philosophy of neural pruning

The difference between math and science is baked into our brains.

Math is perfect knowledge. A mathematical statement is either true or false. But a finite brain can not cope with the incredible amount of data flowing from its environment by using such a perfect knowledge framework. Instead it uses a statistical pattern recognition framework that does not have a black-or-white view on how the world works.

The mind is pragmatic. It builds and discards models on the go without giving a shit about their ontological status. In fact, anything with a sufficiently low probability of occurrence is deemed not to exist at all. (Neural pruning is conducted using non-zero thresholds.)