How This Stunning AI Animation Changes Everything

[@ROHKI]

Really appreciate the shoutout! Thanks for all the love on the episode :)

[@TheinterfaceTvSeries]

When you talk to the devs please tell them that “a man opens a door, walks into a room and sits down at a table” is the holy grail of narrative ai storytelling!

[@TheoreticallyMedia]

Oh, for sure. Multi action, multi location? Weirdly enough, that’s something that we might eventually solve via that AI DOOM (the game) video I did a few videos ago. If you think about the underlying technology there, you’ve basically got just that: action and location.

[@TheinterfaceTvSeries]

@@TheoreticallyMedia Oh, yeah! You need hella processing power for that, right?

[@DavidGreen-n1s]

No way,....the REAL "HOLY GRAIL", is NOT having to "EXPLAIN" a result before achieving it❤

[@DJTechnosapien]

Hey man, I just wanted to thank you. I haven't commented before but I've been watching for about ~5 months..? You've given me so much information and inspiration to create than most people in my life. Just by sharing these videos. You are a G, I look up to you. I started using Udio b/c of your video on it, and I have not stopped ever since. I recently have been using Runway and so many other AI tools you've shared to create these lil music videos on my channel. I know myself and so many others are doing the same thing - it really gives me hope being in a community that supports innovation

[@markgoodman2801]

Thanks for being ROHKI to my attention. I just checked out the full episode and WOW!! Especially love the music/visuals at 8:00. Banger for the eyes and ears.

[@TheoreticallyMedia]

That's one thing I forgot to talk about in this video: The sound design is EPIC. It's so good!! Glad you dug it!!

[@eddysgaming9868]

That was a lot of material to cover, huge advancements. Thanks, Tim.

[@TheoreticallyMedia]

Thanks so much!! Yeah, it was a crazy week! Still Friday to go too!

[@zrakonthekrakon494]

Consistent characters in video should be one of the most important features for the big video generators, not sure why it’s not more quested or prioritized by the likes of gen 3 and kling

[@TheoreticallyMedia]

Overall, it's harder than we think. Devs have been working on solving it-- but by nature of Generative AI, in that it hallucinates a lot-- keeping a consistent character is kind of tough. I've seen a lot of interesting work being done on the problem though. It'll be solved soon!

[@amkire65]

I haven't used Vidu for a little while, and the new feature got me back there again - also my credits have been updated since I was last on Vidu. The reference feature is great. I took a picture of a female folk music singer that I'd created, "targeted" her as the reference, and then I was able to have her singing in other locations. As far as I can tell she looks exactly the same down to her costume. Also, the singer wasn't playing an instrument in the original picture, but I've since changed her from singing in a field to playing a guitar while singing on a beach. Wow.

[@TheoreticallyMedia]

Nice!! I'm glad you got a chance to experiment with it and had some success! Haha, the Batman dance was clearly not the best example!

[@noelwos1071]

Pls Leave the links in the description that would really help 😊

[@evam7964]

I watched one video now addicted to everything you put out

[@ARealPersonNotABot]

I gotta tell you. Id be lost in the AI world if it weren't for your guidance. Your videos are to the point, informative, funny, and its wild how I just started watching you about a year ago, and already it seems like AI has advanced 10 years. Wild times man. Btw Im making a super short AI film. Id like to share it with you and your audience

[@TheoreticallyMedia]

Haha, I know-- I don't often (or I try to avoid it as much as possible) watch my old videos, but every once in awhile, I'll have to go back to check on something I previously talked about. And it's always stunning to see where we were, just 6 to 8 months ago. It's just crazy. But, hey: thank you so much! And please: Send over your video when its ready!

[@giacomobonavera]

Blocking out time in the calendar over the weekend to watch this. Can't wait

[@TheoreticallyMedia]

It's a pretty good one! That Rohki video is wild! One thing I didn't touch on in this video is the insane sound design in it as well. When you get a chance to watch it, make sure you've got some good speakers or headphones to listen to it on!

[@giacomobonavera]

@@TheoreticallyMedia Thanks for yet another great update! Just watched your video and must say I am very keen to check out Rohki's short film. Thanks for the direction!

[@ufologistsguide]

Hi Tim, another great video, thanks. Just for your info the free version of Kling is not working at all, and will hit 99% and keep you waiting FOR DAYS! However the paid version is ok and takes approximately 2 to 3 minutes with pretty awesome results, and you can see it generating slowly, instead of immediately jumping to 99% and staying there for eternity. Also to note, is if you create in professional mode it takes up 35 credits, not 10, and to be honest the quality is not a great difference between professional and standard mode. Keep up the great work. Steve.

[@TheoreticallyMedia]

Hey Steve! Agreed on that Pro vs Standard. Where I think Kling (might) skyrocket is when they introduce the "editing" tab that has been grayed out since launch. On the chinese mobile app, they have a whole suite of features (including upscale to 4k, and color grading) for the videos you generate there. It's insane! But, both the Web (US and Chinese) versions don't have it yet. And yeah, that's unsurprising that the free version takes forever. I mean, it's probably sitting lowest in the queue-- and depending on demand, I'm sure it gets lost at some point. On the plus side, at least Kling has a fairly reasonable monthly rate.

[@ufologistsguide]

@@TheoreticallyMedia Yes Kling will get interesting once its up and running. I maybe late to the party, but I just found 'Katalist' AI story telling script to movie. It's quite awesome what you can do with it, it should be right up your alley. Have a good weekend.

[@CringeCrafter-uv1lt]

I'm learning ComfyUI after putting it off for a while and sticking to comfortable A111/Gradio apps because they were easy, but to do the things I wanted to do, Comfy is necessary and maybe a little command line/python toil. Since I'm fairly committed to open source, I'm really happy with what is happening these days with Comfy and the wrappers that are being put out there. Hopefully, one of these days you'll be able to highlight an AI generated film that I put together. Anyway, thanks for the video. It was really inspiring.

[@SitarSangeet2024]

Thx a lot for your videos, very enjoyable and instructive!

[@TheoreticallyMedia]

Thank you so much! and thank you for watching! Appreciate it!!

[@sidewaysdesign]

Thank you; I have always wanted to see Batman doing interpretive dance.

[@TheoreticallyMedia]

Haha, we now have a go-to generator for our Batman/Joker dance off!

[@AEFox]

I am doing some shorts also, using just Flux, Roop, Kling, Eleven Labs, Davinci Resolve and a little Photoshop. I'll let you know when I finished some to show you. I am a Team of ONE lol!

[@High-Tech-Geek]

Thanks Tim! I like the motion brush concept although I didn't quite understand how Runway's selection tool applies to motion paths.

[@TheoreticallyMedia]

I might have fudged that part-- what I was thinking of was the inpainting brush. Didn't Gen-2 have that? Or...was that another one? It's all starting to become a blur...

[@ECHSBACHS]

Lots going on this week. : )

[@TheoreticallyMedia]

I know, right?! It started off slow. On Monday, I was like: What am I going to talk about this week? By Thursday, I've done 2 videos and I could probably do another one!

[@FramerEscapes]

Great video thank you!

[@TheoreticallyMedia]

My pleasure!

[@GeorgeVenturi]

Awesome video Tim!!

[@mattstory2011]

Thanks Tim!

[@StephenRansom47]

I keep hearing about this ‘duration’ expansion issue … When this technology breaks the 8 minute barrier, the AI could be called FILM MAG. In honor of the old film can barrier. 😅 … thanks for the news. 🥂

[@TheoreticallyMedia]

I love that Film Mag idea!! LOVE IT!

[@churchsantana9462]

Batman’s interpretive dance is crazy 😅

[@TheoreticallyMedia]

Haha. It’s pretty hilarious! A good Batman/Joker dance off could be a viral hit! Yeah, I think there’s something “there” with Vidu. It’s got a secret sauce weapon, I just didn’t find it in my very limited testing!

[@relaxandmeditatestation]

Thanks for your work! I love your reviews!

[@LoopnMix]

haha 8:06! Saw what you did there haha

[@TheoreticallyMedia]

Haha, Vidu was like, "I know what ya'll want to see here!"

[@LoopnMix]

@@TheoreticallyMedia 😂😂

[@Ai-Art-42]

Interesting, as always, good job

[@TheoreticallyMedia]

appreciate that!

[@jzwadlo]

Really want people using these early models to start putting in weirder prompts so we can see how these things think! EG: A cat drinking a sandwich Stuff like that might help us see how these things think cos obviously, we don't even know how that would work so it would be interesting to see! *Just a suggestion!

[@TheoreticallyMedia]

Oh, that sounds fun! I'll pop that into the next set of examples!

[@jzwadlo]

​@@TheoreticallyMedia I think it will be interesting as it's definitely getting tiring seeing all the other channels just review the new models with 'oh lets see how it manages a cat on a window ledge'.... Also the other AI channels that say 'x' is better than 'y' whilst testing a generic prompt that hasn't been formulated specifically for that generator using provided documentation is also something that's starting to blend all the other AI channels together a bit. Long live TM!

[@Apuat]

🤣 just (sort of) took the leap from Premiere to DaVinci earlier this month for a film I was hired to post. Then 48 hours ago Adobe announces new color correction features that automatically detect what flavor of LOG you shot in and automatically converts your footage to Rec 709. You can even use ACES. Proving once again, that working with Adobe products is sort of like joining the mob. You can never really leave the family. Was that piece of footage at the beginning of the Episode of the gentlemen walking towards the tree with deeply muted colors part of a longer project created in AI?

[@jzwadlo]

08:10 I see you Tim ;) As does Vidu 😂 PS: Feed this man VIDU! He has important research to do for science 😎

[@GRNKRBY]

Having fun with Personalization in Niji MidJourney now.

[@LouisGedo]

10:14 Awwww, adorable critter in neat scene

[@RockyBergen]

Hey, I'm in Marketing. I am interested in that stuff too!

[@TheoreticallyMedia]

Haha- I didn't mean it that way! I did some time in creative marketing as well! To be fair, there's a pretty good crossover!

[@gumvue.studio]

thanks for this news, I upload yerterda a scene from a movie.. very close to be a real shot audio also did great, hope you enjoy

[@domehouse79]

Brilliant video

[@jraiber]

Hi, can you please tell me, what is the name of the company, which develops the UI for Ai-Filmmaking you mentioned with the rhoki movie? I ask, because we try to do the same. :)

[@TheoreticallyMedia]

Sure-- but just so your aware, I think they're running fairly stealth right now-- and the platform isn't actually out yet-- but it's called Ikhor.

[@TheBlessingReport]

Go Tim

[@Sasha_sementsev]

Thanks 🙏

[@LouisGedo]

1:11 WTF??? My brain just got injured! 😆 🤣 😂

[@TheoreticallyMedia]

Haha, right?!

[@barryobrien1890]

It will be interesting to see what Sony does with its creative upscaling on the PS5 pro. I get the feeling we are still at the tiktok usecase level for video ai. Huge market but not very useful. Looking forward to the next generation of tools

[@wielandsmith]

I ditched premiere in 2020 and moved to resolve. I do love resolve though the premiere ai is tempting.

[@TheoreticallyMedia]

I'll probably ping pong between both for a bit. At the end of the day, I always say that you edit with your brain, not the tool. I've used Resolve a few times and there's a LOT I like about it-- but, at the end of the day, I'm just faster with all the shortcuts in Premiere. At some point, I'll either be comfortable enough in Resolve-- or, I'm just going to confuse myself and use all the wrong shortcuts in both!

[@MysteryFinery]

my personal workflow is making pure 3d render, sometimes with photo manipulation if I am feeling its needed. then using those as reference. not bad results at all, blown a big amount of money on this though. it isn't cheaper then the 3d stufff surprisingly but it is faster.

[@CanadianVader]

Meanwhile back at Sora AI *crickets* back at Chat GPT *crickets* I think they have fallen so behind, and all these other companies are basically killing it with constant updates. Making things even better and better.

[@EyadishoKahawaki]

00:04 what this style is called? or what prompt is used

[@themightyflog]

How did they do the mouth work?

[@TheoreticallyMedia]

I'll get details on the for the bigger breakdown on Epi 2 when it releases!

[@nicheva417]

I'll never be touching an adobe product ever again. There's only so many times you can buy the same program.

[@TheoreticallyMedia]

Oh, I know. I get it. I kicked and screamed for a long time after CS6 ended. But at some point, I just gave in. Still, totally get it.

[@DavidGreen-n1s]

I will be as HONEST as possible. If we FEAR something, and STILL persist in CREATING "IT", who is to BLAME for its ARRIVAL?

[@krishnaprajapati-hz8vl]

Link in your discription bhi dala kariye sir😢

[@LouisGedo]

8:06 😆 🤣 😂 and what might that be pray tell?

[@TheoreticallyMedia]

Haha. Vidu knows!!

[@muhamadabdallah7960]

you're funny man. thank you.

[@frank-f4w]

someone tell runway to let negative prompts in gen 3 they simply wont listel tell them pls

[@agnesslovehealz]

Can u do video on vidiu w character style reference when get credits

[@paulsheriff]

Hi love your channel can i send you a link to a video i made using Ai noisee ai ..think it turn out great ..i did put link on but ovs it got removed

[@LouisGedo]

👋 hi

[@TheoreticallyMedia]

Heya Louis!! I think I still owe you a response on yesterday’s video! Man, it’s been a grind this week!

[@PakhtoonsGamer]

First and where to run Flux free private

[@TheoreticallyMedia]

Winner!! For Flux, beyond locally-- I mean, you can run it lots of places that don't have a gallery. Fal I'm pretty sure is pseudo private. That said, anytime you're generating on any platform, it's never really private.

[@jjsc3334]

No gallery accepts AI art, what is wrong?

[@MabelYolanda-c9i]

Very disappointed with Vidu’s update. The one thing that made them unique is gone. And as far as I know, there is no other AI video app with character reference…..sometimes I wonder who advice these startups…..I’m sure none of them are advised by real professional artists……anyways, Resolve is about 1 billion miles from getting close to Premiere, as editing platform. I’d never forgive them from destroying the Linux/Da Vinci hardware/software used to color correct celuloide…..

[@MrNobodyX3]

When you talk about the process, please don't assume I know what the hell you're talking about. "yeah that viggle" I don't know what that is. What is viggle?

[@TheoreticallyMedia]

Haha- gotcha! Yeah, that's something I'm working on-- these days it's a tough balance of "this is the new thing" and often I'll end up referring to the "old" thing that a lot of you might have missed. Anyhow, Viggle is a video restylizer. I covered it about 5 months ago here: kzbin.info/www/bejne/ipnPZGuCq52spacsi=r-T8ffoQvni9IuEC The model has improved a lot since then though! You know what I need to do is put up a website with links to all the resources I cover on the channel. hmmmm, weekend project I think.

[@Paranormal_Gaming_]

Hi Tim!! Great video!

[@agnesslovehealz]

We can use comfi in pinokio as well but less options and bells.whisles been looking into it it's ughh kinda feeling lol so rather pay and have easier capability

[@TheoreticallyMedia]

Yup! Pinokio is basically where I learned Comfy-- but, I also didn't want to open up that can of worms in this video, since every time I do a video featuring it, it turns into 400 comments of tech support on how to install Pinokio! Haha. Plus side: Currently the OpenArt Comfy Cloud version is totally free to use. To be fair, it's a little broken (I went over that in the video I did on OpenArt)-- namely, when you try to use other people's workflows-- but, it works just fine if you are building "from scratch"-- So, I think it's a great spot to start learning!

[@lineardrift]

Thanks Tim!!

[@TheoreticallyMedia]

100%!!

[@frank-f4w]

someone tell runway to let negative prompts in gen 3 they simply wont listel tell them pls

[@TheoreticallyMedia]

Have you tried (keyword) or (((keyword))) I can’t be sure, but that’s an old technique that might work. Actually, I’ll go try it out