Always Longer Than Expected

Always Longer Than Expected begins with a line that has haunted engineers, scientists, and planners for decades: "It always takes longer than you expect, even when you take into account Hofstadter's Law."

In an age obsessed with acceleration, where AI breakthroughs arrive every month, timelines compress into press releases, and promises of imminent artificial general intelligence multiply, this book asks an uncomfortable question: why does the future keep missing its deadlines?

Carina Northwood traces the strange afterlife of Hofstadter's Law from a witty observation in a 1979 book to a structural principle shaping modern technology. Drawing on the history of artificial intelligence, from early symbolic systems and AI winters to today's trillion-parameter models, she shows that delay is not a failure of planning but complexity revealing itself only once work is underway.

The book moves from research labs to data centers, from chip supply chains to power grids, from startup hype cycles to geopolitical strategy. Northwood shows how AI progress depends on far more than algorithms, requiring human expertise, institutional memory, infrastructure, regulation, and the slow reproduction of judgment itself. Each attempt to speed things up introduces new bottlenecks. Each correction becomes part of the delay. The recursion is not a bug but the structure itself.

The hardest problems in intelligence, energy, medicine, and climate resist linear timelines because deeper layers of reality emerge only as they are pursued. Delay is how complexity announces itself.

At a moment when headlines promise arrival just over the horizon, this book explains why the horizon keeps moving and why learning to work with that fact may be the most undervalued form of technological wisdom we have.