3 ms·

EU should have its own leading-edge chip foundry. Intel has nothing to do there, not its place for that, they have to restore leading-edge chip manufacturing fo

EU should have its own leading-edge chip foundry. Intel has nothing to do there, not its place for that, they have to restore leading-edge chip manufacturing for america (and I mean the whole america). The issue is the economics of leading-edge chip manufacturing: I cannot see this industry providing affordable leading-edge chips without near monopoly on the global market. Namely, each significant "zone" should have this capabilities, but should not expect to "make money" (sell leading-edge chips at a "loss" to make affordable for all ppl). It should be considered strategic like defence: must be leading-edge, but making money out of it, optional, but nice if it can. The thing, is some world zones are not ready to do that or will sort of fail, then some other zones should be ready to manufacture at scale to cover the needs of those "friendly" zones. There, RISC-V, has a lot to provide: interop at the machine code level, for all those chips. The planned obsolescence of compilers and most computer language syntaxes is really toxic (this is the tip of this iceberg of toxicity) in the end.

YorkshireSeason 4y ago
> EU should have its own leading-edge chip foundry. Yes, "should". But it is unclear that the EU has the organisational capability to pull it off. The last few decades have seen the EU loosing it's technological edge in more and more fields, the more the EU makes (funding) decisions in science and technology. Whether there is causation or just correlation is an interesting question, see [1] for a recent discussion. > planned obsolescence of compilers Which planned obsolescence? Intel is backwards compatible at least to the x86 from 1978 (?). And I think even back to the 8080. IBM mainframes might still be System/360 (1964) compatible, although I have not checked recently. Given the RISC-V extensions and profiles tend to proliferate, and indeed that there is a competitive advantage in pushing your own extension, there is an often articulated fear that the RISC-V might fragment faster than those controlled by a single vendor. [1] https://news.ycombinator.com/item?id=33971577 https://news.ycombinator.com/item?id=33971577

YorkshireSeason 4y ago

> EU should have its own leading-edge chip foundry. Yes, "should". But it is unclear that the EU has the organisational capability to pull it off. The last few decades have seen the EU loosing it's technological edge in more and more fields, the more the EU makes (funding) decisions in science and technology. Whether there is causation or just correlation is an interesting question, see [1] for a recent discussion. > planned obsolescence of compilers Which planned obsolescence? Intel is backwards compatible at least to the x86 from 1978 (?). And I think even back to the 8080. IBM mainframes might still be System/360 (1964) compatible, although I have not checked recently. Given the RISC-V extensions and profiles tend to proliferate, and indeed that there is a competitive advantage in pushing your own extension, there is an often articulated fear that the RISC-V might fragment faster than those controlled by a single vendor. [1] https://news.ycombinator.com/item?id=33971577 https://news.ycombinator.com/item?id=33971577