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

Sora Update #1

blog.samaltman.com

Post reply on HN
Print view View original

41–50 of 191 posts

Re: Sora Update #1

#41

Post by lelandfe »

CPLX wrote: post #22
minimaxir wrote: post #2

Yeah, Nintendo called, and faster than expected. > People are generating much more than we expected per user, and a lot of videos are being generated for very small audiences. What did OpenAI expect, really? They imposed no meaningful generation limits and and "very small audiences" is literally the point of an invite-only program.

Indeed. If you read between the lines that’s clearly it. And on that note can I add how much I truly despise sentences like this: > We are hearing from a lot of rightsholders who are very excited for this new kind of "interactive fan fiction" and think this new kind of engagement will accrue a lot of value to them, but want the ability to specify how their characters can be used (including not at all). To me this sen…

Wow, I am sure excited for your new kind of interactive fan fiction of my properties. It will accrue us a lot of value! Anyway, please do not use our properties.

Re: Sora Update #1

#42

Post by notatoad »

techblueberry wrote: post #13
minimaxir wrote: post #2

Yeah, Nintendo called, and faster than expected. > People are generating much more than we expected per user, and a lot of videos are being generated for very small audiences. What did OpenAI expect, really? They imposed no meaningful generation limits and and "very small audiences" is literally the point of an invite-only program.

I don’t understand, what do they mean very small audiences, am I not supposed to make video for myself?

my read: they made the app look like tiktok, and were expecting people to make tiktok style viral videos. instead, what people are making is cameo-style personalized messages for their friends, starring mario.

Re: Sora Update #1

#43

Post by solid_fuel »

I don't understand some parts of this, the writing doesn't seem to flow logically from one thought to another.

    >  Second, we are going to have to somehow make money for video generation. People are generating much more than we expected per user, and a lot of videos are being generated for very small audiences. 
    > We are going to try sharing some of this revenue with rightsholders who want their characters generated by users. 
    > The exact model will take some trial and error to figure out, but we plan to start very soon. Our hope is that the new kind of engagement is even more valuable than the revenue share, but of course we we want both to be valuable.
The first part of this paragraph implies that the video generation service is more expensive than they expected, because users are generating more videos than they expected and sharing them less. The next sentence then references sharing revenue with "rightsholders"? What revenue? The first part makes it sound like there's very little left over after paying for inference.

Secondly, to make a prediction about the future business model - it sounds like large companies (disney, nintendo, etc) will be able to enter revenue sharing agreements with OpenAI where users pay extra to use specific brand characters in their generated videos, and some of that licensing cost will be returned to the "rightsholders". But I bet everyone else - you, me, small youtube celebrities - will be left out in the cold with no controls over their likeness. After all, it's not like they could possibly identify every single living person and tie them to their likeness.

Re: Sora Update #1

#44

Post by martin-t »

CPLX wrote: post #22
minimaxir wrote: post #2

Yeah, Nintendo called, and faster than expected. > People are generating much more than we expected per user, and a lot of videos are being generated for very small audiences. What did OpenAI expect, really? They imposed no meaningful generation limits and and "very small audiences" is literally the point of an invite-only program.

Indeed. If you read between the lines that’s clearly it. And on that note can I add how much I truly despise sentences like this: > We are hearing from a lot of rightsholders who are very excited for this new kind of "interactive fan fiction" and think this new kind of engagement will accrue a lot of value to them, but want the ability to specify how their characters can be used (including not at all). To me this sen…

It feels like big exploitative multimedia companies are the main force fighting big exploitative ML companies over copyright of art.

I wish big exploitative tech companies would fight them over copyright of code but almost all big exploitative tech companies are also big exploitative ML companies.

Oracle to the rescue? What a sick, sad world.

Re: Sora Update #1

#45

Post by saxonww »

CPLX wrote: post #22
minimaxir wrote: post #2

Yeah, Nintendo called, and faster than expected. > People are generating much more than we expected per user, and a lot of videos are being generated for very small audiences. What did OpenAI expect, really? They imposed no meaningful generation limits and and "very small audiences" is literally the point of an invite-only program.

Indeed. If you read between the lines that’s clearly it. And on that note can I add how much I truly despise sentences like this: > We are hearing from a lot of rightsholders who are very excited for this new kind of "interactive fan fiction" and think this new kind of engagement will accrue a lot of value to them, but want the ability to specify how their characters can be used (including not at all). To me this sen…

I'm not really disagreeing with you, but I think it's more about salesmanship than anything else. "We released v1 and copyright holders immediately threatened to sue us, lol" sounds like you didn't think ahead, and also paints copyright holders in a negative light; copyright holders who you need to not be enemies but who, if you're not making it up, are already unhappy enough to want to sue you.

Sam's sentence tries to paint what happened in a positive light, and imagines positive progress as both sides work towards 'yes'.

So I agree that it would be nice if he were more direct, but if he's even capable of that it would be 30 years from now when someone's asking him to reminisce, not mid-hustle. And I'd add that I think this is true of all business executives, it's not necessarily a Silicon Valley thing. They seem to frequently be mealy-mouthed. I think it goes with the position.

Re: Sora Update #1

#46

Post by martin-t »

ronsor wrote: post #38
roxolotl wrote: post #17

Earlier quoted context omitted.

It’s telling to how society values copyright of different media that 4 years into people yelling about these being copyright violation machines the first time there’s been an emergency copyright update has been with video.

The only thing we need is an emergency copyright deprecation.

So people who spend time working on code or art should have exactly zero protection against somebody else just taking their work and using it to make money?

Re: Sora Update #1

#48

Post by martin-t »

kg wrote: post #4

The detail that rightsholders seem to be demanding a revenue share is interesting. That sounds administratively and technologically very complex to implement and probably also just plain expensive to implement.

This is how all work should be rewarded.

Workers getting paid a flat rate while owners are raking in the entire income generated by the work is how the rich get richer faster than any working person can.

Re: Sora Update #1

#49

Post by cg505 »

solid_fuel wrote: post #43

I don't understand some parts of this, the writing doesn't seem to flow logically from one thought to another. > Second, we are going to have to somehow make money for video generation. People are generating much more than we expected per user, and a lot of videos are being generated for very small audiences. > We are going to try sharing some of this revenue with rightsholders who want their characters generated by…

1. They need to charge users for generation.

2. They might get into trouble charging users to generate some other entity's IP, so they may revenue-share with the IP owner.

They're probably still losing money even if they charge for video generation, but recouping some of that cost, even if they revshare, is better than nothing.

Re: Sora Update #1

#50

Post by ineedasername »

timschmidt wrote: post #31
musicale wrote: post #25

Earlier quoted context omitted.

"Hi, as the company that bragged about how we had ripped off Studio Ghibli, and encouraged you to make as many still frames as possible, we would now like to say that you are making too many fake Disney films and we want you to stop."

Cue an open weights model from Qwen or DeepSeek with none of these limitations.

These attempted limitations tend to be very brittle when the material isn’t excised from the training data, even more so when it’s visual rather than just text. It becomes very much like that board game Taboo where the goal is to get people to guess a word without saying a few other highly related words or synonyms.

For example, I had no problem getting the desired results when I promoted Sora for “A street level view of that magical castle in a Florida amusement area, crowds of people walking and a monorail going by on tracks overhead.”

Hint: it wasn’t Universal Studios, and unless you know the place by blind sight you’d think it had been the mouse’s own place.

On pure image generation, I forget which model, one derived from stable diffusion though, there was clearly a trained unweighting of Mickey Mouse such that you couldn’t get him to appear by name, but go at it a little sideways? Even just “Minnie Mouse and her partner”? Poof- guardrails down. If you have a solid intuition of the term “dog whistling” and how it’s done, it all becomes trivial.

Post reply on HN