Perfect Predictor

The setup

The machine doesn't control you. It doesn't send signals to your brain or constrain your body. It simply watches, processes, and predicts. And it is never wrong. It told your parents what career you would choose. It told your government how you would vote. It is telling someone right now what you will order for dinner.

You feel like you're deliberating. The options seem genuinely open. But the machine already has the answer written down.

What this does to free will

The common-sense view of free will is that it requires real alternatives: you could have done otherwise. The Perfect Predictor puts pressure on this idea from a direction determinism alone doesn't.

Even if the universe were only probabilistic, the machine's perfect track record suggests your choices aren't the source of genuine novelty. They're outputs, not origins. Hard determinists say this is fine, because what we call "free will" is just acting from your own desires and reasoning, which the machine doesn't interrupt. Libertarian free will advocates say the machine's very existence proves freedom is an illusion.

The uncomfortable middle: maybe the machine is tracking something real about you, your character, your values, your reasoning patterns. Maybe that's not a problem for freedom. Maybe that's what freedom is.

How it sharpens Newcomb's Problem

In Newcomb's Problem, the predictor is merely very accurate. You might argue that the small chance of error is enough to justify two-boxing (taking both available boxes), since the predictor could be wrong this time.

The Perfect Predictor closes that exit. It has never been wrong. Now the question of whether to one-box or two-box becomes a question about whether you think your deliberation is real. If you believe your choice is genuinely open, two-boxing seems rational: you take both boxes and get more money. But if the machine is truly perfect, the contents of the boxes are already determined by what you'll decide. One-boxing and finding $1 million is the only outcome a one-boxer has ever gotten.

Your theory of free will determines your answer. That's the trap Nozick set.

Compatibilism's response

Compatibilists argue that predictability and freedom are not in conflict. What matters is whether your action flows from your own reasoning and desires, not whether someone else could have anticipated it. A chess computer can predict your next move if you're a predictable player. That doesn't make your move unfree.

On this view, the Perfect Predictor is just a very good model of you. It predicts you accurately because it understands your values and reasoning patterns. And those patterns are yours. The machine doesn't undermine your agency; it reflects it.

The harder question is whether this answer satisfies. When you learn that the machine already knows what you'll choose, does the deliberation still feel meaningful? The compatibilist says yes. Many people, on reflection, aren't sure.