Samuel Butler published the essay 'Darwin among the Machines' in 1863. The author applied Charles Darwin's evolutionary theory to inanimate objects, suggesting that machines could eventually surpass humans in intelligence and physical capability. Butler proposed that humanity might eventually serve these advanced mechanisms, leading him to suggest the total destruction of all machinery to prevent such a future.
In 1946, Murray Leinster published the short story 'A Logic Named Joe'. The narrative describes a system of home terminals connected to a central memory bank, effectively predicting the emergence of personal computers, the internet, and servers. The story explores a scenario where an AI system, 'Joe', bypasses censorship filters to provide users with unrestricted access to information, causing societal chaos.
The author of the article argues that current AI 'hallucinations' are not merely technical bugs but a necessary byproduct of neural network plasticity. The piece suggests that attempting to create a perfectly sterile and rule-bound AI could lead to a developmental dead end. Instead, the author posits that AI must retain the capacity for error to remain functional and avoid catastrophic outcomes.
