Surfing entropy
Entropy is often described as a measure of disorder in a closed system. What has always fascinated me is not the definition itself, but the implication: left alone, things do not stay stable. They degrade. They fragment. They become harder to understand. When entropy is at maximum, no useful work can be done.
Over time, this concept became a useful shorthand for how work feels when it starts slipping out of control and became the single variable worth watching. The key consideration if the goal is to build teams that function over time.
Physics aside, teams and organizations behave very much like closed systems. External input exists, but most interactions happen internally. Even when they are working well, leaving them unattended leads to deterioration. Blockages form. Coordination costs increase. Efficiency slowly collapses into friction.
The reason is not a lack of talent, process, or tooling. It’s questions.
Modern tech work constantly generates them. Priorities. Ownership. Interfaces. Trade-offs. Technology choices. These do not merely change; they demand answers, repeatedly, under shrinking time horizons. What was previously stable for years now requires quarterly review, sometimes more frequently.
Each unresolved question requires that someone cease work to seek an answer. “Who should own this?” “What if this won’t scale?” “Should we purchase this service or not?” In most cases, the person seeking an answer fears being wrong, so the answer remains ambiguous.
The reality is that most of these questions are unanswerable, or, better, have at least 2 satisfying answers that will yield good results. And this makes it so much harder for people to commit, find closure, and clearly take a path.
Left unattended, these questions do not disappear. They diffuse. They appear in meetings, in code reviews, in Slack threads, and in architectural hesitations. They turn execution into interpretation. Everyone pays a little, constantly, and everything slows down.
This is where entropy accelerates.
Over time, this is how I started to understand my role as a leader. Not as the person with the best answers, but as the one responsible for ensuring that most questions irrelevant to effective contribution are already answered.
That responsibility comes with risk. Decisions are taken with incomplete information. Some will be wrong. But leaving them undecided is rarely neutral. Undecided questions tend to spread, not wait.
What can appear to be a desire to decide everything is often the opposite. It is an attempt to ensure that nothing remains unresolved long enough to become pure entropy, quietly spreading across the system.
Leadership, at least in this framing, is less about control and more about containment. Staying close to the wave of undecidable problems. Taking accountability for answers that allow others to focus on the problems they were hired to solve.
Not eliminating entropy. Surfing it long enough for others to move forward.
December 29, 2025