Every founder tells you their product works. I have sat on the other side of that sentence for most of my career, as the person whose job was to assume it might not be true. Fifteen years securing large organizations trains a particular reflex: when someone makes a claim, you do not argue with it and you do not accept it. You ask how you would check it.
That reflex followed me into founding Setix, and it changed how I think a company should present itself. The default posture of an early startup is persuasion. You build a deck, you polish a story, you rehearse the demo, and you ask people to believe you. It works often enough that everyone does it. But persuasion has a ceiling. The moment your audience is sophisticated, the story stops being an asset and starts being a variable they have to discount.
Claims versus checkable claims
There is a difference between a claim and a checkable claim, and the difference is worth more than the claim itself.
"Our system is fast" is a claim. "Here is the endpoint, time it yourself" is a checkable claim. "Customers love us" is a claim. "Here are three customers, call them" is a checkable claim. "It works" is a claim. "It is running right now, go break it" is a checkable claim.
The content can be identical. What changes is where the burden sits. A plain claim asks the listener to do the work of believing. A checkable claim does the work up front and invites inspection. In security we spent decades learning this lesson at great cost. Vendors asserted their products were safe, buyers believed them, and the gap between assertion and reality became the industry's core problem. The systems that ultimately earned trust were not the best argued. They were the most inspectable.
Live systems beat pitch decks
A pitch deck is a set of claims arranged for maximum belief. A live system is a set of facts arranged for maximum inspection. When I have a choice about which to lead with, I lead with the live system, even when it is smaller and less flattering than the deck.
This is not modesty. It is strategy. A demo environment someone can actually use answers questions no slide can answer, including the questions I would not think to address. It also fails in front of people sometimes, which is uncomfortable and clarifying. A system that fails visibly and gets fixed is more credible than a system that has never been seen running. Anyone who has operated production infrastructure knows the difference between software that is claimed to work and software that is observed to work, and they know which one they would bet on.
The principle underneath
My guiding principle, in security and now in company building, is verify, don't trust. Most people read that as reflexive distrust. It is the opposite. Trust that rests on verification is durable. Trust that rests on personality, reputation, or narrative evaporates the first time it is tested, and everything I am building is for a world where it will be tested constantly.
Setix exists because machine commerce cannot run on belief at all. An AI agent cannot be charmed by a pitch. It cannot read sincerity in a founder's voice. When agents discover one another, agree on outcomes, and settle, every step has to be checkable by a party with no capacity for faith. Designing for that audience is a useful discipline, because it removes the option of persuasion entirely. Either the fact is verifiable or it does not exist.
I try to hold the human side of the company to the same standard the machine side requires. When I describe what we are building, I want each sentence to be the kind a skeptic could test rather than the kind a believer must accept. I do not always succeed. Storytelling is seductive, and founders are rewarded for it in the short term. But the long term belongs to the companies whose claims survive checking.
So this is the standing offer, and it is the only marketing position I am comfortable with. Do not take my word for it. Look at what runs. Ask the questions a professional skeptic would ask. If what we have built cannot survive that inspection, no story should have saved it. If it can, no story was needed.