Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Meta Movie Gen

ai.meta.com

Post reply on HN
Print view View original

391–400 of 1001 posts

Re: Meta Movie Gen

#391

Post by kurthr »

tsimionescu wrote: post #374
intended wrote: post #332

Man, 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…

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. Further, the ability to replicate speech patterns and voice for each of your trusted humans with fabricated material is already commonplace.

True endstage adtech will require attention modeling of individuals so that you can predict target response before presenting optimized material.

It's not just a step back, it's a step into black. Each person has to maintain an encrypted web of trust and hope nobody in their trust ring is compromised. Once they are, it's not clear even in person conversations aren't contaminated.

Re: Meta Movie Gen

#392

Post by bakugo »

sensanaty wrote: post #375
cs702 wrote: post #35

Impressive. It's only going to get better, faster, cheaper, easier.[a] Sooner than anyone could have expected, we'll be able to ask the machines: "Turn this book into a two-hour movie with the likeness of [your favorite actor/actress] in the lead role." Sooner than anyone could have expected, we'll be able to have immersive VR experiences that are crafted to each person. Sooner than anyone could have expected, we won…

Does anyone other than PMs when thinking up user stories do shit like this or finds this kinda stuff desirable? It just sounds like a business person who doesn't have a life other than selling their product trying to think up "real user" usecases every time.

Nope. This is just like when cryptobros would regularly insist that cryptocurrencies would replace banks by the end of the decade. It's safe to assume that anyone who makes such wild predictions is a bagholder who stands to gain financially from said wild predictions coming true, even though they never will.

Re: Meta Movie Gen

#393

Post by genewitch »

denisvlr wrote: post #291
adamc wrote: post #61

Earlier quoted context omitted.

I wouldn't be. How is any of this going to lead to meaningful art? I don't think you get "The Green Mile" from something like this.

Shameless plug: I just created a short AI film (1) and tried to tell an actual story and trigger emotions. I spent countless hours crafting the script, choosing the shots, refining my prompts, generating images, animating images, generating music, sounds, and so on... For me AI tools are just that - tools. True, you have to yield them some "control", but at the end of the day you are still the one guiding and directi…

I'm in agreement with scudsworth here, but i have a little more nuance, i think. I know how long this took, and how many compromises were made. The only reason this works, at all, is because humans have a massive list of cultural memes and tropes that shorthand "experience" for us. it has the "AI can only generate 2-5 seconds of video before it goes completely off the rails" vibe; which allows it to fit in with the ADHD nature of most video production of the last 30 years - something a lot of people do not like. For an example where this was jarring in the video, when they're drawing or painting, you cut the scene slightly too late, you can see the AI was about to do some wild nonsense.

What happened to the mom? Why does the kid get older and younger looking? why does the city flicker in the beginning? which kid is his in the ballet performance? why do they randomly have "lazy eye"? i could keep going but i think we all get my point.

I can intuit the tropes used by the AI to convey meaning, and i'd be willing to list them all with relevant links for the paltry sum of $50. Be warned, it will be a very large list. Tropes and "memes" are doing 100% of the heavy lifting of this "art".

Sorry, human. As someone who stopped creating art on a daily basis due to market dilution (read: it's too hard to build a fanbase that i care about), i am very critical of most "art" produced anyhow.

this is dogshit.

Re: Meta Movie Gen

#394

Post by ein0p »

rolisz wrote: post #367
csomar wrote: post #362

Earlier quoted context omitted.

The solution to that is to make models both open weight and open source. That will equalize the level playing field.

How will that help? How will uncle Joe be able to tell fake videos better with an open source model?

Uncle Joe will just stop assuming that just because there’s a video it is real. That hasn’t been the case for decades. About time uncle Joe caught on.

Re: Meta Movie Gen

#395

Post by benabbott »

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.

So... smash the looms..?

Re: Meta Movie Gen

#396

Post by genewitch »

scudsworth wrote: post #349
denisvlr wrote: post #291

Earlier quoted context omitted.

Shameless plug: I just created a short AI film (1) and tried to tell an actual story and trigger emotions. I spent countless hours crafting the script, choosing the shots, refining my prompts, generating images, animating images, generating music, sounds, and so on... For me AI tools are just that - tools. True, you have to yield them some "control", but at the end of the day you are still the one guiding and directi…

a shameless plug deserves an honest review: this is dog shit

I, uh, gave some more nuance because i had some free time as a sibling comment to this. I hope we don't get downvoted because someone has to call a spade a spade.

Re: Meta Movie Gen

#397

Post by tsimionescu »

dguest wrote: post #387
tsimionescu wrote: post #374

Earlier 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…

It's not all the way back as long as solid encryption exists: Tim Cook could digitally sign his announcements, and assuming we can establish his signature (we had signatures and stamps 200 years ago) video proof still works. So we're not going all the way back, but the era of believing strangers because they have photographic or video proof is drawing to a close.

Cryptography is nice here, but the base idea remains the same: you need to trust the person publishing the video to believe the video. Cryptography doesn't help for most interesting cases here, though it can help with another level, that of impersonation.

Sure, Tim Cook can sign a video so I know he is the one who published it - though watching it on https://apple.com does more or less the same thing. But if the video is showing some rockets hitting an air base, the cryptography doesn't do anything to tell you if these were real rockets or its an AI-generated video. It's your trust in Tim Cook (or lack thereof) that determines if you believe the video or not.

Re: Meta Movie Gen

#398

Post by bee_rider »

Mistletoe wrote: post #213
lairv wrote: post #80

Earlier quoted context omitted.

I'm not sure, I see a common pattern with autonomous vehicle, text to image, llms: the last 10% are hard to achieve

If we judge from AI writing, we can extrapolate what an AI movie would be like. I cannot imagine reading an AI book. It would look and smell like a book but nothing of value or new insight would be inside. Michael Bay might be very interested.

Yes, AI will probably fail miserably for a while at least, at making the sort of well written artistic, clever movie that nobody watches. The only ones that need to be worried are the studios making churning out massive blockbusters…

Re: Meta Movie Gen

#399

Post by bitbasher »

If social media was the scourge of the last decade, the next decade's scourge will be artificial content.

Digital minimalism is looking more and more attractive.

Re: Meta Movie Gen

#400

Post by bee_rider »

lairv wrote: post #80
mywacaday wrote: post #54

Earlier quoted context omitted.

Are we only a few years away from one person/small group made movies where the dialog, acting, location and special effects can be tweaked endlessly for a relatively low cost. If I was a studio exec I'd be worried.

I'm not sure, I see a common pattern with autonomous vehicle, text to image, llms: the last 10% are hard to achieve

A 90% approximation of what somebody wants might be more interesting to that person, than a 99% approximation of what some studio exec wants.
Post reply on HN