Are you for or against Data Centers?

You might appreciate this one. Not sure if others will understand it.

1779651186011.webp
 
We've had data centers around us for many years without issue.

The new data centers are energy and water hogs that also produce noise pollution. Some even have their own power plants.

If a mega-data center moved in, didn't noise pollute, lower the water table, or increase energy/water costs for the community. It wouldn't bother me.



These companies know something many of us don't and they are putting billions on the line, and will continue to do so when the chips need upgrades.

A server rack with 8 Nvidia H200's costs $400k and has the same compute as an entire traditional server ROOM 5 years ago. In fact it would supposedly cost $400k just to build and wire out the room, and now you can get it all in an array of 8 chips.

These data centers aren't for your typical web-hosting or storage. They are essentially for processing massive datasets to improve current services or invent new ones completely. Supporting all AI usage is important but I think developing new services will be primary.

This WILL replace many jobs but will also supplement and supercharge many fields such as medicine, research, and whatever tools companies use.

There WILL be massive turbulence in society. There will be haves and have nots and likely a larger welfare class. Once robots take manual work over, good luck. UBI Universal Basic Income will be commonplace, people will feel useless. Social impact of this will be, unknown but likely not good. We can't be all on welfare, having meaning in life, some of which comes from ones career, and optimism is important in a successful society and for birth rates too.

My conspiracy angle is that some of this compute ( for non-gov't facilities) will be provided for military purposes as well. The AI race is just as important for cyber warfare, developing new weapons, decrypting enemy communications, reverse engineering tech, making political decisions, finding trends and patterns not previously humanly possible. Some high level stuff only our best and brightest know about.

But what are they computing? 5 times nothing is still nothing. beyond virtual things, and communications I see no improvement in day to day tech, than 30 years ago.....really, it has been down hill, with the exception to some things....battery tech being one.

If AI is all super duper powerful, how come they do not direct the AI to make ultra efficient solar panels ending the worlds power "issue".

Some say "Because we have not figured it out yet".......well with the help of super computers that can make billion of decisions per minute, and access to all of the worlds data......coming up with something as simple as a 100% efficient solar panel should not be that far away. Or cure cancer. Or, make a car that gets 100 miles per gallon?


I think it's on par with the Manhattan project, both militarily and privately. Having private industry be a good community steward is the issue, otherwise no one would really care.
Maybe. Such a building would be a very easy target for a near peer to destroy. To bank most capabilities on such a building would be foolish IMO.

More likely the data to be used against us, than a military adversary
 
Yes, GHCP is going from fixed price subscription of $40/mo for an insane amount of usage even with top tier frontier models to charging for your actual usage. My estimated usage is $500+/mo and will continue to rise.

Sure the self hosted open models on reasonably affordable hardware is not going to compete with Claude Opus but for most tasks you don’t need that amount of smarts. For both my servers was about $2500.

The local models can whip up a typical business website or debug error logs or come up with a proper plan based on a lazy description that I can then use in a better model. For large or complex tasks I’m keeping the $10 GHCP plan. Also cancelled SuperGrok so another $30 savings there. The free tiers of Grok, Gemini, Claude, and ChatGPT plus the $10/mo GHCP plan plus my two local AI servers will just about cover my usage. And it will pay for itself in just a few months when compared to the new GHCP pricing.
OK, all electronic stuff. I am talking about physical stuff.
 
If AI is all super duper powerful, how come they do not direct the AI to make ultra efficient solar panels ending the worlds power "issue".
I would wager that's exactly what these companies are using data centers for. We might not hear everything that's happening, just yet. But having a super brain not only run billions of simulations and tests per hour but also recursively and intelligently improve the design/concept is like having a team of 100s of R&D techs.

It's all about bringing a product to market and it will help with that. I bet xAI is doing solar panel research, battery research, anything that increases the bottom line and competitive advantage.

AI is solving 100+ year old mathematical problems in hours. The same being done for any other topic, is certainly happening. We might just not know it yet, but will hear about and hopefully benefit from it in some way. These multi-billion dollar projects are run by people in the know. It's not like the crypto- craze a few years ago.

AI will power everything we do from now on and we won't be able to stop it.

Or cure cancer.
Bezos is starting his own AI company called Prometheus. Specifically focused on medical research IIRC. Some says its so he can invent life extension therapies, which I don't blame him, I would do the same, and this would generate billions or trillions.

AI research into medicine will likely cure many cancers in the next 5 years, cure baldness, and allow us to live essentially forever. Well those with millions to spare. Some futurists postulate that the 1st person to live to 1000 has already been born.

More likely the data to be used against us, than a military adversary
Surveillance, most certainly. Faster correlations between massive data pipelines, in the wrong hands we're screwed and controlled beyond comprehension. Used appropriately, crime goes down and society benefits.

Military uses, I'd say this is the primary focus. I don't believe surveillance tech or even running through massive data sets for this purposes is very AI intensive. Inventing and simulating new weapons, hacking adversaries, decryption, who knows what else, is the likely most important use.
 
If AI is all super duper powerful, how come they do not direct the AI to make ultra efficient solar panels ending the worlds power "issue".

Eventually I think we'll get there; or at least an unobtainable answer as far as current technology and economics allowed. For example, let's just say AI says we'll need 100,000 pounds of silver to create the solar panels we'll need to solve the current crisis with today's current technology. However, current technology make that economically impossible.

It's still in its infant stage though. Next few years maybe we'll see who wins this race.
 
Eventually I think we'll get there; or at least an unobtainable answer as far as current technology and economics allowed. For example, let's just say AI says we'll need 100,000 pounds of silver to create the solar panels we'll need to solve the current crisis with today's current technology. However, current technology make that economically impossible.

It's still in its infant stage though. Next few years maybe we'll see who wins this race.
I hear you partner.....but

That is the same thing that has been being said for the last 20+ years. Always in its infancy. Kind of like Iran being 1 month away from nukes.

The AI could make it economically possible. It should be able to think a billion steps ahead, and every change of course, all possibilities accounted for. Basically, if it was possible, the AI should be able to give us a path. That is what I want to see. An AI you could pose a real question to, and it give the answer. Not "hey AI are there Aliens?" or "hey AI, are we all being surveilled right now"?

How about: "Hey AI, what are the best companies to invest in the stock market"

Maybe there is no infancy. Like a photon. Or maybe it is nothing, a red herring. I can think of a few of those given to the public.
 
I hear you partner.....but

That is the same thing that has been being said for the last 20+ years. Always in its infancy. Kind of like Iran being 1 month away from nukes.

The AI could make it economically possible. It should be able to think a billion steps ahead, and every change of course, all possibilities accounted for. Basically, if it was possible, the AI should be able to give us a path. That is what I want to see. An AI you could pose a real question to, and it give the answer. Not "hey AI are there Aliens?" or "hey AI, are we all being surveilled right now"?

How about: "Hey AI, what are the best companies to invest in the stock market"

Maybe there is no infancy. Like a photon. Or maybe it is nothing, a red herring. I can think of a few of those given to the public.

I agree on all points. Right now it seems like AI is more like an advanced search engine and able to calculate the near future with human-determined formulas. Once the Intelligence in Artificial intelligence starts showing and it starts thinking for itself, that would be quite a thing to experience.
 
But what are they computing? 5 times nothing is still nothing. beyond virtual things, and communications I see no improvement in day to day tech, than 30 years ago.....really, it has been down hill, with the exception to some things....battery tech being one.
This site is undoubtedly hosted at a data center. You don’t think it is any better or useful than a dialup BBS of 30 years ago?
 
This site is undoubtedly hosted at a data center. You don’t think it is any better or useful than a dialup BBS of 30 years ago?
Like I said previously, beyond communications, I dont see much. Also beyond communication, no one has given an example to the contrary applicable to everyday life. Nothing you can buy is cheaper, or better.....nothing.

Take away instant communication, take away smart phones, again I dont see much.
 
Normal hosting data centers are full of storage devises that don’t need much for cooling or power. AI data centers are full of nvidia h200 gpus that suck up massive amounts of power and need tons of water to cool them down.
One h200 chip consumes around 400 watts and each data center can have hundreds of thousands of h200 gpus. It’s not even in the same universe as a hosting data center when it comes to cooling and power.
 
Normal hosting data centers are full of storage devises that don’t need much for cooling or power. AI data centers are full of nvidia h200 gpus that suck up massive amounts of power and need tons of water to cool them down.
One h200 chip consumes around 400 watts and each data center can have hundreds of thousands of h200 gpus. It’s not even in the same universe as a hosting data center when it comes to cooling and power.
And for such wholesome goodness like BitCoin 😵‍💫
 
I agree on all points. Right now it seems like AI is more like an advanced search engine and able to calculate the near future with human-determined formulas. Once the Intelligence in Artificial intelligence starts showing and it starts thinking for itself, that would be quite a thing to experience.
I read an article on a small hedge fund in NYC. It's like 5 guys, computer scientists, finance guys, quants, whatever. They supposedly built an AI so accurate that they are making billions a year. They won't let anyone into the fund, just keep the money for themselves. Makes sense TBH.

All you need is to beat the house by 1% and you can print money, sounds like they are doing better than that.

Also beyond communication, no one has given an example to the contrary applicable to everyday life. Nothing you can buy is cheaper, or better.....nothing.
Massive benefits won't be instantaneous, but they are there, and many more in the pipeline. It's the new way everything will be researched in the future.

We just don't have all the news or know the provenance of every single thing we use. AI is being used for larger more important problems now, here are some examples:

The Simulation Foundation
Before AI, breakthroughs in weather forecasting, nuclear modeling, and protein folding relied on supercomputers running physics equations for weeks. AI learned from those simulation outputs and now replicates in seconds what previously took months, while finding patterns in data too complex for any human or traditional computer to navigate.

AlphaFold
Protein folding stumped scientists for 50 years. In 2020 AlphaFold solved it, predicting protein structures with near-experimental accuracy and releasing 200 million structures freely, fundamentally accelerating drug discovery overnight.

New Drugs in Trials
Insilico Medicine designed a drug for idiopathic pulmonary fibrosis entirely from scratch and reached Phase 2 clinical trials in 4 years versus the typical 10 to 15. Recursion Pharmaceuticals runs AI across billions of cellular experiments at a scale no human research team could match.

Cancer Detection

DeepMind's AI detects breast cancer from mammograms more accurately than radiologists, reducing false positives by 5.7% and false negatives by 9.4%. Similar systems now detect diabetic retinopathy, skin cancer, and lung nodules with superhuman accuracy.

New Materials
Google DeepMind's GNoME project discovered 2.2 million new stable crystal structures, surpassing everything humans had found in the entire history of materials science. Around 380,000 show promise for batteries, solar cells, and semiconductors.

Fusion Energy
DeepMind trained an AI to control plasma inside a tokamak fusion reactor in real time, solving a control problem too complex for traditional computing and directly advancing practical fusion energy.

Chip Design
Google's AI designs the physical layout of its TPU chips in hours, a task that takes human engineers months, and produces layouts that outperform human designs. The chips that train AI are now being designed by AI.

Climate Modeling
AI surrogate models now run climate simulations thousands of times faster than traditional physics-based supercomputer models, allowing researchers to test far more scenarios and make predictions previously out of reach.

Have no fear, this will benefit us all in the end, there will be turbulence but if we make it through it, it will be amazing.
And for such wholesome goodness like BitCoin 😵‍💫
IIRC, some of the bitcoin miners are using their GPUs as data centers now.
 
Data Centers need water, why not build near Lake Michigan…… or Lake Okeechobee (Florida) ?

Lots of water and lots of empty land available.

Just thinking…..

:unsure:
 
Last edited:
I read an article on a small hedge fund in NYC. It's like 5 guys, computer scientists, finance guys, quants, whatever. They supposedly built an AI so accurate that they are making billions a year. They won't let anyone into the fund, just keep the money for themselves. Makes sense TBH.

All you need is to beat the house by 1% and you can print money, sounds like they are doing better than that.


Massive benefits won't be instantaneous, but they are there, and many more in the pipeline. It's the new way everything will be researched in the future.

We just don't have all the news or know the provenance of every single thing we use. AI is being used for larger more important problems now, here are some examples:

The Simulation Foundation
Before AI, breakthroughs in weather forecasting, nuclear modeling, and protein folding relied on supercomputers running physics equations for weeks. AI learned from those simulation outputs and now replicates in seconds what previously took months, while finding patterns in data too complex for any human or traditional computer to navigate.

AlphaFold
Protein folding stumped scientists for 50 years. In 2020 AlphaFold solved it, predicting protein structures with near-experimental accuracy and releasing 200 million structures freely, fundamentally accelerating drug discovery overnight.

New Drugs in Trials
Insilico Medicine designed a drug for idiopathic pulmonary fibrosis entirely from scratch and reached Phase 2 clinical trials in 4 years versus the typical 10 to 15. Recursion Pharmaceuticals runs AI across billions of cellular experiments at a scale no human research team could match.

Cancer Detection

DeepMind's AI detects breast cancer from mammograms more accurately than radiologists, reducing false positives by 5.7% and false negatives by 9.4%. Similar systems now detect diabetic retinopathy, skin cancer, and lung nodules with superhuman accuracy.

New Materials
Google DeepMind's GNoME project discovered 2.2 million new stable crystal structures, surpassing everything humans had found in the entire history of materials science. Around 380,000 show promise for batteries, solar cells, and semiconductors.

Fusion Energy
DeepMind trained an AI to control plasma inside a tokamak fusion reactor in real time, solving a control problem too complex for traditional computing and directly advancing practical fusion energy.

Chip Design
Google's AI designs the physical layout of its TPU chips in hours, a task that takes human engineers months, and produces layouts that outperform human designs. The chips that train AI are now being designed by AI.

Climate Modeling
AI surrogate models now run climate simulations thousands of times faster than traditional physics-based supercomputer models, allowing researchers to test far more scenarios and make predictions previously out of reach.

Have no fear, this will benefit us all in the end, there will be turbulence but if we make it through it, it will be amazing.

IIRC, some of the bitcoin miners are using their GPUs as data centers now.
(n)
yeah I have read the same things, but see no tangible benefits as of yet.

Most notably, weather modeling. Still the only profession that you can be right 100% of the time, and wrong 100% of the time.
 
Data Centers need water, why not build near Lake Michigan…… or Lake Okeechobee (Florida) ?

Lots of water and lots of empty land available.

Just thinking…..

:unsure:

Waterfront property tends to be very expensive and requirements may have to hedge against extreme flooding or other disasters/threats.
 
Here in the KC metro one data center that ground was broken on, just picked up and left when the neighbors in the area protested a lot and vowed to make the data centers life miserable. The other one the city they were approved, got reversed by massive public protest.
 
Data Centers need water, why not build near Lake Michigan…… or Lake Okeechobee (Florida) ?

Lots of water and lots of empty land available.

Just thinking…..

:unsure:
they are and even the Great Lake water basin is not unlimited, hence why there is a pact and certain areas are in the skirt house
 
Ok, so beyond comedy, and making fake youtube videos, is there any tangible (other than the inefficient heat loss and use for space heating)?
I think start at the end of 2025 Claude Code was able to do real productive work that is faster than what human can do efficiently. This is like someone was able to read your design from the last 30 years then become a better engineer than you are, if you are willing to pay for the token cost. Before that sure, a lot of fake videos as tech demo, but recently OpenAI reduced / stopped these video generation AI and focus back on productivity stuff like software development.

It is just the start but from what I heard a lot of Indian contracting firm employee got laid off in the last several months because they can't compete with AI. Many of the translation gigs and testing gigs are now done by AI completely instead of outsourced to 3rd world.

Palentir has been doing a lot for the military as well as automate human labor intensive but boring work but they aren't generative AI, they are ontology based instead.
 
Back
Top Bottom