A Short History of AI as a Business

On 31 August 1955 four researchers signed a proposal to the Rockefeller Foundation. John McCarthy was a young mathematician at Dartmouth College, Marvin Minsky a junior fellow at Harvard, Nathaniel Rochester an engineer at IBM and Claude Shannon the founder of information theory at Bell Labs. They asked for $13,500 to bring "a carefully selected group of scientists" to Dartmouth for two months the following summer, and they stated their premise plainly: "every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it". They predicted "a significant advance" in that one summer. McCarthy chose the name "artificial intelligence" for the project partly to set it apart from cybernetics, the older field of feedback and control.

The workshop met in the summer of 1956 and produced no breakthrough. Its participants, including Allen Newell and Herbert Simon, who arrived with the Logic Theorist, a program that proved theorems from Principia Mathematica, led the field for the next thirty years. Simon predicted in 1957 that within ten years a computer would be the world chess champion. It took forty. The founding left two habits that the business of AI has never lost: a name that works as marketing as well as science, and a confident timetable.

AI as a business did not begin with ChatGPT. It has passed through at least three booms and two winters, and each left lessons that later companies relearned at their investors' expense. This chapter traces those eras, because nearly every argument about AI in the 2020s, over reasoning, hardware, benchmarks or bubbles, restates one from an earlier cycle.