Googles New 'Text To IMAGE Model' Just CHANGED Everything (Now RELEASED!)

Links from todays Video

Welcome to our channel where we bring you the latest breakthroughs in AI. From deep learning to robotics, we cover it all. Our videos offer valuable insights and perspectives that will expand your knowledge and understanding of this rapidly evolving field. Be sure to subscribe and stay updated on our latest videos.

Was there anything we missed?

#LLM #Largelanguagemodel #chatgpt
#AI
#ArtificialIntelligence
#MachineLearning
#DeepLearning
#NeuralNetworks
#Robotics
#DataScience

[madshader]

I put in this prompt "a girl reading a book, sitting on a chair in the library" and it said I violated their policies. Completely useless.

[Wanderer]

Honestly I don’t think anyone is gonna bother with it if it’s censored. We already have Midjourney and Dalle, having another ai image generator is really nothing special, but an uncensored one is definitely intriguing

[SpinX]

You also can’t use it Canada either without a vpn as bard hasn’t been released in Canada at all.

[mattrodriguez]

Can you stop saying "@!%×^×<" just SHOCKED THE ENTIRE INDUSTRY" Its very annoying and clickbaity. That being said i love your vids but please stop hyping your videos up? because when AGI is really realeased you are gonna try to get hype but everyone will think you are just posting another video because thats how all of them are labeled

[jaymethodus]

It’s VERY intriguing how futuristic cars look exactly the same regardless of which model they prompt for it

[MyZombieLick]

"Shocked the industry"? Not really... at least not to anyone who is aware of the other options. Imagen-2 is decent but it's not even close to as good as Midjourney or Stable Diffusion at this point. Maybe Google will overtake them at some point but as of now, they aren't even in the same universe. Midjourney despite not having a super slick interface is still pretty easy to use and they are creating a web version with all the bells and whistles. Stable Diffusion with the latest Web UI is extremely complex but is also the best by far, especially with specifically trained LORAs and other config options. Not even Midjourney can compare to Stable Diffusion when done correctly, it's just the learning curve of it that prevents it from having wider adoption. It's light years beyond Imagen 2 or DALL E 3.

[HappyCamper]

The images I tried making in Bard are terrible. What am I missing?

[TheMrCougarful]

It's difficult to accept that any of this is a real product. I've been doing technology work for 40 years, I've seen a lot of change, but this just blows my circuit.

[pdreding]

And it's still not available in Canada. I think Google's still upset about CanCon.

[Davd-ikhm]

it CHANGED EVERYTHING... then later in the video... its pretty good, pretty cool. very conssistent content

[piteshbhanushali]

Its matching dall e 3 ..but image quality not good as midjourney and stable diffusion.

[skoumastv]

I like this channel but please reduce the clickbaits. Make it more like a newa channel with neutral titles.

[singularityscan]

9:43 it is the same woman! you dint mention that in the video, it's important because A.I. had a problem with the same character!

[stefanjohansson]

14:35 Watermarking is used by every service but they doesn’t always tell us about it. Pointless to use watermark if you inform about them! All those techniques are easy to remove.

[MyrLin]

Hear Hear, Excellent Idea and Viewpoint. Add the Synth Num to all AI suppliers. ... eventually 'they' will of course get around it, but it will help at least a bit.

[Apollyon]

Imagine a generating pictures using AI, imagine gereating a 3D world from that AI generated picture, imagine exploring that 3D world from the AI gereated picture. This and much, much more is what we have to look forward to with AI. The posibilities are endless and all that is left is your own imagination.

[pierruno]

You can use VPN to Sign up and set everythign to US

[toastrecon]

I would love to know more about how the Synth ID works. It seems unlikely that the "watermark" is impossible to remove. If it can be detected, it stands to reason that it can be modified to retain the original appearance until the watermark has been removed. It would also mean that just because an image had the watermark, it wouldn't mean that it was AI-generated, it could be added after the fact. I also wonder if the watermark is added to the entire image, or if it applies to sub regions somehow. Like if you take a picture of a chair with a plant, and then you add in a shelf with objects, is there a way to know that just a certain region was ai-enhanced?

[hearmenow]

"This tool isn’t available in your country yet" No matter whatever country you put in.

[Xzor]

In my head it's pronounced "Imagine", not "Im-uh-gen".