Possible models of modified gravity are being extensively studied now, with most phenomenological motivations coming from puzzles and tensions in cosmology with a natural desire to better fit the known and newly coming data. At the same time, available experimental evidence is limited for testing gravity as a force, beyond the regimes in which the theory of general relativity has proven to be successful. This situation leads researchers to look for "the simplest modification" to general relativity, in a certain class of models, that is enough to solve one or more problems. As a result, we are lost amidst a variety of theories with no deeper guiding principle. In this essay, we openly discuss these problems, and propose some ideas on how to go.