I just did some measurements and am impressed on both fronts: DNS recursive resolution is faster than I anticipated, but also page load times for well optimised sites are also very fast (sub 0.5s). Here's some data:
So it's quite reasonable to do full recursive resolution, and you'll still benefit from caching after the first time it's loaded. One other idea I had but never looked into it was instead of throwing out entries after TTL expiry to just refresh it and keep it cached, no idea if BIND/Unbound can do that but you can probably build something with https://github.com/hickory-dns/hickory-dns to achieve that.
I think you mean "running your own recursive resolver", an authoritative server is one which is authoritative for some zone (e.g. example.net), whilst a recursive resolver is one that goes and walks from the root of the DNS hierarchy to the leaf that you have queried.
It is probably quite a bit slower though needing to have roundtrips at each stage of the resolution, which is also likely a reason that these public resolvers get so much use (latency improvement via caching).
I have my personal LIR with RIPE and they do allow signing up as an LIR without being a company. As an individual you need to provide proof of identity via government issued ID and pay 17% Dutch VAT.
When I signed up I got a /22 of V4 space and /29 of v6 space, although for new signups I believe there's now a wait list to get a /24 and that's it for V4.
Totally agree that the iBGP vs eBGP thing is very confusing, even in some vendor documentation there are some assumptions on what an internal vs external session should do instead of leaving it up to the user to decide how to operate. I get that this is probably for simplicity but it would be nice to give the user the full flexibility of the protocol.