While I'm waiting to hear back from Trezor from my postal-destroyed model T, I thought I'd read up on a few things on how it works – specifically how recovery seeds work. I read through this article on the wiki about how it works and I think I have a decent understanding about it. Note I'm in no way trying to poke holes or be critical about the process, just trying to figure out how it works and practical expectations.
Since a recovery seed can be used to restore a lost Trezor, this means every address can be derived from the recovery seed correct? In other words, while the seed may be random, once the seed is derived, creating new private keys is deterministic in some way using some defined algorithm?
Is there a way to circumvent this or what sort of limitations may I expect in regards to generating various addresses across various cryptocoins?
Can I use my own entropy in addition to the internal RNG to create the seed (thinking about things like rolling dice and other such randomness)?
Answered via this excellent comment.