75 % des US-Aktienmarktwachstums der letzten Jahre kommt von der KI, aber das basierte auf einem Versprechen. Das AGI stand vor der Tür. Jetzt schwenken Unternehmen wie OpenAI auf den Verkauf von Anzeigen und Pornos um, ein sicheres Zeichen dafür, dass sie nicht glauben, dass AGI bald eintreffen wird.
Wenn die KI-Blase platzt, was passiert danach?
Ich vermute, dass es eine Gegenreaktion gegen Big Tech geben wird. Vielleicht ist das Jahr 2025 der Höhepunkt ihres politischen Einflusses. KI ist bei vielen Menschen bereits weitgehend unbeliebt, und das wird sich noch verstärken, wenn sie sehen, ob sie die Wirtschaft und ihre Renten zum Absturz gebracht hat.
KI, die Technologie, wird uns weiterhin begleiten, auch wenn dies bei vielen der heutigen KI-Unternehmen nicht der Fall sein wird. Auch ohne AGI hat es immer noch das Potenzial, transformativ und wirtschaftlich disruptiv zu sein. Regelbasierte Unternehmen – Recht, Buchhaltung, Transaktion und Schadensbearbeitung – könnten überflüssig werden. Humanoide Robotik und autonomes Fahren, beides Aspekte der KI, werden irgendwann Millionen menschlicher Arbeitskräfte ersetzen.
Das Platzen der KI-Blase würde eine Rezession bedeuten. Rezessionen führen dazu, dass Unternehmen ihre Belegschaft abbauen. Ironischerweise werden sie dieses Mal in der Lage sein, viele der entlassenen Menschen durch KI zu ersetzen. Der Absturz, den die KI verursacht, wird also auch ihre Einführung beschleunigen.
If the AI bubble bursts, what will come after?
byu/lughnasadh inFuturology
23 Kommentare
If AI works well enough to replace workers, how is there a crash? Productivity gains would be big.
And AI is a lot more than LMMs.
Re-selling high end processors.
When the dot com bubble burst, Amazon went on to dominate all commerce.
A bubble bursting wipes out all the fake businesses clearing the path for productive use of AI.
Go long on Jacob’s Ladders.
All the evil that’s taken power and high level management positions will buy them as their Zoom backgrounds
> 75% of the US stock market growth of the past few years has come from AI, but that was built on a promise. That AGI was just around the corner.
Sure, hype and overpromising drives prices, but it’s not like if we don’t get some fabled „AGI“ it was all for nothing. If AI can replace massive swaths of employees (either straight up or by making existing ones more efficient) then that’s still a massive gain (edit: for the companies). And that’s already happening.
Growth will not be partially on employment numbers anymore. Ironically it will most likely allow AI and automation to near fully take over production.
The bad times are coming folks.
Some people will lose their jobs. Others will lose their savings. There will be a recession as the economy balances out. A few AI companies will survive and dominate their niche markets. 5 years later everyone will laugh about it and make jokes. 10 years later the next bubble will start to inflate and the cycle will repeat.
There’s nothing new except the widget. It’s the same song and dance.
The ai bubble is going to burst, but that won’t end AI. Someone is going to be left standing, maybe several companies. It will wreck the economy for a while, and we’ll have to add that to people losing jobs to AI and probably increasing climate change. It’s going to be rough, but again, someone will survive. Probably not me
After the dotcom bubble there is still .com in existence, and so it will be with AI.
The poor will get poorer, the rich will be mildly inconvenienced then dry their tears with a wad of cash.
> legal, accounting, transaction, and claims processing could all be made obsolete.
I seriously doubt this. We’ve already seen multiple law firms get in trouble for using AI and having it cite cases or laws that don’t exist or make up quotes. It hallucinates too much to be used in anything that requires a high degree of accuracy and if you have to double check every single thing it produces what’s the point of it? AI experts have said we can never get rid of the hallucinations with LLMs so that’s not going away. You’re still paying someone to spend just as much time double checking it as you would have to have them do the work in the first place.
I think they are actually after the porn market and it will probably just about keep a few of them afloat.
Honestly the amount of people using ChatGPT and blindly trusting its efficacy at completing any task is telling me that it’s unlikely to burst because even if it’s objectively shit people believe it’s amazing and that’s what matters at the end of the day.
Beyond the AI bubble of course there’s the tech bubble and the market’s overall tendency to price in farther and farther out into the future. That’s never a good thing but it’s been going on long before AI.
> If the AI bubble bursts, what will come after?
How could learning be a bubble?
Kind of like asking what comes after the energy bubble bursts. Suddenly people use no energy for anything. What comes after that? What comes after joules?
AI „living up to its expectation“ also means cutting the workforce. There’s so much money invested, because it promised cutting workforce
Hate to say it but the US is already in recession. It’s being propped up by circle investments in AI/Big tech. Once that bubble goes, it will drag down the US and the rest of the world with it. Nothing good will come from this. It will make 2008-2011 seem like holidays in comparison.
> I’d guess there will be a backlash against Big Tech.
It should have happened a long time ago with Microsoft Windows, shitty evil enshittified smartphones and social media platforms. But no, it does not happen. I believe it is mostly because most people are WAY below your level of intelligence. Just meet random people and look at how they use tech. There is very little understanding going on. Just mammals with brainrot brains manipulated by high-tech mind-control corporations. Musk is a nice guy, etc.
Hopefully ssds will start getting cheaper again
AI research will continue, but with less funding. They will substitute cleverness for money. The irrational investment frenzy is driven by hypemongers and pundits, and appeals to investors desperate to find „the next big thing“. This reminds me of the military saying „generals plan to fight the last war“. Investors saw the riches generated by the PC and internet revolutions and hope the AI revolution will make them rich. I suspect that most will not get rich
“AI, the technology, will still be with us, even if many of today’s AI companies won’t be. Even without AGI, it still has the potential to be transformative and economically disruptive. Rules-based businesses — legal, accounting, transaction, and claims processing could all be made obsolete. Humanoid robotics and self-driving, both aspects of AI, will eventually replace millions of human workers.”
So what is the AI bubble bursting to you, just a handful of prominent AI companies’ stock loses value? To me, the AI bubble bursting is AI never gets good enough to replace more than a few hundred thousand workers in the US. We don’t need AI to hit AGI to put millions of jobs at risk, as you note, and presumably that means someone is making and profiting from the AI that lets companies do without those workers.
Now, an alternate “handful of prominent AI companies’ stock loses value” scenario is we have a series of breakthroughs letting good enough open source AI models run on powerful but affordable personal computers so giant server farms aren’t needed for most AI applications (writing code, making videos, etc).
Plenty of use cases for AI, but the revolt against content is encouraging. However, AI improving workflows is here to stay- AI creating significant things from scratch has issues because it doesn’t train on proprietary secrets for the most part. It’s the equivalent of trying to build a strategy from half hearted conference panels.
If? I’d say it’s inevitable the cost alone of having to build the servers for AI and then on top of that the cost of cooling them to make sure they don’t overheat. Also the intelligence isn’t where it needs to be to integrate it in meaningful way. the fact they’re having to pivot to porn already says it all.
Ai and its developments lately have been a good that I hope progresses in the future, but it’s not going to be what tech companies and shareholders want in 2-4 years, rather in 30. The problem is everyone wants the results now, and cause that’s not gonna happen, people are gonna get impatient and throw a tantrum which will tank the market.