Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

'Thirsty' ChatGPT uses four times more water than previously thought

thetimes.com

Post reply on HN
Print view View original

71–80 of 105 posts

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#71

Post by yallpendantools »

fnordpiglet wrote: post #48
MrMcCall wrote: post #8

Earlier quoted context omitted.

ChatGPT is not housing anyone, nor is it providing office space. Even Disneyland is at least providing joy for children and adults.

You know what’s also not housing someone? Office space. This feels like the ultimate waste - massive empty buildings in prime central locations with plentiful public transit designed to house humans during the day in addition to their own houses - which are being pushed further and further away from the central locations because of the office space - that they’re forced to commute long distances against their will to…

I think the question is how efficient is ChatGPT in providing economic value. I don't want to get in a game of anecdotes because I don't want to deny that there are use cases well-suited for LLMs and you might be in such a field. The efficiency question is thus: of all the queries LLMs serve, how many of those are hallucinations (which then need to be reprompted), how many of those are simple queries that could've been solved with more traditional search methods, and how many of those are queries which only LLMs could've solved?

I don't really have an answer to my own efficiency question and will likely never will. But for every company like yours, there are companies who blindly introduce LLMs into their product for no other reason than hype. Spotify, your recommendations ("DJ") are no better because you used LLMs; collaborative filtering worked just fine for music. Meta/FB, your value prop to me as a social network hits rock bottom if my friends are not even sharing real-life experiences. Google, you're already on thin ice with all the SEO spam but your value decreases even more if you give me hallucinations for summaries. Hence, sorry but my outlook isn't rosy.

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#72

Post by MrMcCall »

go_elmo wrote: post #12
MrMcCall wrote: post #8

Earlier quoted context omitted.

ChatGPT is not housing anyone, nor is it providing office space. Even Disneyland is at least providing joy for children and adults.

I agree - it has no direct use to the wellbeing of humans - Id say if the indirect use (e.g. humans saving time [which itself is probably the most energy intensive thing, food, housing, entertainment], gaining knowledge has to surpass the costs in Order for the investment to make sense.

No one is 'gaining knowledge'. LLMs are just black-box tools that compute pseudo answers based upon a somewhat arbitrary set of training data. It's a crapshoot that looks good like ELIZA looks good.

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#73

Post by modeless »

atoav wrote: post #40
modeless wrote: post #27

Earlier quoted context omitted.

Water use is not an externality. All water used is paid for.

[citation needed] Nuclear Power plants who use river water in their cooling cycle and pump the (now heated) water back to the river don't pay for that.

That's not the kind of water use the article is complaining about.

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#74

Post by lambdaba »

MrMcCall wrote: post #67
lambdaba wrote: post #60

Earlier quoted context omitted.

Inference consumes far more total compute than training the model.

By 'inference', do you mean using the model to compute a specific series of results?

Yes, training is a tiny amount of usage, this is well-known.

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#75

Post by ttyprintk »

MrMcCall wrote: post #33
volent wrote: post #4

> Many data centres use water-based systems to cool the plant with towers evaporating the heat, like a huge perspiration system, which means that the water is lost. Someone doesn't know what "evaporation" or "lost" means.

Someone doesn't believe in global heating.

You can simultaneously believe that the energy demand for cooling is unethical and that evaporation is the best source of cooling.

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#76

Post by MrMcCall »

lambdaba wrote: post #74
MrMcCall wrote: post #67

Earlier quoted context omitted.

By 'inference', do you mean using the model to compute a specific series of results?

Yes, training is a tiny amount of usage, this is well-known.

Well, if that's the case, then the situation is worse than I imagined. I thought that training the model would be far more expensive computationally than using it to compute an 'answer'.

Thanks.

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#77

Post by jiggawatts »

yallpendantools wrote: post #37
jiggawatts wrote: post #2

The water doesn’t disappear. Most data centres use water in their chiller units, as do many large industrial facilities. ChatGPT is a minority of the energy, heat, and cooling budget of Azure, which is a minority of the world’s data centre footprint. ChatGPT is useful. This article is not.

TFA: > Many data centres use water-based systems to cool the plant with towers evaporating the heat, like a huge perspiration system, which means that the water is lost. > All of these companies have schemes to put water back into nature using projects that help river flow, capture rainwater, recharge aquifers and modify dams. They have all pledged to become “water positive” by 2030: returning more than they consume.…

Please don’t be that person quoting the rules when you don’t understand the arguments.

This hit piece from someone whose job is directly endangered by LLMs is in the same absurd category as the similar articles about how AirPods contribute to e-waste. just look at them! They’re tiny! Or the breathless editorials blaming Elon for destroying the environment with the damage done to the launchpad by his rocket.. a one time incident that threw a few chunks of concrete on a beach and killed not a single animal.

In a way the more absurd and outlandish the claim the more it is circulated by people that want to stop something they’re scared of, usually in the sense of “this is bad for my career”.

“Windmills kill birds, didn’t you know?!” — someone working for the petrochemical industry.

Write down two numbers before replying: the gigalitres of water used industrially in a similar manner — don’t forget to include literally every skyscraper’s HVAC system that uses evaporative cooling — and then the liters used by ChatGPT.

Divide the two numbers.

Then, and only then, try to make a coherent argument about why ChatGPT is specifically a worse water-usage problem than, say, the rest of the Azure data center where it’s hosted that the NY Times hack who’s about to be replaced by a bot strangely had no problem with.

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#78

Post by ttyprintk »

shakow wrote: post #44
p2detar wrote: post #13

From the article: > Using the chatbot for between ten to 50 queries consumes about two litres of water, according to experts from the University of California, Riverside. From ChatGPT: > A single BTC transaction consumes approximately 1,000–1,500 kWh of electricity, with cooling processes for fossil fuel and nuclear power plants using about 2–3 liters of water per kWh. This translates to an indirect water consumption…

How is it a “consumption” (in either case) if the water is basically just running through a pipe and released a couple miles down the line/evaporated into the atmosphere without being polluted?

It’s irresponsibly common for journalists to use the term for once-through cooling.

But evaporative cooling consumes water just like the human body.

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#79

Post by MF-DOOM »

MrMcCall wrote: post #76
lambdaba wrote: post #74

Earlier quoted context omitted.

Yes, training is a tiny amount of usage, this is well-known.

Well, if that's the case, then the situation is worse than I imagined. I thought that training the model would be far more expensive computationally than using it to compute an 'answer'. Thanks.

Training is much more computationally intensive than computing a _single_ answer. But not compared to running the model over time at scale for millions of users.

Re: 'Thirsty' ChatGPT uses four times more water than previously thought

#80

Post by MrMcCall »

MF-DOOM wrote: post #79
MrMcCall wrote: post #76

Earlier quoted context omitted.

Well, if that's the case, then the situation is worse than I imagined. I thought that training the model would be far more expensive computationally than using it to compute an 'answer'. Thanks.

Training is much more computationally intensive than computing a _single_ answer. But not compared to running the model over time at scale for millions of users.

Thanks. That is precisely how I intuited the situation. 40+ years of programming FTW!
Post reply on HN