Data Center Water Usage as a Home Water Saving Tip

Description

A screenshot of a listicle titled 'HOW TO SAVE WATER AT HOME'. The list contains several standard water-saving tips, such as fixing leaky toilets and taking shorter showers. The final item on the list is highlighted in yellow and reads: 'Delete old emails and pictures as data centres require vast amounts of water to cool their systems.' The humor stems from the juxtaposition of common household advice with a technically true but absurdly scaled suggestion for individual action. It satirizes how the environmental impact of technology is sometimes presented, making a point that while data centers do use significant water for cooling, an individual deleting files has a negligible effect, unlike fixing a major leak at home. This is a joke for those in tech who understand the scale of infrastructure and the often-misguided environmental advice given to consumers

Common Tags

Cloud Datacenter DataCenterOperations CloudStorage CloudComputingServices EnergyConsumption

Comments

Anonymous ★ Top Pick Sure, I'll delete my old emails to save the planet. It's the digital equivalent of turning off a light in an office building that's already on fire

Anonymous ★ Top Pick

Sure, I'll delete my old emails to save the planet. It's the digital equivalent of turning off a light in an office building that's already on fire
Anonymous

Next time the VP asks why the log-retention bucket is only 30 days, just tell them you’re protecting the company’s drought-contingency plan
Anonymous

After 20 years of optimizing database queries and implementing edge caching, I finally discovered the ultimate performance improvement: convincing users their 10-year-old email attachments aren't worth the water bill at our Virginia data center
Anonymous

Ah yes, the classic 'delete your emails to save water' advice - because clearly the 200 liters wasted by your leaking toilet pales in comparison to the nanoliters of cooling water attributable to your 47 unread newsletters sitting on a hyperscale provider's deduplicated, multi-tenant storage array. Nothing says 'I understand distributed systems at scale' quite like thinking your personal inbox has a measurable impact on a facility consuming megawatts and processing exabytes. But hey, at least it's easier than calling your data center's PUE into question or questioning why we're still using evaporative cooling in water-scarce regions
Anonymous

Cute idea, but my inbox won’t move WUE - try fewer 3× replicas, lifecycle to Glacier, and fewer pet clusters before asking me to delete selfies
Anonymous

Brown grass regrows healthy; unchecked blob storage just evaporates the local aquifer
Anonymous

If Legal’s seven-year retention keeps that 2013 “ok, thanks” thread, my S3 lifecycle rule isn’t cost optimization anymore - it’s water conservation
@Glebasya777 1y

AI haters usually say something like this
@SergioEremin 1y

Oh, those famous power hungry storages
@SergioEremin 1y

🙀
@qtsmolcat 1y

Technically incorrect, water is rarely ever actually destroyed. But the bigger point I think is that the AI data centers use way more water in a day than you ever will
1. @kitbot256 1y
  
  IMHO the point is that the datacenters often get their water from the same sources as the city. So if the source is scarce lowering the load on the servers will actually let more water to the customers.
  
  dev_meme 1y
  
  meanwhile on my eu servers so dumbass hoster doesn't shut them down due to idling: */05 * * * * stress --cpu 1 -t 60 -v 1>/dev/null 2>&1
2. @maks_mikh 1y
  
  You do not need AI to store photos on HDD
  
  @maks_mikh 1y
  
  To store "old emails and pictures"
3. dev_meme 1y
  
  I didn’t get it You didn’t knew that data centers disintegrate water and put it on another plane of reality so it’s not accessible to the rest of us anymore?
@shacotustra 1y

Oh that thirsty electronic My home pc consumes more water than my whole family
1. @qtsmolcat 1y
  
  Your computer doesn't use massive amounts of liquid cooling :)
@cacojo15 1y

I believe that what they mean is that for at least for certain datacenters, water is evaporated by the cooling system. It is not release into a river sligtly hotter. (I have seen both systems) But that is even more the case about thermal power plants with big cooling towers that will produce electricity for the datacenter. However I don't know if storage of old data has such of an impact.
@Hyron 1y

While it's true that water doesn't get destroyed obviously, so is true for everything else, especially showers/tap water, but still it makes sense to save up that water. Still the water used for data center and such is heated up, since it's used to cool down the infrastructure, and released into the environment (which additionally damages the environment). I'm not sure you can use the hot water released from this process for other uses immediately or after it cooled down, or how long it takes. Also like, even if it's not much imo there's no downside to sensibilize general public to the fact that data centers consumes a lot and be a bit more aware of energy costs
@Hyron 1y

And yeah if it evaporates than it makes even more sense
@maks_mikh 1y

Looks like someone just used AI to write these recommendations
@deerspangle 1y

Honestly I expect someone wrote this while rather pissed off about corporate consumption being so much higher than individual consumption, but given no power to actually point out how silly it is to ask individuals to solve it And hence this one is tongue-in-cheek But I dunno!
@Kumapawa 1y

The thing is that fresh water is used for cooling data centers. When it evaporates, it's likely to turn into salt water due to the specifics of the hydrologic cycle. So, technically the amount of all water on the planet stays roughly the same, but we reduce the freshwater reserves by converting it into salt water. Running out of fresh water sucks because it takes a humongous amount of energy to turn sea water into drinkable water. Imagine Africa's water scarcity but worldwide.
1. @Algoinde 1y
  
  Why would it evaporate Isn't it just running in closed loops to carry thermal energy away from the servers
  
  @azizhakberdiev 1y
  
  water doesn't disappear, just gets polluted over time and also cooling aint free
  
  @Kumapawa 1y
  
  ~80% of data centers evaporate water: https://www.eesi.org/articles/view/data-centers-and-water-consumption
  
  @Algoinde 1y
  
  what the fuck
  
  @callofvoid0 1y
  
  part of the water usage is accounted for the power plant that produces electricity for the dc
  
  @Kumapawa 1y
  
  Yup, it is. It does not change the point, though. Demand creates supply: power plants generate more electricity to meet data centers' demands. This also pushes for new power plants to be built: https://www.bdcnetwork.com/home/news/55298686/new-trend-gas-power-plants-being-built-exclusively-for-data-centers
  
  @SheepGod 1y
  
  Didn't they already have no water there
  
  @Kumapawa 1y
  
  I believe it's more like, "the state is already cooked, let's suck it dry"
  
  Deleted Account 1y
  
  wth?
  
  Deleted Account 1y
  
  sure they want to keep colding data centers with fire lol😂
  
  @Kumapawa 1y
  
  Jokes aside, it makes perfect sense to build new gas power plants in Texas because it reduces the cost of logistics. Around 25% of American gas is extracted in Texas
  
  Deleted Account 1y
  
  Hmmmm... nice.
  
  Deleted Account 1y
  
  ofcorse it can be colding with gas instead of water, that's better than wasting water. Now i got it🤯
  
  @Kumapawa 1y
  
  Nobody cares about wasting water. Cutting costs is the only priority.
  
  @Algoinde 1y
  
  @devs_chat check this, it's more complex than I thought
  
  dev_meme 1y
  
  This really doesn’t matter Check water consumption of any factory that works with steel There’s factories that are build on rivers (not even biggest ones) and used water literally forms a new river (that eventually goes back to original) I was bathing in one such river when I was a child btw, water there was much warmer even during cool day 🌚
  
  dev_meme 1y
  
  Wtf, I love Internet I just found photo of "the beach with hot water" + city name And it’s exactly like in my childhood memories
  
  @RiedleroD 1y
  
  the difference is that steel is useful
  
  dev_meme 1y
  
  Saying that AI not useful at current stage is just weird 🌚
  
  @RiedleroD 1y
  
  I haven't found any use for it yet - generative art is just automating the creatively fun parts of being human - trying to find facts with LLMs is more work than just doing it with conventional tools (unless you're ok with potentially being lied to) - doing any technical work with LLMs creates more work in the long run, as codebases become unmaintainable with shit code and zero people that are knowledgeable about the codebase
  
  @RiedleroD 1y
  
  ok, there is some utility in the medical field, allegedly, but only in a "maybe we should make extra sure nothing is wrong, because the AI said something is wrong" kinda way. I'm not in the field, so I don't have insight into what's going on, but if it were some kind of revolution in medicine, I'd have heard about that. last major thing I heard is an anticipated revolution in medicine - around 2 years ago.
  
  @Algoinde 1y
  
  And the actual "AI" (ML) usage in medicine existed for several years before the LLM hype cycle, for *actual* inference, tumor detection and so on
  
  @RiedleroD 1y
  
  yea, that's what I'm talking about. again, can't talk specifics because I'm just not in that field
  
  @lambda_coolusername 1y
  
  https://youtu.be/TNeVw1FZrSQ bad news about the state of medical llms 🙃
  
  @RiedleroD 1y
  
  this says relatively little about that kind of AI
  
  dev_meme 1y
  
  p.2 - no idea when I have seen last time wrong factual info from any frontier model p.3 - unironically, skill issue Like any tool, you need to get used to it, and understand pros and cons. What and when works, and when it doesn’t But I got so tired last 6 months convincing people to actually try to think before using that instrument, instead of expecting actual sentient AGI to be provided to them for free… I gave up at this point Happy to help those who curious, esp at work but nothing more
  
  dev_meme 1y
  
  My main point is: it’s not that cheap to get clean water Desalination of water is nothing really expensive too. Its just governments (all of them) that want to steel much more money than actually goes to infrastructure. It’s crystal clear when you see big corps doing the same thing in terms of infrastructure for 1/10 of price or even less There’s no problems with water Only with bandits calling themselves government
アレックス 1y

This but the guy on the left is Sam Altman trying to use a modest country’s worth of water to train GPT-Safer and the guy on the right is selfish grandmas saving a whole two gigabytes of family photos.
@Hermesiss 1y

It just goes to the cloud
dev_meme 1y

At the moment my main idea is that higher in management hierarchy you are / more programmers you had to manage, and more PRs with different styles you had to review - makes it so much easier to treat llm just another colleague I can feel correlation that those who not only write code but also manage programmers are much easier to adopt LLMs into their processes
1. @Algoinde 1y
  
  I'll never treat AI slop as "another colleague" because when reviewing PRs i do it to educate the person so they can start writing better code When reviewing slop, I'm trying to find hallucinations that I can't do anything about, and therefore it's *not* a colleague
dev_meme 1y

That’s what I’m tired of repeating! Stop thinking that it’s sentient AGI, it’s just a tool!
1. dev_meme 1y
  
  And juniors can, and do produce code that works after tweaks and a few reviews But instead of waiting a day, you can start review after 3-5 minutes
  
  dev_meme 1y
  
  And this Junior worked before with all technologies you work now, worked before, and may work for this project
dev_meme 1y

This is obvious bullshit and why would you consider this to be normal/default?
1. @pdsnrc 1y
  
  just personal experience
2. @qtsmolcat 1y
  
  Your average senior developer isn't going to be using ChatGPT, and your average developer putting code into ChatGPT probably doesn't know enough to catch obvious errors. It's a catch-22
  
  dev_meme 1y
  
  Don’t use ChatGPT for coding Guys, really, mentioning anything, but ChatGPT - you didn’t even tried to get into actual vibe coding Nobody uses ChatGPT for vibe coding, like, there’s no even such options You have relatively poor options with tools like cursor and windsurf, Or you go for proper experience that starts from hundreds of dollars: codex (for gpt models) - and you will need 5-pro or 3-pro for proper code writing Gemini-cli Claude code
  
  @RiedleroD 1y
  
  a coworker of mine does sometimes
  
  @deadgnom32 1y
  
  I do too, but I never take the solution as is, it's more for inspiration
  
  @RiedleroD 1y
  
  similar in his case, but I feel he'd be a lot more efficient if he just made the piece of code instead of reading whatever chatGPT spits out and pondering whether that's a good approach
  
  @deadgnom32 1y
  
  I use it more in cases, when I know a certain technical approach exists, but I don't know any libs and tools for it. I explain the math behind a solution I want to achieve — it gives me some code. and this is a starting point, at least I know what to read and have some code sample
  
  dev_meme 1y
  
  I’m just asking to not use it, because gpt models are bad for coding/engineering Of course too many people still do Gemini-2.5-pro and both sonet and opus are way ahead, as long as provide them with proper context in the prompt
  
  dev_meme 1y
  
  Most of senior devs I work with use AI tools, including ChatGPT Documentation, integration specs, debugging - things do actually move faster and it’s great alternative to Google, and it often works better than Google Even with cheap and outdated models like 4o
  
  @qtsmolcat 1y
  
  Until it makes up shit and you don't double check it
  
  dev_meme 1y
  
  What happens when non-fully checked codes is merged into main and reaches production? You deploy prev version, and revert last merge, no? How is it any different? Gradual rollout and canary releases were a thing well before LLMs became a good tool

Use J and K for navigation

Data Center Water Usage as a Home Water Saving Tip

Description

Categories

Common Tags

Comments

More like this

AWS: You Were Supposed to End the Single Point of Failure

If You Die in US-EAST-1 You Die in Real Life

You're So Poor That When AWS Goes Down You Can Still Enter Your House