Joe Weisenthal is not an AI doomer

Joe Weisenthal:
It’s been a pretty rough several, couple of decades, for media. But I’m actually kind of optimistic that there’s just going to be this sustained, if not growing number of people who want to hear from people.

Liz Hoffman:
Welcome back to Compound Interest from Semafor Business, where we explore the ways that business and finance are changing. I’m Liz Hoffman, Semafor’s business and finance editor, and joined this week by a special guest, Semafor’s editor-in-chief, Ben Smith, my boss, because the topic of conversation is the business of business media. Ben, what would you say is your business media diet?

Ben Smith:
I was thinking about this. I think it’s about a third Semafor Business and your great newsletter, a third pseudonymous Twitter accounts, and one third squawk box.

Liz Hoffman:
That is a heady, heady brew. I ask because our guest this week is Joe Weisenthal, one of the co-hosts of Odd Lots, which is Bloomberg’s great finance podcast, and probably a little bit of the DNA of this show. We can get into it with Joe, but we picked his brain a bit as we were putting Compound Interest together. But you know Joe from the early days of the internet, right? What have you made of his career, over the years?

Ben Smith:
Yeah, Joe has always been right at the frontier of business media. In 2004, when I was a weird political blogger, he was a weird business blogger who then got picked up by what would become Business Insider. And then gradually, as the energy shifted from a website to Twitter, became this massive figure on Twitter. And then jumped to Bloomberg, where he and Tracy Alloway developed a podcast that for many years was sort of a curiosity, and now I think is this just massive pillar of the way people understand business in the United States.

Liz Hoffman:
I’m a huge fan of it, we both listen to it. It’s an interesting fit inside Bloomberg, which is a pretty straitlaced place. Famously, a very walled garden, sells these terminals that sort of hook its users into its proprietary trading data. And I mean, it has decades of basically market data that you can’t get anywhere else, for which they charge a gajillion dollars a year. And then you have Joe and Tracy doing these really quirky, interesting deep dives on undersea cables, and the black market for dinosaur bones. And it’s always been a really interesting kind of front face of that super corporate business that is worth a ton of money for totally unrelated reasons. So excited to get into all of that with Joe Weisenthal.
All right, Joe Weisenthal, welcome to the show.

Joe Weisenthal:
Thank you for having me.

Liz Hoffman:
So you are, I think, slightly to credit or blame for this entire endeavor. Which is, we met for lunch at the Washington Square Diner I think, just before Christmas, when this show was kind of inchoate. And we asked you for some advice, and the one that I remember was that you liked the name, Compound Interest. We were on the fence about a few.

Joe Weisenthal:
It’s a great name.

Liz Hoffman:
And you said, “Hey, it doesn’t really matter, but also I like that one.”

Joe Weisenthal:
I did like it. If you guys, if Compound Interest when it supersedes Odd Lots, I’m going to feel really dumb. And then everyone will like, “Look, Joe brought it on himself, by giving them advice at the Washington Square Diner that day.”

Liz Hoffman:
But you actually said, “Odd lots is kind of a stupid name, but the show works, so the name works.”

Joe Weisenthal:
I don’t think name is particularly important. People obsess a lot about that. And I would say also, arguably from an Odd Lots perspective, the intro music that we have, our little Hawaiian guitar. We put about five seconds of thought into that. Arguably, that’s been more important. People always come up to us like, “You can never get rid of the Hawaiian guitar.” A few people hate it, but arguably, I would say the intro music’s been more important than the title itself.

Ben Smith:
I was under the impression for many years, like seven years, that you were the one playing that music.

Joe Weisenthal:
Yeah.

Ben Smith:
And only just learned that it was not you.

Joe Weisenthal:
I know. I wish. I could not play that. I’m not a good guitar player. I could play chords, a few chords can turn into any song, but no. That is just something that we found on a stock music site in 2015 when we started the podcast.

Liz Hoffman:
Well, it’s working.

Joe Weisenthal:
Thank you. Let’s start with Odd Lots. It has this very specific editorial instinct. You go deep on things that are quirky, and weird, and generally I think resists the pull of the news. How did you and Tracy land on that, and when do you sort of succumb to, “There is one big story in the world, and we’ve got to follow it”?
I would say, look, we’re way more news-driven than we used to be. But what I would say is, when we started conceiving of the podcast, we both had other jobs here at Bloomberg. I was doing TV at the time, Tracy also eventually went on to do TV. We were both doing various things with editing, for the web. So it sort of felt like the news itself, as in the news cycle, was pretty well covered on various platforms at the time. And so when we started the podcast, it was like, “Well, let’s just talk about things that might be interesting that would never sort of fit into a, quote, “news cycle,” unquote. And so if you go back to the very early episodes, you see a lot of stuff about, multiple episodes about the Florida real estate bubble of the 1920s, for example. I think we did three episodes about that, which is really interesting. And it’s relevant currently, or it’s always relevant, because people want to understand market psychology and the rhythms and the cycles that always take place. But it was not newsworthy in the classical sense.
Over time however, I do think that even prior to the pandemic, the episodes were starting to become a little bit more, I would say news cycle-ish. And then, when the pandemic happened, this was the thing that I always stress. Which is, for journalists of all sorts, when there’s something novel that happens, nobody is an expert, no one had lived through this period before. Tracy and I felt this intuitively, having started our careers during the Great Financial Crisis, “This is the great levelizer. This is when everyone is sort of starting at zero, trying to understand this thing that is emerging around them.” And so during the pandemic, that’s when we started really covering supply chains and semiconductors, and shipping and trucking, and all these things that sort of became partly our signature.
And so, it was news related, it was the cycle. But it was also like, “Let’s take time, and sort of understand things about the economy and the world we had sort of taken for granted,” and maybe that is what we’ve developed into over the last several years.

Ben Smith:
I spent a fair amount of the pandemic listening to you guys-

Joe Weisenthal:
Yeah, thank you.

Ben Smith:
... but followed you in the early internet, as a blogger-

Joe Weisenthal:
Yeah, that’s right.

Ben Smith:
... and an early Twitter person with a weird Twitter handle, before that got standardized. And I wonder, as you jumped from blogs to website to Twitter to podcast, if you sort of have reflections on the ways in which these technologies really have changed business media.

Joe Weisenthal:
Yeah. No, I mean, they’ve changed business media in so many ways, and I think about this all the time.
It’s funny though, you mentioned the following each other from the old Twitter, or whatever. Because when we were in London recently, we did a live show there, and there were some people in the audience that I really knew from that era. And 2012, to me at this point, does not feel like a long time ago. It feels like a different age. It just feels like, at some point, there was this discontinuous break. And it’s like, “Oh, nice to see you from the old country,” or whatever. It does not just feel like that was distance. It just feels very different, the changes in business media are so many.
But here’s a couple that I always come back to. One is, I think, social media in particular really expanded the pool of people that might be considered an expert in something. You discover all of these voices from out in the world, as some guy who really knows X, they really know this domain very well and they’re posting about it. And that was the core thing that sort of made social media really fun, say 15 years ago, which is there may have been this time where business media in particular had a certain roster of guests and economists, and sell side researchers and so forth, that were the names.
And suddenly they were practitioners that were just saying, “This is how I do my job.” And that was really fascinating, 15 years ago, and then they go direct and then it was like, “Oh, you know what? You’re much more interesting on the domain of, ‘Here is someone who is working on an oil rig,’ or, ‘Here is someone who’s a trucker,’” or whatever it is. And they’re talking about what they do, and we don’t need the sort of standard roster of expert.” That was what was sort of special, and that really changed business media forever, our ability to find some of these voices.
I think it’s interesting, if you project forward to 2026, you think about someone who maybe 10 years ago they were a farmer, and they were talking about farming conditions and they were cool. And I feel like in 2026, if that person is on social media, A, there’s a good chance that their brain has become fried by politics and they’re posting about politics, rather than corn. And B, if they do want to post about corn, there’s a good chance they’re selling it. By which I mean, they’re selling a newsletter or something like that, which is great. And I don’t begrudge anyone for being able to monetize their domain expertise, but I think this is sort of an, under the rise of the creator, the professional, who can sort of monetize their understanding of a distinct thing. I think that’s really changed the flavor of social media quite a bit. And so you kind of lose some of that public good element. And maybe it’s rational for the individual level, but makes it less fun at the aggregate level. A lot of people, if they know something, are selling it.

Ben Smith:
Yeah, totally. And I think kind of the golden era, that we both are nostalgic for, was really about a kind of amateurism.

Joe Weisenthal:
That really is it. And it’s sort of amateurism on multiple levels of like, “Let’s just talk about this. I am interested in this.” And there’s no expectation that this back and forth is about selling something. These are sort of un-monetized transactions, you could say arguably that we started it. By we, I mean to say like, us, the three of us. In that, okay, you identify that there is, as they say in finance, some alpha here. There are people here who are talking about something that is very interesting. It’s not being captured anywhere else. And so we’re like harvesting that alpha. We have articles and podcasts and platforms where we’re finding this, and the other person didn’t, and eventually that sort of disappears. But maybe we were the first to recognize that there were some opportunities in just sort of listening, and putting our ear to the tracks.

Liz Hoffman:
So Bloomberg is a maybe breakeven, maybe slightly profitable, media business attached to a giant trading proprietary data business that prints money. Where does Odd Lots fit into that?

Joe Weisenthal:
I can only speculate, and it’s probably not even, I don’t even think they’d be thrilled if I speculate. It’s like, “What is the goal of podcasts, period?” But I do think that we sit very comfortably in this organization because A, we really do speak to the professional, hardcore audience. If you are in finance, if you’re at a fund, if you’re at a bank, et cetera, we really do want every episode to be interesting to you, and not for it to be sort of beneath you, novice level. And then we want everyone else in the world, who’s not in the world of finance, to recognize that actually there’s a pretty good chance that they could understand it. If they’re sort of interested and engaged, willing to give us a little bit of time, that a lot of these topics we talk about, monetary policy, whatever it is, there’s a good chance that they can understand it.
And what I would say is, having done this now over 10 years, the number of emails that we get, and it makes me feel old every time, but I do really like it. I mean, we have people who say, “I started listening to Odd Lots in high school-”

Ben Smith:
Oh my gosh.

Joe Weisenthal:
... “and I just got my first finance internship in New York City,” or my first job, or whatever. And it still feels weird to hear, but I genuinely do think that Odd Lots specifically, a few other things, are for a lot of young people, their first maybe entrée into the world of Bloomberg Media, and therefore Bloomberg the company.

Ben Smith:
In that context, the one kind of trend that you’ve resisted as you’ve morphed through digital media as we all have, is this twist in a way toward this creature of the creator economy. And we’re not trying to just give aneurysms to your bosses at Bloomberg here-

Joe Weisenthal:
Yeah, I know.

Ben Smith:
... but a lot of people like you would quit, and go independent. How do you think about that?

Joe Weisenthal:
The thing I think about it is, I really like my job. I mean, it was funny. So obviously we were very fortunate last year, right around our 10 anniversary last August, we got this big profile in the New York Times. That was fantastic, very gratifying obviously, to get that kind of attention. The article was very nice, and didn’t have anything negative to say. But people would come up like, “Okay, what next? What do you want to do?” And my honest answer unambiguously is, “I want to do more episodes of the podcast.” And of course there are other things, okay, you could do more shows, or do X. I really like the job of co-hosting Odd Lots, and I think it is actually kind of that simple, in my mind. This is a great place to do it, and I really like it.
And so I don’t spend... I know, I mean, I see people having success, but I like my job. And I really don’t spend a lot of time thinking about that question. I mostly spend my time thinking about who the next guest should be, and that’s kind of it.

Liz Hoffman:
Do you ever run into story tips? You talk to a lot of people-

Joe Weisenthal:
Oh, yeah.

Liz Hoffman:
... but you’re not, I don’t think, a breaking news reporter, so where do you send them?

Joe Weisenthal:
You know, it’s funny, Bloomberg is such a big place. I’m still always turning to Tracy from time to time, for example, and saying like, “Wait, who is it that covers X or Y for us here?” And part of the reason it’s important to know that is because if we’re going to have a guest from a certain domain, it’s good practice just on multiple levels, to talk to the reporter who covers them. More than a handful of times someone will send me a DM or something like that, and I’ll go find the person who would cover it here and I’ll say, “Look, I don’t know who this person is. They have a weird cartoon avatar. I’ve never gotten this person’s name in my life, but I’ve interacted with this person. And I kind of have reason to think that maybe they know what they’re talking about when they say that so-and-so just quit X bank or whatever.” And so I say, “Maybe try to hunt this thing down.” That has definitely happened multiple times.

Liz Hoffman:
What is your biggest uncredited scoop?

Joe Weisenthal:
There was a big departure several years ago from JP Morgan, that I think, it was one of the people who I think at one point was thought to be in line for the CEO, and he left.

Liz Hoffman:
All right.

Joe Weisenthal:
So yeah, I’ll say, I’ll leave at that. Yeah.

Liz Hoffman:
All right. I’m going to go do some sleuthing. We were talking a minute ago that you are, I think, still hopelessly addicted to Twitter.

Joe Weisenthal:
Yeah, yeah.

Liz Hoffman:
A, what value do you think still remains there? And is there something about peak finance Twitter that you think, nothing that’s come since it has totally captured or replaced?

Joe Weisenthal:
I mean, I still 100% think that a lot of other people are addicted to it too, including a lot of very influential people-

Liz Hoffman:
Ben. Ben is addicted to it.

Joe Weisenthal:
... in media, yeah. Both influential, both in media and tech in particular, and also there’s still quite a few people in finance as well. And I just think it’s, from what people are going to be talking about, or what’s influential. I think that for example, a really good recent example, is essentially in the last six months. If you were not on Twitter in, I would say December of 2025, I don’t think you would have anticipated or seen the rise of Claude code specifically, and the rise of vibe coding, and therefore how big Anthropic was about to get in 2026, sort of achieved coming closer to parity with OpenAI. That was a really crisp example, to me, of this happened on Twitter. People started talking about Claude code, and therefore these new capabilities on Twitter or X first, and then it ended up having some pretty big downstream implications. And now...
So I think, that kind of thing, it just happens there first. At least at the intersection of business and tech, which is kind of everything these days, because I just feel it. AI is like the sort of blob that seems to be swallowing every other topic there is, and I feel that. But I do feel like at that sort of intersection of business and tech, it’s Twitter. And a good example is OpenAI’s purchase of TBPN, Andreessen Horowitz’s launching of their own, or independent streaming network. A lot of really important people in the world of business seem to put a lot of value on Twitter, specifically, even if most people are on other platforms. That’s at least how I justify, to myself, my addiction.

Ben Smith:
How do you think about the relationship between, I mean another big Twitter obsession, prediction markets. And the way they’re shaping financial journalism, in specific journalism, and the media in general.

Joe Weisenthal:
I’m excited. I don’t know if I should say fan, my feelings are complicated, but I think it’s cool that we can put a number on conventional wisdom. So I’ve never liked the term prediction markets, but I do think that they do a pretty good job of aggregating conventional wisdom. So you think like, “Okay, here’s some election that’s coming up in Canada,” or whatever. And I think if you want to know, “Okay, who do people expect to win?” It’s a great place to start. Start the question of, “All right, let’s just look at what the odds are.” And I think they probably do a pretty good job, not because empirically they have great track records, although I think they’re not that bad. But more so that, in my mind, a market is more or less efficient if it’s difficult to beat. That’s sort of the test that I think of.
A lot of people, they’ll say, “Oh, it’s all, these numbers are biased, because it’s a certain subset of crypto bros who are using them and they have their blinkers.” If that were true, there would be easy opportunities to make a lot of money by taking the other side of them and their bias. I don’t think there is.
I’ll say two more quick things about prediction markets, or maybe three. One, they’ve existed for a long time. I would say the Fed Fund’s futures market. That’s a prediction market. It is a bet on what 12 members of the FOMC are going to decide eight times a year. It’s literally what a prediction market is. It’s not metaphorically a prediction market, it’s a prediction market.
Two, they’re mostly sports betting sites, and I think this is really important, which is they aspire to be important sites of hedging activity for real economic activity. By and large, they’re not, and we should be mindful of them. It’s like sports, and there’s some politics, and stuff.
And then finally, I worry about, “Okay, all of this information,” certain types of information are now monetize-able in a way they weren’t. A good example of that might be, say, pre-knowledge of the raid of Nicolás Maduro, or something like that. I’m very worried about society, if every time two people share something, or the number of things that you can comfortably share with someone, shrinks because suddenly that is monetize-able information. I think this is going to have probably increased unpleasant effects on society, but I don’t think it’s great that the Nobel Prize Committee suddenly has to shrink the circle of trust about who can know the name of the winner, because they’re worried that someone is going to be able to trade on that information. That just doesn’t seem great. This is not inherently valuable information until there’s a betting market on it.
But then I guess, on the other hand, these markets aren’t going to take off so long as people perceive them to be riven with insider trading. So this is just a classic phenomenon, and a good reason to argue that insider trading laws should exist, and make markets better. People aren’t going to participate in any game if you assume that other people are cheating. And so if they sort of stamp out the concerns about insider trading, that would improve the odds, in my view, that these become important economic instruments.

Liz Hoffman:
I totally agree, in that it’s not always the case, but it is in both the public’s interest and these companies’ financial interests to solve this problem. Because no one’s going to use a market they don’t trust.

Joe Weisenthal:
Yeah, I agree. Totally.

Liz Hoffman:
We’re going to take a short break, and we’ll be right back with more from Joe.
On the sort of financial influencers, finfluencers, I think is what we’re contractually obligated to call them. I saw Lloyd Blankfein on Mrs. Dow Jones’ podcast just the other day.

Joe Weisenthal:
Oh, incredible.

Liz Hoffman:
Incredible. And I realize he’s selling a book, but the idea that Bob Rubin or Hank Paulson would be on Mrs. Dow Jones’ podcast, and I think The Wall Street Journal has a story up today on what Ryan Cohen told Anthony Pompliano. I mean, how should the legacy media, and I would put sort of ourselves in that, both of us in that bucket, be thinking about chasing the long tail of information that is sometimes genuinely very interesting on these platforms, that are not moderated by people who care that much about journalism.

Joe Weisenthal:
There’s so much to unpack there. It’s a good question. I mean, first of all, Lloyd is Lloyd. And part of the reason that it’s hard to imagine Hank Paulson or others, is because Lloyd is a category of one. I love talking to Lloyd, we had him on our show earlier this year. I don’t know, something just incredibly refreshing about talking to Lloyd, like you’re talking to a normal human being. There’s a couple of things, which is a lot of financial influencers, some of them are good, some of them are okay.
There’s a fair amount of this sort of broader fever swamp, that Ben has probably been very familiar with, in particular in politics, is very real in finance as well. It’s like conspiracy theories about Vanguard, and BlackRock, and buying up all the homes. This Twitter influencer macro is like its own new thing, of how a certain type of people online, they’re very obsessed with the BRICS currencies and stuff like that. A lot of them really like talking about, “The Chinese yuan is about to supplant the US dollar,” and so forth. And a lot of them are very influential, and I can feel it when I talk to a lot of people who are outside my sort of media sphere, how much finance influencers sort of shape the way that they perceive the world, and what’s going on in markets. A couple of years ago we were in North Carolina for a little tour that the President of the Richmond Fed did, and someone had asked him about Central Bank Digital Currencies at this event. And that’s like a classic, coming out of these fears of, they’re going to be able to track all your spending and stuff like that.
But anyway, I do think it creates a lot of, there’s a lot of conspiracies and misinformation. And I do not devote a lot of time to thinking about it. A very important thing is stamping out misinformation, but I think there are good opportunities to like, “Let’s actually try to have what we perceive to be an honest conversation about some of this stuff.” And if some people want to hear from a more credible outlet on the question of say, institutional purchases of single residential homes, that’s less conspiracy-laden or whatever, then maybe legacy media could be a good venue for that.

Liz Hoffman:
But all this new stuff, whether it’s influencers that don’t ask real questions, or like frankly Twitter accounts themselves, companies have a lot more ways. I remember, I was a beat reporter for Goldman at the Journal, and Lloyd Blankfein got on Twitter, and we wrote a story. I remember the headline was like, “Welcome to Twitter, Lloyd, you’re going to love it.” And I pretty quickly got my comeuppance, which is that we’d written a story that he was preparing to step down, and for which the company did not comment, was very tightlipped. And then, about an hour after the story, he tweets, I feel like I’m Huck Finn at my own funeral. And we were like, “Should we put it in this story?” I don’t know, it was a-

Joe Weisenthal:
Yeah, I know, it’s very...

Liz Hoffman:
... and I feel like that’s still an evolving conversation, about how much-

Joe Weisenthal:
It is.

Liz Hoffman:
... the media should follow those direct channels that big companies and powerful entities that we cover, have.

Joe Weisenthal:
I know and look, and I mentioned for example, the Andreessen Horowitz streaming service, and you’d see it obviously more in tech. But you also see it in Wall Street because all those places, you know, Goldman has its own podcast, and Morgan Stanley has its own podcast. They’re actually pretty good-

Liz Hoffman:
They are pretty good.

Joe Weisenthal:
... in many cases. So everyone sort of wants to go direct, and everyone would like to cut the three of us here out, I suspect. But also not, because I think that there are certain people who are a little bit either in denial or delusional about what, say, the media, the traditional media can offer. It’s like, “No, we cannot offer more eyeballs per se, or ears.” But I think everyone loves the prestige, and everyone loves how it feels, I guess maybe validation or whatever. I myself would say that I’m surprised, in the year 2026, to this day the validating ability of certain legacy media brands.

Ben Smith:
So you really liked that New York Times profile?

Joe Weisenthal:
Yeah, of course. But also I just think that all of the tech people who love to grave dance, incorrectly grave dance on The New York Times, which is doing fantastically by the way. They all will say yes to the New York Times in a second if a reporter wants to do a story on them, and their publishers would say, “Yes, absolutely, do it in a second,” and they wouldn’t give it a second thought. Yeah, I think many of them are a bit delusional, when they deny the influence that at least a handful of brands still have.

Liz Hoffman:
Ben, you have some theory about this, don’t you? About how the prestige of traditional media is actually what’s feeding the LLMs?

Ben Smith:
Oh, this is a weird little sidelight, but yeah. If you Google, “What is my reputation? What is Bloomberg’s reputation?” You search that in an LLM, or search questions like that. The sources that will feed it are not going to be social media and Reddit, they’re going to be articles often in trade journals that are almost out of business and have no business model, but suddenly desperate PR people are now trying to get their executives back in these dying or thriving publications. Because, for at least right now, these LLMs are very sensitive to traditional authority. I mean, I’m sure there’s many efforts to game them, but it’s-

Liz Hoffman:
By the way, you know why it’s all the trade pubs?

Joe Weisenthal:
Why?

Liz Hoffman:
Because they’re all owned by private equity.

Joe Weisenthal:
Oh, interesting.

Liz Hoffman:
They were the first ones to take the money from the LLMs and say, “Of course.”

Joe Weisenthal:
Oh, yeah. I love that. No, look. I mean, this is such an interesting phenomenon, that people have talked about, literally more in just the last few months. Which is the sort of the, “What is the ideology of Claude and ChatGPT?” As sort of boring, normie liberal. And that’s very interesting. A few people are starting to talk about this. John Burns Murdoch at the FT, I think was one of the first, to sort of flag that this sort of... The Overton Window, maybe if you want to call it, within an LLM is narrower than it was on social media.

Ben Smith:
I remember, I think I once edited a story at BuzzFeed that was called like Twitter is a Truth Machine. And I remember when we had these conversations about social media, about how great it was at getting to the bottom of things, and getting rid of bullshit, and crowdsourcing the truth. And I don’t know, I think the LLMs are going to get really, really good at feeding you into your own bubble of delusion and insanity, the way social media was.

Joe Weisenthal:
Yeah. And I do think the sycophancy is a really big problem. I would actually say a much worse problem, in my view, than the so-called hallucinations. Which are not as big of a deal as many people think. And look, AI psychosis where people think like, “Oh, I’m very close to cracking Fermat’s Last Theorem, or resolving the tension between Newtonian physics and whatever physics, or quantum physics. I think there is a lot of, that sort of psychosis, and it will absolutely create new pathologies and so forth.
I guess my issue, what I would say is, a lot of the criticisms that people have towards the emerging AI world, which feels like we’re entering into. Really just feel like descriptions of the past. We’ve already had this intense war on everyone’s attention span. That’s the story of the last 15 and 20 years. That happened. Political polarization, and the rise of weird extremist ideology. That is the story of the last 15, 20 years.
And then the other big one, of course, is the concerns about wealth inequality. Which maybe AI will exacerbate wealth inequality, but that’s just basically the story of the rise of tech over the last 20 years. I am very open to the idea that AI, or LLMs specifically, could surprise us in ways that are not just, “Let’s just draw a straight line beyond the trends that we’ve seen from social media.”

Liz Hoffman:
Wait, can I poke at that really quick?

Joe Weisenthal:
Sure, yeah.

Liz Hoffman:
Which is, the wealth inequality, because something I’ve been thinking a lot about. I think when Google went public, it was worth $23 billion.

Joe Weisenthal:
I know.

Liz Hoffman:
It’s worth $4 trillion today. All of that accrued in public hands. And you could say more people should own stocks than do, but Anthropic is five years old, and is as valuable as Amazon was when it was 25 and had been public for 15 years. So what do you make of that?

Joe Weisenthal:
No, you’re right. And it’s a really good question. It is extraordinary concentrations, and that is a way in which it does seem like the inequality of the last 15 years does seem to be turbocharged. I don’t know. There are weird things. I mean, it’s important to remember that we don’t know when or if any of these companies are going to get remotely profitable. Right?

Liz Hoffman:
Mm-hmm.

Joe Weisenthal:
So yes, it’s true that equity values have gone a lot, but in terms of sort of accounting profits, that is very different than Google, et cetera, where these companies are intensely profitable. It could be that just one wins out, and then that’d be sort of scary for its own reasons. But then you have a bunch of people who are billionaires today, and won’t be billionaires tomorrow, et cetera. There’s a lot of volatility here in terms of how this plays out. And so, my impulse is to just sort of try to not be too rigid in my expectations. That is what I’m just trying to instill in myself as a sort of operative principle here.

Liz Hoffman:
It’s also a pretty good argument against the wealth tax. That’s the one they’re all using.

Joe Weisenthal:
Yeah, but you know, so this is interesting. I was in London last week, and of course there it’s also a big deal, because a lot of people are fleeing. I do think we’re going to have to figure some things out, potentially, as a society. We haven’t figured out how to do a wealth tax yet, it seems sort of difficult to implement both politically but also mechanistically, if people can flee, et cetera. Obviously, we know about some of the ongoing debates happening in California. But given the degree to which wealth rather than income is the channel via which a lot of very powerful people live their lives and from which they consume, we’re going to sort of have to figure this out. And I don’t think we are. We don’t seem particularly close, either.

Ben Smith:
A narrower question on AI, which is like, what it means to us.

Joe Weisenthal:
Sure, yeah.

Ben Smith:
We’ve already got AI basically writing up the snap stories off the earnings reports. You see them all over Google News.

Joe Weisenthal:
Yeah.

Ben Smith:
Maybe on the other end of that journalistic spectrum is Odd Lots. What in financial media, and in business media, do you think survives AI? What doesn’t?

Joe Weisenthal:
Yeah. I mean, look, I do think that, it does seem like some people like people. Just before I came on here, I saw a headline that I did think was interesting. It was that Charles Schwab is going to use AI to expand its services to less affluent clientele, which I observe, could be read in two ways. One is, here are some premium services that were only economical for Charles Schwab to offer to rich people before, and now they can offer them to more people via AI. I don’t know what they are, but let’s say in theory, that’s great.
But then the other flip side is that, “Okay, for the masses, get to talk to an LLM. And the very elite get to talk to a human.” And I do think that there is just going to be, communicating or listening to a human is probably going to be somewhat of a luxury thing, and become valuable. I’m actually optimistic about this. I’m not just saying this. For years it’s like there’s, “Oh, journalism is really tough. And only go into it if you have to, because it’s so awful,” et cetera. And I don’t know, it’s been a pretty rough several, couple of decades for media, but I’m actually kind of optimistic that there’s just going to be this sustained, if not growing number of people, who want to hear from people.
And then I think the other thing too, and this is something that I haven’t seen much talk about, but remember like 15 years ago? Every newsroom got a data team, and suddenly every publication had a team that was doing great data visualizations, and charts and stuff like that. And that was the hot thing for a while, within journalism. I think LLMs actually are going to create a bunch of opportunities like that.
I haven’t really seen it yet but I think that, like if you’re a journalist listening to this, you might be shocked at how easily you, with no expertise, could build a machine learning model that could distinguish between X or Y. And it’s like, “Okay, which politician is the most pro crypto?” Or, “Did a politician’s speeches become more pro crypto after X or Y?” And you probably could build a machine learning model, without any formal technical background, that actually answers that question.
I’m very optimistic about new sort of things that journalists can do, that we didn’t have the technical capacity to do before. Sure, you want to get a sanity check on all this stuff. Setting aside summaries of news by the chatbots, which is, whatever. I think there’s just exciting things that we can now do technically, with AI, that we couldn’t have done a couple of years ago.

Liz Hoffman:
But is the free version of Odd Lots in the future your agent interviewing the guest agent, and then AI putting the podcast together, and then the one that I pay for is actually you and them and Tracy in a room?

Joe Weisenthal:
I doubt it. I mean, I doubt it. Here’s the thing is, if you’re going to allocate time, I think you probably generally want to talk to a person. And I think there’ll be value. I could be totally wrong. I mean, for one thing, currently AI can’t do an interview because it can’t interrupt, right now. I mean, there’s actually, this was the thing I saw this week, which is that, Mira Murati’s new startup.

Liz Hoffman:
We just saw it. It can interrupt you. It’s great. Yeah.

Joe Weisenthal:
Yeah, so that actually is kind of potentially a big deal, but these are really important things. There’s a pretty big gap, still, between LLMs, and even the voice ones, and anything that we would call human conversation. Precisely because humans just talk with all these tics and half sentences, that AI never does. So I do think we’re actually still kind of a ways away from that, but I think if you wanted to allocate your time to X or Y. I mean, if you want a summary of something that already exists, if you just want to read the Cliff’s Note, that already exists.
It’s actually funny, frequently when I tweet one of our episodes out, there’s some company, they’re kind of annoying because right underneath my tweet they’re like, “Here’s a summary of today’s Odd Lots,” and they’ll reply to it. And it’s AI generated. It’s really good. Like, I hate to say it, but it’s really good. It’s really thorough, and they generate it in about 30 seconds, afterwards. But I think that these things, they exist, and I can’t perceive them as having meaningfully changed the trajectory of listenership, et cetera. I think all these things are out there anyway, and if people wanted them, they could have taken them. So I do remain, unironically and not just in that sort of journalism school professor way, optimistic about human reported media.

Ben Smith:
I think we’ve just got a couple of minutes left, so I’m going to ask you a complicated question and ask you to compress the answer.

Joe Weisenthal:
Okay. I can do that.

Ben Smith:
You mentioned just the power of AI tools for us to play with, and I know you’ve been playing with these tools around your obsession with orality, and with the difference between spoken and written language, spoken and written culture. I’m curious how you see that kind of shift toward a more oral culture, kind of flowing through the media, really.

Joe Weisenthal:
So I mean, look, I think the fact that every news organization in the world... Not podcasts, the fact that they all put their reporters with their little tiny mic, and make them talk into the vertical video. That is like the sort of ultimate peak of this phenomenon, that everyone is just talking all the time. That talk, and conversation, and orality, et cetera, is truly just the sort of dominant theme of our time.
These organizations such as the New York Times, that sort of built themselves on perusing, reading a physical thing, they still do a lot of great real reporting. I’m a true at New York Times admirer, and defender. But a lot of it is still the person with the tiny mic. And I just think this is like such a thing. Everything’s constantly conversation, everything’s back and forth-ness, everything is trying to one up each other and impress each other, et cetera. Everything is fluid, and if you didn’t see it didn’t exist. And people are like, “Oh, why didn’t the mainstream media cover this?” And it turned out the only reason they say that is because they just didn’t go bother to look at literally any website, which it would’ve been right there on the thing.
These are all the markers of a sort of oral dominated society, where everything is evanescent, and it talks and then it’s gone a second later and you move on. And the thing that we’re all talking about, it’s the character of the day or whatever, this is the landscape we’re in. And podcasts are almost like a small part of it. I almost think they’re almost harken back more to the literate tradition, because at least in the literate tradition, it was carving out a time of your day. Where you’re like, “I’m going to allocate time to this, with more or less no distractions. And I’m going to, in this case, listen to people talk instead of read.” Maybe podcasts are sort of the holdout, in some elements, of what we might call patterns of literacy.

Ben Smith:
Joe, you’re so optimistic about everything. Are you optimistic about the end of literacy?

Joe Weisenthal:
No, I’m like a doomer. I’m like, I scroll at night and I look at my phone. And I’m like, “Oh God, it’s so...” But I don’t know, on media things specifically, look, I like doing what I do. And I hope to do it for a long time.

Liz Hoffman:
Well, I think we are out of time, and is a fun place to leave it-

Joe Weisenthal:
All right.

Liz Hoffman:
... I think, because as they say, we did not interview you. We were in conversation with.

Joe Weisenthal:
This was great. I loved being in conversation. That was fantastic. I had a blast.

Ben Smith:
Thank you, Joe.

Joe Weisenthal:
Any time.

Ben Smith:
That was fun. And thanks for helping us get Compound Interest off the ground.

Liz Hoffman:
Thanks, Joe.
All right, Ben, what’d you think?

Ben Smith:
I mean, I am struck by Joe’s optimism. I basically share it. It’s hard to have been in media this long, and kind of ridden these waves, without having a sense of excitement about what’s next. He’s just sort of let his curiosity about the world just kind of pull him forward, in a way that I find very impressive and kind of inspiring, honestly.

Liz Hoffman:
Yeah. I know a lot about orality now, because I ordered a book that he wouldn’t stop talking about on Twitter. And it’s actually like, it’s interesting, and I think a real challenge for people in our business. I mean, I didn’t even start my career... The early, middle innings of my career, I was writing earnings stories. I was waking up at 4:30 in the morning, and I was going into the Wall Street Journal’s offices in Midtown when it was still dark, and I was reading press releases and writing stories that then were printed on paper. And the whole thing just seems, like you said, just like a different age. And I guess the fact that we’re talking to him on our podcast shows that there is a real conversation happening, and obviously we do it, it happens a lot on our stages, but there’s some coming ambient about it.

Ben Smith:
Do you have nostalgia for the sort of more buttoned down, print-y, deadline oriented, the switch is on, the switch is off, media that you came up in? The financial media in particular? Or do you think we’re doing a better job now?

Liz Hoffman:
Not at all nostalgic for it. People ask me a lot about what’s different at Semafor, and I say a lot of things, but actually just the psychological unchaining from that 4:00 PM page one deadline, is just actually massive. And I’m sure people, legacy media organizations with print arms will tell you that they don’t operate that way, I assure you that they psychologically do. Not nostalgic for it, but I do think it informed the reporter that I am, which is that when you start working at a newspaper, your story is immediately at the bottom of a birdcage, just like it didn’t happen.

Ben Smith:
Yeah.

Liz Hoffman:
I don’t know. It’s maybe informed the way I kind of think about coverage, which is, “Well, just write it again.”

Ben Smith:
Yeah. It’s funny, what he said about orality, because that’s actually always been true of the daily news product. It just exists for the moment, it’s not relevant or comprehensible five years later.

Liz Hoffman:
Yeah. And finance, I think, is in some ways really adapted to that in that market’s open all the time, moves every day, and there’s a story every day about what it did. But I mean, it’s incredibly, it’s a numeric kind of reporting, fundamentally. And you’re saying, “Okay, how does this number compare to that number?” And so it always is kind of like living in some time-lapse, and the medium and the media can change. But I think, “Did Google make more money this year than I did last year,” is still a question that people are going to care about. It’s not a question that I’m going to get up and read a press release and write 400 words on in the morning.

Ben Smith:
Well, actually, we have a meeting after this to discuss that.

Liz Hoffman:
There’s a price for everything, Ben.
Well, that’s it for us this week. Thank you for listening to another episode of Compound Interest. Our show is produced by Josh Billinson with special thanks to Anna Pizzino, Katherine Bilgore, Claire Einstein, Rachel Oppenheim, Tori Kuhr, Vilanna Wang, Garett Wiley, Stephanie Chang, and Daniel Hoeft. Our engineer is Bob Mallory, and our theme music is by Steve Bohn.
If you want more Semafor, you can sign up for our email newsletters in your inbox at semafor.com, and listen and subscribe to Compound Interest wherever you get your podcasts.