New AI video model, full motion control, add ANY character to video, new text to audio, 3D heads

[@kipchickensout]

"look how detailed and sharp the video is" i feel like he always only cares about sharpness while the characters disappear and appear out of nowhere and half of the characters have movements that do not even make sense physically

[@jmg9509]

Exactamente

[@rielnezu1281]

maybe full of this video was made by an AI

[@The_Questionaut]

Yeah it has its flaws, and eventually those flaws will be ironed out. Freaky to think about

[@madmantheepic7278]

yea and for the 3d models from Dora, he keeps glazing how amazing the AI is doing when the expressions are forked, and the details are scrambled. Dude is hella AI glazer, like calm down.

[@Furebel]

Wouldn't even call it sharp most of the times he says it...

[@Dexta_mogger]

11:48 bro pretends like he did not click this on purpose

[@TheUnvanquishable]

HAHAH

[@JesusPlsSaveMe]

​@@TheUnvanquishable To everyone in this chat, *Jesus is calling you today!* Repent and turn away from your sins today to obtain eternal salvation🤗

[@FaHiM58]

Oh Myyy...😂

[@jmiles.]

bro couldnt hold back even in editing

[@Dexta_mogger]

@@jmiles. fr

[@vi6ddarkking]

It's funny how in the examples of the 3D Character generation there aren't that many in T-pose or A-pose despite that being how we want them for rigging.

[@fluffsquirrel]

But if it handles the rigging automatic and accurately, there's no need to use that step. I think that just helps guide rigging artists to create more accurate and less deforming rigs for pre-made meshes. Since quads and topology aren't an issue for AI, maybe rigging is the same way. It sees everything at once, 0-occlusions, 0-problems theoretically.

[@solarvoid7853]

@@fluffsquirrel Topology is definitely an issue for Ai.

[@phoenix-heartentertainment7293]

@@solarvoid7853 For sure that would be the next step

[@milo20060]

UVs and retopo 100% issue for ai.

[@HarryJPotter]

trust me, all you have to do is prompt for it.

[@CodyCha]

Hollywood is done in 5 years

[@cutthecheck]

I'd say 3 years

[@thegreatestmoo]

@@cutthecheck more like 6 months

[@AB-cd5gd]

Nah you can't say that if you know how movies are made, which is true is that it will cost much less to make special effects and certain scenes but you will never be able to make a real movie with ai even in 20 years lol

[@Jerry-thefckngman]

​@@AB-cd5gd naa more like 6 months

[@4.0.4]

​@@AB-cd5gd if movies were well written, I'd agree with you. Don't get me wrong, talented writers could make much better movies than AI. Sooo... Hollywood is doomed.

Your videos are amazing, but I think they could be even better if you included an estimated VRAM requirement for running each of those open-source models. ♥ Thank you for providing us with super dense informative content :3

[@NoahHornberger]

as an animator those trajectories look really strange in terms of motion, but I like the direction it's headed in terms of the overall concept. It just needs a little more understanding of how objects rotate around a stationary pivot, like a planted foot.

[@addy666street]

As an animator you should search for new job. As an amateur artist and animator I feel that world doesnt need me anymore. People around okay with it.

[@Chowderchefu]

@@addy666street Lol thats just how you feel buddy, I love being an artist and that wont change regardless of what happens. Dont be so weak-minded, ai will take over programming jobs, waitress jobs, art jobs, writing, etc. I am sure that may happen to some capacity in the near future. However there's no need to cry about it, you have that passion for a reason. Don't lose it over something silly, because as of now we still have our place in the workplace as artists and animators. That reality has not hit us yet, cross that bridge when you get to it. Rather than being too scared to even think of crossing it.

[@JustFor-dq5wc]

You're no1 source of USEFUL AI models on YT👌

[@Arc_Viper]

I'm really looking foreward to the pixverse vid. Looks really intresting

[@woof6367]

He knew exactly what wirh that thumbnail 😂 Thanks for always keeping us update on the latest AI news ❤

[@xandragonist]

"oh my what do we have here?" 🗣🗣🗣🗣🔥🔥🔥🔥🔥🔥🔥🔥 such inspiring words 11:47

[@Dr.UldenWascht]

I was already caught-up with the current news. But the ever-curious scientist in me had to click on that thumbnail.

[@Adam.Magyar]

This guy is good at clickbait... and to be true, he is also very informative. ;)

[@AndreaWatson-o7o]

It's better when the stuff actually exists for us to see and use for ourselves though, rather than just a readme

[@kowaihana]

​@@AndreaWatson-o7owdym?

[@lenny_Videos]

May i ask how you stay up to date with the latest ai news?

[@yesyes-om1po]

@@Adam.Magyar technically not clickbait... those LoRAs DO exist and they DO work very well

[@rjw8631]

great work as usual. a small request: would you change your browser to dark mode for videos? those bright white backgrounds are blinding to this ageing guy's eyes. thanks

[@lionelson7098]

Get ready, everyone! We're jumping into another session of 'Cool AI Stuff You Can’t Test Yet'!

[@JesusPlsSaveMe]

Jesus is the way, the truth and the life. Turn to him and repent from your sins today ❤️

[@Ablumz]

@@JesusPlsSaveMe i sucked off jesus just the other day - he was chill

[@AiFiImAcademy]

Amen to this! Im on too many waiting lists with no real testing yet 😢😂

[@dwainmorris7854]

Good stuff in the right direction but they need to add the ability to change the poses and actions of the figures not just the path it moves in as well as the expressions on human faces like with AccuPose AI

[@DiceDecides]

4:22 cool but the bird in the center got red pants too

[@thewetpen]

Funny that it gave the Koala a tail, most of the time.

[@lemontangs]

That's what we call drop bears here in OZ.

[@capoman1]

24:40 Oh I could see this being used on pro sports; for instance capture all the NBA player animations from game footage. Then NBA 2k or whatever can have real captured animation per player.

[@Threelegs94]

My mind says no, but my body is telling me yeah

[@therandommusicguy4773]

Brother chill lol

[@coreytaylor7367]

Can't wait for all the anime baddies 😎👌

[@handznet]

Dora looks awesome. Need to try it

[@meowwoof5408]

What video editor is this at 16:19?

[@FunDumb]

Much appreciated 👏

[@VaibhavShewale]

damn that looks awesome

[@fitybux4664]

1:30 I could absolutely see this being turned into a video game at some point less than 10 years, once we have the hardware grunt to do all of that in real-time! 😲

[@greenstonegecko]

6:40 Very good compared to the other, but still... I can hear 2 or 3 basketballs bouncing. Just a nitpick but it's crazy realistic sounding.

[@abdessamadbilal-ox5qj]

what we could do with last tool wan you attache image or other 3d models to the captured motions?

[@MaxSMoke777]

Dora would be useful as a base for converting concept art to a 3D model... but hardly for end-use items. It has the same issue as all automatically produced models: It's Sloppy. It has too many polygons, has no refinement, overlooks obvious optimizations that an experienced modeler knows. But it would be very helpful to get the basic shapes started. Reminds me of just about all AI generated art, it's a nice but a hot mess that needs professional help to really make shine.

[@ianwatson5767]

The thing is 3D models outside of video games/VR are used purely for visuals, so for making something like animations, short videos, VFX, etc, AI can just generate videos. I think Ai generated 3D models are kind of a pointless use case right now.

[@Apostate_ofmind]

this is what grinds my gears when someone says AI replaces this or that other job. It doesnt. AI gives you a base, but you need a professional that knows what hes doing to refine it into what we are accustomed to. And that is someone you have to pay for (or do yourself). Ai does not replace artistic jobs. Ai automates the grueling parts that NO ONE wants to do, but that gets relegated to some people that arent talented enough to do something else. No one studies to only do 'inbetweens', but when you get hired in a company, the higher ups dump those bits on you instead, which is good for initial training but then becomes miserable.... Wasted talent. That now would either be free to help with actual artistic endeavours, or be cut from the budget, which means more smaller companies can now afford to hire those youngerlings to do creative stuff, as they are saving on having to hire twice the people.

[@276-]

11:45 "oh my what do we have here" BRO???? 💀

[@evam7964]

For the first time I beat this channel and learned about the new Pixverse before this video came out.I was pleasantly surprised as to how much it improved .

[@israel1231]

😩 now available to the public!

[@ranibro]

Freaky Artificial Intelligence

[@Recuper8]

Is there anything like the first one where you can move the camera how ever you want? Even advanced things like 'dolly zoom'? Or, does that not yet exist?

[@narachi-]

you can render multiple camera angles and zooms then dolly in post?

[@Recuper8]

@@narachi- I'd have to be smarter to pull that off.

[@theAIsearch]

yes, i think i mentioned it in one of my news videos. maybe this one kzbin.info/www/bejne/oImmdIedbqyFo80

[@BinaryBandit4856]

Very interesting stuff. I've been waiting for tools like these for decades. Finally they are here! Subbed.

[@budekins-q3s]

They are here but. .you need a Huge amount of computer memory to use these. .

[@theAIsearch]

Thanks

[@plainpixels]

Microsoft released a 3D pose detection paper / videos that look similarly ready to shake up AI mocap- except neither that work or this work mentions anything anywhere about actually releasing it to anyone ever.

[@Irokusion]

10:35 Well we finally here after few years of developing AI industry we finally have uncensored consumable version of video generator (8 GB vram and less) for generation any Corn... What time to be alive!

[@victorgarciasilva8183]

Hi AI Search project, would it be possible to activate the automatic youtube audio transcription to spanish or another language, the whole world will thank you for it.

[@Thebigpier]

Well, these technologies will allow indie video games to advance in quality almost to the level of triple A, usually the things that you can clearly understand that it is a low quality video game made by a small team are the poor polygonal models and the ridiculous animations, but now even with a small team you can do remarkable things!

[@jzwadlo]

1.4x speed for this channel is perfect

[@jasonhemphill8525]

Good call.

[@PowerRedBullTypology]

The tool I'm lookingto forward most is a sort of tool that combines a lot of these tools into 1 usable tool. At this point it's like everyone s designing one feature of phtooshop. One does clone stamp, the other does brushes, the other the sizeo of a project. It will be really neat if more concepts or ideas would be integrated into 1 workspce where you can kind of control all of them.

[@juanjesusligero391]

ComfyUI integrates almost all of them, though.

[@Cyborg_VideoEditor]

@@juanjesusligero391 EXACT

[@SpartanWolf120]

Is the person making this video AI?

[@oskarmohar6073]

Im sorry, but the quality of the "3d personalized avatar" is awful.. 1985 Max Headroom was better..

[@31eGGy31]

GenHMR immediatly gave me the thought of making funny videos with it by replacing the actual persons with like Voldemord or such, which then plays basketball against Harry or sth. Have you stumbled upon a good tool, with which could realise that already?

[@TalentedKamarty]

GenHMR will be perfect for games. Especially modders.

[@lukasgruber1280]

Thumbnail of the year and its only early january.

[@JayGee1]

Great videos. Are there any models like audio2face that can animate the avatars in realtime ?

[@PowerRedBullTypology]

That Koala is Jesus..he walks on water

[@GamerKittyEuphoric]

they really messed up that koala. lol koala have no tails XD

[@orac229]

Seems low quality, great to be able to make the object move but... the lizard moved sideways and skipped some steps, no foot prints in "most" of the snow and water shots, but you have made each window so small it is not clear. We still seem to be just using advance video editors, nothing I would label as "AI". As a side note I was using Gemini and asked it to make a list of 200 songs (of a specific type) and it gave a list of 10 (further chat continued to fail correctly make the list). I asked for a BPM of 135 and it gave everything from 90 to 160 BPM, really really disappointed so far with what people are calling "AI"

[@Sokratesin_Çaybardağı]

you are a good channel

[@samedits_SS]

Nah man my future plan to be a animator editor is gonna end before it starts

[@Author_SoftwareDesigner]

You can be an Ai editor / Animator

[@madmantheepic7278]

@@Author_SoftwareDesigner man, now that's a depressing idea

[@swannschilling474]

I am really enjoying these round ups! 😊

[@chesswarlock9384]

I for one, welcome our new robot overlords.

[@Rainbowhawk1993]

All hail our lord JC Denton.

[@kinesslop651]

Me too

[@JesusPlsSaveMe]

​@@Rainbowhawk1993 Where are you going after you die? What happens next? Have you ever thought about that? Repent today and give your life to Jesus Christ to obtain eternal salvation. Tomorrow may be too late my brethen😢. Hebrews 9:27 says "And as it is appointed unto man once to die, but after that the judgement

[@3D-User-64]

@JesusPlsSaveMe Dont be annoying with this Jesus sh!t

[@evelingbedoyasalazar2564]

​@@JesusPlsSaveMeNo

[@Lv7-L30N]

Thank you

[@theAIsearch]

You're welcome

[@MichikoAsuka]

Oh wow ai is evolving way too fast it will be everywhere soon i wonder what new jobs are coming

[@loupasternak]

workers at the Unemployment office to handle the influx

[@Rilex037]

this guy uploads a lot

[@BlackTakGolD]

You make me unload a lot.

[@bocheraATC]

@@BlackTakGolD 😭

[@Cyborg_VideoEditor]

HE NEVER SPLEEPS

[@tharixsta-gaming6440]

felt like i was watching a tutorial not a Video about AI

[@jboss1073]

So now we can control lizards and rhinos in the desert automatically but we still need to manually code web front-ends?

[@sgtKrud]

Some people don't want to view this imagery on the thumbnail, I'm gonna have to hide your channel

[@MabelYolanda-c9i]

9:10 Gret Video, but it’s a mistake to have all the foley (sound design) in one track. We usually have 48 to 96 tracks to do what we call M&E (music and effects tracks). If you have everything in one mono or stereo track, you can’t control the mix, therefore Hollywood is not over as many suggest. Why? Because this models are written and trained by computer guys, not by sound designers or post production professionals. So to me, this particular tool is a step back within the fantastic world of AI. I’m tired of advocating for more involvement of real artists in AI, but the developers only listen to KZbinrs, not real professionals. It’s obscure and you can predict how things are gonna end. Monkeys driving a Ferrari.

[@Andricx]

When I asked Pixverse 2.5 for anime style, the first frame looked okay, but in the video, the anime style was forced to be realistic. GET O-

[@HunterSpace-l9n]

Not the gooning😭

[@kipchickensout]

genhmr swapped the identity of some of the poses midway

[@MohdYaseen-q6k]

My question is should someone start their bachelor degree in CS? is it worth it like for me I'm a total beginner or I should better spend my years of Psychology line?

[@JustChillinOnThe5thFloor]

That war scene from 1917 was epic. And if the AI can properly track all that is happening is crazy.

[@740dailyolivia]

Every step you take inspires others, I'm genuinely enchanted by your video! - "Challenges become stepping stones on the path to growth."

[@hansihintersoos7581]

These 3D model generators do a good job with images that already depict a 3D rendered character. If you use real images or anime the 3D model by far isnt that good as in the video

[@raulp2521]

BEST click baits ever

[@DD8842]

Montie Oum would be awesome with this tech if he was alive

[@ArcanePath360]

All very impressive. All very useless since it's always just "A paper is released" and nothing ever comes of it Starting to think half these are just fake

[@computron1]

Exactly. It's all vaporware unless they release the code.

[@byujkt3822]

And next day they release paid ai webs generator

[@DuCaffeine]

❤❤❤❤❤❤❤

[@johnny5240]

What I miss is a common platform for the latest tools. With upgradable modules. Now there are too many cool tools that have to be used separately.

[@justinnine4940]

if there is, the antitrust law would rip it apart

[@cefrio]

So can i use Pixverse to upload a short driving video to let ai change this vid? Ore do you recogmend another software?

[@LorenElara]

Is there a good video to video AI right now? Where it can replicate the models. I've seen some low-quality ones, but looking for something good.

[@Pentafant]

22:30 just showing the model alone does not mean anything. As long as youre not showing the poly count and face distribution, these models are worth nothing, because thats the most important part about 3d design.

[@neorin]

Literally seconds later, it shows you one of its models animated in Unity. Which already irrefutably proves it has a minimally acceptable topology; otherwise, the rigging and skinning would break the model when moved.

[@Pentafant]

@@neorinTrue, but who knows how good this model would run when used in a game, together with other AI-made Models. Optimization is extremely important in animation and game design. So topology is not only about how good a model can be rigged, but how many faces it has and how they are used. This model could have a million faces for what we know, which would crash any game on a mid lvl pc. Or would make animation nearly impossible, cause rendering a single frame would take hours. So as long as the topology is not shown, I really doubt these models are really useable for production.

[@loupasternak]

@@Pentafant you really think that won't be handled gracefully, if not now, then soon ?

[@Pentafant]

@ maybe, but it’s still not even close to that stage and he talks about it, as if it is. Also that would be another hard shot for real artists. So I hope it will still take some time, or never manages to get to that stage. Because it seems quite hard for AI to get an understanding of production ready Modeling.

[@nuca5104]

The AI hangover is incoming I'll bet my money on it.

[@pandamanou]

23:58 finallyyyy i hope this will help some people to actually propose a better game then ea sports fc smh

[@eli_0625]

crazy thumbnail

[@mariopadilla1445]

The only problem with huanyuan is that it requires so much vram. I’m hoping it becomes the sdxl equivalent for video. Hoping to run it efficiently on my 3060 12gb

[@alkeryn1700]

uh, that's a good channel, straight to the point, no bullshit, just the information, thanks !

[@MichaelBaynana]

tangoflux cant give out words : (

[@MichaelBaynana]

i zero understood the ai video composer

[@yusufemrealbayrak2567]

last one is crazy. think this game industry. 3a game about animation.

[@Dis3spectful]

i came for the catgirls

[@ChingueSaganaki]

12:08 --> Don't wear a mask like some of the people in the back, in the scene called "noa (blue archive) ...", only fascists do that 👎🤦‍♂

[@louiewashere3]

2025 is going to be crazy, I'm glad I got me three more 12 Tb hard drives LOL

[@reezlaw]

Not the astronaut riding the unicorn again, what is this, 2022?

[@Nrek_AI_Productions]

You're the absolute MVP my guy. You saved a project I've been working on. Thank you.

[@ThyUnholyBacon]

heck ya

[@stonepiratilla]

Hunyuan not working its say No such file or directory: "ComfyUI_windows_portable\\ComfyUI\\models\\LLM\\llava-llama-3-8b-text-encoder-tokenizer\\model-00001-of-00004.safetensors"

[@ksagosen]

Generating a Lora trained on real people should seriously be illegal. This is just dystopian.

[@imluque]

It's scary and dehumanizing really, it could ruin lifes because of the gooners that could get their hands on these tools I hope there are regulations and restrictions, but i doubt it would happen without the need for something extremelly tragic and impactful to happen irl

[@ksagosen]

@@imluque They are going for celebrities with a lot of power and money, so I'm pretty sure there will be people lobbying for to make it illegal to post material like this. And I get that OP is just reporting on news basically, but the lack of critical commentary is disappointing.

[@madmantheepic7278]

@ksagosen "this is amazing" "highly accurate" "extremely detailed", on repeat, riveting commentary honestly.

[@udayrajpatel9048]

Mindblowing video, thanku so much. ❤❤❤❤❤❤🎉🎉🎉

[@budekins-q3s]

The problem is you need a very expensive computer with a Huge memory to use some of these A.Is.

[@JesusPlsSaveMe]

​@@budekins-q3s *Revelation 3:20* Behold, I stand at the door, and knock: if any man hear my voice, and open the door, I will come in to him, and will sup with him, and he with me. HEY THERE 🤗 JESUS IS CALLING YOU TODAY. Turn away from your sins, confess, forsake them and live the victorious life. God bless. *Revelation 22:12-14* And, behold, I come quickly; and my reward is with me, to give every man according as his work shall be. I am Alpha and Omega, the beginning and the end, the first and the last. Blessed are they that do his commandments, that they may have right to the tree of life, and may enter in through the gates into the city.

[@AB-cd5gd]

Can you do testing of nestmind thinkflow like compare with o1 and google thinking to see the best reasoning model

[@CERITAKASUS70]

i like this channel i like ai

[@FacebookHacks1]

You’ve literally become my weekly source for ai news, keep up the great works!❤️

[@theAIsearch]

Thank you!

[@marwanamv355]

insane

[@Atomic-Purple-Guy]

13:24 my vtube model with be crazy

[@BinaryBandit4856]

Goodbye movie industry. We no longer need you.

[@Nodsaibot]

where the cat girls?

[@fernandoz6329]

Great video as always! GenHMR blown is very promising, you can translate videos directly to 3D, the possibilities are incredible.

[@mksavela]

12:50 PERSE means ASS in Finnish 😆

[@mathbrah]

ai freaky update