You On AI Field Guide · W. Ross Ashby The You On AI Field Guide Home
TxtLowMedHigh
PERSON

W. Ross Ashby

The psychiatrist-turned-cybernetician who proved that only variety can destroy variety, built a machine that taught itself to survive, and left behind the laws every adaptive system—brain or network—now obeys.
W. Ross Ashby (1903–1972) is the most consequential thinker about intelligence that most people working on artificial intelligence have never read. A practicing psychiatrist who spent his days in the wards of Gloucester and Northampton, he kept a private journal for forty-four years, working out in 7,189 handwritten pages the laws of how organized systems hold themselves together. His five-word law—only variety can destroy variety—is the closest thing the study of complex systems has to a conservation law, and it predicts both why scale works in AI and exactly where scale must stop. In 1948 he built the homeostat, a machine of surplus bomb-control units that searched its own configurations until it found stability against disturbance—the first working demonstration of what is now called reinforcement learning from human feedback, sixty years early. His two books, Design for a Brain (1952) and An Introduction to Cybernetics (1956), established the architecture of cybernetics and named the idea of intelligence amplification—the possibility that intellectual power, like physical
← Home0%
PERSONBook →

Keep reading with YOU ON AI

Unlock the full book, 10,000+ field-guide entries, and a 1000+ thinker library. If you have a book code, register now — it takes a minute.

Register with book code Sign in