Hacker News new | ask | show | jobs
by zerosizedweasle 238 days ago
"Whether Sora lasts or not, however, is somewhat beside the point. What catches my attention most is that OpenAI released this app in the first place.

It wasn’t that long ago that Sam Altman was still comparing the release of GPT-5 to the testing of the first atomic bomb , and many commentators took Dario Amodei at his word when he proclaimed 50% of white collar jobs might soon be automated by LLM-based tools."

That's the thing, this has all been predicated on the notion that AGI is next. That's what the money is chasing, why it's sucked in astronomical investments. It's cool, but that's not why Nvidia is a multi trillion dollar company. It's that value because it was promised to be the brainpower behind AGI.

5 comments

What signals have you seen that point to investment being predicated around AGI? Boosting Nvidia stock prices could also be explained by an expectation of increased inference usage by office workers which increases demands for GPUs and justifies datacenter buildouts. That's a much more "sober" outlook than AGI.

In fact a fun thing to think about is what signals we could observe in markets that specifically call out AGI as the expectation as opposed to simple bullish outlook on inference usage.

"Boosting Nvidia stock prices could also be explained by an expectation of increased inference usage by office workers which increases demands for GPUs and justifies datacenter buildouts"

AI is already integrated into every single Google search, as well as Slack, Notion, Teams, Microsoft Office, Google Docs, Zoom, Google Meet, Figma, Hubspot, Zendesk, Freshdesk, Intercom, Basecamp, Evernote, Dropbox, Salesforce, Canva, Photoshop, Airtable, Gmail, LinkedIn, Shopify, Asana, Trello, Monday.com, ClickUp, Miro, Confluence, Jira, GitHub, Linear, Docusign, Workday

.....so where is this 100X increase in inference demand going to come from?

Oh and the ChatGPT consumer app is seeing slowing growth: https://techcrunch.com/2025/10/17/chatgpts-mobile-app-is-see...

Integrations and inference costs aren't necessarily 1:1. Integrations can use more AI, reasoning models can cause token explosion, Jevons Paradox can drive more inference tokens, big businesses and government agencies (around the world, not just the US) can begin using more LLMs. I'm not sure integrations are that simple. A lot of integrations that I know of are very basic integrations.

> Oh and the ChatGPT consumer app is seeing slowing growth: https://techcrunch.com/2025/10/17/chatgpts-mobile-app-is-see...

While I haven't read the article yet, if this is true then yes this could be an indication of consumer app style inference (ChatGPT, Claude, etc) waning which will put more pressure on industrial/tool inference uses to buoy up costs.

My experience suffering with JIRA daily is that the AI is useless and fairly easy to ignore. If it were actually helpful, I could imagine using it more, and the costs would increase proportionately.
On the other hand, I’ve found the integration in Confluence quite helpful, particularly for making sense of acronyms.
I think the motivation for someone like Altman is not AGI, it's power and influence. And when he wields billions he has power, it doesn't really matter if there's AGI coming.
Yep, he just wants to become too big to fail at this point.

I view OpenAI like a pyramid scheme: taking in increasing amounts of money to pursuit ever growing promisses that can be dangled like a carrot to the next investor.

If you owe investors $100 million, that's your problem. If you owe investors $100 billion, that's their problem.

Who else pursued this strategy in the dot-com bubble and failed to the point of going under?

I'm old enough to remember it, but young enough not to remember it well.

What was predicted to be next: AGI

What we got next: porn

Porn has driven everyday tech. Online payment systems, broadband adoption.

Porn (visual and written erotic impression) has been a normal part of the human experience for thousands of years. Across different religions, cultures, technological capabilities. We're humans.

There will always be a market for it, wherever there is a mismatch between desire for and access to sexual activity.

Generate your own porn is definitely a huge market. Sharing it with others, and then the follow-on concern of what's in that shared content, could lead to problems.

> There will always be a market for it, wherever there is a mismatch between desire for and access to sexual activity.

Attractive people in sexually fulfilling relationships still look at porn.

It's just human.

>Attractive people in sexually fulfilling relationships still look at porn.

How do you know those relationships are "sexually fulfilling"?

> How do you know those relationships are "sexually fulfilling"?

You either believe what people report, the clearly-stated position on erotic material of the American Association of Sexuality Educators, Counselors and Therapists (1), or you can just imagine in your head what you think other people’s sex lives are like and just believe whatever you come up with.

1. https://www.aasect.org/our-mission.html

Self-reporting isn't something to be taken at face value. It's not a dichotomy between they say, and you imagine, there's a healthy scepticism option. Also the URL leads to "Our mission" section of some organization's rather clunky website, so if it was intentended as a proof, it failed.
Works for me. You end up travelling for a week away from your partner, they're sick for a while, etc.
I, uh, have a friend, who tells me watching porn with their partner once in a while, can be pretty hot.
>Porn (visual and written erotic impression) has been a normal part of the human experience for thousands of years.

You need to stretch 'porn', 'thousands if years', and certainly 'normal' definitions really hard to believe it. Even my granddad's (not thousands of years ago) exposure to porn was a one-time event when he served in army. My nephew's exposure is everyday. Which he realized to be an addiction at one point way stronger than nicotine.

This is a meme I see online often (and in the show Silicon Valley), but I don't think it holds up in practice.

Re: payment systems, Visa and MC are notoriously unfriendly to porn vendors, sending them into the arms of crooked payment processors like Wirecard. Paypal grew to prominence because it was once the only way to buy and sell on Ebay. Crypto went from nerd hobby to speculative asset, shipping the "medium of exchange for porn purchases" entirely.

As for broadband adoption, it's as likely to have occurred for MP3 piracy and being 200X faster than dialup, as it was for porn.

Pretty sure most of the quotes regarding the connection of porn and tech come from people who just watched that episode of Silicon Valley.

A lot of it was clearly exaggerated for dramatic effect for a comedy TV show, and people run with it as fact.

I don't see how that explanation helps.

Promise: AI will change the world.

Delivery: 1000 year old vice.

To be very fair here, a long time before gpt-5 porn was already being produced with stable diffusion (and other open models). Civitai in particular was an open playground for this with everything from NSFW loras, prompts to fined tuned models.

I had to work for a bit with SDXL models from there and the amount of porn on the site, before the recent cleanse, was astonishing.

to be fair we also got Stephen Hawking bungee jumping | snowboarding | wrestling | drag racing | ice skating | bull-fighting | half-pipe
I can't imagine the republican party is going to be particularly happy about AI being used for mass porn generation
The party of grindr-crashing sexual repression[1] outwardly denounces such depravity, but inwardly rejoices at all the shameful images they intend to generate.

1. Red states are way ahead on porn consumption, based on past annual reports by Aylo.

prompt records = mass blackmail generation
At least the valuations make sense now. ;)
Wait, we got porn?
Yes, Sam A said that "erotica" was coming to openAI. I don't think he's mentioned visual pornography though https://www.axios.com/2025/10/14/openai-chatgpt-erotica-ment...
AGI is like L5 automated driving - academic concepts that have no bearing on the ability of these technologies to transform the economy.
And no bearing on the ability of these technologies to thoroughly screw us.
We were promised AGI and all we are getting is Bob Ross coloring on the walls of a Target store.

The app is fun to use for about 10 minutes then that is it.

Same goes for Grok imagine. All people want to do is generate NSFW content.

What happened to improving the world?

I apologise for talking past the point you're making, but, Bob Ross was a human being, you know, with thoughts and stuff. How could any of these AI toys possibly compare?

I would love to have Bob Ross, wielding a crayon, add some happy little trees to the walls of a Target.