There's another argument (which I think your parent is referring to) which says that the thing that's doing the modeling is the very thing that's being modeled, so it's impossible to separate the data from the model. The radical conclusion people have come up with is that therefore there is no difference between the model and the data. What we perceive to be the model is actually the data, that there is nothing outside of the data because there can't be.
I haven't been able to decide if this is profound or incoherent.