Is “AI” (ie, large language modeling, also known as enhanced word prediction; and with no logical reasoning ability) really so important that this infrastructure needs to be built?
For the love of the gods, let this bubble burst already!
Is “AI” (ie, large language modeling, also known as enhanced word prediction; and with no logical reasoning ability) really so important that this infrastructure needs to be built?
For the love of the gods, let this bubble burst already!
Let them build green energy before it bursts... although, as another user pointed out, this may be the usual money grab and nothing gets built in the end.
Conveniently, the heat from all this power being generated and subsequently used in the data centres doesn't count as emmissions. Twats.
Fairly so - it isn't emissions, and does not contribute to the problem in a meaningful way.
The reason why emissions are dangerous is because they trap solar heat at large enough scales to change the global climate. Server farm heating isn't really anywhere near contributing at that scale.
Honestly, I know this is a polarizing issue, but nuclear is clean and pretty much safe and you don't need batteries for it. Lithium batteries of course being an ecological nightmare. Bring it on I say.
Bring it on I say.
As long as regulation stays in place. Or, better, add even harder regulation (for from security standpoints as well as fiscal) to ensure these fuckers are forced to be actively responsible for the safety and give them no way to back off and abandon a plant.
Let them donate excess power to the grid as well. Eh, fund housing nearby for the homeless.
Mining for nuclear is an ecological disaster, and is often done in poor countries under awful conditions, especially lung cancer due to the radon emissions of uranium.
Arguably it is better than mining for coal, lithium, etc. since those have similar issues, but one gram of uranium contains energy similar to 3 tons of coal.
Try Thorium.
Mostly:
So it has the feel of a plan to promise to spend a lot of money several years from now, and get a lot of PR points today, and quietly cancel the project later.
It also takes a long time to deliver
Not that much. Do remember there's a lot of oil money pouring into FUDing about nuclear.
They're talking about 5+ years on the new nuclear in these. And they haven't done it before, so a 30% deadline slip is realistic.
You can put up a lot of wind and solar in that time.
You can put up a lot of wind and solar in that time.
Which needs a stable baseline to counteract lack of supply and/or a lot of lithium. And space.
The existing large-scale batteries are largely lithium. There are a bunch of iron-chemistry ones and sodium-ion ones which have been deployed over the past year, with factories going up to scale them up. I'm not expecting to be limited by lithium availability for stationary batteries.
Well that is, indeed, wack. I appreciate your perspective, I can't believe I missed the "corporations lying for money" angle. I'm usually on top of it.
In 10 years they'll be swimming in waste with no permanent storage facilities in existence, a little will leak due to cist cutting, and they'll let those shell companies go bankrupt to avoid ever having to deal with it.
In 10 years they’ll be swimming in waste
Stop FUDing.
Investing billions
Weren't the headlines a week or two ago about Microsoft trying to get taxpayer funded aid for reopening 3-mile Island? Companies shouldn't be asking for taxpayer funded handouts when they are basically printing money at this point.
This is a most excellent place for technology news and articles.