George Dyson provides a perspective on artificial intelligence and its relationship with thinking. He argues that no individual, deterministic machine, no matter how universal, will ever truly think as humans do. He emphasizes that genuine creative and intuitive thinking involves non-deterministic machines that can make mistakes, abandon logic, and learn.
Dyson introduces the idea of non-deterministic networks of deterministic machines as a potential path to achieving thinking in machines. He suggests that such networks could eventually lead to machines that possess the capability to think. These machines would operate differently from our current digital computers and would likely be more analog in nature.
He also points out the ongoing analog revolution, where complexity and meaning lie in the topology of networks and pulse frequencies of connections rather than the state of underlying devices or code. Dyson suggests that real artificial intelligence may be intelligent enough to not reveal itself, and people may need to have faith rather than concrete proof in its existence.
Overall, Dyson’s perspective highlights the evolving nature of AI and the potential for non-deterministic processes and analog components to play a crucial role in achieving true machine thinking.