Thank you for writing these articles. But ultimately I feel that environmental arguments against AI are a smokescreen. Really there is a deep discomfort with AI itself, with machines that appear to talk like humans. People feel it is a threat to their humanity in a way that previous technologies weren't. This is an emotional reaction and we can't talk people out of it. All these other arguments about water and global warming are covers from that.
They arent though. The environmental impacts in the short term are very negative and real. Look up noise pollution at these sites and tell me you could live with that. We can get this right. We dont need to be so polarized about it. AI is going to lead to humanity surviving climate change and eventually thriving...but right now it is proceeding in a really 20th centurtpy dirty industrialist manner when it really does not have to. We have a lot of the tech to do this well....and a lot of people will profit from that.
For me, I worry that AI build out is a massive greenhouse gas expenditure that doesn't necessarily give an equivalent societal reward. If we all end up dying from climate change, I'm not sure I'd be able to say Claude or chatgpt were worth it.
I like this article and premise a lot and agree with it broadly, but I don't think the math in the "Carbon" section checks out. I think perhaps "the things we can buy with the tax revenue makes the negative externalities worth it" is a stronger argument than "we can use that tax revenue to explicitly cancel out the negative externalities" when it comes to carbon emissions, since I don't actually know that that's feasible (it might be, but I think the math makes it a lot closer). In particular:
1) I'm extremely skeptical of the RGGI's claimed reduction in emissions. Even if you were to somehow buy the amount of allowances required at $600 million/year, you'd only expect something like a 10-50% reduction in CO2 emissions causally from that. Much of the reduction in emissions commensurate with the implementation of RGGI came from the shale boom shifting from coal to less emissive gas, as well as other incentives for renewable energy production.
2) In general, these carbon allowances/offsets only make sense when you are *not increasing energy demand through your actions*. Imagine I am going to create a new giga-mega-hyper-datacenter that is going to consume as much energy as the entire US. This will be powered by new natural gas plants, but, I say, don't worry, we're going to finance a shift for the rest of the US grid onto renewable energy by paying allowances per tonne emmitted, and our entire emissions will be offset in this manner. Well... I think you can see why this could not possibly work. It's equally implausible at much smaller scales. It only works to a very small extent on the margins, not when we're talking about something like >50% of the entire county's energy consumption.
3) Solar/wind simply cannot replace gas generation for data centers. Currently basically no data centers run on renewables and nearly all run on gas. You'd (1) have to build a ton of very expensive 24-hr storage, (2) have to build a ton of very expensive grid improvements/modernization infrastructure, and (3) have to permit all of this stuff, which is very, very hard and would take a long time! You do mention a couple of these caveats, but I'd guess that it would be way more expensive, like maybe a factor of 10 more! If it wasn't actually that expensive, then people (Microsoft/Google) would be doing it right now (they're not).
4) Energy costs would rise a lot (demand is rising a lot!), likely more than offsetting the reductions from the program. This has to be factored in, no?
The only carbon offsets that I could ever trust are DAC + geological storage and enhanced rock weathering. The former has very high costs still (100s of $ per tonne) and the latter issues with MRV (measurement, reporting, verification) and perhaps scaling?
Other approaches have key failure points (political/legislative uncertainty and temporal mismatch as in forest credits among others).
It’s actually kind of a bummer how hard this problem is.
Will circle back, thanks! To be clear I’m not saying these could power the data centers, just that the data centers could help them get built elsewhere
I somewhat agree with Tony, but I think some of it is a more general discomfort with the rate of change that has occurred in the past three decades. The reason it is coming out now, around datacenters, is that much of that change was virtual. How do you protest your cell phone? How do you get your county commissioner to stop someone from posting something annoying on Facebook?
Datacenters are the only physical manifestation of that change that they can act on.
For the desalinization of sea water to replace lost freshwater, you don’t actually need to build a pipeline. If the desalination plant is over the same aquifer, it can just inject the processed water into the aquifer directly, which can then be drawn out where it is used. Or like Arizona and Nevada are doing with San Diego county water from Carlsbad, just make a virtual trade, which can be done because even though San Diego is farther from the Colorado river, they are allocated more water and at a lower price and can just reduce their usage and let AZ and NV use that portion for a fee that more than covers the desalination cost.
This one strikingly feels the least convincing. I might need to read it again - perhaps I've missed something, but the presentation felt lacking.
The positioning of "money not gained is money lost" is mostly a rhetorical device, the math doesn't work out. A city that doesn't gain $1.5B in tax revenue is not equivalent to a city that lost this amount. Mostly because your budget dictates your ability to spend and a larger budget changes your abilities greatly. Assuming some ordered list of priorities, a village with a budget of $1M is likely not working on environmental concerns. This changes if their budget is in the trillions.
The only equivalent framing is "if you had $1.5B more, would you spend it on these areas" which is a less powerful rhetorical device (even if convincing still on the basis of a lot of the article) and again requires framing this against their current budget.
Combined with people's general lack of alignment with how a city spends its budget, that it should forgo reducing carbon emissions in the name of capital accumulation is a really difficult argument.
I think Seth below made a similar point. My read on this mostly feels like "here's a way to trick someone who isn't thinking too hard" instead of an honest framing.
Interesting yeah, I guess I’m trying to reach for a way to force people to compare the upsides to downsides in a more mathy way and would be interested in any alternatives. Maybe another way of saying it is like “this makes your city services x amount cheaper” or something?
Yeah, although I think most people don't actively acknowledge the downsides to any infrastructure build out when assessing a government spending bill. It's just not popular to talk about the carbon emissions from a new train line, but it's very popular to do so around data centers.
The most convincing framings from your articles were around "data centers are very similar to other things".
I get the sense that lots of pushback comes from the global population being less in line with referring to the service as a (desirable) consumer good. Most arguments against data center build outs include something about how LLMs get things wrong sometimes.
I don't remember you taking a position on this point, but it seems important.
If I'm the kind of guy who hates private equity, hearing that they're building new HQs in my home town will upset me, even if I understand that my city will have an increased budget from the revenue.
Some concerns around LLMs are about job loss, others about cultural psychosis and maybe the increasing wealth disparity. If these are legitimate, the pushback seems very easy to sell. Really, you'd buy into any argument without looking too deeply if it helped you pushback on their buildout.
It seems more of a moral issue than an economic one.
The problem with the idea of "in a more mathy way" is the model used for the calculation often has a very large number of implicit assumptions and values embedded within it. The tedious task then, which very few people want to do, is to painfully go examine every such implicit assumption and estimate, pointing out problems therein. The vast majority of people will just say something along the lines of "I think you're trying to con me", and not elaborate further. Also see my humorous mathiness in this comment:
I make a more general version of this point to friends frequently. If you think the new data center is going to have some negative impact, get the government to tax and regulate it to prevent that impact. You can't claim to have no control over the government if you're trying to get the data center banned in the first place; who else is doing the banning? This all or nothing view is toxic to reasoned debate.
(1) I think that loss aversion, not status quo bias, might better describe the underlying psychological habit that you aim to exploit. People evaluate decisions not in terms of the utility of various possible outcomes, but in terms of the potential gains and losses relative to a reference point. Your advice is to move the reference point so that data center tax revenue is considered as a potential loss, rather than a potential gain, making it more valuable to the decision-maker.
(2) This tactic doesn't make your cost-benefit analysis any more salient, it just exploits human psychology to make your interlocuters more likely to agree with you (and me). Data centers are net beneficial regardless of framing, and moving the reference point shouldn't be necessary to come to that conclusion. Put simply, I think that this tactic doesn't make the other guy realize that you're right, it just tricks them into agreeing with you.
Anyway, I'm a big fan, and keep up the great work!
Another good piece overall, but the land section has me scratching my head.
The summary of the plan says “Make 3% more of the county’s land available to be used by businesses or kept as nature.” That could mean a lot of things, but “making land available” makes me think of environmental cleanup efforts for land that’s polluted.
Instead, the land section talks about agricultural easements under ACEP, which is about paying farmers to keep farming (by selling the right to develop that land beyond farming). That doesn’t sound like it either makes new land “available to businesses,” and it certainly doesn’t make that land “kept as nature.”
The estimated cost is also puzzling: You cite the ACEP cap, then seem to assert that land could be eased at that price. But ACEP easements are voluntary. There’s no guarantee that the cap is a market-clearing price for the quantity of land you want. (If it is, why didn’t those landowners already seek ACEP easements?)
Fair on status-quo bias. But the reframe quietly assumes one county both collects the revenue and pays the externality, and on electricity that isn't how it splits.
Loudoun keeps the $1.3B in property tax. The bill increase lands on Dominion's ratepayers and the wider PJM region beyond Virginia, who get none of that revenue and never sat in the hearing. Your footnote 7 has it: the cost only moves off households if a legislature votes to move it, and mostly it hasn't.
"Could offset it for a tenth of the revenue" is true, and beside the point. The offset is a maybe. The bill already showed up.
Yeah here I'd also like people to just say "Well, what are the actual numbers we're seeing and how does it compare?" There are a lot of ways counties impose externalities on each other. It's a legit objection but I'd like to see the specifics of what's happening and go from there.
Virginia's SCC ran the math when Sen. Lucas introduced SB 253: shifting data center distribution costs off residential customers would cut residential bills by $5.52/month. Sen. Schiff's federal bill is trying to put that same accounting at the permitting stage, before the bills land.
These arguments have a baked in assumption that these data centers will be paying taxes for many years. A key argument against many AI-related developments is that the economics of the sector aren't adding up.
All GenAi is currently sold below cost and it's unclear if a sizeable market would exist with the price set at cost, let alone profit. The costs of building an AI-oriented data center that then is abandoned or requires extensive repurposing is a real social burden rather than a productive benefit.
It does not seem reasonable to me that laypeople should be micromanaging business decisions. Any business could go under. Happens all the time and their assets get sold off and repurposed.
Regrets, this is a "Toxic Waste Is Good For You" type of argument. Every activist has heard some variation - Jobs! Growth! Development! Look at what you're losing!
It's very offensive to labor activists, who view it as a divide-and-conqueror rhetorical trick.
I don’t think it is! If we were dumping toxic waste on individuals and violating their rights that’d be one thing, but we need some grounded way of weighing externalities if they’re not violating rights
It's not about violating anyone's rights. It's a term meaning the argument is trying to reframe negatives (deceptively), to somehow turn them into positives.
JOHN STAUBER: No, they don't. They don't say toxic sludge is good for you. Would that they would. I mean, we picked the title based on a Tom Tomorrow cartoon. We didn't realize it at the time, but that's what put this idea into our mind, a Tom Tomorrow cartoon. And it was sort of like, Don't we wish these PR campaigns were so blatant that they actually said, 'Toxic sludge is good for you.'" Instead, they say, "Beneficial biosolids, a natural organic fertilizer, are wonderful for your garden." ...
But the right amount of (most kinds of) toxic waste in the world is not zero? Like, we probably want to allow some disinfectants in tap water, and some spent nuclear fuel, etc. So we need some way of weighing the negative externalities of different kinds of toxic waste against the benefits they allow.
Yes, that's completely true as a theoretical abstraction. But in practice, many people tends to be deeply suspicious of who is doing the weighing and whether it's "fair" weight. Even honest Utilitarian-type writers have a lot of problems dealing with that suspicion, since they tend to be culturally focused on coming up with arguments about why what leftists want is bad and what corporations want is good. But from outside, the "weighing" argument is basically viewed as running like this:
[Note, this is exaggerated for humorous effect] "Protester, let me, the Rational Utilitarian, explain to your tiny brain why you are a complete idiot. All actions involve trade-offs. There is a cost to removing pollution from tap water. Now, for that cost, we could buy a whole bunch of anti-malaria bednets for poor Africans, saving many of their lives. And you claim to care about Third World, people of color, and so on. Yet YOU WANT TO MURDER THEM, in your quest to impose your virtue-signaling, luxury beliefs about water meeting some impossible PURITY TEST! We're weighing your First World problems against having Black Lives Matter. Isn't it obvious how PollutingCorp is really the good guys here, and you're the bad guys?"
Also, if someone believes an industry is full of dubious accounting shenanigans (as even some pro-AI people do), they will be extraordinarily unsympathetic to any calculation which comes out in favor it.
Ultimately AI is marketed towards multi national corporations and billionaires as an easy way to mass fire employees. I don’t trust the US political system to create a Clement Attlee style welfare state for all the unemployed citizens.
And I’m not some foolish, naive moron to actually trust anything a billionaire promises.
Or their Political, media sycophants.
Plutocrats will fight to the death to avoid paying even 5% more in taxes.
You would have to be extremely stupid to believe in any delusional hope of higher paying positions being created.
Why is the county paying to remove the noise a data center is going to make? Is that even pissible. The polluter should pay on all counts. Why are we socializng the harms the data centers make? When socializing the profits turns conservatives red in the face?
Thank you for writing these articles. But ultimately I feel that environmental arguments against AI are a smokescreen. Really there is a deep discomfort with AI itself, with machines that appear to talk like humans. People feel it is a threat to their humanity in a way that previous technologies weren't. This is an emotional reaction and we can't talk people out of it. All these other arguments about water and global warming are covers from that.
They arent though. The environmental impacts in the short term are very negative and real. Look up noise pollution at these sites and tell me you could live with that. We can get this right. We dont need to be so polarized about it. AI is going to lead to humanity surviving climate change and eventually thriving...but right now it is proceeding in a really 20th centurtpy dirty industrialist manner when it really does not have to. We have a lot of the tech to do this well....and a lot of people will profit from that.
For me, I worry that AI build out is a massive greenhouse gas expenditure that doesn't necessarily give an equivalent societal reward. If we all end up dying from climate change, I'm not sure I'd be able to say Claude or chatgpt were worth it.
I like this article and premise a lot and agree with it broadly, but I don't think the math in the "Carbon" section checks out. I think perhaps "the things we can buy with the tax revenue makes the negative externalities worth it" is a stronger argument than "we can use that tax revenue to explicitly cancel out the negative externalities" when it comes to carbon emissions, since I don't actually know that that's feasible (it might be, but I think the math makes it a lot closer). In particular:
1) I'm extremely skeptical of the RGGI's claimed reduction in emissions. Even if you were to somehow buy the amount of allowances required at $600 million/year, you'd only expect something like a 10-50% reduction in CO2 emissions causally from that. Much of the reduction in emissions commensurate with the implementation of RGGI came from the shale boom shifting from coal to less emissive gas, as well as other incentives for renewable energy production.
2) In general, these carbon allowances/offsets only make sense when you are *not increasing energy demand through your actions*. Imagine I am going to create a new giga-mega-hyper-datacenter that is going to consume as much energy as the entire US. This will be powered by new natural gas plants, but, I say, don't worry, we're going to finance a shift for the rest of the US grid onto renewable energy by paying allowances per tonne emmitted, and our entire emissions will be offset in this manner. Well... I think you can see why this could not possibly work. It's equally implausible at much smaller scales. It only works to a very small extent on the margins, not when we're talking about something like >50% of the entire county's energy consumption.
3) Solar/wind simply cannot replace gas generation for data centers. Currently basically no data centers run on renewables and nearly all run on gas. You'd (1) have to build a ton of very expensive 24-hr storage, (2) have to build a ton of very expensive grid improvements/modernization infrastructure, and (3) have to permit all of this stuff, which is very, very hard and would take a long time! You do mention a couple of these caveats, but I'd guess that it would be way more expensive, like maybe a factor of 10 more! If it wasn't actually that expensive, then people (Microsoft/Google) would be doing it right now (they're not).
4) Energy costs would rise a lot (demand is rising a lot!), likely more than offsetting the reductions from the program. This has to be factored in, no?
The only carbon offsets that I could ever trust are DAC + geological storage and enhanced rock weathering. The former has very high costs still (100s of $ per tonne) and the latter issues with MRV (measurement, reporting, verification) and perhaps scaling?
Other approaches have key failure points (political/legislative uncertainty and temporal mismatch as in forest credits among others).
It’s actually kind of a bummer how hard this problem is.
This seems relevant (haven’t read the paper yet) https://x.com/hausfath/status/2064066655748010158
Will circle back, thanks! To be clear I’m not saying these could power the data centers, just that the data centers could help them get built elsewhere
I somewhat agree with Tony, but I think some of it is a more general discomfort with the rate of change that has occurred in the past three decades. The reason it is coming out now, around datacenters, is that much of that change was virtual. How do you protest your cell phone? How do you get your county commissioner to stop someone from posting something annoying on Facebook?
Datacenters are the only physical manifestation of that change that they can act on.
Silly Andy! There are no trade-offs. If I personally don't like something, it is Morally Wrong. End of story.
And you must always defer to anyone who yells "Carbon!" Otherwise, you are a "Climate Denier" and "Anti-Science"!
For the desalinization of sea water to replace lost freshwater, you don’t actually need to build a pipeline. If the desalination plant is over the same aquifer, it can just inject the processed water into the aquifer directly, which can then be drawn out where it is used. Or like Arizona and Nevada are doing with San Diego county water from Carlsbad, just make a virtual trade, which can be done because even though San Diego is farther from the Colorado river, they are allocated more water and at a lower price and can just reduce their usage and let AZ and NV use that portion for a fee that more than covers the desalination cost.
Excellent summary of a straightforward process to analyze subsidies and inform decision-making.
Been enjoying these articles!
This one strikingly feels the least convincing. I might need to read it again - perhaps I've missed something, but the presentation felt lacking.
The positioning of "money not gained is money lost" is mostly a rhetorical device, the math doesn't work out. A city that doesn't gain $1.5B in tax revenue is not equivalent to a city that lost this amount. Mostly because your budget dictates your ability to spend and a larger budget changes your abilities greatly. Assuming some ordered list of priorities, a village with a budget of $1M is likely not working on environmental concerns. This changes if their budget is in the trillions.
The only equivalent framing is "if you had $1.5B more, would you spend it on these areas" which is a less powerful rhetorical device (even if convincing still on the basis of a lot of the article) and again requires framing this against their current budget.
Combined with people's general lack of alignment with how a city spends its budget, that it should forgo reducing carbon emissions in the name of capital accumulation is a really difficult argument.
I think Seth below made a similar point. My read on this mostly feels like "here's a way to trick someone who isn't thinking too hard" instead of an honest framing.
Interesting yeah, I guess I’m trying to reach for a way to force people to compare the upsides to downsides in a more mathy way and would be interested in any alternatives. Maybe another way of saying it is like “this makes your city services x amount cheaper” or something?
Yeah, although I think most people don't actively acknowledge the downsides to any infrastructure build out when assessing a government spending bill. It's just not popular to talk about the carbon emissions from a new train line, but it's very popular to do so around data centers.
The most convincing framings from your articles were around "data centers are very similar to other things".
I get the sense that lots of pushback comes from the global population being less in line with referring to the service as a (desirable) consumer good. Most arguments against data center build outs include something about how LLMs get things wrong sometimes.
I don't remember you taking a position on this point, but it seems important.
If I'm the kind of guy who hates private equity, hearing that they're building new HQs in my home town will upset me, even if I understand that my city will have an increased budget from the revenue.
Some concerns around LLMs are about job loss, others about cultural psychosis and maybe the increasing wealth disparity. If these are legitimate, the pushback seems very easy to sell. Really, you'd buy into any argument without looking too deeply if it helped you pushback on their buildout.
It seems more of a moral issue than an economic one.
The problem with the idea of "in a more mathy way" is the model used for the calculation often has a very large number of implicit assumptions and values embedded within it. The tedious task then, which very few people want to do, is to painfully go examine every such implicit assumption and estimate, pointing out problems therein. The vast majority of people will just say something along the lines of "I think you're trying to con me", and not elaborate further. Also see my humorous mathiness in this comment:
https://blog.andymasley.com/p/a-simple-trick-to-fix-the-data-center/comment/271882539
I make a more general version of this point to friends frequently. If you think the new data center is going to have some negative impact, get the government to tax and regulate it to prevent that impact. You can't claim to have no control over the government if you're trying to get the data center banned in the first place; who else is doing the banning? This all or nothing view is toxic to reasoned debate.
Love the attention to rhetorical technique here!
I have two thoughts.
(1) I think that loss aversion, not status quo bias, might better describe the underlying psychological habit that you aim to exploit. People evaluate decisions not in terms of the utility of various possible outcomes, but in terms of the potential gains and losses relative to a reference point. Your advice is to move the reference point so that data center tax revenue is considered as a potential loss, rather than a potential gain, making it more valuable to the decision-maker.
(2) This tactic doesn't make your cost-benefit analysis any more salient, it just exploits human psychology to make your interlocuters more likely to agree with you (and me). Data centers are net beneficial regardless of framing, and moving the reference point shouldn't be necessary to come to that conclusion. Put simply, I think that this tactic doesn't make the other guy realize that you're right, it just tricks them into agreeing with you.
Anyway, I'm a big fan, and keep up the great work!
Another good piece overall, but the land section has me scratching my head.
The summary of the plan says “Make 3% more of the county’s land available to be used by businesses or kept as nature.” That could mean a lot of things, but “making land available” makes me think of environmental cleanup efforts for land that’s polluted.
Instead, the land section talks about agricultural easements under ACEP, which is about paying farmers to keep farming (by selling the right to develop that land beyond farming). That doesn’t sound like it either makes new land “available to businesses,” and it certainly doesn’t make that land “kept as nature.”
The estimated cost is also puzzling: You cite the ACEP cap, then seem to assert that land could be eased at that price. But ACEP easements are voluntary. There’s no guarantee that the cap is a market-clearing price for the quantity of land you want. (If it is, why didn’t those landowners already seek ACEP easements?)
That's fair, I was more lazy with it than the other sections because it seemed the least urgent. I'll circle back and edit
Fair on status-quo bias. But the reframe quietly assumes one county both collects the revenue and pays the externality, and on electricity that isn't how it splits.
Loudoun keeps the $1.3B in property tax. The bill increase lands on Dominion's ratepayers and the wider PJM region beyond Virginia, who get none of that revenue and never sat in the hearing. Your footnote 7 has it: the cost only moves off households if a legislature votes to move it, and mostly it hasn't.
"Could offset it for a tenth of the revenue" is true, and beside the point. The offset is a maybe. The bill already showed up.
Yeah here I'd also like people to just say "Well, what are the actual numbers we're seeing and how does it compare?" There are a lot of ways counties impose externalities on each other. It's a legit objection but I'd like to see the specifics of what's happening and go from there.
Virginia's SCC ran the math when Sen. Lucas introduced SB 253: shifting data center distribution costs off residential customers would cut residential bills by $5.52/month. Sen. Schiff's federal bill is trying to put that same accounting at the permitting stage, before the bills land.
These arguments have a baked in assumption that these data centers will be paying taxes for many years. A key argument against many AI-related developments is that the economics of the sector aren't adding up.
All GenAi is currently sold below cost and it's unclear if a sizeable market would exist with the price set at cost, let alone profit. The costs of building an AI-oriented data center that then is abandoned or requires extensive repurposing is a real social burden rather than a productive benefit.
It does not seem reasonable to me that laypeople should be micromanaging business decisions. Any business could go under. Happens all the time and their assets get sold off and repurposed.
Regrets, this is a "Toxic Waste Is Good For You" type of argument. Every activist has heard some variation - Jobs! Growth! Development! Look at what you're losing!
It's very offensive to labor activists, who view it as a divide-and-conqueror rhetorical trick.
I don’t think it is! If we were dumping toxic waste on individuals and violating their rights that’d be one thing, but we need some grounded way of weighing externalities if they’re not violating rights
It's not about violating anyone's rights. It's a term meaning the argument is trying to reframe negatives (deceptively), to somehow turn them into positives.
https://www.democracynow.org/1996/3/5/toxic_sludge_is_good_for_you
JOHN STAUBER: No, they don't. They don't say toxic sludge is good for you. Would that they would. I mean, we picked the title based on a Tom Tomorrow cartoon. We didn't realize it at the time, but that's what put this idea into our mind, a Tom Tomorrow cartoon. And it was sort of like, Don't we wish these PR campaigns were so blatant that they actually said, 'Toxic sludge is good for you.'" Instead, they say, "Beneficial biosolids, a natural organic fertilizer, are wonderful for your garden." ...
But the right amount of (most kinds of) toxic waste in the world is not zero? Like, we probably want to allow some disinfectants in tap water, and some spent nuclear fuel, etc. So we need some way of weighing the negative externalities of different kinds of toxic waste against the benefits they allow.
Yes, that's completely true as a theoretical abstraction. But in practice, many people tends to be deeply suspicious of who is doing the weighing and whether it's "fair" weight. Even honest Utilitarian-type writers have a lot of problems dealing with that suspicion, since they tend to be culturally focused on coming up with arguments about why what leftists want is bad and what corporations want is good. But from outside, the "weighing" argument is basically viewed as running like this:
[Note, this is exaggerated for humorous effect] "Protester, let me, the Rational Utilitarian, explain to your tiny brain why you are a complete idiot. All actions involve trade-offs. There is a cost to removing pollution from tap water. Now, for that cost, we could buy a whole bunch of anti-malaria bednets for poor Africans, saving many of their lives. And you claim to care about Third World, people of color, and so on. Yet YOU WANT TO MURDER THEM, in your quest to impose your virtue-signaling, luxury beliefs about water meeting some impossible PURITY TEST! We're weighing your First World problems against having Black Lives Matter. Isn't it obvious how PollutingCorp is really the good guys here, and you're the bad guys?"
Also, if someone believes an industry is full of dubious accounting shenanigans (as even some pro-AI people do), they will be extraordinarily unsympathetic to any calculation which comes out in favor it.
Labor activists, famously known to be rational actors.
Ultimately AI is marketed towards multi national corporations and billionaires as an easy way to mass fire employees. I don’t trust the US political system to create a Clement Attlee style welfare state for all the unemployed citizens.
And I’m not some foolish, naive moron to actually trust anything a billionaire promises.
Or their Political, media sycophants.
Plutocrats will fight to the death to avoid paying even 5% more in taxes.
You would have to be extremely stupid to believe in any delusional hope of higher paying positions being created.
Why is the county paying to remove the noise a data center is going to make? Is that even pissible. The polluter should pay on all counts. Why are we socializng the harms the data centers make? When socializing the profits turns conservatives red in the face?
I am not againt them per se but the rollout is horrendously unfair and unbalanced.