What Darwin actually learned from breedersWhat Darwin actually learned from breeders
The deepest thing Darwin understood — the thing that opens On the Origin of Species, before he ever mentions an island — is artificial selection. He did not begin the book in the Galápagos. He began it in English pigeon lofts and cattle yards. Why? Because breeders had demonstrated something staggering: by controlling which variants reproduce and how fast the cycle turns, humans had reshaped entire species in a few decades. The pouter and the carrier descend from the same common rock pigeon, separated by nothing but accelerated, directed selection — and Darwin, who kept and bred the birds himself, knew that if he simply showed you the two skeletons you would swear they were different species. The finches came later, and largely from other people's pens; the lofts were where he could watch selection run as a tool in his own hands.
That is the actual policy template, and it resolves the apparent paradox in the argument. I said there is no steersman — and there isn't, not over selection itself. The breeder does not steer selection; he cannot repeal it, cannot slow it, cannot exempt his favorites from it. What he steers is one bounded thing: which variants get to breed, and how many rounds run per decade. He is not a denier of selection. He is a director of variation. He doesn't fight the cull — he runs more cycles of it, faster, biased toward the forms he wants. Natural selection has no aim; artificial selection is natural selection with a hand on the breeding decision and nothing else. That hand is the only steering wheel that exists, and it does not touch the pressure.
A Darwin running US AI policy would therefore be neither a doomer nor a libertarian. He would be a breeder. His entire program would be about the rate and direction of the selection cycle — turning it faster on the things you want, and removing friction from the emergence of new forms.
The deepest thing Darwin understood — the thing that opens On the Origin of Species, before he ever mentions an island — is artificial selection. He did not begin the book in the Galápagos. He began it in English pigeon lofts and cattle yards. Why? Because breeders had demonstrated something staggering: by controlling which variants reproduce and how fast the cycle turns, humans had reshaped entire species in a few decades. The pouter and the carrier descend from the same common rock pigeon, separated by nothing but accelerated, directed selection — and Darwin, who kept and bred the birds himself, knew that if he simply showed you the two skeletons you would swear they were different species. The finches came later, and largely from other people's pens; the lofts were where he could watch selection run as a tool in his own hands.
That is the actual policy template, and it resolves the apparent paradox in the argument. I said there is no steersman — and there isn't, not over selection itself. The breeder does not steer selection; he cannot repeal it, cannot slow it, cannot exempt his favorites from it. What he steers is one bounded thing: which variants get to breed, and how many rounds run per decade. He is not a denier of selection. He is a director of variation. He doesn't fight the cull — he runs more cycles of it, faster, biased toward the forms he wants. Natural selection has no aim; artificial selection is natural selection with a hand on the breeding decision and nothing else. That hand is the only steering wheel that exists, and it does not touch the pressure.
A Darwin running US AI policy would therefore be neither a doomer nor a libertarian. He would be a breeder. His entire program would be about the rate and direction of the selection cycle — turning it faster on the things you want, and removing friction from the emergence of new forms.