How David Perell Uses ChatGPT to Write for Millions

TL;DR: Today we’re releasing a new episode of our podcast How Do You Use ChatGPT? I go in depth with David Perell, a writer, investor, and educator behind the popular online writing class Write of Passage. Watch on X or YouTube, or listen on Spotify or Apple Podcasts.

A lot has been said about prompting artificial intelligence into churning out something magical. David Perell, however, uses ChatGPT to prompt himself. 

David is one of the best known internet writers of his generation. He’s amassed almost a half million followers on X, hosts the popular How I Write podcast, and founded Write of Passage, which has taught thousands of students how to write on the internet. 

Under the warm lights of a studio in Williamsburg, I sat down with David for the first-ever live recording of this podcast. It was a deep conversation that shows the diverse powers of ChatGPT for thinking, reading, writing, and decision-making.

We used it to do deep textual analysis of our favorite books, unpack business strategy questions, understand David’s personality and leadership style—and even co-write an article live with ChatGPT as our sparring partner.

I came away from this conversation feeling deeply moved; it was one of the most inspiring episodes we’ve ever done. Here’s a taste:

• Articulating hidden truths. We feed a list of David’s heroes (which he conveniently had handy, by the way) into ChatGPT and ask it to synthesize the core overlaps among them. ChatGPT comes up with something that David had never articulated before: His heroes exemplify a balance of joyfulness and seriousness. “What GPT can do is it can give you clarity and put words to things that you just had a hazy intuition around,” he says.
• On-demand motivational interviews. Like most of us, David struggles with staying authentic to his beliefs in the face of what others think: “There’s the truth that I know and I’m worried about expressing that truth…because I’m worried about how it’s going to make me look.” We ask ChatGPT to interview David live on the show to help him make sense of this internal paradox. Through the questions it asks, we uncover some deep insights along the way, including into the relationship David and I share.
• Finding ideal collaborators. When you’re working with people who complement your personality type, magical things can happen. So we use ChatGPT to help David discover who he worked best with. His guess before we started was enneagram types five and eight. It admittedly takes some prodding, but ChatGPT eventually comes up with the same answer, validating David’s intuitive conclusion: “Did I call it or did I call it… that is pretty cool, right? That is intuitively what I picked up on!”
• Office hours with ChatGPT. We workshop the New York Times’s business model with ChatGPT to demonstrate the importance of getting really specific with prompts. When we input: “Tell me about the biggest risks in the business over the next 10 years,” ChatGPT spews out “buzzword bingo.” But when we niche down, asking it to focus on how the Times can shift from advertising to subscription revenue through the lens of three famous strategists, it does really well. “[N]ow we can start having a conversation amongst those three people to see where the Mount Rushmore of business strategists agree and…disagree,” David says.  
• Obviating the internet’s biggest flaw. David’s biggest gripe with the internet is that it's pulling us into a never-ending now: “[I]f you look at the way the internet is designed…you open up our social media feeds and basically everything was created in the last 24 to 48 hours.” For David, true wisdom lies in things that have stood the test of time, which ChatGPT can help us unearth: “You can literally, in your custom instruction, say, I don't want anything before 1970.”
• Understanding stories. When David was in middle school, he got tutored in reading comprehension, getting extra help to read books: “I would have to interpret paragraphs…I just couldn't do it.” He says that he still struggles to understand stories sometimes. With ChatGPT, David asks for the views of different experts on the text he’s reading: “Going from [a] kid who can't do it to [a] tech-enabled adult who can do it actually better than even some of the best people in the world at looking at stories is…huge.” 
• Reading old books. David uses ChatGPT to translate a verse from the Bible’s Book of Acts in three different styles, discovering that the Greek word archegos can mean both “author” and “prince.” Sharing his “conflicting desires” as both an entrepreneur and a writer, David says: “[F]or me and my work running Write of Passage, I can now make an argument for writing that is rooted in this Greek word…all these things that I thought were totally different are actually one and the same…GPT is really useful because I can see the different translations, look at the discrepancies, and get my answer.”
• Surfacing anecdotes that spread. David has mastered the art of the anecdote (read the first paragraph of any of his essays and you’ll see what I mean). We decide to write a piece about New Yorkers who have good taste live on the show. In pursuit of a great anecdote for the introduction, David asks ChatGPT to list beautiful buildings in New York. The Woolworth Building, the first one on the list, sparks his memory—and changes the direction of the piece. “I just had an anecdote that showed up in the very first example [of ChatGPT’s output], and we're going to change what it is that we're writing about,” he says.
• Writing from conversation. It comes as no surprise that David sources ideas for his writing through conversations with friends, families, and acquaintances. Now, he uses ChatGPT as well: “Until the development of this technology, you could only do that with another human being…now you can do that with an extremely smart computer.”

You can check out the episode on X, Spotify, Apple Podcasts, or YouTube. Links and timestamps are below:

• Watch on X
• Watch on YouTube
• Listen on Spotify (make sure to follow to help us rank!)
• Listen on Apple Podcasts

Timestamps:

1. Intro 00:53
2. Finding and understanding his heroes 13:42
3. Understanding his personality and leadership style 19:14
4. Who does David work well with? 25:53
5. Workshopping the New York Times’s business strategy 36:52
6. Why ChatGPT is incredible at diversity, accessibility, and speed 52:54
7. Bringing old books like Moby Dick to life with DALL-E 58:50
8. Using ChatGPT for deep textual analysis 1:06:29
9. ChatGPT for writing anecdotes that spread 1:21:04
10. Conversations with ChatGPT as food and drink for the soul 1:25:55

What do you use ChatGPT for? Have you found any interesting or surprising use cases? We want to hear from you—and we might even interview you. Reply here to talk to me!

Miss an episode? Catch up on my recent conversations with software researcher Geoffrey Lit, Waymark founder Nathan Labenz, Notion engineer Linus Lee, writer Nat Eliason, and Gumroad CEO Sahil Lavingia, and learn how they use ChatGPT.

My take on this show and the episode transcript is below for paying subscribers.

This episode was a glimpse into what this show—and this tool—can be. David and I went on a mutual exploration together, live in the studio with ChatGPT. It took us to places we never thought we’d go. I think that’s the magic of this show, and the magic of ChatGPT. 

What stuck with me most about this episode was the Bible analysis we did. As an anxious New York Jew, I was a little apprehensive about doing a Bible study with ChatGPT. But David turned its eye to a particular word in the text, archegos, which means “prince,” “pioneer,” and “author” in Greek. He used that to show that in English, the idea of being a leader and that of being a writer are two separate things. But in Greek, they’re one and the same.

This conflation was particularly meaningful to me, and I think it’s relevant to anyone who’s thinking about becoming a writer. There’s always a sense of loss with the practice involved: If I’m writing, does that mean I’m not doing anything in the world? But this word, archegos, shows that this binary doesn’t have to exist. Writers can be—and are—leaders.

The stories we tell are the businesses, organizations, and movements we build. This episode left me deeply inspired. I hope you feel the same. 

Transcript

Dan Shipper (00:00:00)

I started reading Moby-Dick recently. It can make the books come alive not just through language but through pictures.

David Perell (00:00:05)

Wow. That’s really cool.

Dan Shipper (00:00:08)

Isn’t that cool?

David Perell (00:00:09)

Yeah.

We’re seeing something through the Greek that the English doesn’t reveal, and GPT is really useful because I can see the different translations, look at the discrepancies, and get my answer.

Dan Shipper (00:00:20)

Isn’t that kind of wild.

David Perell (00:00:22)

So now I just had an anecdote that showed up in the very first example and we’re going to change what it is that we’re writing about.

Dan Shipper (00:00:28)

We went in thinking we’re going to write an article about New York City and taste and we came out with this max-min thing and that’s just how the creative process works. And that is not a mistake. 

David Perell (00:00:39)

And that is what GPT is really good for.

Dan Shipper (00:00:54)

David, welcome to the show. 

David Perell (00:00:57)

Thanks, man. I'm excited to be here. 

Dan Shipper (00:00:59)

Yeah, I'm excited to have you. We've been friends for a while. You're an amazing internet writer. You run Write of Passage and you're the host of the How I Write podcast, which is a fantastic show. And yeah, I'm just I'm just really excited to have you to talk about—

David Perell (00:01:11)

And an investor in Every! 

Dan Shipper (00:01:12)

And an investor in Every! Yeah, big supporter. You've been a big supporter for many, many years and I really appreciate getting to spend time with you. So we're gonna talk about ChatGPT, obviously, and we're going to get into how you use it in your creative process, how you use it to think, how you use it to write, how you use it to learn new things.

But where I want to start is I just want to start with you. One of the things that I think is really just special and unique about you is you're on this mission to understand more about who you are and then to reflect that back into your work, and you're on a mission to understand why that's surprisingly difficult to do and why it's hard to accept in some ways who you are.

And so what I want to start with is. … is you. Tell us about you. Tell us about what you've learned about yourself, maybe over the last year, and we'll see where that goes. 

David Perell (00:02:18)

 Well, I think that there's a lot in society right now around “be yourself” and “be who you are,” and I think it's exactly halfway right.

 Because, “be who you are,” look, I just believe that God, Creator, has given us all a series of gifts and has endowed us with a unique fingerprint and basically a unique grain, a shape of our personality that allows us to do certain things easily and other things in more difficult ways, right? Take someone like Michael Phelps, like that guy's just made to be a swimmer.

He has an insane wingspan, super-long torso. LeBron James is made to be a basketball player, but somehow when it comes to intellectual talents, we forget that there are certain things that people really have an aptitude for. And, so I've been thinking a lot about, what am I made to be doing? And, the first thing there is really a surrender.

It's really a surrender. Like, this wasn't my choice. I believe in God, so, what did God make me to do? And actually responding to that. So that's the first thing. And then the other thing is, there's a lot around, “oh, be yourself, be yourself.” And what I find is it, in practice, ends up giving people a moral hall pass to do whatever they want and it leads to a lot of degeneracy, which I'm not a big fan of. And so I'm interested in how do you be yourself in a way that's rooted in integrity? And integrity, I think of as coming from two places, which is clear values and a clear set of principles and also consistency along those principles. And I like the idea of being yourself in a very integrity-heavy way.

And one place we try to get integrity—without people being themselves—is the classic thing where, you see it with a lot of children, where they become lawyers and doctors and they go into a very small list of fields. And what I think lacks integrity about that—I mean, there's a nobility to it of,” Hey, you're going to do your family well.” But what lacks integrity about it is it misses human individuality, which I really value and think is important. 

Dan Shipper (00:04:21)

In your life, what has brought about that focus for you? What made you realize that this was really important to figure out for yourself? To live with integrity, know your values, know your individuality, and surrender to it.

David Perell (00:04:28)

So the pursuit of excellence is my number one value. And I spend a lot of time with people who I think are in pursuit of excellence, who I think are world-class in their field. It's one of the core ways I spend my time. If you really spend time with them, you realize that there's a graciousness to their effort. You know, Djokovic says, “I just love hitting the tennis ball.” You spend time around great entrepreneurs, and there's a way that they're just living fluidly and there's this whole conversation around hard work. And once again, it's exactly halfway right. You know, early in your career, you do have to sort of grit your teeth and grind a bit to get the snowball rolling. But the people who I know who have really done well. Yes, they worked their tails off, but there's a sense that they have where they're just doing what they're made to be doing. I think a lot about the word compulsion. They're compelled to do that thing. Right? You spend time with a drummer. You're hanging out on a Saturday night and they're just tapping their fingers. They're making beats, right?

And I played golf at a fairly high level with many people who now are on the PGA Tour and when I would hit balls with them growing up I mean, there was just a, “Of course, I'm going to wake up and go to the to the golf course all day” and then the other person who comes to mind who is really a model for me here is Tyler Cowen.

Tyler Cowen funded me early in my career. He's a good friend of mine and he's outstanding in terms of the depth, the breadth of his knowledge. I've never met anybody like him, but if you just watch a video with him, spend time with him, there's a joyousness that comes from every interaction because he just has such alignment between his career and the things that he's drawn to and when you find that there's a real attractiveness, a charisma that comes from it because it's so rare. And I look at Tyler and the lesson isn't “be more like Tyler,” the lesson is to think, “What is Tyler seeking and how can I seek that same thing in my own life?”

Dan Shipper (00:06:35)

That's really interesting. I love that. I mean, I think you're mentioning all these things that feel very core to you and also feel core to me, like feeling like you have this sense of joy in the work that you do or maybe a sense of compulsion to do it really well or drive for excellence.

One of the things that I think would be fun to try just going right into ChatGPT is—I think it's really great at… You can do a really good job of articulating who you like, who you look up to, and why, and what that says about you. And I think ChatGPT is really good for giving you more depth and clarity and insight into that stuff.

So one thing that we could try is taking some of the people that you mentioned, so maybe Tyler Cowen or Michael Phelps, writing into ChatGPT who you like and why you like them to some degree, and then seeing if we can push it a little bit to maybe find some more people that you might not have heard of, or maybe give you a little bit more depth of synthesis on what are the things you like about them, and then we can see where that goes.

David Perell (00:07:35)

Great. Let me say one thing, and then why don't you walk me through that. So one of the things that I've always just been drawn by is being the best in the world at whatever it is that I do and I was really obsessed with basketball and baseball and golf as a kid, and I've always just been drawn to excellence. And fundamentally the reason why you have to surrender to your nature and why you really have to think about these questions is you have no shot being the best in the world at what you do if you aren't aligned with the shape of your soul that God has given you. If you're trying to fight that, you have no chance because somebody else has that and now you're competing against them. So the only way to truly be the best is to be aligned with that nature. And if you are, then you have the advantage rather than fighting other people on their terms.

Dan Shipper (00:08:27)

I love that. I think it's such an important lesson. It's one that I've had to learn over and over again, because one of the dangers of heroes is that you try to emulate the things that they do. Right. And then you end up doing something that's not quite right for you because you're following their path and not yours.

And usually there's a thing about heroes that that you like that does resonate with you. And that thing is true and real. But like if you just wholesale take their whole life and try to do it, it doesn't work. And I just found that over and over again, because with Every, for example, I tried to run it like a software company for a long time, and it's not really a software company.

David Perell (00:09:04)

No, it's not. It's a media business. 

Dan Shipper (00:09:07)

It's a creator led media business, right? And so the way that you run a creator-led media business is if I'm one of the top creators, I got to make stuff. I got to be in this room with you, making podcasts or writing or doing all that kind of stuff. But if you're the CEO of a software business, you don't make the stuff—you hire the people to make the stuff and you raise the money and you hold the vision and whatever.

And I had to learn that lesson a lot because it's such a different way of thinking about what to do and how to build a company. And there's a deep symmetry between the way that you build a software business and build a media business, which is that in both cases, the CEO does the thing that they are best at. And in media businesses, that's usually making the content or making the thing because that's the rarest thing. And in software businesses, that's raising the money and having the vision and hiring the team. But if you just wholesale take the software business pattern and apply it to media, it just doesn't work. And I had to learn that from just so much pain and effort over the years because it's so different, you know? 

David Perell (00:010:06)

I gotta say one more thing. Do you know the story of the Bed of Procrustes? Okay, so this myth is amazing. It's an old Greek fable. And there's a hotel owner named Procrustes, and he's sort of this tyrannical hotel owner. And what he does is his entire hotel only has one size of bed, and he wants people to fit in the bed perfectly. So, if somebody is too short, he stretches them out. And if somebody is too tall, he chops off their arms and legs so that they can fit in the bed. And I think that a lot of people take a Procrustian approach to their career where they say, “I need to be this person, I need to fit into that bed,” so rather than conforming the bed to who I am, I'm going to conform myself to the bed, and it creates a lot of misery in the world. 

Dan Shipper (00:10:51)

Yeah, no, totally, and I think the root of that is there's shame in a lot of ways, and guilt in not being able to do something that you think you should do, you know? And I think figuring out what are the overlaps with what other people have done, and what are the things that are just uniquely mine is sort of—It's the work, you know, and I think that that brings us right back to ChatGPT. What I want to do is find a little bit more about who you are and what you like through this. And then, and then maybe we can bring it back to what are the not overlaps or what are the things that are not you about the people you love and what have you had to learn and all that kind 

David Perell (00:11:23)

 Can you help me prompt GPT? 

Dan Shipper (00:011:27)

Let's do it.  Let's start with people that you like, people that you are inspired by, and I want to find some of the overlaps and some of the some of the synthesis there. So I think that's great. Here's some of my heroes.

David Perell (00:11:38)

You know, I think—give me a second. I think I've actually written out all my heroes. 

Dan Shipper (00:011:43)

I love that. 

David Perell (00:11:42)

I think that I can just copy and— 

Dan Shipper (00:011:44)

That would be perfect. 

David Perell (00:11:46)

Yeah so what I did here, Dan, is I had been building this for many months and I gathered a list of people and creators who I really admire and I just had some—I guess you could call them adjectives—on what it is that I admire about them. So, I don't know there's 15 people in here and I just put into GPT, here's a list of my heroes and why I like their personality and writing. Is there anything else that you think I should add? 

Dan Shipper (00:012:16)

Yeah, let's scroll to the bottom and just say, “Can you summarize the vibes of these people?.” And then, yeah, let's just see how that goes. We're just going to explore together.

David Perell (00:12:39)

Yeah, I think one of the things that always strikes me about GPT is not expecting it to give you a good first answer, but thinking of it as a dance and the fourth, fifth, sixth answer is really when you get quality. 

Dan Shipper (00:012:53)

And one of the things I like to have it do is when you ask it to just summarize the vibe, what it will do is it will sort of expand from the name into more about that person. So for each person is saying like, “Peter Thiel known for his curiosity and boldness,” and what we're going to do is take that and then compress it down into something that finds a lot of the overlaps and synthesizes it, but it's helpful to have it in expanded form first.

David Perell (00:13:24)

Yep. Interesting. So what you're trying to do is you're trying to get as much information onto the page, basically give it a lot of receptors, and then it goes into distillation. I like that. 

Dan Shipper (00:013:29)

I'm curious how you feel about the summaries of these people. Do you think it's using the right words for them, the right descriptions?

David Perell (00:13:38)

It's very good. 

Dan Shipper (00:013:39)

Okay, cool. So now what I want you to do is write something like, “Can you help me synthesize these? I want to try to find the underlying overlaps and commonalities between these people.”

David Perell (00:13:54)

So what I did is I said “limit one paragraph. Then share the four core bullet points, then summarize in one sentence.” And I like doing that with GPT. I like giving it very specific prompts and instructions. I basically assume GPT is like a super-fast wizard with like, super-high IQ, but then also just in a funny way not creative at all it just has to do exactly what I tell it to and we'll see what it says. 

Dan Shipper (00:014:19)

So here's  what I love about this, right? It says “heart-led and joyous approach.” So ChatGPT doesn't know that one of the things you said that you love about Tyler Cowen is that he's joyous. Yeah, and it just like found that right? You're self-aware and you think about this enough to know that that's why you like Tyler Cowen. But if you're a person, and many of us are, who, you generally know that you have heroes, but you haven't spent the time to really articulate why. It just did that for you in a way that took, I don't know, a minute or two. And, I don't know how much time you took to identify that about Tyler Cowen, but I think this sort of like fast forwards that process in a really interesting way.

David Perell (00:14:58)

Well, it basically says four things. First, a passion for excellence and innovation. I mean my core value is literally the pursuit of excellence. The second is intellectual curiosity and insightfulness. I mean, that's my job, trying to be those things. Heart-led and joyous approach. I mean, that is exactly how I approach the world. Also quite intuitive in terms of how I move through my interactions. And then finally, commitment to personal and spiritual values. Many of your heroes are guided by a dedication to God, a commitment to truth and fairness, or a drive for personal authenticity. Yeah, I think that's pretty good. I mean, this is a solid B-plus answer, and we got it in five minutes.

Dan Shipper (00:15:38)

Why don't you hit the redo button? I just want to see if you do it again if you get anything different.  That's one of the things I really like to do and it's actually come out of some of these interviews is—I interviewed Linus Lee a couple of weeks ago. He's a researcher who works at Notion on their AI team. And he just does like four of them for each response, just to sort of get the lay of the land. I think one thing that people miss about ChatGPT it is kind of like an intern, but it's not like an intern because it doesn't get angry if you ask it to redo its work, it'll just keep going and going. And what you want to do is traverse all of the different ways it might respond to your question to see if you're missing anything.

David Perell (00:16:20)

 I mean, to your point, I really like what it said here, which is “balance of joyfulness and seriousness.” I think I balance both of those quite a bit. I am a fairly joyous person, but I'm also extremely serious about my craft and I have a really high-quality bar. And I'm always trying to maintain those standards and then every single day do things to raise my quality bar for what excellence looks like. And this idea of the balance between joyfulness and seriousness, this is the first thing I've seen that has now given me something back that I hadn't quite put into words before.

Dan Shipper (00:17:57)

I love that. That's such a good feeling when you, because now when you have words for it, you can do something with it, right? 

David Perell (00:17:02)

You know, it reminds me of, so my fraternity president, my freshman year: I was talking to a guy at a party one night and I said, “Tell me about Artie” and the line was so good. It was exactly this. He's like, “Artie is 90 percent the most slap back, hang out, funny guy, but the 2 percent of seriousness that he is is the most serious of anyone you will ever meet.” And it's funny to read this. I had that conversation 10 years ago. It's so funny. And it's exactly what this is. And something about that sentence really, really resonated with me.

Dan Shipper (00:17:57)

I love that. I love that. I guess there's a couple different places you can go. First of all, I'm curious, what are you curious about? What do you want to know about yourself right now, from having read this?

David Perell (00:17:46)

I want to know the kinds of people I would work well with and who  I need around me in order to—yeah, that's what I'm curious about. And one of the things that we might want to plug in is some of my weaknesses. And I've noticed that I work very well with people who are Enneagram 8s and they tend to be very direct. And I also work well with people who are Enneagram 5s because 5s are all about information synthesis and knowledge.

So some of my weaknesses are basically all my problems in life come down to one thing. And if you even want to write down this sentence, it could even be interesting, which is there's the truth that I know and I'm worried about expressing that truth and saying what I need to say because I'm worried about how it's going to make me look. That's my core weakness in life. So I work really well with people who the 8s tend to be conflict-forward and they say what they mean and they mean what they say. So I'm curious to know, given my Enneagram 3 tendencies, I'm a 3 wing 4, who then do I need around me in the work world? 

Dan Shipper (00:19:08)

I want to start at a previous place, which is before we get to who you work well with. I want to unpack this truth that you know, that you're worried about expressing that truth and saying what you need to say.

So, what I want you to do is start a new chat. Okay. And ask ChatGPT, “I want you to do a motivational interview with me.” And motivational interviewing is a technique in psychology that's about unpacking the sort of conflicting forces within you that produce ambivalence or a behavior that you don't really like and helping you to move into a better understanding of those forces so you can integrate them and do things differently.

So “I want you to do a motivational interview with me. I'm thinking about my tendency to know a truth and be worried about expressing that truth and saying what I need to say. Specifically, I'm worried about how that's going to make me look. I want to try to understand that better, where that and where that comes from. Please do a motivational interview and ask me one question at a time.” One thing I've noticed is ChatGPT will spew 10 questions and it's like, you can't answer 10 questions at once. And so the one question at a time is really good. So it's asking, 

“When you think about situations that hold you back from expressing a truth, what are the specific fears or concerns that come to mind and how might it affect your image or reputation?” And I'd add to that why don't we pick one specific one?

Hmm. “Vengeful.” 

David Perell (00:20:44)

Yeah, so I wrote “being seen as vengeful or uncaring.” 

Dan Shipper (00:20:52)

“Can you recall a specific instance where you felt—”

David Perell (00:20:54)

Oh man, we're getting into it. 

Dan Shipper (00:20:56)

This is getting real. 

Is there anything in this room between us? Yeah, because this it's pretty safe with me.  Even a really small thing because even the smallest thing will be— 

David Perell (00:21:09)

—will be a big thing.

Dan Shipper (00:21:10)

 So like is there something between us that you felt you're holding back and—

David Perell (00:21:14)

Yeah, okay. Here’s one. I think that one of the things, if you're going to make this podcast successful, is how do you figure out how you can get a lot of the information that you're asking for in advance? How do you get it without it being a bunch of work? Right? Because we were going to record this podcast a few weeks ago. 

Dan Shipper (00:21:33)

And I sent you a big document. 

David Perell (00:21:36)

 And then I didn't factor that into my week, and then I canceled on you. And then, yesterday, I was able to do it, but then I couldn't open up the Notion, so then I sent you a text, and, how do you make that prep work really easy so that you can get the prep work, but then as a guest, then it's fluid and easy for me.

Dan Shipper (00:21:51)

And you didn't totally tell me that—you sort of told me, you were sort of like, “Hey, I gotta cancel.” And then you didn't tell me, for example, that you didn't know where the Notion doc was and you just sent me a text. Why don't you throw that in there? So you're saying, “I'm recording a podcast with my friend, Dan. There was quite a bit of prep work in advance. I hadn't mentioned to him that I want to think through how he can reduce the prep work required for the show while also having it be as good as he wants it to be. If he wants to record with high-level and busy guests, there can’t be a lot of prep work. That’s really great.

And it said, “In considering the situation, what do you believe are the potential benefits of discussing your thoughts and reducing prep work with him, and how do you think this conversation could align with your shared goals for the podcast?”

David Perell (00:22:32)

Yeah, I think that first question is better than the second one. I think the benefit is, it's important, I mean, in any experience, for you, as the creator of the experience, to know what it's like to consume the experience. You know what drives me crazy? When you're at a restaurant and the butter is too cold, so it doesn't easily spread on the bread.

And the reason that that happens is, who's ever making the butter isn't eating the butter and spreading it on the bread, so then it gets chunky and blocky. And I think that when you're consuming any experience, whether you're putting on a show, whether you run a hotel, whether you run a restaurant, and whether you run a podcast, is knowing what it's like to be on the other side, because you're just blind to it when you're creating that thing.

Dan Shipper (00:23:13)

 That makes a lot of sense. So what I want to do now: You've given it a bunch of information about who you are, where you are getting stuck, what the pattern is. I want you to ask it to summarize everything it's learned about your psychology from this interview. Yeah, that's great. And then what we're going to do is we're going to start a new chat and we're going to use what we've learned to have it help you think about how to be a leader who you should work with. 

David Perell (00:23:40)

When do you start a new chat? 

Dan Shipper (00:23:43)

I tend to—I think we could do it in this one. And my experience is that with chats, the longer they go on, the more likely it is to sort of go off the rails. And so if you can start with like a really concise piece of information, like a summarized paragraph of like, “Here's the nugget of stuff about me,” and start it with a new chat, it'll be less likely to mess up. 

David Perell (00:24:07)

Okay. Yeah, there's a few things here. I think “sensitivity to perception,” like, “you exhibit a strong awareness of how your words and actions might be perceived by others.” There's no doubt about that. Yeah, I think “balancing honesty with empathy.” I always think of balancing grace and truth. And I actually don't like the word “balancing.” I like full-on both. You want full grace and full truth. I don't like balance because it's not a spectr It's full of both. 

Dan Shipper (00:24:36)

Well, yeah, that's great. I think we should maybe throw that in there. We can copy-paste that and change that little thing. So what I want you to do—and I can start doing this for you if you want. Like, do you think that would make it easier for you? 

David Perell (00:24:52)

That would be a lot easier. 

Dan Shipper (00:24:36)

Okay, cool. So what I'm going to do is I'm going to take what you wrote, and we pulled out some patterns that are about your sensitivities, your perceptions, your past experiences, getting feedback and giving it all that kind of stuff. You mentioned you don't like the words “balancing honesty with empathy.” What would you like to replace it with? Harmonizing or— 

David Perell (00:25:13)

I think “grace and truth,” and full in both of them. So grace is a—there's a tenderness to grace and understanding to grace and a real humanity to grace. And then truth is there's a firmness, there's a logic and there's a strength to truth. And you want full of both of them. I don't like the balance of grace and truth. I like the fullness of both of them. 

Dan Shipper (00:24:36)

Yeah, I love that. “Trying to fully live grace and truth.” Okay, cool. So, what I'm gonna do, so we have a bunch of stuff about you. And I want to help you figure out who are the kind of people that you should be working with. And then we can see if it knows the Enneagram 8s and Enneagram—what was the other one? Enneagram 5. And it would be interesting if it could figure out your Enneagram numbe from just this.

So let’s say, “I'm trying to figure out who are good personalities for me to work with based on my own psychology. Here's a bunch of stuff about me. Based on this, can you help me understand what kinds of people would be helpful for me to work with and who I should avoid?”

Let's just see if that works. I don't know if it will. “Supportive communicators.” “These are people who can provide constructive feedback in a way that's encouraging rather than critical.” 

David Perell (00:26:51)

I almost need the opposite, actually. Because I need people who push me to be more critical. I think five is good. Yeah, five is really good. Like “independent and self confident people.” Yeah, that really helps me.

Dan Shipper (00:27:08)

It says “overly critical or negative individuals are people to be cautious about.” What do you think? Or “those who avoid conflict at all costs?” 

David Perell (00:26:15)

I think two is a big one. I mean, if I'm around other people who avoid conflict at all costs, it's a mess. I'm really good with people who are direct and I can really help them too, because I can come in and help them out. I mean, I've already had that happen today where somebody was direct and then I came in on the other side and I said, okay, hold on here. You don't need to talk like that. 

Dan Shipper (00:27:40)

 Well, let's see. So I said, “I feel like I need the opposite of supportive communicators. I need people who push me. I really like independent self-confident individuals and those who avoid conflict at all costs.” Let's see what it does 

David Perell (00:27:52)

I like people who mean what they say and they say what they mean.

Dan Shipper (00:27:54)

“Challenging communicators.” “Seek out people—” So I mean, at this point it's sort of repeating something that you know intuitively. And I think that's one that's a really interesting part of ChatGPT is it's not going to be 100 percent all the time. But when you get these directives from it, you know, “supportive communicators” or “empathetic and understanding colleagues” or “assertive yet compassionate leaders” or “strategic thinkers” or “independent and self-confident individuals,” some of them are going to be bang-on and some of them are going to be wrong.

And it's really important to take what it says with a grain of salt and check it with your own experience and use it as like, “Okay, here's some things that I like, here's some things I don't like” and keep pushing with it rather than being like, “Oh, it sucks” or like, “Oh, like it's not working.” And also I think to take some of these things and be like, “Okay, this is an experiment. Maybe if you didn't know yourself as well, you could have seen the supportive communicators and been like, “Maybe I need that” and tried it and found that supportive communicators are often conflict-averse and that that doesn't work for you. And so even if it's giving you something that's not totally right, if you're treating what it gives you as an experiment in your life, you can start to make progress knowing who you are.

David Perell (00:29:07)

What I hear you saying is you have to evaluate GPT on slugging percentage instead of batting average. So batting average is what percentage of the time do you get on base? What percentage of the time are you successful? Whereas what you're saying is slugging percentage, which is when you're successful, how successful are you? So venture capital is very much a slugging percentage game, and that's how you have to think about GPT. And I think a lot of people evaluate GPT based on batting average.  So what they do is they say, “This percentage of things is off. Rather than, what you're saying is, “how do you work with GPT to get the one or two nuggets, the one or two threads that you wouldn't have found elsewhere?”

Dan Shipper (00:29:46)

Exactly. It's that diamond in the rough, one out of a hundred, you're like, “Holy shit. That just like completely put into words. Something that I'd always felt, but had never been able to say. And also it's like, you get out of it what you put in, right? So if you're giving it good prompts, you're going to get good results. And so if you're not getting what you want, trying to push it to get more of it is, I think, the best thing to do.

David Perell (00:30:10)

It's sort of like archaeology, right? So what you're doing is you're digging and you're digging and you're digging. And an archaeologist doesn't go out expecting every single second of every single day to yield a new insight. But what they're doing is they're sort of following their intuition for what's interesting where an area could be. And then they end up finding the ruins that they were looking for.

And there's a sixth sense that's created. It's the same thing with how a geologist goes out and looks for oil, right? You look at the map. You have a sense of the map. And then you say, “these sorts of places are more likely to have oil.” And then you strike oil and you say, “Woo, hallelujah.” But you don't expect every single place. And those are very different modes of engaging with the technology. 

Dan Shipper (00:30:56)

Yeah, 100 percent. I think we can do one more thing here, which is “What Eneagram types, do you think I'd work best with?” What's your type by the way? 

David Perell (00:31:08)

I'm a 3 wing 4.

Dan Shipper (00:31:09)

“I'm a type 3. If you had to pick two others, which would you pick as best to work with me?”

There you go. Isn't that kind of wild? There you go. 

David Perell (00:31:34)

Did I call it or did I call it? No, it's, I mean, it's crazy, right? That is pretty cool, right? That is intuitively what I picked up on, and what's really funny is if I look at the people who I enjoy spending time with the most in my life, I love 5s.

I love 5s because they help me understand what's going on. They're analytical, they're perceptive, they're innovative. Wow, that's hilarious. And then I love 8s. They're assertive, they're confident, they're decisive. Wow. ChatGPT. Banger. Alert. Alert. That's really funny.

Dan Shipper (00:32:06)

Isn't that interesting to you? Those words, “analytical,” “perceptive,” and “innovative.” Once you have them, there's something really special about that. 

David Perell (00:32:14)

You know, this is actually really good. So, okay. Now we're onto something. So “analytical, perceptive, innovative.” This is where you started. You basically said, “What GPT can do is it can give you clarity and put words to things that you just had a hazy intuition around.”

So I'm just gonna skim through this. So, yes, I'm “goal driven nature,” and I'm probably too risk averse around perception. Okay, so “analytical,” “perceptive,” and “innovative.” Also I'm a little bit sort of in fantasy land around sort of a more in terms of hopefulness around perfection than being like a realist. So I like 5s because they are realists and they're analytical. “They value knowledge and competency.” I love competency. “They compliment my drive for achievement efficiency.” No doubt. “They dive deep into subjects.” I'm super curious. So they can give me that well structured information. “Then they give me insight, strategy”—yup.

And then the 8s, this is the other side, right? “They're assertive, they're confident, they're decisive, they're not afraid to take charge, confront challenges head on,” right? Like, they're much more in a boxing match with the world. And they help me “push my limits, take decisive action, straightforward ability to handle conflict, tough decisions, direct communication.”

Those are things that I'm very attracted to because I lack those things in myself. And so, of course, there's an element of me where I'm trying to adopt more of that. But if I can just have people around who are like that it's like, “ahh.” It just makes me so relaxed. You know what? I mean? I'm like we have that covered so I can be more of me.

Dan Shipper (00:33:49)

I love that. I think that's so about surrendering to who you are—going back to the thing we started with, which is, yeah, you could totally work on being less conflict averse and that's probably a nice thing for you to do but also different contexts and different relationships and different people will either bring out the conflict aversion or force you to be less conflict-averse because they will just push you on it.

And that is really valuable to know because either way you're making the change. Either way you're creating a context for yourself where you're not being conflict averse. One way is to do it yourself and another way is to create the context and the relationships that do it for you. And I think that that's just a really smart, valuable thing. 

David Perell (00:34:28)

Well to that point. So there's two buckets of self-awareness. There is the first thing of self awareness, which is, these are my weaknesses, and I'm going to basically try to neutralize those weaknesses. The second thing is, how do I get people around me who make it so that I don't even have to, because they can sort of cover that for me.

And I think that the problems that emerge don't come from weaknesses. They come from an ignorance of weaknesses. The worst things that happen are when people try to take on projects that require them to have a strength and a place where they’re weak.When they say, “Oh, I know this,” when they actually have no idea what they're talking about.

And this goes back to Charlie Munger's circle of competence point. Something that's actually more important than how much you know is just being honest about what it is that you know, and then living a life where you are congruent with the shape of your skills and knowledge, and once you try to go beyond your skis, that then is pride. That then is arrogance, and that's really when you get the problem. And I think that people would do better to just be honest about what they are and aren't good at, rather than trying to basically deceive other people into saying, “Oh, you know, I am really good at that,” and sort of puffing themselves up.

Dan Shipper (00:35:46)

I think you're right. And underneath the pride and the arrogance is this sort of shame about having a weakness. And that's what I think is so powerful about doing this with ChatGPT is because it can help you explore that stuff. And also for you, for example, being able to say, “I'm conflict-averse,” a lot of times like I'm conflict-averse too. I want to hide that because I want to be different or whatever, but just being able to stand on that as, it's not an excuse. It's not “I'm just going to do this,” but it is a thing that you should know about me and something that I'm currently working on. And there's a lot of power in that because if you don't have to hide it, you can make changes that make up for it.

Cool. So this is great. I feel like we've done a lot of good stuff, like kind of using ChatGPT to explore a little bit more about who you are, your strengths and your weaknesses, and who are good people to work with for you. I would love to move into the next section and just talk about how are you using it?  Now that you know who you are, how are you using it to make stuff? How are you using it to write? How are you using it to create? 

David Perell (00:36:46)

So one of the other things that I've found with GPT is it's really good at looking at something like a business model and one of the things that people miss is they'll put in their business model and they'll say “GPT, critique this business model.” And what they end up doing is when you're that broad, your answer is going to be very poor. You have to give GPT shape and direction you have to otherwise, it won't give you a good example.

So what we can do is we can put in the Every business model and you can put in here's how we make money, here's how we're growing, here's our product. And let's say that you want to, you know, let's do the New York Times. Let's do the New York Times. That would be really interesting because there'll actually be a tapestry of knowledge that it can build on.

So the New York Times. And maybe you want to write this. So the New York Times, they came out with a memo, I want to say it was 2015 or 2016, and it's actually extremely interesting, where they spoke about how they need to change as a company in order to stay in business. And the New York Times, they make money by being the authoritative source on news stories around the world—politics and culture—and what they have needed to do is shift from advertising revenue to subscription revenue.

So what I would like to know, and maybe we can ask GPT, is what is their annual revenue? How many readers do they have per year? What are the most profitable parts of the business? What are the least profitable parts of the business? And then we can just ask GPT to critique the business model. So have that be the final sentence. And what I want to do is I want to show that it's not going to do a very good job of critiquing the business model

Dan Shipper (00:38:39)

So, I've written all this stuff out. I think it will sort of work, but it might not be able to accurately answer each of these questions in one prompt because, you know, the way it requests information from the web, it—we'll do a search and if it can find one article with all the information, it'll work. But if it requires multiple searches, it's not that good at that. Cool. So, we'll try it, see how it works, and then— 

David Perell (00:39:05)

Yeah, it's a public company.  So, we'll see if it's there. But I'm excited to see if it works.

Dan Shipper (00:39:08)

So, we'll see if it works. And then I'll ask it to critique the strategy afterwards. So, we asked it, “Tell me more about how the New York Times makes money. What we want to know is what their annual revenue is, how many readers they have, and what are the most profitable parts of the business and least profitable parts.” So, the New York Times reached an annual revenue of approximately $2.31 billion, which was an 11 percent increase from the previous year's 2.07 billion. The revenue from subscriptions was significant, amounting to $1.5 billion in 2022, and they had 9 million subscribers in 2022. Advertising revenue was $523 million and, yeah, I don't know if it—Do you feel like it answered all the questions that you have?

David Perell (00:39:52)

That's a really good answer. So why don't we do this? “Tell me about the biggest risks in the business over the next 10 years.” So I just want to set this up. This is the bad answer. And I don't think that it's going to be that helpful. What I want to do is I want to look at this and then I want to compare it to the next answer. And I want to set something up about how to use GPT really effectively as a sparring partner if you're some sort of investing analyst where we're going to get some really specific answers.

Dan Shipper (00:40:23)

I love it. So one of the things that you should know is it just used Bing to search for the answer to your question. So rather than answering from its own sort of knowledge, it just Googled like, “What are the biggest risks for media companies over the next 10 years,” which is something that you might want it to do, but it will affect the answer. Cause it's basically just summarizing things that other people have written directly from its first search rather than just creating it from what it knows. So, that's just one thing to keep in mind. We may wanna redo it and have it not look on the web.

David Perell (00:40:54)

Ilove this. I mean, this is just buzzword bingo. Yeah, and this is ridiculous. “Cybersecurity threats, disruptive innovation ,and digital transformation.” I feel like I'm at the business school 301 seminar just like eye-rolling on a Tuesday morning. It's like, oh my goodness. “Pandemic-related challenges, supply chain, vendor risk.” Like these are just ridiculously broad-swath “climate risk,” like what, no. These aren't the actual risks, right? So let's get super specific and let's say, “Don't search the web, look up the work of Clayton Christensen, Michael Porter, and Ben Thompson, and focus on how the New York Times can continue to shift from advertising revenue to subscription revenue.”

Dan Shipper (00:41:44)

And tell me how you got that, why you want to do that versus the previous prompt, which was, which was much more general.

David Perell (00:41:50)

Great question. So GPT is magical when it comes to names. Names are like the ultimate compression. And that's what you want to do. Whenever you're communicating something, you want to think about compression, right? How much can I deliver in just a few words? And if I say, “Clayton Christensen,” what it's going to do is it's going to look up his Jobs to Be Done framework. It's going to look up his disruption theory. And just in two words I can subsume all of those ideas into all those ideas into that prompt and it can do the same thing with Porter's Five Forces. I can do the same thing with Ben Thompson's aggregation theory. And it'll probably give me something about aggregation theory, about Porter's Five Forces. And now what we've done is we have that tapestry of knowledge, and now we can start having a conversation amongst those three people to see where the Mount Rushmore of business strategists agree.

And then what do they disagree on? And then the fault lines of disagreement. Now we can follow those and we can see what the different vectors of strategy can be. 

Dan Shipper (00:42:52)

I love that. I love that. So we got our answer. So you want to, you want to read it out for us? Sure. 

David Perell (00:42:57)

So, I mean, this is exactly it.

So now we have—look at how much better this is. So we have “Disruptive innovation: Smaller, nimbler competitors who offer cheaper, more accessible, or innovative solutions that gradually move up market.” So that's what you're seeing with somebody like Chris Williamson's podcast is doing this. He is a one man show, basically, and he's moving up market with better and better production quality. Okay, interesting. Not a direct competitor, but these small media companies, that's what we're seeing there. Like Every. There we go. I mean, that's what this is, right? This is you moving up market. Hey, we're on video and stuff like that. Great example. Now, Porter's Five Force. I don't actually know what these are.

So let's look at them. So “threat of new entrants can enter the market with lower barriers, especially in the digital space.” Now that's a great answer. Okay. “So the ability to produce and distribute content has been democratized.” Absolutely. “Bargaining power of customers.” Huh? What's this? “Have more options for news and content their willingness to pay might decrease unless the New York Times offers distinct value.” I love those words “distinct value.” What does that mean? Let's index that, bookmark it, we'll come back to it. “Threat of substitutes.” Alternate alternative sources of news and information. I don't know if you saw what Steven Sinofsky wrote about Twitter a few months ago, but this is what's happening. We're seeing this rise of citizen journalism and media companies can no longer compete on certain vectors. So what happened in the late 20th century is a company like the New York Times. They could have privileged special access to certain high-level people. Now with Google and even just all the people who have boots on the ground, the people who are there can often distribute information faster.

The famous example of this was when the Sully U.S. Airways airplane crashed. The people on the plane, the people who were by the Hudson River, they were reporting better than a lot of the media companies because they were just there sharing information. “Bargaining power of suppliers.” I mean, I've written about this, that the cost of retaining top talent is getting harder and harder, right? You have people who are leaving these traditional media companies and they're going to launch their own thing, right? And the media companies have to do a better and better job of convincing the journalists who they work with to stay. “Industry rivalry.” Okay. I don't think that one is quite as interesting. “Aggregation theory: A risk lies in the potential dominance of digital platforms like Google and Facebook that aggregate news content.” I mean, there's no doubt. That's the big one. Right? So look at how much came from this. We have very specific things. And now it's given us, some strategies.

Now what I would be interested in, maybe for the next question, is “name the number one strategy for the New York Times to use. Period. Tell me a recommendation from Clayton Christensen, Ben Thompson, and Michael Porter. Give me an answer for each one that's no longer than one paragraph.” So now we're going to get a one-sentence strategy, we're going to get a paragraph to summarize, and we're going to get three different perspectives.

Dan Shipper (00:46:07)

Cool. So what we should first do is you want to ask it to first have it write out the recommendations, the bigger ones, and then have it summarize it in one. 

David Perell (00:46:15)

So one thing I'm getting a lot from you is, I'm asking too many things when I prompt GPT. I need to do one thing, one thing, one thing.

Dan Shipper (00:46:22)

One of the ways to think about it is, is, so yes, you want to have it ask it one focused thing if you can. And then also, in the same way that when I ask you a question, it can be easier for you to write out the answer before you—write out everything that you know about something before you compress it. GPT is going to be the same. So what you want it to do is like have it do its thinking out loud with you and then compress that thinking into something that's like, “here's my one recommendation.”

David Perell (00:46:48)

Interesting. So you're doing expansion then contraction and you really want to expand before you contract.

Dan Shipper (00:46:57)

Exactly. It's sort of like, when you do math problems, for example, if you write out all the steps in the math problem rather than doing it in your head, you'll have a more accurate answer. It's a sort of similar idea. 

“Tell me a recommendation from Clay Christensen, Ben Thompson and Michael Porter,” a strategy recommendation, right?

David Perell (00:47:18)

And then I would just say the question being, what is the one strategic move that we should make? Question mark. 

Dan Shipper (00:47:25)

“Is the one strategic move that the NYT should make—”

David Perell (00:47:31)

There's something here where I'm not necessarily trying to get the answer from GPT. What I'm trying to basically do is set up some sort of Socratic debate where I can see what the different dimensions are, the different vectors that people could walk, and then it's like Fitzgerald, right?

“The mark of a first-rate mind is to hold opposing ideas in your head at the same time and still retain the ability to function.” And that's what I'm trying to do. I'm trying to see, there's this idea, there's this idea, there's this idea, and then make some sort of strategic decision from that.

So let's see what it says. “Would likely advise to invest in creating or acquiring lower-end, more-accessible digital media products.” Okay, so what this would be—the way that I see this. So now—see I'm getting new ideas. And this comes back to what GPT isn't necessarily trying to do is give me the answer. What I'm trying to do is have it give me prompts that then give me the answer. So what I just got from this is what if the New York Times basically had a Barstool-style suite of independent creators so that they hit the low-end disruption? Because those are the people who would potentially disrupt them. And now they're under the New York Times. 

Dan Shipper (00:48:43)

I have to say, this is sort of the strategy that the New York Times is pursuing. They've got the Cooking app, they've got Games, they've got Wordle, they bought The Athletic. So they are definitely moving out of the traditional news product into a larger New York Times media bundle that targets non-news consumers and is at a lower price point.

David Perell (00:49:06)

Interesting. There we go. “The New York Times should focus on becoming a superior aggregation platform that delivers unique value to both users and content creators.” That's okay. What I would want It's more specificity there. 

Dan Shipper (00:49:18)

And I mean, I think what it's saying is Ben Thompson would recommend that New York Times become Facebook, which it never will be. So this is one of those things where with ChatGPT, it's going to give you that obvious answer. Like, yeah, I guess Ben Thompson would say that because of aggregation theory, but also Ben Thompson would never say that because the New York Times could never become Facebook, you know, it just wouldn't. It's just not realistic. So you have to like, you have to filter what it's saying through your own logic and intelligence and know that some of the answers are going to be really great like its answer about Clay Christensen and some of them are not going to be, they're going to feel like too obvious or too simple-minded.

David Perell (00:49:50)

And also we're seeing that what Clayton Christensen has written based on what GPT is pulling from is gonna be more useful than what Ben Thompson has said. Yeah, so then we'll get into Porter’s Five Forces. “Differenctiation by leveraging its brand reputation for quality journalism and in depth reporting and Investigative journalism exclusive stories expert analysis,” and this would just have me scratching my head—So basically we're seeing two things here: The first is how do we target the low end disruption? And what are the potential competitors that can come from the low end and this is what you were saying about the apps and what I was saying about the individual creators. And then the second thing here is now the New York Times has a seriously positive reputation and they have a big building on 40th and 8th in Manhattan so they have a certain amount of access, a certain amount of capital that they can deploy to do things that independent creators can't do. And so we're now pulling from both different sides. And then we're basically putting together a strategy there. I wouldn't have gotten there without this help. 

Dan Shipper (00:50:51)

And I really think this is totally right. That's their flywheel to some extent is they have the best reputation and the most money. And so they can get the best writers, which attracts more readers, more subscriptions, all that kind of stuff. And so they can just sort of keep going that flywheel on the high end of reputation for news. And then on the low end, they can acquire other media properties and other verticals, you know, sports or games or whatever, and then bundle them all together. And I think that was actually a really coherent strategy and that's basically what they're doing. So do you want me to summarize it in one sentence to see how it does? 

David Perell (00:51:26)

What I would say is give me a description of Clayton Christensen's recommendation in one sentence and Michael Porter's. So, I want to get one sentence on both of them. 

So we have, “Invest in and develop accessible lower-cost digital media products to attract a broader audience and preempt disruption from new market entrants.” That's what Clayton Christensen is saying. And then Michael Porter saying, “Focus on differentiating the brand through high-quality exclusive journalism and operational efficiency to enhance competitive advantage.”

So what I would just say there is the high-quality exclusive and then the lower-cost digital media products. I'd just be curious to know how much that aligns with what they're doing. 

Dan Shipper (00:52:17)

I think it's very close. I think ChatGPT solved it And that's the interesting thing about strategy. It's honestly similar to some of the stuff we did with personality and understanding yourself. Strategy is often pretty simple. It's just hard to do and it's hard to accept that that should be your strategy. Same for accepting who you are—strategy is just who you are for a company. And I think it's very good at distilling those simple sort of obvious things in a way that you're just seeing it in black and white for you. And then you can't really ignore it. 

David Perell (00:52:49)

One of the things that is becoming clear here is where GPT is uniquely good. So if Clayton Christensen was still alive and he was here and we could have the whole day to workshop strategy with him, we could end up with a better answer. Okay, fine. I'll concede that. I'll concede it. Fine. But, he's literally no longer alive. And We did this in 10 minutes, 20 minutes, right? So what GPT is really good at is along three dimensions—diversity. So we could basically have a room of Clayton Christensen and Ben Thompson and Michael Porter all three of them. So we have a very diverse group of people. I mean we literally couldn't get all those people in the same room now. First thing. Second thing is the accessibility, right? This is $20 a month. Yeah, so it's very cheap. And then the third thing is speed. We did this super fast. And it didn't take the whole day. So if we do this with any other tool where we evaluate it along certain dimensions. You wouldn't try to get a butter knife to cut a steak and I don't know if you've ever tried to get a steak knife to like put it in the little jar with butter—like it's ridiculous, but for whatever reason with GPT, we don't really look at what it's designed for, and I think diversity, accessibility, and speed on those three vectors. It's just fantastic.

Dan Shipper (00:54:13)

I think you're totally right. And I think that that's part of the thing with how much we anthropomorphize it where and all the super intelligent stuff. You're like, “Oh, it could do anything.” And then you try it and you're like, “It's not doing it right” or whatever. And it's like, no, this is a tool. It might feel human. It might feel like you can interact with it in the same way as a human being, but it's good at certain things. It's bad at others. I mean, if you've ever hired people, humans are like that too. If you hire someone and you, and you give them a job and they do it poorly, the first thought shouldn't be, well, I'm just going to fire them.

It's, well, what are they actually good at? And did I set them up for success here? And maybe they're not a fit for the role or maybe there's something else or some other way that they should be working in this role that is good for them. And I think GPT is exactly the same. So I think this is great.

We did the strategy stuff. I know you also use ChatGPT for reading and specifically reading old books. And that's so special to me. Cause like I've been doing the same thing. I'll show you. I started reading Moby-Dick recently. And the thing about reading a book like Moby-Dick, which ChatGPT is: One, it's in the public domain. So ChatGPT has read Moby-Dick already, which makes it a lot easier. 

David Perell (00:55:25)

Do you know exactly—Is it a hundred years where the copyright becomes public domain?

Dan Shipper (00:55:30)

I think it's maybe 75 or 100—something like that. 

David Perell (00:55:35)

Well, let's ask GPT. 

Dan Shipper (00:55:38)

Let's ask GPT. “When do books enter the public domain?” The other nice thing about GPT is I just never correct my typos because it just knows what you're saying.

David Perell (00:55:46)

I've noticed you're typing like a second grader. 

Dan Shipper (00:55:48)

You can just type as quickly as you want and it knows. 70 years. 

David Perell (00:55:53)

70 years. Nice.

Dan Shipper (00:55:54)

So I found these like old chats that I was doing where—you know, with Moby-Dick, it's a difficult book, you know? 

David Perell (00:56:00)

So you take a photo of it.

Dan Shipper (00:56:02)

I take a photo on my phone and then I upload it and then I say, “Can you interpret the last paragraph?” So in this paragraph it's saying, “By reason of these things, then the whaling voyage was welcomed. The great floodgates of the wonder world swung open and in the wild conceits that swayed me to my purpose. Two and two there floated into my inmost soul. Endless processions of the whale and midmost of them all one grand hooded phantom like a snow hill in the air” and I was just like “interpret the last paragraph” because there's so much good language there, right? But like I would literally need to be in an English class to actually know what’s being said.  

David Perell (00:56:37)

“The great floodgates of the wonder world swung open.”

Dan Shipper (00:57:02)

Exactly. So it's saying, “The last paragraph you shared reflects the narrator's deep inner conflict and yearning, Ishmael is expressing his irresistible draw toward the unknown and perilous, the wild conceits that lead him to his purpose.” And so after I read that, it just opened up the whole paragraph for me where I could match the underlying intuition I had that that paragraph was really important in the lyricism and poeticism of it was so great with my actual understanding of what was being said. And I didn't have to do it in the context of a class. I could just do it myself. And I was literally on an airplane when I was doing this. And, and that was so powerful for me, like made this book come alive. And the other thing that really does it—

David Perell (00:57:20)

You know, I had to go to reading comprehension tutoring when I was a kid. It was brutal. So I just couldn't understand books well at all. So my entire middle school, I was in reading comprehension tutoring. And this is basically what I would do is I would have to interpret paragraphs and I just couldn't do it. I couldn't do it. So for me going from—I still struggle with it a bit, like I'm not very good at understanding stories—and now going from kid who can't do it to tech-enabled adult who can do it actually better than even some of the best people in the world at looking at stories is a huge, huge level-up for me. 

Dan Shipper (00:57:57)

That's really funny, because I had the exact same thing. I was very slow to read, and my parents were very concerned that maybe, I wouldn't be able to read or whatever. And so they sent me in elementary school to what they called reading camp.

But really it was just after school, I would go to this place, and they would just drill me on words, and like flashcards and all that stuff to get me up to this, up to the point of being able to read and catch up with my classmates. And it turned out I was just sort of slow but once the light turned on I just like zoomed ahead and I was you know reading at a seventh grade level in like third grade or whatever and I think that that's where all this sort of comes from to some degree is as a as a kid I just had to focus so relentlessly on it that it became a strength.

So one last thing that I want to share, and then I really want to talk you through how you read with it. But one of the things that I think is really interesting about reading with ChatGPT is it can make the books come alive, not just through language, but through pictures.

Moby-Dick has all these scenes where he's in an old inn and there's whale bones everywhere or whatever. And so what I did is I took a picture of some of those scenes and I said, can you turn this into a prompt that I can feed into DALL-E to visualize it?

And it was a scene from the book where he's in a bar and so we got this prompt depicting a dimly-lit antique bar room inside the Spouter-Inn from the 19th century with a vast arched whale's head over the bar, all that kind of stuff. Which is it's taking the text from the book and it's just resummarizing in its own way.

And then I just threw it back into DALL-E. And then it gave me this image which is so evocative. I'm like, I'm there, you know. It makes the book feel like this different reading experience. That's way more engaging and I just love that. 

David Perell (00:59:43)

Wow. That's really cool. Yeah and just make that for you. Yeah, “dimly-lit antique bar…” That's impressive.

Dan Shipper (01:00:01)

Isn't it cool? And literally all I did was I didn't there was no prompting I just took a picture and I was like “turn this into a prompt.” And that's one of the really powerful things I think about the fact that ChatGPT is bundled with DALL-E because you're not prompting DALL-E directly necessarily. It can create a prompt for you from your raw material and then throw that into DALL-E and that prompt it creates is probably going to be better than the prompt you might create on your own.

David Perell (01:00:26)

Totally. The language here is so specific. To your point about reading with GPT, one of the things that I really don't like about the internet is that it's pulled us into this never-ending now. And what I mean by that is if you look at the way the internet is designed, it's all based off a recency bias. So you open up our social media feeds and basically everything was created in the last 24 to 48 hours. Same thing with Instagram stories, Snapchat stories. Even if you do a Google search, the news very much has a recency bias. And I think that this is terrible. Just terrible, because you want to read exactly the opposite way—for the most part. I mean, sure, sometimes it's helpful to know what's going on, but being informed has become way too much of a virtue that we're trying to strive for. And you always have to ask, if you're striving for one virtue, what is the thing that you're giving up?

And the thing that we're giving up is wisdom. And the way that Mother Wisdom shows her head is through time—through Father Time. So Mother Wisdom and Father Time work together. And Father Time basically filters information and what ends up at the bottom of the filter is Mother Wisdom. And so what you want to do if you're going for wisdom is you want to read old things, things like Moby-Dick, people like Nietzsche, things that have stood the test of time. And it's exactly the opposite of what the internet has done. So what I think that we should be doing is when we read with GPT—well, first of all, what GPT does is it doesn't have the same recency bias. You can literally, in your custom instruction, say, “I don't want anything before 1970.” I don't want anything. Okay. So you get out of it there. And also where GPT is really good, to your point earlier about things being in the public domain, is these old books. Because not only can GPT read the book, but also there's a whole consortium of scholars who have done work on books like Moby-Dick. So one of the things we could say to this is, “Give me three different perspectives.” And you just go on Amazon for five minutes, we say, “Okay, who are the three great scholars of Moby-Dick?” And then you can see what their disagreements are. So not only can you get an interpretation, you can see the disagreements in interpretation to see where the fault lines of discussion are and decide what you think. And now you're a layer deep into the text. So what I love about GPT is what it should be doing at a societal level is getting us to read more old books, which we should be doing.  

Dan Shipper (01:03:02)

I love that. And yeah—read old books, get them to come alive, and help them help us to understand them. I think that's really great. I know you said you wanted to talk through an old, an old book in particular. Do you want to do that?

David Perell (01:03:14)

So I want to talk through the Bible. So would you mind getting up, just prompting, GPT opened up a new chat. So, I do a Bible study every day, and I've just fallen in love with the etymology of words, and particularly in the New Testament, looking at the Greek.

Greek is a very—So what's cool about the Bible is Hebrew is a right-brained, much more intuitive language. Greek is a left-brained, much more analytical and logical language. So the languages themselves almost form to the two halves of the cognitive structure in the two brain hemispheres, which is beautiful to think that these two books are God's word. There's something to that. So there is—I think it's Acts 3. Well, here, let's just do this. Where in the book of Acts does the word Archegos show up? 

Dan Shipper (01:04:14)

A-R-C-H-E-G-O-S

David Perell (01:04:16)

Yep. I think it's Acts 3 and Acts 5. And then I think it's in Hebrews 2. Hebrews 12, but I'm not exactly sure. Acts 3. I think it's also Acts 5. Yeah. There you go. So then. Let's say, “Give me the translations of the two Acts quotes in the ESV translation, the NIV, the KGV and N-A-S-B.

Dan Shipper (01:04:47)

N-A-S-B. See if it does it. 

David Perell (01:04:50)

So what it should do is it should give us different translations. 

Dan Shipper (01:04:52)

Okay, let's see.

So give us some context here while it's writing this out. So let's start with Acts 3:15. 

David Perell (01:05:03)

Okay, so the Book of Acts is, so you have the four Gospels, and the Gospels are written by Matthew, Mark, Luke, and John. And Luke wrote two letters, and they were letters to a guy named Theophilus, and the first letter is the Book of Luke, and then the second letter is the Book of Acts. And Acts is about the story of the early church. It's basically a history book. And Luke is particularly interesting because he was a physician, and physicians then, as with now, they're very precise about their details. So what we're getting here is, through Luke in the Book of Acts, we are seeing the different ways that Luke's Greek writing is translated into English.

So the ESV and the NIV are more modern, slightly more accessible versions. Now, they're not as accessible as something like The Message, which really twists the word so that it can be easily understood. Actually, not necessarily in a negative way, so long as you know what is happening. Whereas the NASB is the one that I really like to read, and it's the most faithful to the original text.

Dan Shipper (01:06:13)

So it seems like I can't find the NASB, so maybe we should just find it ourselves. Let’s see—

David Perell (01:06:20)

Well, actually, we can just look at this right here.

Dan Shipper (01:06:21)

So this doesn't give the specific translations except for the first one. 

David Perell (01:06:28)

So, why don't we just say, “Show me how Acts 3:15 is translated in the English Standard Version, the New International Version, and the King James Version. So what we're going to see is the English Standard Version and the New International Version is going to use the word “author,” and the King James Version is going to use the word “prince.” So what I do when I read is I try to look for discrepancies. What are the discrepancies, and what do they reveal?

Dan Shipper (01:07:08)

Here we go. “You killed the author of life whom God raised from the dead,” “you killed the author of life but God raised him from the dead, and “you killed the prince of life whom God hath raised from the dead whereof we are witnesses.” 

David Perell (01:07:22)

Okay. Author. Prince. Yeah. How could author and prince be the same word? An author is someone who sits in a room and writes. A prince is somebody who leads.  So now we ask, okay, what's going on there? “What is the original Greek word for the word author and prince?”

And it's going to give us the word “archegos.” Okay, so now we have an insight. Okay. The word archegos means author, leader, prince, or pioneer. Those are totally separate things in English.  They're totally the same thing in Greek. So now we are seeing something through the Greek that the English doesn't reveal, that Jesus is the author, leader, prince, and pioneer.

He is the writer, type in John 1:1, and what is he there? The text is gonna say, “In the beginning there was the Word, and the Word was with God, and the Word was God.” So Jesus is the Word. And he is like this author of life. Then, from that, he becomes a leader of this new movement, and he's almost this prince, right? Jesus reigns, he's the king, he's the prince, and he's this pioneer of a whole new way of living. So then you're like, okay, we can ask, “What are some examples of an archegos in the modern world.”

Dan Shipper (01:09:17)

That's really interesting. 

David Perell (01:09:09)

And I think you have a few examples. You have someone like Jeff Bezos. Jeff Bezos is an author. He's effectively a prince. He's definitely a leader. And he's wholeheartedly a pioneer. 

Dan Shipper (01:09:22)

There you go. It went right for Elon Musk and Steve Jobs. 

David Perell (01:09:28)

Okay. Now apply this to, ask it to apply to the American founding fathers. Ben Franklin, Thomas Jefferson. This is what they did. You don't just think of them as these people who were founding America. What did they do? They wrote the Federalist Papers. They wrote the Constitution. They were these authors. So you start with being the author. Then you go into being the founder from being the Satoshi, right? He is the author, then he founds the thing, pioneers the new thing, becomes the prince. And you see the same thing with George Washington and Thomas Jefferson, right? So, type that in. And now what you can basically see is for me and my work running Write of Passage, I can now make an argument for writing that is rooted in this Greek word, that the Greek has revealed to me that all these things that I thought were totally different are actually one in the same. And that then is why you should write. And I can only get there by following the etymology of these words. And GPT is really useful because I can see the different translations, look at the discrepancies  and get my answer. 

Dan Shipper (01:10:30)

That's really beautiful. I think in particular why it resonates with me is there's always this tension between being a writer and doing things in the world. And I think one of the reasons why it was hard for me to admit that I wanted to be a writer is because there is this part of me that wants to run businesses and wants to do things. And I have all these ideas and I want to make an impact in the world. And it was hard to reconcile that with the writer part of me. And also, I think there's a lot of social pressure to not follow the writer thing and do this other thing in the world. And this is like a really beautiful embodiment of the path from writing to changing the world. 

David Perell (01:11:05)

It's not the path. It's the same thing. It's the same thing. What you were seeing is two separate things. The Greek is showing you that they're linked together, that through authorship you then get world change. And if that is your tack, you now, through this Greek word of archegos, You have a pin that synthesizes all of these seemingly conflicting desires and is showing you. And so then what you can do is, then we can write, “what are examples of archegos in ancient Greek literature?” And we might get something from Homer, we might get something from somebody else. And you can actually get different archetypes of them. And you can say, I resonate with that one, I resonate with this one. And almost by going back in history and diving into the mythology. There's something that's concrete and real about that. You have a whole tapestry of stories that you can draw from. 

Dan Shipper (01:11:57)

I love that. I think that's really great. Thanks for sharing that. I mean, that really touched me. It's such a deep thing for me to figure out how to be a writer and to do things in the world, and I feel like that's such—it's just been my sort of quest to admit that to myself, that that's what I want. And we have this thing that we say at Every that's basically “the stories we tell become the businesses we build.” And that's been a thing that I have come to myself without this word for it,  which is that, Elon Musk doesn't become Elon Musk without Isaac Asimov. And I think people really underestimate the degree to which stories motivate and inspire to do things in the world. So even if you're not the one going in, going in, building the business yourself, telling the stories that you want to tell about what to build and why is going to change what gets built and I think archegos is a really clean way to express how that works. So I really appreciate you sharing that.

David Perell (01:12:58)

I love that word. I feel the same thing. There's something that's so galvanizing—you hear the author becomes the founder who becomes the pioneer who becomes the prince. And the Greek gets it and we don't have that in English. 

Dan Shipper (01:13:25)

Totally. I want to talk more about the writing stuff so obviously reading old books is the first step. It's like doing that taxonomy, doing that deep engagement lets you swim in these ideas that light you up, but then you're kind of like, “Well, I gotta go do something with that.” I want to reflect that through my own prism into the world, through the work that I do. And one of the ways that you do that. And I'm super curious to hear about how you have used ChatGPT in your own writing process. 

David Perell (01:13:51)

Yeah, so there's a few things. So one of the big ways that I think about my own efficiency gain is I don't really like typing very much. I love talking. So one of the best things that GPT can help me with is if I can just talk through something, and I'll do it when I'm giving feedback, I'll do it when I'm thinking through an idea, I'll talk through an idea, and then I'll ask GPT to summarize it.

“What is the main point that I'm trying to make? What are the core bullet points?” And I talk through ideas, and then I get GPT to help me structure and synthesize them. 

Dan Shipper (01:14:26)

I love that. I have that too. I'll often just go on a walk and then just blab—just free associate. Yep. And then I'm like, “What was interesting about this?” And it'll just make an outline and it'll just pull out those little things that I can use to write a piece or whatever. I think it's also really great for—I don't know if you have this, but I'm sure you do. But sometimes when I write a piece, it's something I've been thinking about for so long. And I have one big note that just a giant list of all the ideas and quotes and whatever. And I'm just like, where do I even start? And it's really good to just paste it into ChatGPT and be like, “make an outline.”  And the outline is going to be mostly wrong. But what it's really good at is I think when you've been noodling on an idea for so long and you're like, “this is so special.” Like your essay on Peter Thiel, it's like you have something that really big that you want to say. You can often forget that there's like a couple of really basic forms that an essay takes and that the form is probably staring you. Staring you right in the face because you're like, well, this is my special thing it'll probably be a different basic structure, but it's like, no, a lot of times it's like "thesis, problem and then solution or something like that, you know? And ChatGPT is really great at taking your long, complicated thing and being like, “Okay, just talk about the problem at the top and then talk about your solution right afterwards.”

David Perell (01:15:43)

Well, there's a deep psychological point here that's been a thread through this entire conversation: Humans aren't very good at creating things from scratch compared to responding to things that already exist. So if you ask somebody, “Hey, what do you want your living room to look like?” and you're a designer. I don't know. I don't know. But if you take that same person and you show them a bunch of furniture, they can tell you what it is that they like and don't like. So the point of that exercise is not to have the perfect living room. It is to create ojects that are concrete that then you can start rearranging and moving around. And just when you go from nothing to something with which GPT is really good at, now you're responding and you can start shaping marble that already exists rather than having to go out and find the marble in the first place.

Dan Shipper (01:16:38)

This is the thing that I think is so critical is I think people listen to you talk and there, and you're so articulate and succinct about what you like and why. And I think there's like a thing that's like, “Well. I don't know if I could ever do that and he's got such great taste” and I think there's this thing that's like it's something that you develop by trying and everyone has everyone whether you know why or not you have things that you like and dislike and you've gone undergone this practice of feeling what you like and don't like and then trying to think about it. And put words to it. And once you have words to it, you can refine and narrow and, and there's so much power there for creating new things. And I think that's the thing that we keep coming back to about why it's so valuable that ChatGPT can name those things because, it can take you from the thing you already know, which is already inside you and is intuitive about what you like and it can reflect it back to you in words that you can then use to narrow and refine, and that's incredibly powerful as a creative tool. 

David Perell (01:17:36)

Yeah, you could almost say that—Alfred Whitehead has a line where he says, “Civilization expands by the number of functions we can perform without having to consciously think about them.” Great line. There's something similar when you're doing creative work is that your ability to lead a creative team expands by how many things that you've made concrete that were initially only obvious to you through feelings.  And if you can take feelings and put them into words, there's always lossiness in a translation. There always is. It's the same way that if you're converting energy from one unit to another, there is something that gets removed. Yeah, it's never perfectly efficient. That's not what we're going for. But if you can take—Say that you have a hundred units of intuition and go from being able to describe eight of them to 74 of them you will be dramatically more productive and that's what GPT is really good for is going from “I can't express this” to “I can” and then telling other people about that. And then get this now that you have it in words that are concrete they can cross-reference ideas with the words that you've given them and now through GPT, you've basically outsourced you as a sparring partner so then they can deliver you something that's of a much higher quality.

Dan Shipper (01:19:01)

Yeah, cause they can be like “I know David likes things that are older and aren't in the never-ending now, and I know that he likes things that are precise and clear and and evocative, and it's like, is this that as opposed to trying to just simulate you in, in their head without having the words.

David Perell (01:19:26)

Right, exactly. So, for example, I really like Rococo design and I really like sort of French interior design and 18th and 19th century. And then I also like decorative arts that are a little bit more masculine in their style. So now I can take those three things and I can, because it's concrete, we can say, and then, then I might like the colors and styles of Persian rugs. And now we just have four things, and GPT can easily pull from that to help us find the shape of what we're going for. And now, like you said, with the DALL-E integration, now we can get images and words in our output. So basically, what we're finding is that the returns to making your thinking legible and clear have skyrocketed with this technology.

Dan Shipper (01:20:13)

100 percent. And it is a tool for making your thinking legible and clear and then getting a return on that investment.  100percent.

So one of the things I love most about your writing is I think you've mastered the art of the anecdote. You're really good at finding a little anecdote that like really hits and like really gets you right at the top of the piece to be like, okay, I want to lean in and understand this. And I noticed that because I feel like I'm terrible at anecdotes. I feel like I should have this library of stories that I can use to elucidate my examples or whatever. And I've just, I've never really had it, but I've started to use ChatGPT for it. And I'm like, it's pretty helpful for it. And I know that that's something that you do. And I just really love for you to talk through how you find those anecdotes and how you use ChatGPT to do that. 

David Perell (01:21:06)

Okay. I'm going to need your help with GPT. Please. So, what we're gonna do is we're gonna write a piece on how New York City, just by virtue of living here, allows you to have better taste. So, what we're gonna do is we're gonna say, “What buildings in New York are the most beautiful and distinctive?” Question mark. “Give me a series of buildings built between 1910 and 1933.” 

Dan Shipper (01:21:42)

And why are you asking that? 

David Perell (01:21:46)

Because I want to find an anecdote. 

Dan Shipper (01:21:46)

Why'd you start there?

David Perell (01:21:47)

So, what I'm looking for is I want to do something specific, and then hopefully, it'll give me a building that I've been to another way that memory works is memory is basically layered and sort of behind doors. So what we do is we outsource our memory to little checkpoints that we find and the way that I know this to be true—I'm not a psychologist—is when you go back to your childhood home and you're back in your childhood room. How many things do you think of that you hadn't thought of in a decade that all of a sudden become clear? That's because those memories are stored in space. How do the world's memory champions think of their ideas? They associate them with spaces, and then they walk through the house, right? Moonwalking with Einstein. He's going through and he's walking through the house and then when he sees that space he thinks of a new memory.

Dan Shipper (01:22:39)

So you're using this basically to prime your contextual memory to surface something that might be interesting. Right. So yeah let's read through this like what are you seeing and what's it what's it making you think of?

David Perell (01:22:38)

So we have the Woolworth building. Here we go. So, now, I just had an anecdote that showed up in the very first example, and we're gonna change what it is that we're writing about. Okay? So, many years ago, I walked into the Woolworth building, and I had a meeting. I was in college, and there was a guy who was big on Twitter who I was very intimidated to meet with, and he was working there, and we were talking about hiring, and he had a great line. He said, “hiring and dating is a max function.” So, here's what that means. It means that certain things what you're trying to do, so for example, if you're designing, furniture in a room, it's not a max function, because you need a bunch of different things that need to sort of play together and stuff like that, whereas if you're just working on hiring one person, you could have a 99 total duds, but it doesn't matter as long as the one person who you're trying to find is good. You only marry one person. You only hire one person to be the new CEO for your company, right? And I remember that conversation so vividly and the Woolworth building basically prompted that. And I could start that with an anecdote and I could basically say there are certain parts about the world where the averages really matter. So ask what is the opposite of a max function? I don't actually know. 

Dan Shipper (01:23:57)

A min function. 

David Perell (01:23:59)

Maybe? Well, there's probably max min, and then there's probably a distributed one. So, there's some functions where all that matters is how good your top value is, and then there's other things where how many things you have are really useful. And what I would just say is now we've just set up a piece that basically presents two different ways of walking through the world in two basically fields that you can play in. Certain things are about maximums and minimums. Certain things are about the broad swath. So now you asked about an anecdote. Yeah, we went into GPT. We found the anecdote. I was this college kid. I was nervous. I was an intern I didn't actually know what this meant and now we can basically write a short piece to say, whenever you enter a working arrangement, whenever you enter a project, you can ask, is it the max function or is it this other thing that we're setting up? And now we've just given someone a world view just off the top of our head. 

Dan Shipper (01:24:53)

I think that's really, really amazing and it's like, and one thing that I want people to notice is that we went in thinking we were going to write an article about New York City and taste. And we came out with this max-min thing or max and average thing. And that's just how the creative process works. That is not a mistake. The mistakes are the things that work that turn into things that are interesting. And what you did is you had a little thing where you're like, “Oh, I'm inspired about New York. I want to think about New York stuff.” You threw it into ChatGPT. It reminded you of this other story and you just follow that thread instead of being like, “Oh, no, this is not on topic.” Like I should filter that out. You're like, “No, this is really interesting.” And I think that's what ChatGPT is really good at is sort of pushing your brain in new, into new areas that you wouldn't have thought of. And one of the things that you're good at and all creators have to get good at is allowing that to happen without filtering it out and being like, that's not on topic is just following the spark. And I think you just did that live and that was really cool.

David Perell (01:25:55)

Well, I have this idea called writing from conversation and that's what GPT allows us to do. So the way that I think about writing is I'm gonna have a bunch of conversations with friends, with coworkers, with mentors, and through those conversations, I always think of conversations as an algorithm for randomness. So the way that the mind works is it gets caught in these grooves, it gets caught in these cycles. What a conversation can do is it pulls you out of those, right? John O'Donohue has a line where he says, “When was the last time that you had a great conversation? A conversation that wasn't just two intersecting monologues, but a conversation, effectively what he's saying is where two people are in new territory, they're actually ascending onto a new plane, finding things inside of them that they never knew that they knew.” And then he ends it and he says, “Conversations like that are food and drink for the soul.” And until the development of this technology, you could only do that with another human being. And now you can do that with an extremely smart computer.

Dan Shipper (01:25:53)

I love that. I think I think you're totally spot on and I think that's a wonderful place to end it. This was an incredible conversation. I learned a lot. I felt touched honestly that we got to do this together. So thanks for doing it.

David Perell (01:26:57)

Thanks for hosting me, man.  

Dan Shipper (01:27:04)

See you soon.

Thanks to Rhea Purohit and Scott Nover for writing and editing support.