For my entire post college life (think financial crisis +) the difference in price between new and used cars has been inconsistent, erratic, and extremely hard to evaluate. While it's true that cars eventually depreciate enough to be cheaper, I have 2012 model year that's probably worth < 20% of what it was new even with reasonable mileage, when I look at used cars less than 5 years old and without a ton of miles, most models that are considered reliable or remotely popular seem to sell above 80% of new model sticker price, and often are marked at or above new sticker at car lots. Choosing between, say a new Coralla vs. a used Camry seems far removed from what I remember in my youth, which is that for model pairings like that maybe your new vs used model overlap would be another mark or maaaybe something like a Yaris, and you'd be like "buying new is for rubes." But today, you have to spend several days of part time work to figure out something like "a new model a, or b, in price range x or a used model c or d with sub n miles in price range z will have similar TCO, let's see what's out there..."
I think there's a reason that that Slate truck seems interesting to some people, and why people think stuff like the Maverick that has gone from a reasonable deal to a questionable one but still remains popular.
That said, I think it's true that most new car buyers are mostly buying expensive cars and that there aren't many cheaper cars despite evidence that they'd sell. So what I'm curious about is that means 5 years from now people will have to sell them at an increased rate of depreciation or they'll just keep them longer.
In the worst case this seems on the verge of cars being driven out of middle class hands (or houses going from two car to one car) if the median income relative to average car price grows more.