We do a lot of software development. We are adopting AI in a big way. It's changing the way I work. I'm old and I remember when I first saw the Internet and thought it was no big deal. Well, AI is similar. It's going to change everything. I cant speak to stocks, but in my experience AI is already making big changes.

However, after the dot com bubble burst, we still had an evolving internet that has changed so much in our economics and culture. AI will likely follow a similar trajectory.

AI stocks will collapse, bringing the stock market down steeply. But that does not mean AI is not real. It will indeed change much in our economy and culture.

Personally, at present, I can’t stand it. It degrades the quality of the internet with multitudes of fakes. But it will be worked out, it will find its place and evolve.

(Retired IT here. My old firm is integrating AI tools already. I kinda wish I was still around. Sounds like it will be interesting (for those in IT implementing AI tools)).

AI companies are selling the service and associated hardware. But they are loaning money to the purchasers to buy their product,

The purchasers will have to make/save enough money from the purchasers to repay the A.I. companies AND make a profit off of the product. One estimate is that it will take decades for all of this to shake out.

The dotcom bubble was a bunch of no name startup companies that had money thrown at them. AI may end up being a bubble, but the company's involved have huge capitalization and long, long track records.

I can absolutely see the stock bubble bursting a bit, but I don't imagine companies like Nvidia will somehow collapse. Not all of their eggs are in a consumer software basket the way early Internet companies were.

And AI is absolutely here to stay. It encroaches on new aspects of technology, the economy, and life in general on the daily.

Three years ago, professors were complaining their students used AI.

Now they use AI to grade papers, create their syllabi, and . . . detect AI in student papers.

Three years. That's about how long protesting it lasted before blatant co-adoption started setting in.

I've been in IT for 31 years, in both public and private sectors, both large and small.

Many organizations are desperate to integrate it into their enterprise environments but struggle to find a use case that makes sense and worth the expense.

Many organizations have implemented it and gone all in and even laid people off. The problem is it's not panning out for many. It's becoming a money pit.
I think if it's an internal AI, using carefully curated internal data and carefully trained in internal processes, it would have some utility but short of that it's a money pit that will never produce anywhere near the value needed to offset the costs and trouble involved.

More and more nursing schools do. Just google Sherpath AI.

And let the implications of that sink in.

...and neither myself nor anyone i knew thought it was """no big deal."""

I’ve never been able to understand how something that cannot be owned or controlled (access to prime numbers) can be the basis of anything secure.

I read here never seem to come true

Burry was correct on the last large scale bubble, timing was a little off

ORCL down down 15% already

all are bad news for the USA's economy, besides tariffs, job loses and health insurance subsides.

Musk and his minions alongwith Tech screwed up.

But hey, insant wealth for the greedy.

Last edited Mon Dec 15, 2025, 04:32 PM - Edit history (1)

bust in the late-1990's to early-2000. It caved as a result of this over-promising without the ability, at the time, to deliver. By 2010, it was not only back, it was established and is now, in 2025-6, THE Establishment.

Like the dot.com's of circa-1999, when the infrastructure required was not yet built and engineered, the AI "Revolution" is not quite ready for prime-time and is somewhat overpromised. Ten years from now, even after it crashes economically now, it will be back and taken for granted as the way things are.

There is definitely an "AI bubble." But AI as a tech isn't going away, anymore than internet commerce went away with the popping of the dot-com bubble.

It's tempting to to just get out of the S&P500, but no one knows when the bubble will pop or the where the bottom will be. I think trying to time the market is a fool's errand. There was a firm that predicted the dot-com bubble and got out of the market in early 1999. But the market went up massively after they exited the market, and bottomed out just slightly lower than their exit position. So they were right.... but what good did it do them? But hey, if you think you can time it and come out ahead, go for it. I'll watch with interest. So yeah, a correction is coming. How big and when? Belly up to the casino table...

New technologies such as the internet and AI are rapidly reshaping the reality of modern life. Acceptance and adaptation are essential.

Whatever else it is.

A lot of it has no utility either. It’s just promoted bullishly.

And a lot of it is downright distasteful. Like those AI mockups of people who have passed.

I thought this morning of how elleng is not with us anymore. Nothing would be creepier than a horrible AI company making mockups of In Memoriam DUers beloved by the community ..
Ewww…that would be so gross.

I would be impressed by an AI that was not built using stolen human output and ..well I’d have to think more about it. But it would be very different from the stuff we see. It’s not so much replacing real human labor (that’s the canard) as getting rid of jobs and services that are essential or useful for ones of no value whatsoever.

I too have no objection when I hear about AI that finds new protein structures or things humans cannot possibly do and which legitimately supplement human effort rather than shilling disgraceful things like the dead peopel thing (rereading this I realized I paraphrased Chuck Palahniuk’s Marla ..I bet that would be a legal argument those creeps would try. Humans themselves use art etc without attribution. No we really don’t. Try again. Lately I want to attribute any piece of writing to its sources when I can remember.
I am not wildly original but nor am I mindless.
I think about creeps way too much :eyes .

And yet that is totally the type of creepy “non-service” they shill and the type of legal argument they will use if a real pushback starts. A lot of what those industries shill makes no sense except in WallE world. I think it’s worse for artists and I totally back serious pushback rather than Google or Facebook approved pushback which would be neutered and tamed.

If they actually ever offered useful stuff I’d be shocked. I think we should distinguish between good AI and bad AI. And even good AI that steals human or animal learning by encroaching into spaces it had or has no business being in is outrageous when so much of earth is endangered and most humans are suffering.

AI should have built in politics to not suck and it can’t be outdated Asimov’s laws. If AI could parse genuine consent and target info in sane ways rather than in oligarch friendly ways, then you have closer to AI that represents the populace or wild spaces than an atrocious disgrace.

Anytime I have second thoughts it’s when I see how it’s all oligarch driven and think of how different it would be if the kinds of scientists I respect worked on it and it had a place in society in the best way not the obscenities we see. After all human intelligence is different when it is yours or mine versus Trump, Zuckerberg etc etc.

Yan LeCun is one AI scientist who made me less reflexively hate filled for the field and he would be a “radical leftist” in America whereas he is merely sane. His focus on common sense should be extended beyond stuff like not bumping into obstacles physically. They should have some comprehension of how humans think.
In this ott world how is finding sleazy and lame ways to repackage the status quo novel. If anything novelty lies in the parts of society Musk etc want to delete.

Present ai fawns because it seems trained on hostages (ie the modern employee who has no rights whatsoever often) because apparently anything else is unstable and violent etc.

At a time of stark and increasing inequality when people resent having to feel grateful for the smallest bits of relief, outlawing anger etc except when it sells viral videos and becomes “safe and profitable” yet again..that’s underestimating human intelligence when it doesn’t fit what’s convenient .

Recently I have started feeling how pointless anger is as much as agreeability when everything somehow gets tamed and goes back into the same safe for them (oligarchs and CEOs) but not for us model.

I myself have pretty scorched earth views about tech companies by now. I wish I could organize with people like some of you guys here or people like Yasha Levine to be part of some broader pushback without attacking scientists and doctors inadvertently, but definitely stopping this fraudulent scourge.

I think I was in the early wave of “life destroyed by AI”. But I am very wary of self-commodification or astroturf like that traffic snarling “artist” Google loves. An artist like Levine’s wife Evgenia I would trust, Simon Weckert not so much.

It’s also pretty much the “optimization problem” to use gross CS speak that DU discusses all the time in a way-electability balancing other stuff. That’s not how DU or grassroots movements work whatever internal squabbling is a hallmark of human groups. But a neoliberal or worse effective altruist prediction scheme would see it that way rather than as a collection of complicated humans each making assessments in ways that you can’t reduce to something as shallow as IBM’s Ocean score. That’s so unscientific and fluffy.

I’ll give you that, but the old quote about solvency and rationality still apply here. We don’t know if these pullbacks are part of a more bearish outlook or just jitters. If rate cuts keep happening, the party may continue (may be an even nastier hangover).

Some companies will survive and likely thrive well into the future. I also wouldn’t conflate AI with crypto. I’m guessing we’ll see nvidia and Alphabet in 5-10 years. These are companies that whether or not people here like them or their products or practices, produce some value (in the eyes of shareholders) and sell actual goods and services. AI itself will continue to move forward in some way or the other. LLMs will evolve. The companies may change, but that’s capitalism. Crypto “treasuries”? I wouldn’t bet on it. They produce absolutely nothing of value.

People should invest according to their own risk tolerance.

I have no reason to sell my semi-conductor holdings. Not only would it be a major tax burden to take on in the waning days of 2025, it contradicts the experts, and ignores the current semi-conductor demands.

but when.

When I was in college in the 70's, there was a computing course offered on Artificial Intelligence Systems.
Modern forms of AI: Siri and Alexa. Some aspects of Google Translate.

https://en.wikipedia.org/wiki/Progress_in_artificial_intelligence

AI does seem to be more "in your face" these days and more bots are available for the general public. (Such as AI summary for searches, widely available apps and deep fake videos, which anyone can make these days) But I think much of the work has been going on behind the scenes for decades.

Also, the newest name for Photoshop is "AI". If someone is airbrushed, it's called "AI".

You said:

Blur, spot repair are the photo editing term coined as "airbrush", but everyone just called blur, blemish removal etc"airbrush".

Nowadays, when I see an enhanced image it's just called AI. It doesn't matter whether traditional photo editing software was used or not. If we see a poster for a politician and it looks enhanced, it's just called "AI". No mention of what process was actually used to remove that wrinkle!

to take advantage of the current AI craze.

First, unloading AI related stocks now because "five or ten" years from now it might not be a good investment is really bad advice.
Second, thinking AI is a bogus pyramid scheme is bizarre.

Most here are “permabears”.

I wish I ignored the noise a long time ago. I would have made much more money. But I’ll admit the fear uncertainty and doubt seeped into my subconscious.

but I don't completely agree about A.I..
I don't think A.I. has lived up to the hype and there are some ways that it has been utilized by some. I'm not a fan, but I think it has it's uses that can benefit humans as long as humans control it and not the other way around. I'm also not sure the benefits outweigh the costs and environmental impact.

It is a huge productivity boost. It saves a ton of time wasted searching for a good answer and then it can help refine and fine tune that answer into a precise, robust professional result. I use it every day.

I don’t disagree there will be huge market fluctuations though. I do plan to invest in Anthropic when they go public as I use their solution on a daily basis and they are actually placing guard rails and alignment technology responsibly as they improve the technology. I recommend you watch the interview with the CEO and members of the company on 60 minutes.

I don’t disagree with your take on bitcoin/digital currencies though. That is a scam and it’s used by criminals both known and unknown. Trump is literally accepting bribes using it.

Must be nice.

People think the bubble, if it is a bubble, will stop AI.
It won't.

So, regardless of however necessary it may be, I am not a fan.

But, I won't be dealing with the ramifications of it either, so who cares what I think, right?

Soon I'll be able to afford both Boardwalk and Park Place.

The AI badge pisses me off just like "the cloud" BS is just a new marketing name for the internet (renamed also the "Internet of Things" ). Its all marketing. "AI" most people interact with is just a sooped up database search engine. "AI" thats beneficial say in medical research has always been there. Its just faster as the hardware slowly catches up with the software. Also more people have learned how to work with software and hardware now. I mean its really only been around for 30 years within the general publics grasp. Think of the advancements in the automobile 30 years after it became widely available. You had "hackers" building street rods. In the next 30 years this will be seen as the same thing.

You are correct though that a massive correction of the AI grift is coming. These billionaire CEO's and boards are marketing the shit out of something most investors dont understand. Hell their already pivoting on the brand from AI to the new buzzword, AI Compute lmao. If you jumped in early and made yourself a wad of cash, good for you. But if you hang around with that gang, like all the market grifters through history, they will end up with your money while you cry. How much is enough? Its certainly to risky of a bet to jump in now IMO.

And it all has to be changed to some "world model" or some shit. The whole thing is bogus anything it does can be done better and more accurately by 'hand'.

Large Language Models (LLMs) can be thought of as giant dialog machines - break a conversation (website, PDF, patterns in an image, etc.) down into individual words, with arrows connecting each word to the next. Put it into the model. Repeat over hundreds of millions of documents. Eventually, you end up with a giant network graph of overlapping conversations. When you enter a prompt, this selects the conversation(s) closest to the one you want that has more details.

There are a few big problems with this - the first being that the shorter the prompt, the more likely the conversation will go down the wrong path and generate something that is entirely nonsensical. The second is that there is no original thought there, only what has been fed into the latent space (the graph), beyond any new concepts coming in from the prompt. This can give the illusion of intelligence. The final is that because the training is such a time-consuming process, most LLMs use a special kind of memory device called a context, and there are limits computationally to how big such contexts can be before they become ineffective.

A world model, on the other hand, is more like a traditional database - it tells the system some information about the world - physical infrastructure, where things are located, changes in organisations, and so forth. It's also called a digital twin, changes in state, etc.. This is preferable than an LLM by itself (though LLMs can us world models) because the WM can be updated dynamically quickly.

A number of key data scientists, including ex-Meta chief scientist Yann LeCun, Gary Marcus, and a number of others (myself included) have all been saying for some time that you cannot have a reasoning system without a world model as something to ground it, and this is beginning to percolate through the AI industry. The Tech Bros aren't particularly happy about it, because such WMs are comparatively far more efficient and work in ways that tend to reduce the importance of their components, but I think that on the tech side as people work with the tech, they're coming to the same conclusion.

Cha-CHING!!!

It's still going. New industries live on capitalism's life cycle, from ideas pitched to venture capital through IPO to growth, consolidation, and more growth. The shares split and split - Oracle split 10 times from 1987 to about 2005. A $3K investment is worth north of $17,000,000 today. And it's that way with the other big companies we know - Amazon, Microsoft, Meta, Google, and more. It's the game that Wall Street doesn't want the public to know about - because they keep the secret hiding in plain sight for themselves. If you work 10 years and bang away investments in the latest growing technologies, you can retire at 30. Bigger return than most any college degree.

If you invested $4.08 (average ticket price when it was released) in Apple stock (0.19 / share at the time) and held ( all the splits, reinvested dividends, up and downs) it would be worth over $10,000 today.

That’s compounding interest for you.

My partner was messing with retirement things, and tech friends were telling us the boom was coming. Since I'm a gamer, I've been using Nvidia products for years, so I told my partner to go ahead (based on my video gaming knowledge, lol - I know nothing). But I know the cards are basically industry standard.

It's . . . going well.

I actually expect this valuation to readjust, but we're middle-aged and looking more towards 15-20 years until retirement. So, for now, whatever happens happens. Not going to panic sell if the bubble bursts. As long as they make hardware, it'll probably be fine long term.

My biggest regret is not buying the stock about 8 years ago when I knew they were in the forefront of pretty much everything. But I’ll admit I was kind of a chickenshit when it came to investing back then.

We both grew up poor and tend to be frugal people generally. During the pandemic, we went through retirement plans and firmed up the foundational baseline 401k, Roth IRA, mutuals, etc and then started setting aside "It's not disastrous if we lose it" retirement money. Not crazy amounts, just a portion that could be subjected to higher risk.

I am not the retirement planning person in the house, so I feel like I can be talked into anything. I truly have no idea, and that's probably irresponsible.

It took me a long time to come around to the idea that even a home could be an investment. We finally bought in 2021 when interest rates were rock bottom.

I always maxed out my 401k but beyond that I had just kept cash until I started paying attention to the current boom. Trying to learn how to manage and hedge risk if we do see the bubble pop.

Became an adult as the DotCom bubble burst and then lost my job in 2008 in that crash, so I've always been skittish about retirement. Just insanely risk averse.

We've had to be coaxed out of it by friends in the last five or so years, for lack of a better word, lol.

We bought our home in the Bay Area six years ago with the crazy low rates as well. It's appreciated about 30% in that time. But we're kind of now in this place where we keep discussing downsizing. Our mortgage is totally fine right now as is. We really just feel like it's too much house for two people (seriously tired of cleaning this place). But the current higher rates are staying our hands on the situation.

Our current attitude is a little, "Let's revisit this in maybe three to five years." I guess I'll eventually buy a Roomba or something.

I think what is commonly called AI (mainly transformer technology) has significant structural issues but there's a fair amount of work going on to mitigate these (and find use cases where such AI is appropriate). However, right now what we're calling the AI Bubble is mainly a data centre and server bubble, and that one definitely is ready to pop. Think of AI not as a tool but as a justification for massive buildout of data centres (mostly on other people's dimes). AI adoption, outside of a few specific areas (code generation, media generation) is NOT getting a wide adoption, and even there, you're seeing pushback from people who were initially gung ho about AI and have become disillusioned. Yet because of the growing anti-AI swell, the technology issues, and the increasing legal minefield, AI is not seeing broad adoption. This means that we're building data centres with servers (mainly nVidia) that I have a shelf-life of only three years, and a growing number of them are becoming stranded even before they are completed. That's where the AI bubble is getting tenuous.

A volatile, unstable, wishful 'currency' with no assets or gov't support backing it.

The big money return in crypto is:::
1. Starting it, hyping it, selling it, and getting out, leaving the
suckers holding the bag.
2. Handling/storing it for others.
...a. Not your own money/crypto. Someone else's money/crypto.
...b. Lots of fraud/money manipulation/theft/lack of regulation.
3. Backed by trust/faith/great promises/optimism... but no
real assets.
4. It can literally disappear and good luck with lawsuits.
------------------
-------------------------------------------
A few have profited by 'investing' but not the majority.

Bitcoin is one of the very, very few that has made profit.
And that is only because people WANT to believe in it.
Even bitcoin has no real assets backing it.

BC is used as a fantastic example of crypto but it is a lucky outlier.

The problem is it's nowhere close to being profitable, and no one even had a business plan leading to profitability. Just maintaining the hardware and power requirements takes like 20 times what it's generating now in revenue. Companies are just blindly throwing investor money at it in the hopes there will be an unforeseen miraculous breakthrough that will somehow be stupendously valuable. Way more valuable than cheaply providing a mediocre replacement for human workers in certain fields, which is pretty much all its achieved so far.

I don't agree that AI cannot create anything new. Most ideas or things that we call 'new' are simply combinations of ideas or things that already exist. Every invention is a combination of a number of different existing technologies. Humans are good at finding those combinations, which is why we have a continuing flow of innovations. But AI will be immeasurably better, because, unlike the human mind, which has limited storage, AI will have almost unlimited knowledge, and can bring together combinations of ideas far beyond what humans can. So AI has enormous potential, though I agree that we are still far from that, and what we have now is an AI bubble that may soon burst.

spot on and well said. 👏

AI is just getting more powerful.

I've been a Gemini user since introduction and in a year it's improved vastly.

This is just the beginning

ChatGPT 5.2 has also improved but sometimes its tone is a little patronizing at times. Perhaps due to guardrails, it can sound a bit too much like it’s your therapist. Reddit has some interesting threads observing how Em dashes aren’t used as often. Instead it relies on a lot of short sentences. It’s odd, but depending on the subject can be a very good reference tool.

I have also used LLMs for small coding tasks and excel macro code. It saved me a lot of time. It beats going through multiple stack exchange discussions, which often didn’t answer my questions. Does it make world class, ready to ship software with efficient code? Probably not, but for my purposes it often worked.

As long as one knows these tools’ limits they can be useful. I don’t use them for image generation, so I have no idea how they compare.

ME:: what is the current hallucination rate for SOTA AI

"SOTA AI hallucination rates vary significantly but are generally low (1-5%) for top models like GPT-4o in standard tasks, yet can jump dramatically (10-30%+) in complex domains (legal, medical) or with newer reasoning systems, with some models showing extremely high rates (even 94% on specific tests). The "hallucination paradox" suggests advanced reasoning might increase errors in certain complex scenarios, with rates often depending heavily on the benchmark, task complexity, and model version."

Would certainly not talk law or medicine.

I was recently watching the BBC series about William the Conqueror and 1066... and because I tend to nerd out, asked lots of questions about tertiary relationships to the prominent individuals.

It was dead on accurate.

Applies to both AI and crypto.

We're heading to an Idiocracy, not because of the fecundity of idiots, but because so many people are eagerly slurping up whatever slop their apps are barfing up on their devices.

The companies involved in the investments are not the xyz.com airy brain farts of the dotcom era. They are major players with high capitalization like Nvidia, Alphabet, Meta, and Microsoft.

Also, unlike the dotcom phenomenon, everyone can see for themselves how impactful AI is.

The issue is not whether productivity is going to be massively increased by AI. It is. The question is whether we as Democrats and liberals in general have the foresight to make sure that all people benefit.

... but created a flood of it instead?

AI is going to be worse than that.

The only AI that actually benefits society will be tightly focused in purpose, and containable within a single machine

Most of the "product" produced by these giant server "farms" will be hazardous waste, corroding the structures that hold our democratic society and communities together.

I don't want to live in a fucking surveillance economy and police state.

Automobile culture was bad enough, restricting our actual freedoms in ways most people do not perceive. I am not "free" if I am forced to own and maintain an automobile under threat of extreme poverty. What's coming with AI is worse.

I'm going to be angry when I'm forced to own a fucking "smart" phone simply to fully participate in this sick society. I've always resented the fact that I must own a car to be considered a fully functional adult.

Here's what's coming: This AI bubble will crash. Our government will bail it out by creating the most powerful surveillance machine that's ever existed.

Big Brother will be watching you, will know everything about you -- where you are, what you purchase, even when and where you piss and shit. You will be guided through life by helpful cheerful voices, and opportunities to develop your own voice will be increasingly restricted.

Most people won't even know they are slaves.

Yes... AI will be impactful... no question. But it's absolutely true that revenues are inflating by circular investments and REAL revenues are WAY lower than needed to support capital investment. There will be a reckoning, and not every player will survive.

10 - 30% hallucinations on complex task is a no go

Simple task can be done with old tech stack for no additional cost

Useless calculations require enormous amounts of energy, cooking our planet and sucking up our groundwater — all for nothing.