There is a specific experience that used to be unavoidable.
You have a thought that is not finished. It is not an opinion yet, it is barely a sentence. You circle it. You put it down and pick it up on a walk. It stays wrong for a while, in a way you cannot articulate, and then at some point it either becomes something or it does not.
Nobody enjoys this. It is uncomfortable, it takes time, and for most of human history there was no alternative to it, which is the only reason anybody ever did it.
There is an alternative now. It is very good, it is instant, and it is free.
I am not building toward a complaint. I use it. This piece was argued out against a machine before I wrote a word of it, and the machine caught two errors I would have published.
I am interested in something narrower, which is what happens to a species that outsources one particular faculty at one particular moment.
Three things, not connected
The shape of this argument is the shape of every crank essay ever written. Somebody notices three large things happening at once and declares a convergence.
Every generation has three. Nineteen forty-five had the atom, the end of empire, and antibiotics. The feeling of standing at a hinge is not evidence of a hinge, and if I could only offer you simultaneity I would not be writing.
The climate is doing what the data said it would do, at roughly the pace the data said, and the pace is not the interesting part anymore.
Machines have become genuinely good at thinking, in the specific sense that they now produce the middle of the process rather than the ends. Not the answer. The working out.
And an eighty-year government record on unidentified phenomena has begun surfacing in fragments, on a schedule set by nobody in particular, with the interesting parts still not in public.
These are unrelated. Nothing causes anything else on that list.
They have one thing in common, and it took me a while to see it.
The same failure
All three require the same cognitive move, and it is the one humans are worst at.
Each of them demands that you reason about something distributed, delayed, and invisible in any single instance.
The climate has no photograph. I wrote about this recently in a different context: the drowned boy on the Turkish beach moved governments in forty-eight hours and the effect was gone in five weeks, and the drownings continued at the same rate for years afterward. That is the identifiable victim problem, and carbon is that problem with no boy available at all. There is no image. There is no moment. There is a number that goes up.
The cognitive shift has no photograph either, and its case is stranger, because the thing you would need to notice is an absence. A capacity that never developed leaves no trace and produces no sensation. It is like aphantasia, where a substantial number of people reach middle age before learning that when other people say picture a beach, they mean it literally. Nobody with aphantasia walks around feeling a hole. You do not miss a room you never entered.
And the disclosure record is invisible for the dullest possible reason, which is that reading it is nobody's job. Not suppression. Not a conspiracy of silence. A very large quantity of paper, in the custody of institutions with other priorities, where the release of any particular item requires somebody to decide it is their responsibility, and nobody has. That is how eighty years pass.
Three problems. One requirement. Notice the thing that is not in front of you.
What we are good at
Set against that, look at what the species has actually solved.
Everything with a face. Everything with a deadline. Everything where the cause is adjacent to the effect and somebody can be held responsible on a timescale that fits inside a career.
We are extraordinary at those. Not adequate. Extraordinary. We cured diseases, we fed billions, we put people on the moon within a decade of deciding to, and we did all of it with the same equipment that cannot make itself care about a number that goes up slowly.
The problems that are left are the ones that defeat the equipment. That is not a coincidence and it is not a tragedy. It is a filter. We solved the tractable ones first, the way anybody would, and what remains at the bottom of the barrel is the category we were never built for.
Which is the actual observation, and I will state it plainly.
We are handing off the faculty at precisely the moment when the only problems left are the ones that require it.
Neither good nor bad
I want to resist the two available endings, because both are lazy.
The first is Thamus. In the Phaedrus, the king tells the inventor of writing that his invention will produce forgetfulness in the souls of those who learn it, because they will trust external marks instead of remembering from within. He was right about the mechanism. Literate people do remember less. He was wrong about everything that mattered, and every alarm since has had to answer him.
The second is the shrug. It always works out. We adapted before.
Both are guesses dressed as conclusions, and the honest position is that nobody knows, because the relevant experiment is running right now on children who are eleven.
What I would say instead is that the trade is not symmetrical this time, and the asymmetry is worth naming without panicking about it.
Writing externalized storage. Calculators externalized execution. Both of those took something off the mind and left the mind's other functions intact. What is being externalized now is the middle: the search for the next move, the sitting-with, the circling. That is not a subroutine. That is the same faculty required by all three of the problems in this essay.
It might be fine. Storage turned out to be fine, spectacularly so, and I would not give up writing to get memory back. It might be that a generation raised with the scaffold develops something else instead, something I have no way to imagine, in the way a literate person cannot really imagine what an oral culture knew.
But it is a different kind of trade, and pretending otherwise because the last four trades worked out is not reasoning. It is pattern-matching, which is the thing the machine is better at anyway.
The last part
Here is what I keep arriving at, and it is not a warning.
The reason this essay is possible is that I learned to think before there was anything else to do it with. The discomfort I described at the start is available to me as a memory of a default condition. I know what it feels like to not know what you think and to have no alternative but to wait.
Somebody eleven years old right now will not have that reference. Not because anything was taken from them, and not because they are lesser, but because the scaffold was there from the first day and there was never an occasion to build the muscle it replaces.
They will not be able to see this. Not because they are inattentive. Because you cannot perceive the absence of a capacity you never had, any more than you can perceive the color you were born without.
Which means the generation that can still notice this is the one that is currently writing about it, and we have maybe one cohort left in which the comparison can be made from the inside.
That is not a call to action, because I do not have one. Nobody is going to give this up and I am not going to either.
It is just an odd position to be standing in. Late enough to see the change, early enough to remember the thing being changed, and out of the way before anyone can check whether we were right.
The thought is still unformed. I am leaving it that way on purpose.
Great piece - thanks very much for writing it.
Good point. Seems adjacent to report that younger generation has lost the mental ability to locate themselves in space since they grew up with Google Maps. Being old, I cannot imagine what that must be like. I could find my way home from anywhere in the state with just the sun and my mental map of landmarks. Inconceivable that people will lose the ability to think about thinking.