Been thinking about: how all the algorithms we use to train humans are directly reapplicable to the training of AIs.
First order worker: does things
O2 worker: trains O1 workers (O2s are trainers)
O3 worker: trains O2 workers (O3s are TTT-capable..
training the trainers)
A meat shop is only scalable after there are enough O3s to maintain business continuity (and a positive growth rate, later) versus the current attrition rate.
You should always have truck numbers for your O[1,2,3] populations.
#operations #risk #process #people