On the Sunday morning at Hacklands, after we’d heard some lovely Chopin, James Tagg got up to speak about his new book “Are The Androids Dreaming Yet?” a popular science look at the history and potential future of Artificial Intelligence. I sat down with James a few weeks later to talk more about the ideas in the book.
Back in August, listening to the talk, I realised just how strongly ingrained our idea is that machines will just get smarter and smarter until they meet and then exceed human capability but also the implication that we’re advanced machines ourselves. Many people are resigned to it, accepting that it will happen sometime in the near future. And we do see machines doing things routinely now that a few years ago would have been impossible – the example I always think of is language translation, I would have argued (even five years ago) that it was just too complex a task for machines to do, and while Google Translate isn’t quite perfect, it does a much better job than I would have predicted. I was wrong.
But. I also believe that there are human qualities and activities that machines might mimic well, but that that isn’t the same as them being intelligent. But what if I’m just wrong (again…)?
In his book and in the conversation we’ve recorded, James moves the argument from beliefs and faith, to the arena of mathematical proof. He shows how the work of Hilbert, Gödel and Turing (among many others), which formed the basis of digital computing in the last century, also holds the key to understanding its limitations.
The good news is that creativity and free will remain something we can reserve for ourselves – and to prove it, I woke up this morning and decided to write this post, and I made up which words to use and the order in which I put them. I think.
Download the podcast (21MB)
PS if you like talking about this stuff and you’re near London, you can join in the conversation at Future of Work: Artificial Intelligence on November 10th.