Say farewell to the AI bubble, and get ready for the crash

Los Angeles Times ^ | Aug. 20, 2025 3 AM PT | Michael Hiltzik

[E. Pluribus Unum]

Most people not deeply involved in the artificial intelligence frenzy may not have noticed, but perceptions of AI’s relentless march toward becoming more intelligent than humans, even becoming a threat to humanity, came to a screeching halt Aug. 7.

That was the day when the most widely followed AI company, OpenAI, released GPT-5, an advanced product that the firm had long promised would put competitors to shame and launch a new revolution in this purportedly revolutionary technology.

As it happened, GPT-5 was a bust. It turned out to be less user-friendly and in many ways less capable than its predecessors in OpenAI’s arsenal. It made the same sort of risible errors in answering users’ prompts, was no better in math (or even worse), and not at all the advance that OpenAI and its chief executive, Sam Altman, had been talking up.

AI companies are really buoying the American economy right now, and it’s looking very bubble-shaped.

— Alex Hanna, co-author, “The AI Con”

“The thought was that this growth would be exponential,” says Alex Hanna, a technology critic and co-author (with Emily M. Bender of the University of Washington) of the indispensable new book “The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want.” “Instead, Hanna says, “We’re hitting a wall.”

The consequences go beyond how so many business leaders and ordinary Americans have been led to expect, even fear, the penetration of AI into our lives. Hundreds of billions of dollars have been invested by venture capitalists and major corporations such as Google, Amazon and Microsoft in OpenAI and its multitude of fellow AI labs, even though none of the AI labs has turned a profit.

Public companies have scurried to announce AI investments or claim AI capabilities for their products in the hope of turbocharging their...

[Common Sense 101]

This dipsh*t Michael Hiltzik probably would have said forget about manned flight, as the Wright brothers were only able to stay in the air for 12 seconds...

[Raycpa]

I have only begun to experiment with AI in my business. I can see it easily raising my bottom line by over 20%. It’s going to reduce cost increase capacity and give me additional services without additional work.

[kvanbrunt2]

My first AI program was called “I got your...”. What ever you typed into the input the program would wrap an “I got your...” in front of it. it was great.

[Lazamataz]

I have only begun to experiment with AI in my business. I can see it easily raising my bottom line by over 20%. It’s going to reduce cost increase capacity and give me additional services without additional work.

Be careful. I've been leveraging AI for almost a year, and I find that the improvements are only so-so.

[KingLudd]

“I have only begun to experiment with AI in my business.”

How so? No need to be specific. I’m curious.

[ProtectOurFreedom]

I just met with a landscape design / build firm. The president and VP business development met with another client yesterday, flew a drone over the site, took some photos, and walked the project site dictating ideas into the Apple “Memos” app. They fed the photos and transcript from the Apple “Memos” recorded discussion to ChatGPT and in seconds had a gorgeous 3D photorealistic rendering of what the customer’s finished project would look like. I’m a bit jaded, but this was incredible. They then put that concept into a traditional landscape design software package.

I met with another competing company who took photos of the property, printed them out, made hand ink markups on the photos (just very rough squiggles and boxes for plants) and provided a plant list poorly coded to the location on the property. She included a sloppy hand written bill of materials. No plans, no real design work, no computer work. It was really junk. She billed me for that “work,” too! What a waste of money.

Guess who won the job?

[plain talk]

AI is just getting started. It will take on many names and be here in some form for the foreseeable future.

[Antoninus]

Sounds like one of those “Internet bubble” articles from 1999.

[Hyman Roth]

People over profit man. /s

[ArcadeQuarters]

Maybe it’s a bubble, maybe not. If you’re not a super genius, dollar cost averaging into the main indexes works best. I’ll come out better off just like after buying into the real estate crash, the Covid crash and the smaller dips. *yawn* The worst thing to do would be to sell and desperately try to get back in later.

[Vermont Lt]

GPT is not exactly the meat and potatoes of AI. It’s merely the face that a lot of people see—kind of like GROK. It’s a cute toy.

Where AI will really come into play is manufacturing and supply chain. That is not the stuff you are going to see in your internet browser.

I used to manage call centers. The AI applications in that field are being used to route calls and present information back to management so that problems can be solved, and anticipated well ahead of the deluge (when things go bad, they go bad fast!). Yes, we used to have people do that stuff. But AI is faster and makes fewer mistakes.

AI is being integrated into almost all aspects of life. They will make almost every process more efficient and reduce errors and mistakes. That means fewer employees working and checking. That saves a bunch of money.

So, AI is not going away. It’s a bubble like the dot.com bubble because speculators are buying EVERYTHING hoping that one of those purchases will be the next Microsoft or Amazon. That stuff will correct. But the infrastructure and software will evolve and it is going to be huge.

Remember: The real money in the gold rush era was made by the people who sold shovels and jeans—not the gold miners; most of them went bust.

[albie]

Oh. The LA Slimes. Never mind

[I want the USA back]

I don’t think so. AI isn’t ready for the trash bin. It’s imperfect, and always will be, but it will be put to use in many sectors — even though it’s not perfect.

[BereanBrain]

I use AI every day and I can do in 2 days what would have taken me a month.

[butlerweave]

AI will have a nervous breakdown and commit suicide

[Dr. Sivana]

AI is just getting started. It will take on many names and be here in some form for the foreseeable future.

Yes. A lot of it will NOT be general purpose AIs like ChatGPT, but medical-specific, robotics, business intelligence, software development, etc. The need will come, and then the AI will be developed (adapting existing AIs) to meet that need. Primitive versions of this were called "Expert Systems". Like computers and the Internet itself, it will need its own infrastructure for numerous tasks. It will not make human employees obsolete overnight, but will be a useful tool as needed, when needed.

[Mariner]

Failure or Abandonment Rates:

Estimates suggest 70-85% of AI projects fail.

RAND Corporation reports a failure rate exceeding 80% for AI projects, twice that of traditional IT projects.

One Gartner report suggests an even higher failure rate of 85% for AI projects.

[E. Pluribus Unum]

Estimates suggest 70-85% of AI projects fail.

That's pretty much on par with IT projects in general.

[9YearLurker]

I guess you bought the Y2K hype too.