What computers still can't do

In this book, Hubert Dreyfus critiques the field of artificial intelligence, particularly the early approaches that relied on symbolic representations to simulate human intelligence.

He argues that human cognitive capacities, such as common sense, moral and ethical reasoning, and contextual awareness, cannot be fully captured by machines.

The book, first published in 1972 and revised in 1992, includes a new introduction that assesses the changes in AI research, including the shift towards connectionism and neural networks.

Dreyfus's arguments emphasize the importance of unconscious processes and the limitations of formal rules in replicating human intelligence.