The water isn’t really consumed. It’s not like the servers are digesting it or boiling it away or something. Doesn’t the water just cycle through the system and then come out warm on the other side? There’s no reason for that water to go to waste!
Seems like these data centers are just dumping perfectly potable drinking water because it’s slightly cheaper than having their own water supply and cooling ponds.
That’s a good point, there’s no reason this couldn’t be designed as a closed cycle system.
it generally gets evaporated because evaporative cooling is efficient and water is cheap.
There’d be billowing clouds of steam coming from data centers if that were the case.
Pretty sure the water just goes through coolant piping and comes out the other side warmer, because water is a good heat sink.
Removed by mod
I’ll be honest, chatGPT has helped a fair bit in my limited coding knowledge when needing to write VBS scripts for Excel or Batch Files for initial system config/app installs at work.
Removed by mod
Fure sure, calling it “AI” in its current form is just marketing
This is a terrible metric. If kWh is too abstract for people, maybe put it in terms of how much water that energy could boil, starting from room temperature at sea level.
For sure, it’s a lot more relatable in these terms.
So many resources spent on services that are objectively not great. I have yet to see any output from Gemini or ChatGPT that sells me on the service.
not every service is for everyone. LLMs are useful for creating filler, answering common, but hard to phrase questions, and creating large amounts of text that is readable but meaningless for filling assets in a video game.
