- cross-posted to:
- PurchaseWithPurpose@lemmy.world
- cross-posted to:
- PurchaseWithPurpose@lemmy.world
cross-posted from: https://lemmy.world/post/44699253
This is clearly a sign that the product failed to draw in enough customers and its viability was overhyped.
Hopefully, it is the start of the AI bubble bursting.
You must log in or # to comment.
Finally, a good news
It’s so they can repurpose that capacity for developing robots. It’s not good at all.
OpenAI told the BBC on Wednesday that it has discontinued Sora so that it can focus on other developments, such as robotics “that will help people solve real-world, physical tasks”.
Robots aren’t like software, it’s immediately obvious when they don’t work the way they’re advertised whereas chatbots can trick people into thinking they’re way more useful than they actually are. The “fake it till you make it” “move fast and break things” ethos of tech doesn’t work when there’s actual, physical evidence that shit’s busted.
Unpopular Opinion Incoming
I was assigned at work to evaluate a few LLMs for potential adoption, so I spent a solid week doing so.
Most of the “AI is broken and doesn’t work” on here is solid echo chamber cope. It’s more competent than several of my coworkers, though it’s thankfully not ready to replace knowledge workers as it requires a knowledge baseline to best direct it and evaluate its answers.
I still advised against using it for multiple reasons, including ethics, but much of Lemmy is playing make believe about the actual capabilities of LLMs.
Cool anecdote. Every time we actually see real data, though, the numbers don’t reflect much in the way of productivity gains or increased efficiency or better output. People say that LLMs are useful because it feels useful, but we aren’t seeing actual usefulness. The most recent study out of Duke University observes “a productivity paradox, in which perceived productivity gains are larger than measured productivity gains, likely reflecting a delay in revenue realizations.”
A delay. Sure.
I think it was build on the original transformer architecture and as such took a shitload of compute and was slooow, so I guess they picked the wrong architecture and had to scrub the whole strategy. Huge loss… That also point to US not having enough cheap compute (compute->combined mem and processing) - likely from missing electricity. Lovely… Die Saltman, die, and Go go China !
Go go China !
Bops the tankie.
Like, I have a Chinese LLM loaded right this second and follow them closely, but holy moly. Curb your enthusiasm.
Anyway, OpenAI has plenty of compute to train a Sora 2 if they want, but apparently they don’t. My guess is some combination of:
-
They couldn’t figure out a more efficient architecture, like you speculated. I buy that. OpenAI’s development is way more conservative than you’d think, and video generation is inherently intense, especially if Sora 1 is the baseline.
-
…Maybe they looked at metrics, saw Sora is mostly used for spam, scams, or worse, and pulled the plug for liability reasons?
-
They’re focusing on short-term profitability, as other commenters mentioned.
-
Why stop there? Just shut the whole company down.
Best I can do is shutting the whole planet down.
OpenAI said it will discontinue Sora, the generative-AI video creation platform it launched in late 2024, without providing a reason for the decision.
That is the strongest indication this is the beginning of the end for the AI bubble. Sora burned a ton of processing power, with no clear value proposition, just to keep the hype cycle going a little longer. Shutting down without explanation leaves the most likely one: they are out of helium to pump into the balloon. And if that balloon isn’t inflating, it’s deflating.
Maybe you can only watch so many nonsense videos. I assume I’m sadly wrong though.
It’s not and probably the opposite.
When Sora launched it was way ahead. Seedance 2’s release was notably better than any of the other video gen models, Sora included.
The market is getting commoditized because there’s no moat and OpenAI hasn’t led on pretty much any release for a while now other than Sora, which they’re probably falling behind on now.
This is the opposite of a burst from a tech standpoint, even if OpenAI as a company starts to pop.
TL;DR: This is likely happening because the tech accelerated across the industry in ways OpenAI can’t catch back up to, not because it’s lagging.
Isn’t spending billions of dollars with nothing to show for it in the end the definition of a popping bubble?
It was a weak attempt to keep relevance when faced against Gemini and Claude. But it’s completely unnecessary now that OpenAI has contracted with the government. They get all that sweet tax payer money and get to repurpose a ton of GPUs making stupid videos to supporting that new gov contract.
my guess is that ai video is not going away, and that sora was largely an expensive marketing project
Chime in if you disagreee, but there’s really only 2 reasons a company like OpenAI shuts down a core service like Sora:
- The service is hemorrhaging money to the point of financial unsustainability.
- The service is not popular enough to drive investor hype as a “loss leader”
We already know that OpenAI is losing money on their generative “AI” products across the board, to the tune of billions of dollars per year, and the economic woes that come from rising hardware prices, oil and gas shortages, and another pointless war in the middle east only make the situation worse for them money-wise.
And so that really just leaves me to conclude that Sora has not maintained the level of popularity and growth needed to impress investors as Q1 comes to a close. Whether it’s users, subscriptions, or time, they must have looked at the numbers and really didn’t like what they saw.
Hopefully this is the beginning of the end of the ridiculous “AI” bubble, and the start of a new tech sector correction.
There’s a third option this time.
It uses a lot of resources they can use immediately for the military contract that will now inevitably form the backbone of the company and effectively will mean they have won the AI war. Anthropic fumbled by not doing what the military wanted immediately, and showing a minimal backbone publicly.
As someone who named their daughter Sora in 2021, this is the best news I’ve gotten this year.
Congrats! 🥳🥳🥳
All people ever did with sora was make doorbell cam footage of dogs watergunning old ladies and gorillas getting sucked into tornados. AI image and video generation is just a tool to make a funny joke, it’s incapable of doing anything serious in its current state, and with the amount of processing power it needs just to be a digital circus clown it’s unlikely to become anything more.
Openai is the canary
It was used almost exclusively for slop and slop-based ads or videos that shouldn’t be slop. I was on there yesterday and some account had 2 videos of a woman in front of a plain wall talking for 15 seconds about tax implications for investments. A real human could have filed it with an iphone in 3 minutes.
But now that’s Google and Grok’s problems, I guess.
If the audio and video is AI-generated I’m going to assume that the script is, too.
Maybe time to move retirement accounts into cash.
you mean giving away billions of dollars of computer with no monetisation strategy was bad? man who would have thought. not sam, apparently. if only there were like, some way to have realised that the goal of business is to earn money
So many people seem to have no idea what they’re talking about. This isn’t ending AI video creation, it just cost them a lot of money to offer it. You can generate a video on your own computer already. AI video isn’t going away because one company isn’t letting people do it on their servers for free any more.
So youtube will be worth watching again right?
Right?
It still is for the creators there. Instead of browsing the algorithm I start on the subscriptions page, to only see uploads from people I actually want to.
There’s sometimes complaints about “I thought you were dead” when the channel has been uploading regularly the entire time. People just never got recommended the videos despite hitting all the buttons.
For example, did you know both Physics Girl and Tom Scott have returned this month - hopefully a sign that the world can still heal.
Some unsolicited add-on recommendations -
Ublock origin - beyond the addblocking, I use the picker tool to filter all the extra sections like “news”, “trending” “you might like” etc.
Unhook - toggles to disable a bunch of features like comments, home screen, end screen etc.
Enhancer for Youtube - Themeing and a bunch of extra settings like setting defaults for each video. speed, volume, resolution, fill screen (which is different than full screen), PIP while you scroll comments. (The author just did a rework, so it can be a little bugged sometimes - reinstalling it fixed it for me last time it went wonky.
