There is an implied relationship between age and interpretation where older sources elicit a more skeptical review. Harley certainly makes this point in his 1988 essay, arguing that maps are neither true nor false (just as a painting or piece of music is neither true or false), but instead represent a discourse of the time. In 2008, Cosgrove adjusts this argument slightly: just as we should interpret maps from the distant past as biased sources, we should also consider the values inserted into modern-day maps.
Of course this is important to keep in mind in our research - as with all sources of data, maps are biased in certain ways. And Cosgrove is right that a map made on a computer is inherently no less subjective than a map drawn by hand. He expresses reasonable concern that maps "permit the illusion" that a place can be represented completely (168). I take some issue, however, with Cosgrove's implication that GIS is somehow problematic because there are lots of people who are not trained as cartographers running around making maps all day. GIS has democratized map-making: there are more people who can make maps than ever before, and these people are far more diverse than the elite cartographers who worked for Kings, Sultans, and Emperors. If the maps of the past were "designed to make a more permanent social order" and "maintain the status quo" (Harley 282), perhaps we have entered a slightly different age of map-making where GIS can actually challenge the existing order of things. Are individual maps still biased? Yes, but maybe they are collectively more representative than they used to be. We should always keep the cartographer's intentions in mind when telling cartographic stories, but I think we can take a slightly difference approach to modern maps.
One other note: I think Cosgrove doesn't give enough credit to spatial analysis that is not represented by a map. Increased computing power hasn't made maps more "true," but it has improved our ability to perform computationally expensive spatial analysis (that may not involve maps at all). If anything, my takeaway from Cosgrove is that as we craft cartographic stories, we should not rely too heavily on maps for data or data analysis - maps communicate important findings but are not themselves a source of truth. If we firmly separate spatial analysis from cartography, we get the most value out of maps by reading them for what they are: storytellers.