Sure, but how do you compare the impact of millions of people (and animal) eating corn (or drinking corn syrup, or using corn in any kind of process...) with the impact of people using data centers ?
Very intuitively, it feels like if I offer the choice of shutting down all data centers, and burning all corn fields, the impact would be dramatically different. I'm really curious how an economist would estimate each effect, and how they would put the two numbers in perspective, but just putting the numbers next to each other feels like doing only half of the job at best. (If an Oracle salesperson was doing it, we would call it downright dishonest.)
But I'm struggling with the fact that I could not objectivize my concern either - and again, I do not claim your figured are wrong.
Corn farming is the go-to example for any comparison of waste.
For another example, if every car in the USA were an EV, you’d only have to use 3.5% of the land area now used for corn to provide 100% of their electricity needs.
But instead, 40% of corn grown goes to ethanol, to provide only 7% of gas car’s energy needs.
Sorry if it is addressed in the main article, but to use the "corn water usage vs AI water usage" as a hook is a bit disingenuous - the metric itself completely ignores the vast difference in usage, value created, etc... Maybe you should compare "water used" over "number of buyers" / "number of consumers" etc... Or even, if you really want to use cold numbers, give me the share of GDP that derives from "corn has grown" vs "AI data center have churned numbers" - and give me the water intensity of both. I'm ready to bet that corn will suddenly look better .
(I suspect this "trick" must have a name - it's not really "cherry picking", and the data is probably perfectly legit, but it's not apple vs orange per se...
But it's the similar tocomparing carbon emission from cars (used by billions) and private jets (used by a few thousands) . Although, here, you can clear the conversation by having a single "per Capita" intensity. For corn vs AI, the right denominator is much less obvious - and there is a "value" element to it (maybe we could find a number like "what would it cost the economy if the corn / ai disappeared tomorrow" ? So of a "counterfactual intensity ?" Probably studied to death, but I don't know the word to lookup...)
It's actually pretty likely that corn would still have a much much higher water intensity than AI if you divide per person consuming it. One involves dumping water into the ground over and over to grow a plant that takes a few minutes to eat, the other involves running water in the vicinity of hot computers to cool them down.
I've tried my best in each comparison I make to give a complete picture using the info I have, but there are always going to be so many variables to compare that any visual is going to be incredibly limited, but imo still useful.
Not sure I get you comment. If there is a comparison between water use in data center and golf course, then I'd also find it hard to interpret, but this way the other way round : many more people use data centers than golf courses ; but golfers are rich, on average (no offense meant, I have golfers friends, they just happen to be squarely in the 10% percentile), so maybe "golf" has a lot of GDP attached ?
Again, my point is that the comparison make some sense physically, but not much in terms of policy...
Sure, but how do you compare the impact of millions of people (and animal) eating corn (or drinking corn syrup, or using corn in any kind of process...) with the impact of people using data centers ?
Very intuitively, it feels like if I offer the choice of shutting down all data centers, and burning all corn fields, the impact would be dramatically different. I'm really curious how an economist would estimate each effect, and how they would put the two numbers in perspective, but just putting the numbers next to each other feels like doing only half of the job at best. (If an Oracle salesperson was doing it, we would call it downright dishonest.)
But I'm struggling with the fact that I could not objectivize my concern either - and again, I do not claim your figured are wrong.
About 40% of corn is used for ethanol for cars, which is an epic waste of food and land.
Corn farming is the go-to example for any comparison of waste.
For another example, if every car in the USA were an EV, you’d only have to use 3.5% of the land area now used for corn to provide 100% of their electricity needs.
But instead, 40% of corn grown goes to ethanol, to provide only 7% of gas car’s energy needs.
Sorry if it is addressed in the main article, but to use the "corn water usage vs AI water usage" as a hook is a bit disingenuous - the metric itself completely ignores the vast difference in usage, value created, etc... Maybe you should compare "water used" over "number of buyers" / "number of consumers" etc... Or even, if you really want to use cold numbers, give me the share of GDP that derives from "corn has grown" vs "AI data center have churned numbers" - and give me the water intensity of both. I'm ready to bet that corn will suddenly look better .
(I suspect this "trick" must have a name - it's not really "cherry picking", and the data is probably perfectly legit, but it's not apple vs orange per se...
But it's the similar tocomparing carbon emission from cars (used by billions) and private jets (used by a few thousands) . Although, here, you can clear the conversation by having a single "per Capita" intensity. For corn vs AI, the right denominator is much less obvious - and there is a "value" element to it (maybe we could find a number like "what would it cost the economy if the corn / ai disappeared tomorrow" ? So of a "counterfactual intensity ?" Probably studied to death, but I don't know the word to lookup...)
It's actually pretty likely that corn would still have a much much higher water intensity than AI if you divide per person consuming it. One involves dumping water into the ground over and over to grow a plant that takes a few minutes to eat, the other involves running water in the vicinity of hot computers to cool them down.
I've tried my best in each comparison I make to give a complete picture using the info I have, but there are always going to be so many variables to compare that any visual is going to be incredibly limited, but imo still useful.
And what of Arizona golf courses? Are they also a threat to human hydration?
You are cherry picking the corn example and not actually using the comparison tool.
Not sure I get you comment. If there is a comparison between water use in data center and golf course, then I'd also find it hard to interpret, but this way the other way round : many more people use data centers than golf courses ; but golfers are rich, on average (no offense meant, I have golfers friends, they just happen to be squarely in the 10% percentile), so maybe "golf" has a lot of GDP attached ?
Again, my point is that the comparison make some sense physically, but not much in terms of policy...