@marcan ehhhhhh. I thought they were merely crippling Adblock users, but arm users… uncool. Gonna use my user agent changer…

@marcan that sounds like a thing that was set ten years ago and never updated; I’d bet a dollar no one currently on the team even realizes this is happening

@marcan Wasn't it Google that was going to phase out user agent strings altogether? I guess that's not going too smoothly.

https://www.zdnet.com/article/google-to-phase-out-user-agent-strings-in-chrome/

@marcan User-agents are the biggest lies on the planet. Because of stuff like this.

@marcan Huh. I just checked, and MacOS does also report Intel. So... yeah.

@marcan I remember reading recently that Youtube was crippling Firefox, and that wasn't even platform-dependent, so the shoddiness is not just for Arm:
https://www.tomshardware.com/news/youtube-responds-to-delayed-loading-in-rival-browser-complaints

@marcan It surprises me that Firefox puts the correct architecture in the User Agent. Didn't we start freezing that a while ago? Does it report Wayland instead of X11 also?

Edit: I assume it might have to do with sites wanting to auto-detect the right executable for downloads.

I believe this only triggers on *software* decoding, which is why it particularly affects us (and not, say, macOS which has hardware decoding).

@marcan because only low powered phones run on arm right

@marcan Are they maybe assuming that aarch64 is mobile only? In that case, it _might_ make sense somehow.

@marcan YouTube product management is dangerously detached from the reality. This whole "guessing the resolution" thing is absolute nonsense. "Yes, downgrade my video to unwatchable because I've just had a 2 seconds interruption of the internet" - said absolutely nobody ever.

If you find any contact, explain to them that 99.9% of their users will never want anything below 720p (and even that only on very slow connections).

@marcan This screams to me like being some sort of decade old hack that was intended to fix performance on an old ARM device. As evil as Google may be, I find it hard to believe that this would be intentional.

@marcan
everyday reminder why should I never return from Invidious :)

@marcan oh, is this why I keep getting 480p^W 360p on android firefox's youtube 😐

...or maybe something else

@marcan Yup, that is always google, first it was lagging the video player for firefox users, now this

@marcan Is this for YouTube only that chrome reports x86 ? Or does it do that on all websites.

Because that would mean that any website relying on this to suggest the download of anything for a certain arch would land by default in the wrong page / tab / whatever in that page.

Seems like an idiotic approach to rely on arch to define codec support / performance / etc...

@marcan ^^ google is always playing "games" like this.

@marcan Question, what happens if no arch is specified, or something exotic (such as ppc, sparc, mips, "Apple Silicon", RISC-V) ?

@marcan i've encountered this with aarch64 in other forms - various websites (which should know better) will point blank refuse to serve you anything but their mobile site unless you tell Firefox to pretend it's an x86 machine. that seems to be the logic - anything arm must be a phone, whatever it says to the contrary

@marcan It strikes me (an inexpert outsider) as super-crude to make decisions like this on the basis of user agent strings. Wouldn't it make more sense to adjust defaults to match the actual client performance? The video player has a lot of "stats for nerds" available once a video starts playing.

@marcan What about websites that host software and provide the default download binary based on the arch from the UA?

@marcan There are probably hundreds of millions of Smart TVs and Android TV boxes having like 8 little ARM cores and not great performance. But yeah, this heuristics is outdated.

@marcan wild guess: They are using this to prevent mobile browsers from trying to play videos in software but they are explicitly excluding Chrome OS as it typically runs on laptop hardware that should be able to play the needed codecs.

@marcan how to test or replicate this? I seem to get the correct resolution based on the when using Firefox.

@marcan motherfuckers

@marcan

Mozilla/5.0 (X11; CrOS aarch64 10452.96.0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/66.0.3359.181 Safari/537.36

I'm counting 1.5 truth and 6.5 lies there. (counting chrome as a half-truth)

@marcan That is ridiculous because if it was designed to sniff a TV, why will they decide to take the low resolution edition on the big screen? Wouldn't it be worse in this case?

@marcan This is very naughty

@marcan what happens if you report arm64 as the architecture?

@marcan Honestly, user agents have outlived their usefulness, all vendors should just agree to use the same one and never change it.

@marcan Thats one of many reasons why I dont use youtube. Its communist un garbage.

@marcan Found Chrome's answer for lying, apparently for privacy reasons https://bugs.chromium.org/p/chromium/issues/detail?id=1500114

@marcan It was recently reported that Google is crippling Youtube performance on Firefox.

@marcan this is sounding more and more like the guy who did this just forgot Firefox exists on ARM outside of the Android app

@marcan so cripples a rival engine for its own benefit?

@marcan Too bad they don't have an option to get 480p but with a decent bitrate.

@marcan I guess that’s not about the CPU but the quality of drivers for hardware decoding

@marcan Total webshit death

@marcan Google disabled U2F based on user agent for a long time as well.