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

81–90 of 191 posts

Re: Sora Update #1

#81

Post by measurablefunc »

noduerme wrote: post #66
piskov wrote: post #19

Broke: cure cancer, new physics, agi, take your jobs, what have you. Please give us a trillion. Woke: AI slop tictoc to waste millions of human-hours.

You make a good point. They may well as admit at this point that curing cancer, new physics, and AGI aren't going to happen very soon. What surprises me a bit is that they'd take this TikTok route, rather than selling Sora as a very expensive storyboarding tool to film/tv studios, producers, etc. Why package it as an app for your neice to make viral videos that's bound to lose money with every click? Just sell it for…

The query data they are collecting can be used for ad targeting. Remember, if you're not paying for it (and in many cases even when you are paying for it) then the data collected from your use of the application is going to be used by someone to make money one way or another. Google made billions from search queries & OpenAI has an even better query/profiling perspective on its users b/c they are providing all sorts of different modalities for interaction, that data is extremely valuable, analogous to how Google search queries (along w/ data from their other products) are extremely valuable to corporate marketing departments that are willing to pay a premium for access to Google's targeting algorithms.

Re: Sora Update #1

#82

Post by JimDabell »

noduerme wrote: post #79
JimDabell wrote: post #78

Earlier quoted context omitted.

The vast amounts of permissively licensed works directly contradicts you. Even if you take away copyright, there are plenty of incentives to create. Copyright is not the sole reason people create.

Vague. Are you talking about reasons to create like the joy of creating? Your bio describes you as a 'tech entrepreneur', not 'DIY tinkerer'. So I'll assume that when you spend a great deal of time entrepreneuring something, you do so with the hope of remuneration. Maybe not by licensing the copyright, but in some form. Permissive licenses are great in software, where SAAS is an alternative route to getting paid. How…

> Vague. Are you talking about reasons to create like the joy of creating?

That’s one of them, but I really don’t have to be specific about the reasons. I just have to point out the existence of permissively licensed works. You said:

> Great, you've just removed any incentive for people to make anything.

This is very obviously untrue. Perhaps you meant to say “…you’ve just removed some incentives for people to make some things”?

Re: Sora Update #1

#83

Post by ojosilva »

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.

And I don't think you can revenue share these generations with rights owners just like that. What rights owner will let their "product" be depicted in any imaginable situation by any prompt by anyone in the planet? Words are powerful and images a 1000 words worth, videos are a millionth fold... I've seen a quick Sora video from OpenAI themselves I believe of the real life Mario Bros Princess, a rather voluptuous one, playing herself on a console and the image stuck. And it's not just misuse, distortion or appropriation but also association: imagine a series of very viral videos of Pikachu drinking Coke or a fan series of Goku with friends at KFC... it could condition, or steal, future marketing deals for the rights holders.

This is a non-starter, unless you own a "license to AI" from the rights owner directly, such as an ad agency that uses Sora to generate an ad it was hired to do.

Re: Sora Update #1

#84

Post by vntok »

lelandfe wrote: post #41
CPLX wrote: post #22

Earlier quoted context omitted.

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.

Nice but there's no need for the "please": it's not a request, it's a demand from an official lawyer-penned, strongly-worded, lawsuit actionable letter.

Re: Sora Update #1

#85

Post by 48terry »

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…

> the writing doesn't seem to flow logically from one thought to another.

Neither has most of the stuff Sam has said since basically the moment he started talking.

It is possible, perhaps, that he is actually a very stupid person!

Re: Sora Update #1

#86

Post by id00 »

signatoremo wrote: post #60
CPLX wrote: post #22

Earlier quoted context omitted.

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…

You may not like their message, but the style can be found in practically any public communication from any corporation. Read a layoff announcement from Novo Nordisk as an example [1]. No difference. This is what I don’t like about HN, manufactured outrage when one dislikes the messenger. No substance whatsoever. When users are given such a powerful tool like Sora, there will naturally be conflicts. If one makes a vi…

I actually really like that comment. It's an example of classic doublespeak and it's a shame that "Open"AI uses it and we as society tolerate that (as well as other companies of course)

Re: Sora Update #1

#87

Post by TeMPOraL »

timschmidt wrote: post #54
ineedasername wrote: post #50

Earlier quoted context omitted.

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 vie…

Absolutely. Though the smarter these things get, and the more layers of additional LLMs on top playing copyright police that there are, I do expect it to get more challenging. My comment was intended more to point out that copyright cartels are a competitive liability for AI corps based in "the west". Groups who can train models on all available culture without limitation will produce more capable models with less fr…

It's a matter of time. I imagine they'll get more effect suppressing activations of specific concepts within the LLM, possibly in real time. I.e. instead of filtering prompt for "Mickie Mouse" analogies, or unlearning the concept, or even checking the output before passing it to user, they could monitor the network for specific activation patterns and clamp them during inference.

Re: Sora Update #1

#88

Post by rsynnott »

piskov wrote: post #19

Broke: cure cancer, new physics, agi, take your jobs, what have you. Please give us a trillion. Woke: AI slop tictoc to waste millions of human-hours.

Well, yeah, but that stuff was all bullshit, whereas the fake tiktok kind of exists and might keep the all-important money taps on for another six months or so.

Re: Sora Update #1

#89

Post by rsynnott »

noduerme wrote: post #66
piskov wrote: post #19

Broke: cure cancer, new physics, agi, take your jobs, what have you. Please give us a trillion. Woke: AI slop tictoc to waste millions of human-hours.

You make a good point. They may well as admit at this point that curing cancer, new physics, and AGI aren't going to happen very soon. What surprises me a bit is that they'd take this TikTok route, rather than selling Sora as a very expensive storyboarding tool to film/tv studios, producers, etc. Why package it as an app for your neice to make viral videos that's bound to lose money with every click? Just sell it for…

> What surprises me a bit is that they'd take this TikTok route, rather than selling Sora as a very expensive storyboarding tool to film/tv studios, producers, etc.

Because it’s not good enough, I would assume. Hard to see it actually being useful in this role.

Re: Sora Update #1

#90

Post by simianparrot »

> In particular, we'd like to acknowledge the remarkable creative output of Japan--we are struck by how deep the connection between users and Japanese content is!

Translation from snake speech bs: We've been threatened by Japanese artists via their lawyers that unless we remove the "Ghibli" feature that earned us so much money, and others like it, we're going to get absolutely destroyed in court.

Post reply on HN