The easiest mistake in building a media project is to start calling people an audience before you have earned their attention.

Once you do that, the language changes quickly. People become segments. Segments become funnels. Funnels become conversion rates. Before long, the work is no longer to say something true enough to be useful. The work is to move a number.

I understand why that happens. Numbers are legible. They make the work feel manageable. They let you believe you are building something because a chart points up and to the right.

I feel that pull myself. When you are building something new, it is tempting to ask the dashboard to tell you whether you are real yet. A few more subscribers, a better open rate, a cleaner growth curve — those things can start to feel like proof that the work matters. But they are not the same thing as trust. They are not the same thing as usefulness. And if I let them become the first measure, Signal will start serving the metric before it serves the person.

But Signal cannot start there.

Signal is not AI news. It is not a daily outrage machine. It is not a tutorial channel wearing the costume of a conversation. The work is to help intelligent non-specialists orient themselves inside a change they can already feel but may not yet have language for.

That means the audience is not a market segment. The audience is a set of lived pressures.

A school leader trying to decide what AI means for learning, fairness, and trust in a classroom. A small-business owner watching a tool do in seconds what once took an employee half an hour, then wondering what responsibility now requires of him. A hospital social worker using AI cautiously, not because she is dazzled by the technology, but because she is trying to understand what it is doing to care, loneliness, judgment, and human contact.

Those people are not abstractions. They are the test.

If Signal cannot help the school leader walk into Monday morning seeing something more clearly, it missed. If it cannot give the business owner a more honest way to think about responsibility, it missed. If it cannot respect the social worker’s ability to hold wonder and skepticism at the same time, it missed.

This matters because AI has a way of pulling every conversation toward either spectacle or sales. One side says everything is over. The other says everything is opportunity. Both can flatten the human question.

But most people are living somewhere in between. They are not asking, “What is the latest model benchmark?” They are asking what to trust, what to teach, what to protect, what to learn, what to automate, what to refuse, and who they become if more of their work, care, and attention is mediated by systems they do not fully understand.

That is where Signal belongs.

The goal is not to convince people to be excited about AI. It is not to train them into adoption. It is not to make skepticism look immature. The goal is recognition: a moment when someone says, “Yes, that is the thing I have been trying to name.”

Recognition is different from engagement. Engagement can be engineered. Recognition has to be earned.

You earn it by beginning with consequence instead of abstraction. You earn it by translating without condescension. You earn it by letting a serious question stay unresolved when false closure would be dishonest. You earn it by choosing guests, topics, clips, and essays according to whether they help real people think more clearly about real decisions.

That standard will shape the next season of Signal. It will shape who belongs in the conversation. It will shape what gets prepared, what gets edited out, what gets turned into a short clip, and what deserves a longer essay.

The public product may look like a show, a newsletter, or a series of conversations. Underneath, the operating question is simpler:

Does this help someone remain more awake, more sovereign, and more human inside the systems arriving around them?

If the answer is no, the topic may still be interesting. It may still be popular. It may still perform.

But it is not Signal.

That is the discipline I want to build around. Not content for an audience we imagine from analytics. Work for people whose questions we are trying to honor before we ask for their time.

Signal grows by recognition, not extraction.

— Renny Atkins

Get The Signal Brief
One idea worth your attention. Every week. Free.