4 ms·

It's a dramatic article likely intended to help sell his book (with an equally dramatic title). Missing from all of this is, well, data. He squarely blames la

It's a dramatic article likely intended to help sell his book (with an equally dramatic title). Missing from all of this is, well, data. He squarely blames large vehicles, SUV's. Now, I have no horse in this race, I prefer small vehicles for practical every-day driving. For moving family, dogs, camping trips, etc., a minivan is great. Yet, at a very basic level, some of what he is saying makes no sense to me. Take this introductory statement: "As a country, we hit the threshold of 1 million cumulative deaths in 1953, 2 million in 1975 and 3 million in 1998. While the past several years of data have not yet been released, I estimate that the U.S. topped 4 million total road deaths sometime in the spring of 2024." So, 4 million by 2024. Fully half of that by 1975. Care to guess how many SUV's existed in 1975? Also, I am going to guess average speeds where likely lower. He is blaming SUV's and claiming that making them 10 cm lower would have a dramatic impact on results. OK, well, if the data up until 1975 is any indication, this is nonsense. How about distractions, drunk driving and, yes, pedestrians and cyclists. Accidents are not always attributable to one side or a single variable. The world is a complex multivariate problem. This guy is trying to sell books by hating on everything and only blaming one or two variables: SUV's bad, wide roads, bad. And then there's are the circumstances of each accident. He gives the example of an SUV losing control, jumping a curb and killing a father and his kid. Well, not every accident is like that. And, no, life isn't going to be perfectly safe. You are not going to go down to zero accidents and zero fatalities, no matter how hard you might try. Which brings me to the last point: He offers not data whatsoever on how things would improve if his world view were to be adopted. If we ripped-up and modified every road and shrunk every vehicle by 10 cm. You see, it is easy to be critical of anything when you don't have to be accountable for any of it. Happens all the time with "experts". As an example, the financial world is full of economists who just "know" all there is to know about finance...yet none of them are billionaires and few (none?) run successful businesses at scale. Yes, of course, road safety is important. I just think that the solutions include far more variables than what this guy trying to sell a book seems to be focusing on in this article. In a few decades, self-driving cars might solve this problem nicely. Don't know.

achenet 2y ago
you've got a good point about experts being critical with no accountability. However, when studying a complex, multivariate problem, if you find one or two variables that can easily affect the outcome (lower hood heights), it does seem sensible to optimize it, while obviously keeping in mind that you can optimize for other things.
- robomartin 2y ago
  The point may have been lost: Fully 50% of deaths occurred during the pre-SUV era. That's 2 million out of 4 million, not a trivial number. If we go out to 1998 and call that a less intense SUV era, that's 3 million deaths, of 75% of all deaths since the beginning of time. This, to me, on first inspection, is pretty strong evidence indicating that the 10 cm prediction is very close to nonsense. The prediction is no different from saying TSLA will go to $1000 because of some randomly-picked variable that will never be tested. Sounds fantastic and cam make me look very intelligent, yet, I would not have any accountability at all in making such a prediction. Now, if I wanted to sell a book about how to make a billion dollars trading Tesla, I would write a bunch of blog posts explaining how this could happen and feature a nice picture of myself holding the book along with an Amazon link. Which is what this article is about.

achenet 2y ago

you've got a good point about experts being critical with no accountability. However, when studying a complex, multivariate problem, if you find one or two variables that can easily affect the outcome (lower hood heights), it does seem sensible to optimize it, while obviously keeping in mind that you can optimize for other things.

robomartin 2y ago
The point may have been lost: Fully 50% of deaths occurred during the pre-SUV era. That's 2 million out of 4 million, not a trivial number. If we go out to 1998 and call that a less intense SUV era, that's 3 million deaths, of 75% of all deaths since the beginning of time. This, to me, on first inspection, is pretty strong evidence indicating that the 10 cm prediction is very close to nonsense. The prediction is no different from saying TSLA will go to $1000 because of some randomly-picked variable that will never be tested. Sounds fantastic and cam make me look very intelligent, yet, I would not have any accountability at all in making such a prediction. Now, if I wanted to sell a book about how to make a billion dollars trading Tesla, I would write a bunch of blog posts explaining how this could happen and feature a nice picture of myself holding the book along with an Amazon link. Which is what this article is about.