Yes but you don’t need 2/3nm chip in a drone! STmicro is a global leader in many sectors for example, like automotive chips, probably Ukraine is using those chips.
If, for even 1s, they get in a position which is threatening, in any way, Big Tech AI (mostly US based if not all), they will be raided by international finance to be dismantled and poached hardcore with some massive US "investment funds" (which looks more and more as "weaponized" international finance!!). Only china is very immune to international finance. Those funds have tens of thousands of billions of $, basically, in a world of money, there is near zero resistance.
Don’t think that’s a fair interpretation of what I said.
Liquid money rich? No.
Can get pulled for big tech packages? Also no, for most of the employees.
AFAIK, big tech didn’t aggressively poach OpenAI-like talent, they did spend 10M+ pay packages but it was for a select few research scientists. Some folks left and came but it boiled down to culture mostly.
I am almost sure it was done so carefully that you can extract it from the abominations which are the whatng cartel web engines with a direct to OS abstraction layer that with only some little amount of work.
What is the model used for maths since it is not LLM? This model would have to be strap on a formal solver as a tool. Are those models training to use "scratch memory"? So many question about the models used for real maths.
Better favor as much as possible RISC-V implementations.
But, I don't know if there are already good modern-desktop-grade RISC-V implementations (in the US, Sifive is moving fast as far as I know)... and the hard part: accessing the latest and greatest silicon process of TMSC, aka ~5GHz.
Those markets are completely saturated, namely at best, it will be very slow unless something big does happen: for instance AMD adapts its best micro-architecture to RISC-V (ISA decoding mostly), etc.
And if valve start to distribute a client with a strong RISC-V game compilation framework...
This is kind of a solution in search for a problem. RISC-V will grow only if people find some value in it. If it solves their actual problems in ways that other architectures can't.
Yeah, the primary reason RISC-V exists is political (the desire to have an "open source" CPU architecture). As noble as that may be, it's not enough to get people or companies to use (or even manufacture!) it. It'll either be economical (costs) and/or performance (including efficiency) that drives people.
It took ARM decades to get to where it is, and that involved a long stint in low-margin niche applications like embedded or appliances where x86 was poorly suited due to head and power consumption.
I don't think that's the primary reason there's momentum there. The reason is to avoid ARM licensing fees and IP usage restrictions.
I think you'll see ever more accelerating RISC-V adoption in China if the United States continues on its "cold war" style mentality about relations with them.
That said we're a long long way from Actually Existing RISC-V being at performance parity with ARM64, let alone x86.
Yep, licensing fee and IP usage restrictions is a massive decision point on some silicon markets.
The other massive point: RISC-V integrates a lot of CPU "we know now" in a very elegant "sweet spot".
And it is not china only, the best implementations are US, and RISC-V is a US/berkley initiative re-centered in switzerland for "neutrality" reasons.
If good large RISC-V implementations do reach TMSC silicon process (5GHz), some markets won't even look at arm or x86 anymore.
And there is the ultimate "standard ISA" point: assembly written code then become very appropriate, hence strong de-coupling from all those, very few, backdoor injecting compilers.
On many of my personal projects, I don't bother anymore: I write RISC-V assembly which I run with a small x86_64 interpreter, that with a very simple pre-processor and assembler, aka SDK toolchain complexity close to 0 compared to the other abominations.
And I think the main drawback is: big mistakes will be made, and you must account for them.
That might be true for the desktop, but RISC-V is wonderful from a pedagogical and research standpoint for academic uses and in the embedded world its license and "only pay for what you need" is also quite nice.
reply