Mr Muse, thanks for the down to earth brass tacks of heat transfer/removal. So refreshing to hear someone talking sense about a subject all people deal with daily without necessarily even realizing it.
I’m ambivalent about AI data centers, believing AIs potential to be far more limited than others, so build away, but don’t be surprised if demand isn’t what you project.
That said, Microsoft can afford to eat 6 to 10% additional cost, can every AI company say the same? They aren’t profitable and have huge overhead costs, so why wouldn’t they build the less expensive option? Also, when isn’t a company (of any type) looking to take advantage of cost savings?
>A single hamburger takes hundreds of gallons to produce
Very annoying that in an article dedicated to correcting one fallacy you repeat another. The study or whatever which came up with the absurd figure for water usage in beef production, from what I recall, arrived at it by measuring rainfall across the entire pasture range and then attributing the total to each individual cow. Same dishonest rhetoric as some of the worst datacenter fear mongering.
Here you go Grug, “The average household flushes their toilet about 20 to 22 times a day, using approximately 32 to 35 gallons of water daily.” 🍔 Lactating cows drink more water. 🥛Is this a better comparison for you? Cows don’t use much water at all if they are grass fed. They get it from what they eat. It depends on the time of year how much they drink vs how much they get from grazing grass.
i clicked because you used "innumeracy" in the title.
literally and numerically, the students in 4th, 8th, and 12th grades are functionally illiterate and innumerate. I'll read tomorrow and have a few more thoughts . . .
Direct use was 17.4 billion gallons in 2023. You left out the much larger amount of water used to generate the electricity that powers data centers.
And the concern is with thirsty hyperscale data centers being built in areas that already face periodic drought or rapid population growth - like parts of the Southwest, Texas, Virginia, Georgia, and North Carolina.
Mr Muse, thanks for the down to earth brass tacks of heat transfer/removal. So refreshing to hear someone talking sense about a subject all people deal with daily without necessarily even realizing it.
I’m ambivalent about AI data centers, believing AIs potential to be far more limited than others, so build away, but don’t be surprised if demand isn’t what you project.
That said, Microsoft can afford to eat 6 to 10% additional cost, can every AI company say the same? They aren’t profitable and have huge overhead costs, so why wouldn’t they build the less expensive option? Also, when isn’t a company (of any type) looking to take advantage of cost savings?
Thank you Alex Muse. The problem is Americans that want to see us fail.
>A single hamburger takes hundreds of gallons to produce
Very annoying that in an article dedicated to correcting one fallacy you repeat another. The study or whatever which came up with the absurd figure for water usage in beef production, from what I recall, arrived at it by measuring rainfall across the entire pasture range and then attributing the total to each individual cow. Same dishonest rhetoric as some of the worst datacenter fear mongering.
Here you go Grug, “The average household flushes their toilet about 20 to 22 times a day, using approximately 32 to 35 gallons of water daily.” 🍔 Lactating cows drink more water. 🥛Is this a better comparison for you? Cows don’t use much water at all if they are grass fed. They get it from what they eat. It depends on the time of year how much they drink vs how much they get from grazing grass.
i clicked because you used "innumeracy" in the title.
literally and numerically, the students in 4th, 8th, and 12th grades are functionally illiterate and innumerate. I'll read tomorrow and have a few more thoughts . . .
Direct use was 17.4 billion gallons in 2023. You left out the much larger amount of water used to generate the electricity that powers data centers.
And the concern is with thirsty hyperscale data centers being built in areas that already face periodic drought or rapid population growth - like parts of the Southwest, Texas, Virginia, Georgia, and North Carolina.