This is a nice low-stakes example of how we can't expect automated driving to just be "like people but with better attention spans and reaction times and no risky stuff."<p>There will be things that are alien to human expectations, novel problems, and differently <i>shaped</i> risk profiles.