Model Building (not ml)

“What is truth? How do you seek truth?”

These two questions represent the theme of questions that plague my mind throughout my day as I take on roles in my day to day life. As an engineer, I often ask “how do I design this system be reliable and scalable? why does this problem re-occur?”. As a foodie, “where can I find tasty food?”. As a human, “how can I live a good life? how can I be a good son/friend/co-worker?”.

To me, it seems that truth can be both objective and subjective. The truth is often objective in the sciences and engineering, and subjective in the humanities. However, engineering does have subjective truths as well. Engineering systems are tools built by humans, and every engineer has their subjective experience building and maintaining those systems.

I don’t think we can actually obtain truth. Similar to Plato’s Allegory of the cave, we can only see the effects of our action, not the cause. At best, we can build models of what we can see. These models are useful in making decisions. Without them, we would get decision paralysis. We just need to careful that these models are not the truth and not be suprised that they can fail.

I have two processes for model building:

1. Taking other people’s models.
2. Build your own: Observations -> Model -> Test if Model Aligns with Future Observations

Each model building process has their own pros and cons.

1. Taking other people’s models. Pros: Faster and easier to take models. Cons: Other people’s models may not work for you. May be biased.
2. Build your own: Observations -> Model -> Test if Model Aligns with Future Observations Pros: Tailored for your unique situation and personality. Cons: Built on only one person’s experience. Takes a longer time than taking someone else’s model.

I tend to incoporate both processess in building my own models. When it comes to more objective truths, I usually take other peoples’ models and refine them based on my own experience. With more subjective truths, I build my own model but take into account other people’s model.

An important part of model building is getting more observations in quantity, quality, and diversity. While I use the word “observation”, you need to also act to create observations. Your model drives how you act and actions creates observations. So, depending on you act, you will only get a subset of all possible observations in the world. For example, if you repeatedly run into a wall, you might conlude that it is impossible move beyond that wall. However, if you climb, you may discover that you can scale and get past that wall.

Be cognizant about what models you have and how have built them. You may discover new things by trying out other people’s models. Find different ways to build and live. It can be miserable yet fun and revolting yet tasty. The underlying truth is unknown and constantly shifting, so be careful about building keeping your models static.