Or is this like Scott's Tots?

The steam deck is AMD.
And most linux users will recommend AMD for a boot & play first experience

> Our governments, at least in theory, are controlled by us, the people

Whenever you go vote, tell yourself "this might be the last election" . Doing that puts technologies and power structures in perspective...

If it made the news, I'm *assuming* it is quite a common issue.
A common issue in a key part of the system.
If your QA team doesn't even log in to the system using most methods while testing, why do you have it? Logging in is functionality 101.

If they hadn't had so many fuck ups in the recent past, I'd be lenient too. But this is becoming a farce.

More than 10 years ago, in my C++ class, every week we had to complete an assignment which we then validated using an online validator. That validator was the single source of truth for our marks.
It ran with mudflap (now ASAN), without, with other sanitizers, and had an extensive battery of tests it validated the work against, also ran performance tests.
It also ran an automatic check for plagiarism. (never heard of a false positive)
It was a love/hate relationship.
And the best: it wasn't binary (pass/fail), we got a score out of 10. Getting above 8 was really hard.
But it was online already back then. Printing code is IMO stupid.

I hear they license it to a few other schools now...

> but does the consumer get a better experience?

Who cares? It's a race to the bottom of the expense sheet. When the consumer has nowhere else to go (either price or everyone does the same) , the consumer experience doesn't matter.
I mean, have you seen windows 11?

They'll learn to run their own instances. At least that's my pipe dream...

If it can find dupes, editors will have a harder time justifying their roles

https://news.slashdot.org/stor...

To quote:

"Water had been deemed a threat to brawndo's profit margin.
The solution came during the budget crisis, when the Brawndo corporation simply bought the FDA, and the FCC..."

You guys started with the EPA, fair enough...

The same way the decreasing number of pirates causes global warming

https://www.spaghettimonster.o...

For instance, I never screencast. yet it is considered a common use case.
OP mentions "plug a sound card in" , last time I did that was around 2000. Since then AC97/Azalia hold the fort.

For me (and my social circle) a common use case is working build system, multiple choices of compilers, and having headers and debug symbols for all software in the OS. Yet almost no OS does that.

I'm trying to say that what is "common" is dependent on which usecase you and your social group have.

> I want GUIs for all common tasks

A common task for you might not be for me...

Yeah, ok, if you reaaaallly want to get pedantic:

The Linux kernel GPU driver will expose the VCE (AMD) engine to userspace, which is then used by mesa's implementation of VAAPI, which is then in turn used by ffmpeg to power any video application you might be using. That's for intel/AMD/nvidia.

Then you get various ARM SoC GPUs (I know of rockchip and Allwinner) which have stateless decoders [1] (available through the v4l2request API) where part of the codec is actually implemented in the kernel. ffmpeg sits on top of that.

better?

The point being that if the hw has it and it's probeable, GNU/Linux OSes will use it.

[1] https://www.kernel.org/doc/htm...

From what I understand, Linux doesn't care. It probes the hw, finds the capability, and uses it.

Intel already does this integration (within the limits you described): https://en.wikipedia.org/wiki/...
It's a real PITA to pick an intel wireless module that will work with non-intel CPUs.