Well, this is going to depend a whole lot on how you define workstation.

There have been multiprocessor PCs going back at least to 486s. SunOS 4 for x86 (which has been a thing for a lot of years) and SCO Unix would both run on a machine with 8x 486DX processors. That was pretty much intended as a server so far as I know, though.

A Sun SS10 (1992) has 2x SBus slots, each of which can be loaded with a 2x hyperSPARC module, available at up to 200 MHz. We had a SS10 and SS20 at Silicon Engineering called seismic and something else starting with sei (we had seine and seidel and seizure and so on, back then on pretty much any UNIX system you could grep the words file) and the SS10 had dual-dual 85 MHz modules, the SS20 dual-dual 125 MHz. Those speeds might be estimates. These machines were both workstations and servers; we used DQS to distribute Verilog and similar jobs to machines with lots of CPU. But people sat at those machines at the same time.

There were also definitely dual Pentium Pro boards from pretty early on, but that was too rich for my blood. I thought about doing a dual celery when that came around, but then Athlon came out and solved my need for more CPU without more Intel.

The same people can be right and wrong, because there's a lot of different aspects to most situations.

Have you ever punctured a lipo cell? I have. Nothing happened. Then I put it in water. There were a few little bubbles. Over the next two years the pack gradually grew until it was about three times its prior size, and hard. At no time did it emit flames. (I kept it in a coffee can.)

I think NCM batteries in particular are fucking terrible and I don't want to downplay that there is a risk of thermal runaway for all lithium cells with liquid (etc) electrolyte, but overstating the case is not a help.

When was their enterprise product solid? When it was Xenix?

I agree we should generally be going EV (I can't, though) but it's convenient to have a gas station in your neighborhood because you might be headed away from wherever else it might be located.

For EVs filling up is more annoying (as it takes longer) so that raises the desire to do it closer to home. And indeed, people do tend to do it there. I don't have anywhere else convenient to do it, and it's not convenient at home, which is why I can't reasonably have one.

He wanted to sell them to business as well. But then Motorola started to choke while Intel and AMD were executing, and they had to port to PC. Then there was no justification for a big price tag.

The BeBox was really a marketing stunt more than anything else. It was built around a Motorola PPC dev board. (sidebad: The "Geekport" a breakout box connector on the original hardware, and was included only because there were other needed ports on the same board that port is on.) With its dual PowerPC 603e processors at either 66 or 133 MHz, it wasn't exactly slow, but it wasn't as fast as any but the slowest Alphastations. What roped people in was the case design. Otherwise it was obvious that you'd be better off running the OS on a PC soon, because they kept getting faster and cheaper and a good one was already faster than a BeBox.

I had a 66MHz BeBox and also ran BeOS on a Pentium Pro 180. The experience was comparable on both machines, with no real leader. All of the same demos that were so impressive on the BeBox were just as smooth on the PC. But no, the BeBox was not way ahead of the NeXT Cube; it was way after it, as in, five years after. That's a long time in computing now, but it was an even longer time (so to speak) back then. The PowerPC didn't exist when they built the cube; the best processor ever in a NeXT machine was a 68040 @ 25 MHz. (another sidebar: That was an extremely respectable processor for its time, but it also represented the last time Motorola would come up with a competitive chip without help from IBM. 68060 had competitive performance, but not competitive cost.)

It seems like a few NeXT machines were in fact sold into higher education. I knew one guy who had a turbo slab as a CS grad student. He really loved Objective C.

Absent his RDF, yes, he would have been. But he was an effective marketing tool. He was also intelligent enough to see that the Newton was overwrought as a portable device and demand something simpler. The market was moving in that direction anyway, and he charged out in front of it successfully.

Yes in fact they are, or at least, any persistent particulates are dangerous. That's what makes automobiles so bad, and why DPFs actually make diesels worse. We've discussed here on Slashdot before that gassers actually make just as much soot as diesels, it's just far finer so it's much harder to detect, which is why this fact went unknown for decades. The reason it's hard to detect is that the particulate sizes are very small. When they get very small (PM2.5 and below in particular) cilia have a hard time removing them from the lungs and they tend to persist. The soot particles are very stable since they are made out of carbon. All persistent irritants are potential carcinogens.

Some of the additives in tires are very carcinogenic.

It's both infeasible and impractical. Even the carbon black and silicates in tires can cause cancer for the reasons explained in the first paragraph. At best you can mitigate risk, you'll never make them non-carcinogenic. It would be better to also reduce the number of vehicles and also make tire compounds harder to reduce wear. This does reduce safe effective speed around corners and such, but most vehicles have a lot of excess in that department these days, and the ones that don't usually aren't going very fast. I like hard cornering, it's where the fun is in my opinion, but I do consider it to be more important to improve health.

Switching facilities are expensive, but you can design grids such that they are able to break into smaller grids, and that does get continually cheaper. Restarting and synchronizing grids can be difficult, but the more battery storage you've got, the easier that gets. So what we'll more likely wind up with is a grid with more compartmentalization, with a lot of people left in weird and unreliable sectors of the network with unreliable power because nobody will force the providers to actually provide them with power reliably.

There's a lot of space available because it doesn't need it all to itself. There's a lot of desert nobody is using, a lot of canals and reservoirs we could cover, a lot of nice safe flat commercial roofs, a lot of commercial glass, a lot of car parks. So since that's not a real problem, can we just legislate it and move on?

Because so many lies are being told about it, yes.

Show me your A*

The "on rails" complaint is indeed goofy any time the game is not open world, and all games don't need to be.

OTOH it used to be common for games to have both single and multi player, and that was good.

Seems like if you were sure of how that came out, you'd have gone and looked it up and had a stronger comment.

Supposedly it's as many as "breast, pancreatic, and prostate cancer combined." https://www.cancer.org/content...

Because 1) handling exceptions is hard and 2) keeping everything updated from the myriad of sources is hard too.

Whether the benefits are worth the risks is surely complicated, but if it means updates get done that otherwise wouldn't (and it does) then there's at least a reasonable chance that it's a net positive overall.

I'm down with practically anything, or at least just about anything practical, over NCM. It just has issues on every level from political to technical.

I've got my fingers crossed for sodium, but sure, nickel-iron would seem to make sense for stationary applications. I'm hoping though that they can come up with a sodium battery that wouldn't be toxic even if you blew it up.