Sakana Evolutionary Model Merge - and other AI News
Вставка
- Опубліковано 7 чер 2024
- These News will Change your Mind: Sakana AI Evolutionary Model Merge, Meta SceneScript LLM for Real World Understanding, Google Virtual AI Vloggers, NeuralLink by Elon Musk implants brain chip into first human.
#### Links from my Video ####
3 Patreon Workflows: / 3-new-reward-100427387
Join my LIVE STREAM: ua-cam.com/users/liveetpOAzhdXV0
Stable Video 3D: • 🔥 Stable Video 3D - Lo...
AnimateDiff Lightning: • AnimateDiff Lightning ...
enriccorona.github.io/vlogger/
sakana.ai/evolutionary-model-...
ai.meta.com/blog/scenescript-...
NeuralLink: • Neuralink Live Update ...
#### Join and Support me ####
Buy me a Coffee: www.buymeacoffee.com/oliviotu...
Join my Facebook Group: / theairevolution
Joint my Discord Group: / discord
AI Newsletter: oliviotutorials.podia.com/new...
Support me on Patreon: / sarikas
00:00 Patreon Reward
01:32 Google Vlogger AI
03:17 Sakana AI Evolutionary Model Merge
04:14 Stable Video 3D
04:52 AnimateDiff Lightning
05:28 Meta Scene Script
07:24 Neurallink Brain Chip
08:06 Future Thoughts - Навчання та стиль
#### Links from my Video ####
3 Patreon Workflows: www.patreon.com/posts/3-new-reward-100427387
Join my LIVE STREAM: ua-cam.com/users/liveetpOAzhdXV0
Stable Video 3D: ua-cam.com/video/4W-yIr_EZd4/v-deo.html
AnimateDiff Lightning: ua-cam.com/video/ktKteQ8X3B0/v-deo.html
enriccorona.github.io/vlogger/
sakana.ai/evolutionary-model-merge/
ai.meta.com/blog/scenescript-3d-scene-reconstruction-reality-labs-research/
NeuralLink: ua-cam.com/video/ZzNHxC96rDE/v-deo.html
👋
The idea of allowing Elon Musk to have any influence over your brain may seems like a dangerous proposition.
Open source neural link
Evolutionary model sounds amazing, im sick of trying out 100 models a month to find only one that works well.
As someone who absolutely hates meta and Facebook... I regret to inform you that I must now give them their props, again. That's pretty cool and impressive. I like it and I hope they release it open source.
What's the platform with those workflow graphics? Looks awesome!
What’s this software that you are using at 0:14 for the avatar generator ? With all the different nodes. Evolutionary Model Merge?
Thanks a lot for the informations!
I was super intrigued when I discovered Sakana. I can't wait to use it soon.
The meta scene script would be super useful for forensic work
Reminds me of the Batman Games :)
VLogger looks promising. I'm going to assume we will never get access to it though.
The crazy thing is: there are companies that already do this is a way where you can't even tell it's AI
I keep on running into 'stock images syndrome'.
Beautiful images, questionable composition and nearly absent story-telling.
A human-painted portrait tells a story about the subject, the artist and their relationship to life.
AI images depict a thing exactly(?) as described, but struggles to tell any kind story.
What do I know? I was raised by a Painter....
That's Strange, because you can create your own composition and elements with image 2 image rendering and a painted input. So you are in control of the art and story you create
Oh Olivio, you're so humble, you are the perfect specimen !
You rock! I swear I’ll get back to you on the layered with alpha video pipeline :)
When you talk about how you can continue a wall in the same way that you can continue a sentence it really got me thinking. Thanks for the great breakdown. I find it fascinating that we're finding ways to tokenize so many things besides word fragments.
Yes, that really is a awesome idea. AI really opens up news was to think about computing data
Thanx 4 sharing this 🙏
You have great taste in shirts.
Explore the captivating realm of AI's rapid evolution and its seamless integration into the fabric of our daily lives. It's truly enthralling to observe the dynamic pace at which technology propels forward, evoking a profound sense of wonderment and perhaps a touch of trepidation. With each stride forward, we eagerly anticipate the unfolding of new horizons and the myriad possibilities they hold. Let us embark on this exhilarating journey together and embrace the future with open arms, ready to embrace the endless opportunities that lie ahead!
Can I use Gligen UI on Comfy UI for this?
its still a steep learning curve to use all the functions in ComfyUI....
Thanks!
first neurolink, welcome "ghost in the shell"
Get the new iBrain XD
There are studies in several universities in the same direction. There is a Brazilian professor and researcher called Miguel Nicolelis who has been studying this (in a less invasive way) for 25 years. The difference is that Musk has plenty of money to make this a product. It is important to emphasize that the advances are not the merit of Musk, but of the researchers he pays. Musk is not Tony Stark.
we need to how to make model for stable diffusion a current video, plz sir!
Neuralink sounds so dangerouns, no? Not even thinking of hacking or whatever, but a magnet EMP or any eletromagnetic device can give a shock to your brain right? idk, hopefully nothing terrible happens
In the future, however, all YT videos will be labeled if tools such as Vlogger AI are used. You have to specify this when uploading and UA-cam then displays a banner in the comment section of your video.
One day, I will 3-D print my waifu, and I will never be alone again!
9:04 that’s exactly what I’m thinking. In the future there will be more and more people who can’t draw but just have the vision and know how to create and manipulate images with AI extremely well. And they’ll have far superior output to someone who has to learn drawing for decades to create same results. And consequently since it doesn’t make sense to learn to draw I believe how it‘ll go is that the AI does all the drawing and the human is solely the art director and visionary.
Edit: but first we need to fix those hands…
I ran across a folder of images I collected from the internet a few years ago because I thought they were awesome. I found that folder the other day and wasn’t impressed. The AI images I can create on my machine make everything else obsolete. The decades of training and talent it must have taken these artists to stand out in their fields has been stolen and fed back to us for our own output.
Yes, same. Things changed so fast. And AI is most of the time really on a master level of detail and technique
THE VLOGGER project also has me more worried about a "deepfake democracy" wherein a false consensus is drawn via lifelike AI of regular people - not just important people
So ?
Much of today's AI art is superficial eye candy. It is sophisticated in composition and content but largely misses out on any coherent narrative. The person behind the AI prompt still has to supply this. Already, some artists are establishing themselves as hybrid AI artists, mixing their traditional skills with the speed and facilities that AI offers. This strikes me as a positive direction.
However, none of this detracts from the fact people will continue to enjoy purely human made art, if anything this is likely to kick-start a new interest in traditional human skills. To compare AI art to human art is fundamentally flawed, unless all you are after is an unconscious dopamine hit. Person made art is all about the artist with the end result being it's expression in some media. AI is all about the end result.
Maybe in simple terms, traditional art is about the journey, AI images are about the destination. To some, this distinction may not matter, but to others it is essential. For me, I hope people stop comparing the two. AI art and human creativity are as different as, say, watercolour painting and cheese. Both are enjoyable and both have their place in our future, but they are not the same.
Just wanted to give huge thank you for your informative and great videos and thought the latest one would be the best place to do it.
Messing with the brain physically is very dangerous. I do not endorse this particular interface especially in association with the deranged mind behind the development of this technology.
Having said that, it is certainly possible to interact with our brain using non-intrusive technologies. Keep this in mind, or rather OUT of your mind 🤪 - this is my honest advice to anyone with a bit of a brain.
- Thank you for your videos and activities - you bring up interesting thoughts having to do with art creation rather than just the usage of new tech for usage sake. Nice!
I guess you would see that a bit different if you would be in a wheelchair, not being able to even your hand. This chip isn't doing things to his brain, it's reading what the brain is doing.
It's important to recognize that brain-computer interfaces have a long history of research both using things like fMRI and more invasive cortical implants (or even stents). What neuralink has done so far is not really new. It's just being branded as such. There's so many ethical considerations that need to be addressed with these implants - like what happens if neuralink goes bust? Or they pivot? Do their patients just get an email one day saying "sorry, we know this gave you agency, but we're recalling the device" unfortunately, that's not an unheard of outcome in this space.
Is the background behind you real?
nope, that's Midjourney. I wish my home was that nice ;)
You can see green fringing around is head/hair.
That's a classical sign of green screen usage, on UA-cam.
Vloggers are looking more and more exposed.
Max Headroom
Bro, give yourself some credit. You are a much better specimen than these plastic AI vlogger examples.
Thank you :)
you can program an ai to make videos of you talking about new ai technology, turning you into the curator instead of the creator
Totally. I was actually thinking about. Especially with the ability to create Videos in different Languages
we all know that you are AI generated. Especially the outfit must be the indicator and evidence for this! No excuses! 😁❤️
I'm happy AI is progressing so fast, soon we can stop capitalist slavery and enjoy our lives of leisure as nature intended. universal income will be happening eventually. people who are afraid to be laid off need to see the bigger picture, and people who don't want to compete by hand, well work on your ego to say the least. there are plenty of ways to express passion through art without a competition mindset, in fact it's the only way to create truly meaningful art. And that's why people who can't for XYZ reason create the art without AI should have the AI to rely upon to unleash their creativity from our mortal shackles of limited potential.
Why are you not answering my question in the comments in yesturdays video? The question was why you did not enable frame interpolation in A1111 as shown in the bottom left of the plugin tab. I clearly pointed out the time from your video.
Maybe he didn't see your question amongst all the other comments. I hope you realise he's under no obligation to answer yours or anyone else's questions. 🙂
@@Elwaves2925 I thought it would be fair if he enabled interpolation on A1111 just like he did in comfy where he claimed only comfy offers more smoother frames. Which reduces his credibility for making these videos because clearly, it seemed biased because if he enabled it, it may have been smooth on A1111 too.
@@DivinityIsPurity Oh I get it, I have nothing against your question itself. I just wanted to point out that you may not get an answer. 🙂
If I'm being honest, as much as I like Olivio, he can be biased. In his recent video about SVD he claimed Comfy was getting it first, yet I've had that in Forge since it launched earlier this year.
The developer that comes up with a reliable way to create quality STLs is going to make a mint.
Is anybody training a model on all the new 3d designs that have come out in the last couple of years. I see that Graphtreon has gone from a dozen 3D designers to hundreds in the same period, and that they are now starting to cannibalise each other.
Totally. That said, if you have a good rotation video, you can create a 3D mesh from that. We already have that tech. I'm kind of surprised this isn't used as part of these Models
@@OlivioSarikas - Maybe a 3d model generator will need to be multi modal? Perhaps it will also need a lot of training data from existing stls. That should not be a problem, as there a now hundreds of designers releasing hundreds of designs every month.
EMO is still better than VLOGGER
"I'm not a perfect specimen." LIES!!!
Never forget to mention Elon Musk whenever you talk about a project developed by one of his companies (just like you usually do with every other company where their owner has nothing to do with the inventions), it could break his mind if you don't👍
Because he founded the company and is it's most prominent figure. Just like we talk about Gates, Jobs or Bezos and other peolpe who had the guts and will to risk their money and start something others said would never be possible. How much did you risk to change the world?
I wonder, @OlivioSarikas, how many of the youtube-commenters these days are bots and not real people. So many 'weird' user-names in every comment-section lately.
a lot of snot, how do you know what's what? That's the least you can do.
I think in the end we will probably regret subcontracting our creativity out to AI but it seems inevitable that interest in mastering skills such as art or writing or making music ect must decline because there will simply be no point in taking the time to learn these things if they can be easily replicated by a machine. Human creations will no doubt still be made purely as a pastime or hobby, but the vast majority of our culture will be produced by machines- with some input from people too of course.
But as I scroll through the endless stream of AI generated images on sites like deviant art I experience a cognitive dissonance- my eyes are sated, fed to bursting on the vivid stunning images on display- but I'm not feeling it, there is void here, a nagging futility. It's like somone gave me a gift, a beautiful puppy, so warm and cute and adorable and then, just as it stole my heart they revealed the truth- it's not a real puppy, it's a machine- a perfect mechanical replica of the living breathing original, so lifelike that- had I not been told- I might never have realised that it was not real.
But I have been told- I know now that a machine made that puppy. And I no longer want the gift.
Check your T-Levels. Especially after 30, T-levels tank.
Until Neuralink is independently verified, I will withhold any celebrations. It isn't revolutionary - there are quite a few companies that have up 20 years head start on Musk - including many more successful, verified implants. Given how dodgy his robot demo was (obviously a waldo), and how little care he has for proper protocols, I won't believe anything he presents until someone external assesses it.
don't agree. in fact I respect hand drawn work more than digital, digital art is no longer respected
Bla bla bla. Nothing interesting
To you 😂
You should stop personalizing the algorithms. You wouldn’t describe yourself while applying a filter in photoshop as “collaborating with the AI” or “working with an AI companion”. Yes, SD and other state of the art machine learning algorithms are very impressive and useful, but that doesn’t make them a person that you can collaborate with. This whole video is a bunch of clickbait, as though we have some sort of generalized “intelligence” algorithm that can come up with new models all on its own. Things like using LLMs for interfacing with the CLIP are very, again, very useful, but at the end of the day it’s still just a dumb machine translating instructions from one language to another, no different from any other machine assisted tool like Excel.
Wow, I learned something new today