Time for all the efforts to converge on RISC-V I guess?
The giant’s open-source contributions from now own “should benefit #Intel first”, according to their tech-illiterate business management.
Those contributions, indeed, already benefit Intel first.
Intel drivers are worked on, tested and improved by an army of kernel developers out there that in many cases aren’t even Intel employees.
Intel’s position as an open-source good citizen so far has meant complete dominance in the data centers market (which overwhelmingly run Linux), without all the headaches given by folks like Nvidia.
If there were viable open-source alternatives to Nvidia/CUDA to train machine learning models you can rest assured that most of the data centers would use those. But there are viable alternatives to Intel. And, unlike Nvidia’s case, those alternatives are also already compatible 1:1 with software that runs on Intel hardware. So this will end up being a very shortsighted decision that will erode Intel’s market on Linux (including data centers, not only on the laptops of geeks like me that represent the 1% of the market), while increasing development pressure on Intel engineers instead of decreasing it (without the support of the community of kernel developers).
And for what? To profit from the firmware and drivers that they build? Those aren’t Intel’s products anyway. The product is the hardware. Unless they are planning to apply a subscription business model also to the usage of your own CPU.
Plus, it should be reminded that Intel has received hundreds of billions in lavish donations from the US government in the past 5 years just to keep it afloat, despite no longer being a competitive player, because America can’t afford to lose its largest hardware manufacturer. Public money means public code. Not locking up the public code that you already have.
Intel rethinking how it contributes to open source community
: Kevork Kechichian says x86 giant's contributions should benefit Intel first