human-computer-collaboration

The chapter discusses the challenges of effective collaboration between humans and intelligent 'creatures' (computers or AI systems), focusing on the dilemmas of communication frequency and detail level in reporting results. It highlights the trial-and-error approach adopted to optimize this interaction and reflects on the difficulties encountered in teaching computers to evaluate their own work intelligently, noting the limitations of artificial intelligence efforts in the 1960s. The text emphasizes the complexity and labor intensity involved in simulating human intelligence computationally, often resulting in inefficient performance despite the computer's speed advantage.

The chapter discusses the contrasting capabilities of humans and computers in problem-solving, highlighting that computers excel at tasks like differentiation which are difficult for students, while humans are better at recognizing simplifications intuitively. It emphasizes the responsibility of the human 'master' to guide the computer 'servant' with clear instructions and error checks. The chapter also outlines preliminary guidelines for effective division of labor between humans and computers, advocating for leveraging computers for repetitive or time-consuming tasks while cautioning against investing excessive time teaching computers tasks humans can perform quickly.