AI boom/bust at least as big as late 90's Internet bubble

intercst

3 Likes

Very different but possibly as big a folly.

The automatic chess board in the 1700s keeps coming to mind. Just because you can get a machine to play tricks does not mean in the long run the public will not catch up with how hollowed out the reality of it is.

MidJourney is making hundreds of millions of dollars. But the liability is massive. The TOU, terms of use, state if an artist/user infringes the company will sue the artist. I doubt it could ever pan out that way. The AI generator by MidJourney does more to create the art than an artist does. The artists do nothing to generate the art.

While MidJourney is making a lot of money per subscriber/artist, the artists are not making much money at all. Art sales are very hard to come by.

Other aspects for the companies developing AI inhouse are lame. Facebook’s ad manager rejects many ads automatically with AI. It is like a receptionist telling a visitor to be seated until a human being is available to see your ad for approval. That was never worth the cost of creating AI. It could have been equally worked as well with much less code or complexity.

I don’t believe the current AI boom is equivalent to the beginnings of the dot-com Internet bubble for the reason that AI is currently being used and making money doing real things. And this is in stark contrast to 1998, 1999. Are there going to be AI failings? Of course. But I really don’t think the fall is going to be comparable to the dot-com bubble. This field has already reached some level of maturity that we never had in '99. That has to account for something.

1 Like

Mixed about that. You would not say Amazon or Google were not mature in 1999 but Amazon was not making money.

Tech historically has a lot of failures for every success. This won’t be any different so we agree. Some dot coms made it just like Intel and MSFT made it at one time. Before that IBM and almost no one made it.

The problem while the AI companies are making money where are the economics for the users? AI won’t have users in the long run if people do not make money with it outside of who is building it.

IBM and then Microsoft had the spreadsheet and more.

Nvidia chips use AI as a component. That is the success a lot of people use.

FSD cars would be next. If it happens. This is where it all could be folly.

In closed systems AI is excellent. Like playing chess. In open systems like road hazards etc…a lot is left to be desired.

Not claiming any expertise but I am claiming a lot of observing. And some training in stats. The end results of AI systems use statistical models.

Classical odds AI works. Other open-ended risks? Not so much.

Classical odds is not how we live our lives.

I have lost all hope in FSD.

1 Like

What is your take?

You realize you are probably dooming humanoid robots as well.

I am more worried about folly at this point. It is also a defensive investing position.

Well, yeah, I’m not putting a lot of stock in humanoid robots either. But for FSD, seriously, how long has this been hyped? Why has Musk not been sued yet over all the pre-sales which have yet to materialize? 5+ years ago I was told about RoboTaxi, not here yet. Same is true for every other manufacturer, with GM’s system possibly the best out there and still no where close to Level 4, let alone 5. I’ve given up.

Humanoid robots? So much has to happen there, and a lot of it feels like the challenges to FSD. So…

2 Likes

Classical odds versus genuine risks.

Genuine risks in business are financial risks, business risks, and economic risks. You never set up classical odds to invest in a corporation.

You can but it is not a good idea.

Road risks? There are weather risks, road condition risks, traffic risks, human and animal risks, and mechanical failure. Then the bigger risk the onboard digital systems (hardware) are not good enough. The software and AI are not good enough. Against the big killer speeds on the road.

Just getting the onboard systems to talk to each other is a mess. Computer components are imperfect at that.

The Internet was making money and doing real things in the late 1990’s, but they were also funding a lot of stuff like Pet.com.

AI is going to make investors a lot of money, likely much more than Internet 1.0. But in a frenzy, we’re going to see a lot of AI-Pet.com, too. Just don’t be surprised to see a 50%+ market drop at some point. It will recover like it always does, but as a retiree, you need to survive the trough.

intercst

3 Likes

I do not know if MidJourney is public or private. Saw a report from last year $200 million in revenue with 40 employees. In fact some antivirus companies have probably done that as well.

The problem comes with artists figure out they can not make their money back. LOL MidJourney is not the most cost friendly software.

https://youtu.be/CcGS-gD2Gho?si=NvVEeogRqf7YI7AA

Programming note:

If you are a subscriber of streaming service “Curiosity Stream”, I recommend “iHuman”, a 2 hour documentary about AI which threads many of the thoughts about the future together in a compelling (and occasionally over-the-top) way.

The doc is on the “leaving soon” list, so watch it sooner rather than later if you care to. While much of it trods things we already know, the way it threads them together, along with presenting new (to me, at least) predictions and concepts was compelling.

It is, according to Google, also available free on Tubi, or with premium subscriptions on YouTube, Sling, or Roku. Highly recommended.

It’s hard to remember, and perhaps I will watch again and take notes, but there are segments on “facial profiling” (not a misspelling: predicting bad actors by their facial features, ala phrenology), showing how AI is being directed/corrupted by corporate actors, involvement by the CIA and/or repressive governments for population and political control, and more. Worth your time if you think AI is going to be big. (Hint: it is.)

2 Likes

I think Apple did too. I don’t think the Apple car was ever more than a platform to develop automobile tech that could either be branded or licensed. When the board realized how far into the future the payoff for anything was, they pulled the plug.

4 Likes

Exactly. Which companies are the ones in the AI bubble? We all know about NVDA but its’ earnings are surging. MSFT is making good use of AI in its copilot assistant, which it charges for, and which companies will buy because they can reduce headcount. In the future I can see call centers being fully staffed by AI bots.

I don’t think that was the primary issue. The primary issue is that Apple realized that high gross margins were never going to happen in the auto market. The Tesla high gross margin period was fleetingly short, and even those margins didn’t reach the threshold that Apple prefers. They essentially realized that there would never be a sufficient payoff.

1 Like

A friend who is a Ph.D. in aeronautical engineering does a lot of complex computer coding. He believes a perceptual level needs to be added to today’s AI to make FSD work well.