Here is why you use simple models.
The blue crosses are the data. The red line is the line of best fit. The black line is a polynomial of degree 50 of best fit. High dimensional models have a tendency to fit the data by wiggling wildly.
That problem would be handled by cross-validation; the OP is saying that a simple model doesn’t have an obvious advantage assuming that both validate.
Given that both models validate, the main reason to prefer a simpler model is the sort of thing in Gears vs Behavior: the simpler model is more likely to contain physically-realistic internal structure, to generalize beyond the testing/training sets, to handle distribution shifts, etc.
It depends on what cross validation you are using. I would expect complex models to rarely cross validate.