Ezra Klein is one of the Internet’s most influential journalists. I asked him about the state of the media and why he doesn’t write with AI.
Some lessons from our interview:
Conversations are where you explore ideas. Writing is where you make them rigorous.
Sometimes you don't need to develop your taste, but have more faith in the taste you already have.
Some writers fade into the background to capture objective truth. Others (like Hunter S. Thompson) crank up their personalities and become the story itself.
Beware of reading summaries instead of the actual book. It's more dangerous to think you've read something that you haven't than to not read it at all.
Writers who outsource their learning to AI operate on a flawed model of how the mind works. They think people can download information like you see in "The Matrix" but that's not how people learn.
So how do people learn then? They need to grapple with information. And that takes time.
People will say that books should be articles and articles should be tweets, but the truth is that a lot of information should actually be longer. There are ideas you can only explore when you let them breathe, and sometimes that means a 300 page book.
Do the work that others skip: Good journalists are the ones who actually read things like the Congressional Budget Office report instead of skimming the summary.
Say it straight: Journalism doesn’t need rhetoric. It needs clarity. And directness. Say what you're trying to say without embellishment or jazz fingers.
Vox’s cautionary tale: Ezra co-founded Vox, which did some of the best journalism in the world during the 2010s. But their business model was unsustainable because they were yoked to advertising, and the idea that competing on platforms like Facebook would convert scale into revenue. The problem is that the platforms took the money. When they did, Vox (and other media startups like BuzzFeed) weren’t able to keep their top talent and stay at the frontier.
Why has the New York Times business been so resilient? They built a digital bundle. Games, cooking apps, and sites like Wirecutter fund their journalism. This is how journalism always worked. In the age of print newspapers, international bureaus were funded by classified ads that told you there was a sale at the mall.
Transcript
Table of Contents
The Genesis of a Journalist: Early Mentorship and Voice [00:00 - 06:30]
Crafting a Narrative: From Columns to Books [06:30 - 10:04]
The Intuitive Nature of Taste and Editorial Trust [10:05 - 13:00]
Journalism's Evolution: Digital Media and Institutional Roles [13:01 - 26:04]
The Business of News and the Value of Deep Work [26:04 - 37:54]
The Role of AI and the Embodied Nature of Knowledge [37:54 - 48:49]
The Nuance of Communication and Writing with Uncertainty [48:52 - 54:08]
David:
This is my interview with Ezra Klein, who's one of the Internet's most influential journalists. He co-founded Vox in 2014, and it was there that he really pioneered this new era of online journalism.1 Remember those Vox explainer videos that were so popular?
As we recorded this, he's got the number one book on the New York Times bestseller list. It's called Abundance. We talk about the state of media, why pieces need to be longer, not shorter, how he reads to understand the world, and why he's skeptical that AI will ever radically change the craft of deep and meaningful journalism.
When you were at the American Prospect, you had a mentor, and he said, "Ezra, pick up the damn phone." What does that mean?
Ezra:
This is my first editor in journalism, Michael Tomaski, who's currently the editor of the New Republic and of Democracy, a journal of ideas.2 He's still out there preaching the gospel.
I've been a blogger, and started blogging in my freshman year at UC Santa Cruz. As a blogger, you just sort of spew, at least at that point if you were me, spewed your opinions out onto the Internet. Then I got this fellowship, my first job at the American Prospect.
I was making $21,000 a year, and it was, as my dad always said, like a grad school, but they paid you a little bit. I remember I turned in the first draft of maybe my first feature piece to Mike, and he said something at the time that felt very harsh: "This isn't even a piece of journalism. This is only you. You have to pick up the phone. You have to be talking to people. Part of journalism is integrating voices and thoughts that are not your own."
Even now, I'm an opinion columnist, and so within the structure of journalism, I am allowed to just sort of spew my opinions out. But most of my columns are filled with quotes, filled with voices that are not my own, in part just because I'm just not that interested in what I think. I think these are so much richer when they're this panorama.
David:
I was amazed as I was prepping for the interview how much your voice has changed. Early on, you really just let it rip with your voice.
Over time, I want to say it was around probably early 2010s, I don't want to say you toned it down, and I don't want to say it's more refined, but it became something more professional maybe. But what was that?
Ezra:
It's not that I don't do this for fun. I'm not saying it isn't sometimes fun, but it's not my hobby. My voice carries weight.
I treat it professionally, and I treat it as a responsibility. Still, honestly and authentically, it's my voice, and I try to be pretty true about what comes into it.
I remember I forgot who I was having an argument with on the Internet. I think it was John Shade, who was then at the New Republic, and he said something to me I thought about a lot after that: "You don't realize you've gotten big and your punches will really land."
I took that to heart. I wouldn't do the kind of takedowns that I did when I was a young blogger. As I got bigger, part of it would just be punching down.
Part of it is that I'm more interested, or I try to be more interested in construction than destruction. Around 2010, two years before that, I moved to the Washington Post. Now I work at the New York Times.
When I was at Vox, I sort of embodied Vox as one of its co-founders. I write with more institutional weight. More people are on the hook for what I write, and so I treat it more carefully.
One thing I always notice is I've gotten really bad at writing short. I used to be really good at writing short, just sort of throwing something off the cuff. In part, that's because I feel a need to make sure that what I'm writing feels like it justifies itself, that it isn't just me shooting from the hip.
On the other hand, I do plenty of podcasts like this, which are just me shooting from the hip. So maybe I've let that part of myself unleash in other ways.
David:
How does working with different editors impact your voice? You're at the Times now. I don't know if there's an institution in the world that has probably more weight than the New York Times.
How does that impact your voice? Do you feel like the New York Times brand sort of becomes a lens through which you just naturally write?
Ezra:
I don't write through the New York Times brand. That isn't a weight I feel. I don't think it's what they want from me.
My show, though, is called the Ezra Klein Show, not the New York Times Opinion Show. I'm on the opinion side.
My editor there, Aaron Retica, is like an institution and a genius. And the thing that I have with Aaron is trust.
The most important thing for a writer to feel about their editor is that their editor knows what is good and what isn't. I think people think of editing as a technical task. Like you're going through the sentence and is the grammar off, and could it be better, and is the argument clear? I barely give a shit about any of that. I'm a good writer. I can tell you if my sentence is clear.
What I need before something pretty hot goes out into a very big world where there's a lot of attention on it is to believe that the other person who saw it has really good judgment.
The thing that is hard to find is taste. The thing that is hard to measure is taste. The most important thing is taste, both with writers and with editors.
On some technical level, if you have an intuitive sense for what is good and what is not, you can probably upskill to get there. The biggest problem is not knowing what is good and what is not, and it's a very hard thing to teach, in part because it's subtle, it's textured, it's different from person to person. It doesn't need to be the same.
I'm not saying there's one objective good, but you can have two editors saying really different things. Who's right about what is good is what separates the editors.
David:
So, as you think about being an editor, editing others, do you feel like you have a sense of what you're going for that's explicit in your opinion pieces, or is this more of an intuitive dance that you and your editor are feeling out together and that's part of the relationship?
Ezra:
No. I know when I'm writing a piece what I'm going for.
David:
Okay.
Ezra:
By the time I am writing, I've done a lot of work.
David:
No, but I mean just with your voice in particular, like what an Ezra Klein opinion piece is.
Ezra:
I don't spend time thinking about that. If an Ezra Klein opinion piece had more variation than it currently does, I would think that was good. If I was writing more experimentally, if I was taking more flyers, what keeps me from doing that is not a straitjacketed belief about what my voice should be. It is an absence of time.
So I don't go into things and think, this needs to be an Ezra Klein piece. I go into things and think I am trying to execute this piece. It is in my head. And pieces that are in my head tend to be, I guess, Ezra Klein pieces.
It's hard to get away from yourself. But I would really welcome more capacity to play around than I have, but the limiting factor is time and energy.
David:
Do you feel like you have that with a book?
Ezra:
The limiting factor is time and energy.
David:
Okay. Okay.
Ezra:
I don't take two years off to do books.
David:
Right.
Ezra:
I got two months of book leave on this. I was going to do four months and October 7th bisected the book leave. I think if I do a book again, I will only do it if I think I can take six or eight months, which I don't know how I would give myself. I'd have to be in a very different point in my career.
I'm proud of the book and I'm proud of the craft of the book. I think the book is, in the parts where I was lead writer, the best writing I have done at a craft level. It's a step forward for me, but something gives and I seem to be good at making it, not the work, but it doesn't mean nothing is giving.
David:
When you say the craft is a level up, what about it?
Ezra:
I think the book has a very different cadence. If you read the introduction, it's almost a march. I remember throwing out a number of introductions and realizing that the voice of this thing had to be quite different. I think it's been a real step forward in my own work, marrying reporting that is extremely granular, more granular than what you normally see in my profession, which I'm saying there, opinion journalism.
If you look at what I'm doing in that book or in the columns that led to some of it in the Notice of Funding opportunity for the CHIPS Act, or how the Tanahan housing development worked, or what was going on in Prop H in Los Angeles on affordable housing and how people had to stack tax credits, I'm on the one hand getting extremely granular than to make some highly generalized points. And I think that marrying of the 30,000 ft and the three inch is very effective.
But this book is a manifesto I have not really written. If you look at my first book, Why We're Polarized, that's a highly descriptive book. That book is almost completely non normative. And the part that is normative where I'm saying, here's what I think we should do, you can almost read that I'm doing it under duress. I have this whole thing in the concluding chapter where I basically say, I don't really like concluding chapters.
And this book is much more prescriptive. It's much more the way I think the world should be. It's descriptive too, of problems, but it is pointing towards, you know, the output of Why We're Polarized, so to speak, is a model of polarization. The output of abundance is a different vision for how to do government and how to describe what liberalism is pointing towards. And, you know, that required a very different kind of writing.
David:
Well, earlier you used the word kind of hot when it came to thinking about writing a piece. And even abundance has that kind of hot energy of there's like a tempo to it, the kind of tempo that has ideas spread. There's a kind of writing that actually leads to that.
Ezra:
I think that's right. I think it's also ideas. I think it's having a sense of texture, of the moment or zeitgeist.
You tend to know when you're writing something that is going to Spark. I'm not 100% right. I've definitely had things I'm like, this is going to be big, and it wasn't. And I've had things where I thought I was just putting out one for me and it blew up. But I had done enough.
I had watched enough around abundance catch fire already in columns and Derek's original piece on it in the way that I already saw it being taken up by nonprofit groups, by politicians that I knew that, like, I was plugging into a live socket. Writing with Hot Energy and Abundance
David:
Yeah, that's a nice way to put it. So early in your career, I guess you'd wake up every single day, you'd read like 60 news stories before you'd start writing. I guess your assistant would send you these stories, and then, boom, you'd kind of get right to it. Do you think that that sense of t taste was honed by volume? And if that doesn't account for all of it, what were the other factors?
Ezra:
I don't think anybody can describe their own taste. I've thought about this a lot.
It is the nature of podcasting to ask people on some level who are at least by some definition of what they do, good at what they do, how they do it. And it is one of my observations or learnings as a podcaster that they're the people least capable of explaining it.
Because my taste is intuitive to me. Why does something spark for me and something else doesn't? I don't think I could quite always tell you. Why do I read something and think, ah, there's a conversation, and I can read something similar and not see it?
It is honed by all that. Part of the honing process was coming to trust things I already felt. One of the key things for me as an editor, that took a very long time. I wish it had been true.
I think my judgment was good when I started Vox, but I didn't trust it, and I wish I had trusted it more. So I think part of taste is not always its development, but your faith in it.
Think about artists, not true for all of them, but a lot of them do their best work right at the beginning of their careers. You think about musicians, you think about a lot of writers. Sometimes their most vivid, vibrant work is at the beginning. And I think that can often be, because in those professions, often the first thing you do is the most yours. After that, the audience is in your head. The editors are in your head. Living up to yourself is in your head.
For me, at least as an editor, which interesting, I think is more what we're talking about here, I was much less willing to credit my own editorial instinct when I was younger. I felt like I needed external validation. Oh, we're doing this because I can see it's popping on Chartbeat. I know that the rest of the media is covering it. I have some reason to believe that the place I'm charting us towards is a place that it makes sense to go.
Whereas certainly now, on my own show, in my own columns, I have much more faith that even if where I'm charting us is weird, that if I really feel it, it's likely to be the right answer for my show. That would not be, though, of course, how you'd want to edit the entire New York Times.
David:
So you were talking about Vox and how at the beginning you didn't trust your editorial voice as much. Tell me about that.
Ezra:
I just think that there were things I trusted. I mean, Vox was built on an idea of what it would mean to do explanatory journalism.3
I definitely trusted my view that we didn't do enough of that in the media. I trusted my view of how to hone that.
If I could go back as an editor, I would say no to a lot more. I would just say, this isn't right. This doesn't work. We shouldn't be going this direction.
My ability, it's not so much what I said yes to as what I should have said no to. And it's not that it's any one kind of thing. I just know that I wasn't that confident.
And when you're not that confident, it's hard for other people to know what it is you really want from them.
David:
What had me really excited for this interview was if I was thinking about the last 20 years of the Internet, I feel like there's one way that you could just understand how journalism has changed by just following your career. And Vox, there was like 2016, 2018, like Vox explainer was another level.
I remember the Barack Obama interview you guys did with that dark background with the grasping in front.
Ezra:
Joe Posner, video genius, yo.
David:
Like, I saw that and I was like, this is something from the future you don't usually see. It's sort of like the car race ahead that much.
Ezra:
Yeah. I mean, one thing that's always hard about having, I think, any kind of editorial innovation is that if you do a good job with it, it gets fast followed so quickly that then it's hard to maintain that differentiation.
Vox Explainers, which we did across many, many different formats, were really good thought that then every other organization began doing, which is good in a way. It's a way you have influence over your profession, which ultimately I think most of us want to do.
Then also you're really good people get hired by organizations with deeper pockets than you have. I mean, the difficulty of digital media in that era is we were yoked to a business model that didn't work. Advertising.
Not just advertising, but the idea that if you competed on all these platforms, Facebook, et cetera, and you built these huge audiences, you could eventually turn the scale into money. The platforms took the money.
And so the idea that more audience would lead to more money, which would lead to more money to pay journalists and fund journalism, didn't happen. It happened to some degree, but we did ultimately turn print dollars into digital dimes.
When you say that the car couldn't keep racing forward, I don't exactly think it's because we ran out of ideas. It's because our ability to keep hiring at the rate that we were being poached, it just couldn't keep up.
Now Vox is still there. It's still doing great work. Vox Media is still there.
But when I think of what went wrong for that class of digital media, the editorial was supposed to feed into a business model that was supposed to throw off more and more money that would allow you to keep up leveling your newsroom. If you can't keep paying the talent over time, you will degrade.
David:
How much original writing were you doing versus being an editor in chief?
Ezra:
I did a lot of both. I probably wrote two to three pieces a week at Vox. I also did the Weeds and the Ezra Klein show at Vox.
One of the problems for me as a manager is that fundamentally, I'm a writer and now a podcaster. I remember coming home one day and talking to my wife and saying, I don't think I got anything done today.
She's like, well, tell me about your day. And I told her about all these meetings.
And she sounds like it sounds like you were managing all day. I'm like, yeah, but I didn't get any writing done.
And she said, this is not going to work great for you if you can't in yourself feel that managing is a day of work. On some level, I never felt it, and it's why I don't think I'm cut out in the long run to mainly be a manager.
For me, what feels like a day of work is reporting, writing, podcasting. It is the creation of the direct editorial product. It's not the management of the people creating the direct editorial product.
David:
How is your consumption changed over the years? Because I do just get the sense that you started off just reading absolutely voraciously.
And do you think that that's the best way to understand what's happening in government and policy and society and economics, or has that changed over time in terms of what I guess is trying to get a model of reality that's actually accurate?
Ezra:
I don't think there's any one best way. There are ways that different people think and learn, and I combine a number of them in my own work.
But I know people who aren't super heavy readers, but they're amazing at being in touch with everybody at all times. They're better reporters than I am, but they're always on the phone, they're always making calls, they're always texting. Their model of reality, or their model of what they're covering, or their source of new ideas, is coming from that.
They're not sitting around reading the history of liberalism that I happen to be reading right now, but they are talking to people that I'm not talking to. So everybody has their own process.
For me, reading is the wellspring of all the rest of the work. If you said you could only have reading and reporting, I'd keep reading.
When I am reading books or think tank papers or academic reports or articles printed out in a fairly undistracted way for fairly long periods of time, my mind settles almost meditatively into a more associational mode. I'm reading, thinking, highlighting, looking up, and making connections. It's that mind state that I'm actually trying to achieve, and I don't really find that anywhere else.
Even conversation doesn't really do that for me because I can't stop enough. If we're having a conversation and it goes somewhere generative, I can't just be like, wait, wait, wait, like, I'm just gonna sit here and cogitate for a minute.
That's like, I feel this pressure right now. I'm out promoting the book. I'm here doing this with you today. It's been squeezing out the time for reading. We're recording in May. I've decided in June and July, I'm going to do no work travel. I'm going to do very limited media appearances or aren't things I am myself making at the times, because I just need to get back to where I'm spending two to three hours a day with wood pulp in front of me and a pen in my hand. Right now, I'm spending intellectual capital. That's how I build it.
David:
We're talking a little bit about this at the beginning, but what else has been the things that you've been working on in terms of how you go about the work? Voice, style?
Ezra:
I don't tend to work on things in that defined way, to be honest with you. I don't have years where I think to myself, I'm really working on sharpening my voice. I don't even have pieces really where I do that.
I'm pretty connected to the idea itself in the piece, and I'm chasing that. The piece is done, I guess, sometimes when the deadline hits. But when I feel I've communicated what I'm trying to communicate, I get better through repetition and constancy.
I don't tend to go through long periods where I'm focusing on a skill. I'm not saying it never happens. Sometimes I'll listen to an episode of my own podcast and think, your questions are so wordy.
Although, I've then had the experience where I try to do short questions for a while, and on the one hand, my questions aren't so wordy. On the other hand, the episode is worse because it turns out I'm not able to elicit the same things I was eliciting through more table setting.
So I wouldn't get the feedback, you never stop talking, but I would get the feedback your episode sucks.
So you sometimes never know when you begin pushing on something that you think is a weakness, if it's actually part of a slightly subtler strength.
Different episodes or different, different columns are different. The book was interesting because it was one extremely extended project. Within the confines of that, there was a real necessity of thinking about structure and craft.
But for the rest of my career, I'm pretty high output, so I have a lot of at bats, so I don't treat things too preciously. Balancing Concise Questions and Rich Content
David:
Do you feel like your hit rate needs to be higher at a big institution or no?
Ezra:
That my hit rate needs to be higher? I think it is higher because at Vox, even when I was running Wonk Blog, I felt like I needed to be on everything that whatever the institution needed me to be that day, particularly at Vox, I had to be.
Right now, I just do the thing I think I'm going to be the best at doing. And so I think that's made my hit rate much higher. It means I'm letting a lot of pitches go by. Things that I don't touch or cover, not because I don't think they should be covered, but because other people are covering them at the New York Times.
I don't feel the need to be led by the news cycle in the same way. I am led by the news cycle, but in the sense that the news cycle actually connects to me emotionally and so things in it will really pull me. So I think that has made my hit rate higher.
Vox, when I was helping to start it—it's different now—but when I was helping to start it and was one of its main voices, it really mattered for Vox what my work at Vox was. If I just suck for the next year, the New York Times is going to be fine. It might not feel okay for me if I start feeling my work isn't connecting, but the organization is supporting me. I don't have to support the organization.
David:
Your career is a kind of prism for what's happened in online journalism and how things have gone. And with the way that people like Jonah Peretti and all the big names in media we're talking the early 2010s, the big institutions are crumbling.
There's going to be a new rise, but we've seen kind of the opposite. Maybe it comes back to the business model idea that actually these big institutions do really have an authoritative voice about them that just hasn't disappeared.
Ezra:
People always had this wrong about me, though, and I pushed back on it when I was leaving the Washington Post. I'm an institutionalist. I like institutions. I think they're important.
When I left the Post to found Vox, I wasn't just going independent. I was trying to build something. I was excited in the building of that thing. I wanted to build a big, grand, great institution that would add value to journalism.
Being at the Times, I'm part of institution that I really value. I love the New York Times. I think it is an irreplaceable journalistic institution.
I was never a Martin Gurry revolt guy. I thought there was an opportunity in that period of the 2010s when we suddenly had access to all these new digital tools. I thought we could build things that hadn't been built before.
There are some of those guys who I think actually had a pretty friction filled relationship to even the idea of institutions. They really did want to topple them. Everybody's like, do you want to topple the Post? No, I think they're great. I want to build this thing.
Once I felt like I had gone as far as I could go in that, I was thrilled to join the New York Times. That wasn't some change in me. It was all part of the same project. And right now, the thing I'm building there through the show or through my columns is a thing that they're particularly well suited as a home for.
I like institutions, and I think we need more of them, not fewer.
David:
When you talk to young journalists who come to you and they're saying, hey, I want to become a journalist, how do you think that answer is different now versus 10 years ago, versus 20 years ago when you really got started?
Ezra:
I have more trouble with that question now than I used to. I think journalism is in worse shape than it was even 10 or 15 years ago, certainly than it was 15 years ago.
I think that the number of places that seemed top-tier to me when I became a professional journalist in 2005, that you would have just thought about moving between a little bit seamlessly, was a much wider group of places. A lot of places I love have been somewhat hollowed out, magazines that have really storied pasts. You think about what Newsweek was and what it is.
Sports Illustrated. There are things that are zombified out there. But even things that are not zombified have just been degraded, they're working on shoestring budgets compared to what they once were. The former editor of Vanity Fair, Graydon Carter, has a new memoir out about being in magazine journalism when the budgets were gigantic.4 Then he was succeeded. The editor of Vanity Fair just moved on or is moving on, and there's just a Times piece about, does anybody actually even want this job anymore because it's become so much tougher under the current budgetary environment?
So there are bright spots. I'm not saying nobody makes a good living on Substack, although that did not create a journalistic middle class. I think there's a pretty limited number of people able to do well there. I think that model has real limits for how many people it can support because it's pretty expensive. How many 60 to $80 single-person subscriptions can a person afford?
David:
I'm going to share a thesis, and I want you to poke holes in it and tell me where you agree or disagree. One way that I see an institution like the Times is basically what's really cool about it is it's basically a bundle of a bunch of different writers.
The middle class is propped up by a lot of the really big names. Take someone like David Brooks. I know so many people who have read David Brooks every single week for 20 years, right? And what David Brooks does is he then gets people to subscribe to the Times, and then people subscribe to all these other people. So there was this really effective income sharing inside of journalism that now Substack doesn't really have.
Ezra:
There are two dimensions to the way the bundles has broken, the bundle we used to have. One bundle was you were supporting foreign bureaus by letting people post classified ads for used skateboards. We just dominated classified advertising. It was a huge income stream, just absolutely gigantic. We funded investigative journalism by letting people know that there was a sale at the mall.
I remember when I was a kid, I would open up the LA Times, and the movie theaters paid money to put their schedules in an advertisement in the Times. If I wanted to see what movies were playing when, it was in the LA Times, and that was a revenue stream.
Now, all that is unbundled. One of the things that has made the Times strong and somewhat unique is that it has actually rebuilt a digital bundle.
So, it has the games app, the cooking app, and Wirecutter. First, we have that cross functional bundle. We're not actually just selling you journalism; we're selling you a bunch of other things too, where we have all this audio. Then, as you're saying, inside the organization, there is a bundle.
The bundle of the writers who are high audience and fairly low investment and the writers who, all the way on the other side of it, might be low total audience and high investment. That might be the most important work your organization actually does. So, war reporting is really expensive. Investigative reporting, where people spend six to eight months on a piece, is really expensive.5 And that work has a huge public benefit.
Yes, it's cross subsidized by the people churning things out, and me and some of my colleagues in opinion, but that's always been really important. One of the things that always worried me about Substack as an answer to journalism is it was only ever an answer to one kind of thing, which was high frequency take writing. It was never going to be an answer to investigative journalism.
There was the functional cross subsidy of classifieds and local advertising monopolies and these other things we did. And then there was the way that the sports section subsidized the foreign bureaus, and as we've pulled those things apart, we've thrown a lot of things into crisis.
David:
Have you seen the idea that journalism is the fourth estate?
Ezra:
Right.
David:
Very classic idea. Have you seen studies that have shown basically the social cost of not having good investigative research? I'm trying to basically quantify what's going on.
Ezra:
I have seen studies like that. There are a lot of studies about local journalism and corruption. I wouldn't try to tell you numbers from them off the top of my head because it's been a while since I read them, but there's definitely work on this.
David:
In three hours you're surprised at how much writing you can get done, but in 30 minutes, you're kind of shocked at how little you get done.
Ezra:
I think about this on the book. At some point, what I started trying to do on the book was taking five-day chunks. Basically, I'd go to a cabin for a week, a month.
And the rhythm of that often was for three or four days. I felt like I made no headway, or the headway I made got thrown out. And I was like, "Fuck, I've wasted a week."
And then over the next two days, I would get so much done that I hadn't wasted the week, but it was very nonlinear.
That structural work had to happen, that mental work had to happen, those false starts had to happen. This is a reason I am completely… I think it is very dangerous to use ChatGPT in any serious way for writing.
Because what ChatGPT will never tell you is that the problem with what you're doing is that it's the wrong thing entirely. ChatGPT is always going to give you the answer of how do you tweak, how do you rewrite, how do you do the thing that you think you should be doing? It doesn't know that you're just wrong.
Right? And to what I was saying earlier about good editing, and this is also true for a writer to themselves, you have to be attuned to that voice in you. It's like, "Not right, not right, not right, not right, not right." You're not trying to bypass that or get around it or get to where it's soft. You're trying to get to the point where you're like, "Ah, got it right." And that's usually intellectual labor.
It's not that you've messed up stylistically; it's that you haven't actually done enough reporting, you haven't read the right thing. Oftentimes what would be wrong with a chapter was I hadn't read enough books. I just had not absorbed enough content.
Book writing for me is constantly finding something to pass the baton to. It's like, I read this book, you know, "Homelessness is a housing problem," and I can pass it the baton for 1,600 words. And then I did this reporting on high-speed rail, and I can pass that the baton for 3,000 words. And I had this interview, and that gets a baton for... You're constantly…
Very rarely is what I wanted to do or felt like I should have been doing solving a problem through rhetoric, that it wasn't about not having words. It was about, did I have something at the core of those words, some vehicle, something to pass the baton to, something I was actually communicating about? And when things weren't working, it's usually that at the core, it didn't matter how well I wrote it, that the core of the thing wasn't there. ChatGPT Can't Identify Fundamentally Wrong Ideas
David:
What I'm taking from that is actually one of my biggest lessons from doing this show, is just the most important thing in writing, especially nonfiction, is just having something to say that's important.
And if you have something to say that's important, that's profound, that's vivid, that's hot—I love that word. I'm not going to forget that. But if that then covers up over a lot of other things, and you can wax poetic about storytelling and rhetoric and whatever, but if the tip of the sword of what you're saying can't pierce through the fabric of culture, then it's just going to fall flat.
Ezra:
Look, writing is communication. The question with communication is, what are you communicating?
It's actually important not to gussy that up too much. I'm not saying it's not great to have beautiful style. It's great if you do. I wish I did.
But the most important thing, at least in nonfiction writing, is having the thing you're actually trying to communicate and having done the work to have something worth communicating. Trying to hide that in embellishment and ornamentation is a terrible habit. I've definitely read books and known writers who, they're using a lot of jazz fingers.
David:
Yeah.
Ezra:
To distract from at the center of the thing, they haven't done enough work.
I was once, for my first book, on Pete Holmes's podcast, You Made It Weird. He's a comedian, he's great. I love Pete Holmes.
I'd been on some nerd thick tangent about 1940s America and whatever, and he's sort of looking at me. He's like, "It's fun what you do. You're like a dog who keeps running up to the back door and be like, like dropping something, like dropping a dead bird or something, or a bone."
I was like, that is exactly what I do. At the core of what I do is I try to go out and get you something and bring it back to you. The writing is just a vehicle for the thing.
A lot of the time when I was writing Abundance, I would have a sticker above the keyboard or the screen, and it would just say, "Say what you're saying." A lot of what I think is good about the writing in that book on my side is that it is stripped down. There's very little rhetorical flash. I am trying to say as directly as I can to you what I'm trying to say to you. It's much more declarative in some ways than other things I've done.
David:
One way to get what you're saying out in writing is to think about what we're trying to say and then pretend that now you're summarizing what you're trying to say to somebody at a coffee shop or something, and then just go write that. Just say the damn thing without all the dressing up.
Ezra:
I don't agree with that at all, actually. The way I would say it at a coffee shop, the way I explained the book in an interview, that doesn't work in writing.
David:
Yeah, yeah, yeah.
Ezra:
One of the things I like about writing books is that when you stretch your ideas over that much larger scaffold of 300 pages, you realize there's a lot you thought you knew that you didn't. For me, what works as writing is writing. When I write for the podcast, when I write for audio, it's a very different form of writing. And I wouldn't use that when I'm talking to a person either.
David:
Right.
Ezra:
Communication should be looser and more open and less dense than writing should be. I can't hide in writing the way I can in conversation, and that's part of what I love about it.
If I was just talking to you at a coffee shop and you probed into some place that maybe had a hole in it, it's very easy for me, even without noticing I am doing it, to rhetorically glide right by it. Conversation is not built for rigor; it's exploratory, it's probing, that's beautiful. Writing, I think, is a technology of rigor.
It's much harder. It's not impossible, but it is harder, certainly for me with myself, when I'm writing a piece, to skate by something that I just haven't done the work on. I can see it. Worse than I can see it, I get stopped by it.
I very rarely just stop in conversation because I haven't done the work to keep talking.
David:
That'd be weird.
Ezra:
But in writing, it happens all the time.
David:
I've been asking you a lot about process, and you've said, well, you know, I don't know, but the word that you keep coming back to is work: do the work, do the work, do the work. And that's been the biggest thing I've taken away so far.
A lot of what gives you ideas, a lot of what gives you good ideas, is just consuming a lot of information. What I'm getting is that a lot of times when you're stuck, it's just there's not a deep enough well to draw from.
Ezra:
Writing is the end point of the process. Having created the product, like researched the product, created the product, manufactured the product, it's like selling the product.
I don't think of writing as the core of anything. If you cleaved me off from the research and the reporting, you should not read the writing anymore. That's not what you should get from me: an endless rhetorically elegant recitation of the things I already believe.
David:
Right.
Ezra:
I had written a commencement address for something that got canceled, maybe during COVID.
The whole commencement address was called "Just do the work," or do the reading in this case, actually. The point I was making was that you just would not believe, as a young person going out into the world, how many shortcuts your elders are taking.
When I got into journalism, the idea that I would just actually read the Congressional Budget Office reports, which are not complicated, are not usually longer than 30 pages, but people are just reading the executive summaries, if that. People just weren't doing the reading, and that was a huge opportunity for a young person.
You could just beat people by outworking them, by reading the things that they had ignored, that they found too boring, that was not the part of the job they enjoyed. It happens to be the part of the job I actually enjoy, so bully for me. I read the book. That's why the podcast is good.
It's the writing, the conversation. I would say to my team on the show that 80% of the conversation happens before we walk in the room. It's the choice of guest. It's the preparation for the guest. Us thinking through the conversation.
You know, I always like this, you ever see the movie Tar about the composer, the fictional composer Lydia Tar? I think I got the movie's name right. It was a big deal a couple of years ago, but it begins with this fictionalized interview between her and Adam Gopnik, who's a real life writer for the New Yorker.6
He's saying to her, and she's a classical composer or conductor, I'm sorry. He's saying, how much revelation happens to you when you're conducting? And she says, almost none, but a lot happens in practice.
I was like, yeah, that exactly, that I actually have relatively little revelation when I'm taping a podcast conversation because so much of my brain is absorbed in the work. Even when they're good, I walk out and I don't really remember what happened. If I really like it, I have to go listen to it again myself.
The revelation, the epiphany, the learning, it happens before I walk in the room typically.
David:
Tell about Tom Wolfe.
Ezra:
There's sort of two kinds of immersive reporters, and I'm not an immersive reporter, but one kind of immersive reporter tries to disappear into the background. They try to make you basically forget they are there.
Tom Wolfe was the other kind, and he was always in an all white suit, and he completely stood out anywhere he was. He was like, I want to stand out. I want people to come to me. I want them to be interested in me. I'm like putting up a flag of being somebody you should talk to.
I always just thought that was a fascinating anecdote and also just a good reminder that you can do the same job in such profoundly different ways.
David:
Well, also, I think that there's certain kinds of journalists who kind of fade into the background and give you the objective truth, as how people talk about it.
Then there's other kinds. I think of Hunter S. Thompson. Like, only Hunter S. Thompson could have written "Fear and Loathing." It's not like you just get this sense of this guy with an absolutely wacky mind's perspective on what happened, and that's what makes it vivid and entertaining.
Ezra:
He also, in a way, Hunter S. Thompson's topic so often was actually Hunter S. Thompson.
David:
I want to hear about AI and journalism. Are you just completely against it? How are you feeling about that? What are sort of the fault lines of debate right now?
Ezra:
I'm not completely against anything, and I have not, and not for lack of trying, and I think not for lack of being informed or interested in the issue, found a way that I consistently use AI in my work. I'll sometimes use it right now as a replacement for Google searches, and it's valuable for that.
Deep research has been good on some kinds of preparatory work for me. It's just very hard to not be outsourcing the part of the work I need to do the most. AI is better at the things I need to do than the things I don't need to do.
Having AI step in, which I will sometimes do in a crunch, if I don't have time to make the spreadsheet myself and gather the data in the way I'd like to gather it, having AI summarize a book or a paper for me is a disaster. It has no idea what I really wanted to know. It would not have made the connections I would have made.
That is, again, to my stance on this as somebody who thinks it's the time I spent researching that was the most important time. Having AI do research for me is a bad idea, but if you had other people on the show, they'd give you something different. I think somebody like Tyler Cowen has made AI much more fundamental to his workflow than I have.
I'm interested for the thing I will see that other people would not have seen, and I think AI typically sees what everybody else will see. AI's Usefulness Versus Outsourcing Thinking
David:
I guess what I'm trying to figure out is how much. On one hand, I hear what you're saying: that is the most important part of my work, and I agree. I think the years have proven that out.
On the other hand, I wonder how much of it is a prompt issue of can you prompt it in distinct ways? I think of AI almost like looking at the world through a kaleidoscope. I agree that the default prompt gives you the way that everybody else is looking.
Ezra:
That's assuming that I know what my prompt should be. There's no prompt that would have told me, "Just go read more books, man. You're not there. You haven't gone far enough."
I'm not saying AI can't be useful or that it won't be useful, but I think I'm pretty against shortcuts. You have to limit the amount of work you're doing; you can't read literally everything. But in some ways, I think it's more dangerous to think you've read something that you haven't than to not read it at all.
As somebody who knows the people at these labs and has talked to them about how to use it and has access to the good models, it has not become an honored partner in my creative workflow.
David:
Right. So then that seems like a pretty important fault line: there is your writing, and you're like, "Ah, it's not that useful." Then there is the "I'm doing research; I'm trying to figure out what's going on."
I'll use a super concrete example. We're talking about homelessness as a housing issue. You know, the kind of thing that is really great for AI.
I haven't really thought about that. I grew up in San Francisco, so it's very near and dear to my heart. I kind of watched the city get destroyed by homelessness. I would use AI to basically say, "Hey, 03, can you give me a list of how housing has changed over the years? List of papers that have been written." But what I'm hearing you say is, "Yeah, that's like a better Google."
Then there's something else: I want you to write this for me, and that you wouldn't use.
Ezra:
Yeah. I wrote a piece about this a couple years back; it was called "Against the Matrix Theory of the Mind."
I think I used to conceptualize knowledge the way you see it in the movie The Matrix, where it's like I wanted the port in the back of my mind that the little needle would go into. I had read John Rawls's Political Liberalism. I thought that what you were doing was downloading information into your brain.
Now I think that what you are doing is spending time grappling with the text, making connections. It will only happen through that process of grappling. So, the idea that you could speed run that, the idea that it could just be summarized for you...
Part of what is happening when you spend seven hours reading a book is you spend seven hours with your mind on this topic. The idea that O3 can summarize it for you, in addition to all this stuff you just will not have read, is it you didn't have the engagement. It doesn't impress itself upon you. It doesn't change you. What knowledge is supposed to do is change you, and it changes you because you make connections to it.
David:
What you're saying is knowledge is more embodied than just looking up something on AI. Because knowledge is embodied, it has a way of becoming a part of you and flowing through you into the writing.
Ezra:
Yes.
Ezra:
Yeah, into the writing, into your thinking. Not very much that AI has given me has really changed me very much.
I think that the time you spend with things is pretty important and a reason you get changed over time. I've been reading Matthew Iglesias for how many years now, since he was in college and I was in college. In some ways, it's the length of that relationship I've had with his work that matters most. Any one piece is not that important.
When I sat down and read a RAND report recently on housing construction costs in different states and California and Texas and Colorado, if I just had O3 spit that out, it would have been some useful factoids. It actually mattered that I spent more than an hour on the report marking it up, like looking at it. I just think there's something about your process of grappling with something that integrates it into the way your mind actually works.
I think shortcuts often operate off of a misguided metaphor of how our minds work. We sort of have this download model; I'm downloading the information.
Sam Bankman-Fried is one of many tech people I've seen who would be like, what's the point of all these books? Shouldn't every book be a magazine article? Ideally, should every magazine article be a tweet?
The answer is no. It shouldn't be. In part because that actually doesn't contain the information. I can tell you that for sure having written books and then always thinking about how much fidelity is lost to them when I do the excerpts, when I go talk about them on podcasts.
But also because it cannot change you. A magazine article will change you less than a book will. A tweet will change you less than a magazine article will. You may not realize that at the time, but I think for most people, most of the time it's true.
It's a reason why podcasting is powerful because when they spend two hours with you over and over again, it's like what changes you isn't the wave, it's the erosion.
David:
The way this is also very vivid in my life is travel, where if I go somewhere, somehow that place just becomes completely alive to me in a way that it wasn't before if I just read about it.
What I'm really taking from you here is actually what a book is isn't just giving you a lot of information. It's giving you a container to think about a narrowly defined scope of ideas. Part of reading the book is obviously downloading the information, but the other part is just the kind of rumination that happens in your brain in the process of consumption and just thinking about what it is that you're reading.
Ezra:
That's true for a lot of things. Travel is a good metaphor here. For me, reporting travel often works like this. I had a trip to Israel and the West Bank a year ago in June.
I remember coming back, and that trip was really searing for me. It was one of the most emotionally intense reporting trips I've ever done, the most. I'd come back and people were asking, "What did you learn that you didn't know going in?" In a way, the answer was almost nothing. But now I understood it, or at least understood it in a different way than I had before—not in the way the people I was talking to understood it.
You can have epiphany through writing. Writers always love the Joan Didion quote, "I write to find out what I think."
But weirdly, I think you have to be careful with that, because sometimes writing is a process of persuading you of what the piece needs you to think. You've got to be careful not to become accidentally persuaded by the formalism of whatever your own assignment is.
I notice when I'm doing a podcast, I don't feel a need to come to one best answer on something. So I can hold a lot of different ideas in tension and end with them unresolved. I often think that's a good intellectual place to be.
When I'm writing, there's more of a need to chisel down to my one best argument. In doing that, the person I always persuade the most is myself because I'm going through this whole background process of weighing things, discarding them, and I'm not saying that's not valuable.
But sometimes I think afterwards, I over persuaded myself. I talked myself more into something than I should have. I didn't leave enough doubt. I didn't leave uncertainty. Instead of believing three or four things might be true or all were making a reasonable point, I talked myself into the idea that one thing was true.
It's not even that I think I was wrong. It's just I didn't leave room for the opposite. Sometimes you need to at least be holding things in tension to understand very complicated currents in life. You've always got to be a little careful with the way that writing, particularly if you haven't done enough and complex enough work behind it, will persuade you of a simpler reality than you should really believe.
David:
It's funny because one of the things that you were talking about earlier is the virtues of making something longer. We're also talking about Substack and the take economy.
Ezra:
Right.
David:
We very much live in a world right now where the best way to get attention is to basically have a piercing idea that is a definitive philosophy of how something works, that people say, "Oh, I never thought about that." It's super compressed, super chiseled, and what you're saying is there's room for the conclusion that hasn't even been fully resolved.
Ezra:
Yeah, if I was to describe things I wish I was better at writing, uncertainty is one of them. I think it's very hard for writers to write in a place of uncertainty without falsely resolving it.
Only a couple over the years to me really do it well. Tomasi Coates has done very well at points, but there aren't many writers who can say in a compelling way, "I am not sure of the answer to this. I'm not done with the process. This is not a finished conversation or a finished product."
David:
What I'm hearing is maybe there's a kind of pressure in writing. There's something about the medium of having the authoritative take, and there's no good mechanism. So if I'm talking to you and I say, "Okay, this is how it is now," what I just did is I just, through my voice, increased my level of confidence.
If I do this and I'm sort of thinking like this, you can intuitively understand that. Now I'm maybe at a three or four, but there's no real technology in writing for doing that. There's some subtle stuff with sentence structure, but it's much more difficult.
Ezra:
The structure of the opinion column is, "I got my take and I'm telling it to you."
I always mean to do more things. A case for something and a case against it, and it's kind of hard to do. It feels weird to the audience. It's not impossible. I've done it before.
Not talking yourself prematurely out of uncertainty is difficult. As a writer, you always think you're convincing the audience. You're usually not convincing the audience of anything. You're convincing yourself, though. You always believe your own bullshit first, so you've got to be cautious about that. Opinion Column Structure and Persuasion
David:
Ezra, thanks for doing this. There were so many things that you said that meaningfully surprised me, and I'll send you an email because you have a way of not remembering the podcast.
Ezra:
It doesn't move into long-term memory.
David:
There were seven things that you said where I was like, "Okay, interesting. Let's roll with it."
Ezra:
I'm glad. Thanks. You don't want it to be too obvious from the outset.
Share this post