Intelligence We Now Build

There is a final reason this argument — which would matter in any era — has become impossible to defer in ours.

For most of history, the tools a civilization built were slow. They extended human reach, but they did not perceive, decide, or coordinate at scale on their own. That has changed. We are now building intelligence itself: systems that model, predict, persuade, and act faster than any deliberative human process can follow.

And here the thesis becomes its own proof. Technology amplifies the state it is given. Artificial intelligence does not resolve incoherence; it scales it. Placed inside dysregulated systems optimized for extraction and engagement, it multiplies exactly those dynamics — faster persuasion, finer prediction, more efficient capture of attention and behavior. Placed inside coherent systems, the same capabilities could extend understanding, surface what fragments us, and coordinate care at scales human attention alone could never reach.

The technology does not choose. The conditions surrounding it choose.

This is why the coherence crisis and the technological crisis are not two crises. Artificial intelligence is not arriving from outside civilization; it is arriving as civilization — trained on our outputs, our pathologies, our wisdom, and our fragmentation alike. The real alignment problem may therefore be larger than aligning machines. It may be civilizational: whether humanity can become coherent enough to wield the intelligence now entering its hands.

This argument mattered before artificial intelligence. After it, the argument becomes existential. The question is no longer only whether we can build intelligence. It is whether we can wield it.

That question — how intelligence might be wielded toward humane rather than extractive ends — is the whole subject of Volume II. Here it is enough to see that AI rescues nothing on its own. It magnifies whatever it is given. An incoherent civilization will build incoherent intelligence, faster.