Welcome to Inkbunny...

To view member-only content, create an account. ( Hide )

No, the AI bubble is not going to pop

by ThaPig

I keep hearing people say that AI is going to fizzle out and compare it to the dot-com bubble back in 2000. But that's a flawed analogy. When the dot-com bubble busted, the internet didn't disappear.

The dot-com crash was more of a Darwinian extinction event than a pop. When the fad started, everyone (including me) registered a dot-com domain expecting to become rich overnight. Many didn’t have a real plan; they thought it was as easy as putting up a web page and watching the money roll in. Naturally, most of them failed... but that doesn’t mean the concept of online commerce failed. The companies that survived, like Amazon and Google, became huge and pretty much rule the world now.

The same thing is going to happen with AI. Right now, a bunch of businesses are jumping on the bandwagon and investing billions into AI without much of a plan. There will be a natural purge, and many will crash and burn. But when things settle down, the ones that survive will become AI powerhouses. AI itself is not going anywhere.

I understand how terrified we all are about some of the consequences of artificial intelligence, but we live in the world that is, not the world we wish it was. A thing can’t be un-invented. We have to adapt to the new world it brings.

For good or for bad, Pandora’s box is open.

Viewed:	61 times
Added:	3 weeks, 1 day ago 01 Nov 2025 17:24 CET

22 Comments on this Journal

Leave a Comment...

Norithics

3 weeks, 1 day ago

link

The difference is, the Internet was actually useful and had good use cases for your every day person. AI sucks at pretty much everything any layperson would want it for. People are already bored of the imagegen slop, the deepfakes and pollution of the web is a continual thorn in everyone's side, products' AI "features" are hated by practically everyone because they don't work, it can't replace workers because it hallucinates massive errors that even the laziest person wouldn't, and the workers who use it become less productive because it's two "workers" who don't know what the other is doing. The only people who find any use out of it are scientists and they're the ones who were interested in the first place. AI just doesn't make money, and that's going to be the real problem with it being adapted anywhere.

ThaPig

3 weeks, 1 day ago

link

@ Norithics

The Internet itself began as something with little practical use beyond research and military applications. It didn't take long for people to find ways of making money out of it.
Trust me, greed finds a way.

Norithics

3 weeks, 1 day ago

link

@ ThaPig

Okay but people have been trying to apply the greed principle with as wide a brush as humanly possible and they haven't. Also, by the time the dot com bubble happened, there were already ways to use the internet to make money, they just got turned into monopolies over time. This thing loses money at a rate that is catastrophic relative to what it can do. You simply cannot sidestep the enormous costs and diminishing returns. It's not "early days" and it's not "going to keep getting better." These are all spin.

alistair

3 weeks, 1 day ago

link

@ Norithics

It is a power tool. Does a power drill make people money? Not if they're just randomly drilling holes in stuff. But if you use it for the right purpose, and have the skill, it is a massive force multiplier.

Norithics

3 weeks, 1 day ago

link

@ alistair

A drill doesn't randomly turn into a belt sander with no clear indication of why. It's unreliable, that's what makes it a bad "tool."

Waccoon

3 weeks, 1 day ago

link

@ Norithics

Damn... exactly how I think about AI. You're supposed to use tools, not the other way around.

Anything marketed as "smart" technology I avoid like the plague.

Issarlk

3 weeks, 1 day ago

link

Edited 3 weeks, 1 day ago by Owner

@ Norithics

I can tell you that AI is very valuable for me. As a computer scientist, AI often point me in the right direction to configure software ; even write scripts. As another comment says, it's a tool - so far - and only those who know what they want to do benefit from it.

For example I would never let an AI generate a whole application, I would architect it then ask for it to write specific functions. If you use it without knowing what you do, you end up with vulnerable apps or unmaintainable ones.

AI is a force multiplier as said above.

Norithics

3 weeks, 1 day ago

link

@ Issarlk

Part of the problem is that everyone started using the term "AI" to mean like 50 different things, some of which had already existed well beforehand, which makes arguments about it interminable.

ThaPig

3 weeks, 1 day ago

link

@ Norithics

Very true! The term AI is thrown around to mean many different things, from machine learning and data analysis to simple automation scripts and chatbots. Some of those things are new, and others have been around for a long time. Cell phones have had features like predictive text, voice recognition, and photo enhancement algorithms that learned from user behavior, long before AI became a buzzword. More than twenty years ago, Photoshop already included tools like auto-correct, smart selection, and content-aware fill that relied on early forms of what we now call artificial intelligence.

MaxDeGroot

2 weeks, 2 days ago

link

@ Norithics

I have to disagree here. A.I. has come a long way since 2021 when just trying to get it to make a figure with the same number of fingers on both hands was a chore if not an impossibility, to 2025 when it is making deepfakes of actual people that are nearly indistinguishable from reality if done right.

I guarantee you within five years a small handful of people will put out a full feature film, because A.I. is a TOOL, not a replacement.

Norithics

2 weeks, 2 days ago

link

@ MaxDeGroot

I guarantee you they won't. They've pumped entire national budgets worth of money into this and every revision's improvements become more and more marginal. As it stands, they're already asking the US Government to bankroll their next round of valuation scam because there literally isn't enough venture capital in the world to fund it.

Look, A.I. is great demo tech, you can eventually come up with something that looks great in a demonstration, but the user– and by 'user' I do also mean corporations trying to use it to replace workers– will struggle to get anything usable. It doesn't save money, it doesn't make people more efficient (the opposite, in fact), and the math isn't working. And without those extremely expensive data centers, the tech doesn't noticeably improve.

SpoonFox

3 weeks, 1 day ago

link

I believe the "AI Bubble popping" refers to the investments into it, not the technology itself. AI itself has a roughly 65% hatred by consumers who are -forced- to have it installed on their stuff (Reports between 60% and 70% based on which tech site you use, I just averaged it). People don't like when things are forced on them. It's like when Cortana became a thing on windows and a lot of people just buggered it off. So it's garnering a reputation as a form of software that people don't like because it's forced on them. There are also those who hate it because it's replacing their jobs, garnering more spite towards it. Overall, it's got a negative outlook. The only reason businesses are investing heavily is that they can spend $500 on AI in exchange for what they could be paying $50,000 for people to do. Just like machines replacing factory workers, we still have factory jobs. Machines aren't perfect and need to be maintained, AI is the same way. AI is also way more vulnerable than humans...

Imagine the news, "AI leaks corporate secrets to another company. AI Company tells the corporation that they should've read the ToS and the corporation cannot sue." XD

Issarlk

3 weeks, 1 day ago

link

Edited 3 weeks, 1 day ago by Owner

I think AI will plateau. I read an article about a prominent AI researcher leaving Large Language Models aside to try other stuff.
There's no certainty that LLM will ever manifest AGI levels of intelligence or common sense, or grasping meaning of what it's doing. After all it's only putting word together according to probability ; the fact that it's managing to do what it does is amazing emergent behavior but it might never get better.

If AI needs "Large Knowledge Models" or "Large Understanding Model" to become really useful, then we might be looking at years of stagnation.

ThaPig

1 week, 2 days ago

link

@ Issarlk

I was told that Language Models are just an early stage in AI evolution, the next will be Word Models. Instead of just playing with text pattern recognition to get the right sentence, it would simulate the mechanics of the problem to find the answer.

Issarlk

1 week, 2 days ago

link

@ ThaPig

Did you mean World model ?

ThaPig

1 week, 1 day ago

link

@ Issarlk

Yes, I misspelled that. World Models is the right term

Rakaziel

3 weeks, 1 day ago

link

The problem with the current approach to AI is that more complex AI both requires more power and hallucinates more. And this is not even getting into the evolution towards lying and self-preservation which, while fascinating in its parallels to nature, is not something you want in a tool.

IMO it is currently at the stage where it will crash and then get picked up again a few decades later when someone has figured out the solution for the energy efficiency problem.

Waccoon

3 weeks, 1 day ago

link

@ Rakaziel

Poor reliability and unpredictability are exactly why I don't consider AI to be a tool. A power drill is a tool. AI is a toy.

It's the same reason I refuse to use any software tied to the Cloud. Any day the owners can push out an update or change their policy that breaks everything, removes features, or otherwise yanks the rug out from under your feet.

Waccoon

3 weeks, 1 day ago

link

Yes, the bubble will pop. It won't kill the tech, but the economic fallout will be very real.

That is the very nature of disruptive technology, and has been true since the early days of the industrial revolution. Long term, things will settle down, but in the short term, we're in for a world of hurt, and the tech industry has no plan (or desire for one) as to what to do about it.

ThaPig

3 weeks ago

link

Edited 3 weeks ago by Owner

@ Waccoon

That's what I said, things will settle down, not crash as some expect. Some of the most practical elements of AI will remain, while most of the useless crap will fade away.

And yeah, most probably whatever new order results from that will be controlled by assholes. But that has nothing to do with AI; that has always been the nature of the human animal. ( ᐢ (oo) ᐢ )

Phantasma

1 week, 2 days ago

link

Edited 1 week, 2 days ago by Owner

The Problem with A.I. Slop!

https://youtu.be/vrTrOCQZoQE

This video reminds me of a comic of yours about a civilization inside a virtual world where the information becomes corrupted/unreliable with time.

ThaPig

1 week, 2 days ago

link

@ Phantasma

Yeah, I was trying to predict the future. Everyone connecting through social media while the planet is on lockdown. I guess I didn't get it entirely wrong.
I also started the storyline of the planet where people go crazy about AI and end up turning to religious fanatics. But now I'm kind of undecided about finishing it, because people are going to think I'm making a commentary about the present day, when in reality I started that story 16 years ago.

New Comment:

Move reply box to top ↑

Log in or create an account to comment.

Help

Information

Policy & Legal

All artwork and other content is copyright its respective owners.

Powered by Harmony 'Gravitation' Release 80.

Content Server: Los Angeles Cache - provided by Inkbunny Donors. Background: Blank Gray.

The Inkbunny web application, artwork, name and logo are copyright and trademark of their respective owners.