Rodney Zadeh

In 1864, Dostoevsky gave voice to a man who could not stand an idea fashionable in his time: that once science uncovered the laws behind human desire, everything a person did could be worked out in advance. Every action calculated and entered in a table, like logarithms. A life explained rather than lived as an open question.

The underground man's objection was not that the calculations might be wrong. It was that they might be right. If everything a person will do can be calculated ahead of time, then even an irrational choice becomes a kind of declaration of independence, a way of proving he is not a piano key being played by forces outside himself. He puts it mischievously. Twice two makes four is an excellent thing, but twice two makes five can be charming too.

The underground man does not defeat the table. He refuses to belong to it.

More than a century later, Joseph Weizenbaum arrived at a related problem from inside the machine. At MIT he built ELIZA, an early program that could imitate certain kinds of conversation, and he knew exactly how limited it was. The people using it often behaved as though it were not limited at all. They confided in it. His own secretary, who knew he had written the program, asked him to leave the room so she could continue her conversation with it in private. What unsettled him was not that the machine had secretly become intelligent. It was how readily people supplied the intelligence themselves. In 1976 he published a book arguing for a distinction that has only grown more important: what a machine is capable of doing is not the same question as what we should allow it to decide.

Neither man was making a prediction. Both were describing something already visible in their own time. One imagined a man refusing to be reduced to a machine. The other watched people treat a machine as though it were a person. What interests me is not that they saw the future. It is that the problem was visible this early, this clearly, and the building continued anyway. Faster each decade, and it is continuing still.

Taking things apart is my default. Objects, people, systems, claims, motives, arrangements, explanations. I want to know what something depends on, what governs it, what is producing the behavior, and what remains when one of its assumptions is removed. Look carefully. Then ask what your way of looking has already decided for you.

That habit was installed early. My family has been moved by history more than once. Governments changed, and ordinary people carried decisions they had no part in making.

Growing up inside that, you learn quickly that the official version and the actual one are rarely the same. Whatever a system says it is doing, the thing actually driving it is usually somewhere else. I have never had the luxury of assuming otherwise, and I have been looking at that gap ever since.

For a long time I had those questions without the language for them. Most of what I know I taught myself, by following questions rather than a curriculum. One book led to another, one subject exposed a gap underneath it, and I kept going backward to learn what the rest was standing on. That longer story lives on its own page. What matters here is where it led: to a group of writers who had spent entire careers on the questions I had been circling.

Read together, they pick up where Dostoevsky and Weizenbaum left off. Jacques Ellul argued that the real force in modern life was not any single machine but what he called technique, the drive to find the most efficient way of doing everything, until efficiency becomes the standard every other value has to answer to. That is Dostoevsky's table, no longer a thought experiment but a way of running a society. Ivan Illich asked what happens to the people inside it. He described a threshold where a tool stops extending what people can do and starts requiring them to organize their lives around it. Langdon Winner moved the question into the objects themselves, asking whether artifacts have politics, whether a design can carry an arrangement of power regardless of who uses it or why. Neil Postman described the last stage, when a culture stops merely using a technology and starts letting it set the terms by which everything else is judged. They disagree with each other in places, which is worth knowing before anyone tells you the question is simple.

Melvin Kranzberg put the part they share in a single line. Technology is neither good nor bad, nor is it neutral. A thing can be built with no bad intent behind it and still arrive carrying an arrangement that decides something for you.

If a design can carry an arrangement of power, then who builds these systems matters as much as what the systems can do. Money moves things. Influence moves things. Incentives move things. None of that is new. What is new is the reach. We are building systems that increasingly sit between people and the decisions they make, and they are not being built outside human nature. They are built by people, inside institutions, under pressure from the same familiar forces. That is why integrity in these systems interests me. Not as decoration. As structure.

The deeper I have gone into tech, the less abstract Winner's question has become. Code occupies a strange territory between thought and force. An idea becomes a rule. A preference becomes a default. An assumption can be embedded in a system and repeated millions of times without ever having to present itself as an assumption again. An assumption in the mind is still only an idea. An assumption inside a system becomes a condition, and other people have to live inside it.

That distinction is most of what I write about here. Technology, people, interfaces, incentives, business models, and the assumptions buried inside them. I return to the same questions. What is this actually doing? What must its makers believe about the person using it? What has been made easy, and what has been made difficult? What behavior is being rewarded? And when something is free, what is the person actually paying with? Usually it is their data, their attention, and a record of their behavior that gets sold or used to shape what they do next.

I have little interest in performing certainty. Certainty can be the result of serious inquiry. It can also be what appears when inquiry stops too soon. Some questions resolve. Others become more precise. Precision is also progress.

None of which means the trajectory is doomed. The same systems capable of distorting behavior can raise human capability at a scale that would have been hard to imagine not long ago. The question is what gets built into them. A system can work and still not hold. Holding is a property somebody has to build in on purpose. I think integrity will prove to be more than a moral preference. It will be a design advantage. Build something people can understand. Test what it actually does. Know where it fails, and say so.

What gets built into these systems is being decided right now, by a fairly small number of people. I want a hand in it.

Which is also why I think the creative part matters more than it usually gets credit for. Building good systems is not only an engineering problem. It takes people who can imagine what is not there yet, who notice what everyone else has stopped noticing, who are willing to ask whether an arrangement everyone accepts was ever necessary in the first place. Criticism can tell you what is wrong. Construction asks the harder question: what else is possible? Sometimes what looks inevitable is merely familiar. Sometimes what looks like a constraint is only a convention nobody has bothered to challenge.

So I build as well as write. Software, hardware, experiments, prototypes that work and prototypes that do not. I also have ideas I intend to build, not corrections to what already exists but arrangements I suspect are possible and have not yet seen anyone attempt. Some will turn out to be wrong in ways I cannot see from here. That is the price of finding out, and it seems reasonable. A principle that survives construction has earned the right to be taken more seriously. One that does not has still taught me something.

The tools changed. The question underneath them never did. What is actually happening here, and what happens if we build it differently?

What I want to build now are things that benefit people without taking more from them than they know they are giving, and that open ground they never had access to before.

This site is where I document the thinking and the products.