Trying to build and apply models is by far the most important strategy humans can employ when trying to understand and change the world. Having a structure of thought helps us anticipate behaviors and phenomena we have never personally observed. This allows us to know without experiencing; a necessary mechanism when we only have one lifetime to live. A new model can help illuminate unexplored gaps in our understanding and provide new solutions to old problems. One model's fudge factor can be another's insight. By exploring and developing new models, we give ourselves the tools of progress. Those who construct them well are able to communicate ideas that endure; holding up their thinking on the scaffolding of the model.
I hope I have peaked your interest in the criticality and usefulness of models, so let me now disabuse you of any hope that they can deliver us to some high-minded well-thought through utopia.
Models can be wrong, or more precisely, not always right. Turning off your brain, applying a model, and forgetting everything that went into it, may work fine until it doesn't. Models have to trade off between usable and absolute correctness. Models that incorporate all factors become increasingly hard to apply, so baking in assumptions and biases to a model can do a lot to simplify its usage at the cost of increasing its risk. We humans rarely miss an opportunity to take a mental shortcut, and applying a model is a great way to do that. If we do so implicitly, our thinking can appear to have all the biases and assumptions necessary for the model's functioning. Communicating with those who don't share your models can be confusing at best and cause them to question your motives at worst.
These drawbacks are what I hope to address with this page. By sharing the models I use, I can be rigorous in explaining their assumptions and biases, and even more importantly, help communicate how I think to those that think differently. This page is a tool for communication and thinking. Other posts of this blog I will attempt to keep as static snapshots, but this page will evolve as I do.
Systems develop over the course of many years shaped largely by the same force that governs evolution: what survives stays. While what dictates survival is slightly different across natural, business, political, and cultural systems, survival is not necessarily predicated on good functioning. An example can be found on your own body. Knees are, biomechanically, a bad joint, as anyone who is an orthopedic knee surgeon or a patient of the former can tell you. They are, however, a result of millions of years of evolution, and they have thus far not stopped humans from flourishing. The same can be said about the functioning of many systems.
This contrasts with designed systems, where even a first-year biomechanics student could create a better knee. While we don't have that sort of control over our orthopedic system (yet), we do have that sort of control over the systems we make. Designed systems, however, aren't guaranteed to have the defining characteristic of evolved systems: survival. Big changes, even those that come with drastically better performance along one dimension, can cause unintended consequences or missed dimensions.
Using this model well means you are able to see evolved systems with dysfunctional or suboptimal parts, for what they are. They can both teach you a lot about what survives, while still underperforming; something to neither revere nor vilify. The key is to know that when a designed system has a high chance to survive and thrive, you can substantially change or replace the evolved system to make progress. The catch: when the environment where it lives continues to change, your designed system must evolve with it.
...Full blog coming soon...