I was down with flu recently, and it was not the kind you work through. I could not do anything. What frightened me more than the fever was that I could not think. A thought would start and slide away before I reached the end of it. When thinking is most of what you do all day, losing it, even for a few days, is a strange and quiet kind of fear.
When I finally looked at a screen again, the world had not paused for me. Content kept coming from everywhere, new models, new tools, new announcements, as if my empty head made no difference to any of it. Take only the week of 6 to 10 October. One tracker I looked at counted eight confirmed model releases. A new Mistral, a new Claude Haiku, a Qwen image model, a Google image model, a coding model from Microsoft. None of them were the launch of the year. They simply kept arriving, one after another, like trains at a station.
The Part That Stayed With Me
What stayed with me was not the number. It was that nothing waited. My own intelligence went offline for days and the amount of intelligence being produced in the world did not move at all. A few years ago that would have felt impossible to say. Now it just feels true.
The Price Of Thinking
Sam Altman said in a conversation at the Federal Reserve last year that the cost of a unit of intelligence has fallen by more than ten times every year for five years. That is his own estimate, not a standard measure. He also described a future where intelligence is a utility, like electricity or water, bought on a meter.
The phrase he keeps reaching for, too cheap to meter, is actually from 1954. It was said about nuclear power, that electricity from atoms would one day be almost free. It did not turn out that way. I am not certain it fails the same way here, but I notice it is a promise that has been made before.
It is also not as simple as one price falling. From what I have read, the cheap end keeps getting cheaper while the very best models have gone up in price. So the part turning into a commodity is the good enough part. Which is exactly the part most of us spend our working days on.
What Keeps Me Awake
If good enough thinking costs almost nothing, what do people do? I do not mean the film version where robots take over. I mean something quieter. For most of my working life, being the person who could work it out was the thing I was paid for. If working it out becomes something anyone can switch on for a few pence, that is not a small adjustment. That is the ground moving.
I wrote a few weeks ago about how wonder wears off. I remember being amazed when an AI named everything in my fridge from a photo. Now nobody blinks. I think intelligence is going the same way, from wonder to tap water. Nobody stands in a kitchen marvelling at tap water. That is fine for water. I am less sure it is fine for the thing we built our whole idea of a career on.
What I Think Is Left
I do not have an answer, and I would rather say that plainly than invent a tidy one. But I notice a few things that do not seem to get cheaper when thinking does.
Deciding what is worth building at all. A machine can hand you ten thousand options, and somebody still has to choose one and live with it. Being answerable when it goes wrong, which is the point I made in my piece on Musk, that a person still has to be the one who answers for it. And knowing the people in front of you, what they are actually asking for underneath what they say.
In my book, nobody in the village was ever thirsty for water. They were thirsty for the morning it was costing them. I keep coming back to that now. If intelligence is the water, and it is about to become nearly free, then the real question was never the water. It is what we do with the morning it gives back.
I do not know yet what that looks like. I only know it is the question that stuck in my head while the rest of me was too sick to think, and it has not left since.