I run it on my RK3588 based MNT Pocket Reform. I have to force the GL ES 2 backend because of, presumably, Panfrost bugs, but otherwise it runs well despite the fairly weak CPU and GPU.

They did improve this in more recent generations, but you need a W series chipset to use it.

There are 16" laptops with ECC, you can get a ThinkPad P16 with it for example. I've yet to find any 14" devices with ECC though.

Interesting, I actually have a thinkpad p16s, surprised I didn’t notice ECC availability.

Similar experience. I played with overclocking the DDR5 ECC memory I have on my system, it would appear to be stable and for quite a while it would be. But after a few days I'd notice a handful of correctable errors.

I now just run at the standard 5600MHz timing, I really don't find the potential stability trade off worth it. We already have enough bugs.

I believe caches and maybe registers often have ECC too though I'm sure there are still gaps.

50W average doesn't seem absurd, peak power is going to be an order of magnitude higher, but computers are often running pretty close to idle...

Are the instructions missing, or does iOS just not run the hypervisor?

I suppose we'll find out pretty soon, supporting virtualization would be important if they wanted to sell these to CS students that need eg. Docker.

So my observation comes from the fact that UTM webpage: https://getutm.app/faq/#what-are-the-limitations.

Now this webpage may be out-of-date, so take my claims with a grain of salt.

A bootloader exploit at that, and we haven't had one of those in years.

You don't need to be writing assembly. Anything sharing memory between multiple threads could have bugs with ARM's memory model, even if written in C, C++, etc.

As someone that uses a YubiKey for WebAuthN - I really wish Bitwarden didn't do this. I know I can turn it off, but it's a bad default.