Uncanny Valley: A Memoir

"I was looking for a book that would talk about Silicon Valley and these new ways of information processing. I wanted to move from Red Plenty through into what’s happening in Silicon Valley at the moment and draw out both the tangencies and the overlaps. There’s a minor publishing industry on Silicon Valley, which has gone from books about how ‘Silicon Valley is awesome’ to books about ‘Silicon Valley is evil.’ I was looking for something a bit different. Initially I proposed the Zuboff book, but it’s 700+ pages, so I felt guilty imposing it. Also, like many authors, Zuboff depicts Silicon Valley as an all-devouring Moloch, a monstrosity of surveillance with a single logic. Books like that get at very, very important aspects of the politics of Silicon Valley, which I don’t want to discount, but there’s also something too totalizing about their understanding of how these technologies are reshaping the world. What I really like about the Wiener book is that she uses her life story and the various situations that she finds herself in as a way of trying to describe this world that is emerging, but not necessarily to try and pin it down with a simple argument saying ‘this is this’ or ‘this is that.’ She starts as an underpaid publishing person in one of those miserable jobs that we all know about. Then she goes to a minor startup, and moves on through another startup, through to working for a major platform, GitHub. She never names any of the companies, but it’s often quite clear who she’s talking about, and one of the fun parts of the book for people who hang on the outside of this world is figuring out who is who. Wiener uses her own personal experiences and those of people around her to capture how the logic of information and engineering, the Marc Andreessen “Why Software is Eating the World” logic actually intersects with the real world, with real people, with real lives. Also, she wants to get at the messiness of the people who are actually implementing this stuff, the complicated ways in which they try to deny to themselves that they are part of the surveillance industry and to capture the role of women. There’s a great history of Silicon Valley that came out about a year ago, The Code , by Margaret O’Mara, which talks about the way in which women get systematically written out of the story. The stuff that they did was not considered core because many of them weren’t engineers. Wiener gives a very, very good sense of what it is to be on the receiving end of that. Silicon Valley is built around the idea that if you’re on the hard, cutting edge of mathematics and engineering you’re doing good stuff, but if you do anything softer and fuzzier your contribution is discounted. You’re somebody who is necessary, but in the way that the plumbing is necessary in a building. You don’t want to think about it more than you have to and you certainly don’t think about it as something that adds to the business’s value proposition. Uncanny Valley is a very good book about the system of Silicon Valley. She has this bit at the end where she says, “I was looking for stories; I should have seen a system”, but I think that underestimates what she has accomplished. She’s doing what Spufford does in a very different way, which is why I think the books are comparable. She’s telling the story of the system filtered through the story of her own life, and this gives a sense of the system as something which is not totalizing. The book is wary and oblique, getting at the system from a variety of different perspectives, looking at the different ways in which it works and which it doesn’t quite work as advertised. She describes the primate hierarchies that are associated with it and she gives you a far better sense of how it actually unfolds, in this complex and messy way, than more deliberately big-picture books that are organized around a statement or thesis. Yes, it’s very clearly extraordinarily important because of how it is reshaping the world. These companies have enormous clout and their actions have consequences. But they are not providing an engineering solution or an optimal world. They are half-accidentally creating an ecology with unexpected consequences and unanticipated evolutionary niches, occupied by strange new predator and prey relations. Within the technological possibilities that are being created, people find their own ways of doing stuff. I was reading Uncanny Valley together with William Gibson’s science fiction novel Agency . They make for an interesting pairing. They’re both about Silicon Valley; Gibson continually uses brandnames for effect, while Wiener, as I said, never mentions them on principle, referring to Facebook, for example, as ‘the network that everybody hates.’ Gibson has this great phrase from one of his early books ‘the street finds its own uses for things.’ That’s what I felt about Wiener’s book. You have these systems which are supposed to effectively replace hierarchy and standard ways of doing things with a simpler, cleaner, more efficient world, but it doesn’t work cleanly and efficiently. So GitHub is supposed to be a platform for people to collaborate on building software, but it works—like all of these Silicon Valley businesses—by using algorithms and other means to replace employees. So you’ve got Wiener and her team of four people who are supposed to be the moderators controlling the content that is put up by nine million users. Obviously that isn’t going to work. What happens is that people begin to turn the system to their own uses. She describes how the creepier elements of the alt-right and neo-Nazis begin to colonize GitHub as a means of first of all, organizing protests against the perceived dominance of the man-hating-women in the technology industry and then as a platform to organize support for Trump. This is not something that the designers of the system ever expected. But when they create these open systems, which have a minimal degree of human oversight, people are of course going to start finding unexpected ways to use them. Some of these unexpected uses can be wonderful and extraordinary and transformative, but given the way human beings are, some are going to be self-interested and some downright creepy and nasty. The story behind the story that Wiener tells, the story that she’s hinting at and pushing towards through her personal experiences, is the story of how we ended up in the world that we’re in today? What role does this Silicon Valley engineering mentality play in this? When Francis Spufford talks about why he wrote Red Plenty as a novel rather than non-fiction, he says that the novel form gives you negative capability in the Keatsian sense. It allows you to have your cake and eat it, laying out a variety of different perspectives and giving each of them the opportunity to make their case in an overlapping narrative. Red Plenty has room for both the people who are excited about planning and the Hayekian argument against their excitement. He’s able to use the framework of the novel to give each of these perspectives. Get the weekly Five Books newsletter Wiener is doing something similar with autobiography. She describes this character who she meets, who is very clearly Patrick Collison, one of the founders of Stripe. She describes her dialogues with him, presenting him as someone who is bright and smart and compassionate and believes in the Silicon Valley ethos and is fully committed to it. He ends up being one of the billionaire founders of this world. Wiener clearly comes at it from a different perspective but she tries to give both of these perspectives a space that they can play out in. The result is a book that you can certainly disagree with and I’m sure plenty of people will disagree with. Still, it seems to me to present a much more recognizable version of the world that is being built than the more standard, ‘here is why Silicon Valley is the be-all and end-all of everything’ or ‘here is why we are all going to become puppet-slaves to the machine that is consuming us all.’ So the Anna Wiener book is a little like a modern version of Red Plenty , written as autobiography, because like Spufford she’s trying to talk about the collision between these abstract and delightful ideals of how information transforms everything and the real and complicated messy ways in which human beings actually work and live. We began this conversation by talking about political science and whether political scientists talk about information systems. I said basically no, not very much, but they should. I think political science and social science more generally ought to be entering these debates. We can’t leave it all to the engineers and the lawyers. In this interview, I’ve stolen liberally from work I’m doing with Marion Fourcade, who is a wonderful sociologist. Anything intelligent I say, she should get full credit for. Her idea is that you can think about all this as classifications and categories and feedback loops. Machine learning processes involve feedback loops where they are supposed to optimize on something and they throw out stuff, based on their classifications of the situations that they’re dealing with. They measure how people respond and then they adjust the classifications and so on and so on. But we also have the actual feedback loops of people’s lives, of the way people use these systems in all of these unexpected ways. What we don’t have much understanding of is how this intersects with the feedback loops of democracy. In this new information architecture, we don’t have ways for people to really express collective voice, or have collective input into these platforms that shape how we organize the world. So one of the interesting questions, then, is: can you start to think about ways we aren’t just interfacing with these systems as users, as consumers, as paid product—because our eyeballs are being sold to advertisers—are there ways in which we can participate by saying collectively, ‘No, we don’t like this. Yes, we do like that’? And how do you even begin to start building the structures that might allow this? Now again, there are ways in which you might start to think about using some of these new forms of information system as ways to enhance democracy. Glen Weyl talks about this a lot. The politician Audrey Tang is doing work along these lines in Taiwan. There’s also a professor called Federica Carugati at Stanford who is thinking about stealing ideas from the way that democracy worked in classical Athens to try to build up towards councils of randomly chosen people who might be able to have direct input into algorithmic decision making, like that constitutional convention in Ireland. How this would all work is hard to say right now. We barely even have the vocabulary to begin to articulate the questions. That is really where I would like to see the conversation between political science and technology going."