Are We Becoming More Connected to Reality, or Less?

A natural question for anyone who thinks connection to reality matters: is the human species getting better at it over time?

I want to resist the two available narratives. The progress story says we are steadily replacing superstition with knowledge. The declinist story says we are drowning in misinformation and were more grounded when life was harder. Neither survives examination, and the honest answer is genuinely mixed.

The Case for More

Science connects some of us to reality. That is real and it is not small.

But qualify it immediately. Scientists are subject to cognitive dissonance like everyone else, and can hold rigorous standards in their speciality alongside remarkable nonsense outside it. The discipline is domain-specific and does not automatically generalise, which is disappointing but well documented.

More people work as engineers, physicians, and technicians — occupations that force daily contact with facts that do not care what you believe. A bridge either stands or it does not. That is a real increase in the number of people whose working lives involve reality-testing, though it does not follow that the habit transfers to their politics.

And religion has moved the right way. Go back a millennium and people believed considerably stupider things, with more confidence, and with more willingness to act on them. In that narrow domain the improvement is unambiguous.

The Case for Less

Here is the mechanism that troubles me most, and it is new.

Flat-earthers and conspiracy theorists can now find each other.

Consider what happened to such a person before. In a town of a few thousand, someone convinced of an eccentric theory had no one to discuss it with. The belief was socially unsustainable — not refuted, just unsupported, dying of loneliness. Isolation was doing epistemic work nobody had designed it to do.

That mechanism has been switched off. A belief held by one person in ten thousand now has a community of hundreds of thousands, with content, moderators, internal status hierarchies, and a shared vocabulary. It is sustainable indefinitely, and it feels from the inside exactly like membership of a scientific community: peers, disputes, jargon, a sense of knowing something outsiders do not.

We removed a filter without noticing it was a filter. Nobody could have predicted that connection would work this way, and I do not think it argues against connection — but the cost is real and largely unremarked.

Why I Refuse the Summary

There is a temptation to net these out and declare an overall direction. I think that is a mistake, because the changes are not commensurable.

The gains are concentrated in domains where reality pushes back quickly and unambiguously — engineering, medicine, anywhere with a short feedback loop. The losses are concentrated in domains where it does not: politics, history, nutrition, anything where being wrong costs you nothing for decades.

So the picture is not a species getting better or worse at reality. It is a species getting much better at it where the feedback is fast and no better at all where the feedback is slow — and, thanks to the loneliness effect, actively worse in the slow domains than it was.

That is a specific claim, and it suggests a specific response: the interesting question is not how to make people more rational in general, but how to shorten feedback loops in domains that currently have none.

Whether we should be more connected to reality is a separate question, and a larger one. I have assumed the answer throughout, and I keep noticing that I have never actually argued for it.