Essays/Building

Risk is an experiment

A single risk is a bet that can go either way. What decides the outcome is the ability to keep experimenting until something holds.

I stopped calling them risks and started calling them experiments. A single risk is a bet, and a bet can go either way. What decides the outcome is whether I can keep going: use what the last attempt showed, run the next one, and continue until something holds.

Deciding on too little information

As a founder I often have to decide on information I do not have, or have far too little of. That is what risk means in practice: a decision made on thin information, or a picture of a future with no data to say it will happen. Some of these decisions are small. Some commit money, time or energy for a long while to a product that might work. Sometimes, as in a card game, a large part of what has been built has to be risked at once.

My rule is that any decision is better than no decision. Waiting feels careful. But while I wait, the opportunity moves, the market shifts, and the choice I was weighing expires. A wrong decision at least gives me feedback I can use.

Risk is not recklessness. Recklessness is acting without the information I had time to gather. Risk is what is left once I have gathered what the time allowed and something is still unknown, which it always is. Preparation removes the first. Nothing removes the second, and waiting for it to go away means waiting forever.

Fear makes this harder, because fear argues well. When I am afraid, my mind does not refuse to analyze the decision. It analyzes it thoroughly and returns a well-reasoned case against. I call these pseudo-decisions. The sign is that every argument points the same way. Analysis done from curiosity finds arguments on both sides. Analysis done from fear finds only reasons to stop.

So I change the question. I stop asking whether I should do it and ask how I would do it if I had already decided to. My mind then produces methods instead of objections. Often, once the methods are written down, what looked like a cliff turns out to be a staircase.

The loop has no implementation step

Once, in a conversation with a few accomplished people, someone asked what the secret of success is. The answers were the familiar ones: work hard, spend time with strong people, have a brilliant idea. The question stayed with me. Later I sat down and tried to describe what had actually happened in fourteen years of business. What came out was a loop.

  • I gather as much information as I can, from outside and from my own intuition.
  • I refine it to the situation in front of me.
  • I experiment, repeatedly, until something is stable.
  • I let the whole system evolve from there, and then I start gathering again.

There is no step called implement, and that is deliberate. Implementing would mean I know the solution is correct, and I almost never know that in advance. The aim is to run experiments that raise the chances, hoping one of them turns out to be the answer. A failed experiment is then one attempt among many I was always going to make.

The step people skip is the third. Gathering and refining both happen safely in the head, and they look like diligence. Feedback comes only from contact with the world. That is why I ship a rough first version early, as I describe in Ship the prototype, then listen. I use the loop where the horizon is weeks, months or years. For a choice that has to be made today, I rely on intuition and on what I already know.

What does not giving up mean in practice?

I went bankrupt twice in the marketing business and started again each time. What those years felt like is in There are no hero stories. My wife once asked me to define the secret of success, and I told her: never give up. I would add a condition now. Persistence spent hitting the same wall is wasted when there may be a door a few centimeters away. I do not give up on where I am going. I give up on a failing route as soon as I can see it is failing.

How much risk should a team carry?

I am sometimes asked whether risk belongs only to the person leading a company or also to the people working in it. I think it extends to them, in proportion to their power to decide. A company where nobody risks anything learns nothing new. A company of gamblers blows up. I try to build a culture in which each person takes on slightly more risk than is comfortable, thinks it through in the time available, and does that thinking from curiosity.

Back to the essays

Field Notes

A letter from Andrei Hasna on building with AI agents

Subscribe by email

By clicking “Subscribe”, I confirm that I want to get email delivered to my address and agree to Andrei Hasna’s Privacy Policy and Terms.

In each letter

  • New essays on AI agents and the systems behind them
  • What Hasna is building, and why
  • Costs, routing and results from real work
  • What did not work
  • News of the book and the podcast