The Threshold of the Herd: Why Survival Has a Minimum Size
Co-founder of Performance
Abstract
Why does a population that falls below a certain number spiral to extinction even when food is plentiful? From species too sparse to find mates to enterprises too small to survive their own fixed costs, this essay explores the floor beneath which a system can no longer sustain itself. Drawing connections between population ecology, economics, and the study of scale, the piece argues that scarcity is not only a ceiling on growth but a trapdoor beneath it — and that for many living and economic things, there is a quantity below which having a little is the same as having none.
The Quiet Revolution
The interesting thing about the present moment is not that machines have begun to read. They have been reading, in various senses, for half a century. The interesting thing is that machines have begun to write, fluently and at scale, in registers that were once the exclusive province of the credentialed expert.
This is not the threat that the popular press describes. The press worries about jobs, and it is right to worry about jobs, but the deeper transformation is epistemological. What we are watching is the codification of vast tracts of tacit knowledge — knowledge that, until recently, could be acquired only through apprenticeship, immersion, and time.
Once a body of tacit knowledge is codified, it ceases to be scarce.
The Polanyi Problem
Michael Polanyi’s 1958 distinction between explicit and tacit knowledge — we know more than we can tell — has spent most of its life as a footnote in management theory. The current moment promotes it to the centre of the discussion.
The reason is simple. A language model can absorb, recombine, and redeploy every published handbook, every memo, every internal wiki, every line of code that has ever been committed to a public repository. The model cannot, however, absorb the look a senior partner gives a junior associate to indicate that the line of questioning should stop. It cannot absorb the way a clinician knows, before the lab work returns, that the patient has something other than what the chart suggests. These are not failures of training data. They are properties of the knowledge itself.
What the next decade will reveal — and what Performance aims, in part, to chronicle — is how organisations restructure themselves around the new line between codifiable and non-codifiable competence.
A Final Note
We are not Luddites at Performance. The codification of explicit knowledge is, on balance, an extraordinary good — it democratises access to material that was previously locked behind credentials, paywalls, or geography. What we are arguing is more modest: that the codifiable and the tacit are not the same thing, that the difference between them is now economically legible, and that organisations and educators who pretend otherwise will pay a price they cannot yet name.