It might be outside your area of interest as a data guy, but I'd love to see you explore how to actually convince the relevant populations to accept data centers in their communities. You previously showed that data is not effective. From your last article:
>Given how much disdain people hold at baseline, how strongly they hold their most errant beliefs, and how unconvinced they are by statements of fact, you might expect supplying facts to have little effect. You’d be right! People’s opinions moved in the direction of supporting data centers, but by a barely significant amount
As I read that quote again, it's clear you find them frustrating. That makes sense since they are saying something you know is false and operating in a way that clashes with your values (using data instrumentally rather than as a means to truth). It's really hard to build cooperation without a foundation of goodwill though. Should you trust someone who hates you and has little curiosity for you (who, in place of a clash of values, sees you as simply bad and wrong)?
That article does talk about methods to get people to agree to allow data centers. The most persuasive thing is just putting money in their pockets and patting them on the back.
That makes sense, yeah. I wonder, if someone did a study of the psychology, if they could find a way to mitigate the fear that costs less (by addressing the unspoken/unconscious concerns that are non-obvious to outsiders), but projects like this are tough to carry out and can veer unethical. I don't personally care about data centers; I just find the premise interesting!
The key observation is that data centers will cause a substantial, and unprecedented load growth.
The capacity in CA will be procured by the Load Serving Entities under a February order by the CPUC. The LSEs will enter into *long-term* contracts for that generation to serve load growth. These are obligations held by the public, since they are allowed to recoup their costs from rate payers.
On the other hand, the data centers have very, very short term risk. They must succeed in a short amount of time. If the AI business model works (and I think it ultimately will) then it is likely that average per kWh prices will go down, due to scale effects.
However, if the data center model doesn't work, or if AI demand stumbles, *the public* holds the long-term obligation to pay for the unused generation capex.
This is a classic risk duration mismatch and a shifting of risk to the public.
Since your analysis only applies to data centers that currently exist, you're naturally biasing the data to the prior case. Wald's bombers.jpg
It’s some kind of text input issue with Safari on iPad. It defaults to 106 and autocompletes if I try to delete. But, if I select all and type, I can replace it.
It might be outside your area of interest as a data guy, but I'd love to see you explore how to actually convince the relevant populations to accept data centers in their communities. You previously showed that data is not effective. From your last article:
>Given how much disdain people hold at baseline, how strongly they hold their most errant beliefs, and how unconvinced they are by statements of fact, you might expect supplying facts to have little effect. You’d be right! People’s opinions moved in the direction of supporting data centers, but by a barely significant amount
As I read that quote again, it's clear you find them frustrating. That makes sense since they are saying something you know is false and operating in a way that clashes with your values (using data instrumentally rather than as a means to truth). It's really hard to build cooperation without a foundation of goodwill though. Should you trust someone who hates you and has little curiosity for you (who, in place of a clash of values, sees you as simply bad and wrong)?
That article does talk about methods to get people to agree to allow data centers. The most persuasive thing is just putting money in their pockets and patting them on the back.
That makes sense, yeah. I wonder, if someone did a study of the psychology, if they could find a way to mitigate the fear that costs less (by addressing the unspoken/unconscious concerns that are non-obvious to outsiders), but projects like this are tough to carry out and can veer unethical. I don't personally care about data centers; I just find the premise interesting!
This is amazing! Every time I hear/read the broad statements like examples 1 and 2, my mind defaults to “ok, based on what”?
Did you see the commercial of Jason Kelce promoting Garage Beer and Liquid Death? Look it up on YT if you haven’t seen yet…
Hi!
Energy pricing analyst here
The key observation is that data centers will cause a substantial, and unprecedented load growth.
The capacity in CA will be procured by the Load Serving Entities under a February order by the CPUC. The LSEs will enter into *long-term* contracts for that generation to serve load growth. These are obligations held by the public, since they are allowed to recoup their costs from rate payers.
On the other hand, the data centers have very, very short term risk. They must succeed in a short amount of time. If the AI business model works (and I think it ultimately will) then it is likely that average per kWh prices will go down, due to scale effects.
However, if the data center model doesn't work, or if AI demand stumbles, *the public* holds the long-term obligation to pay for the unused generation capex.
This is a classic risk duration mismatch and a shifting of risk to the public.
Since your analysis only applies to data centers that currently exist, you're naturally biasing the data to the prior case. Wald's bombers.jpg
Your site doesn’t allow bills higher than 199, ours are well over $400
The site allows you to pump the bill up as high as you want. Not sure why you're seeing a limit.
It’s some kind of text input issue with Safari on iPad. It defaults to 106 and autocompletes if I try to delete. But, if I select all and type, I can replace it.
Fixed!