New AI video generators, AI detects cancer, Robots with muscle, AI animation tools

[@theAIsearch]

Level up your academics with Aithor.com: bit.ly/AIsearch_Aithor Sign up for a FREE & use code AISEARCH10 to get 10% off a Pro subscription (code expires: 11/17/2024)

[@JBDuncan]

This is a techonology advancing so rapidly that multiple youtube channels can make 30-40 minute updates on the advancements every few days. Not months or years... Days.

[@Seriouslydave]

I have noticed this too, youtube throws 7 month old vids at me and if i dont see the date it only takes me a few seconds to realize its oooooold news, if it isnt for the last 4 days then its probably outdated or already in another channels video. Its crazy how fast these things are progressing.

[@cosmic5241]

Within less than 5 years, we gone from very little advancement or news of the advancements in ai coming out to now having everything become outdated within a week or two just about. It's impossible to keep up with everything unless it's your job too tbh

[@MyAmazingUsername]

11:17 Super funny how you point out the nearly invisible warping on the building, but not the paddle that is melting like butter in the air. 😂❤

[@TzarBomb]

Yeah, that was the first thing I saw, "wait, that paddle looks weird." hahah

[@sa_3od]

Thanks for the info :) It's crazy how fast AI is evolving! We’re waiting for the Omnigen video

[@theAIsearch]

you're welcome!

[@isakisak9989]

We got the Dronehosts from Westworld before gta 6, unexpected but cool.

[@theAIsearch]

lol looks exactly like it!

[@UlyssesDrax]

A surprise, to be sure, but a welcome one.

[@Wasserganger]

22:16 - Damn, it reminds me of androids from Alien franchise.

[@Octamed]

Pretty close to having a "Home video to VR" holodeck type thing. Imagine getting an 80s VHS video and experiencing it in VR

[@theAIsearch]

interesting idea

[@ernesto.iglesias]

11:44 and the building moves too

[@Seriouslydave]

Motion builder for ai. They hated motionbuilder when it first came out, it was going to end animators but it became pretty big, i see this as mocap for concept animations etc maybe even eventually end product.

[@ZorenStudio55]

Ai never sleeps 😂

[@theAIsearch]

😃

[@orbetymo]

But us mortals still have to 😮‍💨. It's not fair!!

[@joajoa9584]

About the AI cancer, it's scanning images of pre-processed samples of human tissues. So it's not just scanning your body and telling you what it is. You need to go there, where they open your body or insert a device that will tear a part of it, extract it, process it, color it, and then input it into a microscope.

[@theAIsearch]

"where they open your body or insert a device that will tear a part of it" damn that sounds invasive

[@joajoa9584]

@@theAIsearch the only exceptions are typical lab samples, such as saliva, urine, feces, or blood. These are very useful but not always as informative as taking microscopic pictures of the cancer tissue, as is the case for this AI.

[@Aman43086]

Please bring Tora installation tutorial

[@chariots8x230]

28:09 I like that AiOS is picking up the poses on images & videos that have multiple characters in them, but I just wish we could put our consistent characters in those poses. Does anybody know if we can work with consistent characters here by being able to apply our consistent characters to each of the poses? I want to be able to create images & videos with multiple consistent characters interacting with each other. This would be a great tool for projects like AI filmmaking and comic creation.

[@theAIsearch]

i mentioned a body swapper here which might help kzbin.info/www/bejne/d4i1oGVnm8yskLs

[@manonamission2000]

I would wait until the US FDA approves that foundational model for cancer. Sure, the research was published in Nature, but there are often challenges replicating outcomes from research articles, and retractions are also a thing.

[@SimpsOfHumanProgress]

I always love how you highlight exactly what you will explore in the video ❤

[@theAIsearch]

thanks

[@HarrisonBorbarrison]

Vidpanos looks cool

[@theAIsearch]

yes!

[@apatsa_basiteni]

Amazing stuff.

[@-Burs]

Man.. This news is getting crazier each day. Hard to follow them unless you have very specific interests. And I want to try them all, yet so little time to do that. Thanks for the overview, saves a ton of time than researching all of them myself 👍

[@theAIsearch]

yeah, i constantly feel overwhelmed and dont have enough time to try them all

[@zuliusakbar]

Thanks for info

[@theAIsearch]

Welcome

[@Astronyan1]

Can you make a video about fine-tunning a LLM model with your own data?

[@pmriche]

Merci !

[@pmriche]

Thank you very much for searching and sharing so many information about AI. Very interesting and useful information.

[@theAIsearch]

Thank you so much for the super!

[@u2bAriel]

Hi. Love your content. Is mimo ai body swap available yet?

[@budekins542]

Impressive A.I breakthroughs but. .can you do some step by step tutorials on how to access them. .

[@theAIsearch]

coming soon! i'm working on an omnigen tutorial rn

[@28Corner]

There is an AI that could keep consistency of a character being generated so the look of the character will be the same on every different prompt we gives so we could create a broader story using the same character from prompt to prompt, but I couldn't exactly remember what's the AI name is

[@cdlamagiela]

best chain youtube !!! thx for all , just add subtille Mister the goatAIchain , big kiss

[@theAIsearch]

thanks

[@xenn2996]

Thanks man

[@theAIsearch]

You're welcome!

[@hiddendrifts]

it seems like generative ai by itself will only have creative application. but its technology and research can be used to power more technical uses

[@Storytelling-by-ash]

wow awesome tool and opensource

[@theAIsearch]

yes!

[@thewatersavior]

360 videos in Tora will be trippy

[@garylcamp]

The Mochi Playground misses a lot of the text. Looks to me like half the results are missing some elements mentioned in the text. The realism in my 2 experiments was not very good too.

[@theAIsearch]

thanks for sharing! i'll post my review video tonight on mochi & allegro

[@thewatersavior]

It will be interesting to see what complexion these video generators assume your character will be when given a prompt that just asks for a person

[@crawfordscott3d]

Why... This is just political bullshit. This is like saying you can't wait to see the racial diversity in a new movie. You miss the whole point in exchange for pretty social politics

[@thewatersavior]

@@crawfordscott3d Lol not quite the same. its questioning the data used to support the model. There isn't a right answer, its a dialogue that uncovers reality. Doesn't have to be offensive - just needs to be understood. If every image of a "human" looks like one characterization, perhaps it suggests the data is overfitting somewhere? Good research questions its own data and outputs, friend. The same model, when applied to health conditions, might struggle if its missing a large swath of the population. Appreciate your feedback :-)

[@pieugedher257]

If it's fine with you can you also share hardware reqvirement for any opensourse project you talk about, thanks for the video bty

[@chariots8x230]

17:02 Can Mochi do *‘Image to Video’?*

[@theAIsearch]

not yet

[@GOD_AND_FLAT_EARTH_MUSIC_VIDEO]

Hotshot still the best video generator and crazy resolution, how can you say mochi (480p btw xD) is the best when Hotshot is not in the list 😄

[@theAIsearch]

hotshot is closed source. also, see this artificialanalysis.ai/text-to-video/arena?tab=Leaderboard

[@peacetoall1858]

Omnigen looks amazing. When do we get to play with it?

[@theAIsearch]

its out now! i'll post an installation tutorial soon

[@peacetoall1858]

@@theAIsearch Awesome! Waiting for that.

[@henrythegreatamerican8136]

About that A.I. detecting over 96% cancer. I'm curious how accurate it is in finding pre-cancer stages that have a high chance of becoming malignant.

[@yoagcur]

and how many false positives

[@Irokusion]

There is always only two questions with all this new fancy AI models, can I run it locally and how much vram do they require...

[@zachary3603]

yes and MORE xD

[@MrDonCoyote]

great video as usual

[@theAIsearch]

Thanks!

[@chenc9945]

4:45 that iron man punching lol

[@XDgamer1]

mochi 1, AIOS ❤

[@BBadakhshon]

11:53 ah that's Iran Isfahan's "Nesfe Jahan"

[@fizbonetv2816]

All these AI models would be super cool to use... if my pc could fun any

[@Seriouslydave]

Talk to IT dept from any graphics or data driven business they can probably sell you everything but the storage cheap.

[@TicTac-g7m]

​@Seriouslydave Good idea for those needing a computer. When I was fixing computers, I often had 2 or 3 laying around that I put back together, making them available for cheap if someone needed one.

[@psyraproductions]

Hello, when is Sony's AI model going to be published? I mean the music motion thing...

[@TomTheEnglishPicker]

I’m not sure if it’s been removed but I can’t find the same motion inversion on hugging face , just one from another creator

[@hiddendrifts]

39:06 bro explained that so effortlessly i'm lowkey wondering if he studied medicine before

[@theAIsearch]

👀👀

[@depralexcrimson]

Imagine he used AI to explain that

[@hiddendrifts]

@@depralexcrimson would be very on topic lmao

[@rogeliocorral1472]

incredible.....

[@theAIsearch]

indeed!

[@eastmanwebb5477]

We’re going to need a leaderboard to rank who has the best leaderboard.

[@theAIsearch]

lol

[@Noredia_Yuki]

Please make an AiOS installation tutorial🙏. I have come up with too many ideas when i don't even have it installed yet🤤

[@XDgamer1]

I want a way to make AiOS with AI rig the generated mesh automatically

[@Twizzforsberg]

Im just waiting for a free or low cost place where i can make my shortfilm. So i can use same person straight thru the short. Right now i get different every tine and i need to generate atleast 10 to 15 second clips

[@theAIsearch]

mochi is your best bet. it's free & unlimited if you have a good enough gpu. i'll post a full review video tonight

[@Twizzforsberg]

@theAIsearch Perfect i will watch

[@alsamuef]

Overwhelming.

[@Bahamamos]

They said diffusers coming soon! This means civitai models! [OmniGen!] Hopefully AIOS also does this.

[@theAIsearch]

that'd be awesome

[@jonmichaelgalindo]

There's so much... 😐 I just finished testing OmniGen (had to write a whole new testing set). Gotta test mochi now that it's doable.

[@theAIsearch]

yes, it's overwhelming!

[@diego.coutinho]

most insane week so far - pretty much what all ai channels say everyday.

[@DanIel-fl1vc]

Does AiOS allow you to export a human skeleton with animations. That'd be really useful.

[@creeperboy3755]

SDXL + Flux + Omni Gen = 💥

[@theAIsearch]

killer combo

[@OkabeRintarou-us9qi]

Is omnigen tools works with anime?

[@theAIsearch]

i'll test it this week and make a video on it asap

[@lukasgruber1280]

human summary: "WOW!"

[@NugrohoAdi_S.]

List AI tools that can assist game developers?

[@Walter5850]

19:13 it's not going to Literally kill your gpu

[@Allplussomeminus]

12:14 China is definitely winning the humanoid robot race in terms of mobility.

[@BigSources]

I read 38:52 skill issue because of skin and soft tissue lol

[@PriyanshuSingh-sd2dc]

he is like some one mak 19:20 then shows legendary kijai’s 😅 git hub

[@WorldCuppppp582]

Tora install tutorial please

[@giri.goyo_yt]

Hee hee. They used iStock for the video source. I need this model for my music videos!

[@theAIsearch]

didnt even bother to remove the watermark lol

[@lpqlbdllbdlpql]

26:22 how fun 26:39 we're all fucked.

[@TheOneMaddin]

Look at the kayaking video more closely. The main dude is paddling with half a paddle and the other functioning side is weirdly deforming. How can you notice the building but not that? :D

[@firefox8713]

Hold on to your papers

[@robertanderson2424]

Fellow scholars

[@sebastianb.1926]

I would rather be interested to know the people that actually enjoy AI content. Who are they, where do they come from, are they in tune with the world outside their screens, do they have any friends? I'm not being sarcastic, but genuinely worried about AI bringing with it disconnection and mental health issues among the disenfranchised.

[@tuckerbugeater]

it's over

[@Michael_H_Nielsen]

hello terminator

[@BBadakhshon]

ha, I see 19:15

[@pdjinne65]

The last one is the only actually interesting one. The rest is but a diversion.

[@theAIsearch]

coming soon!

[@williamcastillo3929]

You're the f*cking best man, holy shit!

[@a31-hq1jk]

Hii

[@theAIsearch]

hi!

[@TLabsLLC-AI-Development]

I want to bring genmoai to the Temporal Prompt Engine and bring it as a web app. If anybody wants to work on this with me, please reach out.

[@SS801.]

that new tool is cool but it is so low fps

[@theAIsearch]

which one?

[@thewatersavior]

that rabbit trick had one tell... the flick of the wrist.. or rather the wristlet changing colors.. Maybe that could be fixed with more prompting. Awesome video

[@BBadakhshon]

39:00 get out of here bro! what

[@ArtificialMaze]

Can You clean the mess with sources... name them dont just put link after link. Robots does not look fluid still, and Mochi is just WOW, any that i used could not make a proper video but this one just wow! BTW is there any AI Video upscaler? Like for images Upscayl which i can use on my PC?

[@robertanderson2424]

Bossy boots over here

[@manonamission2000]

Is pika still even in use by anyone? my goto's continue being minimax and kling

[@theAIsearch]

minimax and kling are also my gotos

[@customswitchblade6924]

as a 3d artist, id say, were cooked

[@NoenD_io]

Cant wait for ai to feel pain

[@ClipCompass-ex5ip]

OMNIGEN IS OUT!!!! OMNIGEN IS OUT!!!! OMNIGEN IS OUT!!!! OMNIGEN IS OUT!!!! OMNIGEN IS OUT!!!! OMNIGEN IS OUT!!!! OMNIGEN IS OUT!!!! OMNIGEN IS OUT!!!! OMNIGEN IS OUT!!!! OMNIGEN IS OUT!!!!

[@theAIsearch]

tutorial coming soon!

[@ClipCompass-ex5ip]

@@theAIsearch YESSSSSSSSSSSS

[@ClipCompass-ex5ip]

THANKSSSS