12 August

The emergence of a near-transcendent force of destruction in the Manhattan Project created the fear of global annihilation. Instead that transcendence ushered in an era of almost unprecedented peace, via the threat of mutually assured destruction. There’s something here in the thinking about AI today. Another possibly near-transcendent technology, in terms of the qualitative leap in capacity. And more apocalyptic fears. Technological transcendence is not a controllable force. It’s a force that supersedes human rivalry, suppressing it, an access to forces that overwhelm fights for status. It’s possible, and obviously conjectural (but also a pattern in history), that AI may become not a destroyer of worlds, but a creator and protector of them. There was a profound hubris in Oppenheimer’s fear. The human belief that we control fate, and are master of the things we create, including the belief that any behaviours and capacities they exhibit must have been put there by us. If you have children, how much control do you have over them, and their destinies and ideas and behaviours, as time goes on? Our creations have agency. And sometimes so much agency that we become humble before them.