I think it's super cool that Olympia HS has a student run newspaper, but I don't think this is something that should be posted to HN. The only source quoted on the water issue is an EE professor from a school in California, who I am guessing is not a subject matter expert on water in Washington state.
FWIW, as a Washington resident, I can say that we're not exactly a state worrying about water shortages. We're probably one of the more reasonable places to build data centers due to cheap green energy and pretty plentiful water. Obviously, we need to manage it responsibly, but I haven't seen any evidence of looming issues here (please feel free to correct me, though).
> I think it's super cool that Olympia HS has a student run newspaper, but I don't think this is something that should be posted to HN. The only source quoted on the water issue is an EE professor from a school in California, who I am guessing is not a subject matter expert on water in Washington state.
FWIW, as a Washington resident, I can say that we're not exactly a state worrying about water shortages. We're probably one of the more reasonable places to build data centers due to cheap green energy and pretty plentiful water. Obviously, we need to manage it responsibly, but I haven't seen any evidence of looming issues here (please feel free to correct me, though).
I agree the lack of source in TFA is less than ideal, and the author is essentially saying "just trust the professor bro".
But you have to admit it's ironic your claim has the same problem, essentially "just trust me bro".
I pointed out the very specific major flaw, which is that it had one not entirely relevant source, which makes sense for a high school newspaper.
I'm sure that high school journalism has some great outliers, but in general, I don't think we're the intended audience, nor that the journalistic standards are up to what we'd expect from a better source.
Did you read the article? It's six tiny paragraphs and provides hardly any actual data or reporting.
If this was a positive piece about AI of similar quality, I can't help but suspect you'd be responding differently.
> The only source quoted on the water issue is an EE professor from a school in California, who I am guessing is not a subject matter expert on water in Washington state.
No they don’t. Do you really believe that? Maybe on certain niche issues the opinions of a HS student are useful, but mostly they are still growing into some understanding that can contribute in a meaningful way. Which means mostly their opinions are dumb and useless.
I mean, take your position to its natural conclusion: there are people who understand more than you about basically any given topic, which means your opinions are dumb and useless.
How much higher would the energy cost be without evaporative cooling? It doesn’t seem that hard to use water-air heat pumps to get rid of the heat without any water use, so the reasons it’s not used are probably economic. I suppose you could just make water more expensive?
Indirect evaporative systems average around COP 17.5 and dew-point systems can hit ~30, so your numbers check out for the best cases. Worth noting that direct liquid cooling with dry heat rejection is now achieving PUE 1.03-1.06 with near-zero water, which narrows the effective gap considerably for the high-density AI racks that are driving most new builds.
> Despite efforts to enforce clean energy, the bill died in committee
I wish it was easy to force issues like this into ballot measures. The citizenry should be able to rip control out of the hands of their representatives when so motivated.
The amount of water mentioned in the article is completely inconsequential. Per the article, across 126 data centers they consume several foot-acres of water per day. That is incredibly efficient!
Annualized, that is 0.0001% of the water used to produce subsidized corn ethanol. If we can afford to waste that much water on corn ethanol subsidies then we can definitely afford the water for data centers.
HB 2125 was killed by the Democrats because it was a deeply unserious bill unrelated to this. For example, it required data centers to turn off their power during ordinary periods of high electricity usage. Because, you know, we can just randomly turn off the Internet during the day and there will be no bad consequences.
That's the thing with evaporation: you don't want your water to leave stuff behind after it evaporates because that will foul your equipment and cause lower efficiency.
You could in principle design systems with enough fouling mitigations that you'd be fine, but its likely that the cost of those mitigations is roughly the same as just purifying the water up-front.
It's funny that in movies like the matrix they imagine that humanity would fight back against the machines. In reality the first thing ai will do, which it has already done, is capture our governments through the application of money, and then the humans would first have to defeat their own institutions before they can even begin to fight the machines. Neoliberalism is profoundly unable to deal with threats if the threats produce short term profits. That goes for housing shortages, global warming, health care costs, falling birth rates, across the board if it produces short term profits that can be used to bribe politicians its impossible to address. AI is no different.
29 comments
FWIW, as a Washington resident, I can say that we're not exactly a state worrying about water shortages. We're probably one of the more reasonable places to build data centers due to cheap green energy and pretty plentiful water. Obviously, we need to manage it responsibly, but I haven't seen any evidence of looming issues here (please feel free to correct me, though).
> I think it's super cool that Olympia HS has a student run newspaper, but I don't think this is something that should be posted to HN. The only source quoted on the water issue is an EE professor from a school in California, who I am guessing is not a subject matter expert on water in Washington state. FWIW, as a Washington resident, I can say that we're not exactly a state worrying about water shortages. We're probably one of the more reasonable places to build data centers due to cheap green energy and pretty plentiful water. Obviously, we need to manage it responsibly, but I haven't seen any evidence of looming issues here (please feel free to correct me, though).
I agree the lack of source in TFA is less than ideal, and the author is essentially saying "just trust the professor bro".
But you have to admit it's ironic your claim has the same problem, essentially "just trust me bro".
Why shouldn't it? The thoughts and opinions of high schoolers matter just as much as those of adults.
I'm sure that high school journalism has some great outliers, but in general, I don't think we're the intended audience, nor that the journalistic standards are up to what we'd expect from a better source.
Did you read the article? It's six tiny paragraphs and provides hardly any actual data or reporting.
If this was a positive piece about AI of similar quality, I can't help but suspect you'd be responding differently.
> The only source quoted on the water issue is an EE professor from a school in California, who I am guessing is not a subject matter expert on water in Washington state.
>The thoughts and opinions of high schoolers matter just as much as those of adults.
But...it's not purporting to be an opinion piece. It seems intended to be a factual news article.
If this was an opinion column, I'd almost be more inclined to give it a pass.
> Despite efforts to enforce clean energy, the bill died in committee
I wish it was easy to force issues like this into ballot measures. The citizenry should be able to rip control out of the hands of their representatives when so motivated.
Annualized, that is 0.0001% of the water used to produce subsidized corn ethanol. If we can afford to waste that much water on corn ethanol subsidies then we can definitely afford the water for data centers.
HB 2125 was killed by the Democrats because it was a deeply unserious bill unrelated to this. For example, it required data centers to turn off their power during ordinary periods of high electricity usage. Because, you know, we can just randomly turn off the Internet during the day and there will be no bad consequences.
You could in principle design systems with enough fouling mitigations that you'd be fine, but its likely that the cost of those mitigations is roughly the same as just purifying the water up-front.