what is the point of it all?
deciding what to do. in techinical terms: a policy, a function from states to actions.
reinforcement learning is how i think of it.
the main loop in RL is between perception and action, observation and decision.
observations split into sensory (basically raw observations of state), and intuition. intuition is maps between these internal sensory observations.
on the decision side of an agent, there’s model based and model free.
model based has a signature like (action, state) to state. basically, what happens next? this is the function of Thinking. the failure mode of it is that the states don’t actually matter. for an extreme example, think of ppl who comment about spelling on reddit. correct, but useless.
the model free side has the type signature: (action, state) to real number (aka ‘how good is the state/action’). the failure mode of it is delusion: like Chad in ‘Tucker and dale vs evil’, u just want to be right so much you force it and essentially ignore the dynamics of what’s going on.
in personality terms, ppl are observers xor deciders. actually, just like in RL, everyone is a decider in the end. the point of RL is determining a policy. model based/free are the 2 big approaches to it, but everyone just wants to figure out what to do in the end.
observers tend to go from Thing issue to Thing issue. the cable broke so i couldn’t plug in the computer so i got stuck.
deciders are always on about people, like when he was mean to me or she was nice to him or yadda yadda.
For an example of a decider, Heather Dale the singer. She does the Huron carol engaging deaf singers and singing in 4 languages is nothing to her. but she’s weird about the intros of everyone and why she’s doing it and …
a lot of comedians are observers, especially Se (external sensing). Take Kevin Hart, an Se-Fe. Denzel Washington too (I saw Man on Fire last night so I’m shoehorning in Denzel).
The basic function flow is S->N -> T -> F. Just like in RL, the model based methods are for the benefit of the model free, and like in RL, mixing approaches is best.
a big thing about personality is swings. ppl have natural preferences for thinking/feeling and sensing/intuition. maybe this is the easiest way to decide if you think personality theory is real: are thinkers and feelers real, different kinds of people? what you believe about that will color your view.
Each function has an internal (“introverted”) and external variant. Internal variants tend to be convergent (Ni) and depth first (Ti) vs the breadthwise nature of external functions like Se and Te.
I am still a bit fuzzy about exact type signatures for Sensing and Intuition. Intuition has a function like aspect in that senses/impressions should be pluggable as input and mental impressions should come out. Maybe a map from Sense to Sense (Sense -> Sense)?