Chapter 6: The Monkeys That Didn't
Believing the Headline
There's an invaluable lesson here too, about how much trust can reasonably be placed in the media, and the internet more broadly, for disseminating the truth. Several things are worth pulling apart.
First, there's the blind, uncritical forwarding of Anderson's own announcement, essentially unchallenged, by most of the outlets with the resources and the obligation to check it.
Second, there's the reaction of the numbers expert quoted in the piece, Tim Harford — an economist and journalist who had spent years presenting the BBC's own radio program about statistics, More or Less. Asked about Anderson's project, Harford offered that "if he's running an evolutionary approach, holding on to successful guesses, then he'll get there."
Either Harford didn't know what Anderson was actually doing, or he said something that means nothing. Anderson's program wasn't holding on to anything useful. It was matching random strings against a text that already existed, and ticking off the ones that matched. Nothing was being built. A matched string didn't do anything, didn't lead anywhere, and didn't make the next match one bit more likely.
The word "evolutionary" brings a picture with it. In evolution, as it's usually described, a change that helps an organism survives and spreads, and a change that doesn't help dies out, or at least loses out to its competitors. Whatever you think of that picture, it has nothing to do with what Anderson did. But borrowing it makes his program sound like something real — as if a text were somehow being grown, piece by piece, the way life is supposed to have been. That's a farce, and someone of Harford's caliber should have spotted it in a moment.
In fairness to Harford, he may only have been repeating what he was handed. The coverage was already describing Anderson's method as "an evolutionary approach," one that kept the useful strings and threw the rest away. If so, that's the first lesson all over again: an expert passing along a framing without stopping to check whether it fits. Either way, it doesn't.
I suspect the word got past him — and past everyone else who used it — because of what it is. Evolution carries so much weight that attaching its name to something makes it sound like science, and almost nobody stops to ask whether the name fits. When evolution is the subject, the obvious is remarkably easy to overlook.
Last revised 2026-11-17