We only hear about the bad things about A.I. data centers. How much will the good out-weigh the bad? Will they really be that awful to the environment and jobs and society, or is this just paranoia and mass hysteria?
>>16987551> How much will the good out-weigh the badPretty much not at all. That's the biggest reason people are so upset.There are already plenty of industry that are environmentally destructive and bad for the local quality of life, like power plants, paper mills, chemical factories, mines, etc, but people put up with them because they're beneficial and profitable.AI data centers, by contrast, are utterly worthless. Building them doesn't actually benefit anyone anywhere in any real way, the way that getting some paper made or some power generated does. AI products are NOT profitable and they are NOT useful. These factories just make the local area worse for no benefit.
>>16987561But what if we make the AI make the AI useful?
The more telling aspect is this realization isn't supported by the same media industry that trumpets climate change catastrophism; it's gets the pass from the Jewish panopticon for, reasons
>>16987579how much money must i give you for that revolutionary idea
>>16987551This surely won’t be hilarious to watch in action.
The tragic Irony is it's all pointless even if you assume AI will be a life changing technology, it won't. Just like the original computers were the size of small warehouses just to run some punch cards at 1% the speed of a TI-83 graphing calculator. You can call it Moore's law if you want but 10 years from now they will be able to match the power of these inefficient data centers with a simple 3000 sqft office space filled with GPU racks, or a more advanced and efficient chip set up than we have today. The super large one in Utah will take 10-15 years to build and The pace of technological advancement is way outpacing the ability to build infrastructure for it. Meaning in 15 years when the Utah AI data center is complete it will be outdone by the cutting edge design that can do it for 1/10th the cost in 1/100th the space generating 5x the output. But most of the shifty AI hype men will be cashed out and on to the next grift by then.
>>16987604Sure the concrete won't be obsolete but maybe it will. Physical building shapes and layout also change radically over time to acomodate new ever changing demands of cutting edge technology. Just like older traditional data centers are completely unable to handle the new AI load. You couldn't even strip an old data center to the concrete and steel beams, hollow it out, and fill it with new cooling, power, and GPUs. You just can't. The physical building is unable to accommodate it. https://medium.com/b8125-fall2025/are-we-building-tomorrows-digital-ruins-the-looming-obsolescence-of-data-centers-in-the-ai-age-1057f7b1dc7ahttps://builtin.com/articles/future-of-data-centers-aiThe articles above help explain this. Even physical location of the building closer to it's main user base could one day render these massive buildings obsolete and when you think of the Utah center, in the middle of no where Utah, it's physically far form the vast majority of the US population. Not building it on the heavily populated coast lines is a death sentence for 10 years from now. Screencap it for 2036 when the thing shuts down after 4 months of uptime to start building in Virginia or Central California. Hell they might have to build an AI data center in LA just to handle the AI porn industry that will be a thing soon.
The most depressing thing for me is how apparently everyone quantifies intelligenceFor me, intelligence is figuring out something completely new from combining prior experience.Every LLM is just a Google aggregator
from my understanding, the datacenters will only accelerate deep learning models, matrix multiplication etcno other AI
>>16987561>Building them doesn't actually benefit anyone anywhere in any real way, the way that getting some paper made or some power generated does.So AI algorithms can't ever possibly be used to write papers or aid power generation hardware to make power generation more efficient?
>>16987551How is it compared vs the heat output of an active volcano?
>>16987551>How much will the good out-weigh the bad?Data centers are required for the roll out of central bank digital currency to enslave the masses, therefore the massive draw of electricity and freshwater is worth the investment. Through the control of digital currency you control production, distribution, sale, and use (central planning). If there are sufficient technical advancements in quantum computers using AI, or non quantum AI then that drastically reduce energy consumption (and therefore generate less heat and need less water for cooling) then the potential impact on the environment will be a lot less and less of a concern.
>>16987551is this one that much worse than the surveillance ones long established there
>>16987551Data centers are an investment scam. You're paying through the nose to buy supercomputers that you don't need all to train a LLM that doesn't do anything useful. Congrats, your 401k got invested into a server farm that can't find customers. Enjoy holding the bag when they turn no profit.BTW there's no more training data. Current models already are trained on the entire internet.
>>16987561>AI products are NOT profitable and they are NOT useful. These factories just make the local area worse for no benefitDatacenters also have very few jobs. A paper mill employs thousands of full time workers. A data center employs a dozen security guards and a few techs. There is no financial return in taxes or local employment.
>>16987873AI data centers have significantly more GPU compute. The old surveillance was more REGEX matching, speech to text and fingerprinting with a searchable archive of every email, text message and phone call.
>>16987551just use the heat to generate energy, ez
>>16987551>How bad will A.I. data centers actually be?Only as bad as the pedo mass murder cult having total and permanent informational dominance over you can be.
