Concept · part of Groups, splits and scale

Put a dozen people in a room to crack a hard problem and let everyone blurt out a first hunch. The room tends to stampede toward whichever answer was said first and most confidently. Let them think a while in smaller, looser clusters before comparing notes, and they land on the right answer more often.

That shape, rising and then falling, is what the book calls the inverted-U: a best point in the middle, with losses on either side. The room example is the Zollman effect, which the book takes from O’Connor and Weatherall: past a point, more connection just spreads early mistakes faster.

The book applies the same shape to network size. Its first instinct was that small communities cure what goes wrong in huge ones. But small networks shrink ideas too, and they tend toward echo chambers. Very large ones suffer under the pressure of what gets passed on and a falling limit on what an idea can carry. So the book stopped prescribing smaller networks. It argues for a shape of connection in place of a quantity: bridges sparse enough to avoid a stampede and dense enough to let evidence cross.

Read the full argument

How it connects


← The Zollman effect · Groups, splits and scale (6 of 7) · Part IV: integration or fragmentation? →

Explore all the ideas · Read the book in order · Open the interactive map