I’m still slowly making my way through Superintelligence by Nick Bostrom, and a few pages into Chapter 8 – titled as above – is this brilliant exclamation from the imagined reader:
But wait! This is not what we mean! Surely if the AI is superintelligent, it must understand that when we asked it to make us happy, we didn’t mean that it should reduce us to a perpetually repeating recording of a drugged out digitised mental episode!
It’s an oddly fun book, when read in the context of Douglas Adams or sci-fi books. Though perhaps best not to spend too long contemplating exactly how accurate it might be.
Thinking more about the similarity to the work of Douglas Adams, the book constantly reminds me of Deep Thought and Marvin, the original paranoid android. Especially the last few chapters with titles and subtitles such as ‘The superintelligent will’, ‘The relation between intelligence and motivation’, and ‘Existential catastrophe as the default outcome of an intelligence explosion?’.
I think Tome Yorke would enjoy it.