We can use older processes if we have to. We'd be taking a step back of... maybe 5-10 years? Computers 10 years ago were not that much slower than they are today. Volume would be a bigger concern than performance. Maybe it'd force the tech industry to start writing more efficient user-facing software instead of depending on the incremental advancements made by chip designers and semi fabs.
> We'd be taking a step back of... maybe 5-10 years? Computers 10 years ago were not that much slower than they are today
There’s more to the world of computing than your laptop.
Stepping back to 10 year old GPUs and server CPUs would be a massive handicap on the country.
> Maybe it'd force the tech industry to start writing more efficient user-facing software instead of depending on the incremental advancements made by chip designers and semi fabs.
It’s not about the speed of your laptop loading Slack. Large scale compute is already squeezing as much performance as we can out of server hardware.
> Is “don’t buy stuff with TSMC chips” really a valid option we have?
Not sure that TSMC would want to do that either! We're probably their biggest market, even allowing for China.
> Isn’t that basically “stop buying high technology” to a large degree?
I think you're right, to an extent, at leastt in the near term.
However, we do have (and especially used to have) various fabbing here in the States, from Samsung to Intel. Especially the latter has been neglected, but these changes would probably accelerate on-shoring and perhaps bring some of it back here.
Don't forget that TSMC is in a country that is probably going to go through some significant instability in the next few years. From a business continuity perspective, we'd need to consider availability and supply chain management with the strong possibility of a major vendor being located in the middle of a hot warzone.