no dice.