Thanks for those excellent points. It is almost certainly as you say, that right now, manufacters simply can't keep up with demand for HBM from AI datacenters. And that means they have diverted their production away from consumer-level DRAM.

But that also is figured in their profit calculations. They decide what proportions of their manufacturing effort go to HDM and DRAM, in order to maximize profits. And right now, the big bucks are in HDR, so they take priority.

Relief is unlikely until 2028, when new fab plants are expected to come online. Or demand for HBM levels off. Or consumers are willing to accept higher prices for DRAM. I expect we'll see the third outcome for now.

If your job is filling out forms or collating information to produce reports, if it's taking notes, if it's taking inventory, if it's managing schedules, if it's producing documentation...

All those jobs are going to fall to IT. Not entirely, but it'll be human oversight and an AI replacing a team of white collar workers.

At the same time, it'll be embodied in robots and unskilled manual labor jobs will evaporate (this is already happening).

Good luck adjusting when the disruption is broad, deep, and rapid throughout the economy and workers can't retrain as quickly as jobs are eliminated. This isn't the automobile, this is "cheap obedient slaves with almost no support cost for those who can afford the upfront price tag".

Maybe not on a Tesla with cheap sensors, but I think FSD is inevitable, even desirable. Humans can be remarkably bad at driving, even while awake and sober. Some FSD systems already demonstrate better attention to safety than humans, and I expect they will continue to improve. I welcome our self-driving overlords.

Oops, sorry. That was GrahamJ, not SniffTheGlove.

You ignored the "or how much" part of SniffTheGlove's post. And there's the rub.

Businesses control supply of their products in order to maximize profits. They do the math, taking all factors of the market into account, in order to set the "how much" part. Right now, people want memory. Desperately. And businesses have set price -- and supply -- accordingly. Controlling supply just the right amount lets them keep prices at a level that maximizes their profits.

You'll see prices come down when demand goes down, or some competing technologies arrive.

Boosters are risking their freedom and even their lives. If it was easier for them to find work then they'd do legitimate work instead of boosting. Selling at flea markets is a job itself, so they're clearly willing to work.

Because the website helpfully provides structured data with tags which make it clear what it is, it's very easy to just use old fashioned and cheap filters to exclude content like that from the training set without losing anything else.

Protesting something that doesn't happen is better PR for you than protesting something that does...

It's a trade off: you get abundant free energy to run the server, with extreme constraints on cooling because your server is running in the most perfect Thermos bottle ever.

Others are taking the opposite tack: undersea data centers for abundant free cooling at the expense of having to get the power down to your servers.

If had to bet on which one is more practial, I'd go with undersea servers. Build them off the coast of Chile, run cables out from batery-backed solar plants in the Atacama desert.

If you've got 1 bad cop, and 9 "good" cops who are covering for him, what you've actually got there is 10 bad cops.

You have it exactly right. I can see why you didn't post with an identity, you'd get punished by the reich wingers. Wage theft exceeds all other theft combined but maggots are still crying about shoplifters

Thanks! That was a twist to the story I really didn't anticipate.

Most satellites aren't. But a satellite that is a data center certainly would be.

Computers don't suddenly get more efficient or produce less heat just because you put them in orbit.

Oops:

That's better.

I'm not sure what you asked Gemini. That would help us understand the answer it gave you.

With some simplifying assumptions, all that matters is the solid-angular subtense of the mirror or the moon, and the albedo (the percentage of light reflected from the object.) The closeness of an object affects its solid-angular subtense.

The light flux from the sun (photons per metre squared) is essentially the same at low earth-orbit and the full moon, because the full moon is not much further away from the sun than the earth is.

For an object twice as far away from us to have the same solid-angular subtense as the mirror, it would need to have twice the linear dimensions or four times the area. So it would reflect four times as much light. But because it's twice as far away, the inverse-square law would reduce its apparent brightness by a factor of four also. So, two objects with the same solid-angular subtense and albedo would be equally bright, if illuminated from the same source.

I made some simplifying assumptions, but overall I think my calculation is sound.

Now, if the mirrors are paraboloidal or ellipsoidal, then perhaps they could direct their light more efficiently than a planar (mirror) or convex spherical (moon) object could. That's another discussion.