Hi, Great video thanks! i was wondering how did you manage to lock the camera in a campfire shot? is there any specific keyword for this? all my attempts to lock the camera in one place are failed. the camera seems always moving which is not what i want. would be happy if you could help with this. Thanks
9:26 From what I’ve heard, SD3 isn’t just $20 a month for the time that you’re subscribed to it. If you ever end your subscription, neither you or your clients can continue using any of the images that you generated with it. In other words, you’ll have to be subscribed forever if you want to keep any of the work you created with it at any point in the past. So, this is definitely not a good choice for client work, or for building a business on.
Great video! Looking forward to a Runway Gen-3 with image prompts. For your Films of the Week, do films that are submitted have to be generated completely by AI or can they have live action in them as well?
8:28 playground has this system , for training . You need choose around 100 images . . P.S. Make sure audio ai equals video ai, sometimes video is ahead of audio ai .
We still need to convince a rich person to fund our actual film projects. With the hundreds of iterations required for every image which then get iterated again when attempting to create a useable video clip, and the costs of every tool being $20-40-100-499/month, and there being dozens of tools, the costs for a solo creator are still out of reach ($1000s per month just in subscriptions). Using just Luma as an example, it will cost about $500 in credits just to get a single useable 5 minute clip from which you can trim one shot. And by usable, I'm referring to Shy Kids claim that they had to run 100s of iterations for each scene in their short balloonhead video from a few months ago... and the consistency in that was not great. And they had a ton of post production to do as well. And that was in Sora which purportedly takes hours to generate a single short clip. So hundreds of hours and 1000s of dollars to create each scene. Oof. I guess we'll be "creating" random little 2 minute scenes that we have very little control over for the near future. They don't tell my story, but they look cool! /s
Preach! Totally feel your pain. Currently AI video generation is amazing for quick stuff and messing around [memes], but yeah, for serious filmmaking on a budget, it's still not quite there. Between the iteration costs and the post-production needed, it's a hefty price tag for solo creators.
for video advertisements I'd still use real people and scenes but if there's something it's like simulating then climbing everest or sailing through treacherous deep waters on a boat (to build the scene) I would try and implement AI there.
I guess I’ll start talking about this. Sora had given a few different versions of outputs. They showed the top versions with more compute. The lower versions with less compute look like luma. They simply said more compute equals more stable quality and details of what’s happening in the 3d representation created. So all runway or even luma has to do is increase the compute. Luma explained that the more people that use theirs the compute will raise due to some sort of factor on how it runs. Look into it. I just keep waiting for people to realize it and they never did. So I will start talking about this.
The big issue with Adobe is not AI training or ownership of content. The big issue is the the perpetual royalty free license to use your content any way Adobe wants.
such horror. learning off other humans work... without giving them money!?!?!? why, every good occidental knows that knowledge and data must be gate kept instead! accessing information to learn from is 🚨"STEALING"🚨, right-wing think tanks have been saying that for decades sheeple! you wouldn't download a car right?!!1!! 👢😝
Gen 3 Alpha will be superior to Dream Machine in many respects, but will most likely fall short in some aspects compared to Sora AI. Only testing will provide definitive answers to these questions.
from what i can see overall runwaygen3 is better but luma does pretty good with image to video (which runway 3 wont have at launch YET) i cant wait though im suprised there is not more hype for runway gen 3 . the clips ive seen on twitter thats NOT on their website is amazing too!
Midjourney personalization is something I will never use. Every project I do has a different style and look depending on 1000 factors. I don't have "a style". I don't want everything to look the same. I want things to look unique. I wish Midjourney would focus on improving their models, giving us 3D and video rather than things no one is asking for.
1. Runway gen-3 2. Luma Dream Machine 3. Runway gen-2 I suppose that you can use that Midjourney feature on separate projects and if will be added as a different style sets that will be convenient to choose working on the project and keeping up the style coherence
AI results in general have too many subjects breaking the fourth wall, staring into or acknowledging the camera- a big no-no in conventional filmmaking unless there’s a specific story driven need for it, and commonly overused by novice filmmakers.
This is what I keep telling people! I think I could watch AI films for hours and never need to watch the uninspired bureacratic mess that Hollywood films have become ever again.
Here's a thing. When everone's creating, who's gonna consume? 🤔 Why should you watch someone else's film, when you can make your own in a second? Hence, who'll watch yours apart from you?
Even without AI tools, there is endless amount of content out today. The idea applies. I believe there are days/months/year uploaded to UA-cam every minute.
1. Luma's Dream Machine
2. Runway Gen-3
3. Runway Gen-2
runway gen3 is better than luma i think at this point, but its hard to tell when we dont have access to it yet!!
Luma is not so bad , ❤
We love them all!
When will runway v3 be available?
Thoon!
It's available now!
1 - Runaway gen 2
2 - Runaway gen 3
3 - Luma Dream Machine
Hi, Great video thanks! i was wondering how did you manage to lock the camera in a campfire shot? is there any specific keyword for this? all my attempts to lock the camera in one place are failed. the camera seems always moving which is not what i want. would be happy if you could help with this. Thanks
We'll take a look a the prompt and report back!
9:26 From what I’ve heard, SD3 isn’t just $20 a month for the time that you’re subscribed to it. If you ever end your subscription, neither you or your clients can continue using any of the images that you generated with it. In other words, you’ll have to be subscribed forever if you want to keep any of the work you created with it at any point in the past. So, this is definitely not a good choice for client work, or for building a business on.
Great clarification! We'll have to look into it!
1. Luma's Dream Machine
2. Runway Gen-3
3. Runway Gen-2
Great video! Looking forward to a Runway Gen-3 with image prompts. For your Films of the Week, do films that are submitted have to be generated completely by AI or can they have live action in them as well?
Agreed! I2V should be soon! For AI Films we typically do mostly AI rather than LiveGen but we would consider anything.
8:28 playground has this system , for training . You need choose around 100 images .
.
P.S.
Make sure audio ai equals video ai, sometimes video is ahead of audio ai .
thanks for the tip!
do you know if RUNWAY will create an IMAGE TO VIDEO like Luma?
Undoubedly it will and we hope it will soon!
when will be available gen3?
It's available now!
haha just received the mail 🎉
We still need to convince a rich person to fund our actual film projects. With the hundreds of iterations required for every image which then get iterated again when attempting to create a useable video clip, and the costs of every tool being $20-40-100-499/month, and there being dozens of tools, the costs for a solo creator are still out of reach ($1000s per month just in subscriptions).
Using just Luma as an example, it will cost about $500 in credits just to get a single useable 5 minute clip from which you can trim one shot. And by usable, I'm referring to Shy Kids claim that they had to run 100s of iterations for each scene in their short balloonhead video from a few months ago... and the consistency in that was not great. And they had a ton of post production to do as well. And that was in Sora which purportedly takes hours to generate a single short clip. So hundreds of hours and 1000s of dollars to create each scene. Oof.
I guess we'll be "creating" random little 2 minute scenes that we have very little control over for the near future. They don't tell my story, but they look cool! /s
Preach! Totally feel your pain. Currently AI video generation is amazing for quick stuff and messing around [memes], but yeah, for serious filmmaking on a budget, it's still not quite there. Between the iteration costs and the post-production needed, it's a hefty price tag for solo creators.
for video advertisements I'd still use real people and scenes but if there's something it's like simulating then climbing everest or sailing through treacherous deep waters on a boat (to build the scene) I would try and implement AI there.
Building a planned scene would take hundreds of hours prompt experimenting and it would still be nit there.
True, this is certainly not cost efficient for the average creator right now.
1: Runway gen-2
2: Runway gen-3
3: Luma's Dream Machine
1. Runway Gen-3
2.Luma's Dream Machine
3. Runway Gen-2
2 weeks are a long time in our new AI world! : )
So true!
1.Runway Gen 2
2. Runway Gen 3
3. Luma's Dream Machine
This video is incredibly creative in showcasing new AI tools like Runway Gen 3 and its ambitious vision!
Thanks!
I think your Mike volume is a bit low .
Sorry, we'll work on that!
1. Runway Gen-3
2. Luma's Dream Machine
3. Runway Gen-2
1. Luma's Dream Machine
2. Runway Gen-3
3.Luma's Dream Machine
Great vid. Exciting time now!
1. Luma 2. RunwayG3 3. RunwayG2
Do you have acces to gen3?
We do now :)
1.Runway Gen-2
2.Runway Gen-3
3.Luma Dream Machine
I guess I’ll start talking about this. Sora had given a few different versions of outputs. They showed the top versions with more compute. The lower versions with less compute look like luma. They simply said more compute equals more stable quality and details of what’s happening in the 3d representation created. So all runway or even luma has to do is increase the compute. Luma explained that the more people that use theirs the compute will raise due to some sort of factor on how it runs. Look into it.
I just keep waiting for people to realize it and they never did. So I will start talking about this.
We appreciate your perspective!
1. Luma Dream Machine
2. Runway gen 3
3. Runway gen 2
1 - Luma Dream machine
2 - Runway Gen 3
3 - Runway Gen 2
1 - Runway Gen 3
2 - Luna dream Machine
3 - Runway Gen 2
The big issue with Adobe is not AI training or ownership of content.
The big issue is the the perpetual royalty free license to use your content any way Adobe wants.
such horror. learning off other humans work... without giving them money!?!?!? why, every good occidental knows that knowledge and data must be gate kept instead!
accessing information to learn from is 🚨"STEALING"🚨, right-wing think tanks have been saying that for decades sheeple! you wouldn't download a car right?!!1!! 👢😝
Interesting! This is definitely a big problem that will need to be solved this year.
Are you angry from stable diffusion 😅 ?
Angry? Not at all!
1 Luma
2 Runway Gen-3
3 Runway Gen-2
Runway Gen 2
Runway Gen 3
Luma
correct at least someone has a brain and an eye
@@petercarrick2678 to think that luma is third is wild
@@bbmusic6443 Luma is number one until I can use Runway Gen 3
I’m still not getting near the amazing useable quality outputs compared to their highlight reels, just saying quite frustrating
What prompts are you using? Let us know and hopefully we can help troubleshoot
1 - Luma Dream machine
2 - Runway Gen 3
3 - Runway Gen 2
The creative possibilities are pretty credible actually. :) The q is, what do you do with random scenes?
It's up to the creator! :)
@@curiousrefuge I meant that from a production pipeline, they're hardly as useful as one might think. :)
1. Luma Dream machine
2. Runway Gen 3
3. Runway Gen 2
1.- Runway gen-3
2.- Luma Dream Machine
3.- Runway gen-2
Video 1 : Runway Gen3
Video 2 : Luma Dream Machine
Video 3 : Runway Gen2
1-Luma
2-Runway Gen 3
3-Runway Gen 2
1. dream machine
2. runway gen 3
3. runway gen 2
Video1- Runway Gen 3
Video2- Luna dream Machine
Video3- Runway Gen 2
1. Luma dream machine; 2. Runway gen 3; 3. Runway gen 2
1. Luma Dream machine
2. Runway Gen-3
3. Runway Gen-2
1. Luma Dream Machine
2. Runway Gen 3
3. Runway Gen-2
1. Luma Dream Machine
2. Runway Gen-3
3. Runway Gen-2
1. Runway Gen 3
2. Luna dream machine
3. Runway Gen 2
1.RUNWAY-GEN3 2.LUMA 3.RUNWAY-GEN2
Gen 3 Alpha will be superior to Dream Machine in many respects, but will most likely fall short in some aspects compared to Sora AI. Only testing will provide definitive answers to these questions.
We'll be excited to see how it all plays out in the next few months!
Video 1: Runway Gen 3, Video 2:Runway Gen 2, Video 3: Luma
1. Runway Gen 3
2. Runway Gen 2
3. Lumas
I hope I live to see a time when you getq to put your favourite book into one of these and it recreats everything perfectly.
It won't be too long from now :)
: Runway gen-3
2: Luma Dream Machine
3: Runway gen-2
great update _ i am wating for Runway gen-3, Luma Dream Machine is not good enough.
It's out now! How do you feel about it?
1. Gen-3
2. Luma
3. Gen-2
1 - luma
2 - gen 3 runway
3 - gen 2 runway
1. Luma Dream machine
2. Runway Gen3
3. Runway Gen2
1. Luma
2. Gen 3
3. Gen 2
1. Luma
2. Runway gen 3
3. Runway gen 2
1. Luma dream machine
2. Runway Gen 3
3 Runway Gen 2
Congratulations! You are the winner of the AKOOL annual subscriptions. Please send us your email here: hello@curiousrefuge.com
@@curiousrefuge i have sent it.
1. Dream Machine
2. Gen 3
3. Gen 2
1 :luma dream machine
2: runway gen 3
3:runway gen 2
Video 1: Luma's Dreame Machine
Video 2: Runway Gen-3
VIdeo 3: Runway Gen-2
1. Luma
2. Runway Gen 3
3. Runway Gen 2
1: Gen 3
2: Luma
3: Gen 2
1. Runway gen 3
2. Luma
3. Runway gen 2
Game answears:
Runway Gen 3, Luma dream Machine, Runway Gen 2
Runway Gen 3
Luma Dream Machine
Runway Gen 2
1 Luma
2 Runaway 3
3 Runaway 2
1 luma dream machine
2 runaway gen 3
3 runaway gen 2
This is all cool. but I write horror and sci-fi and the restrictions will result in me not being able to create content.
It's true - it's very difficult at this point to generate material like that.
Runway Gen 2
Luma
Runway Gen 3
Good finds!! 😊
Thanks! 😊
the intro was so long i lost interest and left at 56 secs
And yet you wrote a comment! :) Thanks for tuning in!
guitar video to audio sounds nothing like it
1: Runway gen-3
2: Luma Dream Machine
3: Runway gen-2
1. runway gen 2, 2. luma, 3. runway gen 3
1- Luma's Dream Machine
2- Runway's Gen-3
3- Runway's Gen-2
No words on censoring, licensing, ban on civitAI on SD3 medium, you are really missing the point here, kinda wack
We'd be happy to do a video dedicated to those topics for you.
1: Luma ai
2: Runaway Gen 3
3: Runaway Gen 2
1) Luma 2)Gen-3 3)Gen-2
1. Gen 3
2. Luma
3. Gen 2
Its fine on my device
We're glad to hear that!
1 Luma
2 Gen-3
3 Gen-2
they said few days and it's still not here. same story as sora. what a sc,,am
It's here!
@@curiousrefuge is it public now? not the beta thing which is only accessible to some people
What an amazing time to be alive. In just a week we have Luma Labs, and now Hedra.
So many fun tools!
1. Luma
2. Gen 3
3. Gen 2
from what i can see overall runwaygen3 is better but luma does pretty good with image to video (which runway 3 wont have at launch YET) i cant wait though im suprised there is not more hype for runway gen 3 . the clips ive seen on twitter thats NOT on their website is amazing too!
We're excited for Runway Gen3 for the image2video :)
1.Luma
2. Gen 3
3. gen 2
Midjourney personalization is something I will never use. Every project I do has a different style and look depending on 1000 factors. I don't have "a style". I don't want everything to look the same. I want things to look unique. I wish Midjourney would focus on improving their models, giving us 3D and video rather than things no one is asking for.
True - it's certainly not for everyone !
Runway 3
Luma
Runway 2
luma dream machine
runway gen 3
runway gen 2
luma, gen 3, gen 2
1. Runway gen-3
2. Luma Dream Machine
3. Runway gen-2
I suppose that you can use that Midjourney feature on separate projects and if will be added as a different style sets that will be convenient to choose working on the project and keeping up the style coherence
AI results in general have too many subjects breaking the fourth wall, staring into or acknowledging the camera- a big no-no in conventional filmmaking unless there’s a specific story driven need for it, and commonly overused by novice filmmakers.
That's a really good and fair point! Something new filmmakers need to work on :)
2 is clearly runway gen 3 i know cuz i saw that clip lol
Shhhhh :)
in b4 curious refuge reveals his entire channel is literally ai including him.
Hah! Not true but that would be trippy!
1 Gen 3
2 Luma
3 Gen 2
ok guyz gen3 is out . time to do a cool video about it ;-) lol
Just released! Go check it out :)
1- Luma’s Dream machine
2- Runaway gen 3
3- Runaway gen 2
This is what I keep telling people! I think I could watch AI films for hours and never need to watch the uninspired bureacratic mess that Hollywood films have become ever again.
We agree - it's great to see a lot of new ideas!
Here's a thing. When everone's creating, who's gonna consume? 🤔 Why should you watch someone else's film, when you can make your own in a second? Hence, who'll watch yours apart from you?
Even without AI tools, there is endless amount of content out today. The idea applies. I believe there are days/months/year uploaded to UA-cam every minute.
definitely not anywhere close to your intro hyperbole... still a long way from anything narrative.. animated stock footage at best
We appreciate your perspective!
@@curiousrefugeokay, where are these AI generated blockbusters?