Did this website kill anyone else’s phone?
Meta Movie Gen
401–410 of 1001 posts
Re: Meta Movie Gen
#402Earlier quoted context omitted.
I don't agree. While "just" audio, I've made a few AI songs that have made people tear up and trigger strong emotions. I think you can do this with video too, just more challenging right now.
I'm sure eventually you can, but I don't think triggering emotions is the correct "KPI" so to speak. On social media platforms, typically the most popular content triggers the strongest emotions. It's rage-bait however, or sadness bait, or any other kind of emotional manipulation. It tricks the human mind and drives up engagement, but I don't think that is indicative of its value. To be clear, I'm certain that's not…
Right now there is a ton of stigma around AI art. That stigma fuels a ton of poorly-informed rhetoric against it. There is also tons and tons of casual use of AI art being shitposted for funsies everywhere that reinforces that rhetoric that AI art means "Push button. Receive crap. Repeat."
Meanwhile, as someone who has been engaged with the AI art community for years, and spent years volunteering part-time as a content moderator for Midjourney, the process of creating art via AI with intentionality is deeply human.
As an MJ mod, I have seeeeeeen things.... It's like browsing though people's psyche. Even in public portfolios people bare their souls because they assume no one will bother to look. People use AI to process the world, their lives, their desires, their trauma. So much of it is straight-up self-directed art therapy. Pages and pages, thousands of images stretching over weeks, sometimes months, of digging into the depths of their selves.
Now go through that process to make something you intend to speak publicly from the depth of your own soul. You don't see much of that day to day because it is difficult. It's risky at a deeply personal level to expose yourself like that.
But, be honest: How much deeply personal art do you see day to day? You see tons of ads and memes. But, to find "real art" you have to explicitly dig for it. Shitposting AI images is as fun and easy as shitposting images from meme generators. So, no surprise you see floods of shitposts everywhere. But, when was the last time you explicitly searched out meaningful AI art?
Re: Meta Movie Gen
#403Why 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…
I'm confused the demo let me press a button and generate a video, was it not supposed to?
Re: Meta Movie Gen
#404Earlier quoted context omitted.
We are at a crossroads of technology, where we're still used to the idea that audio and video are decent proof that something happened, in a way in which we don't generally trust written descriptions of an event. Generative AI will be a significant problem for a while, but this assumption that audio/video is inherently trustable will relatively soon (in the grand scheme of things) go away, and we'll return to the his…
I'd argue it's a step or two more manipulative. Not only do bad actors have the ability to generate moving images which are default believed by many, they also have the ability to measure the response over large populations, which lets them tune for the effect they want. One step more is building response models for target groups so that each can receive tailored distraction/outrage materials targeted to them. Furthe…
> Further, the ability to replicate speech patterns and voice for each of your trusted humans with fabricated material is already commonplace.
Just like the ability to emulate the writing style of your trusted humans was (somewhat) commonplace in the time in which you'd only talk to distant friends over letters.
> Once they are, it's not clear even in person conversations aren't contaminated.
How exactly could any current or even somewhat close technology alter my perception of what someone I'm talking to in-person is saying?
Otherwise, the points about targeting are fair - PR/propaganda has already advanced considerably compared to even 50 years ago, and more personalized propaganda will be a considerable problem, regardless of medium.
Re: Meta Movie Gen
#405Earlier quoted context omitted.
Llama 3 is licensed under "Llama 3 Community License Agreement" which includes restrictions on usage, clearly not "Open Source" as we traditionally know it. Just because pytorch is Open Source doesn't mean everything Meta AI releases is Open Source, not sure how that would make sense. Datasets for Llama 3 is "A new mix of publicly available online data.", not exactly open or even very descriptive. That could be anyth…
Sure, the Llama 3 Community License agreement isn't one of the standard open licenses and sucks that you can't use it for free if you're an entity the size of Google. Here is the Llama source code, you can start training more epochs with it today if you like: https://github.com/meta-llama/llama3/blob/main/llama/model.p... It's rumored Llama 3 used FineWeb, but you're right that they at least haven't been transparent…
> Here is the Llama source code
Correct me if I'm wrong, but that's the code for doing inference?
Meta employee told me just the other day: "However, at the moment-we haven't open sourced the pre-training scripts", can't imagine they would be wrong about it?
https://github.com/meta-llama/llama-recipes/issues/693
> For models I prefer the term "open weight"
Personally, "Open" implies I can download them without signing an agreement with LLama, and I can do whatever I want with it. But I understand the community seems to think otherwise, especially considering the messaging Meta has around Llama, and how little the community is pushing back on it.
So Meta doesn't allow downloading the Llama weights without accepting the terms from them, doesn't allow unrestricted usage of those weights, doesn't share the training scripts nor the training data for creating the model.
The only thing that could be considered "open" would be that I can download the weights after signing the terms. Personally I wouldn't make the case that that's "open" as much as "possible to download", but again, I understand others understand it differently.
Re: Meta Movie Gen
#406Earlier quoted context omitted.
> I don't think you get "The Green Mile" from something like this. But maybe you do get Deadpool & Wolverine 3 Guess where the money is?
If it becomes easy to make "Deadpool & Wolverine", it will no longer be where the money is. Everything that becomes a commodity attracts competition and ceases to be special. (You can see some of that in super-hero movies, which have started to be generic and lost some of their audience.) But, in reality, even making that kind of film is miles away from these examples.
> If it becomes easy to make "Deadpool & Wolverine", it will no longer be where the money is. Everything that becomes a commodity attracts competition and ceases to be special. (You can see some of that in super-hero movies, which have started to be generic and lost some of their audience.)
Well, given the studios still hold the copyright, they can severely constrain supply to keep profits up.
My suspicion is that this kind of stuff gradually reduces some of the labor involved in making films and allows studios to continue padding their margins.
Re: Meta Movie Gen
#407Man, everyone is happy with these advancements, and they are impressive. I’m here looking at users and wondering - the content pipelines are broader, but the exit points of attention and human brains are constant. How the heck are you supposed to know if your content is valid? During a recent apple event, someone on YT had an AI generated video of Tim Cook announcing a crypto collaboration; it had a 100k users before…
It's just something to put ads next to. Selling ad spots is the business, and investors demand an increase even if they already have 3.5 billion pairs of eyeballs. https://www.404media.co/where-facebooks-ai-slop-comes-from/
Re: Meta Movie Gen
#408Earlier quoted context omitted.
Sure, the Llama 3 Community License agreement isn't one of the standard open licenses and sucks that you can't use it for free if you're an entity the size of Google. Here is the Llama source code, you can start training more epochs with it today if you like: https://github.com/meta-llama/llama3/blob/main/llama/model.p... It's rumored Llama 3 used FineWeb, but you're right that they at least haven't been transparent…
> Here is the Llama source code Correct me if I'm wrong, but that's the code for doing inference? Meta employee told me just the other day: "However, at the moment-we haven't open sourced the pre-training scripts", can't imagine they would be wrong about it? https://github.com/meta-llama/llama-recipes/issues/693 > For models I prefer the term "open weight" Personally, "Open" implies I can download them without signin…
The source I linked is the PyTorch model, should be all you need to run some epochs. IDK what the pretraining scripts are.
Re: Meta Movie Gen
#409Impressive. 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!
Yes, and like pretty much every AI release I've seen, even these cherry-picked examples mostly do not quite match the given prompt. The outputs are genuinely incredible, but if you imagine actually trying to use this for work, it would be very frustrating. A few examples from this page: Pumpkin patch - Not sitting on the grass, not wearing a scarf, no rows of pumpkins the way most people would imagine. Sloth - that's…
The one thing that immediately stood out to me in the ghost example was how the face of the ghost had "wobbly geometry" and didn't appear physically coupled to the sheet. This and the way the fruit in the sloth's drink magically rested on top of the drink without being wedged onto the edge of the glass as that would require were actually some of the more immediate "this isn't real" moments for me.
Re: Meta Movie Gen
#410Man, everyone is happy with these advancements, and they are impressive. I’m here looking at users and wondering - the content pipelines are broader, but the exit points of attention and human brains are constant. How the heck are you supposed to know if your content is valid? During a recent apple event, someone on YT had an AI generated video of Tim Cook announcing a crypto collaboration; it had a 100k users before…
This will just lead to people not taking videos as evidence anymore. Just like images of war crime aren’t irrefutable evidence due to staging and photoshop, videos will lose their worth as evidence. Which is actually a good thing in some instances. If someone blackmails you with nudes/explicit videos, you can just ignore it and claim it’s fake.