I took a long time to write the book. At the start, I was caught up in a lot of the optimism around computers and politics. The Obama era of campaigns. I was trying to find my voice in those discussions and I happened to be walking through the Toronto Metropolitian University library (then Ryerson) when I chanced on a 1972 book called Computers and Political Campaigning. It sort of hit me that people have been talking about how computers are going to revolutionize campaigns for at least 40 years.

One of the cases in the book was the Simulmatics Corporation, that has been explored in depth in Jill Lepore's fantastic If Then. The firm promised in the early 1960s to simulate politics with computers. But Simulmatics only gets a footnote, and the book moves to looking at databases, direct mail, and all these more routine, but at the time, better fantastic technologies for computers. So on one side, what might sound fantastical was really routine, but what was routine was really fantastical.

What I hope people take away is both a skepticism that computers can fix politics, and better demands for claims that it will. Especially in the next election cycle, how can we assess accurately the hype. My book helps readers distinguish between what's new and what's just someone who knows the rules of the game, trying to oversell the new tool or piece of software.

Now, obviously, the moment is AI. What is AI going to do? Well I want readers to ask, what AI? How is that different from the last so-called AI. And I think I give readers a vocabulary to make a distinction.

Ongoing thread. More from Fenwick McKelvey to follow.