[Permit]

> You should consider dropping that instinct.

This is the reason we have people mistakenly repeating the conclusion that AI consumes huge amounts of water comparable to that of entire cities.

If you make any other assumption than "I don't know what's happening here and need to learn more" you'll constantly be making these kind of errors. You don't have to have an opinion on every topic.

Edit: By the way, I also don't think we should trust big companies indiscriminately. Like, we could have a system for pesticide approval that errs on the side of caution: We only permit pesticides for which there is undisputed evidence that the chemicals do not cause problems for humans/animals/other plants etc.

[peppersghost93]

"If you make any other assumption than "I don't know what's happening here and need to learn more" you'll constantly be making these kind of errors. You don't have to have an opinion on every topic."

I can do this and still start off by assuming the corporation is in the wrong. The tendency to optimize for profits at the expense of everything else, to ignore all negative externalities is inherent to all American corporations.

[Workaccount2]

The main thing that people snag on is scale and frequency.

If you are super into "ACAB" (all cops are bastards) you can easily "research" this all day for weeks and find so many insane cases of police being absolute bastards. You would be so solidified in your belief that police as an institution are fundamentally a force of evil.

But you would probably never come across the boring stat that less than 1 in 500,000 police encounters ever register on the "ACAB" radar.

This is almost always where people run aground. Stats are almost always obfuscated for things that people develop a moral conviction around. Imagine trying to acknowledge the stat there are effectively zero transgender people perving on others in public bathrooms.

[JumpCrisscross]

> I can do this and still start off by assuming the corporation is in the wrong

You really can't. You can start off with a prior that it's more likely the corporation is wrong than not. But if you're assuming your conclusion, you're going to find evidence for what you're looking for. (You see the same thing happen with folks who start off by assuming the government is in the wrong.)

[api]

When you go shopping and see two items for sale that seem nearly identical, do you buy the cheaper one?

If you have long term savings do you want it to earn interest?

The desire to optimize for profit exists at all levels among all participants in the economy. Everyone does it. We are the system and the system is us.

Regulations are usually the only way to fix these things because there are game theoretic effects in play. If your company spends more to clean up and others don’t, you lose… because people buy cheaper products and invest in firms with higher profit margins. The only way out we’ve found is to simultaneously compel everyone. But that doesn’t remove the incentive.

[ljsprague]

Many individuals optimize for profits too.

[dmos62]

It might seem like bias will get you to where you're going faster, but at the end of the day it's just bias.

[CGMthrowaway]

>people mistakenly repeating the conclusion that AI consumes huge amounts of water comparable to that of entire cities

Does it not?

"We estimate that 1 MWh of energy consumption by a data center requires 7.1 m3 of water." If Microsoft, Amazon and Google are assumed to have ~8000 MW of data centers in the US, that is 1.4M m3 per day. The city of Philadelphia supplies 850K m3 per day.

https://iopscience.iop.org/article/10.1088/1748-9326/abfba1/...

[jeffbee]

Why do we need to assume so many things, when we can peg it to reality.

Worldwide, Google's data centers averaged 3.7GW in 2024. Globally, they use 8.135e9 gallons of water in the year, which is 30.8e6m³ per year, which is 84e3m³ per day. Double that to meet the assumed 8GW data center capacity, 168e3m³/day. QED: the estimate 1.4e6m³/day is high by a factor of 10x. Or, in other words, the entire information industry consumes the same amount of water as one very small city.

I believe this is why Google states their water consumption as equivalent to 51 golf courses. It gives a useful benchmark for comparison. But any way you look at it the water consumption of the information sector is basically nothing.

[marcyb5st]

Yeah, but that is for everything. YouTube, Amazon itself, AWS, Azure, GCP, ... not just AI stuff. I mean, it is still a lot of water, but the numbers are not that easy to calculate IMHO

[shadowgovt]

Resource consumption of AI is unclear on two axes:

1) As other commenters have noted: raw numbers. In general, people are taking the resource consumption of new datacenters and attributing 100% of that to "because AI," when the reality is generally that while AI is increasing spend on new infrastructure, data companies are always spending on new infrastructure because of everything they do.

2) Comparative cost. In general, image synthesis takes between 80 and 300 times fewer resources (mostly electricity) per image than human creation does. It turns out a modern digital artist letting their CPU idle and screen on while they muse is soaking significant resources that an AI is using to just synthesize. Granted, this is also not an apples-to-apples comparison because the average AI flow generates dozens of draft images to find the one that is used, but the net resource effect might be less energy spent in total per produced image (on a skew of "more spent by computers" and "less by people").

[baq]

how much is it in burgers and steaks? serious question

[MSFT_Edging]

AI water usage is pretty bad on a local scale where a large water consumer(Data centers) start sucking up more water than the local table can bear at the expense of the people living there.

Even if the general takes seen on water use is wrong, it's correct in that these companies don't have the best in mind for the average person. It's correct that these companies will push limits and avoid accountability. It's correct that they're generally a liability creating a massive bubble and speculation based on an immature tech designed to automate as many careers away as possible without a proposed solution to the newly unemployed besides "deliver fast food" or "die".

Despite legally treating corporations as people, there's no consistently enforced mechanism that can punish them like people. Monsanto can't be sent to jail for murder. Their C-Levels will never see a cell the way the average person can have the book thrown at them for comparably minor crimes.

Because companies cannot be held accountable legally and effectively, it's important to assume the worst, to generate the public blowback to hold them accountable via lost business.

[gruez]

>Even if the general takes seen on water use is wrong, it's correct in that these companies don't have the best in mind for the average person.

That just sounds more like cope than anything else. eg. "AI companies sucking up all the water might not be a real issue, but I still think they're evil for other reasons".

[SoftTalker]

"undisputed evidence that the chemicals do not cause problems"

Impossible standard. You cannot prove a negative.

But, I think it's fair to assume that any chemical that is toxic to plant or insect life is probably something you want to be careful with.

[hgomersall]

Nonsense, if we view proving as providing evidence for, then absolutely we can prove a negative. We have our priors, we accumulate evidence, we generate a posterior. At some point we are sufficiently convinced. Don't get hung up on the narrow mathematical definition of prove (c.f. the exception[al case] that proves [tests] the rule), and we're just dandy.

[JumpCrisscross]

> Impossible standard. You cannot prove a negative

It's also a deep incumbency advantage. Of course the guys selling the existing stuff are going to dispute the safety of a competitor.

[jjgreen]

You cannot prove a negative.

How about Fermat's last theorem?

[christophilus]

Your edit was a good one.

It's a rational default position to say, "I'll default to distrusting large corporate scientific literature that tells me neurotoxins on my food aren't a problem."

As with any rule of thumb, that one will sometimes land you on the wrong side of history, but my guess is that it will more often than not guide you well if you don't have the time to dive deeper into a subject.

I'm not saying all corporations are evil. I'm not saying all corporate science is bad or bunk. But, corporations have a poor track record with this sort of thing, and it's the kind of thing that could obviously have large, negative societal consequences if we get it wrong. This is the category of problem for which the science needs to be clear and overwhelming in favor of a thing before we should allow it.

[jollyllama]

Indeed. Every rule has an exception but heuristics are useful.

[qarl]

Not at all. NOT AT ALL.

There are shades of gray here. But you are absolutely not required to extend benefit of the doubt to entities that have not earned it. That's a recipe for disaster.

Personally, I find myself to be incredibly biased against corporations over people. I've met a lot of people in my life, they seem mostly nice if a bit stupid. Well intentioned. Selfish.

Are corporations mostly well intentioned? Well, consider that some people tried to put "good intentions" into corporations bylaws and has been viciously resisted.

Corporations will happily take everything you have if you accidentally give it to them. Actual human beings aren't like that.

[JDEW]

> …undisputed evidence… do not cause problems…

This is unworkable in practice; nothing will ever be completely safe. Instead, we need a public regulatory body that makes reasonable risk/reward tradeoffs when approving necessary chemicals. However, this system breaks down completely when you allow for lobbying and a revolving door between the public and private sectors.

[pajko]

https://www.reddit.com/r/explainlikeimfive/comments/yzlsoj/e...

[more_corn]

Ai does us a crap-ton of water. Most data centers use closed loop liquid cooling with heat exchangers to water cooling. (At least all the big ones like Google and Amazon do)

I’m curious what evidence you think you’ve seen to the contrary. from my side, I used to build data centers and my friends are still in the industry. As of a month ago I’ve had discussions with Google engineers who build data centers regarding their carful navigation of water rights, testing of waste water etc.

[seg_lol]

> Most data centers use closed loop liquid cooling with heat exchangers to water cooling.

If these data centers are so water efficient, please explain the Dalles data center use > 25% of their water supply?

https://web.archive.org/web/20230130142801/https://centralor...

https://web.archive.org/web/20251014013855/https://www.orego...

[SV_BubbleTime]

Parent says consume, you write use.

I’ve been unclear on this. What datacenter out there is using an open loop cooling system that does not return the water after cooling for other uses?

It seems extremely inefficient to have to filter river water over and over then to dump it into the ground so deep it doesn’t go back to getting into an aquifer.

[gamblor956]

AI does consume huge amounts of water comparable to entire cities. A single AI facility consumes more water than most cities.

That AI consumes somewhat less water than cities of millions is not a defense.

[nradov]

No that's incorrect. Now you're just lying and making things up.

[deanishe]

> We only permit pesticides for which there is undisputed evidence that the chemicals do not cause problems for humans/animals/other plants etc.

This is broadly how the EU operates. If companies want to start putting new stuff in/on our food, they first have to demonstrate it's no dangerous.

Not the best approach to everything perhaps, but I'd rather not have capitalists freely innovating on my food.