I have been sitting on Masters of the Universe He-Man and She-Ra movie storyboards for almost 2 decades. I doublt it can ever be done for a number of reasons. I see these AI videos and I get a tiny tiny tiny amount of hope. Good morning.
At this point, it's hard to say if we really need Sora. Right now, it doesn’t seem essential, and unfortunately, ClosedAI appears to have missed an opportunity. Their competitors have outpaced them, offering solutions that resonate more with what people are looking for. It's a tough space, and sometimes timing is everything. Sorry, ClosedAI, but it looks like you’ve been left behind in this race.
happens! even more so because less people on the AI model means less testing, less correcting itself, less learning. Almost a must these days to not wait on such things if the intention is to improve a model. I could also be wrong, who knows! If a trained model can make another untrained (trainee) model advance without all the hassle the trainer model had to do, then all this discussion is a bit of a mindFCk. I'll wait until these guys released something then my iteration trains on that model's failures, then my new version will be better///and so on and so on, not an expert but explained in very simplistic terms this is what I see with most models, LLM or image, music, and now video. What do you think?
@erkron You bring up an intriguing point! The process of continuous iteration and use is fundamental to the development of AI models. The idea that a trained model could guide an untrained one without requiring it to go through the same challenges is particularly exciting. If this concept can be fully realized, it could fundamentally change the way we think about training artificial intelligence, offering new avenues for innovation. Traditionally, training AI models involves exposing them to vast amounts of data and allowing them to gradually improve through multiple cycles of learning and correction. This is a time-consuming and resource-intensive process, which often requires substantial computing power and human intervention to fine-tune and adjust the model's performance. However, if we can leverage a pre-trained model to accelerate the learning of a newer, untrained model, it could dramatically reduce the time and resources needed to produce high-performing systems. The potential implications of this for various AI applications are enormous. For example, in language models, where context and nuance are critical, a trained model could pass on insights to its successor, allowing it to understand complex patterns much more quickly. In fields like image recognition or video processing, this kind of rapid training could lead to systems that can accurately identify objects, environments, or even emotional expressions with much greater efficiency. Moreover, this approach aligns with the broader trend in AI toward creating models that can not only learn but also self-improve and transfer knowledge. The idea of transfer learning, where one model helps another by sharing its learned experiences, is already being explored in many areas of AI. If this can be perfected, it could streamline the development process across various domains, from natural language processing to autonomous systems. Another interesting aspect of this concept is how it highlights the importance of adaptability in the AI space. As AI continues to evolve at an unprecedented rate, the ability to quickly integrate new methods and improvements becomes crucial for staying competitive. Models that can rapidly improve themselves and assist others in doing so could lead to a new generation of AI that evolves faster than ever before, driving advancements in areas such as healthcare, robotics, and entertainment. This kind of innovation also pushes the boundaries of our understanding of what AI can do. As we see more advancements in areas like reinforcement learning and unsupervised learning, the idea that AI systems could autonomously learn from one another becomes less of a futuristic notion and more of an achievable reality. It’s an exciting time for AI development, and the speed at which these systems can now evolve is staggering. I’m very curious to see how these innovations will shape the future of AI and its applications across multiple industries. With continuous iteration and the integration of these advanced learning techniques, the potential for AI seems virtually limitless.
You mean they're the ones not actually scamming morons with a non-functional product, so they're losing out? If they do come out with a working product with a satisfactory output rate, they're the ones I will give money to.
@@RetroFutureArts You're right, a working product is obviously the bare minimum. But it's more than that-it's about staying competitive, meeting needs faster, and evolving with the market. Even if ClosedAI isn't scamming anyone, their timing and ability to keep up are critical. The real challenge is delivering when it matters most.
@@erkronOne benefit OpenAI has is it can use its GPT models for validating the output. GPT is extremely good at describing images. I’d imagine the GPT models are much faster and better when they have a top grade processor that’s dedicated to specifically training models. After seeing what o1 can do, I’m convinced their models could self improve or enhance other models.
This update is a game-changer! The improved realism and precise control over characters, elements, and motion in Cling AI 1.5 is mind-blowing. It's incredible to see how even complex actions like eating and dancing are now handled so naturally. The new features, like motion brush and end frames, offer creators unparalleled creative control. I'm really excited to see where this takes AI video creation!
6:20. I love how he puts a “Real” hand for reference there, for those of us who don’t know what real hands look like and therefore couldn’t tell that there was anything wrong with the picture on the right
I tried Kling AI and it worked fine for a few videos then stopped working. The video generation is stuck at 99% done while the message "generating video" appears for several days without any video appearing.
Great videos you have :)) From my perspective, it is almost impossible to imagine, now when we watch these videos, that 3D engine tools are not already being used in the AI video generation process to ensure everything aligns and works as it should.
How would you do lipsync with Kling? This is an extremely useful feature, which made me prefer Runway so far. Is there a way to add lipsync? Please let me know.
I haven't seen you around on Instagram lately and wanted to let you know that you've been missed! Your content always brings such a fresh perspective, and I love following your journey. Keep up the fantastic work - it's always inspiring to see what you're creating next!
I have a video I set up more than 24 hours ago, and it's still processing. Is this common? Did yours go quickly because you have a paid membership? I want to try it before I figure out how many credits I need for a particular project.
Kling 1.5 does have a real slow motion problem, alot of videos delivered in slowmo even when you try to prompt it out. I think we need speed settings Kling!
Hollywood is soo toast... had problems prompting KlingAi v1 though, it didn't really do what i asked it to do - usually just pan and zoomed, totally ignoring the action in the prompt i stopped using v1 now is it worth it to try v2? i don't need pan and zoom in FHD tbh but maybe my prompts are just missing the mark
@@heartshinemusic still tryna wrap my head around all the latest advances. My main concern, when will we be able to generate a whole 30 second clip . Instead of 5-10seconds at a time
I tried Kling, but there is no "generate" button. I uploaded an image for an image to image video, but I don't see any way to generate it. Can anyone help?
Hi AI Samson , each time i try kling 1.5 image to video , image are created with copilot ai and upscale rework on leonardo ai than i go to kling to animate , i have negative prompt , error , artifact ect ect... i always have some weird small artifact in the video , any tips , solution? Thanks
Hi Samson, How are you today, Now I want to buy premium Kling subscription, but I don't know how to make consistent characters and consistent style in Kling image generator, and I don't have money to buy Kling and Leonardo together or Kling and Midjourney together. can you make a video about what Kling image generator can do for consistent characters and consistent style. thank you very much.
Hi I tried to download the prompt document but there is an error. Any help would be most appreciated. - this is the text message I am getting- {"error":{"code":"INVALID_PARAMETERS","message":"This form has missing or invalid fields."}}
MiniMax is still the King (for now) in most areas of text-to-video, and for a 100% free service (for now) it cannot be beat. Kling and Runway still are mostly slow motion, and even most of their "fast motion" is just a faster slow motion, whereas MiniMax by default renders videos in normal speed motion, which is really the whole point of having characters move in films, so that they move in normal speed. 99.9% of almost every film is in normal speed, for that is what our brains and eyes are most used to. MiniMax is the A.I. filmmaking KING, overall, even without image-to-video, which they are getting soon...
@RetrowaveMemories Since MiniMax is currently 100% free (as well as better overall at following prompts and animating correct human actions without morphing), it is in its own category in this sense and is THE KING for now, and Kling and Runway are quite a bit behind. Their morphing is a big 👎, unless one deliberately likes morphing for their A.I. films, hehe.
Hi I have a question which I have already asked on so many dialogs but couldnt managed to find the answer yet . I want to subscribe klings pro plan but I want to know if they will continue to give me 66 per daily credit besides the 3000 monthly credit
🤣Yes, just what I thought, he looks nothing like Will Smith. MiniMax can render a perfect Will Smith video with just text, no image to video needed (which is a feature they will get soon, though).
Kling sux. Have 2 videos that have been at 99% for a week. If you pay for a standard plan theres not enough credits for anything good because the new model uses like 3 times the credits and its the only model that gives you a good result. If you try to upgrade your plan itll make you puchase it as a second plan and not upgrade your current plan
If only I could get a consistent character (look and clothing) across video generations. It seems impossible to create different scenes with the same character 😞 That's the problem with AI, its just too creative, not enough consistency.
Is that Will Smith video going to become part of our history, something we’ll always look back on when it comes to AI? I mean I can’t imagine always having my face distorted for the whole world to see on a consistent basis. Did they start putting that video up before the slap or after ? CHATGPT Are you referring to the Will Smith AI-generated videos or deepfakes? If so, these types of AI content involving him have been circulating both before and after the infamous slap incident at the Oscars. However, the frequency of AI-generated content involving Will Smith, likely increased after the slap due to the surge in media attention and internet culture trends.
Ok, I guess in less then 1 Year, a single Person is able to produce a blockbuster movie within a month, with extremly low costs. Its so amazing, how we will be able to create our own movies. I am 100% going to try this!
@@melanp4698 yes of course. But that can be streamlined with AI also. I still think that, although Video generation will go far very fast, the small details and morphing(a fly flying behind the head and coming out of the other side for example) will take AI some time to perfect it.
@@MircoBlumenthal Yes and consistency, emotions, voices etc. Making a movie is way more than just pretty stuff. You need full creative control over every scene to have characters interact, environments etc. Plus like i said, have the ability to also actually write a good script, which is the biggest issue of all. I'd say no chance within 5 years at least.
@@melanp4698 you are aware even the script can be written by ai. the times are a changing the the movie industry will change too. i have been making films for 20 years and the power we have now is crazy and should not be downplayed
@@todaychange5-7783 MiniMax does it better with just text to video, and overall follows prompts better than Kling and Runway. Runway and Kling are mostly only slow motion, and even their "fast motion" is just a faster slow motion in most cases. MiniMax renders NORMAL SPEED by default, for they do not have the morphing that Kling and Runway has, in most cases. MINIMAX for the WIN... 😆👍👍
I really want to test it out but kling 1.5 requires a PRO account which is 300$ a year!! that's a lot to invest in a tool that was mostly disappointing up until now :( I wish they would come out with a testing period, so I can see if it really is so good.... the 1.0 version is pretty crappy to be honest
good, but explain how to record videos with the same person in each video, how not to lose the created character, whether it is a tool that allows for continuity of creation, because only in this case the tool will be really useful.
You really can't, thats why Samson (and I) are suggesting to use an image generator like Midjourney. Midjourney support character referencing, a way to keep people mostly constant based on a (generated) reference image. You can see that effect (not perfectly, but good enough) in the latest music video 'Clear Blue Sky' on my bands youtube channel. FYI: that movement was crated using runway gen3, not kling, because I still find kling too expensive. Only runway has an unlimited plan.
Here is a real and honest review. Ref image: Model HR image, Caucasian, black hair, stand up, portrait, torso normal, nothing special, create 4 images. Image from image: simple prompt ask for, smile, different background , reference 0 Pro: Consistent, Hands are reasonable good Con: Simple questions in prompt not working like kiss, kiss lips, smile; tries to make model thinner whilst not in prompt, and not realistic, often fail error - Image from Image: same but reference -- (minus), Pro: Nice looking model, hardly like model, hands etc. are good. Con: None Tried to color black hair to blond: in prompt Good realistic result but 1 out of 4 and the others fail. To make them smile is doable, to make slips look like a kiss with different prompts, no success (which is bad IMO) - Generate model from prompt, NO reference: Pro: Nice consistent result with different clothes, prompt 90% visible in image Con: Smile no problem but again no kiss and winking one eye also not showing. But in general usable results - Sometimes process will keep spinning and there is no image generated, not even after hours, can't find a way to stop it. - Image to video Model , face facing camera, total body facing to the left, enough space at left side to move (walk). Simple prompt: looks at camera, smile, walking. Took 2 days to finish (rolling eyes), 5 sec video Face changed too much, became somebody different, hand highly deformed. This is far for good . This is not worth paying for IMO
So far even Runway seems to be a more useful tool, Kling is terribly inconsistent, it is almost impossible to make it do everything according to the script or at least not change people's faces
Realism isn't boring. Real life is rendered in "realism," for it is what our brains and eyes are most used to seeing on a daily basis, not cartoony stuff. Also, the realism is Kling and Runway is still not as real looking as the real world. Fantasy doesn't mean "cartoony" in itself. ALL fictional movies are forms of fantasy, by definition. Try again...
@wizards-themagicalconcert5048 No one needs to "prove you wrong" because you have done that yourself. 😅 Kling is already being used for animating awesome "fantasy art," as so is Runway and Luma. What planet have you been on? Try again...
It's fascinating to see how Samson's mind works when creating something like Kling 1.5. The attention to detail and the process behind bringing such realism to life is truly impressive. Every step reveals a deeper level of creativity, pushing the boundaries of what's possible with AI. Keep up the amazing work, this is a real game-changer!
@@sunwalker4017 yeah I’ll be cancelling before then most likely, as usually w overhyped Ai latest, my outputs already aren’t at the level of the teaser vids
It is not all that great when you have to use “only” what Kling has conceived. Try to get it to work on an image you upload as the model of your character or you want to enhance your image in some way and it still be the image you started out with. This one should be easy, but it is impossible. Here is a test. Upload a not so good face shot of an orangutan, ask Kling to enhance it to be more life like. After a few rounds tell it you want the image with a smile, a grin, a puzzled look, instead of a huge gaping mouth open image. Good luck. Kling likes to invent stuff. Just not your stuff.
Who's going to watch AI generate imitations of real world? I immediately switch off this type of content because there is absolutely zero benefit. No value at all. How stupid it is to promote own demise. Soon AI will replace you talking to the camera promoting AI stuff and the algorithm will favor content made by AI instead of human.
your channel is great, can you record a video where you turn a written story, e.g. about dogs or other animals, into a video in several episodes, present how you work on a specific task step by step, even in several videos
Because they want you to pay subscription - common practice. They lure people in with free accounts, but then they do things like that so you buy subscription.
@@CarlVicars Got it, thanks a lot. Will give it test and see wether it is as good as they say cause the examples shown are tested before. I will do my basic simple tests (where most of the AI apps will struggle) with new AI Apps and see what we get.
unfortunately the motion features are not available in 1.5, and there was a huge leap in quality between both versions , 1.0 even with the feature ends up being kinda awful. Be ready to lose a bunch of credits before you got one that isnt too bad
Shtp u f00l Ai samson... kling is not king he still very stpd i know very much If if u think its king then I challenge u to make use it perfectly in 3 example of famous new trend on internet name ::fruits turn into animals/birds:: Will u??????
I see people walking way more strangely than the guy walking backwards in my neighborhood 😅
I have been sitting on Masters of the Universe He-Man and She-Ra movie storyboards for almost 2 decades.
I doublt it can ever be done for a number of reasons. I see these AI videos and I get a tiny tiny tiny amount of hope.
Good morning.
At this point, it's hard to say if we really need Sora. Right now, it doesn’t seem essential, and unfortunately, ClosedAI appears to have missed an opportunity. Their competitors have outpaced them, offering solutions that resonate more with what people are looking for. It's a tough space, and sometimes timing is everything. Sorry, ClosedAI, but it looks like you’ve been left behind in this race.
happens! even more so because less people on the AI model means less testing, less correcting itself, less learning. Almost a must these days to not wait on such things if the intention is to improve a model. I could also be wrong, who knows! If a trained model can make another untrained (trainee) model advance without all the hassle the trainer model had to do, then all this discussion is a bit of a mindFCk. I'll wait until these guys released something then my iteration trains on that model's failures, then my new version will be better///and so on and so on, not an expert but explained in very simplistic terms this is what I see with most models, LLM or image, music, and now video. What do you think?
@erkron You bring up an intriguing point! The process of continuous iteration and use is fundamental to the development of AI models. The idea that a trained model could guide an untrained one without requiring it to go through the same challenges is particularly exciting. If this concept can be fully realized, it could fundamentally change the way we think about training artificial intelligence, offering new avenues for innovation.
Traditionally, training AI models involves exposing them to vast amounts of data and allowing them to gradually improve through multiple cycles of learning and correction. This is a time-consuming and resource-intensive process, which often requires substantial computing power and human intervention to fine-tune and adjust the model's performance. However, if we can leverage a pre-trained model to accelerate the learning of a newer, untrained model, it could dramatically reduce the time and resources needed to produce high-performing systems.
The potential implications of this for various AI applications are enormous. For example, in language models, where context and nuance are critical, a trained model could pass on insights to its successor, allowing it to understand complex patterns much more quickly. In fields like image recognition or video processing, this kind of rapid training could lead to systems that can accurately identify objects, environments, or even emotional expressions with much greater efficiency.
Moreover, this approach aligns with the broader trend in AI toward creating models that can not only learn but also self-improve and transfer knowledge. The idea of transfer learning, where one model helps another by sharing its learned experiences, is already being explored in many areas of AI. If this can be perfected, it could streamline the development process across various domains, from natural language processing to autonomous systems.
Another interesting aspect of this concept is how it highlights the importance of adaptability in the AI space. As AI continues to evolve at an unprecedented rate, the ability to quickly integrate new methods and improvements becomes crucial for staying competitive. Models that can rapidly improve themselves and assist others in doing so could lead to a new generation of AI that evolves faster than ever before, driving advancements in areas such as healthcare, robotics, and entertainment.
This kind of innovation also pushes the boundaries of our understanding of what AI can do. As we see more advancements in areas like reinforcement learning and unsupervised learning, the idea that AI systems could autonomously learn from one another becomes less of a futuristic notion and more of an achievable reality. It’s an exciting time for AI development, and the speed at which these systems can now evolve is staggering.
I’m very curious to see how these innovations will shape the future of AI and its applications across multiple industries. With continuous iteration and the integration of these advanced learning techniques, the potential for AI seems virtually limitless.
You mean they're the ones not actually scamming morons with a non-functional product, so they're losing out?
If they do come out with a working product with a satisfactory output rate, they're the ones I will give money to.
@@RetroFutureArts You're right, a working product is obviously the bare minimum. But it's more than that-it's about staying competitive, meeting needs faster, and evolving with the market. Even if ClosedAI isn't scamming anyone, their timing and ability to keep up are critical. The real challenge is delivering when it matters most.
@@erkronOne benefit OpenAI has is it can use its GPT models for validating the output. GPT is extremely good at describing images.
I’d imagine the GPT models are much faster and better when they have a top grade processor that’s dedicated to specifically training models.
After seeing what o1 can do, I’m convinced their models could self improve or enhance other models.
This update is a game-changer! The improved realism and precise control over characters, elements, and motion in Cling AI 1.5 is mind-blowing. It's incredible to see how even complex actions like eating and dancing are now handled so naturally. The new features, like motion brush and end frames, offer creators unparalleled creative control. I'm really excited to see where this takes AI video creation!
6:20. I love how he puts a “Real” hand for reference there, for those of us who don’t know what real hands look like and therefore couldn’t tell that there was anything wrong with the picture on the right
I tried Kling AI and it worked fine for a few videos then stopped working. The video generation is stuck at 99% done while the message "generating video" appears for several days without any video appearing.
Same here
@@Kali-dw5cq I'm afraid it's being overwhelmed with interest. But might be trying to encourage you to spend some money. Try different times of day.
I have had the exact same experience! Very, very disapointing
It does that for free users... Won't do that if you pay for it
@@AdamReese-wl1fz Try throwing money at it
Are all of the AI clips seen here being created from a still image? Is that the first step... still image and then Kling 1.5 for the animation?
Great videos you have :)) From my perspective, it is almost impossible to imagine, now when we watch these videos, that 3D engine tools are not already being used in the AI video generation process to ensure everything aligns and works as it should.
Best overview with great advice I have seen. Thank you
I am glad i got in early and paid for the 1 year sub
How would you do lipsync with Kling? This is an extremely useful feature, which made me prefer Runway so far. Is there a way to add lipsync? Please let me know.
What do you thing of Kling 1.5?
I haven't seen you around on Instagram lately and wanted to let you know that you've been missed! Your content always brings such a fresh perspective, and I love following your journey. Keep up the fantastic work - it's always inspiring to see what you're creating next!
Awesome
@@I-Dophler Thanks Dophler!
@@aisamsonreal I've missed you.
👋
I cant wait to try KLING on my next music video!! ! WOW! Thanks so much for your review of this AI tool 😍😍😍😍😍
It is unclear what is the Plan required for 1.5. They say premium but what is premium exactly? Is it the Premier Plan or any pay plans?
I have a video I set up more than 24 hours ago, and it's still processing. Is this common? Did yours go quickly because you have a paid membership? I want to try it before I figure out how many credits I need for a particular project.
Amazing video as always ,Thank you and yes Klingai 1.5 made a very step up about realism and quality
Can you use the same character across different scenes?
Samson, another excellent video. I come to you to see what the latest improvements are. thank you.
The security verification puzzle doesn't work. I can't log in. I entered a US address via VPN but it still doesn't work
The only problem with a sign up is that captcha doesn’t work so it’s not possible to use it.
Kling 1.5 does have a real slow motion problem, alot of videos delivered in slowmo even when you try to prompt it out. I think we need speed settings Kling!
HI. You mentioned that there is a guide in Kling AI. Can you tell me where it is and how to access it?
In the description 👍
Uhhh so I'm trying to generate with Kling and it has been saying will be finished soon... since yesterday.
Hollywood is soo toast... had problems prompting KlingAi v1 though, it didn't really do what i asked it to do - usually just pan and zoomed, totally ignoring the action in the prompt
i stopped using v1
now is it worth it to try v2? i don't need pan and zoom in FHD tbh
but maybe my prompts are just missing the mark
Thanks! Nice overview... Damn, can you imagine Kling 2.0!
@@heartshinemusic still tryna wrap my head around all the latest advances. My main concern, when will we be able to generate a whole 30 second clip . Instead of 5-10seconds at a time
So what they are saying is the most you can create is 1.5 minutes per month, how can this possibly be usable?
Excellent video 😮!
Major problem - you can't get past the security verification procedure here in the UK to use Kling. .
Is it safe? Considering its a Chinese software?
14:40 - the splash happens before it hits the water though, lol
It's the Apple's force field.
Does kling offer unlimited generation plans like runway? This is the only thing holding me back from trying it.
I tried Kling, but there is no "generate" button. I uploaded an image for an image to image video, but I don't see any way to generate it. Can anyone help?
Hi AI Samson , each time i try kling 1.5 image to video , image are created with copilot ai and upscale rework on leonardo ai than i go to kling to animate , i have negative prompt , error , artifact ect ect... i always have some weird small artifact in the video , any tips , solution? Thanks
This is amazing! Thank you for this video. I love your teaching 🙏
Everytime I try to use Kling image to video, the video gets stuck on 99% and never finishes I’ve tried several times in the last month
I tried it and it downloaded an exe file instead of a video file! VIRUS ALERT!
Can you make Ai videos of a band playing a concert with all band members playing instruments 🥁 and the crowd reaction?
still no way to bypass the nsfw filter though :(
What about mixing 1,5 then 1,0 ( form motion brush) and back? Ah, theres no vid to vid..
Hi Samson, How are you today, Now I want to buy premium Kling subscription, but I don't know how to make consistent characters and consistent style in Kling image generator, and I don't have money to buy Kling and Leonardo together or Kling and Midjourney together. can you make a video about what Kling image generator can do for consistent characters and consistent style. thank you very much.
I watch your videos to see how fast AI progresses.
Great video, well done :)
Every week a new advance. Those dogs were pretty realistic
Great video!
does free account still stuck at 99% for days?
Of ccourse, free will give you free. Standard will give you results in a minute and Pro will give you in seconds.
Yeah, haven't been able to generate anything for a week now
Yeah, pay for service 😄 They don't care about people who are freeloaders.
@@wizards-themagicalconcert5048 As they should😂 and the free results isn't even that good
Actually managed to create a video just now, so there is hope.
Hi I tried to download the prompt document but there is an error. Any help would be most appreciated. - this is the text message I am getting- {"error":{"code":"INVALID_PARAMETERS","message":"This form has missing or invalid fields."}}
Hi, you need to submti your email to get it
Amazing new competetion to minimax
I can't wait for their features whenever its ready
MiniMax is still the King (for now) in most areas of text-to-video, and for a 100% free service (for now) it cannot be beat. Kling and Runway still are mostly slow motion, and even most of their "fast motion" is just a faster slow motion, whereas MiniMax by default renders videos in normal speed motion, which is really the whole point of having characters move in films, so that they move in normal speed. 99.9% of almost every film is in normal speed, for that is what our brains and eyes are most used to. MiniMax is the A.I. filmmaking KING, overall, even without image-to-video, which they are getting soon...
@RetrowaveMemories Since MiniMax is currently 100% free (as well as better overall at following prompts and animating correct human actions without morphing), it is in its own category in this sense and is THE KING for now, and Kling and Runway are quite a bit behind. Their morphing is a big 👎, unless one deliberately likes morphing for their A.I. films, hehe.
@Retro "...new competetion [competition]..."
Is this video even real? 👀 Or is Samson actually AI-rendered? 😱
;-)
Why are these apps all limited to 10 second clips? Super tedious to even make a 3-minute video.
if I want to use a specific yacht model and enter it from a photo, can I use it in multiple recordings?
When will we be able to generate 30 second clips? The whole 5 to 10 second generations really have limited use in real film
Mouth matching on *Kling* !
I would never pay for a art program that censors creativity.
Hi I have a question which I have already asked on so many dialogs but couldnt managed to find the answer yet . I want to subscribe klings pro plan but I want to know if they will continue to give me 66 per daily credit besides the 3000 monthly credit
Please teach me how you do this!
Hi Ai😊👍
I'm not sure that "Will Smith" shot in your intro is carrying the water you think it is--unless you meant Will Smith from accounting. Then yeah, okay.
🤣Yes, just what I thought, he looks nothing like Will Smith. MiniMax can render a perfect Will Smith video with just text, no image to video needed (which is a feature they will get soon, though).
Kling sux. Have 2 videos that have been at 99% for a week. If you pay for a standard plan theres not enough credits for anything good because the new model uses like 3 times the credits and its the only model that gives you a good result. If you try to upgrade your plan itll make you puchase it as a second plan and not upgrade your current plan
If only I could get a consistent character (look and clothing) across video generations. It seems impossible to create different scenes with the same character 😞 That's the problem with AI, its just too creative, not enough consistency.
Is that Will Smith video going to become part of our history, something we’ll always look back on when it comes to AI?
I mean I can’t imagine always having my face distorted for the whole world to see on a consistent basis.
Did they start putting that video up before the slap or after ?
CHATGPT
Are you referring to the Will Smith AI-generated videos or deepfakes? If so, these types of AI content involving him have been circulating both before and after the infamous slap incident at the Oscars. However, the frequency of AI-generated content involving Will Smith, likely increased after the slap due to the surge in media attention and internet culture trends.
How is nobody saying how scary this is?
Nothing on the Internet will be believable within the next 12 months.
Ok, I guess in less then 1 Year, a single Person is able to produce a blockbuster movie within a month, with extremly low costs. Its so amazing, how we will be able to create our own movies. I am 100% going to try this!
Waaaaaay more than a year. Also writing a script takes a long time. Also a lot of other things :)
@@melanp4698 yes of course. But that can be streamlined with AI also. I still think that, although Video generation will go far very fast, the small details and morphing(a fly flying behind the head and coming out of the other side for example) will take AI some time to perfect it.
@@MircoBlumenthal Yes and consistency, emotions, voices etc. Making a movie is way more than just pretty stuff. You need full creative control over every scene to have characters interact, environments etc. Plus like i said, have the ability to also actually write a good script, which is the biggest issue of all.
I'd say no chance within 5 years at least.
@@melanp4698 you are aware even the script can be written by ai. the times are a changing the the movie industry will change too. i have been making films for 20 years and the power we have now is crazy and should not be downplayed
Do you consider Kling beginner friendly. I’m not a digital native but I’m fascinated by AI.
I don't want realism; I want AI anime!
Then use Runway
Then use image-to-video for that. Kling and Runway both animate anime characters decently, but MiniMax can do it better with text to video.
@@todaychange5-7783 MiniMax does it better with just text to video, and overall follows prompts better than Kling and Runway. Runway and Kling are mostly only slow motion, and even their "fast motion" is just a faster slow motion in most cases. MiniMax renders NORMAL SPEED by default, for they do not have the morphing that Kling and Runway has, in most cases. MINIMAX for the WIN... 😆👍👍
@@FilmSpook It's still not good enuff. Way too much warping and distortions.
@@todaychange5-7783 Are you getting solid consistent results with Runway and anime images?
I really want to test it out but kling 1.5 requires a PRO account which is 300$ a year!! that's a lot to invest in a tool that was mostly disappointing up until now :(
I wish they would come out with a testing period, so I can see if it really is so good.... the 1.0 version is pretty crappy to be honest
avery one can open kling successfully, because I face difficulty.
❤❤❤❤❤
But is not unlimited
Dear New UA-camrs, May We succeed 🥰🔥🔥❤❤❤❤
🥰🔥🔥❤❤
see how movies will be able to be made with 10.000€ in a couple prompts but we still need to pay the same amount monthly as subscription. kek.
work on the audio, that wet clicking is annoying
good, but explain how to record videos with the same person in each video, how not to lose the created character, whether it is a tool that allows for continuity of creation, because only in this case the tool will be really useful.
You really can't, thats why Samson (and I) are suggesting to use an image generator like Midjourney. Midjourney support character referencing, a way to keep people mostly constant based on a (generated) reference image. You can see that effect (not perfectly, but good enough) in the latest music video 'Clear Blue Sky' on my bands youtube channel. FYI: that movement was crated using runway gen3, not kling, because I still find kling too expensive. Only runway has an unlimited plan.
Here is a real and honest review.
Ref image: Model HR image, Caucasian, black hair, stand up, portrait, torso normal, nothing special, create 4 images.
Image from image: simple prompt ask for, smile, different background , reference 0
Pro: Consistent, Hands are reasonable good
Con: Simple questions in prompt not working like kiss, kiss lips, smile; tries to make model thinner whilst not in prompt, and not realistic, often fail error
-
Image from Image: same but reference -- (minus),
Pro: Nice looking model, hardly like model, hands etc. are good.
Con: None
Tried to color black hair to blond: in prompt Good realistic result but 1 out of 4 and the others fail.
To make them smile is doable, to make slips look like a kiss with different prompts, no success (which is bad IMO)
-
Generate model from prompt, NO reference:
Pro: Nice consistent result with different clothes, prompt 90% visible in image
Con: Smile no problem but again no kiss and winking one eye also not showing.
But in general usable results
-
Sometimes process will keep spinning and there is no image generated, not even after hours, can't find a way to stop it.
-
Image to video
Model , face facing camera, total body facing to the left, enough space at left side to move (walk).
Simple prompt: looks at camera, smile, walking.
Took 2 days to finish (rolling eyes), 5 sec video
Face changed too much, became somebody different, hand highly deformed.
This is far for good .
This is not worth paying for IMO
So far even Runway seems to be a more useful tool, Kling is terribly inconsistent, it is almost impossible to make it do everything according to the script or at least not change people's faces
Kling sucks for concept art, fantasy art or other fantastic concept art imho ! Please profe me wrong ! It's just good mostly in boring realism !
Realism isn't boring. Real life is rendered in "realism," for it is what our brains and eyes are most used to seeing on a daily basis, not cartoony stuff. Also, the realism is Kling and Runway is still not as real looking as the real world. Fantasy doesn't mean "cartoony" in itself. ALL fictional movies are forms of fantasy, by definition. Try again...
@wizards-themagicalconcert5048 No one needs to "prove you wrong" because you have done that yourself. 😅 Kling is already being used for animating awesome "fantasy art," as so is Runway and Luma. What planet have you been on? Try again...
It's fascinating to see how Samson's mind works when creating something like Kling 1.5. The attention to detail and the process behind bringing such realism to life is truly impressive. Every step reveals a deeper level of creativity, pushing the boundaries of what's possible with AI. Keep up the amazing work, this is a real game-changer!
But by god it seems slow on generation for me
That is why i do not want to watch news, because they could be manipulated with this tool.
I think your upload might have broken the website 😅 nothing is generating
Takes forever
Signing up for the $12.88 now!! 🎉
but only for the first month :) after that $32.56
@@sunwalker4017You can cancel before that😏
@@sunwalker4017 yeah I’ll be cancelling before then most likely, as usually w overhyped Ai latest, my outputs already aren’t at the level of the teaser vids
looks nice, explain how to record videos that are episodes of a story, whether it is already possible or not yet.
can someone just use an old real photo please
The first profession that AI will take away is Pornstars.
It is not all that great when you have to use “only” what Kling has conceived. Try to get it to work on an image you upload as the model of your character or you want to enhance your image in some way and it still be the image you started out with. This one should be easy, but it is impossible. Here is a test. Upload a not so good face shot of an orangutan, ask Kling to enhance it to be more life like. After a few rounds tell it you want the image with a smile, a grin, a puzzled look, instead of a huge gaping mouth open image. Good luck. Kling likes to invent stuff. Just not your stuff.
Not comfortable sharing personal information with a chineze company, there are plenty of ai outthere to choose from.
Chinese so def not subbing to that.
Who's going to watch AI generate imitations of real world? I immediately switch off this type of content because there is absolutely zero benefit. No value at all. How stupid it is to promote own demise. Soon AI will replace you talking to the camera promoting AI stuff and the algorithm will favor content made by AI instead of human.
your channel is great, can you record a video where you turn a written story, e.g. about dogs or other animals, into a video in several episodes, present how you work on a specific task step by step, even in several videos
I just registered for free and the web always gets stuck at 99% generating for me.
And takes credits in advance.
Because they want you to pay subscription - common practice. They lure people in with free accounts, but then they do things like that so you buy subscription.
@@ToneVirtue
Yeah...
It's like if we sell 1mil for just 1 month we are golden even if our product is shit.
Perfect startup mindset 😎
You can't even get a free trial account cause the verify puzzle doesn't work (for months)
Hold and drag arrow. Not puzzle piece. It took me a while to figure it out.
@@CarlVicars Thanks in advance I will try it in the morning and do some test.
I test all those AI apps and write about it.
I'll get back to you
@@CarlVicars Got it, thanks a lot.
Will give it test and see wether it is as good as they say cause the examples shown are tested before.
I will do my basic simple tests (where most of the AI apps will struggle) with new AI Apps and see what we get.
you failed the puzzle because you're an ai
@@Gaze73 Should have known. Also failed my own Turing test ;-)
6:57 dude are you for real she looks like skeletor !!!! 😂
unfortunately the motion features are not available in 1.5, and there was a huge leap in quality between both versions , 1.0 even with the feature ends up being kinda awful. Be ready to lose a bunch of credits before you got one that isnt too bad
me seeing my power bill at 7:55
Shtp u f00l Ai samson... kling is not king he still very stpd i know very much
If if u think its king then
I challenge u to make use it perfectly in 3 example of famous new trend on internet name ::fruits turn into animals/birds::
Will u??????