Pit-traps for rationalists: the seductive appeal of seemingly-universal frameworks that hardly ever work in practice.
You can waste your entire career on one. You can waste an entire academic discipline and billions of dollars of research funding on one.
ribbonfarm.com/2019/01/10/rem
Conversation
Replying to
Feel strange as i call myself a system thinker *because* i believed that a grand model is not possible and that models are at best local (in all dimensions).
That said, i may just live in different circles than the author ;)
1
3
Replying to
Yes I think “systems thinking” is used two mean two almost opposite things. Rooted in disagreements among the founders of cybernetics in the 1950s I think. Cc
2
2
The two opposite things being?
1. A grand model to capture all models?
2. No grand model. Just local models?
1. Epistemic more important?
2. Instrumental more important?
1
Maybe something like “we can understand and control complexity if we just use better math” vs “the world is so complex and indefinite that it can be understood only partially, and only influenced, not controlled, so we’ll get better outcomes if we accept that”?
1
You’re getting at a more important distinction obliquely: how much a controller needs to “know” about thing being controlled. Two seminal results. The internal model principle in control theory (engineering) and Ashby principle of requisite variety, both state basically that.
1
1
sciencedirect.com/science/articl
and
en.m.wikipedia.org/wiki/Variety_(
(there are a few other formalizations)
3
2
Tryin simplify & learn. Is the distinction obliquely referred by between
1. I can have a perfect model or combo of models of the thing I want to control
Vs
2. It’s not possible to have that perfect model.
/1
1
And that both principles you cite show the extent of how much info / feedback needed for such a model to work?
/2
1
Btw I find the Ashby easier to understand than the internal model. And I base my current understanding of the internal model by its wiki article.
Cause cannot grok it from the original paper.
Happy to get feedback where I might be mistaken
/3
1
That’s good enough if you’re not doing anything technical



