Notes on systems,
technology, human behavior
and building.

A broken experience rarely belongs to one person. Ask a family planning a celebration, a team struggling to coordinate, or a customer facing a dozen contradictory instructions. People are often responding rationally to an irrational arrangement of tools, incentives and information.

The question that interests me is not merely who made the mistake. It is what the system made easy, what it made difficult, and why the same failure keeps returning.

Better outcomes sometimes begin not with demanding better behavior, but with redesigning the conditions around a decision.

TAKE THE QUESTION WITH YOUWhat behavior is your current system quietly rewarding?

A marketplace can say it values trust while rewarding only speed and volume. A company can celebrate customer service while measuring only the number of tickets closed. A person can say they value learning while arranging every day around public approval.

I try to separate stated intention from measurable reward. Follow the reward long enough and a system begins to reveal its real operating instructions.

That is why I look at incentives before slogans, and consequences before applause.

TAKE THE QUESTION WITH YOUIf every participant followed the reward, what would break first?

I am interested in the moment a belief stops being something we have examined and becomes something we merely protect. A familiar belief can feel like identity; challenging it can feel like betrayal.

But disagreeing automatically is simply another form of obedience. A more demanding freedom is to inspect evidence, understand context, listen to expertise and accept responsibility for a conclusion, even when the conclusion is uncomfortable.

That is the distinction I want Obey Nothing to explore: not rebellion as performance, but judgment as a personal responsibility.

TAKE THE QUESTION WITH YOUWhich belief would you still hold if nobody rewarded you for holding it?

Entrepreneurship attracts certainty. The pitch sounds complete, the diagrams connect beautifully, and the future looks inevitable on a slide. But an elegant explanation is not customer behavior, and a feature list is not a business.

I want the freedom to hold an ambitious thesis and still let a real conversation, a failed assumption or an inconvenient constraint change the implementation. MatroMet is a practical place where I keep examining that distinction.

The goal is neither endless planning nor reckless shipping. It is a cycle of observation, small tests, correction and accountable building.

TAKE THE QUESTION WITH YOUWhat evidence would persuade you to change the product, not just the pitch?

Every notification makes a tiny request for priority. Every feed arranges the next thought before the previous one has settled. Repetition can make the familiar feel important, even when it has little to do with the life we meant to build.

I examine where my attention goes, who benefits from keeping it there, and whether the information I consume changes any decision I can actually make.

Protecting attention is not refusing technology. It is choosing when technology deserves access to the steering wheel.

TAKE THE QUESTION WITH YOUWhat is receiving your attention without earning your intention?

A system that produces an answer in seconds can make uncertainty feel resolved. But speed does not establish truth, consent, fairness or context. Those remain questions of evidence and judgment.

I am interested in AI as an amplifier of useful work: finding patterns, reducing repetitive effort and surfacing overlooked possibilities. I am equally interested in where it must stop and ask a human to decide.

Automation should earn trust through verifiable outcomes, not through fluency alone.

TAKE THE QUESTION WITH YOUWhere should a human remain accountable even if software can act faster?

Space, time and the scale of the universe remind me how little intuition alone can settle. That is an analogy for how I approach difficult questions, not a claim of scientific discovery or expertise.

When something is unfamiliar, I want to know what we have observed, what model explains it, what that model cannot explain, and what new evidence would change the conclusion.

The most useful question may not be whether I sound convinced. It may be whether I have made it possible to be corrected.

TAKE THE QUESTION WITH YOUWhat observation would change your mind?

Some choices look efficient because their costs arrive later. A hurried shortcut can save a day and create years of maintenance. A system that produces quick growth can also make future trust harder to earn. I am interested in the distance between an immediate win and the consequence that follows it.

Thinking in decades is not a reason to postpone evidence. It is a reason to make today's experiments honest, reversible where possible, and useful even if the larger ambition takes longer than expected. In building MatroMet, that distinction matters to me: ambition sets a direction, while what people actually need must shape the next decision.

Time does not automatically improve an idea. Feedback, discipline and repeated correction might. I want to build the kind of judgment that can survive both excitement and delay.

TAKE THE QUESTION WITH YOUWhich decision would you make differently if you had to live with its consequences for ten years?