I have no objection to model-making, and I agree with your perspective on the potential benefits of creativity in that regard. So I have no objection to the economist's act of modelling economies.

The problem is that modelling an aspect of society isn't like modelling an aspect of nature, whereas the neoclassical tradition elides and obfuscates the differences. Hence the distinction between the hard and the social sciences. Economics is a social science that presents itself as a hard science.

The question is whether models of society can be decisively tested, such that critics are silenced and the evidence speaks for itself. If the tests are always debatable because of ambiguities in the evidence and the politicization of the field, we'd only be dressing up as scientists and our talk of "testing models" would be vain. The models would just be formalizations of intuitions and would be subject to GIGO, and the interpretation of the data would be subject to the shenanigans of statistics.