I think your arguments are sounds. Don't break existing functionality is good but you have to take it into context. Sometimes you have to in order to make the product as a whole better.
"You don't like it? Too bad, we're doing it anyway" is what such statements are actually saying.
Emotionless calculation: X% people like it current backspace behaviour, Y% people use it only by accident (they may actually have telemetry to know the percentages). If they change the behaviour, Y% will be happier, which X% will split into "installed extension to revert", "won't care after a month", "switches browser". If the overall change is positive enough, they made a good product decision, even if you and I don't like the new behaviour.
If this never happened to you, good for you, you lucky bastard. Now get out of the way, this is a quality of life improvement for millions of real people out there. Hard working people. Mothers and Grandmothers. Think of them while you are bothered to work around this slight inconvenience of yours.