I went to the Snowflake Summit last year. And Altman sat on the stage saying LLMs would be coming up with new chip architectures and medical solutions within a year, and that AGI was around the corner. I turned on GPT and let it listen and respond to some of what he said. It replied, in less flattering terms than I am about to use, that he was -- being too optimistic. When you are bouncing around between AGI, Mickey Mouse videos, ad algorithms, and p@rn-bots, I think it's appropriate to question your motives.
I think the only logical conclusion is that many of these tech leaders are liars or have absolutely no idea what they are talking about. Maybe somewhere in between.
On here and else where there are people who see AI for what it is and are absolutely blown away by it and defend these people without realizing that they are regularly promising something much more to investors that can never be fulfilled. The idea that LLMs can ever reach any sense of true AGI is delusional.
It makes total sense to me that this would happen. The economics around Sora and video generation in general are just not there right now, and if you're a company that's also doing research into these things, that's basically a bottomless pit for money. I think OpenAI ceding the space to Google and others for the moment is probably the smart move.
I had fun using Sora and I'm bummed to see it will get removed from the API as well later this year, but no biggie. Veo is plenty good.
It really must cost so much money to generate these videos. That they can generate 12 second videos that are high quality in such a short amount of time - that takes some serious horsepower.
If anything, Sora was an experimental question: giving away video generation is expensive, but is the voluntary user labeling and engagement data, which can be fed into RLHF, accretive enough to model training that it's a meaningful trade to make?
The shutdown of the service makes it clear that the answer was "no."
(It's not a particularly useful signal, though, in evaluating OpenAI's future. It could mean that OpenAI is less interested in video data, which might have implications on their AGI ambitions. It could equally mean that OpenAI has enough data that it's hit diminishing returns, or has found a cheaper source of labeling, or doesn't consider it meaningful one way or another. So there's a lot of thoughtpieces that the shutdown is a sign of weakness, but I don't think it's worth jumping to conclusions.)
I'm actually very surprised, even if it was costing money. Their technological moat has turned out to be much more shallow than expected and competition fiercer. At the moment I think their greatest asset is brand and engagement. With a popular product and a deal with Disney seemed like a slam dunk on remaining prominent in the brand space and retaining user engagement.
Not only have they thrown out a name everyone knew, and exited the market segment, but they've also triggered Netflix/Google graveyard woes. "We may not maintain products you like". This could make people wary of buying into new products, "will it be there in a year?"
It took off rapidly but that was hardly because of any hyping and almost entirely due to word of mouth and people actually liking the product, until the press picked up on it.
From what I remember they still had an invite process when they were getting popular and the demand clearly overwhelmed their servers several times, indicating a much bigger response than they expected. If anything I think OpenAI was downplaying the product at the time.
ChatGPT is one of the most hyped tech products ever. We’ve had nearly 4 years of hype. From claims of AGI to claims we need UBI because of the ways it will devastate the work force.
What are you talking about? It has nearly a billion users. It's the fastest growing consumer app in history. What about it is hype? What does hype even mean?
You might not like the discussion around the tech, but to say the tech is hyped like something like NFTs where they had a crazy amount of media coverage and a tiny minority of people actually using it, is just wrong.
> It took off rapidly but that was hardly because of any hyping and almost entirely due to word of mouth and people actually liking the product, until the press picked up on it.
Not my experience. A whole lot of breathless mentions in media, LinkedIn, and especially top-down company emails. Far fewer cases of people actually liking/using it.
I think the way Sam Altman talked about AI. The framing of it. That they had to hold back the real version because it is just too powerful; they don't even know how it is working; it is already doing these incredible things that would change the world, but we can't / won't release it was all cleverly orchestrated.
The hype cycle for AI products is brutal. Going from "this will change everything" to silence in a few months is rough. Makes you wonder how many other AI products are riding the same wave right now without anyone noticing.
They're just removing it from public access and selling it to big money instead. Think large advertising companies, government agencies, Coke-Cola, Hollywood, etc. The scary part is now that they've removed it publicly, it's going to be harder to keep a pulse on what is real and what is fake. We can't trust any video, audio or text content now.
54 comments
On here and else where there are people who see AI for what it is and are absolutely blown away by it and defend these people without realizing that they are regularly promising something much more to investors that can never be fulfilled. The idea that LLMs can ever reach any sense of true AGI is delusional.
I had fun using Sora and I'm bummed to see it will get removed from the API as well later this year, but no biggie. Veo is plenty good.
It really must cost so much money to generate these videos. That they can generate 12 second videos that are high quality in such a short amount of time - that takes some serious horsepower.
The shutdown of the service makes it clear that the answer was "no."
(It's not a particularly useful signal, though, in evaluating OpenAI's future. It could mean that OpenAI is less interested in video data, which might have implications on their AGI ambitions. It could equally mean that OpenAI has enough data that it's hit diminishing returns, or has found a cheaper source of labeling, or doesn't consider it meaningful one way or another. So there's a lot of thoughtpieces that the shutdown is a sign of weakness, but I don't think it's worth jumping to conclusions.)
Not only have they thrown out a name everyone knew, and exited the market segment, but they've also triggered Netflix/Google graveyard woes. "We may not maintain products you like". This could make people wary of buying into new products, "will it be there in a year?"
It took off rapidly but that was hardly because of any hyping and almost entirely due to word of mouth and people actually liking the product, until the press picked up on it.
From what I remember they still had an invite process when they were getting popular and the demand clearly overwhelmed their servers several times, indicating a much bigger response than they expected. If anything I think OpenAI was downplaying the product at the time.
It's had a ton of hype since then of course.
You might not like the discussion around the tech, but to say the tech is hyped like something like NFTs where they had a crazy amount of media coverage and a tiny minority of people actually using it, is just wrong.
> It took off rapidly but that was hardly because of any hyping and almost entirely due to word of mouth and people actually liking the product, until the press picked up on it.
Not my experience. A whole lot of breathless mentions in media, LinkedIn, and especially top-down company emails. Far fewer cases of people actually liking/using it.
Some people lost their minds over the popular sentiment: https://www.washingtonpost.com/technology/2022/06/11/google-...
Was it fun while it lasted? Sorta, but it got old pretty quick.
Is this a business? Hell no.
Goodbye to Sora
https://news.ycombinator.com/item?id=47508246