There is so much nuance and context missing that I can't see this as anything other than astroturfing.
I can get behind "AI water use is not a serious concern" if all you are talking about is selling inference, and you're comparing some sort of usage metric (e.g. "water use per request"). Water and power use for inference is on the level of other heavy Internet products like video streaming or cloud compute.
There is a lot I can't ignore, though. Model training is incredibly demanding, so much that OpenAI was trying to get $1 trillion in investment to practically double the number of data centers in the United States by 2030. That is a serious concern when we have to make decisions between, say, consumer water availability and tech investment in water-scarce areas like Arizona and New Mexico.
In Oregon, there are some unique problems with Amazon's water deals in Umatilla, where they are increasing nitrate concentration of the local groundwater through evaporative cooling, and refusing to pay for on-site treatment.
I can go on about other environmental harms, but I think you should take a more nuanced look at the issue. Having ChatGPT summarize a news article is not an unreasonable demand compared to other compute activities, but AI in general is driving compute demand so high that the general public is forced to reckon with a problem that's been there since the beginning: the expansion, operation and use of the Internet has physical environmental consequences.
It's completely trivial when you compare it to livestock agriculture's water usage (and even more trivial when looking at overall environmental damage), yet mysteriously 99% of people clamoring against GenAI's water usage have never opened their mouths about that. Or ever skipped a hamburger due to it. It's a joke. There are a hundred other good reasons to criticize GenAI, go pick one of those. The truth is that people are embarrassed of the real main reasons they dislike GenAI, yet they shouldn't be. Unlike this one, they're perfectly valid reasons.
I've known this for a long time and ironically this very article proves me right. The article is straight LLM slop, right from the mouth of GPT, Gemini or Claude. It's overt. Out of all 30 articles on the current HN frontpage, this one is by far the highest % LLM-written of all of them - I took the time to skim through each of them. This tells you all you need to know. With this in mind it's poor behavior to accuse GP of all people of astroturfing.
The word "fake" draws attention but I think the article obscures two real problems:
Training is missing from the analysis entirely (as someone else noted)
Inference water use is indeed minimal per prompt no argument there, but training the old GPT-3 consumed roughly 5.4 million liters of water. LLaMA 3: ~22 million. These are huge events, happening multiple times a year across the industry, folding them into national averages seems like the statistical simplification he article criticizes everyone else for doing…
"Small nationally" ≠ "fine locally"
The Dalles, Oregon is the clearest example. In 2012, Google used 12% of the city's water supply. Today it consumes a third, around 1.19 million gallons per day, and well a sixth data center comes online in 2026, in the same area.
The city is now pursuing a $260 million reservoir expansion into a national forest (!), where 95% of the projected new water demand will be industrial, not residential. Residents are looking at a potential 99% rate increase by 2036 to fund infrastructure that may exists primarily to serve one company. Apparently the city fought a 13-month legal battle just to keep those numbers secret, that’s like a community being reshaped around a single tenant.
Hays County, Texas residents sharing the Edwards Aquifer with incoming data centers voted to block one. Memphis is watching xAI draw 5 million gallons per day. Bloomberg found two-thirds of new U.S. data centers since 2022 are sited in high water-stress zones. Arizona have already passed ordinances capping data center water use.
This to me looks like a problem in the making, AI water use isn't a national crisis for now, but local impacts are already real, training costs are systematically underreported, and the five year trajectory in water stressed regions deserves serious attention indeed
There are many good criticisms against data center. And yet, the water issue always comes up first. Must we spew false/untruthhood just so our political message is catchy? I suppose yes - in times of war/politics, the laws/truths are silent. But it doesn't have to be so here.
I've never had it come up first. Neat how 2 people can have 2 opposite experiences based on their different life paths.
Anyways: Between our 2 opposite experiences, it might as well be totally random, so I don't think the ordering of concerns is that important. Better to focus on substance, like the concerns themselves.
> Why not let Amazon take a stab at turning things around, then?
The point the article is making is that iRobot's bad decisions are the reason the company was failing. Blaming regulators for a poor acquisition outcome may be fair, but they were a very minor part of the outcome.
According to the article, there is a defence invoking that exact sentiment if a merger is about to be blocked, bu iRobot decided not to invoke it (likely because it would have caused the price to be lower).
Why the price would be lower? Presumably the price was already agreed upon. Having a provision in the contract that the price is reduced if this argument is made to the antitrust authorities makes no sense.
I now realized that I recently saw some old tweets from this guy where he first opposed this merger and then celebrated the cancellation of the deal. So it seems he's just grasping at straws to look less like an idiot.
If the buyer is publicly listed, they have to consider their own share price,that is - the perception of their shareholders. The buyer will be punished for overpaying for a company circling the drain, they will also be punished from waling away from purchasing a seemingly healthy company at a good price.
The buyer and purchaser have to agree on which of the 2 options they publicly present to their shareholders and regulators - Amazon wouldn't be fooled because they had access to the financials and had its own judgement on viability (which may not be material Amazon's plans). However, the approach would decide the corresponding offer, and typically wouldn't be a retrospective decision, but would lock-in the higher or lower from the start.
They'll just hire the engineers they need out of the failed iRobot and not compensate the investors / founders for building something worth acqui-hiring.
The existing Roomba revenue stream probably doesn't matter. The expertise or maybe the brand (not a great brand imho) aligns with some company priority.
Free money? I honestly don't understand comments like this. It's as if you aren't even trying to make sense. Amazon would have been bailing out Roomba so if anything this would have cost Bezos money.
If by "open and freely accessible" you mean there should be no rules of the road, then I suppose yes. Personally, I'm glad CF is pushing back on this naive mentality.
If you actually bother to click through and read the article, you'd find the court expressed sympathies with the intent of the rule, but the FTC "is required to conduct a preliminary regulatory analysis when a rule has an estimated annual economic effect of $100 million or more", and they did not do that.
The blame here belongs to the FTC for its rushed and sloppy process that put the rule on shaky ground legally.
Depends on how accurate you think the >$100 million estimated impact from the lower court is. When the FTC did the analysis they came up with a lower impact so they didn't have to do it. I'd be more willing to believe they got it right than a single judge did.
Why does it matter? As far as I can tell () the law asks the FTC to do an estimate, they did, and now the argument was ‘some one else thinks it’s wrong’. But does the law require an actual estimate?
If they are worried about this… either mandate some third party do the estimate, or mandate the study. This is just confusing.
() - of course I haven’t read the actual law or ruling yet…
The FTC did their estimate, it came out to less than $100 million so they, to their estimation were not required to take the additional steps of some types of impact analysis and the required comment period etc about it, the courts are disagreeing with the FTC and kicking the rule out. They only have to do that extra step when the economic impact would exceed $100m USD.
The main reason I think the court got it right is that with ~33 million businesses in the US you could argue that sending every business an email would cost them >$100 mil in just labor cost if they forward it a few times and several employees spend a reasonable amount of time reading it.
Luckily not all 33M of those businesses are wringing subscriptions out of their customers (yet), so it might be fairer if we could narrow it down to the subset who do?
What's more interesting to me is the court is basically admitting that doing the right thing for customers will cost unscrupulous businesses more than $100M they're currently fleecing those customers for, so they won't let this go ahead.
> Luckily not all 33M of those businesses are wringing subscriptions out of their customers (yet), so it might be fairer if we could narrow it down to the subset who do?
This is quite correct. The FTC estimated that there are 106,000 businesses offering subscriptions to Americans. Those are the ones that would have to comply with the new rules.
A hundred million dollars is a lot of money, but divided by a hundred thousand businesses it suddenly is only $1000 each. Not actually that much! Since the new rules proposed by the FTC include a lot more than just a button on a website, complying with the rules would in fact require every one of these companies to do a fair amount of work. They’ll have to review all of their existing marketing material, and all of the forms that they use to sign up customers. They must ensure that all material facts are disclosed to every prospective customer, and that consent is obtained from the customer correctly.
Certainly these are good rules for businesses to follow, but the question now is the cost. Can your business review all of its marketing material for less than $1000? I doubt it. So the judge rightfully noted that the FTC’s estimate of the impact was insufficient.
And all they have to do as a result is to allow additional public comment on the proposed rules, with the specific intent to find alternatives. If these alternative rules would be just as effective but cheaper to comply with then the FTC is supposed to drop their own proposed rules and adopt the alternatives. They had already done some of this in the earlier phases of the process, and the result was that several unworkable rules were indeed dropped. They could have spent a few more months doing the final review and analysis, but they decided to rush it through instead.
> What's more interesting to me is the court is basically admitting that doing the right thing for customers will cost unscrupulous businesses more than $100M they're currently fleecing those customers for, so they won't let this go ahead.
I’ll say it again that this has nothing to do with how much the unscrupulous are getting away with, or whether it would cost the unscrupulous businesses anything at all. This is entirely about the cost to the legitimate businesses.
Why do you think the FTC analysis was more accurate than the opposing sides? The judges, of whom there were multiple, were going off of opposing side argumentation not just their own subjective opinion. That's how courts in the US work.
The companies suing to stop this have every reason to massively inflate the difficulty and cost of compliance to continue their long established dark patterns of trapping people in difficult to cancel subscriptions. Judges are not experts in the field and have a hard time evaluating the actual credibility of various presented estimates, you see it all the time with long debunked forensic evidence techniques being accepted still years later by judges and courts.
To figure out who's right, we would need to do research, rather than choosing the judges versus the FTC based on vibes.
I'm hardly going to do that research myself, so I have no opinion. There are legal bloggers whose opinions I'd respect. I assume comments on Hacker News are no more informed than my own, unless they show they have relevant expertise.
Can you point to specific examples where the courts interpretation and reasoning wasn't rooted in the law and the arguments from the lawyers in the court specificall? Because I can't. I might disagree with some of the opinions but I can't point to anything where they were clearly not basing it on the law and the various lawyer's argumentation.
If you believe the judges are all being as objective as possible and that the systemic flaws in our judicial system are not being exploited by the Trump admin/lawyers representing the Trump admin without pushback from said judges than you can’t be convinced by me. I can cite an example, and if you aren’t personally convinced you can simply say “the court of law said x was innocent or y was held liable” or whatever you consider the correct arbiter.
You can appeal to the system that I am actively saying is not working but you get the benefit of “well the law/eatablishment says so.” You can lean on your opinion or the results, whichever serves you better, in a way I can’t as you cite the case itself. The entire discussion is poisoned out the gate.
No, that is not my position. I don’t know where you got that from. My position is advocating for healthy, much-deserved skepticism of the current judicial system. More specifically, I am saying that an appeal to some alleged neutrality/objectivity judges are supposed to maintain when carrying out their duty is particularly flimsy in 2025.
I don't know how you advocate for such a thing without holding judges to a standard of neutrality and objectivity and acknowledging/celebrating it when it happens.
I don’t understand what you’re saying anymore. It sounds like you’re agreeing with me? If we have to celebrate when it happens, then that means it’s not a standard expectation. It’s special and worthy of celebration.
Good catch. That whole pipeline is very robust too - I remember mocking the Burke society folks in college, not realizing at all that they were figuring out who was headed to work with federal judges/think tanks for the next generation.
Why would this have any economic impact? These dark patterns don't generate any net value, they just move money from one pocket to another. The money will be spent somewhere else, instead.
The economic impact here is only factoring in how much it would cost companies to comply with the measure which is inherently designed to give an extra hurdle by not counting the money saved by consumers not trapped by dark cancellation patterns.
Right. It doesn’t count the money consumers save by not being trapped primarily because most businesses are legitimate. This threshold is all about identifying rules that are costly for legitimate businesses to implement and allowing them time to suggest alternative rules. If the alternative rules would be just as effective while being cheaper to implement, then the FTC is supposed to drop their own proposed rules in favor of the cheaper alternatives.
I don't understand, they did an estimate and found it below 100$ million. That seems to have followed the process. An estimate can always be challenged and is just a best effort prediction. Now it seems this create a pretty flaky ground for precedence that the FTC simply can never estimate less than 100$ million as it could always be challenged in court, what if it was more? And they now have to always follow the more effortful process of assuming it is more than 100$ million.
It really seems like a weird line in the sand that the court will just randomly decide on a case by case now, with the FTC having no way to know if the court will agree with their estimate or not.
> you'd find the court expressed sympathies with the intent of the rule
And you'd find such sentiments to be completely worthless, except insofar as they act as cover for a ruling on a technicality in favor of the same corporate interests that fund the politicians that appointed these judges.
Sure, but that's a different argument. OP wasn't carrying water for the companies that would be affected by this change, they were carrying water for the rule of law. If the FTC had sued saying that the $100M limit was too restricting and had no valid basis, then sure, this would be a valid argument. But the judges have to rule on the law as it's written, not as HN commenters would like it to be. Is the law wrong? Corrupt? Maybe! But that's a different conversation.
Believe me, I'm incredibly disappointed that this didn't work. I paid a Planet Fitness membership for a year after I had moved to a place too far away from any PF location to reasonably use it, just because the cancellation process was so convoluted that it took me ages to figure out how to cancel. I think that companies should be held liable when they employ predatory business practices like this. I agree with your premise, that the limit is too low and there's nothing to stop companies from lying about the cost to implement the rule. But the law is the law is the law is the law. Courts exist to interpret the law, and in this case, the law they were asked to interpret was whether the FTC had abided by the $100M cap. They found reasonable justification to rule that they had not.
Again, I get the desire to be up in arms over this. But recent events have shown just how fragile our legal system is when people decide that the rules can just be ignored, and I wish that people would be more hesitant to throw the baby out with the bathwater, even when doing so would mean I wouldn't have to pay planet fitness $20/mo for a year.
Tidy logical explanations of rule systems that click for people are very powerful when they come from authority. There’s a comfort in this sort of bureaucracy that appears to have taken broad considerations to protect us from complex dangers and second order effects.
A shoddy implementation would just mean later problems. Hopefully the FTC gets the memo and does it "the right way" to make it watertight, otherwise people will just get away with doing whatever they want.
I find it quite absurd to suggest that a ruling that forces corporations to simplify their business practices would cost $100 million to implement. I would be willing to bet the companies would significantly reduce their spending by following these rules, not increase them. But it's not in any way surprising that they'd find a judge who buys their massively overinflated estimates over the FTC's calculations.
This is a pretty narrow view. A lot of businesses--whose bread and butter (well maybe just the butter) is keeping people locked into subscriptions they don't want--put a large effort in challenging this rule. They would have fought it like hell during the "analysis" which would have stretched into the Trump presidency were it would surely would have been killed. Even if the analysis had been completed, it's likely the courts would have struck it down for overreach (like Dept of Education's student loan forgiveness). It died because a lot business interests are opposed to it.
I don't know what you mean by "narrow" here. It sounds like you're saying that they did it at the last minute, and failed to finish. But you're saying that since the next administration would "surely" never do click to cancel, that somehow should immunize the FTC from following their own regulations? The next administration was elected.
The reason they have to do studies is so they can't rush things through. We don't want them to be able to rush things through. They're creating law.
In this case, I'm guessing the FTC knew it was a long shot and took the Hail Mary pass anyway. If they had done the preliminary regulatory analysis the ruling wouldn't have been completed during the Biden administration. So they gambled that it would be better to take their chances with the courts than with the next administration, given both Republican commissioners voted against the rule. Which makes this less of a disappointment to me that it would otherwise.
It’s amazing how twisted the term “anti-competitive” has become. Where anti-competitive companies push for anti-competitive regulations under the false pretense of preventing anti-competitiveness.
Standard Oil is being competitive.
The U.S. oil refining and distribution industry is being anti-competitive.
Because they suck at competing against Standard Oil and they want to get unfair, unethical advantage themselves.
Imagine spending years and billions building something and then I show up and say “hey man that’s not fair, give me a slice of that thing”.
And before someone tells me “that’s the Sherman Act”, I don’t care. If that’s the law then it should be changed. Laws have been written (and lobbied) for all sorts of reasons and surprisingly not all of them are fair and ethical.
(I hope this illustrates how easy it is to make this exact argument about literally any monopoly.)
I don't think browsers will ever let web code affect things outside the viewport because scammers would cook up some truly zany things with that power.
Even rendering arbitrary pixels inside the viewport is bad enough. Something that went out of fashion but is apparently now back in fashion is detecting the user's operating system and browser, then displaying a pixel-perfect replica of a second browser window open to Paypal and asking you to log in... displayed within the bounds of the first browser window.
This is why the new login prompt of 1Password is worse than the old one, it appears at the center of the screen where the website could easily put a replica. The old one opened at the height of the extension icon, a bit above where the browser opens the alert dialog
It is a completely fake concern. See here: https://blog.andymasley.com/p/the-ai-water-issue-is-fake
reply