I find the edit video with text the most fascinating aspect. I can see this being used for indie films that doesn’t have a CGI budget. Like the scene with the movie theater, you can film them on lounge chairs first and then edit it to seem like a movie theater.
These are not movies, these are clips. The stock photo/clip industry is surely worried about this, and probably will sue because 100% these models were trained on their work. If this technology ever makes movies, it'll be exactly like all the texts, images and music these models create: an average of everything ever created, so incredibly mediocre.
I imagine a movie maker where you say "use model A and put them in scene 32f, add a crowd and zoom in on A. They should look very worried." Then they can just play with it. Then save a scene, onto the next. Since AI can continue an animation, I don't see why it can't faithfully recreate given models with more development
What'll happen in both industries is the same that will happen everywhere else: adopt or die. The huge winners will be those that creatively use this new tool without 100% relying on it to do everything.
There have been several AI short film festivals, as well as several AI music videos that have been produced. The caveats are that quality varies, the best ones simply employ solid production in general (good editing, strong directorial vision, etc), and I don't know that anything feature length is out or even in the works.
the problem is that these stock footage companies are up against the richest corporations to ever exist. Legal recourse will take a monumental amount of money and time.
Hate to say this, but as things stand, tech companies stand to become all pervasive and all powerful if AI keeps growing the way it has
I'm seeing weird bank on a Pixel 6a / chromium browser as well. I'm on mobile so I can't check the source, but this can't just be static HTML.
When I scroll the page, sections of text are missing then pop in, randomly though not as a scroll driven animation. It almost feels like something is blocking the browser's render loop and it can't catch up to actually paint the text. That'd be an insane bug on such a simple page, though I put nothing past react these days if they used it here.
We humans are so excessively dependent on vision input and with entertaining through visuals, too. But more and more all those visuals become meaningless to me and it all just feels like fast-food-junk to me.
As any pre-schooler will be able to produce anything (watch out parents) imaginable in seconds doesn't make it better to me or is of any real value.
Ok, i needed to edit it again to add: maybe this IS the value of it. We can totally forget about phantasizing stories with visuals (movies) because nobody will care anymore.
Yeah I agree, I never understood the appeal of photography, it's so easy, you don't need to paint for hours to produce something original, you just need to buy a camera and click on a button. That's it. And people pay for that, I don't get it.
I've been saying for years that generated content is an impending tsunami that's going to drown out all real human voices online. The internet may become effectively unusable as a result for anything other than entertainment.
This is interesting and i see some of this now. Even here on HN and other forums i thought were mostly "human". Even one of my group chats i can tell one of my friends is using ai responses, but one of the other members cant tell and replies earnestly.
I am grossed out by this. my instinct is to avoid ai slop. The interesting part to me is: What next? Where do we go? Will it be that "human" forums are pushed further into obscurity of the internet? Or will go so far as that we all start preferring meeting in person? Im clueless here
Maybe that's a good thing. The internet never reached its potential as being the connective fabric of humanity. Mostly it's just marketing and spam. If the internet died and we all went back to smaller communities, that really wouldn't be the worst thing IMO. We're not really evolved for global communications at scale anyway.
The Internet used to be a sort of hideaway for nerdy people to hang out and have fun. Ever since the invention of the smartphone, possibly before (see “Eternal September”) it’s gone to shit. These days I would rather spend time offline.
Are there any other Internet-based hideaways to retreat to? Somewhere where ads, clout chasing, and AI slop doesn’t exist?
All the vids have that instantly recognizable GenAI "sheen", for the lack of a better word. Also, I think the most obvious giveaway are all the micro-variations that happen along the edges, which give a fuzzy artifact.
I assure you that's not enough. These are high quality videos. Once they get uploaded to social media, compression mostly makes imperfections go away. And it's been shown that when people are not expecting AI content, they are much less likely to realize they are looking at AI. I would 100% believe most of these videos were real if caught off guard.
That sheen looks (to me) like some of the filters that are used by people who copy videos from TV and movie and post them on (for example) facebook reels.
There's an entire pattern of reels that are basically just ripped-off-content with enough added noise to (I presume) avoid content detection filters. Then the comments have links to scam sites (but are labelled as "the IMDB page for this content").
At least all the humans in these videos seem to have the correct number of fingers, so that's progress. And Moo Deng seems to have a natural sheen for some reason so can't hold that against them. But your point about the edges is still a major issue.
I wonder how much RLHF or other human tweaking of the models contributes to this sort of overstauration / excess contrast in the first place. The average consumer seems to prefer such features when comparing images/video, and use it as a heuristic for quality. And there have been some text-to-image comparisons of older gen models to newer gen, purporting that the older, more hands-off models didn't skew towards kitschy and overblown output the way newer ones do.
So I’m probably going to be too closed minded about this: but who the f*ck asked for this and did anyone consider consequences of easily accessible AI slop generation?
It’s already nearly impossible to find quality content on the internet if you don’t know where to look at.
My kids both have creative hearts, and they are terrified that A.I. will prevent them from earning a living through creativity. Very recently, I've had an alternate thought. We've spent decades improving the technology of entertainment, spending billions (trillions?) of dollars in the process. When A.I. can generate any entertainment you can imagine, we might start finding this kind of entertainment boring. Maybe, at that point, we decide that exploring space, stretching our knowledge of physics and chemistry, and combating disease are far more interesting because they are real. And, through the same lens, maybe human-created art is more interesting because it is real.
Why do these video generation ones never become usable to the public. Is it just they had to create millions of videos and cherry pick only a handful of decent generations? Or is it just so expensive there's no business model for it?
My mind instantly assumes it a money thing and they're just wanting to charge millions for it, therefore out of reach for the general public. But then with Meta's whole stance on open ai models, that doesn't seem to ring true.
A lot of folks in this thread have mentioned that the problem with the current generation of models is that only 1 in (?) prompts returns something useful. Isn't that exactly what a reward model is supposed to help improve? I'm not an ML person by any means so the entire concept of reward models feels like creating something from nothing, so very curious to understand more.
Off topic but some day you could live off grid with your own solar fusion mini reactor powering your own hardware that enables creating your own stories, movies and tales. No more need of streaming services. Internet would be to obtain news, goods and buy greatest and latest (or not) data to update your models.
Decentralization could be for once not as painful as it is now; however, I still believe every single hardware vendor would try to hook to the internet and make you install an app. Looking forward to this AI revolution for sure.
Always important to bear in mind that the examples they show are likely the best examples they were able to produce.
Many times over the past few years a new AI release has "wowed" me, but none of them resulted in any sudden overnight changes to the world as we know it.
VFX artists: You can sleep well tonight, just keep an eye on things!
Are any image / video generation tools giving just the output or the layers, timelines, transitions, audio as things to work with in our old fashioned toolsets?
The problem:
In my limited playing of these tools they don't quite make the mark and I would easily be able to tweak something if I had all the layers used. I imagine in the future products could be used to tweak this to match what I think the output should be....
At least the code generation tools are providing source code. Imagine them only giving compiled bytecode.
Photo sharing websites (including Facebook) used to be wrappers around ImageMagick with extra features. I love how the backbone of their training involves calling out to ffmpeg. It gives a little hope to those of us who, too, are working on a smaller scale but with similar techniques.
Scale? I have access to an H100. Meta trained their cat video stuff on six thousand H100s.
They mention that these consume 700W each. Do they pay domestic rates for power? Is that really only $500 per hour of electricity?
Things are about to get weird. We can't control this at any level:
At the level of image/video synthesis: Some leading companies have suggested they put watermarks in the content they create. Nice thought, but open source will always be an option, and people will always be able to build un-watermarked tools.
At the level of law: You could attempt to pass a law banning image/video generation entirely, or those without watermarks, but same issue as before– you can't stop someone from building this tech in their garage with open-source software.
At the level of social media platforms: If you know how GANs work, you already know this isn't possible. Half of image generation AI is an AI image detector itself. The detectors will always be just about as good as the generators- that's how the generators are able to improve themselves. It is, I will not mince words, IMPOSSIBLE to build an AI detector that works longterm. Because as soon as you have a great AI content classifier, it's used to make a better generator that outsmarts the classifier.
Seriously though. This is the company that is betting hard on VR goggles. And these are engines that can produce real time dreams, 3d, photographic quality, obedient to our commands. No 3d models needed, no physics simulations, no ray tracing, no prebuilt environments and avatars. All simply dreamed up in real time, as requested by the user in natural language. It might be one of the most addictive technologies ever invented.
[+] [-] syntaxing|1 year ago|reply
[+] [-] deng|1 year ago|reply
[+] [-] sethammons|1 year ago|reply
[+] [-] DirkH|1 year ago|reply
[+] [-] jkolio|1 year ago|reply
[+] [-] spaceman_2020|1 year ago|reply
Hate to say this, but as things stand, tech companies stand to become all pervasive and all powerful if AI keeps growing the way it has
[+] [-] Aeolun|1 year ago|reply
The video’s look cool, but I can’t really enjoy reading about them if my phone freezes every 2 seconds.
[+] [-] _heimdall|1 year ago|reply
When I scroll the page, sections of text are missing then pop in, randomly though not as a scroll driven animation. It almost feels like something is blocking the browser's render loop and it can't catch up to actually paint the text. That'd be an insane bug on such a simple page, though I put nothing past react these days if they used it here.
[+] [-] Kailhus|1 year ago|reply
[+] [-] rudasn|1 year ago|reply
[+] [-] hnben|1 year ago|reply
[+] [-] runeks|1 year ago|reply
[+] [-] kosolam|1 year ago|reply
[+] [-] fasa99|1 year ago|reply
A: because your phone is a potato
The first free tech support from HN is free, subsequent questions will be $29.99 per.
[+] [-] reneberlin|1 year ago|reply
As any pre-schooler will be able to produce anything (watch out parents) imaginable in seconds doesn't make it better to me or is of any real value.
Ok, i needed to edit it again to add: maybe this IS the value of it. We can totally forget about phantasizing stories with visuals (movies) because nobody will care anymore.
[+] [-] ShrigmaMale|1 year ago|reply
[+] [-] baby|1 year ago|reply
[+] [-] heurist|1 year ago|reply
[+] [-] boogieknite|1 year ago|reply
I am grossed out by this. my instinct is to avoid ai slop. The interesting part to me is: What next? Where do we go? Will it be that "human" forums are pushed further into obscurity of the internet? Or will go so far as that we all start preferring meeting in person? Im clueless here
[+] [-] solardev|1 year ago|reply
[+] [-] nl|1 year ago|reply
Have you seen what most humans say? If an AI says more intelligent things I'm all for it.
[+] [-] shortrounddev2|1 year ago|reply
[+] [-] cedws|1 year ago|reply
Are there any other Internet-based hideaways to retreat to? Somewhere where ads, clout chasing, and AI slop doesn’t exist?
[+] [-] TiredOfLife|1 year ago|reply
That already happened, even without AI.
[+] [-] chpatrick|1 year ago|reply
[+] [-] Andrex|1 year ago|reply
[+] [-] skywhopper|1 year ago|reply
[+] [-] TrackerFF|1 year ago|reply
[+] [-] lopis|1 year ago|reply
[+] [-] dekhn|1 year ago|reply
There's an entire pattern of reels that are basically just ripped-off-content with enough added noise to (I presume) avoid content detection filters. Then the comments have links to scam sites (but are labelled as "the IMDB page for this content").
[+] [-] newaccount74|1 year ago|reply
[+] [-] DebtDeflation|1 year ago|reply
[+] [-] blargey|1 year ago|reply
[+] [-] Rinzler89|1 year ago|reply
That's something that can be fixed in a future release or you can fix it right now with some filters in post in your pipeline.
[+] [-] alana314|1 year ago|reply
[+] [-] wiseowise|1 year ago|reply
It’s already nearly impossible to find quality content on the internet if you don’t know where to look at.
[+] [-] nthdesign|1 year ago|reply
[+] [-] lwansbrough|1 year ago|reply
[+] [-] Animats|1 year ago|reply
- Every script in Hollywood will now be submitted with a previs movie.
- Manga to anime converters.
- Online commercials for far more products.
[+] [-] turblety|1 year ago|reply
My mind instantly assumes it a money thing and they're just wanting to charge millions for it, therefore out of reach for the general public. But then with Meta's whole stance on open ai models, that doesn't seem to ring true.
[+] [-] tqi|1 year ago|reply
[+] [-] clvx|1 year ago|reply
[+] [-] sourraspberry|1 year ago|reply
Always important to bear in mind that the examples they show are likely the best examples they were able to produce.
Many times over the past few years a new AI release has "wowed" me, but none of them resulted in any sudden overnight changes to the world as we know it.
VFX artists: You can sleep well tonight, just keep an eye on things!
[+] [-] thinkingemote|1 year ago|reply
The problem: In my limited playing of these tools they don't quite make the mark and I would easily be able to tweak something if I had all the layers used. I imagine in the future products could be used to tweak this to match what I think the output should be....
At least the code generation tools are providing source code. Imagine them only giving compiled bytecode.
[+] [-] gorgoiler|1 year ago|reply
Scale? I have access to an H100. Meta trained their cat video stuff on six thousand H100s.
They mention that these consume 700W each. Do they pay domestic rates for power? Is that really only $500 per hour of electricity?
[+] [-] benabbott|1 year ago|reply
At the level of image/video synthesis: Some leading companies have suggested they put watermarks in the content they create. Nice thought, but open source will always be an option, and people will always be able to build un-watermarked tools.
At the level of law: You could attempt to pass a law banning image/video generation entirely, or those without watermarks, but same issue as before– you can't stop someone from building this tech in their garage with open-source software.
At the level of social media platforms: If you know how GANs work, you already know this isn't possible. Half of image generation AI is an AI image detector itself. The detectors will always be just about as good as the generators- that's how the generators are able to improve themselves. It is, I will not mince words, IMPOSSIBLE to build an AI detector that works longterm. Because as soon as you have a great AI content classifier, it's used to make a better generator that outsmarts the classifier.
So... smash the looms..?
[+] [-] msp26|1 year ago|reply
It being 30B gives me hope.
[+] [-] throw310822|1 year ago|reply
Seriously though. This is the company that is betting hard on VR goggles. And these are engines that can produce real time dreams, 3d, photographic quality, obedient to our commands. No 3d models needed, no physics simulations, no ray tracing, no prebuilt environments and avatars. All simply dreamed up in real time, as requested by the user in natural language. It might be one of the most addictive technologies ever invented.
[+] [-] bitbasher|1 year ago|reply
Digital minimalism is looking more and more attractive.