That's actually a position of reasonable people who engage in non-greedy reductionism, mostly replying to greedy reductionists (to use Dennett's terminology).

Trying to understand reality is futile is a narrow and trivial sense: the map will never completely match the territory. That's not the notion I'm criticizing.

In the case of Boyd, when he says "any inward-oriented and continued effort to improve the match-up of concept with observed reality will only increase the degree of mismatch," he seems to imply that the harder we work to create a model of observed reality with our concepts, the worse the match will be. That's a truly weird notion.

Maybe his quote goes from being an example of thinking gone horribly wrong, to thinking gone horribly explained, if we try to figure out what he means by "inward-oriented."

When this orderly (and pleasant) state is reached the concept becomes a coherent pattern of ideas and interactions that can be used to describe some aspect of observed reality. As a consequence, there is little, or no, further appeal to alternative ideas and interactions in an effort to either expand, complete, or modify the concept.(19) Instead, the effort is turned inward towards fine tuning the ideas and interactions in order to improve generality and produce a more precise match of the conceptual pattern with reality. (19) Toward this end, the concept—and its internal workings—is tested and compared against observed phenomena over and over again in many different and subtle ways.(19) Such a repeated and inward-oriented effort to explain increasingly more subtle aspects of reality suggests the disturbing idea that perhaps, at some point, ambiguities, uncertainties, anomalies, or apparent inconsistencies may emerge to stifle a more general and precise match-up of concept with observed reality.(19) Why do we suspect this?

If we are charitable and creative, perhaps Boyd means something like this: "given a bad theory, additional ad hoc modifications increase the mismatch between the theory and observed reality." Though I don't think that's true either: ad hoc modifications of a bad theory don't "increase" its mismatch with observation. Rather, they stretch the theory until it does match the observations, making the theory more strained.

A much better framework for discussing matches of theory with observation than Boyd's 9th grade philosophy paper is Imre Lakatos' work on "progressive" vs. "degenerating" research programs.