I had heard that conceptually, they couldn't find a benefit for having the FPGA on-package but it was a big thermal/power challenge. So the integrated case was actually worse.

On the Omnipath side, similar story though they actually shipped a version... It got watered down to PCI-e attached even while on-package and so it was just a pain with zero benefit for being on-package. Very weird board/cooling design needed to let cable go straight to CPU, having to share the TDP budget between the loosely logically coupled, but tightly physically coupled functions...

But yeah, Intel doesn't seem to know what to do except processors, and they've squandered even that by now...

There's a huge difference:

Our governments, at least in theory, are controlled by us, the people. Ok, the 1% who make the major campaign contributions. But that's still a lot of people.

The number of SpaceX or Amazon shareholders who have enough shares to have a say in these matters is single-digits. So power is concentrated in much fewer hands.

Given dynamic battlefield, I don't think that is as easily done as you think, and the moment SpaceX makes a mistake and knocks out a Ukrainian drone on a mission, they'll be guilty for everything. There's not really a winning position for them here.

Russia isn't hiding that it targets civilian infrastructure. They still wage war the way everyone did it in 1939-1945. US and UK bombers essentially just opened the doors above German cities and let the unguided bombs fall wherever. We're not doing that anymore because most of the world learned that despite all this, they didn't exactly surrender. So it's a huge waste of resources. Russia, on the other hand, still thinks that Ukrainians will agree to becoming Russians due to a few cold and dark winters.

Hiding money flow from public view. It is trivial to automatically trace all transfers on the blockchain. And the same way I don't post my banking history to the Internet, I have a reasonable need to not have all of my Bitcoin transactions fully transparent to everyone in the world.

So tl;dr: The legal use is: Protect my privacy.

That doesn't mean I am doing anything illegal. I might be doing something perfectly legal but socially controversial - maybe I make campaign contributions to the communist party, or consume an unhealthy amount of furry porn. It might also be legal but I have a need to hide my finances from someone specific - maybe an abusive spouse, maybe overly controlling parent, maybe a stalker.

For the moment, Bitcoin is still a bit of a niche thing, but the more it moves into mainstream, the more people will have the interest and the capabilities to use Bitcoin to breach people's privacy when they use Bitcoins to pay for something.

I really hope that Waymo's cars aren't relying on their Nader-beepers to avoid killing people. They should be (and AFAIK are) relying instead on their video cameras, LIDARs, and other sensors to stop the car before it hits the wayward kid/technician.

Yes, they should have done it the right way instead!

Err, what was the right way? It's not obvious to me, given that Microsoft doesn't have design control over the hardware its software runs on.

At this point, my biggest fear is that she will -- and then call me over to debug the AI-generated codebase, when it inevitably doesn't work quite right.

Particularly striking as they started from a pretty solid premise, that mismanagement broadly is the cause. Especially citing Boeing, which was *well* documented that the changes can be traced back to acquiring McDonnel Douglas, which was ripe for the taking after being mismanaged into failure and Boeing having the genius idea that the best thing they can do with a leadership team that tanked their former company is to put them in charge of the still viable Boeing. People who wanted to scream DEI pointed to DEI initiatives that started *after* the troubled MAX program was already in the air.

I wager engineers are willing to agree, as they see their work as solid but the business mismanaging things to make good engineering infeasible.

"We (the broader company) doesn't know how to engineer, but *I* still do" I could easily imagine being the takeaway. I think most of us can relate to being part of a broader mismanaged whole.

I'd say the big thing is they took their core product as granted, and focused a great deal of their income on almost anything else, aiming/hoping for some horizontal growth instead of investing to preserve their processor market share. Intel is flush with cash and could either invest in CPUs, or, say, buy McAfee, a brand that had lost most of it's value a decade prior. Or maybe acquire some HPC products to try to build an in-house all-in-one HPC solution to compete with their partners, then decide that was a bad idea and mostly abandon that expensive effort. Or maybe buy an ethernet switch chip company, and then promptly do nothing with it. Since it worked out so swimmingly the first time, do the exact same thing with another ethernet switch chip company and again just shrug and never do anything with it. Maybe spend a boat load of money trying to make "Optane" a thing, including heavy evangelizing to try to convince people to fundamentally rework core concepts of how they work to justify the apparently awkward in-between of PCM which was never going to be as fast as SDRAM nor as cheap as NAND. Along the way spend money on all sorts of weird random projects someone had without any target customer expressing interest in the hopes they stumble upon some unexpected Model T moment in a new market segment.

Intel just assumed their position in the market was unassailable and went about trying to start *something* else because protecting their core business wouldn't deliver adequate growth (they pretty much had the market cornered). So you have a lot of big 'lottery ticket' investments with inconsistent execution on top of dubious justifications in the first place. They failed to coalesce around a common accelerator/GPU strategy leaving them critically disadvantaged compared to AMD and nVidia, their CPUs surpassed by AMD, their fabs long passed by TSMC and none of their gambles paid off, so now they are just boned. I suppose on the upside, *now* they have growth opportunity in their core competency since they ceded so much ground...

Let's assume for the sake of argument that OpenAI and its competitors are trying to do the right thing here and make their AIs as harm-free as possible.

Not everyone will be that responsible, however. Now that it has been demonstrated that a suitably sycophantic AI can compromise the psyches of significant numbers of people, it's only a matter of time before various bad actors start weaponizing their own AI models specifically to take advantage of that ability. "Pig butchering" will be one of the first job categories to be successfully replaced by AI. :/

>The ruling elite has decided they do not want you to be educated

No, the ridiculous price of tuition is actually the result of the government encouraging people to go to college. They will pay whatever 'need' is, and 'need' is whatever the price of college is... so colleges keep raising the price, year after year, much higher than inflation.

It's not going to the professors, either. At my local state college, only 13% of the budget goes to professors (including benefits).

At community colleges here in California, 50% goes to direct instruction. Why? Because state law mandates the 50 percent rule. So community colleges are reasonably priced and often do a better job actually teaching with smaller classes than state colleges, but I guess we have a good 9-3 football team at state, so...

I didn't do anything crazy like actually read the article, but I did go as far as to read the third sentence of the summary, which began like this:

.... and that would seem to indicate that IPv6 is currently handling around half of Internet traffic.

AS I recall, someone tried using chemical SNAP generators with lasers and Neural Networks in the 1980's or 90's.
The system tried to shoot down the moon. Again. Happened before too, with a straight radar system (the Soviets tried? Might have been the US, I forget).

Pretty much. The lady that did the spreadsheet was shocked that it worked much better as a web application. She asked for PHP and SQL training. I helped her get the first basics, then since she wanted to change professions pointed her to professional training. I was concerned I'd show her bad habits, not being expert in it.