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

Ben Affleck's surprisingly comprehensive take on LLMs for video

twitter.com

Post reply on HN
Print view View original

61–70 of 187 posts

Re: Ben Affleck's surprisingly comprehensive take on LLMs for video

#61

Post by spiderice »

impomura wrote: post #45
spiderice wrote: post #28

Earlier quoted context omitted.

> Why should I bother to read something you didn't bother to write? The answer seems obvious: because it's better. Obviously it isn't better now . But it's easy to imagine a day in the not-too-distant future where AI can shit out 1000 hours of content that is better in every way than what humans create. And it will even feel more human than the human made stuff because the AI will have learned that we like that. What…

it's impossible to answer to this line of reasoning without wasting time so I'll just start right away with the ad hominem. you just don't like art, you don't understand it and you want slop, admit it and don't feel compelled to enter the discussion with your growth oriented bullshit mindset

You're conflating the reality of the situation with me. I didn't say I wanted AI generated content. Just that it seems like it will inevitably win. All the insults in your comment just stem from an imaginary and inaccurate picture of me, a stranger, that you created in your head.

> don't feel compelled to enter the discussion with your growth oriented bullshit mindset

Then why respond?

Re: Ben Affleck's surprisingly comprehensive take on LLMs for video

#62

Post by ricoxicano »

moralestapia wrote: post #37
aeturnum wrote: post #18

Earlier quoted context omitted.

You're mis-reading what he says. People create new stuff, AI (so far) remixes old stuff. Down the line you will be able to order a remix of succession and AI should be able to do that for you but it won't be able to add any interesting additional spin - it's just gonna make it to order (obviously a big deal! which he says). >implying the non-AI content we have nowadays is highly creative and unique You just inserted…

??? Show me a movie in the last decade where the premise of it couldn't be generated by some AI prompt. (I'll even make it easier by stating that said AI prompt can be generic and to some degree trivial; i.e. not the premise explicitly laid out in the prompt).

???

show me an AI prompt that can generate a movie premise in the last decade where the underlying model hasn't already been trained on that movie's plot.

Re: Ben Affleck's surprisingly comprehensive take on LLMs for video

#63

Post by andrewinardeer »

spiderice wrote: post #28
bdndndndbve wrote: post #12

My favorite quote on this topic is "Why should I bother to read something you didn't bother to write?" Especially when you're talking about fiction and reading/watching for enjoyment, what does it matter if you can shit out 1000 hours of AI content? Maybe it's good to keep babies entertained? Studios have gotten so into the habit of treating "content" as a fungible commodity, but the fact is that even blockbuster mov…

> Why should I bother to read something you didn't bother to write? The answer seems obvious: because it's better. Obviously it isn't better now . But it's easy to imagine a day in the not-too-distant future where AI can shit out 1000 hours of content that is better in every way than what humans create. And it will even feel more human than the human made stuff because the AI will have learned that we like that. What…

Mechanizing the expression of artist endeavour seems silly. Does an LLM know the pleasure and pain that love can instill or does it just regurgitate tokens in a pattern it thinks is best fit?

Re: Ben Affleck's surprisingly comprehensive take on LLMs for video

#64

Post by mempko »

nwgo wrote: post #19

He is wrong. Movies will be automated and when they will be, all these actors will be first one to go. I say this despite him and Matt Damon my favorite actors. When it pours it rains. They don't know what is in store and what we are working on. Hint: codename: Project Hailey (named after Ho--y)

My prediction. As more and more movies are made with AI, and when they start replacing real actors with AI, people are going to value movies less, and live performances more.

I expect AI will bring a resurgence of live performances like live theater. Why? Because we value what people do more than what machines do.

Re: Ben Affleck's surprisingly comprehensive take on LLMs for video

#65

Post by dmortin »

Maybe the main actors, the stars of the movies are safe for now. But background actors who just sit in the background or have some small extra role will be decimated, because those roles can be replaced with AI.

Since many actors start with small roles or background work, it will have an effect on the industry if entry jobs are eliminated.

In the long run even the main stars can get into trouble, because it may be true that AI won't be able to reproduce what they do, creating a new, creative scene, but most movies are trash even today, so if AI can make an average, good enough movie in the future which is cheap to generate, but people still watch it then the bottom of the industry will fall out, lots of people who work on movies will lose their jobs which can have a huge impact on the industry.

Re: Ben Affleck's surprisingly comprehensive take on LLMs for video

#66

Post by xenospn »

moralestapia wrote: post #16
moogly wrote: post #11

Earlier quoted context omitted.

> He even contradicts his whole premise of "actors will always be necessary" with his example of a custom generated series episode. He said it would be "janky and weird" and likened it to a remix.

People who think generating language is very much different from generating audio/video are in for a ride, which I think is the flaw with his reasoning. (Also, yeah, there's a very strong interest for him in saying actors will not be replaced, but let's assume he's being frank here). All the "AI personalities" that are already popping up here and there are already decent enough, again, it will only take a couple year…

I have a strong suspicion that people who think AI-generated characters are "similar" to people are borderline autistic and simply can't grasp nuance.

Re: Ben Affleck's surprisingly comprehensive take on LLMs for video

#67

Post by no_wizard »

moralestapia wrote: post #13
sottol wrote: post #9

Earlier quoted context omitted.

How soon though? 2 years? 10? 50? He starts by saying this won't happen soon, _maybe_ ever.

I'd bet that in two years you could do something equivalent to that "make me a new episode" use case. I already see really convincing deep fakes on Facebook, etc., it's not an impossible problem anymore, it's just a matter of increasing their quality.

Is there any deep fake video that is possibly convincing that is 100% generated synthetically?

All the deep fakes I've seen that don't look incredibly weird at first glance essentially re-use clips (like Back To The Future[0]) and they replacement an element, like a face, but the rest of the footage already exists.

If you look closely though, particularly at the mouth of the video I linked, something isn't quite natural about it still.

Same deal with this Taylor Swift fake[1]

[0]: https://www.youtube.com/watch?v=8OJnkJqkyio

[1]: https://x.com/McAfee/status/1745226438641602866

Re: Ben Affleck's surprisingly comprehensive take on LLMs for video

#68

Post by xenospn »

moralestapia wrote: post #13
sottol wrote: post #9

Earlier quoted context omitted.

How soon though? 2 years? 10? 50? He starts by saying this won't happen soon, _maybe_ ever.

I'd bet that in two years you could do something equivalent to that "make me a new episode" use case. I already see really convincing deep fakes on Facebook, etc., it's not an impossible problem anymore, it's just a matter of increasing their quality.

Getting to 90% is easy. The last 10% can take 50 years. Don't forget we're the product of evolution and subconsciously look for things current technology will not be able to replicate.

Re: Ben Affleck's surprisingly comprehensive take on LLMs for video

#70

Post by stego-tech »

impomura wrote: post #45
spiderice wrote: post #28

Earlier quoted context omitted.

> Why should I bother to read something you didn't bother to write? The answer seems obvious: because it's better. Obviously it isn't better now . But it's easy to imagine a day in the not-too-distant future where AI can shit out 1000 hours of content that is better in every way than what humans create. And it will even feel more human than the human made stuff because the AI will have learned that we like that. What…

it's impossible to answer to this line of reasoning without wasting time so I'll just start right away with the ad hominem. you just don't like art, you don't understand it and you want slop, admit it and don't feel compelled to enter the discussion with your growth oriented bullshit mindset

Literally this. Ben hits the nail on the head that these tools can “write convincing Elizabethan language but can’t write Shakespeare”, along with his metaphor about craftsmen vs artists.

These tools can never create art because art is the imperfection of reality transposed from the mind’s eye using the talent of the artisan and their tools. Writing a convincing enough prompt to generate an assortment of visual outputs that you “choose” as the final product can never be art, because your art skills ended with the prompt itself - everything after was just maths, and not even maths you had a direct hand in. Even then, you cannot really shill your prompt as art either, because you wrote tokens to ingest into a LLM to generate pseudorandom visual outputs, not language to be interpreted by other humans and visualized on their own accord.

Art is one of those things you cannot appreciate until you make it, and generating slop is not creating art. A preschooler with a single, broken crayon and a napkin makes better art than anything generated via tokens and math models - and to really drive that home, I’d argue that the teenager goofing around with math formulas on their graphing calculator to create visually beautiful or interesting designs is also superior art than whatever the LLM can spew forth using far more advanced maths.

If you really want art, then make it. Learn to draw, practice photography, paint some scenery, experiment with formula visualizations, layout a garden, or heck, just commission an artist to bring your idea into reality. Learning to articulate your vision with language in such a way others can illustrate or create it is a far more valuable skill than laying out tokens for an LLM.

Post reply on HN