“The only constant in life is change.”

- Heraclitus

There’s a pretty well-known article by Dan McKinley called Choose Boring Technology. It describes innovation tokens as a model of approximating an upper bound on new technologies that a company can reasonably use.

I love this article. While it’s great at describing the technical side of things, I also find the human side of things quite interesting. In this post, I’ll talk about the human side of choosing alternates.

A Thought Experiment

Let’s consider an example. Suppose we own a boba shop, and we’re considering switching from using regular cow milk in our drinks to using oat milk. I imagine most readers fall into one of two camps:

  1. Of course! That sounds great for X, Y, and Z reasons.
  2. Why would you ever do that?

Now, consider the opposite: for our boba shop, we’re considering switching from using oat milk to using regular cow milk. I imagine that the categorization above still applies.

What’s more, I often observe that those who fall into group 1 in the first example aren’t always those who fall into group 2 in the second example, and vice versa.1 That is, some people want change for the sake of change, and some people hate change for change itself.

Naively, we might believe that if choice A is better than choice B, then we should always want choice A, regardless of where we start. It would mean, therefore, that the people above are completely irrational. Are they?

Adding Friction to the Model

In a perfect world (frictionless surface, etc.), they are irrational. However, let’s consider the effect of friction on decisionmaking. Suppose we add the following constraint to the problem above:

When switching between milk choices, there is a fixed cost required to retrain employees and source ingredients.

Now, we can easily see: if we make this fixed cost high enough, it might never make sense for a company to switch choices. Conversely, if there is a negative cost associated with switching (publicity, subsidy, etc.), then it might always make sense to switch.

This extends beyond fixed costs. Friction can result from a number of factors, including irrational ones. For example, extending the analogy above, it’s possible for the boba shop to never wish to change due to the owners being maximally risk-averse.

Software Engineering and Corporate

Back to Choose Boring Technology. The article assumes that companies are perfectly rational and write from the perspective of choosing risk or not. But what about the failure modes of companies to choose new technology?

In order for a company to choose a technology and benefit from it, the friction incurred must be greater than the benefit of the old technology.

This technical friction comes in a few forms: education (needing to teach everyone a new language along with all of its idioms), inexperience (understanding the failure scenarios of the new tool), and external popularity (does an ecosystem exist?).

Then there are also the pathological, irrational causes for friction:

  • Dislike for change, or group 2 from above
  • Overstatement of risk
  • Bad experiences from the past

When choosing new technology, it’s worth considering them through the lens of alternates. If the world in which the alternate was chosen is a better world, yet the choice to switch is a poor one, consider the causes of friction. It might be the case that there are tons of pathological causes for friction, in which case the reasons not to switch are illegitimate, or it might be that the benefit gained is far too small to offset the normal causes of friction.

Below are some ideas that helped me form these ideas.

  1. We might call the group 1s the small-l liberals and the group 2s the small-c conservatives.