Once we hit the 90s, what are even the sources for added consumption in American households? We hit all the big ticket power sucking items already, like refrigeration and AC.
this will allow the hyperscalers to build more DCs to make you completely redundant, so that bezos can have a 3rd fleet of jets
But they're not, they're building out data centers because the world demands more AWS capacity. Shouldering the emissions from this on Amazon, unless you believe they're being particularly inefficient about it, is shooting the messenger.
Its the same as when Uber launched. VCs suppress prices, to create demand ($5 Uber rides anywhere in the city) which generates more transactions, which generates more pollution. Instead of a alternative lower cost transportation (like BART or muni), SF residents chose the highest environmental impact and lowest cost option.
Uber hasn’t raised from VCs in years, and their business is far bigger than it was back when they were losing money.
The idea that SF residents choose to use Uber rather than BART because Uber is cheaper is simply wrong - Uber is much more expensive than BART, and with some notable exceptions for shared rides, that was true during the VC funded growth period as well.
The cost of these services is artificially suppressed to drive adoption, at the cost of our environment.
> The idea that SF residents choose to use Uber rather than BART because Uber is cheaper is simply wrong
When I lived in SF. Uber and Lyft cost between free and $5 to go anywhere in the city. Yes, $5 is more than the $2.75. But for price of a cup of coffee more, Uber would pick you up and drop you off exactly where you needed. Taking muni or bart involved walking, waiting, more waiting, and then more walking.
Really the only reason why we don't consider making light websites and tooling these days like in the past where hardware constraints mandated it, is that all we think it is unsexy. That is it. All this waste for a sexy css and an ai app. Funny how we make a big stink of only calling libraries we actually use in code then we reach for a chatbot trained on all the corpus of man to tell us how many oz in a cup.
"Data Center Water Usage: A Comprehensive Guide" - https://dgtlinfra.com/data-center-water-usage/
https://www.businessinsider.com/how-data-centers-are-deepeni...
>this will allow the hyperscalers to build more DCs [...]
This juxtaposition makes zero sense because the datacenters are getting built regardless of whether people are eating beef or not. By and large, there's no global cap on carbon, so everyone is free to emit whatever they want, and someone eating less beef isn't allowing amazon to train more AI models or vice versa.
People who don’t understand the scale of the problem came up with that.
JCM9•3h ago