Very informative! I believe your closing argument was well stated. Learn and understand this new technology is the best path forward for any individual, company or country.
This is awesome ,as later anyone can direct their own movies with their own stories, values and customs from anywhere in the world. No restrictions and zero gate keeping. Anyone will be an actor or actress or their own movie studio,regardless of geography or connections.
Just a heads up: your audio is peaking severely when you speak at higher volumes and gets very distorted/crunchy. This can be fixed easily by adjusting your recording volume settings (lower the recording volume so that you don't go in the red by testing out how loud you want to speak before you set your recording volume), but I've noticed it on other videos of yours so I thought I'd mention it!
Not sure I agree with your point about striking a pose at the end not being possible. For the AI to do what it does, must have figured out tracking against time. Therefore it has an internal timeline, by which it fits the video. My suspicion is it does not create the videos like humans would. It knows the end time stamp and the starting, and works forward, backwards, and in-between, to ensure all the objects and motion fit, before it begins to animate. Sora would have an internal database that tracks all the objects in a scene, which it may dump at the end of the task ( or keep for permanence as in remembering long prompts and relating new prompts to previous Ai can now observe an image, but visuals are just for humans. I'm sure it focuses on the mathematical construct within its NN. Therefore it will have all the info it will need to impromptu edit ( front, back, and middle). I'm almost certain everything in the generated world is tied to a key co-oord value, as is currently done in video games, giving access to rotate, yaw etc.
I lived 15 years before the internet was mainstream and I saw simple concepts like email initially inundated with FUD by the media and the public in general. I'm sure there's some eventual downside to AI but it won't be 1/10 of what we fear it to be today.
WHY SHOULD MICROSOFT PROFIT OFF OF A MACHINE THAT LEARNS FROM OTHERS" WORK? Here's a bunch of films based on motion from unreal engine 5. Pay us, not them. How is everything OpenAI does not derivative? And why the fuck is everything it's focused on ENTERTAINMENT? Like why not AI boring and tedious jobs????
@@Gerlaffy LOL It sees and perceives nothing on it's own. It takes datasets and based on what it sees it reassembles images from bits of all the work it's looked at. Not in any kind of creative or artistic way at all. It's a literally producing the average of what's in the dataset. Flood the dataset with trash and you destroy AI's output.
@@Kimberly_SparklesJust wrong. AIs have been given all sorts of content and yet they perform better than you and I in many tasks, they can comprehend and focus on the details you know? How do you think an AI can play Minecraft successfully autonomously without being told what to do or how it works?
@@Kimberly_Sparklesgarbage in garbage out is true broadly speaking but it isn't simply making a collage of previous ideas it is generating new content and was trained/tuned with a given dataset and maybe human reinforcement and or working with external data sources. Saying it just gives you the average is super reductionist and just not even true, it depends on the specific AI model you're talking about and parameters used when promoting and positive vs negative prompts, lots of factors go into what the output is for a given input and a given model. Talking about details of AI systems as though they are all one thing or work in one simple way is just wrong though.
You example of an "impossible" task for SORA is not a good one. OpenAI is using the Vision Transformer research by DeepMind which should be able to do the equivalent of picture infill but with video frames. In much the same way that Dalle and StableDiffusion allows users to delete parts of an image and have the AI fill it in with something plausible or even extrapolate outside of the picture frame, SORA is capable of extrapolating a few frames of video to a full minute clip. So to do your task, just delete the last few seconds and prompt it to have the walking lady do a twirl. You can even load you own video clip in and ask Sora to create a seamless splice between the two very different video clips. However I'd consider this prompt to be impossible for SORA: "Carly Reilly's next video she proves that SBF is innocent and frees him from prison, produced in the style of Mr Beast"
I'm watching your lips very closely to figure out whether you generated the video in chunks of 60 s and then recorded the audio to match the lip movements…
This technology has the potential to fundamentally change the movie industry. Suppose that the movie you go to see is unique to that particular viewing. So you and your friends will need to go to the theater together.
"Dear Sam: I'm too stupid to invest or set aside any of the millions of dollars I make so if I lose my sweet youtube income I'll starve and it'll be all your fault."
Wait until you can ask the latest Sora to make your next episode by asking it to create your video using the outine you have prepaired. You took that production time from 8 hours to 30 minutes. What do you do with the rest of that time. If you have a production staff, do you need them anymore. Based on the current trajectory, I give it two years.
I know this is slightly off topic, but Elon Musk just sued Open AI, he helped create Open AI, which has gone “Closed AI” to financially benifit the instigators, you will have amazing fertile ground reporting the outcomes. 🥰👍👊
I'm sorry but that mr. Beast tweet was just a funny joke and should not be taken seriously, he was just making a funny joke about progression, not a sign of fear upon AI. Scaring people with "Well it's over, everyone" is just mean. Being scared of something that has never been used for bad things on this earth is just misleading. I don't know how you became so delusional about AI.
The equivalent of the 90s DJ revolution is happening now with Sora. From the homeless to the new rich, as long as one knows how to read and write (that might an obstacle), laziness and lack of talent will not be an obstacle for you to create crappy videos!! Crappy audiences will form around these new AI artists and filmmakers. Christopher Nolan unfortunately is going to miss out on the revolution because he is too old school and too concerned with quality and storytelling. :D
You say it "doesn't understand" stuff, and yet you can compare this to state of the art AI video from just 6 months ago, and it obviously understands it way WAY better than it did then. You can also compare still image generation from this year, compared to a year ago and a year before that. At what point are you going to see where this is going? You call it out for being "one shot" (not allowing edits).... do you seriously think that that will be true a year or two from now? Why wouldn't it? Why won't it start working like ChatGPT works, where you can have a conversation with it and keep tweaking things? You need to extrapolate, not just look at what it does currently. Sure Mr. Beast doesn't have to worry, for one he's already rich and he'll be fine. But this stuff is getting better at maybe five times the speed a grade schooler is getting better. I'm saying this video is going to look mighty dated a year from now.
Thank you for the very nice shout out Carly!!
Confused about why this channel doesn’t have a million followers
Because Carly started with nfts/crypto. Shes grow if she continues her broad topics.
It probably will one day. Then you can look back and say, I followed her, way before she was THE shit
Maybe due to her support for anti vaxxers like Musk, Ackman etc.
Very informative! I believe your closing argument was well stated. Learn and understand this new technology is the best path forward for any individual, company or country.
This is awesome ,as later anyone can direct their own movies with their own stories, values and customs from anywhere in the world. No restrictions and zero gate keeping. Anyone will be an actor or actress or their own movie studio,regardless of geography or connections.
does anyone else see the frame rate tank on her vids lately? Could be something on my end
I think there's an error when they export the file from the editing software. Wrong codex for the camera maybe?
Yes, record or export settings probably wrong
Ah another case of "It's not good enough" and "It won't be better soon" that'll seem utterly out of date and insanely wrong within half a year!
Just a heads up: your audio is peaking severely when you speak at higher volumes and gets very distorted/crunchy. This can be fixed easily by adjusting your recording volume settings (lower the recording volume so that you don't go in the red by testing out how loud you want to speak before you set your recording volume), but I've noticed it on other videos of yours so I thought I'd mention it!
This is just the beginning of A.I. and still in its early stages. Imagine in 10-20 years, we haven't seen nothing yet.
Not sure I agree with your point about striking a pose at the end not being possible.
For the AI to do what it does, must have figured out tracking against time. Therefore it has an internal timeline, by which it fits the video. My suspicion is it does not create the videos like humans would. It knows the end time stamp and the starting, and works forward, backwards, and in-between, to ensure all the objects and motion fit, before it begins to animate.
Sora would have an internal database that tracks all the objects in a scene, which it may dump at the end of the task ( or keep for permanence as in remembering long prompts and relating new prompts to previous
Ai can now observe an image, but visuals are just for humans. I'm sure it focuses on the mathematical construct within its NN. Therefore it will have all the info it will need to impromptu edit ( front, back, and middle).
I'm almost certain everything in the generated world is tied to a key co-oord value, as is currently done in video games, giving access to rotate, yaw etc.
Your videos are great. You do a great job.
I lived 15 years before the internet was mainstream and I saw simple concepts like email initially inundated with FUD by the media and the public in general. I'm sure there's some eventual downside to AI but it won't be 1/10 of what we fear it to be today.
you could mention how worried some people were about y2k, but it turned out to be totally manageable.
Neurotypicals are obsessed with things staying the same and not changing. Just like their own minds, they are stuck in the status quo
I personally look forward to our silicon overlords.
How long have you been studying Mr. Ballen videos for your presentation?
Fair enough ChatGPT has more brainpower than SSSniperwolf (low bar) but good luck to whoever tryna make the Artificial Carly !
It is literally reassembling hundreds of people's tiny bits of works into one.
WHY SHOULD MICROSOFT PROFIT OFF OF A MACHINE THAT LEARNS FROM OTHERS" WORK? Here's a bunch of films based on motion from unreal engine 5. Pay us, not them. How is everything OpenAI does not derivative?
And why the fuck is everything it's focused on ENTERTAINMENT? Like why not AI boring and tedious jobs????
No... That's wrong. It's understanding a deep and wide range of content and making it's own new stuff, like people do...
@@Gerlaffy LOL It sees and perceives nothing on it's own. It takes datasets and based on what it sees it reassembles images from bits of all the work it's looked at. Not in any kind of creative or artistic way at all. It's a literally producing the average of what's in the dataset.
Flood the dataset with trash and you destroy AI's output.
@@Kimberly_SparklesJust wrong. AIs have been given all sorts of content and yet they perform better than you and I in many tasks, they can comprehend and focus on the details you know? How do you think an AI can play Minecraft successfully autonomously without being told what to do or how it works?
@@Kimberly_Sparklesgarbage in garbage out is true broadly speaking but it isn't simply making a collage of previous ideas it is generating new content and was trained/tuned with a given dataset and maybe human reinforcement and or working with external data sources. Saying it just gives you the average is super reductionist and just not even true, it depends on the specific AI model you're talking about and parameters used when promoting and positive vs negative prompts, lots of factors go into what the output is for a given input and a given model. Talking about details of AI systems as though they are all one thing or work in one simple way is just wrong though.
You example of an "impossible" task for SORA is not a good one. OpenAI is using the Vision Transformer research by DeepMind which should be able to do the equivalent of picture infill but with video frames. In much the same way that Dalle and StableDiffusion allows users to delete parts of an image and have the AI fill it in with something plausible or even extrapolate outside of the picture frame, SORA is capable of extrapolating a few frames of video to a full minute clip. So to do your task, just delete the last few seconds and prompt it to have the walking lady do a twirl. You can even load you own video clip in and ask Sora to create a seamless splice between the two very different video clips.
However I'd consider this prompt to be impossible for SORA: "Carly Reilly's next video she proves that SBF is innocent and frees him from prison, produced in the style of Mr Beast"
I’m hard at work on a Carly Oh Really AI
I'm watching your lips very closely to figure out whether you generated the video in chunks of 60 s and then recorded the audio to match the lip movements…
THANKS FOR THE INFO.!!
This technology has the potential to fundamentally change the movie industry. Suppose that the movie you go to see is unique to that particular viewing. So you and your friends will need to go to the theater together.
Kinks aside
Kink aside, cowboy.
I for one welcome our new AI video generator oveerlords.
How come Mr Beast isn’t a Super Chad and is just kinda a passive, nice, wimpy guy?
"Dear Sam: I'm too stupid to invest or set aside any of the millions of dollars I make so if I lose my sweet youtube income I'll starve and it'll be all your fault."
Wait until you can ask the latest Sora to make your next episode by asking it to create your video using the outine you have prepaired. You took that production time from 8 hours to 30 minutes. What do you do with the rest of that time. If you have a production staff, do you need them anymore. Based on the current trajectory, I give it two years.
still not scared. although moderately curious.
😮 so when are you going to be scared when the Terminator army gets here or the Nana drone swarm that will eat the flesh off your very bones😮
The really crazy thing is that this technology is doing the same thing your brain does when you read a book.
WOW
I know this is slightly off topic, but Elon Musk just sued Open AI, he helped create Open AI, which has gone “Closed AI” to financially benifit the instigators, you will have amazing fertile ground reporting the outcomes. 🥰👍👊
Like, comment and subscribe to feed the YT gods
Keep up the good work
What an unfortunate name. Nobody noticed that the groundbreaking, video-generating software shares a name with the Sex Offender Registration Act?
It's basically a massive theft machine. It's not "creating" the footage, it's taking it from wherever, like, this video for example.
Eine ziemlich gruselige Frau
I'm sorry but that mr. Beast tweet was just a funny joke and should not be taken seriously, he was just making a funny joke about progression, not a sign of fear upon AI. Scaring people with "Well it's over, everyone" is just mean. Being scared of something that has never been used for bad things on this earth is just misleading. I don't know how you became so delusional about AI.
Wow ... Sora AI clips of Mr. Beast ripping people off ... in 3-2-1. "Godd#m, the Pusherman". Thanks
Wen opj revival 😢
The equivalent of the 90s DJ revolution is happening now with Sora. From the homeless to the new rich, as long as one knows how to read and write (that might an obstacle), laziness and lack of talent will not be an obstacle for you to create crappy videos!! Crappy audiences will form around these new AI artists and filmmakers. Christopher Nolan unfortunately is going to miss out on the revolution because he is too old school and too concerned with quality and storytelling. :D
You say it "doesn't understand" stuff, and yet you can compare this to state of the art AI video from just 6 months ago, and it obviously understands it way WAY better than it did then. You can also compare still image generation from this year, compared to a year ago and a year before that. At what point are you going to see where this is going? You call it out for being "one shot" (not allowing edits).... do you seriously think that that will be true a year or two from now? Why wouldn't it? Why won't it start working like ChatGPT works, where you can have a conversation with it and keep tweaking things?
You need to extrapolate, not just look at what it does currently.
Sure Mr. Beast doesn't have to worry, for one he's already rich and he'll be fine. But this stuff is getting better at maybe five times the speed a grade schooler is getting better. I'm saying this video is going to look mighty dated a year from now.