Get in touch

Constant tweaking and improving: This week on Track Changes we are joined by friend and tech writer Clive Thompson, to talk about his most recent book Coders: The Making of a New Tribe and the Remaking of the World. We chat about coder culture, its influence on society, and why the search for efficiency can sometimes be a double-edged sword. Paul and Rich also share their worst bug stories.

Transcript

Paul Ford Would you freeze your head, Clive? 

Clive Thompson No, I would not freeze my head. 

PF What about you, Rich? 

Rich Ziade Just the head? 

PF Well, that’s the thing—I mean, I know your head. I don’t think you should freeze it. [Music plays alone for 18 seconds, ramps down]. Rich? 

RZ Yes, Paul? 

PF People who write about technology. It’s an unusual category of human being. 

RZ They’re mostly nice people. 

PF Mostly. Some could do a little bit better. But we have one of the nice and more thoughtful ones here today. 

RZ We do. 

Clive Thompson Oooh. 

PF On the podcast [music fades out]. Clive Thompson, welcome to Track Changes. 

RZ Welcome, Clive. 

CT Good to be here and thank you for the totally delightful sign in. 

PF That’s what we do [chuckling in background]. Now, a little while ago you did a thing, namely finished and published a book. Something I seem to be incapable of [Clive laughs]. And the book is about something that we know very well: programmers. But it’s not called Programmers. [Clive laughs] It’s called Coders. What’s the subtitle? What comes after the colon? 

CT The Making of a New Tribe and the Remaking of the World. 

PF Oh my Lord! 

RZ Boom. 

CT Yeah, yeah [laughs]. Exactly! 

PF Ok. 

RZ Pow! 

PF Wow! Ok. 

[1:13]

RZ So what’s this book about? 

PF What the hell made you decide you wanted to write about programmers? 

CT Ok, here’s the reason why—and I’d been—I’ve been writing about technology . . . and its effect on everyday life for like 25 years. 

PF You started like the Canadian Wired, essentially. 

CT Yes! Yes! 

PF What was it called? 

CT It was called Shift Magazine. 

PF Shift! I remember Shift! 

CT Yeah, it was like Wired except weirder. In a good way. Like it was—you know they would write articles about just weirdos doing weird web art and stuff. 

PF Is it frustrating as a Canadian that everything has a comparative in the United States? Like McGill! Good college, it’s the Harvard of Canada. 

CT You know why they think that? Because it sounds private. 

PF Mmmm!

CT But they’re all public universities. 

PF That’s right because it’s socialist. 

CT Socialism, exactly. 

RZ Oh is that true? 

PF Mm hmm. 

CT Yeah, yeah. 

[1:53]

RZ McGill is not a private university? 

CT No. 

RZ Ahhh. 

CT It just sounds private. Yeah, no, there are these weird, hilarious analogues. I mean I edited this thing called This Magazine which is like Canada’s progressive magazine, and we used to say it’s like The Nation but funny. 

PF Canadian. Funny and Canadian. 

CT Exactly. So yeah we do wrestle with the constant we are the sort of take home game. If you can’t actually play Wheel of Fortune they give you the box game to take home. That’s the Canadian version. 

PF Ahhh! 

RZ [With Paul] Ahhh! 

PF But [Clive chuckles] so many good things come out. Literally, half the comedians that we love come from Canada. 

CT And deep learning. 

PF What do you mean? 

CT Oh it came out of Canada. 

PF Deep learning?!?

CT Yeah, sure. So dig this: they were working on deep learning—the big pioneers—

PF Tell the people what deep learning is. 

CT So deep learning is basically the kinda hot new form of AI where instead of writing down a bunch of individual instructions and rules, you know if this happens, do that. You basically set up a system and you train it. You say—if you want it to recognize cats, you show it a million cat pictures. When it gets it wrong, you feedback that information, and it readjusts all the little mathematical weights to say, “Hey, alright, don’t do that again.” When it gets it right, you feed that back in, readjust it. After a million of those, it can now recognize a cat perfectly. Right? That’s deep learning. The idea of that had been around for decades but no one really could get it to work. And they thought it was never gonna work. Except for the Canadian government that put a couple hundred thousand dollars into a few people—

[3:15]

PF Canadian dollars. 

RZ They asked for two million but the Canadian government was like—

CT [Laughing] Exactly! 

RZ “Listen, let’s take it easy with this.” 

CT Exactly! 

RZ “Ok?” 

PF It’s a stack of coins with loons on them and they’re just like [Rich chuckles], “Here.” 

CT It’s also Canadian dollars, right? 

PF Yeah. 

RZ Yeah. 

CT And so that’s all it took. They gave it to Geoff Hinton, long term guy had been toying around at the University of Toronto. Yann LeCun and a guy—I think it was at the University of Montreal? And, you know, they cracked it! 2012 they went to the Image Net Contest, the kind of a shootout for AI. And they killed it! And then Google is like, “Woah! Can we do that?” The rest is history. So basically all the deep learning stuff; all the AI that fuels 90% of what Facebook and Google are doing came out of the Canadian government getting like scraps of money—

PF Thanks Canada. 

CT Yeah, there ya go. 

RZ Nice. 

PF You know what’s great for me right now? Is watching another professional technology explainer and watching your brain work because you feel alone having to explain this shit to everybody. You’re just like, “Ok, I gotta make somehow deep learning accessible to an audience. Ugh, what the hell? It’s mathematical weights—ok . . .” 

[4:19]

RZ Cats. 

CT Cats. 

PF Cats, yeah [Rich chuckling], ok. And I’m just watching your brain turn and it’s like looking into a mirror. 

CT Yeah, I know. 

PF Cuz I have to do this too. And it—there’s a part of you that dies when you do it but there’s no other way through. 

CT I think we should actually find a bar and find like the nine of us that do this and [laughs]—

PF Just explain technology—

CT Only we can understand that horror but and this comes back to what you asked: why did I write the book? Well, I realized after 25 years of writing about software that I’d sorta spend a lot of my time saying, “Ok, you know, here’s how x, y, or z piece of software or social network is kind of warping and deforming and reshaping everyday life.” But the average person, you know, they could get that but they had no idea where software came from. 

PF Right. 

CT You know? It’s like as if it was just magic, like a UFO landed and just disgorged it. 

RZ Sure. 

PF Mm hmm. 

CT And they didn’t understand that, you know, because it’s made by humans there’s a whole ton of decisions made along the way. You know, what to do, what not to do. So I kinda wanted to say, “Alright, let’s pull back the hood and show a little bit of what that culture looks like and who are the obsessives and weirdos that decide it’s really fun to sit around all day long—actually I was gonna say telling machines what to do but there’s a wonderful, the best description of programming I ever heard was telling rocks what to think. 

[5:30]

PF Yeah, that’s right. Literally a chemical substrate that has to light up. 

CT So that’s why I decided to write the book, basically. I wanted to bring that stuff mainstream. 

PF Well, you know, there’s a mass—the mass media way of approaching tech is to just focus on incredibly wealthy people who run companies—

CT Right, yeah. 

PF And go like, “Mark Zukerburg, Bill Gates,” and talk about like what they’re up to [yeah] and somehow that explains technology. And it’s not. 

CT Well, you know, after writing this book, I mean, I tried to count it up. I think I—there’s around 200 developers I sorta spoke to, you know, ranging from—

RZ Uh huh. 

CT—hanging out with a bunch of them at a bar at a conference to like talking for hours to one person to profile them. And so one of the things I was trying to do is what are the actual common threads? And what makes ‘em tick? Why are they like this? You know why they get into it. And, you know, some of the things people think are kinda, you know, maybe kinda obvious, like, you know, they’re very logical thinkers! Well, you know, they have to be. You know, if you can’t systematically and logically, this isn’t gonna work out very well. So that was obvious stuff but what actually kind of [chuckling] intrigued was kind of the more unexpected things, right? I think actually, Paul, you and I might’ve talked about this at some point in time was I began to realize the more I talked to them that these are people who have an ability to endure a grinding level of frustration that’s like an order of magnitude higher than most normal human beings can handle. 

PF Yeah, my joke is always that if someone who works in this field, if they have a callus on their finger from reloading . . . 

CT Yeah. Exactly. 

RZ Mm. 

PF Over and over again. Like it’s just—you have to get into a zone where you can look at something five or six hundred times over and over and over again over the course of a couple of hours and just incrementally change or update it. 

[7:00]

CT Exactly. And that’s because, again, what people don’t get about the writing of software is they have all these images in Hollywood of someone sittin’ down and typing—

RZ Clanckin’ away. 

CT Yeah and it just pours out. And they don’t know what I think everyone working for you guys knows which is that actually what you’re doing is fixing broken stuff. Like literally something you just—a function you just wrote already isn’t working, it’s not passing its test. And then there’s a codebase you inherited uncommented weird cobal from 20 years ago and you gotta figure that out. So basically the daily life of most developers is just sitting around staring at things that don’t work and somehow being able to smash their forehead through the wall to get that—and I found that kind of existentially interesting. Like also what it does to you. You talk to the ones who are like 40 years in versus the ones that are five years in. 

RZ But even the stuff that does work, a good, curious engineer still thinks, “That’s not the best way to do it.” 

CT That’s right. That’s right. And that’s the flipside of the other thing I found is like, you know. Ok so there’s all this diversity definitely in the psychologies and personalities, introverts, extroverts, you know, all sorts of stuff. It’s not like the sort of—

RZ It’s not just the introverts. 

CT It’s not just he shoegazers. Exactly! There’s this idea that it’s all about a bunch of shoegazers. Not true. But what were the things that really were true? Nearly across the board. One of them was kinda what you just said which is this desire to constantly tweak and improve and render more efficient and more optimized. 

PF Oh God, they’re annoyed by everything. 

CT Yeah I have this chapter called “The Cult of Efficiency”. And it’s basically like everyone would tell me about how it happened early on like in their learning to program where they sort of realized, “Oh Christ. Computers are really good at taking these repetitive actions and this sort of annoying—things that are being done and speed ‘em up.” And then they just start developing this x-ray vision and they can’t turn it off. Like everything they see, they want—

PF Needs to get more efficient.

CT Needs to get more efficient, needs to get faster, right? You know like this one woman who worked for Zillo. I was like [laughing]—we were hanging out in San Francisco and she’s like, “Oh the way here, I was standing at the corner and watching people cross the street and I almost felt like screaming at them cuz I thought they were doing it in like suboptimal way.” Like—She’s like, “I’m outta my mind.” Right? You know. 

PF Eh. It’s San Francisco, I think people would welcome it. 

[9:03]

CT Maybe true. 

RZ This all makes sense. What else? 

CT What are some of the other high points? Yeah. Like—

PF What were you looking for? 

CT I was looking for the stuff that would not just help people understand the people but understand the decisions they make architecturally and engineering wise that kind of result in the downstream effects that we live with, right? And this is actually why the efficiency obsession which turned out to be unexpectedly interesting and productive as an explanation. Because the more I began to look at it, the more I began to talk to people, you begin to see that like that sort of desired, optimized, speedup tweak is a fractal across not just the great things that software gives us . . . but like a lot of the really weird externalities that we’re living with, right? You know? It’s like ok, Microsoft Word. I’m old enough to have written essays with a manual typewriter, right? 

PF Mm hmm. 

CT And that was [chuckling] a very inefficient process. It was gruesome. So, Microsoft Word comes along and says, “We’re gonna automate a lot of the blocking and tackling,” and suddenly I can actually think in a—literally think in a better fashion, a smoother fashion. 

RZ Sure. 

CT That’s your upside for efficiency. One of the down sides is, you know, companies that are now like saying, “Well, let’s render efficient basic, everyday activities, you know: the ordering of food; the ordering of cars; how we allocate—” 

PF Of people. 

CT Of people. Airbnb, like, “Here’s an empty room. Could we optimally match this with need?” And it all kinda starts off like you’re like, “Ah, that’s a great idea!” But then holy Moses when the rubber hits the road you begin to realize that it’s the everyday world of people doesn’t respond so well to being optimized all the time. And it’s kind of almost—I began to realize I’d stumbled into the ER problem of all software, you know? Whether it’s a newsfeed, you know, like the newsfeed was literally an optimization ploy. It was Zuckerberg saying, “I wanna make it faster for people to see more stuff of what everyone’s doing.” First blush that was great. Like suddenly you this ESP. You knew everything—the stuff that was going on in the world, it felt great. But, you know, they had to start algorithmically sorting that stuff. 

[10:58]

PF You know what it is? I think about this a lot. You start to see the world in data structures. And so you go, “Oh, everything’s a list.” And when you talk to people from the west coast and they look at media and think about, you know, what is The New York Times to the west coast? It is a feed. It is a list of articles in chronological order with tags for sections. If you say that to the people at The New York Times, they’re like, “You are missing—” 

RZ You’re an animal. 

PFall of the subtlety that goes into organizing and juxtaposing and looking for connections.” 

RZ How many times has some technological idea come forward to fix media? Again and again and then they start these consortions and then this little startup shows up. 

PF Is there anything that tech wants to fix more than media? Maybe [Clive laughs] health care. 

RZ No . . .

PF No, no, no. Like personal wellness is the other one. 

CT Yeah, personal wellness. And dying. 

PF Dying. Yeah. 

CT It’s the ultimate inefficiency, right? 

PF They wanna stop dying. 

CT Yeah, they wanna stop dying. 

RZ Is this all bad? Like Uber? Airbnb? 

CT Again, like, I think it’s frequently very good, right? Like I mean any economist would argue and many people in the world of culture would argue—

RZ Well, let’s not leave to the economists to determine what’s good or bad but—

[12:03]

CT All I was gonna say is that humans—the progress of civilization has been in creating ways to let individuals punch above their weight by doing more with less, you know? 

PF Right. “I can cut down a tree by holding a chainsaw up against the tree.”

CT Exactly, yeah. 

PF “That’s a hell of a thing, it used to take me like a whole day.”

CT Yeah, there you go. 

PF “Now I can cut down a forest. I can kill pigeo—”

CT [Chuckling] By yourself! 

PF “I can kill squirrels.”

RZ So, this is good. Is this good? I can’t tell if it’s good or bad. 

PF Good and bad are very reductive terms. Shit gets weird at scale and humans are a mess! 

CT Ah ha! Now scale! That’s the other thing I found which is super interesting which is alongside obsessions with efficiency, a lot of developers are super obsessed with scale, right? 

PF Oh nobody even wants to build anything for like—just for an individual to use and enjoy anymore. 

CT Right, yeah, exactly—

PF That’s trash. 

CT And you can sympathize with why, right? Like once you’ve solved the problem for yourself with software, there is almost zero marginal cost in rolling that out to the rest of the planet. And there’s something exciting about that. Like I—

PF Well, that’s also where all the revenue is. 

CT The revenue, yeah. 

RZ There’s also impact, right? You get a high out of that impact on the world. 

[13:07]

CT That’s the thing! I chatted with Ryan Olsen who is an Instagram engineer. Literally two days after they launched Stories, right? So we’re hangin’ out in the office and he was completely burned out because they’d gone on some, you know, three-week long march of not sleeping. But he was like, “Yeah,” he gets on the subway, the day after they launch, he’s lookin’ around and everyone’s like using Stories on the subway. And he’s like, “That feels—” He was literally like, “I can’t even understand how to process that feeling that I’ve done this thing.” 

PF Mm, and here it is. 

CT “And suddenly like people all around the damn planet!” And it was a small team. It was like five engineers working on that, right? 

RZ It’s probably the same feelings that—

PF And Snapchat did all the product management for ‘em. 

RZ Pff! Zing. 

PF Zang! Just a little memory goin’ back. 

RZ I think that’s power! That’s called power! That’s the same feeling a dictator feels on the balcony when he’s givin’ that speech. 

PF That is not how the engineers at Instagram feel themselves, like on a little balcony, goin’—

RZ I think the fact that you’ve [Clint chuckles] injected your opinion [yeah], creation into everybody’s lives in such an intense and profound way is very empowering. It just feels—

CT Totally, totally. Yeah, yeah. 

RZ Let’s put aside whether it’s good or bad or better for you or worse for you. 

PF It’s a wonderful feeling. I’ve had it. And you feel like—

RZ It’s a high. You get a high out of it. 

PF “I put my thumb on it a little bit. The world’s a little different cuz I’m here.” 

CT Yeah, totally! 

[14:25]

PF That’s a good feeling. Writers do that. 

CT Yeah, and frankly, again, like a lot of the time this has been I would argue to the benefit of the human race, basically. I mean really I think what I came to, I began to realize is one of the problems we’ve got is you asked, Rich, is this good or bad? The real answer is: we actually I have, I think, socially, culturally, economically, we have pretty good ways of recognizing the advantages of efficiency . . . and optimization. We have really no language that’s very good for talking about the occasional advantages of friction and efficiency, right? We don’t have as many paradigms for talking about that and so it’s why, actually, there’s this guy, Dave Guarino, terrific developer for Code for America, with his team, when they’re rolling stuff out and it’s affecting like, you know, food stamp recipients, you don’t wanna screw that stuff up. He was like, “Automate as little as you can.” 

PF Yeah, of course. Of course. 

CT “Like slowly figure it out.” He would do all these Wizard of Oz protocols with live products because he wanted to know—

PF What’s a Wizard of Oz protocol? 

CT That’s where like the software looks like it’s doing it but really a human’s doing it. You know the first thing was—what they were handed was this 50 page nightmare of an unresponsive site for applying for food stamps that no one could actually use on their Android phone. They couldn’t go in and tear it down. So they had to build something and put a layer of crud on top of it, right? 

RZ Mm hmm. 

CT So they basically just put a little responsive page . . . and they would ask people like five questions that would give them 80% of what they needed. And they would get that, they would manually enter it, and it would call the person up, and they would finish the process. So it was Wizard of Oz, like the person thinks they’re using software and they get a phone call from their team. And they slowly figured out how to automate the rest of it. And his goal was move slowly and only automate what you need to when you need to. Produces—

RZ Because you can cause damage if you go too fast. 

CT You can cause damage if you go too fast. 

PF This is real. I remember talking to someone who worked in juvenile justice and they had put a default setting in that ended up keeping a kid in jail overnight. Cuz it had sort of rolled around 24 hours. And it was like not the greatest tragedy in the world but it was extremely serious. 

RZ Yeah. 


CT Sure. 

[16:21]

PF Like they were like, “Ok, we have to address this immediately.” And it was like full-stop moment because the ramifications are so real. Like it’s not like, “Oh boy! They didn’t get their pictures in a timely manner.” 

CT Yeah, yeah, yeah. 

PF It’s like, no, a kid stayed overnight in a juvenile justice facility. 

RZ Most of the time technologists don’t bother, they just go. [Clint laughs] And they’re like, “Let’s see—woops! What did I just do to Syria?” 

CT Yeah I mean the flipside to the power of software to like have that big impact is yeah you can [laughs]—you can do something pretty bad pretty easily, right? Some of my favorite things in the book was actually just asking people, “So what’s your worst bug story?” 

PF Mm hmm. 

CT Oh my God! It was so much fun. I could’ve just done an anthology of the worst bugs—

RZ I bet. 

CT—cuz people had—what’s the worst bug you guys have had? 

PF Early days I destroyed 30,000 fantasy sports accounts for basketball fans. I didn’t know how to update the database and I shouldn’t have been in that position [Rich laughs under breath]. This was like ‘97. 

CT Yeah. 

PF And we were able to roll it back. I had to send a lot of emails and there were a lot of people around me who had delegated to me going like, “Ok, ok, ok. We knew you shouldn’t have done this and we let do it anyway. Ok. Ok.” 

RZ Money bugs are the wo—[mm!] it’s like, “I’m building a transactional system that’s touching bank accounts and whoops! It’s just like one line off. But that one line—the cascade effect is huge and when you’ve gotta think about all of that, right?” 

[17:43]

CT I will say that of the category of my worst bug, oh my God, 80% of them are database. Like, “Oh my God, I did something and it wasn’t backed up.” 

PF No, it wasn’t transactional—

CT It’s basically gone. Yeah, exactly. 

PF That’s the thing is if you follow the rules, the rules are actually set up like, “Turn on backups, use full transactional behavior in your database so that you can roll things back, and—” If you follow the rules, you actually can’t get into too much trouble [no] because everybody makes the same mistakes but no one will follow the rules. 

CT Some of the bugs are kinda fun though. Like where it’s like it’s not fatal, like nothing’s getting destroyed but they just—they didn’t anticipate a form of user behavior and now it’s super weird, right? So, Lance Ivy, early architect of Kickstarter, basically kind of built the frontend, backend all by himself. As he told me, “I should’ve let someone else help me for a long time.” You know that type of thing? 

PF Yup [Rich scoffs]. 

CT And their first one kind of million dollar thing blew up. Right? So it’s going to a million dollars and this was kind of exciting for the early Kickstarter. 

RZ Sure. 

CT It was some guy, I think he was doing a video game or something. The campaign was getting there and everyone was having these like—these moments where they were like—all the supporters around the world were all sitting there and refreshing the page non-stop cuz they wanted to see it go roll to a million. Well, unfortunately, that was essentially a denial of service attack on Kickstarter with all these people refreshing. And so the server’s [chuckles] like red lining and red lining and it’s going down and they’re taking it back up! And going and taking it back up. Cuz they hadn’t yet made a really simple refresh that didn’t require to refresh the whole page, right? 

PF Sure. 

CT Nowhere did they think that the cultural behavior of Kickstarter was gonna be, “Oh my God, I wanna see it roll to a million.” And so that was a terrifying night for them, right? So there’s a lot of fun stories like that. [Rich coughs]

[19:21]

PF You know the thing that I think about like nothing is easily good or bad? Is agriculture. People are like, “Everything must be organic.” And it’s like, well, if everything’s organic, this whole continent’s gonna starve. 

CT Yeah. 

PF This kind of modified rice is what people in order to sustain itself. 

CT Cuz we can do it at scale, basically. 

PF Yeah we can do that at a scale and it works and it is efficient and it’s good to go. At the same time, there’s all sorts of things with giant modern agro-business that aren’t good for people, right? And so it’s like balancing that out—but you can’t sit down and say, “Everything must be organic.” 

CT No, exactly. As with efficiency where you kinda need a new language to talk about, “Alright, so when is inefficiency good? When would you wanna actually engineer it in?” The other question is: when is scale good and when is it just gonna break stuff? Right? [Mm hmm] That’s another thing that we don’t have a lot of language to talk about. What things are like massively improved by scale? What things are kinda wrecked by scale? And I think what one thing—the ER lesson of a lot of social media is certain forms of human conversation just break when you try to do ‘em at scale basically. 

PF Yeah, no, it’s true. 

CT I mean I’m sure you guys find this all the time in the work you’re doing in media, right?

PF I mean it’s an easy one, right? Cuz Facebook is fine when you have 20 friends and with 700 people that you’re connected with through all of your different aspects of life, some professional, some family, it’s almost unusable. It just locks up. 

CT And it’s kinda funny, like you talk to Twitter people who’ve been around since the beginning, and they’re all like, “Ah, the first two years was great.” Well, that’s because it was small scale. And sometimes I’ve wondered like is—is maybe that just—maybe that’s like an iron law of social media. Like you just need to keep churning out new ones because the first two years when it’s small scale it will be great and then it’s just gonna be like a tire fire afterwards. Maybe that’s just the way it goes. 

PF It’s a couple hundred people with similar inclination. Twitter is now an ongoing national referendum on absolutely everything. And so, you can’t really participate in that unless you’re willing to be part of the referendum. 

CT Right, right. Yeah. 

RZ You can observe. That’s about all—

[21:15]

PF Observe—

CT You can lurk. 

PF You know, I put out silly jokes and you know just sort of have my fun there and promote stuff. But like you can’t—I don’t wanna get into the conversation cuz it’s just endless. 

CT Well, and you created that little temporary community, right? What was that? 

PF Tilde Club. 

RZ Tilde Club. What does that mean? Tilde Club?

PF And this is kind of a deep cut from the Paul Ford archive. Tilde Club is a single Unix server with lots of different accounts. You can have an account on any computer, right? You know how if you’re using a Mac and it’s like, “Here’s your guest and here’s you.” That would actually scale to 500 people. You could plug that into the internet and you could say like, “Anybody who wants to connect to this,” and you could send each other mail just on that machine. 

RZ You’re getting a command prompt to a Unix box. 

PF This is built in to every computer [ok] in 2019. It’s just—

RZ So what are people gonna do with it? 

PF Oh you can make webpages, have accounts, send email to one another. 

RZ Got it. 

PF All the things you do on a social network are basically—

CT Right. 

PF Facebook and Twitter and everyone—those are all riffing off of things that used to exist 25 years ago. 

CT Right. 


RZ Right. 

[22:12]

PF It’s built into the system. 

CT It’s a tiny social network run by Paul Ford. 

PF That was the original form of it and then it kind of blew up and then it imploded. 

RZ But these are small and quaint and nice but like obviously it reaches a point where if someone feels like, “Oh my God, I can actually influence other people.” 

CT They want scale. 

PF Well who doesn’t? You’re a writer, right? 

CT Yeah, yeah. 

RZ That’s success! Scale is often associated with success. It’s the hockey stick chart, right? 

CT Yeah, yeah. 

RZ Or whatever you wanna call it. Do they feel bad? . . . You went into this coder world—

CT Do they feel bad about what happened? 

RZ Well they screwed up the world! [Clive laughs] The world is a mess! It’s a sloppy mess. We can’t have elections anymore and it’s their fault! Cuz they thought it would be really cool to give everybody a newsfeed. 

CT They’re divided down the middle. Half of them feel bad. They are self-reflective about it. They’re like, “Oh, wow, we were naive.” 

RZ Maybe optimistic about how humans work. 

PF I think Twitter in particular. The Twitter crew went on and a lot of them looked in the mirror and went—

CT right. 

PF Like touched—

RZ “What have we done?” 

[23:07]

CT So that’s that half, the other half they’re being paid too much money to confront it. They have too much power. It’s like—

PF Oh yeah. 

CT—that old—you know, you can never convince someone of a fact if their job requires them not to understand it, basically. 

PF Right, yeah, it’s like Upton Sinclair or H.L. Mencken or something—

CT That’s it. Yes. 

PF One of those. 

RZ Just denial [Paul sighs]. 

CT The half that I think are like the ones that Paul said look in the mirror and kind of going, “Oh man,” and they have like a sweat mustache. 

RZ Mm hmm. 

CT They kinda got out and they’re doing something else and they can be reflective about it. 

RZ Oh they got out?!

CT Yeah—

PF That’s critical! When you’re in there—

RZ They left the tribe. 

CT And the other ones that are there or moved on to like another company and are kind of—they have a lot more trouble, I think, being—talking turkey about it. 

[23:45]

PF I think that you really do believe, if you—there is a price point for someone who—for everyone to believe something. Like there is a price point—I could make you believe all sorts of things, Rich, for the right number. 

CT Oh my goodness. 

RZ Mm. I think that’s pretty much Postlight’s mission statement [laughs, Clive joins]. 

PF That is our mission. It’s a high number. Mine was a lot lower when we got started [Rich laughing], it keeps going up, as I get more and more experience at running a company with you, I’m like—cuz it used to be like, “Paul, here’s 50 dollars. Please believe everything about capitalism is good.” And I was like, “Ok!” [Others laugh] And now—now I’m like, “Mm. It’s gonna take a lot more than that.” 

RZ Do they feel like they kind of . . . not just feel bad but almost got it wrong? I mean look—

CT Yeah, yeah, yeah . . .

RZ Yes, we can recognize the cat. I think that’s awesome, the computer can recognize the cat but there are still, you know, thousands of people in the Philippines looking at pictures of images posted on Facebook cuz they can’t solve that one. 

CT Ok, here’s where I would take that: obviously I spent a lot of time doing like a post mortem on social media because people care about that but if you wanted to say right now . . . who are the engineers who are looking in the mirror or should be looking in the mirror, and going like, “Ok, what are we doing?” It actually is a lot of those deep learning people cuz they’re workin’ on systems, you know, anyone that’s working on facial recognition at this point in time . . . should probably be looking in the mirror and going, “What the hell am I doing?” 

RZ “What am I unleashing?” Yeah. 

CT Exactly, yeah. Like stuff like that where like—

PF Well this is also one where it’s not like they can look back ten years from now and go, “How could we ever have known?” 

CT “How could we ever have guessed?” 

PF Yeah. Everyone’s very excited to apply this technology in all kinds of places where it’s gonna mess with civil liberties. 

CT Yeah, yeah, precisely. 

[25:18]

RZ But the train’s left. 

CT The train has left and so some people say, “Well, I don’t wanna necessarily roll the state of the art forward but in China . . . they are rolling it forward gleefully and rapidly with government money and with government having these datasets of billions of shots of faces that they hand back and forth between industry.” They’re going to be the world masters at this stuff. 

PF Yeah. 

CT We should at least know how to do it as well as they do. 

PF A moral slip in an arm’s race: if we don’t have it, they’ll have it. 

CT It literally is like nukes. 

PF And frankly it’s not necess—it’s not wrong. It seems very likely that China will continue to invest vast resources [yes] in good and efficient facial recognition. And so as a nation-state do we have a response? Right? And now you’re like, “Well, here we go.” Like we need to do a congressional hearing. And Amazon’s gonna show up and say, “We’re ready to help.” 

RZ “And don’t you need socks?” [Paul and Rich laugh]

PF “You looked at [others laughing] facial recognition in Congress, maybe you need a new weapon—” 

RZ “A hat.” 

PF “—and a hat.” 

RZ “You need a hat.” 

PF Oh God. It’s gonna start selling us congressmen [Rich laughs]. 

CT If the EU and the US both jointly banned face recognition in like a bunch of consumer cases, that would have a strong global effect [right] because if you can’t sell that stuff here, then that would curtail some of China’s growth stuff. But that type of coordinated political action is difficult to envision. 

[26:40]

RZ It’s getting harder and harder to get that kind of coordination. 

PF It’s not a moment where you could get a quick outcome. 

CT Exactly. 

PF Well, that’s good—

RZ Tell me something good. Gimme a feel—let’s close this with a feel-good story, Clive. 

PF That’s right—it’s a marketing podcast. 

RZ “Coders are charming!” And then it was like, “We’re goin’ down!” 

CT Here’s some fun stuff: one of the things—people—I’m writing this book about coders and a lot of people who are not developers would say, “So should I like learn this?” 

PF Yeah. 

CT “Is this the type of thing where I like should give up my job in media where everything is dying—” 

RZ We hear that a lot. 

CT Yeah, exactly. I’m sure you do. And so the answer that I sorta came up with is if you’re beyond a teenager, you’re probably not gonna shift your career around this type of thing. You’re not gonna work at it hard enough but if you’re an intelligent person with a bit of free time and the ability to apply yourself, you can kinda learn a little bit of it. And it turns out that knowing a little bit is weirdly fun and even becomes quickly useful in your personal life. So, you know, I learned a lot of Basic and stuff when I was a kid. I didn’t do much programming in my twenties or thirties. Then in my forties, I started doing more of it so I could talk knowledgably to the coders I’m talking to, but just like stuff for like hobbyist projects but oh my God I started discovering all these weird, fun things that were useful for me. So like my book comes out. And what does an author do when their book comes out on the first day? 

PF Usually they sink into despair and [Clive laughs boisterously] and start to panic and no one will tell them how the book is selling. 

CT Ok and because they won’t tell you how the book is doing, what you do is you sit there . . . at your Amazon page, refreshing it every five minutes to see if it has appeared on any best seller lists, right? 

PF God! Why would a human being do that, Clive?! That just sounds—

[28:13]

CT And then you do that every five minutes for the whole day until you pass out like in the evening. And then the next morning you show up, or at least I did, going, “I’m not gonna do that. I’m gonna actually get work done.” And of course every five minutes, you refresh the browser. So after my third day—solid day of this, I’m like, “Ok, this is ridiculous. It’s also a repetitive behavior that I can outsource.” So I just wrote a quick scraper: scraped the page, find any information about the lists I’m on and text it to me. This robot will do it for me and sure enough it kinda cured me. I stopped doing it. And so what I tell people is like, “You’re probably not gonna quit your job and become a developer . . . but it’s pretty easy to learn just enough Python to do that.” And then there’s all these weird little fun things you can kinda do in your daily life. And it also kinda give you—Paul and I have talked about this, it’s more fun than writing. It’s probably more fun—if you’re an accountant, it’s probably more fun than being an accountant. If you’re in hospitality, it’s probably more fun than being in hospitality. So it’s sort of a fun thing to do with your brain

RZ It’s very satisfying. 

CT It’s very satisfying. 

RZ A little game. Yeah. 

CT Exactly, yeah. Like there are a few things—

PF And it’s never done. It’s never done. 

CT It’s never done. 

PF You can always make it a little bit better. 

CT Always make it a little better. So that’s—this is the craft and the art and the delight of programming that I encounter with all the coders I talk to. That I—I do just enough in my daily life to really enjoy and treasure. Like I actually look forward to it. And that’s—if you wanted an upbeat note to end on: I tell people, “Go out and learn a little bit of Python and see what that feels like.” 

RZ I love this little bit of advice. 

PF Nothin’ wrong with that! 

RZ Nothin’ wrong with that. 

PF The whole world might having—it’s goin’ through a pickle but you can still have a little Python. 

[29:39]

CT Yes. Get your little robot doin’ what you want it to do. 

PF There’s actually a pun in there, you know that? And I didn’t mean there to be. The Python serialization format is called Pickle. 

CT [Jinx with Paul] Pickle. 

RZ That’s really funny. 

PF Ha! 

RZ That’s really great. 

PF Some great material. Well, on that note—

RZ We tried to end on a high note and then you had to throw that in at the end—

PF—just threw this whole thing out the window at this point [Clive laughs]. 

RZ Very cool. 

PF I don’t even know who I am [others chuckling]. Alright, Clive Thomspon has written a book called Coders. It’s been out for a while. He’s here as a friend, not really to market and it was great to have a nice conversation about what in the world is happening with technology and programming—

RZ And the world of coders. 

PF Oh my goodness! People can just buy this thing on Amazon. Where would you prefer they buy it? 

CT You know, actually, I say buy it anywhere that you’re normally buying books. 

PF That’s right [music fades in]. Ok.

CT Just don’t sweat it. Although, if you do buy it, I’ll know because my robot is gonna text me my sales rank every day. 

[30:32]

RZ Clive, thank you so much for doing this. This was great. 

CT Yeah, it’s great to be here. 

PF Rich!

RZ I love coders. 

PF Yeah, me too. I like making the thing. 

RZ I’m ok. I’ll forgive them. Just ok, so what? 

PF I’m increasingly accepting that even though I don’t get to make the thing, creating an environment in which good software can be built is a valuable use of time. 

RZ It’s a beautiful thing. 

PF Yeah, it’s really important. But I gotta put my ego away. I don’t get to make it anymore.

RZ Well, we get to work with great people who make it. 

PF We do. We get to work with better programmers, product managers, and designers than I ever was. 

RZ Than we’ll ever be. 

PF Yeah, yeah, that’s true. 

RZ And they work at Postlight, Paul! 

PF They do and you know what? You can build a relationship with them over time because we are your product partner. 

RZ Yes. 

PF At Postlight. The three Ps. 

RZ Reach out! hello@postlight.com

PF All you gotta do, you gotta send an email and we will read that email and do the best we can to help you. We wanna make it work. So get in touch if you got something to build and if you’re looking for a company that can really stand by you as you build and ship your products and your platforms over time. That’s us. Let’s get back to work. 

RZ Have a great week! [Music ramps up, plays alone for five seconds, fades out to end.]