Podcast: Play in new window | Download | Embed
Subscribe: Apple Podcasts | Spotify | Android | Pandora
DAVID PAKMAN — political commentator, YouTube host, and author of the new book Pay Attention — joins me to talk about Algorithms, AI, and the Attention Economy. We break down how attention became the core commodity of the modern economy, why algorithms reward outrage over nuance, and what happens when AI starts cloning creators without their consent.
David walks through the real story of “The Daniel Carter Show,” a YouTube channel that used AI voice-cloning to impersonate him and was algorithmically recommended to his own audience. From there, the conversation covers why right-wing messaging has a structural advantage on algorithmic platforms, what “behavioral surplus” means and how platforms monetize it, and the media-literacy habits families should be teaching the next generation before handing over unrestricted social media access.
We also discuss anonymity and accountability online, why the “calm middle” is checking out of social media, and a simple “media diet” framework for staying informed without getting consumed.
Themes
Why attention is the defining commodity of the digital economy
How “rented ground” platforms (YouTube, TikTok, podcasts) shape what creators can say
The real story behind an AI deepfake impersonating David Pakman
“Behavioral surplus”: how platforms monetize what you don’t even realize about yourself
A practical media-diet framework — and how to talk to kids and aging parents about safe media use
Timestamped Show Notes / Chapters
00:00 — Cold open: “You just have to learn how to play the game” — David on the political right’s platform advantage
00:15 — Wealth Actually intro & disclaimer
00:40 — Introducing David Pakman and today’s topic: algorithms, attention, and responsible media consumption
00:59 — David’s new book Pay Attention: why he wrote it and the “confessional” behind-the-curtain angle
02:38 — Attention as the “lingua franca” of the modern economy
03:25 — Creating content on “rented ground”: YouTube, TikTok, and podcasting platforms as landlords with no clear lease
04:44 — How Facebook’s shift from chronological feeds to algorithmic “For You” feeds rewired the entire incentive system
06:36 — Why boring-but-important topics (like housing policy) need an emotional hook to break through
07:37 — What the political right does better on these platforms — and why it’s about style, not substance
08:01 — The COVID vaccine messaging case study: fear and scapegoating vs. accurate-but-boring nuance
10:10 — From the traditional GOP to the “Trump party”: the John McCain/POW moment as an inflection point
11:05 — The pre-Trump roots of modern right-wing messaging (AM/low-power FM religious radio, Rush Limbaugh) — and the recent press-corps ban of CNN, Politico, and MSNBC from the White House briefing room
12:08 — Why the “calm middle” checks out of social media, and how outrage drives engagement
12:40 — The trap of over-weighting vocal-minority feedback: “98% of my Twitter replies are anti-social screeds”
14:29 — Online anonymity, accountability, and Jonathan Haidt’s real-identity proposal
14:55 — David’s own experiment: requiring a real email address cut toxic messages by 95–99%
16:10 — AI and deepfakes: the danger of synthetic content built on real people’s data and personalities
16:44 — The “Daniel Carter Show”: how an AI voice clone of David was algorithmically pushed to his own audience
18:59 — “Behavioral surplus” — how platforms monetize the gap between what you say and how you actually behave
21:31 — David’s media-diet “food pyramid”: proactive, vetted sources as the staple; algorithmic scrolling as junk food
24:09 — Advising families and next generations on safer, smarter media habits
25:17 — David’s own household rules on kids, screens, and delaying unrestricted social media access
27:12 — The financial-literacy parallel: media literacy as a life skill that should be taught early
27:58 — The media-literacy pyramid — and the Pew Research finding that most Americans (across party lines) struggle to tell fact from opinion
29:10 — The AI risk multiplier: outsourcing critical thinking to algorithms
29:32 — Where to find Pay Attention and follow David Pakman
30:19 — Closing thanks
Transcript
[00:00] “So I actually think that from a political standpoint, if what people are trying to do is use these platforms to get the ideas they believe in out, you just have to learn how to play the game. I don’t think there really is any other solution on the political side.” — David Pakman
[00:15] [Show Intro] Welcome back to the Wealth Actually podcast, the show that features experts, entrepreneurs, and commentators that will give you the right knowledge, planning, and guidance so you can preserve your assets and enjoy your wealth. Learn more and subscribe today at WealthActually.com. This podcast is for educational and entertainment purposes. It is not investment, legal, or tax advice. It does not represent the opinions of the employers of the host or guests. David Pakman has a multimillion person following and an outsized presence with his online political talk show. He knows a lot about how the algorithms work across platforms like YouTube. We’re gonna talk a lot about that and also how to responsibly consume media in this day and age
[00:56] Frazer Rice: David Pakman, thanks for coming on.
[00:58] David Pakman: Thank you.
[00:59] Frazer Rice: You’ve got a new book out. Pay Attention. By the time this podcast launches, it will be fully out and about on Amazon and wherever else everybody gets books. Tell us a little bit, just briefly, from the book writing process, your major voice in media, and what got you writing on this topic?
[01:18] David Pakman: Well, there’s a confessional aspect to the book — just kind of explaining the compromises as a content creator that I have to make. And I say “have to” in the context of, if I want this to be my full-time job, if I wanna succeed at it, there are compromises that have to be made, as is the case in most jobs. So one aspect of the book is kind of pulling back the curtain a little bit. The other aspect is about best practices at the individual level, when so much of what we come to believe is true — and sometimes isn’t — is mediated by these digital platforms that are algorithmically delivering content in a way that is not impartial. It’s so important to understand why we are fed the things we are fed rather than other things, as a starting point for having conversations about what we want the world to look like. And then I also have a political interest in this — as someone on the political left, I think the political right naturally benefits from the way these algorithms work, not because of anything special they’ve done, but from their communication style. So if the political left wants to succeed — which I think it does, at least it claims to — it has to better learn how to conduct itself on these platforms. We’ll dive into that in a second.
[02:38] Frazer Rice: One thing in your book — and I think generally that you talk about — is the commodity of attention, and how that has become really the lingua franca of the modern economy. Maybe dive into that a little bit about why that has become the case.
[02:55] David Pakman: Tell me from which perspective do you want me to talk about it? ‘Cause that could go in a bunch of directions.
[03:00] Frazer Rice: Well, I think the main thing is to say: if people are sort of used to oil or plastics or gold being a commodity, the idea of attention really driving, in many ways, economic decisions or allocation decisions around productivity — that seems to be something that’s a little bit more nuanced in the last, certainly five years, but maybe even back the last 10 or 15.
[03:25] David Pakman: Well, I think that conversation kind of starts with understanding incentives. I create content on what I describe as “rented ground” — YouTube, TikTok, podcasting platforms — all of these platforms are basically lending me space, for as long as it’s at their pleasure to do so, to publish my content. And they’re landlords that don’t really give me a clear lease that tells me what I’m allowed to do and expected to do. I kind of have to figure it out by doing it. Their incentive is to keep users on the platform as long as possible — the longer users are there, the more money the platform makes. So when we look at the content we get from these platforms, it’s not because the content is true that we’re shown it. It’s not because the content makes society better or makes us better citizens. We’re shown the content that the algorithm determines is most likely to keep us on the platform. That’s really the starting point. One of the critical changes: I don’t know how early you were on Facebook, but at the start, when I was in college, if you were friends with 20 people, your home feed was a chronological feed of what your friends posted. Very simple — if you looked on Saturday, went back Sunday and scrolled, at some point you’d hit the last thing you saw Saturday, and you were kind of done. There was a clear break point — “okay, I’m caught up, I’m getting off Facebook.” That’s a real problem for a platform that wants to keep you on there as long as possible. So eventually Facebook and Twitter switched to an algorithmically recommended feed. TikTok calls it the “For You” page, and it’s just the default. What that does is show you content not in chronological order, but based on what keeps you watching.
[05:52] Frazer Rice: At a much, much smaller scale than you do, I get frustrated oftentimes saying, I’m putting out things that I think are interesting and good that get a good response from a narrow set of people. But if I were Sydney Sweeney or a dancing bear and able to proliferate content in those terms, I’m sure it would be a lot more, I guess, metrically successful in terms of eyeballs and duration. And it sounds like as you go through and create your content, and you get the tools YouTube gives you — retention curve and things like that — that shapes, in many ways, not just the content but the delivery from your end.
[06:36] David Pakman: I try to limit the degree to which it shapes the content, but it certainly shapes the delivery. If you take a topic like housing — extraordinarily important, because everybody needs a place to live, and it interacts between municipal, state, and federal law — how do we talk about housing in a way that’s interesting? Definitely not starting with zoning or building permits and regulations and Section 8. There has to be some emotional core to it. One of the ways the political left can more successfully address that: “Do we agree that if you work full-time somewhere, you should be able to afford to live in that same place rather than commute an hour away?” That gets to the emotional core, and then we fill in the “boring stuff.” So the way I try to say it is — we’re thinking about presentation style, not about ignoring important issues.
[07:37] Frazer Rice: So as we start to think about left and right, and who’s — let’s call it — good at the platform and good at what’s happening, and maybe less good: what does the right wing have as an advantage in this space that the left wing doesn’t? And I’d go so far as to say the left wing’s experience in traditional media isn’t translating over into this new way of communicating.
[08:01] David Pakman: Yeah, that second part is really important. I want to make clear — I’m not making value judgments here, just trying to analyze what is. One of the things right-wing messaging over the last couple of decades in the U.S. has done very well: number one, pithiness and simplicity, which naturally work better, especially in short-form, low-attention-span formats; generating anger or fear as an immediate visceral response, in a much better way than the political left does; and often implying — or sometimes overtly naming — the scapegoats, enemies, and villains we should be pointing to. I’ll give an example — and I know some people on the right will say “that’s not what I believe,” totally fine, but we’re talking at scale. Imagine the COVID vaccine messaging: there was a period where you had people like Tucker Carlson and others saying “this vaccine has killed millions of people.” It’s extraordinarily short. It makes people afraid — “wow, I could have been one of those millions.” It makes people angry — “who did that to us?” And it creates villains: was it Dr. Fauci? Was it Biden? Even though Trump was taking credit for the vaccines when convenient, then it became less convenient — but put that aside. It creates villains and scapegoats out of pharmaceutical companies. It happens to be completely untrue, but it naturally does really well on these platforms because of those characteristics. Now compare a message like: “It would be great if this vaccine prevented transmission — it’s not great at that, but it does keep people, for the most part, alive and out of the hospital. It’s also safe. It might make your arm hurt, you might feel flu-like for a couple days.” It’s long. I’m explaining. There’s no obvious person to be angry at. There’s no particular reason to share what I’m saying. So aside from the truth of any of this, that second message just doesn’t do as well by the incentive structure of these platforms.
[10:10] Frazer Rice: Back on the Republican side for just a second — maybe not “Republican,” but right-wing. I get back to an example where the Republican party I grew up watching changed and turned into more of the Trump party. The messaging and mode of messaging — I can almost pinpoint a spot: when Trump went after John McCain and basically said, “I don’t want my war heroes to [have been] captured.”
[10:40] David Pakman: Exactly.
[10:41] Frazer Rice: And I looked at that and said, “well, game over for Trump.” It didn’t happen that way — he swept aside a whole party, and then his — let’s call it Roy Cohn-inspired — way of communicating took over. Is that really the right-wing component, or is there something more GOP-oriented, or a combination that gives them an advantage?
[11:05] David Pakman: A lot of this stuff predates Trump. In my first book, I tell the story of how a lot of modern Republican messaging, even before Trump, came from the right wing picking up AM and low-power FM radio licenses early and putting religious content out. Churches were very early to AM and low-power radio — and eventually that led to Rush Limbaugh. The origins of this communication style are older than Trump. Now, I think you’re right that Trump changed it — he went much more populist in his messaging, and ultimately became significantly more authoritarian. As we’re recording this, CNN, Politico, and MSNBC were just banned from the White House press briefing room by the administration. George W. Bush wouldn’t have done that. So there’s definitely a significant change there — but the style, if not the substance, really predates Trump by several decades.
[12:08] Frazer Rice: So as we get back to how the economics of this work — ultimately the platforms have to make money. They have to get the eyeballs, and then the attention beyond the eyeballs. We talked about how outrage drives acquisition and retention. What makes the “calm middle” close the app and move on, out of fatigue, I think? Does that make sense to you, or do you analyze it differently?
[12:40] David Pakman: Yeah, I don’t know — to the extent the “calm middle” exists on social media, they’re still there, but underrepresented in participating. This is tough for me as a creator: I may be motivated to look at the feedback I see online, but I have to remember that represents a tiny portion of my audience. I’ve been on the internet for decades and maybe left 10 comments in my life, never written into a show — I have to remember I’m not hearing from most people. If I look at my Twitter replies, 98% of them are anti-social screeds — the kind that would get you fired or blacklisted from your social circle if you behaved that way in real life. That’s one of the really pernicious effects of these platforms: they narrow the scope of topics in the discourse, because a lot of stuff doesn’t do well; they generate disproportionately negative reactions to differences that, if we met in person, would probably feel far smaller; and they make people think disagreement is inherently uncomfortable, when it doesn’t have to be. I’m both Jewish and Argentinian — two cultures very much known for arguing and debating politics and religion all the time, and then going back to “how’d the brisket come out?” I think the way we engage in these conversations online makes people believe disagreement is so uncomfortable that we need to wall ourselves off from anyone with a different view — and that’s a dangerous impact.
[14:29] Frazer Rice: Where do you come out on quasi-anonymity when people engage online? If you have a Twitter account that isn’t quickly traceable back to your actual self, I assume it unlocks fiery opinions that people would otherwise feel accountable for. Is there a world where that accountability shows up more than it does now? Jonathan Haidt has put out a number of ideas—
[14:55] David Pakman: —as to how he’d change the way people exist and engage online. One thing he said is there really should be no fully anonymous users — that doesn’t mean you post under your real name, but the platform has to know who you really are. Even that would change the level of accountability people feel. On my own website, I used to have a contact form that just checked the email field for something that looked like an email address — so “abc@123.com” would go through. When we changed it to require you to actually send a real email from a real address, the volume of angry, toxic messages dropped at least 95%, if not 99%. If you’re really motivated you can make a Hotmail account just to write to me, but most people won’t — most people are lazy. Just adding that layer of a traceable email address totally changed people’s behavior. There’s a lot to that idea.
[16:10] Frazer Rice: Let’s talk a little bit about AI — the robots creating content under the auspices of somebody’s identity. I just did a podcast with some estate planners who talked about the idea that Facebook, et cetera, are going to have the tools to generate comments and responses out of people’s personalities even though they may be dead 10 years prior. The danger of generative AI using the data underpinning different users’ personalities — how do you relate to and diagnose that risk?
[16:44] David Pakman: Well, it’s sort of the last section of the book. For creators, it’s a huge risk. We just dealt, over the last week, with a YouTube channel called “The Daniel Carter Show.” If you looked at it, there was a guy who looked sort of like me, in a studio that looked sort of like mine — and it was quite literally my voice, generated by a voice-cloning AI system, which are very good now. Written scripts were fed into the machine, which spit out this guy, “Daniel Carter,” who sounded just like me. Some audience members came across this — interesting, because it means it was algorithmically suggested to them: “if you’re watching the real David Pakman, here’s something else you might like.” They wrote to us about it; we reported it for using my likeness, and it was shut down. But it’s going to be very easy to generate a thousand of those, ten thousand of those — essentially a nearly limitless supply. I don’t know what the technological solution really is — it’s going to be AI detecting that stuff while AI is being used to make it. It feels like a race to the bottom. The one thing I write about in the book that I think is relevant: people do seem to recognize the importance of real connection. I’m doing some in-person book events right now, and I’m like, “who’s going to come out to hear me talk about my book?” — and all the events are oversold, selling out, because there’s a real desire to connect with people in a genuine way. I don’t think people are going to embrace synthetic creators and news people once they can tell that’s what they’re being fed. Now, if it gets so good no one can tell — that’s maybe a different story.
[18:59] Frazer Rice: You talk a little bit about “behavioral surplus” — I’d like you to get into that concept, where the platform isn’t just extracting actual data, but the gap between data and behavior.
[19:10] David Pakman: Yeah, there are a couple of different ways this happens. One of the things I try to do in the book is not become conspiratorial — there’s really no conspiracy behind this, it’s just that if you follow the incentives of the platforms, this is naturally what they’d do. Platforms monitor user behavior not only during a specific session, but over periods of time. What we want to consume can be very mood-dependent, and often tied to time of day — are you at work or home, is it a weekend or weekday. There are times you’re more likely to buy a product advertised a certain way than at other times; times you’re looking for political content versus musical content. What these platforms are learning to do, in a programmatic rather than conspiratorial way, is figure out when there’s a “behavioral surplus” — an opening to nudge you into actions beyond just what you’re watching, in ways beyond even our own ability to perceive or understand. That gap — that delta — is directly monetizable. AI and big data are only going to make it easier to exploit that in ways so subtle and sophisticated that many of us wouldn’t even realize it.
[20:47] Frazer Rice: Interesting, and scary. For some of us it wouldn’t be — I stay curious, and I view social media from a peculiar lens; it’s a great way to interact with people I have no idea about, and even useful with people I know as friends. But there are some people who just say “I give up, delete the account, I want no part of this, I’m going to hide under the covers and hope this all goes away.” That doesn’t seem like your direction either. How do you think about staying engaged without being defeated by the enormity of what we’re talking about?
[21:31] David Pakman: It’s funny — a lot of people react to some of the suggestions I give in the book, which are: yeah, scrolling is fine, but think of it like a food pyramid — it should be the junk food. Mostly, your media diet should be things you’re proactively vetting: “these are the four podcasts, these are the four online magazines I’ve evaluated and trust” — pulling their content in rather than being fed by scrolling. That should make up the bulk of your media diet. But people sometimes write and say, “David, aren’t you worried you’re advocating people not watch shows like yours, putting yourself out of business?” I laugh, because if this book is wildly successful, maybe 100,000 people read it — relative to my existing audience and the number of people scrolling incessantly on these platforms, that’s such a small fraction that even if all 100,000 never go on YouTube again, I’ll be okay. And these platforms are so addictive, so tuned to sucking people in, that I have zero concern about saying “you really shouldn’t consume that much of this.” On a personal level, spending less time on these platforms is great — or getting off them entirely is a completely healthy choice, the only downside being the modest social component some people use them for, which most can find another way to maintain. Now, there’s also the political side of this…
[24:09] Frazer Rice: …when the right wing would find themselves frustrated by traditional media, their solution was to take advantage of a new technology and opportunity — so it’s probably cyclical in terms of who benefits from the platform going forward. I deal a lot with advising people across generations, and there’s always a comment from the senior generation: “what can I do to make my kids safer, better users of platforms, better consumers of media, or protect them from the dangers out there?” There are a lot of things you can do, but the curated media diet is interesting to me — I like having different sources, different opinions I might not agree with, to test my priors and stay rational. What do you do yourself on this front?
[25:17] David Pakman: I do try to get most of my content proactively — I’ve decided what I want to listen to, and that’s the bulk of my media diet. My own content isn’t on a lot of the scrolling apps, so I spend some time there but try to limit it. I have young kids too young for devices right now, but at some point, sooner than later, it’ll become an issue. I don’t know exactly where I’d put the age number, but what Cal Newport and Jonathan Haidt describe as “unrestricted” social media access shouldn’t come anytime soon for kids — and it may be a bigger problem for girls than boys, given how boys versus girls tend to interact with each other, though that gets into the weeds. One thing really missing in U.S. public education is media literacy — understanding: is this neutral content? Is this an advertisement, and what’s it trying to do? Is this opinion or a news report? And understanding the mechanics of why you see what you see on something like Instagram — that mom might see completely different content. That education has to happen, ideally well before kids get unrestricted access to these platforms. I don’t know exactly how we do that at scale, but that’s what I’m thinking about.
[27:12] Frazer Rice: It dovetails nicely with the idea of financial literacy — the sooner you get ideas like compound interest into people’s hardwiring, the better able they are to spot deception and distortion out there. I really like the idea of a journalism class — something that lets you distinguish between the reporting of news and what goes into creating a story, the impact of editorial versus the writer, and what it means to be a responsible story versus purely opinion versus shouting into the wind via tweet. I might think about developing that myself.
[27:58] David Pakman: The other thing, which is more from my first book: I think about this as a pyramid. At the top is algorithmic content you’re scrolling — below that is good journalism you’re proactively vetting; below that is the media literacy we’re discussing; and even below that, like a Maslow’s hierarchy, is basic critical thinking. There’s a Pew Research Center study where they showed people 10 statements and asked them to categorize each as a matter of fact or a matter of opinion — not whether you agree, but whether, say, “chocolate is the best ice cream” belongs to the universe of opinion or fact. The numbers were a little better for Democrats than Republicans, but pretty poor for everybody. Americans, at a basic level, aren’t that good at distinguishing fact from opinion — and how are we going to have nuanced, productive conversations about complex issues if we’re still fighting that battle?
[29:10] Frazer Rice: The other thing that scares me is that as AI proliferates more, people outsource or delegate critical thinking — just punching in “is this right or wrong” and blindly trusting a different algorithm to tell them. That leaves you more and more exposed. All right, we’re coming to the end of our half hour here. How do we find your book, how do we purchase it, and how else do people stay in touch with you and your show?
[29:42] David Pakman: Yep, the book is at davidpakman.com/attention. If people want signed copies, there’s a link for that. You can get it on Amazon, Barnes & Noble, Kindle — there’s also an audiobook, which I recorded myself. Unlike, say, Maggie Haberman or Jonathan Swan, who had a million-dollar publicity budget behind their books, mostly it’s me going on shows like this to promote it, and that will determine whether it succeeds. So I really appreciate people ordering it — by the time this goes live, it should actually be out. That’s also my website where people can see my other content.
[30:19] Frazer Rice: David, thanks for being on. I’ve read the book and really enjoyed it — I think you’ve got a lot of pearls of wisdom in there that are not only educational, but that people can act on in their own lives to better themselves and their families going forward. Thank you so much.
[30:34] David Pakman: I really appreciate it.