Where in the World Are We?

Where in the world are we?

We keep trying to answer by deciding what AI is. That question splits at once: intelligence, consciousness, creativity, replacement, progress, threat. Each has its camp, and the argument between them starts before anyone has described anything. Perhaps we have not met a new someone. Perhaps we have entered a new somewhere.

The longer you use a tool, the more of your thinking travels through it. Where thinking travels, there is terrain. What you notice depends on how far you have travelled, and what you find there can be compared with what other people find. Someone deeply immersed is already far inland. Someone who uses ChatGPT now and then has visited a port. Someone who has never used it is hearing reports from home. Researchers chart it. Companies build settlements. Governments draw borders. The reports conflict, and the reports coming back from inland are starting to form a map that people with very different beliefs can examine together.

Here is one of mine. For weeks, working with these systems, I kept arriving at the same conclusions: that judgment was becoming the scarce thing; that evaluating would matter more than producing; that the ability to question the machine grows more important as its answers get cheaper. Then I read an article about raising children alongside AI. The most-liked comment beneath it said nearly the same things, in nearly “my words”.

“Epistemic supervision.”
“Access to judgment.”

I recognized the thought before I knew whether a human had written it, and I still don’t know whether it was written by a human, an AI, or both. For this essay, it doesn’t matter. I had reached a place by one route and found evidence that someone else had arrived there too.

Some landmarks are still there when you return, even if two people reach them through very different terrain. They can disagree about why the landscape looks the way it does and still recognize where the other person has been. But sometimes a traveller comes back with a mountain nobody else has reported, and there is no way, from inside, to know which you are holding: a landmark others will recognize, or a mountain only you have seen. I do not know what others have encountered. That is not a failure of the reports. It is why a map is needed.

One doubt belongs here, and I will leave it standing. Finding the same landmark does not yet tell us whether we discovered the same mountain or followed different roads into the same city.

Every report comes from somewhere, and every traveller brings assumptions with them. So the question is not whether a report is free of those assumptions. It is whether someone who brings different ones can still recognize what the other traveller is pointing to. Someone frightened by AI and someone exhilarated by it may disagree about nearly everything important. But if both encounter the same dangerous crossing, their disagreement about the continent does not prevent them from adding the river to the map. Common geography before common philosophy. We can begin mapping the terrain around AI without first agreeing on what AI is.

The proposal is modest: before asking everyone to agree about intelligence, consciousness, replacement or progress, see whether we can agree that there is terrain worth describing. Then the question is no longer a verdict anyone has to reach first. It is an invitation.

Where in the world are we?