> For example gating new things which could become a stability issue.
And yet we were somehow able to ship stable software to millions of people in the days before telemetry/studies. What telemetry/studies truly provides is offloading QA costs to users.
Sure we were. I have video games that were pressed to CDs with no possibility of updates and they still run to this day. On a complete play through of many of them you wouldn't notice a single bug.
Using users for QA has made us lazy and borderline incompetent.
I think you're thinking of extreme cases here like games which are released with bugs these days for holiday rush, because the patching is easy / built-in. That's not the case for everything and not the case for Firefox.
And the stable software in the past may be due to nostalgia. They had bugs. Even before dialup was common, games were patched: Diablo 1 had a number of patches for example, Quake 3 went up to 1.32, Dune 2 (1992) got 1.07. We just accepted bugs, because... what can you do?
And yet we were somehow able to ship stable software to millions of people in the days before telemetry/studies. What telemetry/studies truly provides is offloading QA costs to users.