This is completely wrong, counter to all contemporary scientific theories on the subject. Exposure therapy works exceptionally well and we know how it works: through desensitization. We can measure this in multiple different ways including blood-serum levels of various substances, fMRI, HRV, etc. The physical evidence for that and how it works is staggering.

A recent experience I had came while walking in a forest with friends about an hour from Seattle. While walking, we heard an animal sound of some kind that cause an immediate "hairs on the back of my neck standing" up reaction along with what felt like an adrenaline rush (time slowing, narrowing of focus, etc). All of this happened immediately, and it took a few moments for my conscious mind to catch up. My friend and I looked at each other and walked briskly away from the sound. It wasn't until we encountered another hiker that we found out a mother bear and cubs had been spotted in that area.

This was unlike any other fear of animals, jump scares, or even near-miss types of experiences I've had in the past. It was a very primal and immediate reaction to a sound I've never heard before.

> Fear is fundamentally a hardware attribute and not a software issue (ie not trainable).

The fun thing is that this hardware we have can be rewired so it could be trainable in many instances. Maybe in all of them if we only knew how.

> not trainable

Counterexample: crowd-control horses.

[EDIT: we can quibble about terminology; I think it's fair to say that while apprehension may always remain, fear, and certainly panic, can be trained away]