I've been burned by them, so I experienced the blunders first hand.
I have a Centrino 6235 Wifi N/Bluetooth adapter, it works just fine except it needs Intel drivers for the Bluetooth part. Their answer seems to be "just upgrade it, lol", suggesting some newer Wifi AC adapters. Higher speeds, they say.
Well, first of all, I won't get anywhere near those speeds for various reasons, and second, why do I have to throw away a perfectly functioning piece of hardware? Ridiculous.
Lol. A hardware manufacturer doing things that results in consumers buy more hardware? That's like a software company discontinuing sales of an old OS in order to sell its latest abomination. Intel and MS deserve each others. Toss apple on that pile too imho. Synergy between AMD and linux/foss is a good thing, consumer-friendly behavior that should be encouraged.
Which is about all we can ask of companies doing open source work. And it's still vastly more than most are doing.
Not sure I am reading this correctly.
AMD already has GPU RDNA drivers in Linux kernel. So It doesn't need Samsung to kick start this.
The partnership with Samsung ARM and AMD RDNA GPU SoC only goes for Phone and Tablet, or any market that AMD silicon currently does not currently operate in.
I think that was the GP's point, ie Samsung wouldn't need to invest in writing new drivers but should rather be able to do relatively easy adaptations to existing ones to get things off the ground.
AMD has an interesting portfolio of processors, gpus and now fpgas. So here's another fantasy: Amd Risc-V APU. AMD hasn't had a totally unique architecture since the am29000 (29k). So it would be interesting to see them build their own CPU arch using an open instruction set with their GPU. Start with a nice quad core that can handle a laptop or mini desktop a la NUC. Another fantasy would be an AMD Zynq.