Remember when some of y'all got real mad at me for doing some back-of-the-napkin math about how much water "AI" likely uses, in light of the fact that they refuse to release their actual numbers?
Anyway.
Google’s emissions are up 48% in 5 years. Microsoft’s are up 30% since 2020. Those soaring figures are being driven by mass investments in data centers to power AI tools.
Generative AI is a climate disaster and data center expansion must be stopped.
Because people seem to not understand how "AI" uses so much water, here's a quick+dirty breakdown:
The more data centers you have & the more traditional power you use, the more water you 1a) use To cool & 1b) lose, because they arent perfectly sealed systems, & 2) will need to boil to run turbines.
Remember when some of y'all got real mad at me for doing some back-of-the-napkin math about how much water "AI" likely uses, in light of the fact that they refuse to release their actual numbers?
Anyway.
New game: First Large Language Model/Deep Learning/Generative "AI"/"Machine Learning" firm w/ only renewable energy sources, an open source & radical consent model for the collection & use of training data, & an output correction process that's not just "don't talk about prejudicial bias, ever" wins
I mean even "closed" cooling systems have loss points and leakage, but again , the more data centers you Have, the more power they Use (≈the more water they require to turn to steam to run turbines), the more heat they Produce, the more cooling they Need= the more Water they need. It's use and scale
Because people seem to not understand how "AI" uses so much water, here's a quick+dirty breakdown:
The more data centers you have & the more traditional power you use, the more water you 1a) use To cool & 1b) lose, because they arent perfectly sealed systems, & 2) will need to boil to run turbines.
