That's quite a stretch. First, I never committed to agreeing with Keen on everything. I watched an interview with him in which he defends the use of fancy math in economics, even as he criticizes the neoclassical use of math.

But second, the problem has never been the use of math as such in the modeling of economies. The problem is with utopian assumptions that have the appearance of being propagandistic or apologetic rather than scientific. And Keen's models are heterodox precisely to the extent that they're more realistic and even pessimistic about capitalism, so they don't fit so well with the neoliberal narrative.

Third, all models simplify, and the test of whether the simplifications are useful is if they help predict behaviour. Heterodox models predict better how capitalist systems go wrong. Oh, I'm sure "neoclassical" models can likewise account for those downsides, but only by adding a million concessions that have rewritten the original, utopian inspiration of the model.