Very fun idea. I do think this puts the water argument into a bit more perspective. Your average person should drink ~3L of water per day, or about 22mL per 11 minute average-length chat that Quili quotes on their site. It's hard to pin down exactly how much water AI uses, but something on the order of 5mL per interaction seems reasonable [1]. So the people doing this are using about 4x more water than an AI model would. Of course, it does matter what happens to that water, whether it turns into water vapor or urine and how much of that water re-enters the water cycle, but it's an interesting comparison.
That's true, but the people could also be doing some other job with their effort too. If you factor in the water consumption of the energy used to power their laptops while answering these questions (~30W of power for 11 minutes per interaction, 1mL of water per W of power) you get another 5mL of water per interaction purely in electricity use.
Not saying it's a 1:1 comparison but it puts it in perspective.
That would translate to 11mL of water per interaction. Plus the 5mL for running the laptop that I pointed out in my other comment. I think the comparison still holds up with that math.
This is a neat idea but some of the claims on that website seem extremely propagandistic. I'm not particularly gung ho about datacenters but if datacenters are causing water issues the solution is to shine a light on alternative technologies to solve the problem. Perhaps Chile could become a leader in building cleaner datacenters but thats not the goal I am taking from this initiative.
I feel like I'm living in an alternate reality, because I have no idea what this page is about. Something with data centers using up all the water? Reading the comments here and people arguing dcs dont use that much water compared to regular people who drink water, confuses me even more. Like whats next? People are told to drink less and fork over their RAM? That would be wild
Whoa. This is amazing. I think What is interesting is whats driving this: the psychologically formed habit of asking questions to bots rather than search engines.
If the median query on the internet is now and forever answered by an AI bot, humans could just as easily become really important as an "authentic" source of answers.
I agree with this, but I can’t help but reflect on how we poisoned our own well by destroying the human internet before AI.
The answers to almost every search are already out there, but we made the public internet so toxic and user hostile, you now need a whole new layer of AI to sift through the bullshit, ads, and clickbait.
I don’t think for a second that AI will not go the same hellish direction.
Of course it will go the same direction - many of the leaders in the AI space are the very same entities either technologically or financially that are responsible for that destruction.
Unironically this is a "return to tradition" opportunity in the making to me.
If this really is the case, then there's an engineering solution here to re-train people by answering questions with search results first, leaving AI reconciliation for those cases where someone would be clicking the "next page" button or scrolling further.
Of course, that may not align with the perverse incentives driving whoever is behind each bot.
Is there a way to read some of the responses that i’m missing? It sounds like a lot of fun to see a what i expect is a Pictionary style human drawn image in response to a prompt. I want to see it.
As soon as I saw "Why are hot dog buns sold in packs of 8 and sausages in packs of 12?" I wanted to know the response! My immediate reaction was "because sausages tend to last longer than buns" but I'm curious to see how Quili.ai answered.
Excluding the Google AI summaries included with every search, most requests are likely going to be coming from enterprise users using it for code.
So, probably no.
Sounds like misinformation, let's compare: "The clothing and textile industry uses roughly 170 times more."
The difference is even larger if you look at the state of the art: "As of November 2026, the state of the art is direct-to-chip liquid cooling combined with closed-loop dry coolers. This allows a data center to operate with essentially zero water consumed for cooling, even for high-density AI infrastructure."
You might not want to put an older/less efficient data center in a very dry area, but other than it seems quite reasonable.
Who the hell would think putting a datacentre in one of the most actively seismic countries on the planet is a good idea? I'm puzzled. Pretty good initiative other than that.
Well, usually its not a big problem, Chile in general has a good earthquake resistant infrastructure, because of its building code which has been built based on past experiences.
Chile is a good place, if you look at the submarine cable map (which allows connections over the internet between regions) you'll see that in the west part of South America, Chile is the only country with a submarine cable that goes to Asia, and is going to be working in 2028 and Chile in general over land also has a very good internet connection and is very advanced technologically.
"BECAUSE OF THAT, COUPLED WITH AN ONGOING CLIMATE CRISIS, CHILE IS FACING A SEVERE DROUGHT EXPECTED TO LAST UNTIL 2040."
I like the initiative as a life-affirming activity, but spreading falsehoods about water usage is not a good thing. Lying and deceiving makes the world a worse place. This sentence could just as easily read:
""BECAUSE OF [Quili], COUPLED WITH AN ONGOING CLIMATE CRISIS, CHILE IS FACING A SEVERE DROUGHT EXPECTED TO LAST UNTIL 2040."
Without explaining what proportion of the drought comes from the climate crisis, or the mechanism by which data centers cause drought, you could substitute almost anything for [Quili] here.
The water hype needs to go. It's caught on with people who get their news off tiktok, but we need to have better epistemic models if we want to actually make the world a better place.
If you have the same rainfall today as you do tomorrow, but tomorrow someone steals all of the water and leaves you with none, is that a drought? Not technically...but who the fuck cares about the technicality?? You don't have any water!
On a related note, many famous famines went from stressful-but-survivable bad harvests to mass casualty events because of mismanagement: the Irish Potato Famine, the Holodomor, the Great Leap Forward...
It seems the villain here could be the Chilean government, for prioritizing the tech economy over the people at a time when it matters most.
I don’t quite get the classification of alien considering they were grown with human data (and humans were the ones that invented the architecture to make it work in the first place)
Very fun idea. I do think this puts the water argument into a bit more perspective. Your average person should drink ~3L of water per day, or about 22mL per 11 minute average-length chat that Quili quotes on their site. It's hard to pin down exactly how much water AI uses, but something on the order of 5mL per interaction seems reasonable [1]. So the people doing this are using about 4x more water than an AI model would. Of course, it does matter what happens to that water, whether it turns into water vapor or urine and how much of that water re-enters the water cycle, but it's an interesting comparison.
[1] https://www.seangoedecke.com/water-impact-of-ai/
The people doing this are drinking water either way.
That's true, but the people could also be doing some other job with their effort too. If you factor in the water consumption of the energy used to power their laptops while answering these questions (~30W of power for 11 minutes per interaction, 1mL of water per W of power) you get another 5mL of water per interaction purely in electricity use.
Not saying it's a 1:1 comparison but it puts it in perspective.
I can't seem to find the laptop power usage bit in the TFA, is that just supposed to be presumed? Isn't that kind of high?
30W is on the low end, based off of web browsing and standard office usage. I didn't link my source so here it is: https://solartechonline.com/blog/how-much-electricity-does-l...
My urologist told me off for drinking 3L of water per day, said I should be aiming for 3 pints, maybe 4 at the most.
That would translate to 11mL of water per interaction. Plus the 5mL for running the laptop that I pointed out in my other comment. I think the comparison still holds up with that math.
You sure they weren't just confused about imperial unit conversions?
This is a neat idea but some of the claims on that website seem extremely propagandistic. I'm not particularly gung ho about datacenters but if datacenters are causing water issues the solution is to shine a light on alternative technologies to solve the problem. Perhaps Chile could become a leader in building cleaner datacenters but thats not the goal I am taking from this initiative.
This is an alternative technology.
This canard about AI and water usage really just isn't going to go away, is it?
> This canard about AI and water usage really just isn't going to go away, is it?
What does canard mean in this context?
I feel like I'm living in an alternate reality, because I have no idea what this page is about. Something with data centers using up all the water? Reading the comments here and people arguing dcs dont use that much water compared to regular people who drink water, confuses me even more. Like whats next? People are told to drink less and fork over their RAM? That would be wild
Whoa. This is amazing. I think What is interesting is whats driving this: the psychologically formed habit of asking questions to bots rather than search engines.
If the median query on the internet is now and forever answered by an AI bot, humans could just as easily become really important as an "authentic" source of answers.
I agree with this, but I can’t help but reflect on how we poisoned our own well by destroying the human internet before AI.
The answers to almost every search are already out there, but we made the public internet so toxic and user hostile, you now need a whole new layer of AI to sift through the bullshit, ads, and clickbait.
I don’t think for a second that AI will not go the same hellish direction.
Of course it will go the same direction - many of the leaders in the AI space are the very same entities either technologically or financially that are responsible for that destruction.
Unironically this is a "return to tradition" opportunity in the making to me.
If this really is the case, then there's an engineering solution here to re-train people by answering questions with search results first, leaving AI reconciliation for those cases where someone would be clicking the "next page" button or scrolling further.
Of course, that may not align with the perverse incentives driving whoever is behind each bot.
This is cool. It's worth noting that there has been a much wider Chilean water crisis: https://en.wikipedia.org/wiki/Chilean_water_crisis + https://www.bnamericas.com/en/interviews/chile-has-no-time-t...
Is there a way to read some of the responses that i’m missing? It sounds like a lot of fun to see a what i expect is a Pictionary style human drawn image in response to a prompt. I want to see it.
As soon as I saw "Why are hot dog buns sold in packs of 8 and sausages in packs of 12?" I wanted to know the response! My immediate reaction was "because sausages tend to last longer than buns" but I'm curious to see how Quili.ai answered.
Really interesting. This makes me wonder if it would cost less to have human answers for like 90% of every-day AI requests.
Excluding the Google AI summaries included with every search, most requests are likely going to be coming from enterprise users using it for code. So, probably no.
> Each chat lasted an average of 11’ 4
What does this mean?
11 minutes 4 seconds
11 foot 4 inches
11 minutes, 4 seconds
11m4s
Sounds like misinformation, let's compare: "The clothing and textile industry uses roughly 170 times more."
The difference is even larger if you look at the state of the art: "As of November 2026, the state of the art is direct-to-chip liquid cooling combined with closed-loop dry coolers. This allows a data center to operate with essentially zero water consumed for cooling, even for high-density AI infrastructure."
You might not want to put an older/less efficient data center in a very dry area, but other than it seems quite reasonable.
"Clothing uses more!" ... yes. Because humans need clothes.
"Closed loop systems solve this problem!" ... most DCs are not using, or planning to use, a closed loop system.
The beginning of the video made me really thirsty!
Isn't there a bigger water crisis in Chile? datacenters are easy to blame probably
Did they track their water usage for that day?
Humans have neural networks, too!
People still don't understand that
a) datacenters don't use water for computation, they use it for cooling
b) the water isn't obliterated, it's still there
c) the water can be recycled and rereleased into the ecosystem, which it mostly is.
Who the hell would think putting a datacentre in one of the most actively seismic countries on the planet is a good idea? I'm puzzled. Pretty good initiative other than that.
Well, usually its not a big problem, Chile in general has a good earthquake resistant infrastructure, because of its building code which has been built based on past experiences.
Chile is a good place, if you look at the submarine cable map (which allows connections over the internet between regions) you'll see that in the west part of South America, Chile is the only country with a submarine cable that goes to Asia, and is going to be working in 2028 and Chile in general over land also has a very good internet connection and is very advanced technologically.
[dead]
"BECAUSE OF THAT, COUPLED WITH AN ONGOING CLIMATE CRISIS, CHILE IS FACING A SEVERE DROUGHT EXPECTED TO LAST UNTIL 2040."
I like the initiative as a life-affirming activity, but spreading falsehoods about water usage is not a good thing. Lying and deceiving makes the world a worse place. This sentence could just as easily read:
""BECAUSE OF [Quili], COUPLED WITH AN ONGOING CLIMATE CRISIS, CHILE IS FACING A SEVERE DROUGHT EXPECTED TO LAST UNTIL 2040."
Without explaining what proportion of the drought comes from the climate crisis, or the mechanism by which data centers cause drought, you could substitute almost anything for [Quili] here.
The water hype needs to go. It's caught on with people who get their news off tiktok, but we need to have better epistemic models if we want to actually make the world a better place.
Why aren't you asking why they are trying to build a datacenter which requires an abundance of water in a place where water is scarce?
Why focus on epistemic models instead of ethical ones?
If you have the same rainfall today as you do tomorrow, but tomorrow someone steals all of the water and leaves you with none, is that a drought? Not technically...but who the fuck cares about the technicality?? You don't have any water!
Yes, in your example, the person who stole the water would have stolen the water.
In real life, in Chile, who "stole all the water" ? How did they "steal" the water?
On a related note, many famous famines went from stressful-but-survivable bad harvests to mass casualty events because of mismanagement: the Irish Potato Famine, the Holodomor, the Great Leap Forward...
It seems the villain here could be the Chilean government, for prioritizing the tech economy over the people at a time when it matters most.
This underscores how vastly underrated humans are by the people who are pushing this obnoxious, inhuman, alien technology onto... humans.
I don’t quite get the classification of alien considering they were grown with human data (and humans were the ones that invented the architecture to make it work in the first place)
I have a question: have you used one of these models for code, CAD, general 3d work? The vast chasm will take on a very different perspective.
Totally. This is way more exciting than any new frontier model. It is kind of crazy how good an idea it is...