The bubble is definitely gonna burst and even the dog knows it's a' coming so looking to find an opening to retreat to in safety.

 

AI bubble triggers global stock market sell-off in 2026

Global stock markets slumped on Tuesday as growing fears of an artificial intelligence bubble triggered a broad sell-off in technology stocks. The Nasdaq Composite fell 2.2% and the S&P 500 dropped 1.4%, while South Korea’s Kospi index plunged 10%, triggering a trading halt. The downturn underscores growing concern that massive investments in AI infrastructure might not yield the expected returns—reminiscent of the dot-com bubble.

Why are investors concerned about the AI bubble?

Several factors have contributed to the turbulence. First, AI spending is projected to exceed $1.6 trillion between 2026 and 2029. Second, the Shiller P/E ratio for the US market has risen above 40 for the first time since the dot-com crash, signaling extreme overvaluation.

Third, prominent figures such as OpenAI CEO Sam Altman and JPMorgan CEO Jamie Dimon have warned of a bubble. The economic impact of an AI bubble burst could be far-reaching. Nevertheless, some analysts remain optimistic: Tom Hulick of Strategy Asset Managers foresees no catastrophe, and Wedbush analyst Dan Ives describes the sell-off as a buying opportunity.

Source: https://informedclearly.com/de/wirtschaft/56673/ki-blase-boersenverkauf-2026

 

Last edited: Aug 29, 2026

Thank God! Abominable Intelligence for music and graphics should have been banned long time ago!

 

Yep, that's why I advocate all the time I can to all the people I can reach to fucking unionize, but feels like a madman in de desert. Since my youth, firts in the CNT and later on the CGT, the only thing I've seen is a absolute drops in numbres across all unions in my country and worldwide.

So I'm ecstatic about AI tools of all type and also absotule terrified about our inability to articulate any from of efective collective resistance. Stickers and empty phrases looks cool on the internet but without strikes, resistance boxes and taking huge risks we are going to get crushed like worms.

 

The very thing we cannot talk about openly on here and at best anyone will be cautioned for doing it, has way too much control in way too many areas they do not belong in. Kick 'they who shall not be named', off the A.I. bus and we may be better off.

 

I feel the same way @Demloc : it’s like watching a slow-motion car crash while the driver keeps hitting the gas. On one hand, it’s amazing to have an assistant that can translate, summarize, and compare articles (everything under your critical eye), on the other, AIs can follow biased algorithms or partial guidelines that distort inquiry. That’s why it’s essential to know at least the basics of the subject in order to exercise critical judgment.

A psychologist colleague of my wife’s reported frustration because she uses AI to read and summarize book chapters, articles, and even formulate hypotheses in grad school, basically, everything you shouldn’t do, especially when you're evaluated through the lens of a knowledge producer! With repetition, this can make the mind lazy, sluggish, and poorly trained, especially for people in literature, linguistics, and the arts (the humanities, which deal with a high degree of abstraction, but honestly, I think this applies to any field). So I told her, "—Look, as a grad student, you really need to know your stuff, because you're the one who's gotta exercise judgment over what the AI throws back at you." And then, she hit me with: "—Ugh, what a drag... I thought this was supposed to be some advanced thing that would let me study less!" (Holy shit!). Again, in practical terms, humans want to delegate more than they should to the tool and, worse, are reducing and/or losing their power of inference, choice, and pragmatic action, because they expect the AI to decide, broadly speaking, what’s best for them.

In short: the big problem is excessive, passive use, which generates an atrophy of thought and critical action. It’s not just about using tools like Unison MIDI Wizard, but about routinely adopting the "—do it for me", and ultimately "—decide for me!" method, which in this case atrophies creative capacity. I’ve been hearing reports from teachers and pedagogues working with middle and high school students who express serious concern that interaction with AI is making students more passive, less critical, and more dependent. And this confirms that cognitive atrophy is a real and urgent concern in the educational field.

Which leaves the question that won’t go away: what will happen to this horde of AI dependents on the day the TECHs get together and drop the hammer: "—That’s it, folks! It's over!"? Will the crowd that traded autonomy for convenience wake up to reality, or will they keep trying to type "how to think/produce for myself" into ChatGPT? But then, at that point, the damage may have expanded into cognitive, emotional, and even worse, existential dependence.

If students and professionals are already outsourcing judgment to AI, imagine when musicians start delegating composition entirely to it (which is already starting to happen, at the amateur level). The issue isn’t just licensing, the real questions are: what’s the scale of this production, what’s the quality of it, and what do we lose when we set aside our autonomous capacity for critique and creativity?
Licensing is just one layer... the deeper issue is cultural and human impact.

 

Last edited: Aug 30, 2026 at 4:17 PM

You could also hold a seperate competition for AI music, with the judges themselves.
A competition was held for humanoid robots in China, where all the participants were robots. Such a competition could be held for AI music.

 

Last edited: Aug 31, 2026 at 9:30 AM

@Mynock - Props for just layering your thoughts into a meaningful form.

I´m not sure what will be the tipping point to be honest.

Recently i started playing with growing seeds just to get some grip on it, i´m not a prepper or anything like that but i´m very aware that people who have large datasets that can cross corelate them with the "human experiments".

A lot of specific knowledge brings sorrows and multiplies them.

 

Well, zoom out a bit.
Kospi has been climbing pretty much non stop since april 2025. End of april 2026 it was at 6598, then within two months it got rallied up past 9000, dropped ~44% within 40 days, and is now sitting almost exactly back where it was before the rally began.
In sum: +42 % up and -27 % down.
Price is now running a bit behind actual performance, which isn't surprising after a rally like that. Wouldn't call that a bursting bubble though.

Shiller PE doesn't really tell much more than that the market is expensive compared to last 10 years avg. Young positions get methodologically skewed by that anyway, which is exactly why experts still argue what shiller actually says about the market. Yes, numbers looked similar during dot-com, but that doesn't mean it works the other way around and dot-com is repeating right now.

Tbh, I absolutely think a big chunk of the AI sector may well be overvalued atm, but currently I don't see any hard data that supports it's a bubble. It’s a matter of time whether revenues catch up to expectations and turn into profit or if costs just keep rising. Could be a bubble, could not be. If that's enough to call it a bubble, the claim by itself is a bubble. All you can say for certain is that AI is in a bullish phase imo.

 

Last edited: Sep 2, 2026 at 1:13 AM

So maybe we can get rich quick betting against the AI bubble by creating and buying up some AI market bubble swaps. All we need now is an approximate predicted crash date and we're set! pffft...

 

Hey all you 'Bubble Bursting' geniuses!

You are writing this on the internet, which had its' BUBBLE burst back in 2000! That is 26 years ago.. Umm so what if the bubble bursts. It does not mean the end of AI. It will all just get cheaper.

Get a grip.

 

Being right about the direction isn't enough. Someone would have to be right about the direction, timing, but also have enough capital to ride out the period between the two. Wildly different. People always hear about someone being right on a position after the fact, but not that often when they end up margin called and closed out of the deal a year before, and possibly well over 100% in the hole before their position eventually proves correct.

There are lots of rules about companies doing buybacks of their own stock to keep the prices up. But when you are talking about companies like the hyperscalers involved in all of this, they will get "creative" doing circular financing of one another. They will outlast the "bubble genius's" attention spans.

 

Thanks for your detailed and intelligent take!
Since neither of us can see into the future, we’ll just have to wait and see whether there’s a market correction or not!

AI Bubble: ECB Expert Group Warns of Consequences of a Stock Price Slump

Immense expectations surrounding AI are driving stock prices to ever-higher levels. ECB researchers warn that a price correction is “likely.”

Several experts at the European Central Bank consider it "likely" that the stock market rally driven by AI-related stocks will lead to a "correction." Research into similar trends surrounding past technological revolutions suggests this, according to an ECB blog post. The research group warns that a sharp drop in prices would affect the eurozone in two ways: firstly, through extensive indirect investments in the shares of the "Magnificent Seven"—the tech giants Nvidia, Apple, Alphabet, Microsoft, Amazon, Meta, and Tesla—and secondly, through "the extent of overheating in eurozone stock markets themselves." According to the group, a correction would be expected even if valuations within the sector were rational.

Source and more: https://www.heise.de/news/KI-Blase-...r-den-Folgen-eines-Kursrutsches-11421468.html

 

I´m just waiting for the cheap hardware flood, which will eventually come when the trainings done. Some companies have already stopped scaling datacenters, so we´re pretty close.

 

If an "AI bubble" bursts expect the unexpected.
It's not unlikely that hyperscalers will buy up bankruptcy estates and freed-up production slots for cheap, you get nothing either way.
Even when that doesn't happen and production slots become available for consumer hardware, the shift might take up to 2 years.

 

Stopped? Which ones?

 

Were they stolen cars?

 

because it's surprisingly easy to flag genuine content as AI slop, either someone can tag your work or more likely those AI-driven content management/delivery systems will - it already happens with youtube and facebook, now most recently I think instagram automatically limits your reach if it decides you are AI

 

Nope modern IP and data laws didn't exist back then to weaponize. Instead, they had to invent "public safety" Red Flag Laws to choke out the competition.

Calling AI "theft" is just the modern version of that red flag: an incumbent industry using whatever legal lever currently exists to outlaw technology that threatens its business model.

you can even look at the dataset that the atlantic put out and said was acutal audio used in training.

https://huggingface.co/datasets/laion/LAION-DISCO-12M

they used suno and udio in the article but those tools didnt use the datasets from the article.

https://www.theatlantic.com/technology/2026/06/ai-music-generators-suno-google-udio/687485/

its a "dataset" of song information as text. no actual song data. "Stolen" is just information thats on the internet put into a dataset format which is just a fancy spreadsheet.

its all just a way for us to support our personal view and get mad at whats to come.