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.

That's good advice for individuals building a home system for personal use. It's not really applicable for businesses and companies, though, since they likely don't have the expertise or the man-hours required to cobble together their business-critical systems from used parts. They're going to want to buy new, from a company that give them good support if/when anything goes wrong.

It'll get used, if using it is more profitable than letting it go dark. Given the infrastructure costs of keeping all that hardware running, it's not a given that it will be. Once the investor $$$ stop flowing and the debt limits are hit, we'll see how much of the AI hardware build-out can really pay for its own room and board, and how much was just 'peacock feathers' whose only real purpose was to impress gullible investors into handing over their money.

The weird thing about that scenario is, they made it into a kid's movie.

You're not thinking it through -- the goal isn't just to put everyone on benefits and make them spend the rest of their lives clicking the TV remote and waiting for their next welfare check. If you want to do it right (and the robots provide sufficient surplus resources to support it), you go a step further and hire people to do the job they always wanted to do, whether it makes a profit for anyone or not. If that means we have 100,000 ski instructors and 300,000 mediocre artists, then so be it; the robots do the grunt work, and the people are paid to do their preferred avocation.

Not that I expect that to actually happen, of course; in the event the robots actually can replace all labor, the upper classes will make sure that economic surplus goes to themselves, with only the absolute minimum getting distributed to anyone else.

Honestly, they don't "need" robots or anything else; they could just keep doing what they've always done and hope for the best.

However, unlike some countries I could mention, the Chinese government has a vision of what it wants its future to be like, and is willing to work and invest to realize that vision. Hence robots, and other economic development.

There has been plenty of progress in using AI to control robotics; they use robotics-specific AIs for that, of course.

The fact that ChatGPT (or even LLMs in general) isn't particularly useful for robots shouldn't be a surprise, since robots (other than maybe C3PO) are about physical manipulation of objects, not about language generation.

the hunger by the 1% to remove as much humanity from the workplace is sickening.

they fully know they are destroying the middle and lower classes (even more than they already have).

they, like the R party, just dont care. they think they are rich and insulated enough. they never cared what their own people need. the 'let them eat cake' time has come back again, but even worse.

there will be no thought to social systems needed to support the unemployed (which will be many of us, given enough time).

I'm glad I'm retiring soon. I would not want to compete in a job market that bosses think can be done by computer, alone.

and I would not want to be the 'prompt meister' to try to coax answers from the machines that make sense.

some see a great future with AI. I see nothing but doom and gloom. the greed factor is strong in humans and the class disparity will cause rioting and civil wars.

maybe not wars. the US has created a special police force that is above the law, so any uprisings will EASILY be dealt with. they thought about that. ICE is not just for foreigners. its a general purpose police force answerable only to 1 person.

people, please show me I'm wrong. but all signs point to a very bad future for 95% of the 'thinks for a living' workforce.

Dig a bit deeper and you can save money by skipping the nuclear-reactor part; just heat the water for your steam turbines with the geothermal heat that's already present down there.

Compared to what was available before, it is quite impressive.

The negative feedback is prompted by the fact that AI is constantly being shoved into every one of our orifices 24/7 by every vaguely tech-related company as if it was the second coming of Jesus. To justify that amount of social pressure, it would indeed have to be quite a bit better than it actually is, and that's why people aren't impressed.

You mean, why does Linus keep agreeing to be interviewed, and then reply to straightforward questions with the obvious answers?

What would you rather he do? Refuse to be interviewed, or maybe make up unexpected answers just to be edgy?

... which, to be clear, is a good thing. Working around the compiler is dangerous and a code smell, so it shouldn't be something that is easy to do. It usually indicates that either the compiler's capabilities aren't sufficient to meet your needs (in which case, a better solution would be either a better compiler, or to re-evaluate the wisdom of your approach), or that you are doing something the wrong way and should find a way to do it that works with the compiler, rather than around it, so that you get the benefits of the compiler's co-operation.

The economy does not rely on ever-increasing amounts of debt to function. With a reduction or reversal of debt growth, the economy would be somewhat different, not damaged. The economy runs on production and consumption. Debt is a side issue.

I'm going to make some rough approximations here.

There are difficulties in dissipating power in high speed processors. Assume that the power that can be dissipated is proportional to the area of the chip. Relative to a single active layer chip, the power that can be dissipated per layer is 1/(number_of_layers * thermal_conduction_to_coolant). Thermal conduction to coolant is dominated by copper in the heatsink and SiO2 in the chip. Copper is at least 200 times more thermally conductive than SiO2. Assume that the maximum acceptable temperature rise is 50 Kelvin across a 1 cubic centimeter copper cube; that corresponds to 200 Watts. Assume that diminishing returns occurs when the thermal drop across SiO2 equals the drop across the copper. Since they add, if we keep the limit at 50 K the limiting power is 100 Watts. The implied thickness of SiO2 is (1 cm)/200 = 50 microns. How many layers can be squeezed into 50 microns?

A brief internet search seems to yield a minimum layer thickness of 100 nm (0.1 micron) for gate logic -- (1 active layer plus many interconnect layers.) Thus 500 active layers can be squeezed into 50 microns. What happens then?

Power dissipation in CMOS logic, ignoring leakage, is proportional to freq * V^2. Let our single layer CPU performance be 1 unit, limited by 1 cm copper and running at 1.2 volts (There's very little SiO2 for the heat to pass through.) At first glance, our 500 layer CPU with same voltage limited by 1 cm copper plus 50 micron SiO2 is 1 * (500 layers) * (1/500 heat per layer) * (1/2 thermal conductivity) = 1/2 unit. Layering loses. However, that is not the whole truth. Layering allows many more transistors, thus more clever circuitry, which might be enough to improve the performance some. 3D means shorter interconnects, shorter interconnects means less capacitance, less capacitance means less power dissipation. (The other major contributor to capacitance is the FET's gate.) I can only guess how much lower heat (more speed) that allows. Maybe 1.5X? speed is then 3/4 unit. That (1/500 heat per layer) is (1/500 speed) and with CMOS reduced speed allows reduced voltage.

Over a limited range, CMOS speed is proportional to voltage. By lowering voltage, heating is reduced. Thus reducing voltage means speed does not have to be reduced to 1/500 of the single layer CPU. With a supply voltage of 1.2 x 1/10 = 0.12, speed reduced to 1/10, power per layer is reduced to 1/1000 compared to the single layer CPU. 500 layers operating at 1/10 the speed is a 50x performance improvement.

Alas, we can't do that. Huge CMOS CPUs can't be made to operate at 0.12 V, and I don't know if it will ever be possible. I'll guess and say that somewhere in the range of 0.3 V and 0.6 V will some day be practical. If it's 0.6 V, speed could be 1/250, times 500 layers = 2 units. If it's 0.3 V, speed could be 1/62.5, times 500 layers = 8 units.

The above is too optimistic, because of difficulties in controlling threshold voltage and leakage, and the difficulties in massive parallelism and massive multi-threading.

I'd like to repeat the calculations for 10 layers and 50 layers. I'd like to check my work. I've already spent about 2 hours on this reply, so I'm giving up. Have fun.

In my estimate, the speed improvement in AI chips is going to see the same slowdown we've already seen in CPUs: single threaded performance is almost at a standstill and multi-threaded performance is increasing much less rapidly than it used to. If this slowdown occurs, there will be less pressure to replace existing AI machines with faster AI machines. This means a longer life cycle for existing machines.

Whether new facilities continue to be built will depend upon the degree to which AI is useful, and whether AI's usefulness requires more hardware. Nobody really knows.