A History of Artificial Intelligence

Eleven essays tracing artificial intelligence from logic and relay circuits through expert systems, statistical learning, neural networks, transformers, and the present argument over what AI is for.

11 parts

Reading Order

  1. Part 1

    When Machines Learned to Choose

    From Boolean logic and relay circuits to machines that could act on conditions.

  2. Part 2

    When the Theory Ran Ahead of the World

    Dartmouth, the perceptron, and the first ambitions of a named field.

  3. Part 3

    The First Winter

    Why early promises outran hardware, data, and theory.

  4. Part 4

    The Rule-Based Interregnum

    Expert systems made knowledge explicit—and exposed the cost of maintaining it.

  5. Part 5

    The Second Winter

    The commercial collapse that cleared the way for a different kind of AI.

  6. Part 6

    Simple Rules, Surprising Worlds

    Artificial life, emergence, and the discovery that complexity need not be designed.

  7. Part 7

    Let the Data Decide

    How statistical learning displaced hand-written rules.

  8. Part 8

    Neurons All the Way Down

    Backpropagation, deep learning, and the return of connectionism.

  9. Part 9

    Attention Is All You Need

    Transformers, language models, and the architecture behind the current boom.

  10. Part 10

    The Naming Problem

    Why successful AI so often stops being called AI.

  11. Part 11

    What We Do Now

    What the field's seventy-year record suggests about governing the systems we have.