We live in the sort of weird world where people would rather look into ways of stopping those people dying from CO poisoning than snarkily dismiss their deaths as statistically irrelevant.
Pretty much all of those 35000 deaths involve mistakes, faults or cases of pure bad luck whose individual contribution to fatalities is comparatively small; pretty much all of them get investigated to see what went wrong and whether it's something that's likely to be repeated or something which can be avoided in future without disproportionate cost.
Your comment is the snarky one, and you attempt to dismiss 35,000 deaths rather than 2. In every single one of those 35,000 cases, someone decided to drive that day, on purpose. You almost got it right: In this crazy wacky weird world we live in, people would rather talk about 2 CO deaths caused by some dumbshit accessory nobody needs, than the 35,000 caused by their own complicity and laziness. Nobody asked whether those deaths were intentional, but you answered anyway, so that's interesting.
Way to totally disregard my point that pretty much all the 35000 deaths consisted of one or two individual deaths, some of which see junctions redesigned or products recalled as a result without anybody haughtily dismissing their interest in seeing a fix with "how can you consider this to be deadly when many more people die from many other things related to vehicles?"
Personally I think there's more value in fixing the dumbshit accessories nobody needs which are a recurring cause of relatively small numbers of injuries and death than focusing on many accidents attributable to discrete random errors and deciding we're all powerless to change a thing, and I think dumbshit accessories nobody needs are a more easily corrected problem than need for motorised transport or humans (and AI) are not infallible. Evidently your mileage varies.
Pretty much all of those 35000 deaths involve mistakes, faults or cases of pure bad luck whose individual contribution to fatalities is comparatively small; pretty much all of them get investigated to see what went wrong and whether it's something that's likely to be repeated or something which can be avoided in future without disproportionate cost.