Sora is here • Hacker News Board

Re: Sora is here

#811

Post by krainboltgreene » Tue, Dec 10, 2024, 9:34 AM UTC

dumbfounder wrote: post #728

runarberg wrote: post #712

Earlier quoted context omitted.

The Blair Witch Project was a (surprise) creative masterpiece. It worked with very limited technology to create a very clever plot which was paired with an amazing marketing. The combination of which the world hadn’t seen before. It took some creative geniuses to peace the Blair Witch Project together. Generative AI will never produce an experience like that. I know never is a long time, but I’m still gonna call it.…

Why does the idea need to be generated by AI? Let people generate the ideas, the AI will help execute. I think soon (3-5 years) a determined person with no video skills will be able to put together a compelling movie (maybe a short). And that is massive. AI doesn’t have to do everything. Like all tech, it’s a productivity tool.

> Why does the idea need to be generated by AI?

This is the at-first-fun-but-now-frustrating infinite goal move. "AI (a stand in for literally anything) will do (anything) soon." -> "It won't do (thing), it's too complex." -> "Who said AI will do (thing)?"

Re: Sora is here

#812

Post by coffeebeqn » Tue, Dec 10, 2024, 9:35 AM UTC

lelandfe wrote: post #285

pen2l wrote: post #80

Every day that passes I grow fonder of Google's decision to delay or otherwise keep a lot of this under the wraps. The other day I was scrolling down on YouTube shorts and a couple videos invoked an uncanny valley response from me (I think it was a clip of an unrealistically large snake covering some hut) which was somehow fascinating and strange and captivating, and then scrolling down a few more, again I saw someth…

I saw my first AI video that completely fooled commenters: https://imgur.com/a/cbjVKMU This was not marked as AI-generated and commenters were in awe at this fuzzy train, missing the "AIGC" signs. I'm quite nervous for the future.

One of the clearest signs in the current gen is that the typography looks bad still.

Re: Sora is here

#813

Post by coffeebeqn » Tue, Dec 10, 2024, 9:36 AM UTC

krick wrote: post #795

lelandfe wrote: post #285

Earlier quoted context omitted.

I saw my first AI video that completely fooled commenters: https://imgur.com/a/cbjVKMU This was not marked as AI-generated and commenters were in awe at this fuzzy train, missing the "AIGC" signs. I'm quite nervous for the future.

Looks dope though. But what impressed me recently was some crypto-scam video, featuring "a clip" from Lex Fridman Podcast where Elon Musk "reveals" his new crypto or whatever (sadly, the one I saw is currently deleted). It didn't really look good, they were talking with weird pauses and intonations, and as awkward these 2 normally are, here they were even more unnatural. There was so much audacity to it I laughed out…

If someone were to train a model on Joe Rogan podcasts whole run, I’m sure it would spit out extremely impressive fake results already

Re: Sora is here

#814

Post by krainboltgreene » Tue, Dec 10, 2024, 9:38 AM UTC

robotresearcher wrote: post #707

dbspin wrote: post #691

Earlier quoted context omitted.

I think you've misunderstood the objection. Lets pick something concrete. It's a medieval script, it opens with two knights fighting. OK so later in the script we learn their characters, historic counterparts etc. So your LLM can match nefarious villain to some kind of embedding, and doubtless has trained on countless images of a knight. But the result is not naively going to understand the level of reality the scrip…

> but the result is not naively going to understand the level of reality the script is going for… We can already get detailed style guidance into picture generation. Declaring you want Picasso cubist, Warner brothers cartoon, or hyper realistic works today. So does lighting instructions, color palettes, on and on. These future models will not be large language models, they will be multi-modal. Large movie models if y…

This is such an incredibly confident comment. I'm in awe.

Re: Sora is here

#815

Post by lambdaone » Tue, Dec 10, 2024, 9:38 AM UTC

KennyBlanken wrote: post #791

superfrank wrote: post #420

Earlier quoted context omitted.

I know there are people acting like this is obvious that this is AI, but I get why people wouldn't catch it, even if they know that AI is capable of creating a video like this. A) Most of the give aways are pretty subtle and not what viewers are focused on. Sure, if you look closely the fur blends in with the pavement in some places, but I'm not going to spend 5 minutes investigating every video I see for hints of AI…

You can see the perspective/angle of the objects changing slightly as the camera moves in a way that makes it pretty obvious they're CG, AI or otherwise. That's always been a problem with AI generated imagery in video/animation; it changes too much frame to frame. If researchers figure out how to address that, yeah, we've got a problem. Until then - this looks worse tha Then there's the usual giveways for CG - sharpn…

Yes. The lack of diffuse reflection from the pink train is the clearest giveaway, and AI videos in general have problems with getting shadows and radiosity right. There's also the existence of the real-world Hello Kitty Shinkansen and the APM Cat Bus in Japan that makes this image more plausible.

Re: Sora is here

#816

Post by troupo » Tue, Dec 10, 2024, 9:39 AM UTC

robotresearcher wrote: post #707

dbspin wrote: post #691

Earlier quoted context omitted.

I think you've misunderstood the objection. Lets pick something concrete. It's a medieval script, it opens with two knights fighting. OK so later in the script we learn their characters, historic counterparts etc. So your LLM can match nefarious villain to some kind of embedding, and doubtless has trained on countless images of a knight. But the result is not naively going to understand the level of reality the scrip…

> but the result is not naively going to understand the level of reality the script is going for… We can already get detailed style guidance into picture generation. Declaring you want Picasso cubist, Warner brothers cartoon, or hyper realistic works today. So does lighting instructions, color palettes, on and on. These future models will not be large language models, they will be multi-modal. Large movie models if y…

So, we went from "just hand off movie script to automated director/DP/editor" we're now rapidly approaching:

- you have to provide correct detailed instructions on lighting

- you have to provide correct detailed instructions on props

- you have to provide correct detailed instructions on clothing

- you have to provide correct detailed instructions on camera position and movement

- you have to provide correct detailed instructions on blocking

- you have to provide correct detailed instructions on editing

- you have to provide correct detailed instructions on music

- you have to provide correct detailed instructions on sound effects

- you have to provide correct detailed instructions on...

- ...

- repeat that for literally every single scene in the movie (up to 200 in extreme cases)

There's a reason I provided a few links for you to look at. I highly recommend the talk by Annie Atkins. Watch it, then open any movie script, and try to find any of the things she is talking about there (you can find actual movie scripts here: https://imsdb.com)

Re: Sora is here

#817

Post by krainboltgreene » Tue, Dec 10, 2024, 9:39 AM UTC

int_19h wrote: post #472

alpha_squared wrote: post #433

Earlier quoted context omitted.

How would that even work? A dog has physical features (legs, nose, eyes, ears, etc.) that they use to interact with the world around them (ground, tree, grass, sounds, etc.). And each one of those things has physical structures that compose senses (nervous system, optic nerves, etc.). There are layers upon layers of intricate complexity that took eons to develop and a single photo cannot encapsulate that level of com…

A single photo doesn't have to capture all that complexity. It's carried by all those countless dog photos and videos in the training set of the model.

Actually, it does have to capture all of that complexity because it's a photon-based analysis of reality. You cannot take a photo without doing that.

Re: Sora is here

#818

Post by krainboltgreene » Tue, Dec 10, 2024, 9:42 AM UTC

skydhash wrote: post #552

wongarsu wrote: post #256

Earlier quoted context omitted.

And as AI oversaturates the cliched average, creators will have to get further and further away from the average to differentiate themselves. If you pour a lot of work into your creation you want to make it clear that it isn't some cliched AI drivel.

You will basically have to provide a video showcasing your workflow.

I promise you that the artists can outlive the VC money.

Re: Sora is here

#819

Post by logicchains » Tue, Dec 10, 2024, 9:49 AM UTC

phtrivier wrote: post #799

Not available in France yet, I'd we interested to know if it's a matter of progressive rollout, or some form of legislation (EU or otherwise ?) that's making OpenAI cautious ? Something like the EU AI Act [1] ? In a sane world, any video produced by Sora would be required to have a form of watermarking that's on par with what intellectual property owners require. We've put people in jail for sharing copyrighted movie…

>any video produced by Sora would be required to have a form of watermarking that's on par with what intellectual property owners require

It's a completely different thing. IP owners want watermarks on their IP so they can prosecute people who use their IP without giving credit, nobody's forcing them to watermark it.

Re: Sora is here

#820

Post by lmm » Tue, Dec 10, 2024, 9:50 AM UTC

TeMPOraL wrote: post #789

lmm wrote: post #717

Earlier quoted context omitted.

I mean, no? None of the AI-generated images managed to be indistinguishable. Some people were much better than others at spotting the differences. He even quotes, at length, an artist giving a detailed breakdown of what's wrong with one of the images he thought was good.

Did you read the article? Respondents performed barely better than chance. Sure, no one was actually 100% wrong[0]. Just almost always wrong, with a noticeable bias towards liking AI art more . The detailed breakdown you mention? Maybe it's accurate to that artist's thought process, maybe it's more of a rationalization; either way, it's not a general rule they, or anyone, could apply to any of the other AI images. Mo…

Yes, I read the article. Did you?

> The average participant scored 60%, but people who hated AI art scored 64%, professional artists scored 66%, and people who were both professional artists and hated AI art scored 68%.

> The highest score was 98% (49/50), which 5 out of 11,000 people achieved. Even with 11,000 people, getting scores this high by luck alone is near-impossible.