A high mutation rate and an aggressive crossover policy for splicing in the genetic algorithm is optimal policy for when you know you have two highest examples of human life, vs you knowing you have two lowest examples of human life and the nanotech says you're barely clinging to survival.
Said another way, tuning hyperparameters of the genetic algorithm to be context-sensitive to whether you have a dimepiece babymaker or a 2 of 10 babymaker in the genetic fitness department.
The Genetic algorithm always need mutation rate and crossover to be as agressive as possible sinc what we have is known not-good and so we need to get something else to be pretty-good faster.
Said another way, exploration and exploitation is not something to be decided on intelligently based on how much you know about the payout structure of the arrangement currently had.
https://youtu.be/mo96Nqlo1L8?t=99Mutation = maximum and inflicting breakages on crossover = maximum. This is always ideal.
