Summary
Hubert Dreyfus's central thesis in "What Computers Can't Do" is that the formal, rule-based, symbolic manipulation approach to artificial intelligence, prevalent at the time of its publication, fundamentally misunderstands the nature of human intelligence and will ultimately fail to replicate it. He argues that human intelligence is not a matter of applying explicit rules to explicit representations of the world, but rather an embodied, intuitive grasp of context that cannot be formalized.
The book's key ideas include the distinction between skilled, knowing-how intelligence and rule-based, knowing-that intelligence. Dreyfus highlights the crucial role of embodiment and the background context of human experience, which computers lack. Readers learn that genuine human intelligence relies on a tacit understanding developed through continuous interaction with the world, a process AI, as then conceived, cannot emulate through symbol manipulation alone.
Full text isn't indexed yet — this overview draws on general knowledge of the book and its metadata, and chat works the same way.
Key concepts
- Symbolic AI — The approach to artificial intelligence that assumes intelligence can be achieved by manipulating symbols according to formal rules.
- Tacit Knowledge — Knowledge that is difficult to articulate or formalize, often gained through experience and practice.
- Embodiment — The idea that intelligence is not purely abstract but deeply tied to having a physical body and interacting with the physical world.
- Background Context — The unstated, implicit understanding of the world that humans possess and use to interpret situations.