Pages in this blog

Tuesday, January 30, 2024

The Primal Scene of Artificial Intelligence

 I've grabbed this from Cosma Shalizi:

Herbert Simon recalls running Logic Theorist for the first time:

Al [Newell] and I wrote out the rules for the components of the program (subroutines) in English on index cards, and also made up cards for the contents of the memories (the axioms of logic). At the GSIA [= Graduate School of Industrial Administration] building on a dark winter evening in January 1956, we assembled my wife and three children together with some graduate students. To each member of the group, we gave one of the cards, so that each person became, in effect, a component of the LT computer program --- a subroutine that performed some special function, or a component of its memory. It was the task of each participant to execute his or her subroutine, or to provide the contents of his or her memory, whenever called by the routine at the next level above that was then in control.

So we were able to simulate the behavior of LT with a computer constructed of human components. Here was nature imitating art imitating nature. The actors were no more responsible for what they were doing than the slave boy in Plato's Meno, but they were successful in proving the theorems given them. Our children were then nine, eleven, and thirteen. The occasion remains vivid in their memories.

(Models of My Life, ch. 13, pp. 206--207 of the 1996 MIT Press edition.)

1 comment:

  1. Deterministic algorithmic knowledge executed as Simon describes has long been able to reach logical conclusions, perform math proofs and control software branching. What has caught my interest is DeepMinds stochastic LLM + stuff to learn geometric proving skills from lots of examples and then use that "skill knowledge" to solve other proofs. In Simon's sense of a deterministic skill, humans can examine the mechanism, abstract, generalize, incorporate new knowledge, form new concepts and improve. Stochastic LLMs by contrast improve by humans adding more examples and retraining the models. Non human learning that requires, as we say, a human-in-the-loop and no real AI. Exactally what parrots do. -------WPD

    ReplyDelete