NOTES ON AI

Thinking in Systems

Thinking in Systems by Donella H. Meadows

1. The Basics

Elements. Interconnectedness. Purpose. These come together to create a system. Meadows presents these building blocks in the very first paragraph. Our primitives in this world of system tracking. What strikes me is the need for purpose. Interconnected elements without a purpose (e.g. sand scattered on a road), do not constitute a system. So, it seems, purpose imbues interconnected elements with something special. Who determines the purpose? What are the edges of purpose? I imagine this book will give room to test these limits.

3. Why Systems Work So Well

I jumped to this chapter, because I want to know! Well…why do they?!

Three characteristics of highly functionling systems are explored: resilience, self-organization, and hierarchy. The last two seem to be in some form of conflict to me, but I think it is because I implicitly associate self-organization with non-hierarchical structures (which I’m realizing is not necessary - as systems can self-organize to a functionaly hierachy, or process of hierarchy-ing).

Resilience

We know that resilience is the bounce back from a system that has been stressed beyond its equilibrium, but what are the mechanisms that make this possible? Why are some systems resilient and others not? “Resilience arises from a rich structure of many feedback loops that can work in different ways to restore a system even after a large pertubation”. It’s worth noting that, “systems that are constant over time can be unresilient”.

And this was one of my favorite images:

“I think of resilience as a plateau upon which the system can play, performing its normal functions in safety. A resilient system has a big plateau, a lot of space over which it can wander, with gentle, elastic walls that will bounce it back, if it comes near a dangerous edge. As a system loses its resilience, its plateau shrinks and its protective walls become lower and more rigid, until the system is operating on a knife-edge, likely to fall off in one direction or another whenever it makes a move.”

Self-Organization

“The capacity of a system to make its own structure more complex is called self-organization”. This resonates deeply with me. I look around the practice of my life and notice all the attempts to make my life more rich and filled with life. I sit in our cabin at home, a space we carved out as a sanctuary within the broader sanctuary of our home. Why does someone need a sanctuary within a sanctuary? Because I have returned from a war of the soul, and am learning to feel love, express love, and learn about the world beyond fear. The cabin is filled with books about inner and outer worlds. I layer systems of meaning the way we layer acids in a delicious meal (with aged vinegars splashing into lemon juice and pickles).

“Out of simple rules of self-organization can grow enormous, diversifying crystals of technology, physical structures, organizations, and cultures”.

Hierarchy

“Complex systems can evolve from simple systems only if there are stable intermediate forms”.

Hierarchies “reduce the amount of information that any part of the system has to keep track of”. I have felt this in my role as an engineer, where my exploration of business value emerges in conversation with product owners, but does not require me to extend the context I hold out past them (to our customers directly, for instance). Although part of me thinks that would be fun, another part of me (the part that usually wins out) knows that I just need to know enough to do my job successfully, and cultivate trust with others in my system that they have the knowledge and diversity of perspectives necessary to do their work successfully. In fact, that is typically how I see work communities that are shy of creating such hierarchical separations - members of the team fear or doubt others will arrive at the best solutions.

So many gems here. For instance, when hierarchies break down, they usually split along their subsystem boundaries. These can sometimes behave as systems unto themselves (i.e. the liver can be studied independently from the system of the body). The term suboptimization is introduced here to describe what happens when “a subsystem’s goals dominate at the expense of the total system’s goals”. The inverse poses challenges in the other direction - central control preventing effective functioning of the subsystems. This arena of Meadows’ research seems particularly pertinent to tracking the emergence of hierarchy in fast-growing startups.