AI Is a Mass-Delusion Event | Hacker News Reader

AI Is a Mass-Delusion Event

theatlantic.com

Great piece.

Also, thinking about Sam Altman talking to Theo Von about Dyson spheres kinda makes me want to live on a different planet.

I believe in many respects I already do live on a different planet than these people.

explaining dyson spheres to theo von is like trying to explain art to a shih tzu

Not the first time someone said something insane about data centers.

I present to you: underwater data centers! https://news.microsoft.com/source/features/sustainability/pr...

Realistically, with the cost of spaceflight coming down steadily, this might actually be physically doable. We'd have to figure out the whole physics of it all, or maybe not, OpenAI can just ask ChatGPT for help!

Hard to say if you are being ironic or you really believe that bullshit. "/s" helps with this situations. So, I will assume that you are for real believing that or you have invested a lot of money on AI and the hype profits you. In that second case it seems immoral to write such claims without any disclosure.

Lol I wish I could grift as hard as these guys /s

No, I don't believe that they are serious about building a DC in space lol

It's a dumb idea IMO

It wasn't at all difficult to tell that comment was sarcasm.

The first clue was this line:

> I present to you: underwater data centers!

"I present to you" is clearly lighthearted, and then it gets rounded off with an exclamation mark.

And then this final sentence is also clearly sarcastm:

> We'd have to figure out the whole physics of it all, or maybe not, OpenAI can just ask ChatGPT for help!

I'd suggest to you that in situations where you think the meaning is unclear a better approach might be to assume the best rather than the worst. This is how we can extricate ourselves from the shoutfest that the world has become.

https://archive.md/qCTF4

Surprisingly, this isn't clickbait. It's more that it's a qualitative rather than quantitative analysis. It's more of a cultural and psychological impact piece. I expected something basically dismissing AI as a mass-delusion.

I can't say the author is _wrong_ in their chosen examples like "re-animating a dead kid to interview an AI version of them in the news" That's flat out bizarre. The AGI hype, the whole "every job will be displaced" moving bar.

I say this as an AI maximalist. Progress <> permission. We do need more accuracy, safety, and the players to disclose synthetic content.

The real danger is rebuilding social institutions around "mid" models and only later noticing we've gone wrong -- not Skynet -- but information quality, consent, energy, labor, etc.

Words aren't specific, they're arbitrary. Any model using the arbitrary to stand in for action or specifics will always be mid. Words are in this way gibberish. Only actions are real. That's the achilles heel Warzel can't see, he's a wordsmith.

No word is ever accurate/specific, it's only a metaphor for the event. No software can ever resolve metaphors to real, irreducible events. This had to be taught day one in LLM/NLP/ML for the field to be real. It wasn't the field is false.

Unsurprising. Your virtual girlfriends are just linear algebra and the ELIZA effect.

archive.org

https://web.archive.org/web/20250818211916/https://www.theat...

archive.today and siblings are putting up full captchas with motorcycles and crosswalks for my phone and I prefer dinner to debugging.

Are you using a VPN?

I am only using icloud private relay, country and time zone, no more local than that. But so is the mac, and I get at worst, a checkbox there, though I use Firefox there rather than safari, and extensions vary.

OK just checked. Safari on the mac requires just the one checkbox, but Firefox gives me the stop sign puzzle. Same relay.

Retrying on the phone, this time, the checkbox suffices. So there's apparently some history.

Since I am a scout's honor non-bot, I really am fed up with cloudflare and/or Google, or any other entity making me pay the price for someone else's bad behavior.

Thanks for asking.

Looks like you prefer debugging to dinner :-)

I had an early dinner!

You’ve gotta complain about the state of the modern web once in a while!

Tell me that Cloudflare ISN'T the de-facto web. I use reading glasses to see the monitor, and I have had it with these all-day eye exams.

And yes, I was "there" before the world wide web, with uucp, and this constant interruption sucks even worse.

Who remembers jplvax?

Mass-delusional events remind me of the dancing plague.

Reading the article, or having a good model rewrite it, substituting with the dancing plague is surprisingly grounding.

A model spat out for me an interpretive dance version of the first example. It remained hard to read.

> The feeling of instability she describes is a hallmark of the generative-AI era.

This is by design. Power-hungry billionaires thrive in chaos. God governed countries and the rule of law do not allow for the level of inequality that we are seeing today.

AI is just another tool to sell fear to citizens in the developed world. Fear to lose their jobs, fear to lose the opportunity to invest in the next big thing, fear to be outgun on wars, fear to not be part of “A new era for humanity”.

I love when well articulated and informed journalists put into words things that many of us only understand instinctively.

> Earlier this month, OpenAI released GPT-5, to mixed reviews. Altman had promised “a Ph.D.-level” intelligence on any topic.

When investors put their money on words and fantasy not on sound business we are in for a hard landing of the economy.

The problem is the industry hasn't solved the conduit metaphor paradox, so AI will always remain a magic act in public view that goes nowhere and a a specific tool that can help solve real problems in chemistry and biology. As a general tool it can never work. Words are arbitrary, they're meaningless.

Jim Acosta is a mass delusion event.

The second Tulip mania passed by almost unnoticed yet it's AI getting all the hate?

Anyone seen some NFT tulips still for sale?

The problem with AI is the marketing is both cringe and retarded even though the product is genuinely useful.

All media technologies seem weird at first. 19th century Europeans fled from the first movies of trains, expecting to be struck by the moving image. Deep jungle tribes believe that photographs steal a piece of your soul. I myself remember spending hours on YTMND 20 years ago. The early internet was Weird and Fresh and exciting. So if GenAI makes established, successful people feel weird and confused, then to me that seems like a good sign.

> Europeans fled from the first movies of trains,

But the best guesses I have seen, this is an urban legend.

> Deep jungle tribes believe that photographs steal a piece of your soul.

That one is about religion more then about technology. Some Islam branches believe depicting humans in any form is wrong and they are aware of technology for years.

> But the best guesses I have seen, this is an urban legend.

Could be. OTOH at that time, even with jerky motion in B&W, on a large screen in a darkened room? With people never having experienced that before? Seems plausible.

Is this an urban legend, too?

https://illuminatingfacts.com/the-train-speed-panic-why-peop...

Reply on news.ycombinator.com