how do we sort a school yard of hundred students based on their height?
we can simply tell them to check theirs with the person in front of them and if they were reverse, change their positions. give them time for enough iterations and soon they will be sorted. it might not be the most optimal solution but with applying a simple small principle to the system, some big results will happen.
in the same way we created LLMs via RL. we just told them to do random things and if that thing resulted in a specific result, reward it with doing more random things in that direction. again we didn't need to teach them every single concept from math, literature, jokes, image understanding, etc. we just gave them enough energy for enough epocs of training.
i'm thinking now of applying this to my life. and specifically at the moment: the micro principle of exchanging
maybe sth similar to what ryan trahan did. he just walked down the streets and exchanged different things with each other. a penny with a pencil and a pencil with a 1 dollar. he continued his micro principle and ended up traveling the whole country.
similarly i just need to constantly exchange one thing with a another (better???) thing. night time with sleep, sleep with morning time, morning time with running in park, running in park with hunger, hunger with food, food with dopamine, dopamine with code, code with money, money with time, and so on...
i just need to make good micro decisions. maybe after enough iterations, they end up big creations.
another one is about ai destroying jobs. many people will lose their jobs and that's inevitable. another microprinciple can be to just gain a skill others don't have. know more languages than others, gain better social skills than others, have a healthier body than others and so on ... in the long term if anybody is going to be replaced you would be the last one.