It truly is. In early medschool we had a group student session on the adaptive immune system where we had to brainstorm on antibody generation and nobody in the group even came up with an idea remotely close to how it actually works. Robust and amazing. Talk about brute-force searching.
Still, at this day my hope is that we'll be able to do the body's work even better at some point using computer designed antibodies (or similar) at least for therapies. There are some notable bottlenecks in the adaptive system where antigen fragments have to be broken down and shown to potential antibodies and this is a bit different between humans etc. The human antibody molecule is also very large and it's particular design doesn't fit some epitopes you'd want to hit (not an expert on it but I remember one of the HIV proteins having an area you'd want to train on but that can't accommodate the antibody variable fragments well).
Still, at this day my hope is that we'll be able to do the body's work even better at some point using computer designed antibodies (or similar) at least for therapies. There are some notable bottlenecks in the adaptive system where antigen fragments have to be broken down and shown to potential antibodies and this is a bit different between humans etc. The human antibody molecule is also very large and it's particular design doesn't fit some epitopes you'd want to hit (not an expert on it but I remember one of the HIV proteins having an area you'd want to train on but that can't accommodate the antibody variable fragments well).