Is the shine already coming off ai & data centers?

[mstateglfr]

Why arent there bombings of data centers and related buildings in the US? I am not absolutely not advocating for this, I just find it interesting that it isnt happening at all.

One of the largest Microsoft Azure centers in the world is in my town like 5mi from my house, another large Microsoft center is in my town a few miles away, and a massive Microsoft center that is in the early construction stages is on the edge of town 10mi away. Apple has one in a neighboring town, Facebook has one in my metro, and there are a handful of smaller ones scattered around the metro's edge.

Seeing one a week ago got me thinking about all the articles and threads I read about how furious people are across the country that data centers are being built in their town/land/neighborhood.
But violent destruction of these centers doesn't exist. There aren't groups bombing and destroying data centers.
Drones readily exist. Explosives readily exist. It could be done from afar since local surveillance is high.
...yet it doesn't happen.

- are people just not angry enough yet?
- is there something about Americans in general that keeps us from going to the 'nuclear' option(doubtful)?
- do people who would likely do such destruction realize they benefit from the center existing and are really just NIMBY?


I would think that with all the crazies in this country, and all the outrage over centers being built and the perceived/real downsides to them, a crazy or a group of crazies would have struck by now.

 

[dorndawg]

Why do people give a ****? I mean, if they literally have their house next to the data center, sure, that's not ideal just like it's not ideal to be next door to any industrial and most commercial activities. But outside of the additional trucking during construction and a temporary boost in the cost of employees and probably some local goods during construction, it just doesn't seem to be that big of a deal unless people are scare mongering about water or electricity rates.

I've ready the noise can be really unpleasant, even pretty far away. I've never been especially close to one so I do not know this for certain.

 

[johnson86-1]

Often it’s the regular folks. The smartest experts in the auto industry just lost 10s if not 100s of billions rolling out failed EV initiatives that they’ve now cancelled. I could have told them it was a bad idea 10 years ago.

That was a little different because that was a politically driven push, not experts really thinking EVs were going to win on their own merits.

It's very possible that AI proponents are early and this is going to be more like the dot com bubble. Too much investment too early in technology that people aren't able to figure out how to monetize quickly enough for the early investments to payoff. But we'll look back in a decade or two and realize it really was transformative.

I'm sort of predicting in between. I think AI is oging to be useful much faster, but that the current investment is basically an arms race for a leading position based on the idea that they will reach something like a small scale singularity where the AI is self improving enough that nobody can catch up afterwards. I don't think LLMs are the path to get to that place, and you're going to end up with a lot of overinvestment, but that there won't be a point like after the dot com bubble where you have people like Paul Krugman calling it the equivalent of the fax machine.

 

[Maroon Eagle]

Since we’ve got another pre-shìțťìng thread going on, I’ll just be watching these arguments…

 

[johnson86-1]

Water,

I think this is being addressed. Companies don't want to spend billions of dollars on a facility and then be dependent on water that isn't there.

noise,

This is legitimate, but it's a reason for buffer property. My understanding is that Data center hum is basically equivalent to heavy traffic. Are they building these where there is not at least a 1,000 feet of buffer with trees and fencing between the data centers and residential?

eyesore,

I just don't get this one. Do data centers look worse than factories in general? I mean, a buffer with a screen of trees between data centers and residential seems like a prudent thing to require, but if you're talking about driving down the road or seeing them from other industrial properties, are they really worse than any typical factory or plant?

traffic

Real during the construction phase for sure, but one of the complaints is that they don't have enough full time jobs, so people are simultaneously complaining that they generate too much traffic and not enough traffic.

and perhaps most important, everything else negative that we simply don't know about yet. People complain about windmills on unused farmland that are right next to communication towers that have been there for decades with no complaints, so go figure.

 

[johnson86-1]

I've ready the noise can be really unpleasant, even pretty far away. I've never been especially close to one so I do not know this for certain.

Well, I googled how loud AI centers are and the AI generated answer said it was like heavy traffic. I don't know why you'd question that.**

 

[mstateglfr]

...it just doesn't seem to be that big of a deal unless people are scare mongering about water or electricity rates.

I think you are really downplaying the impact some people experience.

In my city, Microsoft is the largest user of water for like the last 6 or so years.
Microsoft is the largest user of water...and we have have multiple water usage bans in that time due to droughts and due to high level of contaminated water because my state has what seems like 0 regulation for Ag runoff.

- People are paying more for water that is less clean(we have a reverse osmosis unit in our house), people are restricted from using water at times, and data centers are chugging water without a care.

- People are paying more for their electricity, and meanwhile data centers require infrastructure to be laid and often get a break on the cost of electricity through low-cost tariffing, tax exemptions, and more.

- People are subjected to the environmental changes in light and noise. This can be dismissed as people just being little bitches, but it is real.

There is definitely scare(fear) mongering. There is also definitely legit and real concerns that you seem to not grasp.

 

[patdog]

That was a little different because that was a politically driven push, not experts really thinking EVs were going to win on their own merits.

It's very possible that AI proponents are early and this is going to be more like the dot com bubble. Too much investment too early in technology that people aren't able to figure out how to monetize quickly enough for the early investments to payoff. But we'll look back in a decade or two and realize it really was transformative.

I'm sort of predicting in between. I think AI is oging to be useful much faster, but that the current investment is basically an arms race for a leading position based on the idea that they will reach something like a small scale singularity where the AI is self improving enough that nobody can catch up afterwards. I don't think LLMs are the path to get to that place, and you're going to end up with a lot of overinvestment, but that there won't be a point like after the dot com bubble where you have people like Paul Krugman calling it the equivalent of the fax machine.

Pretty much agree. I've always been of the mind that when it comes to any kind of new technology you don't want to be on the leading edge. The occasional company may hit a grand slam, but most will go bankrupt. You don't want to be in the last half to adopt it either. The sweet spot is the second quartile. Let the early adopters make the mistakes and then get in right behind them. I do think AI in some form is here to stay. But I don't think it's going to reliably be able to do what a lot of people expect it to.

 

[ckDOG]

Just had this conversation at work. I leverage it. It makes me more effective and efficient. It's a no brainer investment at the current cost. But will it in the future? At some point the tech companies are going to get us hooked and then massively increase the price - like a drug dealer. Will we go all in at that point? Or push back and go back to more man power / less productivity bc the dollars no longer make sense?

 

[CochiseCowbell]

Possibly a little weathered, but still shiny.

 

[L4Dawg]

Well, first of all it's not in St. Louis and nowhere near an Armory. . . . .

I googled Amory Data Center. Everything that came up was about that one in St. Louis. One of them even had Armory spelled Amory.

 

[DoggieDaddy13]

It'd be more fair to say China is successfully running a astroturf campaing. No different than Russia funding green groups in the US.

documented FSB/Kremlin directed funding for U.S. environmental groups - $0.00

document FSB/Kremlin funding for Right-wing pundits, MAGA political candidates, and your run-of-the-mill authoritarians around the world - $300 million and counting.

• The Tenet Media Scheme: The U.S. Justice Department indicted employees of the Russian state broadcaster RT, alleging they funneled nearly $10 million to a U.S. media company to pay prominent, right-wing influencers to spread pro-Russia narratives and political messaging.
• Illegal Campaign Contributions: Federal authorities prosecuted Republican operatives for illegally funneling $25,000 from a Russian national into Donald Trump's 2016 campaign through U.S.-based consulting firms.
• Global Influence Operations: U.S. intelligence officials released findings that Russian agencies spent over $300 million since 2014 to covertly fund political groups in the U.S. and nearly a dozen countries worldwide.

 

[POTUS]

When the government is seizing people’s land for data centers as a public good and using imminent domain, yeah I’ve got a problem with that. People will not listen to him because of political differences, but if you know who Matt Walsh is, his YouTube video on data centers is pretty good.

 

[horshack.sixpack]

Well, I googled how loud AI centers are and the AI generated answer said it was like heavy traffic. I don't know why you'd question that.**

Perhaps a moratorium on temporarily being way too loud? You can't build it if you have to run jet engines for months on end. Even weeks would drive me nuts. https://tennesseelookout.com/2026/0...p-along-the-mississippi-tennessee-state-line/

 

[HRMSU]

Yet all the tech leaders in the world are fully convinced it's the future and thus are investing billions of dollars and betting their company's future on AI being the next big thing. Who do you think is more likely to be correct, the "regular folks" or the smartest tech experts on Earth?

Ya, like they're stupid or something. Cracks me up that people think something they use on the internet is the real AI. If they saw the real thing it would scare the crap out of them.

 

[JackReacherDawg]

My finger-in-the-wind sense is that lots of regular folks are deciding ai may next not be the next greatest thing. This polling sure seems to back that up. Where are you on it, what are you hearing and seeing?

View attachment 1296772

Americans Oppose AI Data Centers in Their Area

Seven in 10 Americans oppose the construction of an AI data center in their local area, including 48% strongly opposed.

news.gallup.com

Search results have gotten way better, but this still isn't AI. Every single one of these "AI"s cant do basic geometry problems. Because it has to actually think to do that, not match text. Which is all it really does.

The results I get as an SME sound good....but often have very basic mistakes. I had one last week insist ASTM 304 steel was stainless steel. Its not, its confused with 304 stainless steel, which is not the same thing. I had a mechanical engineer very confused and close to screwing up a drawing until I got involved. And I dug into its sources to see why, and it was seeing a website from a manufacturer that made that mistake. That's the thing. Its only as smart as what it was trained on, which is the entirety of the internet.....and there's a LOT of dumb, wrong **** on the internet. And it has no ability to know the difference.

Try this that I learned this week. Get a result from an AI, then tell it that its wrong and try again. Youll get a different result. If you tell it why it's wrong, you'll get a better result. Because its just telling you what you want to hear based on your prompt.

 

[HRMSU]

It's a tool. It should be wielded by SMEs. It should not be trusted implicitly. My take is same as it is on guns and chainsaws. Good if you are hunting or cutting down trees. Bad if you are killing people with them. Just instruments that are useful in the hands of the competent and dangerous in the hands of the malicious or incompetent.

It's our nuclear generation....we either lead or we become number 2 to another country. You don't want to live in a world where we are #2.

 

[HRMSU]

You know what would be killer? If cities that are crumbling in the delta that has old downtowns, that are dead, pitched an idea to these companies that if they renovated these buildings and upgraded the infrastructure to handle the water needs, they could use space in these buildings for the servers. As long as the bottom floors were at least partially available to people to open businesses for minimal lease. Let the city maintain control of the locations, but give the data companies free space in return for renovations and upkeep.

Not that big of a stretch. They've talked about putting data centers in the ocean. Why couldn't the delta take advantage of the Mississippi River?

eta: not putting it in the river but using the water

 

[POTUS]

It's our nuclear generation....we either lead or we become number 2 to another country. You don't want to live in a world where we are #2.

If we need to "beat China" then I'm sure these data centers aren't getting components from them right? insert Anakin & Padme meme here.

 

[patdog]

Search results have gotten way better, but this still isn't AI. Every single one of these "AI"s cant do basic geometry problems. Because it has to actually think to do that, not match text. Which is all it really does.

The results I get as an SME sound good....but often have very basic mistakes. I had one last week insist ASTM 304 steel was stainless steel. Its not, its confused with 304 stainless steel, which is not the same thing. I had a mechanical engineer very confused and close to screwing up a drawing until I got involved. And I dug into its sources to see why, and it was seeing a website from a manufacturer that made that mistake. That's the thing. Its only as smart as what it was trained on, which is the entirety of the internet.....and there's a LOT of dumb, wrong **** on the internet. And it has no ability to know the difference.

Try this that I learned this week. Get a result from an AI, then tell it that its wrong and try again. Youll get a different result. If you tell it why it's wrong, you'll get a better result. Because its just telling you what you want to hear based on your prompt.

Had a client who used AI to recreate a spreadsheet I had created. He had to tell it that it was wrong about a half dozen times at different stages in the process before it finally got it right. It's good for crunching data, but it's only as good as the data it crunches, and it can draw the wrong conclusions from that data.

 

[HRMSU]

Funny because if you look at the donations of the tech guys, it's almost completely to left wing causes. But somehow the right is playing ball with them. It's almost like we don't have two parties, but one.

Ya, like Silicon Valley was ever conservative. Money has no ideology but people with money do and it varies.

 

[HRMSU]

Often it’s the regular folks. The smartest experts in the auto industry just lost 10s if not 100s of billions rolling out failed EV initiatives that they’ve now cancelled. I could have told them it was a bad idea 10 years ago.

Huh, don't you think that was a little forced?

 

[johnson86-1]

Perhaps a moratorium on temporarily being way too loud? You can't build it if you have to run jet engines for months on end. Even weeks would drive me nuts. https://tennesseelookout.com/2026/0...p-along-the-mississippi-tennessee-state-line/

That's not the data center itself. I'm assuming those are aeroderivative turbines. That sounds like a real issue but its an issue of where power plants are allowed to be located. Not a practical difference to any residents that have to deal with it, but a real difference with respect to whether you should be against data centers or for zoning/permitting requirements for power companies (or others) locating new plants.

 

[CochiseCowbell]

Its only as smart as what it was trained on, which is the entirety of the internet.....and there's a LOT of dumb, wrong **** on the internet.

I'd be interested to see how often AI has referenced SPS and more importantly, why?

@dawgstudent you got that info?


**

 

[HRMSU]

That was a little different because that was a politically driven push, not experts really thinking EVs were going to win on their own merits.

It's very possible that AI proponents are early and this is going to be more like the dot com bubble. Too much investment too early in technology that people aren't able to figure out how to monetize quickly enough for the early investments to payoff. But we'll look back in a decade or two and realize it really was transformative.

I'm sort of predicting in between. I think AI is oging to be useful much faster, but that the current investment is basically an arms race for a leading position based on the idea that they will reach something like a small scale singularity where the AI is self improving enough that nobody can catch up afterwards. I don't think LLMs are the path to get to that place, and you're going to end up with a lot of overinvestment, but that there won't be a point like after the dot com bubble where you have people like Paul Krugman calling it the equivalent of the fax machine.

Some of you smart engineers on here could probably critique this but its definitely interesting.

Humanity May Reach Singularity Within Just 4 Years, Trend Shows

When will the singularity occur? According to new research, humanity may achieve singularity within just 5 years—much sooner than we originally thought.

tech.yahoo.com

 

[bolddogge]

Yeah, lets just pull the plug on all this AI stuff. I've always wanted to learn to speak Mandarin anyway.

 

[turkish]

Are we no longer concerned with climate change from greenhouse gas emissions? That’s their biggest impact. The incremental power they require is coming from fossil fuels nearly exclusively. If there was an expansion in the petrochemical industry demanding this much more power generation over such a short time period, it would never get off the ground.

 

[Perd Hapley]

I didn’t realize that there was ever a shine to begin with.

Recently read about a mega data center being built in Utah that was expected to consume twice as much power / energy by itself as the whole rest of the state of Utah combined. Ridiculous. We’re certainly figuring out ways to destroy the planet way faster than we ever did before.

 

[dawgman42]

The More You Know!

 

[Villagedawg]

civil liberties

What are these? I think I remember a history teacher talking about them one time?

 

[Villagedawg]

Yet all the tech leaders in the world are fully convinced it's the future and thus are investing billions of dollars and betting their company's future on AI being the next big thing. Who do you think is more likely to be correct, the "regular folks" or the smartest tech experts on Earth?

Let me take a stab at it. The ones with money?

 

[Perd Hapley]

It's a tool. It should be wielded by SMEs. It should not be trusted implicitly. My take is same as it is on guns and chainsaws. Good if you are hunting or cutting down trees. Bad if you are killing people with them. Just instruments that are useful in the hands of the competent and dangerous in the hands of the malicious or incompetent.

You leftist SOB…..Pry it from my cold, dead hands….

 

[paindonthurt]

Can you tell me the upside here ? Massive depletion of water sources, more power requirements than what it takes to run San Francisco or other large cities, enabling mass surveillance that may erode civil liberties, huge bills to be funded by overstretched tax payers, no popular support, who voted for this ?

what am I missing here ?

There isn’t a massive depletion of water.

We need more energy infrastructure in this country anyways. Good catalyst to get more energy infrestructure.

yeah let’s put AI on the back burner and let China do it before us! GREAT IDEA!

 

[paindonthurt]

That's easy, Republicans.

Not picking or choosing here, just facts. The AI/tech lobby has got them in their pockets.

 

[Villagedawg]

It's our nuclear generation....we either lead or we become number 2 to another country. You don't want to live in a world where we are #2.

We much prefer a world where we aren't even in the top 20 in Happiness (23rd), Basic Human Needs (28th), Life Expectancy and Health (40th to 50th), and Press and Personal Freedom (28th). *

 

[patdog]

We much prefer a world where we aren't even in the top 20 in Happiness (23rd), Basic Human Needs (28th), Life Expectancy and Health (40th to 50th), and Press and Personal Freedom (28th). *

As imperfect as things are here, I would be very skeptical about a lot of these rankings.

 

[85Bears]

Let me take a stab at it. The ones with money?

I’m afraid a few posters on here would normalize, rationalize anything. Critical thinking and intellectual curiousity are becoming a rare thing.

 

[johnson86-1]

As imperfect as things are here, I would be very skeptical about a lot of these rankings.

Happiness one is plausible because that’s subjective. Basic human needs is almost certainly a measure of preferred policies, not actual human needs. Life expectancy and health is very probably correct if there is no adjustment for demographics. Press and personal freedom is ludicrous. As bad as we occasionally are, we are still the tallest little person there.

eta: one of the tallest little people. Places like Switzerland may be better. Maybe some developing countries that are embracing freedom have an argument. But most developed countries are pretty awful.

 

[paindonthurt]

Yet all the tech leaders in the world are fully convinced it's the future and thus are investing billions of dollars and betting their company's future on AI being the next big thing. Who do you think is more likely to be correct, the "regular folks" or the smartest tech experts on Earth?

No no
It’s clearly the regular folks and the Sixpack warriors who know what’s better!

 

[paindonthurt]

That is a really interesting question.

It’s not really though.
It’s pretty clear who is smarter.