Totally! Second voice sounded robotic, lacked a subtle inflection of emotion that was in first clip. This is the worst it will ever be, i can imagine a time when the AI voice will be detectable because it sounds TOO real!
5:00 - A use for this could be in prompt engineering. I know different AI generators will work a bit differently based on training but if you want to work out a general guide and you are not sure how you should describe the image you are creating, then this is a solid way of editing and updating your prompt on the fly while seeing the results in real time to test the functionality of your prompt. Solid.
thanks for a fantastic video, it's been a fun journey with pika, I'm currently combining AI with Blender and other tools to create a mini series done in a way that I don't think anyone has seen. Been using them since the early days, made quite a few things with them and have made lots of friends in the community. Are you on the Discord?
Thnx for this info, so if i want to create a 15 min "movie" where the generator doesnt change/evolve characters and background rapidly theres a a bunch of pre-making right? Im an AI noob, so id like your advice to cut it down into a "to do" list, which and how many programs do i actually need? (This takes a rewatch) 😅
There is still quite a bit of editing that needs to happen. You'd also need to use some kind of faceswapping program to make sure the characters are consistent. In addition to that, you'd need to make sure your prompts are probably consistent with the same descriptions so that way you have the same cinematic style throughout the film :)
Zero ten ai wardrobe would be a good application in clothing stores if u implement the code and ai to work real time with mirrors as screens and 3D depth cameras u can try on clothes right at the mirror
They are still to short. When we get clips around a minute or so and u have good control over the outcome, then it becomes a real game changer. But we ofc dont get there over a night.
I like to propagate these smaller changes into the future. like by 2025 we will have full high quality movies, we will have a Google assistant that's on "Her" quality, we will be able to instantly translate any language to any other in any voice. just looking at what's gonna happen, you can estimate the bigger picture. and these are just products, what's happening in labs is months to years ahead
I am a hobby 3d artist, and it looks like I might ditch computer all together in the future and go paint and sculpt something for real, the future is crazy
I think it's best to do it in chunks rather than pasting the entire book (if that's what you're doing...). That way the AI tool can focus better on the nuances of the text
hi really impressive work and research... can you please help me out i want to make animation movies using ai for my own stories can you suggest me any site or method by which i can do it... waiting
Imagine being able to literally create a blockbuster short/film from the comfort of your bedroom and having said creation be seen by millions, if not billions of people! The future is looking to be extremely exciting!!
no, because if 100s of thousands of ppl are uploading movies no one will be seeing them because of home made movie overload... We separate rocks from diamonds because EVERYONE has access to big old rocks, big diamonds are extremely rare
I think it will end up being the same as TH-cam. There is far more content than what anyone could ever watch, but using algos the better creators end up rising to the top. I also don't think the majority of films in the future will be feature-length. They will be mainly short films because audiences will prefer shorter form content. We already see that as true across most mediums and it will only continue.
@@armondtanz Well first off, most people don't even know how to create a compelling story. Just look at how bad most books/screenplays are. So only the people with actual skills and talent are going thrive with this technology
As someone who has spent 15 years learning VFX compositing I am not looking forward to when tik tok becomes ai driven content because people without the skills are able to make crap.
I've been very disappointed in Pika 1.0, it feels like they really haven't developed much at all, a lot of half-baked features that have no use beyond gimmicky videos on twitter by influencers. If that is the market they are going for then good luck with that, Midjourney is going to blow them out of the water once they launch video.
I love seeing this every week, thank you. I love AI although I am concerned about the ethical and security implications of those voice cloning softwares
2:02 "this technology will be the underlying technology that is used for set extensions in the future". You don't know that and I think you are making some huge leaps from mediocre results. Just look at how deepfakes have been integrating into VFX pipelines. Artists still have to come in and clean that stuff up a lot to make it use-able on the big screen. I just have no faith that raw outputs from text prompts will be used on professional productions without being heavily modified or projected on geometry. You are demonstrating zero understanding of how this stuff is actually used in the industry.
You have a fair challenge, but I think we'll have to agree to disagree on this one :). It's true that the results from this demo are *extremely* early and have lots of room for improvement, but we still stand that the underlying technology will be the norm in the future. We'll see!
As a vfx CG environment artist we go though lot of pixel fking from client and it never get approved coz of whole other reason and also clients won’t agree to the use of Ai in mostly all cases coz it can make them have future issues with copyright coz it’s not what is created by humans. And as per latest copyright laws in us anything which is not created by humans won’t be able to get copyright so your title is very misleading or I can say a nice click bait one. I have been doing this for 15 years and me and my fellow artist can say that this might look great but to get it pass for a movie is a pipe dream Atleast for now. Thanks
We're just at the beginning. The equivalent of pong and the CGI hand and Ed Catmull's CG hand in 1972. Copyright cleared models are here including the Getty news I mentioned in the video. But within the next three years (max) AI VFX will be silver screen ready. These tech demos are just showcasing what will be here very soon.
how are you going to talk about gemini without talking about the fact that google admitted to editing the video and that it's results were not reflected accurately? Cmon man
this is all so insanely amazing i am actually scared wow.. just in awe idk. new tech is gonna be build on top of this and even more crazy can you imagine? I grew up with the first game boy, my children will have vr. i cannot fathom it xD
I can already see an evil very effective way to use one of these tools to get money from inocent people. Hopefully people don't fall for it a lot. So sad how these tools can be used for evil in a big scale.
I’m a big fan of your work but I’d love to see you more often address both sides of these advancements, the promise of the gift and the risks of the tech. Empowering us with info on how to handle both sides of the coin would be super appreciated! Get us thinking about how can approach this ethically and responsibly as well as creatively and ambitiously!
Noted! We appreciate the comment and suggestion - there is certainly a tricky line that people need to be aware of so that people know how to ethically navigate AI tools in the future. Thanks again for the suggestion!
@@curiousrefuge thank you for the kind and considered comment! I think if you love something and wanna see it succeed there is absolutely nothing wrong in pointing out where it could be misused or improved. Helps to safeguard it and make it even better! 💕
@@Also_sprach_Zarathustra. it’s ok to be 360 in your views. Last thing CR will become is a doomer, this is one of the best places in the world to come for the celebration of AI generative art advancements! If you love a thing, it’s ok to point out blind spots or areas it can grow or improve 🫶🏼
I'm not so impressed anymore. Okay I'm still impressed with what it can do, but to make anything worthwhile we need to be able to do multiple characters on screen and able to do action. These things are still far behind and nobody mentions that
There is certainly a long way to go before anything hits the silverscreen. And you're right there is a high expectation that people require. Luckily these tools are getting better at a exponential rate! :)
@@curiousrefuge they will get there and we need to take care of the great creative overload and the depression inside the art communities. Sometimes it looks a bit like the triumph of the nerds who are fascinated by this technology without looking at the implications for our society. Don’t forget that there are highly intelligent people designing gears of war every day without feeling guilty.
I don't think aí generated Videos have much of a future. Too expensive to run, poor results, lack of consistency, you can't control exactly what you want.
We disagree. The tools get exponentially better every week, competition is driving the features to be incredible, and it's far less expensive than live studios. There is certainly a lot of room for improvement but we are getting there relatively quickly.
@@curiousrefuge No. People after power will take things to the next level. It willl move further and further away from humanity until it's too late. You are being hypnotised by the promise of something you will never attain.
@curiousrefuge technology is vaguely entertaining as the ai companies boost the hype. At some point, people will realize that it's not going to meet expectations. When it does, I wouldn't want to be holding any ai stock that's for sure.
Hey, it's me! Also looking forward to seeing the Holiday Film Competition! I'm on the edge of my seat to find out who won!! (hmmmmmm....)
Hey Tim! The submissions this year were INCREDIBLE. The fans won't be disappointed :)
1ST one is real - 2ND one generated - but MAAAN this is getting soo close! exciting!
Totally! Second voice sounded robotic, lacked a subtle inflection of emotion that was in first clip. This is the worst it will ever be, i can imagine a time when the AI voice will be detectable because it sounds TOO real!
Whoa that's an interesting way to think about it.
We'll have to dial in the 'fakeness' slider a bit. haha
Thanks so much for this content! You are my favorite channel these days!
Great video! Appreciate the shout out on “Humans”. Glad you enjoyed it.
Any time! Glad you enjoyed it :)
I REALLY appreciate how you take the time to show and explain every part of the process. SO helpful. Bravo!
Glad it was helpful!
12:17 So where did they find the reference footage? Did an AI also generate the reference footage?
Good question! It seems like some of it could be in these demonstrations
5:00 - A use for this could be in prompt engineering. I know different AI generators will work a bit differently based on training but if you want to work out a general guide and you are not sure how you should describe the image you are creating, then this is a solid way of editing and updating your prompt on the fly while seeing the results in real time to test the functionality of your prompt. Solid.
Good point!
Amazing advances indeed. I will say I wish they would put time into making the clips longer than a few secs.
Agreed! I suspect in the next few months we'll see more and more tools have native features with longer outputs ;)
9:08 The media might be calling a chatGPT killer but the audience is calling it a fake.
True!
thanks for a fantastic video, it's been a fun journey with pika, I'm currently combining AI with Blender and other tools to create a mini series done in a way that I don't think anyone has seen. Been using them since the early days, made quite a few things with them and have made lots of friends in the community. Are you on the Discord?
Combining workflows with Blender opens up a ton of possibilities! Yes, all the CR team is on our discord :)
@@curiousrefuge I meant to ask if your were on the Pika discord sorry I wasn't clear, that being said I'll jump onto your discord too!
How the hell is this even possible!? Trippy!
I can finally start in my own action movie lol!
That's right!
Amazing insights. Thank you!
We appreciate you watching. Stay tuned for more in our more recent episodes!
Thnx for this info, so if i want to create a 15 min "movie" where the generator doesnt change/evolve characters and background rapidly theres a a bunch of pre-making right? Im an AI noob, so id like your advice to cut it down into a "to do" list, which and how many programs do i actually need? (This takes a rewatch) 😅
There is still quite a bit of editing that needs to happen. You'd also need to use some kind of faceswapping program to make sure the characters are consistent. In addition to that, you'd need to make sure your prompts are probably consistent with the same descriptions so that way you have the same cinematic style throughout the film :)
@@curiousrefuge much appreciated 🙏🌞 thnx Champ, i see its quite some efforts, but if it is possible its worth a try to learn more 👊💥🤙
Zero ten ai wardrobe would be a good application in clothing stores if u implement the code and ai to work real time with mirrors as screens and 3D depth cameras u can try on clothes right at the mirror
Sounds interesting!
Pika really needs a lot of work to be viable for real use
It certainly still has a ways to go but it's come a long way and we are excited for all the updates :)
This shit needs to be stopped.
16:34 this technique will evolve into a workhorse, and eventually come to rule them all
You might be right :) We'll see!
They are still to short. When we get clips around a minute or so and u have good control over the outcome, then it becomes a real game changer. But we ofc dont get there over a night.
Very true, it makes it difficult to have such short clips but luckily these tools keep improving at such a rapid pace :)
thanks for shout!!! :)
Of course!!
Cool that your using artlist footage
Which section?
Caleb, great update ! Each week they get better and better. 😊
Ayee! Thank you so much friend. Cheers!
The text to voice is interesting. It is trained with her voice i'd imagine, and it does lose her accent. But definitely cool.
It's super crazy how fast all of this is evolving!
I like to propagate these smaller changes into the future. like by 2025 we will have full high quality movies, we will have a Google assistant that's on "Her" quality, we will be able to instantly translate any language to any other in any voice.
just looking at what's gonna happen, you can estimate the bigger picture. and these are just products, what's happening in labs is months to years ahead
It's crazy how fast we are moving in this ecosystem. You're absolutely right that we'll start to see all those big changes in the near future!
So if your a Game developer, will this simplify the creation of Cutscenes??
It's certainly a good use case for this technology as it's still early!
Thank you very much!
You're welcome!
ZeroTen integrated into clothing shops - Gap, H&M, etc. - will be so cool
Yes that is certainly a probable use case!
Hey, aren’t you late like couple of weeks with the “news”?
We try our best! :). It's tough staying on top of AI news in this speedy industry!
@@curiousrefuge in deed you’re right. So damn fast
I am a hobby 3d artist, and it looks like I might ditch computer all together in the future and go paint and sculpt something for real, the future is crazy
We've seen 3d artists fuse their craft with AI tools to create ultra amazing work!
Hey Caleb, Fantastic Video ! We need explore video editing tool to help with youtube channels, thanks.
Great suggestion!
How do we get this demo
It should be rolling out to most people now! Make sure you are on the waitlist :)
Amazing new thank you❤
Neither GPT or Bard can edit a draft book that needs editing and ideas for enhancement?
I think it's best to do it in chunks rather than pasting the entire book (if that's what you're doing...). That way the AI tool can focus better on the nuances of the text
What are you saying "cree ai?" "cri ai"? I can't figure it out.
Krea.ai :)
A is fake; B is real.
hi really impressive work and research... can you please help me out i want to make animation movies using ai for my own stories can you suggest me any site or method by which i can do it... waiting
Check out our website for our recommend tools :) (www.curiousrefuge.com)
i'd imagine a Prompt VFX and Movie director job would be like the closest to being god you can get
Interesting take! People will be able to create some truly incredible cinematic experiences in the future.
Looks like the generations are only 3 or 4 seconds long. Still very limiting imho.
Very true - so has some time before it can be used more seriously but we're so excited for it!
Imagine being able to literally create a blockbuster short/film from the comfort of your bedroom and having said creation be seen by millions, if not billions of people! The future is looking to be extremely exciting!!
no, because if 100s of thousands of ppl are uploading movies no one will be seeing them because of home made movie overload... We separate rocks from diamonds because EVERYONE has access to big old rocks, big diamonds are extremely rare
@@armondtanzok doomer
I think it will end up being the same as TH-cam. There is far more content than what anyone could ever watch, but using algos the better creators end up rising to the top. I also don't think the majority of films in the future will be feature-length. They will be mainly short films because audiences will prefer shorter form content. We already see that as true across most mediums and it will only continue.
@@armondtanz Well first off, most people don't even know how to create a compelling story. Just look at how bad most books/screenplays are. So only the people with actual skills and talent are going thrive with this technology
As someone who has spent 15 years learning VFX compositing I am not looking forward to when tik tok becomes ai driven content because people without the skills are able to make crap.
I've been very disappointed in Pika 1.0, it feels like they really haven't developed much at all, a lot of half-baked features that have no use beyond gimmicky videos on twitter by influencers. If that is the market they are going for then good luck with that, Midjourney is going to blow them out of the water once they launch video.
We'll be very excited to what MIdjourney can do once it releases video!
I don't have much faith in midjourney, it still can't follow prompts, and it still can't do hands reliably.
Stop whining it's free okay 👍 now I do admit that blurring is some shit that's out of sick that gets on my fucking nevre
I am unable to accès Pika from discord even the old one can anyone help? 😢❤
Have you had any luck since you commented this? You may want to reach out to Pika too with more details
I love seeing this every week, thank you. I love AI although I am concerned about the ethical and security implications of those voice cloning softwares
Second audio clip is the fake.
8:49 The first one is real.. I'm talking about the voice and i stopped the video here just to comment. I will put a PS with the true answer ;-)
We'll reveal the answer shortly :)
your video card titles are in the wrong order
Thanks for the heads up! We went ahead and fixed them! Sorry about that!
@@curiousrefuge all good 👍
Animate anyone seems superior.
Both tools are crazy. Animateanyone does seem pretty powerful though doesn't it
I’m guessing you misspoke because things are progressing so fast but “Rock, Paper, Scissors” was only a few MONTHS ago, not years.
Woops!
I think that A is the ai version
whats the difference Pika web based and in discord version?
Pika web version has more features. :)
2nd one was fake, I think? But that was a damn good test.
It's getting veeeeerrrry hard to tell :)
2:02 "this technology will be the underlying technology that is used for set extensions in the future". You don't know that and I think you are making some huge leaps from mediocre results. Just look at how deepfakes have been integrating into VFX pipelines. Artists still have to come in and clean that stuff up a lot to make it use-able on the big screen. I just have no faith that raw outputs from text prompts will be used on professional productions without being heavily modified or projected on geometry.
You are demonstrating zero understanding of how this stuff is actually used in the industry.
You have a fair challenge, but I think we'll have to agree to disagree on this one :). It's true that the results from this demo are *extremely* early and have lots of room for improvement, but we still stand that the underlying technology will be the norm in the future. We'll see!
Cool what it can do so far, but still nothing usable, even for amateur creative projects
Disagree! We have student being hired to use these skills at agencies and contract work.
❤❤❤❤❤
Glad you liked it!
I wanna comment for the algorithm. But at the same time I wanna keep theses infos for myself 😅
Understandable! The algorithms ruin all the fun these days :)
As a vfx CG environment artist we go though lot of pixel fking from client and it never get approved coz of whole other reason and also clients won’t agree to the use of Ai in mostly all cases coz it can make them have future issues with copyright coz it’s not what is created by humans. And as per latest copyright laws in us anything which is not created by humans won’t be able to get copyright so your title is very misleading or I can say a nice click bait one. I have been doing this for 15 years and me and my fellow artist can say that this might look great but to get it pass for a movie is a pipe dream Atleast for now. Thanks
We're just at the beginning. The equivalent of pong and the CGI hand and Ed Catmull's CG hand in 1972. Copyright cleared models are here including the Getty news I mentioned in the video. But within the next three years (max) AI VFX will be silver screen ready. These tech demos are just showcasing what will be here very soon.
only time will tell, i would suggest you to title your videos appropriately. @@curiousrefuge
how are you going to talk about gemini without talking about the fact that google admitted to editing the video and that it's results were not reflected accurately? Cmon man
Sorry - some of this was filmed and edited prior to the news. We appreciate the comment!
@@curiousrefuge it’s ok, still love you 💁🏻♂️
this is all so insanely amazing i am actually scared wow.. just in awe idk. new tech is gonna be build on top of this and even more crazy can you imagine? I grew up with the first game boy, my children will have vr. i cannot fathom it xD
It's truly amazing how far everything has come and how fast it's moving!
there isnt much time left for actually making moner... sell those shovels
...huh?
@@curiousrefuge what are you confused about... Sell those shovels
I can already see an evil very effective way to use one of these tools to get money from inocent people. Hopefully people don't fall for it a lot. So sad how these tools can be used for evil in a big scale.
Very true - people need to use AI tools ethically!
2nd one is the ai generated one.
PIKA should keep it free without points.... they should keep long time until AI master its own art.
That would be an interesting strategy!
I’m a big fan of your work but I’d love to see you more often address both sides of these advancements, the promise of the gift and the risks of the tech. Empowering us with info on how to handle both sides of the coin would be super appreciated! Get us thinking about how can approach this ethically and responsibly as well as creatively and ambitiously!
Noted! We appreciate the comment and suggestion - there is certainly a tricky line that people need to be aware of so that people know how to ethically navigate AI tools in the future. Thanks again for the suggestion!
@@curiousrefugenooo, don't become a doomer please 🥺
@@curiousrefuge thank you for the kind and considered comment! I think if you love something and wanna see it succeed there is absolutely nothing wrong in pointing out where it could be misused or improved. Helps to safeguard it and make it even better! 💕
@@Also_sprach_Zarathustra. it’s ok to be 360 in your views. Last thing CR will become is a doomer, this is one of the best places in the world to come for the celebration of AI generative art advancements! If you love a thing, it’s ok to point out blind spots or areas it can grow or improve 🫶🏼
Original content that makes sense will always be preferred. Ai will just be a garbage maker.
AI content can be original
I think you meant to say TLOU not GTA.
Good point!
Gemini video got entirely debunked over the weekend tho?
AI news happens so fast - we did the filming/editing prior the new news. We appreciate the heads up and comment!
@@curiousrefuge you’re not wrong about the speed. Thank you!
it wasn't gta but last of us?
True! Those screenshots were the last of us - good catch!
potter not podder
:)
A Real , B Fake?
1:00
Thanks for watching!
I think B is "fake"...but it is just a guess.
We'll see! ;)
I'm not so impressed anymore. Okay I'm still impressed with what it can do, but to make anything worthwhile we need to be able to do multiple characters on screen and able to do action. These things are still far behind and nobody mentions that
There is certainly a long way to go before anything hits the silverscreen. And you're right there is a high expectation that people require. Luckily these tools are getting better at a exponential rate! :)
3 Seconds, poor quality (looks at the underwater details). It is far away but shows where the industry is heading in the near future.
Agreed, still some ways to go but it'll get there soon!
@@curiousrefuge they will get there and we need to take care of the great creative overload and the depression inside the art communities. Sometimes it looks a bit like the triumph of the nerds who are fascinated by this technology without looking at the implications for our society.
Don’t forget that there are highly intelligent people designing gears of war every day without feeling guilty.
B is AI
I don't think aí generated Videos have much of a future. Too expensive to run, poor results, lack of consistency, you can't control exactly what you want.
For now yes but you'll see
We disagree. The tools get exponentially better every week, competition is driving the features to be incredible, and it's far less expensive than live studios. There is certainly a lot of room for improvement but we are getting there relatively quickly.
the second one is the fake one
its just making everything so easy. is that good for our brains though?
The content that wins in the end will be those that include human elements like storytelling and empathy :)
Blocking really will be a dead art.
The next 1-2 years will be very interesting!
2 sounds like the fake one
B is fake.
Thanks, but from a creative trying to make a living, fuck AI
Creatives that use AI will take everything to the next level!
@@curiousrefuge No. People after power will take things to the next level. It willl move further and further away from humanity until it's too late. You are being hypnotised by the promise of something you will never attain.
B
Please, no ww3, no next pandemic yet, not unless I make my own AI movies come to life huhuhu
We want to see your work!
@@curiousrefuge Np mate.
A
lets be honest and not hype this, all the examples were garbage
Well we appreciate you tuning in either way!
Bard still sucks compared to chatgpt 3.5
Thanks for watching and leaving your thoughts!
First!!!
You get the first place emoji:🥇
Yawn. Not trustworthy
Either way, we appreciate you watching, Neil!
A is fake
umm. yeah. looks like crap.
Still has a little ways to go but the technology is interesting!
@curiousrefuge technology is vaguely entertaining as the ai companies boost the hype. At some point, people will realize that it's not going to meet expectations. When it does, I wouldn't want to be holding any ai stock that's for sure.
B is AI
A