The Car Safety Feature That Kills the Other Guy (slate.com)

Interesting chap, been aware of them for a while.  Seems to favour the measured approach, which we need more of.

You're a few decades late to the party!  It's only the latest arms race to improve things on the inside without regard to the outside (bigger, faster, better accelleration, more screens and toys ...).  Depending on how much of a mitigation you think "emit elsewhere" is you might even add in "batteries included" - although it may be too soon to tell.

Isn't "my own private Chashitsu" half of the promise of the car? (The other half being "powered exoskeleton" or "magic carpet / teleporter").

I suspect this is true - bit like how wars throughout the 20th century seemed to produce more people with severe injuries - because during e.g. WWI and WWII those people simply didn't survive to reach the more capable medical services at the rear.  (Helicopters and improved first response trauma care meant that they did).

It would be interesting if someone has some numbers.

In the UK we've made roads "safe" in 4 ways: removing non-motorists from the roads (mostly by "education" / dislike and lack of convenience or even fear - but sometimes also by physical barriers).  The road designs and infrastructure have improved*.  There have been some "soft infra" changes e.g. better training, some legal tweaks, there's now some social pressure against some dangerous practices e.g. drink driving (but not speeding or mobile device use...).  Finally there's the protection by vehicles themselves and maybe emergency response after it's all gone wrong.

The first has had a big effect, not so sure about the training / rule / social changes.  Better infra likely helps but I'd suspect like you that improved vehicle safety - in general and in a collision - may be a major thing.

Auto express had an article reminding everyone of some of the many improvements:

https://www.autoexpress.co.uk/car-news/90221/the-evolution-of-car-safety...

* The list is quite extensive actually.  Road lighting, rumble strips, turn lanes, cats' eyes, more traffic lights, energy absorbing crash barriers, standards for widths / turn radii at different speeds etc.  I'm amused by "but cycle infrastructure is too expensive" / "we'll have to build a lot of stuff" when put in the context of what the average junction involves / costs...

Good point re infrastructure. Although I suspect that also has unintended consequences. I vaguely remember much money being spent on straightening the bends on main roads (hence all those nice big laybys) to speed the traffic up; now we're spending more money to slow it down again. Similarly, junctions on residential streets were given an 'apron' to smooth the turn and provide better sightlines; and people either park on them or use them to swing round corners at speed. Do either make life safer I wonder? And both these disadvantage walking - faster traffic, a bigger junction to cross.
I think I read somewhere that some junctions have been improved by making it less easy to see, so drivers are forced to slow down and actually look.
I worry about all this 'improved' safety generally making drivers less careful. Road markings you just follow mindlessly, big metal boxes, in-car clever stuff that removes the need to watch for hazards, etc. And back to the point in the OP.

I'm more hazy on "risk compensation" but as far as infra goes the UK has gone a fair ways for motorists and for everyone else essentially just made it so they're out of the way, without making that either convenient or particularly safe.

I'm pretty happy for motorists to have further safety improvements to both vehicles AND infra (e.g. banning and even physically preventing overtaking into an oncoming lane as that's a common source of cock-ups) ... IF the same is applied for other modes of transport.  Ideally in a system which aims to comprehensively address the safety of all modes ([1] [2]).  Strangely, where this has been done lots of people find that they don't need to use their cars for every journey.