But AI is going to solve all the important problems. RIGHT?

But when you ask for details, "important" keeps coming back to shareholder value, now measured in new and improved cryptocurrency!

Funny thought of the day: Think of major changes motivated by positive teleological objectives. (I don't want to say "good intentions", but...) Stuff like the abolition of slavery, the NHS, or even the EPA. Ancient history, right? I'd like to hear your more recent examples to prove me wrong, or even better, something along such lines that is currently in any pipeline with any prayer of implementation RSN. (Certainly not aware of anything in America where they are deadlocked between the Rumplicans and Dumbocrats. Orange-nosers versus hunters of flying elephants...)

Unthanks.

If I were a gambling man, I'd give HIGH odds against AC not wasting time. My time and even your time to tl;dr it.

Mod parent funny.

Even funnier is the book Tech Monopoly by Herbert Hovenkamp, especially when it comes to the area of solutions. The book Careless People was also pretty funny, especially in the parts where it tried to portray Zuck as a real human boy.

Root of the problem thoughts? Good intentions and a charged blaster will get more done than good intentions alone, and Microsoft has plenty of well-charged blasters. Now think about major changes that were linked to good intentions. Stuff like the abolition of slavery and the NHS. Ancient history, right? Anything in the pipeline now? Of course Microsoft and f[r]iends are claiming that AI will solve all the important problems, but when you ask for more details it seems the "important" stuff keeps coming back to shareholder value. "You shall not press down upon the brow of labor this crown of [Copilot-enhanced Microsoft Office]. You shall not crucify mankind upon a cross of [shareholder value]."

"I don't drink a spreedsheet every day, but when I do, I drink LibreOffice. I'm the most boring man in the world."

I'd want to see some numbers... I guess that means I'm questioning some of your assumptions. For example, liquid cooling is assuming that the servers run hot, but I think there is a complicated tradeoff there and it's possible to design things differently so that air cooling might be viable. But if you assume there is a race condition and we need to run everything as hot and as fast as possible, then the room for negotiating new solutions becomes much more limited. Perhaps it's more of a philosophic problem? I think the important questions are not urgent? The big problems are going to wait as long as it takes for their solutions?

Actually links to my thought of the day about major intentional changes. All of this AI stuff seems to be driven by opportunity. We can do this AI stuff now so we HAVE to do it, and on the largest possible scale. It's going to fix everything, but when you ask what "everything" means it keeps coming back to the problem of putting more fantasy money in the pockets of a few exceedingly rich folks who already have more money than gawd and no good uses for it, in stark contrast to gawd's advertising. I've been thinking about the big changes that were deliberately created for teleological reasons... Stuff like the abolition of slavery and social security and voting rights for everyone. On the medical side stuff like Medicare in the US and the NHS in England. Seems like that "fashion" has gone out of style now and I can't think of any recent examples of such grand ambitions for change that are being taken seriously... File me with the nutters who care too much about "humanity", whatever that means?

I'll be back as a consultant at triple the rate in less than 18 months to clean up the mess...

Even if you're automating something you yourself do on a regular basis.

This is a general property of the universe: put something under a (metaphorical or literal) microscope and you discover things that are at once obvious and at the same time fantastical.

Making Things is fun and frustrating that way.

Being unaware of where the opponent is and going through the motions of kicking a phantom ball was very much par for the course.

Let's ask Michael Hastings. Oh, wait.

Just joking. Of course it's only a joke, but it's the obvious one I was looking for when the topic came up. I'm sure Obama's deep state never would have done such an unethical thing and you'd have to be a real conspiracy nut to blame someone left over from Dubya's time.

Oh, and I'm also sure that this is one of those problems that will just go away if we wait a bit longer. That's the other low-hanging joke I was looking for.

Probably a PROFIT! joke hanging around somewhere, too.

Not disagreeing, but on this line of approach, it seems better to just put the data centers in the dessert rather than transmit the power. I'm even trying to imagine a design where the solar panels are providing shade to help keep the data centers cool... What kind of air gap would you need to make that work? Kind of a jungle high canopy design?

Another option might be burying the computers, but I think that would lose out because of the cooling problems... But possibly convection powered cooling could be made to work?

in the free market businesses form and compete for consumer money with other businesses, it is the competition that is pro consumer. Unions, like any other political formation, exists to protect special interests, preventing businesses from competing to the best of their ability and pushing prices up - anti consumer.

Cupcake, 45% of government work entails coming up with workarounds to problems created by the other 50%. The remaining 5% (at best) is something you might call useful without having to supress laughter.

This extends to contractors. Microsoft is a big one.

Businesses are pro-consumer by definition, unless they are government subsidized. Being anti-union (as I am) is being as pro-consumer as possible. Unions push prices up, consumers want lower prices, unions are anti-consumer.

I think you're oversimplifying because the supply pattern for the renewable energy isn't that complicated. Depending on various factors, I'm thinking you could get relatively high usage numbers with relatively low expenditures on storage, but part of that optimization would involve idling some of the servers some of the time. However even in that case you would want to tweak the system to keep the best servers doing as much of the work as possible.

Eh? Now the spelling check has decided that videocide is problematic after all? Were my eyes napping before? Eh and double eh?