The vast data centers that power artificial intelligence are so energy hungry that they’re heating up their surroundings, according to new research. It’s an alarming finding given the number of data centers is predicted to explode over the next few years.
No statistical significance, control groups, numbers, or models. Just rolling average temperature vs temperature.
They don’t say how they filter out trends or seasonality. They don’t look at switch point models or do difference in difference. And it’s a non reviewed preprint.
This research needs an actual statistician to use their data and fix it, because as is, this is not worth seriously discussing.
Edit: okay downvotes. I’m not saying there isn’t something here, I’m saying this study does nothing to support the hypothesis seriously.
We have data picking and choosing to throw out data, we have no statistical rigour, and we have no comparison against other rural land transformation activities.
We already know about urban heat islands, is this just that? That’s an important question when trying to say data center use is the issue.
It probably is an issue, the chart is compelling, but without better work none of this is actionable.
If you click though to the study instead of reading the drivel summarizing it for profit they have two graphics about time to data center and temperature.
This study is crap.
No statistical significance, control groups, numbers, or models. Just rolling average temperature vs temperature.
They don’t say how they filter out trends or seasonality. They don’t look at switch point models or do difference in difference. And it’s a non reviewed preprint.
This research needs an actual statistician to use their data and fix it, because as is, this is not worth seriously discussing.
Edit: okay downvotes. I’m not saying there isn’t something here, I’m saying this study does nothing to support the hypothesis seriously.
We have data picking and choosing to throw out data, we have no statistical rigour, and we have no comparison against other rural land transformation activities.
We already know about urban heat islands, is this just that? That’s an important question when trying to say data center use is the issue.
It probably is an issue, the chart is compelling, but without better work none of this is actionable.
Chart?
If you click though to the study instead of reading the drivel summarizing it for profit they have two graphics about time to data center and temperature.
https://www.researchgate.net/publication/403073048_The_data_heat_island_effect_quantifying_the_impact_of_AI_data_centers_in_a_warming_world/fulltext/69c21ebcac3812287560e46e/The-data-heat-island-effect-quantifying-the-impact-of-AI-data-centers-in-a-warming-world.pdf?origin=publication_detail&_tp=eyJjb250ZXh0Ijp7ImZpcnN0UGFnZSI6InB1YmxpY2F0aW9uIiwicGFnZSI6InB1YmxpY2F0aW9uRG93bmxvYWQiLCJwcmV2aW91c1BhZ2UiOiJwdWJsaWNhdGlvbiJ9fQ
That… Is consistent
I wonder if there were trees or shade something in the before and direct sun in the after
And that’s why I want to see someone else take their data and do this justice