durability of inferred knowledge
Think of DNA and reimplementation. Question: how often on average are things reimplemented in DNA? Program durability is one thing, but
existing code is capable of solving the problem AND existing code is used to solve the problem
Edit distance efficiency, tentatively
Is quite another. I'd be interested to see
good ways to measure that
good ways to improve it
how DNA performs on that scale.
Also, reading the deductive uncertainty paper would really help right about now, since it hooks up real nice.
(Above notes sparked from Samo Burja, Functional Institutions are the Exception. Civilization: Institutions, Knowledge and the Future - Samo Burja - YouTube)