[SQueeeeeL]

Not OP, but it's kinda like boiling or alcohol versus poisoning (which is essentially how most drugs work). Certain attack vectors are literally unevolvable, as they would require fundamentally different building blocks to be able to be able to handle, which is too big of a leap for evolution to make.

[lucb1e]

Alcohol resistance is also a thing, though. https://duckduckgo.com/?q=alcohol+resistance+bacteria

[r3trohack3r]

This was a light bulb comment for me, thank you.

The challenge is: 1) finding something that kills mostly/only cells we don't want in or around our cells 2) kill them in such a ruthlessly efficient way that there are no survivors

As a sibling points out, the vacuum of space does #2 quite well. Actually we have a lot of stuff that handles #2 quite well. But #1, differentiating between "good" and "bad" with our murder machines, is pretty much an unsolved problem.

[midoridensha]

We need to create nanides: nano-scale machines that can differentiate between microbes and destroy the bad ones. Just don't piss them off with a phaser blast when they get into the ship's computer core.

[salawat]

Look into immunology, it's a biologically mostly solved problem with the mammalian adaptive immune system.

...In the sense we're here because an implementation evolved, however, the physics and minutiae of the problem domain are so numerous that even nature with billions of years of time, and the happy accident of intelligent tool using life haven't been able to make significant inroads on nailing down the problems space in an "ahead of the machine" sense yet.

[midoridensha]

The Covid pandemic shows that immunology is very much not a "solved problem". Any time someone dies of a disease caused by a virus, bacteria, or even cancer, this shows that immunology is very much unfinished.

[SonOfLilit]

GP means that evolution solved it to a certain extent and we human engineers are still way behind it in terms of the performance of our solutions