@augieray The interpritation here isnt a great one.
The wording is proper to say that we expect arithmetic growth but get exponential. But it is incorrect when the section starts with "in other words"... We do not **underestimate** the risk in approach, we misunderstand the growthrate of approach. This isnt the same because a linear growth can exceed an exponential growth during the relevant range of the function.
I attached an image as an example. Here I picked an agressive exponent to show the effect more clearly, the plot is of x vs x^5. Notice how during most of the graph the linear one actually significantly exceeds the exponential. This is my point, seeing linear growth will actually "overestimate" the danger of covid as you close distance not underestimate it, at least at first.