The last page.

Herbert Simon, the same person who warned in 1971 that a wealth of information creates a poverty of attention, also gave us a word for how people usually decide. He called it satisficing. Most of the time we don't look for the best possible option. We look until we find something good enough, and then we stop. We rely on context, on a conversation with someone we trust, on the mood in the room, on whatever happened to be in front of us that day. Decisions made this way are a bit messy and limited, but most of the time they work well enough.

What I find interesting is that satisficing depends on a place to stop. When the shelf ends, good enough is easy to recognise. When there is always one more option, one more review and one more tab, settling starts to feel careless, as if we didn't do our homework. So we keep looking, and we get tired. Then the decision either stalls or goes to whatever is easiest to reach for.

Lately, the easiest thing to reach for is often an AI system. I use it too, and often it helps. But I see a loop forming that I think is worth naming.

AI makes it cheaper to write and to read, so we produce more: more drafts, reports, meeting notes, summaries, and summaries of summaries. More text makes us more tired, and the more tired we are, the more we want something else to read it for us. In the 1860s the economist William Stanley Jevons noticed something similar with coal. When steam engines used coal more efficiently, the total amount of coal burned went up. Efficiency invited more use.

Each step in this loop makes sense on its own. Taken together, they move something important. First we hand over the reading, then the comparing, then the shortlisting, and then the framing of what the options are in the first place. At some point the tool that filters becomes the tool that decides, and it is hard to say exactly when that happened.

I see the same pattern at work. Dashboards multiply. Every meeting produces an automatic summary that few people open. Strategy decks arrive with long appendices produced in minutes. When everything can be documented, the scarce thing becomes judgement: someone who can say this matters, that doesn't, and this is why.

If that judgement slowly moves to whoever, or whatever, writes the summary, the organisation has changed the way it makes decisions without ever choosing to. This is one of the quieter parts of AI adoption, and one I think leaders should look at closely, because it doesn't show up in any roadmap.

For most of human history food was scarce, and both our bodies and our cultures were shaped by that. When food became plentiful in many parts of the world, we needed new habits: cooking, mealtimes, shared ideas of what is enough. We are still learning them. Information abundance is much younger, and our sense of "enough" is still thin.

So I think the work ahead is a design question: how do we put stopping points back in? A few places I would start:

Think of one decision you asked AI to help with recently. If there had only been three options in front of you, would you still have needed help? And if you would, did you want it to choose for you, or to help you find the last page?

I would like to hear what you noticed. Write to me →