#whereas say the problem with large language models is conceptual; it's an inefficient approach to ~intelligence~

alright guys, what we've got from solar power is going to be lots of power that you can't store available at random ass times. So what are we going to see? We're looking for industrial processes that are

hilariously wasteful of energy
equipment is cheap
you can turn them off and on easily

I've only thought of two things here.

Splitting water without good catalysts, just paying the overpotential. Why buy an expensive electrolyzer containing platinum or iridium as a catalyst, when the catalyst is expensive and power is cheap? Especially since you'll only be able to run the electrolyzer a minotrity of the time, when generation is especially high, so it's just going to be sitting idle most of the time. So I'm thinking we're going to start seeing inefficient green hydrogen production
Desalination of water by distillation. Or if not that, "Mechanical-Vapor-Compression", which seems to be the method with the worst energy cost but advertised as having cheap units. Right now I think people usually use reverse osmosis, where you use pumps to push water through a membrane. It's the most energy efficient, but it's capital intensive... I think a lot of the cost is buying and maintaining the membranes? Idk. Whereas you can just boil it. The MVC method seems to use heat for evaporation somehow so maybe it's not that different.

I was trying to think of a data center angle, but chips are expensive, you don't want to turn the computers off at night. Maybe ice-based cooling? Freeze water when power is cheap? Or water-based cooling, since water has a high heat capacity so maybe it will stay cool between spikes in solar generation?

raginrayguns

yeah, nevermind, they're not going to do this. I mean to some extent they are or already do (see for example ice storage air conditioning. But the way to make money off a few hours per day of low electricity price, without too much capital investment, is just to charge a battery. I was still of the mindset that battery prices couldn't keep going down because they were ultimately limited by the availability of lithium. No, there's plenty of lithium, and sodium batteries aren't as speculative as I thought, they're ready to go at scale.

#raginrayguns #Solar Power #Batteries #^^^^^^#My take on this was that the problem with Batteries were Materials problems and Material problems are generally solvable #whereas say the problem with large language models is conceptual; it's an inefficient approach to ~intelligence~#maybe not always! If you can make their processing and memory as cheap as it is in bioforms maybe it might be better!!#of course: they dont want to actl make something self directing and intelligence in bioforms IS self direction so that's ALSO a problem #but anyway #Technology #Materials Chemistry #Engineering #appreciative reblogs #zA Opinions #minitagrants