This is insanely, embarrassingly false. There's no limitation to the number of parallel execution pipes and it has absolutely nothing to do with how "complex" the instructions are. I mean for fuck sakes, you just have to look at Zen3 (link: https://www.anandtech.com/show...) for a counter-example to some mythical "limit" to parallel x86 pipes.

If all that was necessary for better performance was more parallel execution pipes, they'd have shoved 69 of those in there long ago. One of the wider designs I've ever worked on was Project Denver (link: https://www.anandtech.com/show...) with 7 wide superscalar pipes. Embarrassingly, that thing performed like a schizophrenic sloth.

There's much more to CPU performance that just how much execution resources you have. And that's why Intel hasn't spent a lot of time increasing them in their designs: there are bigger bottlenecks to be battled.

Almost every modern ARM implementation does instruction decoding and not something much more complex than the x86 front-end of most Intel or AMD CPUs. The complex x86 instructions in modern x86 CPUs are handled by ROM lookup tables.

And ARM is anything but a "simple" instruction set -- especially prior to v8. And most (except Apple) modern ARM implementations still has to support the older v7 ARM instructions. That means things like Thumb, load-multiple, NEON's more....eccentric vector load/stores, etc.

And you know what? There's a reason ISA's can be complex. Instructions get added because someone needed them and it was more performant and efficient to let the compiler specify (through an instruction) what to do rather than the uArch figuring it out at runtime.

The "purest" RISC ISA today is RISC-V (brainchild of Patterson, who loves simple RISC ISAs) but that simplicity has led to all sorts of limitations of where it can be used. There's no security model, for example. Or virtualization. Or wide SIMD...

NUMA works in certain specific HPC but general purpose OS's don't have the insight into app behavior to realistically be able to schedule well in NUMA configurations. Caching turns to work out a lot better, which is why the big iron Power and Xeons don't have NUMA and instead use on-package DRAM chips as caches.

I question this assertion. Do you have data behind this? Most places I've worked at (that wasn't Apple) used Office by default and mixes Windows and Mac machines. The default Office suite was MS Office.

And I said nothing, because I did not make headphones.

Actually, Airpods haven't meaningfully dented 3rd party headphones sales. And bluetooth works better on an iPhone than most devices I've used.

I don't know that this is true. Sure the cotton weavers put up a stink at first but they eventually were paid more (and happier) in factories. Crop field farmers put up a stink when the tractor and havest machines came along but they eventually found jobs (and were happier) in office jobs.

Even grocery clerks find it better being small shop managers (because of the lower cost, there can be more stores) rather than just checking people out.

At some point in human history, reading and writing were considered skilled labor. The idea that there are classes of work that are "skilled" vs "unskilled" is, I suspect, a giant myth. There exists very few jobs in modern day that don't require some degree of specialization.

Even the Chinese kid assembling iPhones had to take some (grueling) time to learn how to do it, which puts him/her above his/her peers who are trying to escape the farmlands.

The problem was never "skilled" vs "unskilled". The problem in the US is that what was considered "unskilled" was de-emphasized and everyone told to go to a 4-year university to get "skills". All of the trades (including advanced manufacturing) that normally fall under an apprenticeship model was left to rot and die.

Let's take the most extreme example we have today: Elon Musk. In 2020, he qualified for a $2.9B pay package (all of it in company stock). Tesla has roughly 70k employees. So that amounts to about 40k/year in company stock to each employee.

The key though, is that Musk is paid primarily in company stock -- his salary is actually the State mandated minimum. So the company doesn't actually allocate that much cash (as in USD) for executive pay. Were you to dock his salary, you'd end up with less than $1/employee.

The stock also is award because the company's stock is growing so much (that's the benchmark).

So you're right that the company could take Musk's shares of stock and give it to employees instead. But company stock isn't cash -- the company can freely print more of it (as it did for Musk). And it does so regularly as compensation for its employees as well. There's no "freeing up" to do in this case.

How do you deal with an increase in real wealth (houses, goods, population) not being followed correspondingly with an increase in currency supply?

That house used to need 3 cars for each employee to get to work. Now they probably only need 1 car for groceries. Maybe a second like a van for family trips -- that would previously have eaten gas.

So you replace at least one ~10k car with a ~3k VR headset and an extra ~50/mo in internet tier costs.

You can also tell your insurance you're driving a shit ton less (which I did). The insurance premium alone for 2 cars vs 3 and driving next to 0 miles on one of those car is probably more than your additional internet access.

Communication is definitely lacking with even a good webcam. For one, there's hand-gestures, even facial expressions that get lost.

It depends on what your'e discussing. A rigorous whiteboard session where algorithms are discussed is definitely harder. Not impossible, but harder.

VR (with low-latency internet) would solve this.

Good for you.

I log in in the morning after interacting with my wife (who sits a few ft from me) and my pets. Maybe my kids some day. The idea that I spend most of my waking hours servicing other people rather than being with my family seems like a sad sad dystopia.

I pity people who don't sit at their own desk and home; only interacting with others when they choose.

They probably consider the stench of the unwashed masses unpleasant long before they heard of COVID.

I seem to recall during the WDCC announcement that they demonstrated Docker on Mac on Apple Silicon...

I don't believe Apple PR has, even once, referenced any specific benchmark. Simply a vague "faster than" line graph.