PRINCIPLES TO UNDERSTAND OUR MIND:
1) Our mind is an ensemble of parallel, bottom-up processes. It cannot be described as a single top-down process, even if this is how we perceive it.
2) Our mind is not made to perceive correctly, but to act correctly.
1) Our mind is an ensemble of parallel, bottom-up processes. It cannot be described as a single top-down process, even if this is how we perceive it.
2) Our mind is not made to perceive correctly, but to act correctly.
(Corollary to #2: perception biases are features, not bugs.)
3) Perception is proactive. We choose what to sense in order to facilitate action.
(Corollary to #3: one can’t build an *effective* mind whose only job is to perceive. This is a limit of many current AI approaches.)
3) Perception is proactive. We choose what to sense in order to facilitate action.
(Corollary to #3: one can’t build an *effective* mind whose only job is to perceive. This is a limit of many current AI approaches.)
4) Our brain is specialized, not modular, just like our cells are specialized *in their outputs* but their inputs are general (they get info about the general state of the body) and could not live as separate modules.
(Corollary to #4: current AI approaches will fail if they’re kept as modular. Our brain works because most of it receives feedback from most of our actions and reactions; brain modules do not talk *exclusively* through limited bandwidth APIs).
5) Our conscious thoughts are a “simulator app” used by our brain to run bottom-up simulations to predict results it cannot infer with experience-based *direct* inferences alone (aka intuition).
6) As much as an app cannot fully understand the machine it is mounted on, not the processes that called for its execution, so our conscious mind cannot fully understand our unconscious one.
7) Admissibility is our mind’s proxy for truth.
(Admissibility of simulations is our conscious mind’s proxy for their truthfulness.)
8) The admissibility of stories is (also) measured by how well they match our internal emotional state.
(Admissibility of simulations is our conscious mind’s proxy for their truthfulness.)
8) The admissibility of stories is (also) measured by how well they match our internal emotional state.
(#8 explains how we make sense of dreams, why BS stories stick, the adoption curve and many of the emergent behaviors of groups of humans; however irrational it might seem, it makes evolutionary sense. Perhaps one day I’ll write a post about it.)
9) Stories are justifications to explain the mismatch between the 2+ events we observe and our simulation of it; by their property of increasing the admissibility of two observations appearing together, they reduce surprise and improve acceptance of what we perceive.
10) The ability of our brain to formulate stories to make sense of what we observe is a critical element of learning. Those who are, on average, less likely to make stories out of incoherent observations are better at detailed (*) fields and impaired at contextual (*) fields
(*): this paper explains the differences between detailed and contextual fields psyarxiv.com/xm5ca
As example: computer science and physics are detailed, social conversations are contextual.
(of course, we’re talking about a spectrum and averages here)
As example: computer science and physics are detailed, social conversations are contextual.
(of course, we’re talking about a spectrum and averages here)
11) TL;DR for the paper: the more we can make sense of noisy data with stories, the better we are fields where incoherencies are noise (such as human interactions) and the worse we are at fields were incoherencies are signal (such as STEM).
12) The inability to ignore noise is one of the factors contributing to the emergence of behaviors that often we classify as autistic (as I explain on autosmclarified.com), whereas…
13) the ability to make sense of locally incoherent data is one of the factors helping in making sense of noisy human behavior (an ability which we sometimes classify as neurotipicity, even when in some individuals it is too strong and slows the mastery of non-noisy fields.
14) In my 2nd book I demonstrated how the learning rate of a person in a given field is proportional with how closely his mind’s sensitivity to noise matches that field’s. A highly underestimated teaching technique is removing noise from a contextual field as written below.
15) Why are contextual fields noisy, and why is noise-tolerance required to become proficient at them?
Explanation below
(read “people on the Spectrum” as “people who are more likely to notice noise as incoherent”)
Explanation below
(read “people on the Spectrum” as “people who are more likely to notice noise as incoherent”)
16) As much as heliocentrism greatly simplified models and allowed us to get a more intuitive understanding of how things actually work, managing to make your mind work at the right resolution and to properly differentiate signal from noise is the key to facilitate learning.
