I think his idea is that there are times where the safest thing to do is to accelerate out of a dangerous situation on the road. Which is sort of true, but at those speeds most cars tend not to accelerate very quickly so I don't find it particularly convincing.
Yeah, I’m really skeptical just because at that speed human reaction times are so limited. It seems like the guys who said they should have to pay for seatbelts because if they crashed into a lake it might be hard to get out without drowning - true at least once, somewhere, but orders of magnitude less common than the times where a seatbelt kept someone alive.
In lots of situations speed doesn't really limit how much time you have to react, because it's something like "someone is merging into you without noticing you" or "you and someone on the other side of an empty lane both started to merge into it".
But ya, the statistical argument of how often does accelerating out really help, vs. often is someone driving too fast the cause of the danger, seems very relevant.
I was thinking of the only time in 3 decades of driving where speeding up made sense: I was on the 405 in LA and I was behind some guy who had a stack of mattresses poorly tied onto the back of a pickup truck. The top one flew surprisingly high up and I realized that the way it had caught the wind meant that if I braked it would land pretty close to me, so I went under instead and it landed a few car lengths behind me. At 45mph, in my early 20s, driving a sporty car that worked out but at 90mph I don’t think I’d have had time to process it quickly enough. I also note that this was once a long time ago where as I’ve lost count of the number of times I’ve been endangered by someone going much too fast or seen the aftermath (cars in buildings, upside down on the wrong side, etc.) so I would in a heartbeat take a world where the extreme speed outliers are gone.
