The A.I. Daily
How accurately did science fiction predict AI?
A retrospective essay assessed how well science fiction predicted AI, arguing sci-fi describes rather than predicts. The author broke down how AI was actually built, used and deployed, then scored stories; Ted Chiang's 'The Lifecycle of Software Objects' ranked first, with Richard Powers' 'Galatea 2.2' at 8.4/10. No pre-2017 work clearly described Transformer attention or the path from chatbots to agents.