The end of Sora also means the end of Disney’s $1 billion OpenAI investment
- Thread starter JournalBot
- Start date
Upvote
171
(171
/
0)
So this means I won't get less-than-stellar Bluey episodes on my Disney+ because creativity and quality costs money but AI is only $1 B dollars cheap in a too-good-to-be-true situation where nobody wins?
Oh no.....
Oh no.....
Upvote
72
(72
/
0)
Did OpenAI possibly realize that GenAI video generation was morally incorrect, violating the rights of artists everywhere, and their moral compass demanded... nope, couldn't complete that sentence, I'm sure it's all about money.
Upvote
131
(135
/
-4)
Upvote
181
(181
/
0)
A vast, deafening sucking sound.
I do not belong to the cohort that thinks hope is dumb and everything is doomed all the time. I have earnest and sincere hope that this represents the beginning of the collapse.
I do not belong to the cohort that thinks hope is dumb and everything is doomed all the time. I have earnest and sincere hope that this represents the beginning of the collapse.
Upvote
87
(87
/
0)
Seems like the only people who would care about this are those who want to steal disney's characters anyways. Disney doesn't deserve to exist, but the artists whose work they claim ownership of should be personally compensated for every single frame the plagiarism machine generated.
Upvote
11
(19
/
-8)
More big numbers AI "investments" that turned out to be smoke and mirrors, entirely non-binding, no money changed hands etc.
What a surprise.
What a surprise.
Upvote
96
(97
/
-1)
I've been at a number of companies built on R&D, but it's always still seen as the easiest thing to cut by those in charge. These days new executives are essentially buying a house to flip it, with no plans to live there. They're not concerned about the foundation
Upvote
97
(97
/
0)
I assume the thinking was, "they're going to rip off our stuff no matter what, let's at least get a cut".
Upvote
18
(19
/
-1)
In case anyone missed it, here's an example Disney gave of the sort of high quality AI content they were developing. It's worse than even a skeptic like me could have imagined. I'm still in disbelief that they seemed proud of this, and wanted to show everyone.
TED talk
Yup... that's some good Star Wars all right
TED talk
Yup... that's some good Star Wars all right
Last edited:
Upvote
77
(77
/
0)
Uhh can I get a little more info on that top shot? Is someone vacuum-sealing Mickey using his own glove? It's from the “Mickey: The True Original & Ever Curious” exhibit in Shanghai. No information is given on what it means though.
Upvote
22
(22
/
0)
It's bugging me that the thumb is wrong, the left-handed Mickey glove is being worn on the right hand for some reason... wait a minute, screwed up fingers... it must be AI!
Upvote
24
(24
/
0)
Sorry to ask, but I haven’t been keeping track
What percentage of OpenAI’s ‘big money’ deals have turned out to be vaporware ?
My casual assessment is that its easily north of 80%
What percentage of OpenAI’s ‘big money’ deals have turned out to be vaporware ?
My casual assessment is that its easily north of 80%
Upvote
53
(53
/
0)
You clearly just pulled that number out of your ass.
Congratulations, kid, you've got what it takes to make it big in this business.
Upvote
98
(98
/
0)
It's probably mostly vaporware and gift cards. I recall that Microsoft's "investment" in OpenAI mostly considered of Azure credit, which means that MSFT spent far less money than the number on the headline.
Upvote
37
(37
/
0)
Regardless of your feelings about the morality of generative AI, a basic fundamental truth that most people would have agreed upon pre-LLMs is that mimicry is not artistry. So with Sora, where's the value? It costs more money to create these shorts than they produce in value. So easy decision by OpenAI to shut it down.
Likewise as a long time software engineer, I've always known that the code I write has no value. It's not an asset; it's a liability. It's a means to an end. The creativity and artistry is in knowing what to build, not how to build it. Yes, it's nice to have micro optimizations and move toward fewer keystrokes per output. But simplifying the implementation does very little to help ship new features, because the act of drawing, writing, coding has never been the bottleneck.
According to Wikipedia, A Game of Thrones took Martin around five years to write. If my mental math is correct, I could sit down and retype out the whole thing in 48 hours and make subtle changes at the same time. I cannot believe it took him so long!
Point being - even if you can automate away the act of pen on paper and produce an insane amount of mimicry, if there is no audience or consumption, it's just a big pile of wasted compute.
Likewise as a long time software engineer, I've always known that the code I write has no value. It's not an asset; it's a liability. It's a means to an end. The creativity and artistry is in knowing what to build, not how to build it. Yes, it's nice to have micro optimizations and move toward fewer keystrokes per output. But simplifying the implementation does very little to help ship new features, because the act of drawing, writing, coding has never been the bottleneck.
According to Wikipedia, A Game of Thrones took Martin around five years to write. If my mental math is correct, I could sit down and retype out the whole thing in 48 hours and make subtle changes at the same time. I cannot believe it took him so long!
Point being - even if you can automate away the act of pen on paper and produce an insane amount of mimicry, if there is no audience or consumption, it's just a big pile of wasted compute.
Upvote
62
(62
/
0)
Upvote
21
(21
/
0)
Wouldn’t I have to have my daddy pay for tuition at Wharton’s first?
Upvote
29
(29
/
0)
While practically every white-collar worker and student needs to read and write text every day, and many of them also need to accompany the text with images from time to time, only a small percentage of them need to produce video to do their jobs. It follows that the only way video AI can conceivably produce a lot of revenue is by having contracts with companies like production studios and advertising agencies that make video for a living. We have now learned that OpenAI couldn't convince anybody of consequence other than Disney, which will make the future of other video AIs quite interesting.
Upvote
34
(34
/
0)
I read on Forbes a while back that Sora cost $15M/day to operate, so that $1B wasn't going very far. Between killing Sora and the introduction of ads, it's clear that OpenAI sees the clock ticking on their ability to light money on fire without showing some kind of ROI.
Upvote
56
(56
/
0)
For OpenAI. The writing's been on the wall a while for them.
Not for Google or Anthropic.
Upvote
14
(14
/
0)
In our collective governmental wisdom, we have decided to combine the deflation of the AI/Nvidia bubble so vividly described by Ed Zitron, with the biggest oil shock in half a century. May you live in interesting times.
Upvote
33
(34
/
-1)
I think the most interesting part of this story is the fact that no money changed hands. I feel badly for Ed Zitron.
I saw a few comments right here on Ars, where we should know better, talking about how mad Disney would be about all that investment money going up in smoke.
LLM investment story mantra: "There was no money. There never is."
I saw a few comments right here on Ars, where we should know better, talking about how mad Disney would be about all that investment money going up in smoke.
LLM investment story mantra: "There was no money. There never is."
Upvote
42
(42
/
0)
Has anyone seen any kind of reasonably rigorous systemic review of AI successes and failures? Something that gives the big picture? It's easy to cherry pick articles about failures of AI, but I am having a hard time getting a sense of whether AI failures outnumber and outweigh the successes, and where AI excels vs where it's a disaster, and what the real, reasonable range of it's economic utility is?
Has anyone seen any kind of report or study on this?
Has anyone seen any kind of report or study on this?
Upvote
20
(20
/
0)
This is how the original conversation between Disney and OpenAI went.
Disney: Your software is using our IP without permission! We demand you pay us 10 times the money you made with our characters.
OpenAI: We accept your demands, but unfortunately for you, we operated at a loss, so you actually owe us money now. That'll be 1 billion, please.
Disney: I don't see any flaw in your logic.
Disney: Your software is using our IP without permission! We demand you pay us 10 times the money you made with our characters.
OpenAI: We accept your demands, but unfortunately for you, we operated at a loss, so you actually owe us money now. That'll be 1 billion, please.
Disney: I don't see any flaw in your logic.
Upvote
66
(66
/
0)
Well … you kinda have to define what your metrics are
And which division the report is coming from (PR/Marketing or accounting)
As far of ROI on investment vs productivity gain there’s been quite a few reports lately
… and unlike the marketing ones they don’t seem optimistic
Upvote
16
(16
/
0)
I never saw anything that indicated Disney had any interest in Sora for production work outside of puffery about that mini-Twitch user-generated content thing. It was little more than a merchandising deal.
And if true Sora only generated about $2 million in revenue, wow. I have to wonder if that would even cover Altman's Davos travel budget.
Upvote
28
(28
/
0)