Very narrow? The masses could get a computer for almost all the time integrated circuits have been produced.
And the entire time, better manufacturing pushed prices down at least every few years. Including RAM going up multiple times. We would have to buck that entire trend on top of having a failure of supply and demand to end up with the current prices being permanent.
(And don't be dramatic. Even at current prices, computers are available for the masses.)
> Very narrow? The masses could get a computer for almost all the time integrated circuits have been produced.
The “personal computing for the mass” era hadn't even started when I was born, and I don't even consider myself old yet.
> And the entire time, better manufacturing pushed prices down at least every few years.
It's not because something happened before that it will necessarily happen again.
> Even at current prices, computers are available for the masses.
Computers are still available, but for the first time they are significantly less affordable than just a year ago. And nobody's saying the personal computer era is over yet, we're arguing that it's a possibility in the near future, and you won't convince me otherwise by saying “but it has always been like that” because it's not how it works, and because I'm just old enough to know it hasn't.
The “personal computer” has already been mostly replaced by locked-in terminals in the general public, so it's not like this worry is coming out of nowhere.
The masses could get one in the 70s or 80s depending on your exact threshold. Integrated circuits had only existed since about 1960.
And the manufacturing improvement and price drop ramp has been going for 65 years. It's not guaranteed but better lithography machines keep being built and it's still happening with CPUs and GPUs.
> You are embarrassing yourself. «Ce qui est excessif est insignifiant»
Or you could not be an ass.
I think you're saying it would take an unreasonably extreme threshold to make that statement true? It does not. Note that "could" is different from "sufficiently motivated". Even with the limited abilities of computers then, 15% of US households had one by the end of that period, let alone the 46% and 37% numbers for school and work use. If computers were as useful as they are today, the household number would have been well over 50%.
> If the hyperscalers are able to make more money from these chips
There's a limit on how much volume they can absorb.
> 15% of US households had one by the end of that period,
You realize it's like saying that automobile became mainstream in 1900-1950 and using figures from the late 50s to justify your the range you picked?
> If computers were as useful as they are today, the household number would have been well over 50%.
What are you doing here? The question is “how long personal computers has been a mass phenomenon”, not “how much people from the 19th century would have bought one if it existed by then”…
> There's a limit on how much volume they can absorb.
And there's a limit in manufacturing capacities. Nobody will invest to build a additional $50B fab to satisfy a demand that pay less than the (hyperscaler-driven, in my scenario) market price.
Again, I'm not saying it's what will happen, I have no superpowers to know the future, but neither do you.
> What are you doing here? The question is “how long personal computers has been a mass phenomenon”, not “how much people from the 19th century would have bought one if it existed by then”…
Oh that's where the miscommunication is.
I'm treating "so cheap it was available for the masses" as the ability to purchase. Is there a reason I shouldn't? And yes they did exist. What didn't exist was an internet to plug them into.
They became mass affordable in that period, and tens of millions of people did buy them even without online access.
The number of actual purchasers is significantly less important than the number of possible purchasers, because we're taking about price regressions on modern day computers, not actual computer use 50 years ago.
> And there's a limit in manufacturing capacities. Nobody will invest to build a additional $50B fab to satisfy a demand that pay less than the (hyperscaler-driven, in my scenario) market price.
Why not? Most of the existing fabs were built with the expectation of, let's say 50% markup over costs. And they got similar investment.
The price is now 4x higher or whatever, for 500% markup. Even if a new entrant can't get any of that hyperscaler money at all, just being able to sell RAM at 1.5x the old price to consumers would give them >100% markup. That should be able to get billions of dollars of investment shouldn't it?
And the entire time, better manufacturing pushed prices down at least every few years. Including RAM going up multiple times. We would have to buck that entire trend on top of having a failure of supply and demand to end up with the current prices being permanent.
(And don't be dramatic. Even at current prices, computers are available for the masses.)