About a year ago I really enjoyed reading Edge.org’s What to Think About Machines That Think: Today’s Leading Thinkers on the Age of Machine Intelligence. What I liked about it was getting the wide range of perspectives on the issue. It felt like I was getting a lot of insight into the topic, and I felt like I was having a discourse with the people at the forefront of AI research. The experience was overwhelmingly positive, and thus I was pretty delighted to find out that there was more of it available. Not really being sure which one to tackle next, I dove into This Idea Must Die, the 2015 book published by Edge.org editor John Brockman, which attempts to look at the ‘Scientific Ideas That Are Blocking Progress.’ This seemed interesting enough, and I was pretty happy to pick it up.
Unfortunately, it didn’t hold up. While I was happy to return to What to Think About Machines That Think day after day to get someone else’s opinion on AI, here it was very hard to find the motivation to continue. The topic’s were pretty diverse, and while John Brockman did a pretty good job organizing them in a pretty logical way, the seams always seemed to show. While that is a minor gripe, it does lead to a more substantial one: the topics here are too far reaching. I don’t imagine there are any kind of 17th century polymaths that can have solid opinions on all these fields. Those people no longer exist (only because these fields are not too complex), and a person who could actually gain something from every essay in here likely doesn’t exist either. For me personally, there were several entries that touched on modern physics that simply made my eyesight blurry. It was really no fault of the authors involved at all. The topics were just too out of my range.
And no one wants to feel stupid. That doesn’t really make for a good reading experience.
This can be elaborated on with another comparison to What to Think About Machines That Think. In that book, each individual author brought their specific expertise to a common theme (AI) and that theme could function as a scafolding for understanding the author’s point. They had to make their knowledge relatable to some common argument. But in this one no such scaffolding exists with This Idea Must Die. With What to Think About Machines That Think: , even though I likely can’t remember many of the authors and arguments know, at the time of reading I felt like I was getting something from it. I felt like I was engaging with the material. Here, there were several pieces that merely made me say ‘Um, yea. Ok.’
This was somewhat balanced by the the couple of essays wherein the author was speaking to something that I have some expertise. In some cases, I agreed with them, which is pleasant. In a few other cases, some of the essays where not only wrong, but laughably so. That too took me away from this one. A few bad apples shouldn’t take from the credibility of the whole pot, but sometimes they can.
I’ll likely pick up another Edge.org book, but certainly with a lot less enthusiasm.