My understanding is that the "deliberate slowing" reported, was such that if you have firefox send a chrome user agent, that one gets faster (and functional) results.
Is that consistent with the "it is because FF doesn't support new thing yet" explanation?
No because again, IIRC, sending a chrome user agent to YT in FF breaks a lot of the functionality.
Ah, “the parts that work are faster but a lot of it is broken” does seem like it suggests the benign explanation
Does it? What stops working, since I haven't noticed anything?
I believe it was that YT was using the legacy v0 shadow DOM API that only Chrom[e|ium] supported, with a slow polyfill for other browsers, rather than the much better supported v1 API that all other major browser makers agreed to implement.