We'll be able to increase chip capacity eventually, and we're also still doing pretty well at clean energy conversion. Eventually we'll get there.
I'm much less pessimistic about this.
We'll be able to increase chip capacity eventually, and we're also still doing pretty well at clean energy conversion. Eventually we'll get there.
I'm much less pessimistic about this.
There is LOTS of reporting about this. One example:
https://www.tomshardware.com/tech-industry/big-tech/microsof...
Again, I ask, where are you getting the 60-80% number?
https://www.oracle.com/news/announcement/blog/closed-loop-co...
I wasn't born yesterday dude, it's obvious you're deflecting. You cited a sloptistic generated by an LLM. Take the L.
What?
Like the water situation east of the rockies has been pretty bad... It just ain't the desert that you get over in California or Arizona
Maybe "drought" doesn't mean what I think it means. In this case, I amend my comment to say "There's clearly lots of water there" instead of "drought."
When these days centers say they use (x) MW/GW, that is all turning into heat, with some additional (%) on top for inefficiencies.
And don't forget, they are literally making the largest natural gas power plants. Which in turn generate a fuck ton of heat to make the power because they burn gas to heat water to turn it into steam to turn a turbine... Which in itself is another +30%-60% of heat on top of what ever number they quote for power because of the inefficiencies.
So for every MW, it's really 1.3-1.6MW of heating, then they use that 1 MW to power the cooling systems or the server clusters. So for every 1MW of power use, it's on the order of 2.3-2.6x of heat generated.
Again, the scale is the issue at hand. These data centers are consuming more supplies and power for something that is not necessary to human existence