Part II: How Things Find Answers

"Nature is trying very hard to make us succeed, but nature does not depend on us. We are not the only experiment." — R. Buckminster Fuller

Every complex system faces a version of the same problem: how do you find a good solution when the landscape of possibilities is too vast to search exhaustively?

Bacteria finding food, venture capitalists choosing startups, jazz musicians improvising, engineers optimizing designs, and evolution shaping species — all of them are running search algorithms. They don't call them that. Most of them don't know they're doing it. But the strategies they independently discover are the same, for the same mathematical reasons.

Part II explores seven of these universal search strategies. Gradient descent (Chapter 7) is the art of finding solutions by feeling downhill — the strategy shared by water, neural networks, and market prices. Explore/exploit (Chapter 8) is the fundamental tradeoff between trying new things and sticking with what works — a dilemma solved (and re-solved) by organisms from bacteria to toddlers. Distributed vs. centralized (Chapter 9) is the oldest architectural debate in every field, from neuroscience to military strategy. Bayesian reasoning (Chapter 10) is the optimal way to update beliefs with evidence — and the story of how every field discovers it, forgets it, and rediscovers it again. Cooperation without trust (Chapter 11) shows how game theory's insights appear identically in bacterial colonies, Cold War diplomacy, and open-source software. Satisficing (Chapter 12) explains why "good enough" reliably beats "optimal" in systems from evolution to grocery shopping. And annealing (Chapter 13) reveals why systems need randomness — disorder, noise, creative destruction — to escape bad solutions.

The unifying insight: the universe has discovered a small number of search strategies that work, and every complex system converges on them independently. Understanding these strategies transforms how you see problem-solving in any domain.

Pattern Library checkpoint: After Part II, your library should include entries for each search strategy, with notes on which ones your own field uses — and which ones it should be importing from elsewhere.

Chapters in This Part