WEBVTT

1
00:00:00.100 --> 00:00:12.160
[on-hold music] Welcome to the Rebooting show. I am Brian Morrissey.

2
00:00:12.390 --> 00:00:26.760
This week, I was joined by BuzzFeed CEO, Jonah Peretti, for a wide-ranging discussion about the anti-SNARF manifesto that he just dropped, and also BuzzFeed's plans to build out its own social network, and also just how we got to this point.

3
00:00:27.280 --> 00:00:34.800
But before we get to that, I wanna tell you a little bit about an upcoming online forum that we are holding. This is our version of a webinar, but it's just way better.

4
00:00:35.060 --> 00:00:40.920
It's an hour-long interactive deep dive into a particular challenge that publishers are facing.

5
00:00:40.980 --> 00:00:55.320
And on February twenty-sixth at one PM, I will be joined by Valnet's Director of Marketing for Gaming, James Kosur, and Chris Hart, CEO of TruAnthem, a social media automation platform that is our partners for this online forum.

6
00:00:55.900 --> 00:01:06.020
The, the f- online forum is about how to win at social in twenty twenty-five, and basically the, the cheap traffic era is over. But social platforms are more dominant than ever. That's just the reality.

7
00:01:06.040 --> 00:01:15.980
And in the more l- with less era, publishers just need to be more strategic in how they deploy scarce resources to these platforms and the goals that they have for being on there.

8
00:01:16.360 --> 00:01:28.880
James is gonna go into a lot of detail in this session focused on how Valnet's brands, which are across entertainment, gaming, and tech, use Facebook, TikTok, X, and even Dark Social. So please do come to that.

9
00:01:28.920 --> 00:01:35.360
You can find out more information. I'll leave it in the, uh, show notes, and it's also in the Rebooting newsletter, which I hope you all get and read. Please do.

10
00:01:35.780 --> 00:01:52.190
So as I said, this week, I was joined by Jonah, and last week he issued a three-thousand-word manifesto that decried how broken social media is and BuzzFeed's plans to build its own social network that will be optimized for joy rather than the addictive wares that Jonah calls SNARF.

11
00:01:52.360 --> 00:01:59.600
It's an acronym, and we get into what it means and... But basically, think about how the algorithms have weaponized themselves to get our attention.

12
00:02:00.160 --> 00:02:09.440
So I wanted to talk to Jonah actually less about this hazy plan for BuzzFeed Island, BF Island, excuse me, 'cause it isn't even a product yet, and, and it doesn't even have a demo video.

13
00:02:09.539 --> 00:02:12.720
We talk a little bit about it, but it's in broad strokes, to be honest with you.

14
00:02:12.880 --> 00:02:23.100
But I wanted to talk more about this evolution of social media and how it has pretty much overwhelmed the information space and a lot of just regular people.

15
00:02:23.290 --> 00:02:33.920
There's a lot of books out right now, I noticed, about how we're losing our attention. And Jonah has built his career basically at the collision of media and social technology.

16
00:02:34.100 --> 00:02:47.859
All the way back in two thousand and one, he had his first experiment in virality by writing Sweatshop on a customizable pair of Nikes, or at least he tried to, because he got rejected by Nike, and then the back and forth, uh, with the company went, went viral.

17
00:02:48.060 --> 00:02:48.890
Remember when we used to say that?

18
00:02:49.500 --> 00:03:01.210
He went on to start BuzzFeed as a company after Huffington Post and would basically ride the rising tides of social media as Facebook was really just getting going and becoming dominant.

19
00:03:01.720 --> 00:03:15.180
And of course, tides go out, so BuzzFeed's had a rough go of it. We get into a little bit about the plans for turning the company around. But I think I, I see this as, you know, really pivoting hard into tech.

20
00:03:15.340 --> 00:03:24.940
And, you know, I asked Jonah about this, about, like, whether, you know, this is just sort of heavy operating the publishing business while trying to, to build a new business.

21
00:03:25.020 --> 00:03:35.120
Because publishing is, is a rough business and, you know, just BuzzFeed's stock, it has been down. It's been decimated. It's down, like, ninety-four percent since, since they went public.

22
00:03:35.200 --> 00:03:43.480
So, you know, the stock market doesn't necessarily love digital media companies at this moment. But I can remember meeting with Jonah during BuzzFeed's early days.

23
00:03:43.630 --> 00:03:52.100
And it's somewhat hard to remember now, but social media was really exciting back then. And, and many people, and even me at, at, at times, really saw it as the future.

24
00:03:52.280 --> 00:04:00.920
I mean, Barack Obama used it, and it was given a lot of credit for his, what we thought was a historic election in two thousand and eight that would change America forever.

25
00:04:01.460 --> 00:04:10.900
You know, as Facebook grew, it, it took on an expansive vis- vision and mission to connect the world. And Twitter was even credited with helping the Arab Spring get going.

26
00:04:11.340 --> 00:04:23.740
And, and BuzzFeed really became a pioneering company of, of that time. Okay? And publishers would copy BuzzFeed endlessly in its rapid-fire approach to mastering what it took to get people to share content.

27
00:04:23.780 --> 00:04:35.440
I mean, it was basically hacking social media in a very handcrafted way, combing through the data and, and basically capital- capitalizing on how people are, you know, frankly, suckers for content that says something about them.

28
00:04:36.020 --> 00:04:47.080
That's when they like to share it. So, you know, those were, as Jonah calls it, the beer and wine days, and, and those have been replaced by, in, in his view, uh, the fentanyl of AI-driven algorithms.

29
00:04:47.100 --> 00:04:54.500
And this really started when, when TikTok burst on the scene, and all the social platforms have basically followed TikTok's lead.

30
00:04:55.100 --> 00:05:06.600
And, you know, that along the way, these platforms also, by the way, ditched the publishers it once would, would... I mean, it would much rather have millions of creators than deal with a bunch of ornery publishers.

31
00:05:07.240 --> 00:05:15.440
And you know the rest of the story. So anyway, Jonah and I talk about how it all went wrong and whether now might just be the time for an alternative.

32
00:05:15.450 --> 00:05:24.260
And we also get into my favorite topic, which is whether we're gonna have websites in the future. This is a fun conversation. As always, I would love to know what you, what you think of this and the show.

33
00:05:24.380 --> 00:05:33.880
You can email me at bmorrissey@therebooting.com. And if you do like the show, please give it a rating or review. I, I always like to see those. So here's my conversation with Jonah.

34
00:05:34.050 --> 00:05:49.350
[on-hold music] Jonah, welcome to the podcast. I think this is our fourth or fifth. We, we do these every few years. Good to be here. So you just came out with, uh, with one of your manifestos, SNARF... Anti-SNARF.

35
00:05:49.440 --> 00:05:57.480
You're not pro-SNARF. Anti-SNARF manifesto. And I always enjoy your, your manifestos. But I wanted to, like, sort of just...

36
00:05:57.620 --> 00:06:04.310
Before we get into the manifesto, because y- I wanna get your perspective about how we got here, right? Like, I mean, and not go...

37
00:06:04.969 --> 00:06:16.440
'Cause I think you're scratching at an itch a lot of people have, and that is being articulated quite a bit. I'm, I'm going through the Chris Hayes bookYou know, about attention and the siren's call.

38
00:06:16.640 --> 00:06:26.080
It's, it's the name of it. It's a good book. And this has been talked about a lot, and I think AI is making it just more prominent in, in the discourse, if you will.

39
00:06:26.120 --> 00:06:33.960
But you've been, you know, at this at, you know, the forefront of it, right? When I, when I think we f- probably first talked, you know, it was really when BuzzFeed was starting.

40
00:06:34.260 --> 00:06:47.260
And, you know, I always thought about, you know, you, before that at, at, at HuffPost, or Huffington Post then it was at the time, but you really, like, understood the sort of viral mechanics of the time, right?

41
00:06:47.660 --> 00:06:57.520
And, and it was, like, fun back then. Like, and I think you get at that, and I think a lot of... I always try not to be nostalgic, right? But, like, [laughs] things...

42
00:06:57.900 --> 00:07:07.820
It is, it is undeniable, like, things seemed a little bit lighter and, like [laughs] on the internet back then. So very simple question, if you were to look at, like, what went wrong?

43
00:07:08.000 --> 00:07:09.900
What were, like, the three or four things that went wrong?

44
00:07:10.580 --> 00:07:23.720
Uh, I think the biggest thing is when you move along a gradient of, of optimization or virality, it might be that when you have a quantitative different, at, at a certain point it becomes a qualitative difference, right?

45
00:07:23.920 --> 00:07:35.720
So you might have fun viral content that, that is persuasive to people to share and, and you're pl- kind of playing that game, and we certainly w- did a lot of the early work in that. You did a lot of that, Jonah.

46
00:07:35.800 --> 00:07:40.700
A lot of it. Yeah. Yes. The 37 things only the middle child understands. Yeah. Yes.

47
00:07:40.980 --> 00:07:57.660
And, and then when it gets to a certain point, it, it flips from being, you know, a little bit of persuasion here or cleverness there to being something that is, like, deeply manipulative and, and pushes people to, to do things that are against their own interests or their own happiness.

48
00:07:58.280 --> 00:08:07.360
And so I think, you know, there's lots of, there's lots of examples of this. You know, if you're a basketball fan, the Euro step is pretty cool innovation, but if you can take- Yeah...

49
00:08:07.370 --> 00:08:12.180
four or five steps, you know, then all of a sudden you're like, "Wait, th- there's no traveling rules anymore?" Yeah.

50
00:08:12.220 --> 00:08:40.860
You know, so there's a, there's a fine line between something that is, uh, advancement in, say, viral engineering or viral contagious content that helps people spread something that they feel happy about, to something where you're, you're just being completely manipulated by AI that is smart enough to, to manipulate you better than, you know, maybe any, any entity has been able to manipulate people before.

51
00:08:40.880 --> 00:08:48.820
And so I think that's the, the, the, the sort of difference. Like, I kind of describe it as the, there was the beer and wine era of- Right...

52
00:08:48.830 --> 00:08:57.420
of, of the internet, and then there's the fentanyl and crack cocaine era of the internet. And, you know, there's problems with beer and wine. People keep...

53
00:08:57.460 --> 00:09:01.240
There's alcoholism, there's people drink too much, people, you know.

54
00:09:01.260 --> 00:09:29.900
But, but there's also, you know, more clear benefits where people relax and socialize and, you know, fall in love and have fun and, and, you know, you can take that kind of, that kind of metaphor of, like, you know, being a little buzzed might be a positive thing, but being, being to the point where you're, you're, you're not able to control yourself and you're in an addictive loop where you can't s- you know, control your own behavior is something that I think- Yeah...

55
00:09:29.940 --> 00:09:30.740
is, is, is a problem.

56
00:09:31.140 --> 00:09:43.780
And, you know, the other thing which I think is kinda underappreciated is that these algorithms, you know, once they became these deep learning algorithms, they become black boxes where the people working at the companies aren't even sure why the content recommendation is happening.

57
00:09:43.840 --> 00:09:51.140
So, you know, when I would go visit Facebook in the early days, they would say, "Hey, could you help us with feature selection that we could put into our algorithm?"

58
00:09:51.220 --> 00:09:57.140
Meaning, you know, what are the features of content that, that, that, that we think are important to optimize for?

59
00:09:57.520 --> 00:10:02.880
And then once it became this deep learning model, it was like, "We don't even know what's going on, but it's doing an amazing job." And it...

60
00:10:02.920 --> 00:10:27.810
And, and so I think the, the, the challenge with that is maybe it's doing an amazing job 'cause it's figured out how people with ADHD's brains work and is essentially manipulating that, or how people with autism's brains work and they're manipulating that, or how people who are, you know, have trouble with impulse control or maybe, you know, have a level of, like, of, of impulse control around sexuality where they're the f- the- Yeah...

61
00:10:27.840 --> 00:10:32.450
the deep learning algorithm's like, "Oh, we found that this person is gonna click on anything that has this kind of content."

62
00:10:32.950 --> 00:10:51.560
And so you end up with whatever someone's weakness is in their brain is being exploited by algorithms that the people who work at these companies don't even understand to create addictive patterns and compulsive patterns because if you're running a business, you want predictability, and the most predictive, predictable thing is addictive behavior.

63
00:10:51.700 --> 00:10:58.789
If you look at human behavior, there's all kinds of, of, of behavior that's unpredictable, that has a lot of human agency and creativity in it. Mm-hmm.

64
00:10:58.800 --> 00:11:05.980
And then there's addictive behavior where you know exactly what that person is gonna do, and if you can monetize that, it's a much more predictable thing.

65
00:11:06.020 --> 00:11:22.860
And so if you just tell a, a kind of black box algorithm, like, make the user time, you know, the, the user time spent go up and make the conversions on ads go up, you're gonna have a lot of weird emergent effects that the people who work at these companies don't even fully understand.

66
00:11:22.940 --> 00:11:31.860
Yeah. I mean, I always say it's, like, kind of we're likely going to optimize ourselves into oblivion. Like, I mean [laughs] we, we started optimization and it was like, "Oh, this is cool. We can, like...

67
00:11:32.320 --> 00:11:34.760
It's like this is improvement. We're improving things."

68
00:11:34.920 --> 00:11:45.260
And then all of a sudden, like, we gave it, we gave it over to machines and ultimately, you know, whatever, it's the paperclip thing that the- we're gonna be optimized in- into not existing anymore.

69
00:11:45.720 --> 00:11:57.680
But, I mean, you were... I mean, I always th- thought about BuzzFeed as, I mean, almost like a social science, like, back, because you were trying to crack open what got, what got things to go viral, right? Like...

70
00:11:57.700 --> 00:12:09.660
And it was before algorithms were as prominent, right? Like, it was done by personal sharing, you know? Forwarded emails and stuff, links that get passed around.

71
00:12:09.680 --> 00:12:14.320
And then it was, like, things that were gonna be posted on, on Facebook or on Twitter then.

72
00:12:14.360 --> 00:12:27.960
What are you having for lunch and whatnot, but you would also share links and people would get traffic, and it was, it was also, I think what you guys recognized was that the content was a form, sharing was a form of social expression, right?

73
00:12:28.020 --> 00:12:42.288
The, the 37 things only a middle child can understand, it says something about you, and so you wanna share something about you to someone because that, you know, I guess builds some kind of s- connection with other people.Fair?

74
00:12:43.048 --> 00:12:52.138
Yeah, no, absolutely. I mean, we spent a lot of time and, you know, thinking about what's the difference between content and the informational value in content- Yeah...

75
00:12:52.188 --> 00:13:18.528
and social content and, you know, when you share something, you might be sharing it 'cause you're trying to tell the world about yourself, you might be sharing it 'cause you're thinking of the other person, you might be sharing it because, you know, you're, you're trying to show that you're someone who has a kind heart, so you care about cute animals, or you care about a touching story, and then that lets other people in your life know that, that, that they can connect with you emotionally.

76
00:13:18.668 --> 00:13:22.908
You know, there's, there's a whole host of things that, that drive sharing.

77
00:13:23.268 --> 00:13:33.517
I think one of the things that we didn't fully anticipate in those early days is that some of the most predictable sharing comes from anger and fear and outrage. Sure.

78
00:13:33.517 --> 00:13:38.238
And so if you're competing- But to be fair, the platforms nece- didn't necessarily...

79
00:13:38.268 --> 00:13:47.848
It didn't seem like, I mean, at least when Facebook of that era, when Mark Zuckerberg was talking about it, I remember asking him, the only question I ever asked him was like, "Why don't you have a dislike button?

80
00:13:47.908 --> 00:13:55.368
Like, that's how people like, you know, go through life." And he's like, "That's not part of the product." I'm like, "Not the question." But, you know, they were thinking, "We're gonna connect the world.

81
00:13:55.528 --> 00:14:01.798
Everyone's gonna be together, digital Times Square," and all this kind of... uh, Times Square. Di- digital, it has become like the old school Times Square. Mm-hmm.

82
00:14:02.048 --> 00:14:07.768
But digital square, and, um, people are gonna be connected to rear square, et cetera.

83
00:14:07.868 --> 00:14:16.788
And I miss that era of, of optimism in some ways, but you get a lot of people together, and I kinda think this is, like, a little bit inevitable, right?

84
00:14:16.828 --> 00:14:31.368
I mean, there were people who were warning that this kind of utopian view of what social media was gonna do would end up getting controlled by autocrats and used to control and to manipulate people rather than connect them.

85
00:14:31.988 --> 00:14:43.568
Yeah. I mean, I definitely think that a lot of the things we assume are just human nature are a moment in time, you know? And so... and are about the architecture of the systems that people are using.

86
00:14:43.708 --> 00:14:55.408
So, you know, back in those sort of Obama-era days of social media, you had Facebook thinking, "Oh, we're connecting the world," and that was gonna be this, you know, positivity and people connecting and mutual understanding.

87
00:14:55.888 --> 00:15:07.628
BuzzFeed making content that was similar to, you know, to that kind of worldview where it's like, you know, you wanna understand people who are different than you, but the purpose of understanding them is to connect with them and realize we're all in this together.

88
00:15:07.688 --> 00:15:17.008
So there was that period where that was the kinda content that went viral. This just wasn't, like, do-gooder stuff that, that, that we ma- that we made that didn't go viral.

89
00:15:17.048 --> 00:15:22.308
This was some of the most viral content on the internet- Mm-hmm... and it was super positive and kinda brought people together.

90
00:15:22.388 --> 00:15:37.088
And I think that had a lot to do with the early days of the social platforms, that they were more open and it was easier to make things kind of go viral during those days 'cause they were trying to in- you know, incentivize people to upload content.

91
00:15:37.148 --> 00:15:51.197
It was kind of culturally the Obama era, which was, you know, defined by multiculturalism and people, you know, connecting with each other, and new groups of people who no- weren't part of the conversation joining the conversation and being part of, part of something bigger.

92
00:15:51.208 --> 00:16:06.838
And you had, you know, researchers like, uh, Jonah Berger writing a book called Contagious that, that, y- you know, analyzed what goes viral, and it was like, "Positive things go viral," [chuckles] and, you know, and, and occasionally there's an exception of negative things- Things, things, ThingsThatDidn'tAgeWell.com.

93
00:16:06.868 --> 00:16:13.628
Yeah, so yeah. But then a c- you know, you had the one exception, which was negative things where there's outrage, and you say you share it because of the outrage.

94
00:16:14.068 --> 00:16:23.388
But, but then, you know, time passes, and if you just did all the same research and all the same studies about, like, what goes viral, like, today it would be totally different types of things.

95
00:16:23.958 --> 00:16:31.448
And so, you know, does that mean- When, when... Let me ask you this. When did you start... When did you... 'Cause again, you, you've been, like, studying this for, for forever.

96
00:16:31.508 --> 00:16:42.958
Like, when did you start to s- to s- either see that shift in the data? Because, like, I get the, the sense that you're a little conflicted, right? Because, like, BuzzFeed to me was always about...

97
00:16:43.048 --> 00:16:44.928
Like, it was really good at, at...

98
00:16:45.668 --> 00:16:55.128
When the content that was being shared was, was the, the puppies and the cats and the, the dress and the, you know, the, the middle child stuff, you know, that was, like, sort of what you wanted. You wanted to be...

99
00:16:55.228 --> 00:17:00.058
It was the bored at work network, you know? And it was, like, a thing to... It was, it was a good thing, right?

100
00:17:00.148 --> 00:17:07.828
And, and now if you're just f- if you're just chasing, if you're just following the signals, right, it's, like, pretty clear what the signals are saying, you know?

101
00:17:08.228 --> 00:17:20.168
[chuckles] Like- Right, and we have- It's like, "Get into the fighting pits." Yeah, and we have, you know, a company that has a mission and values, and so we weren't just making anything to make it go viral.

102
00:17:20.208 --> 00:17:21.538
You know, we were- Right...

103
00:17:21.568 --> 00:17:34.288
we were spreading truth and joy, and, and we, you know, continue on that mission even when it's, it's, it's kind of suboptimal now d- 'cause the fighting content is what's, you know, goes viral.

104
00:17:34.338 --> 00:18:06.008
And it's partly why we're building our own platform, and it's partly why we are really working on optimizing for the platforms in a way that allows us to fight against, you know, what, what I call SNARF, which is, you know, the kind of content that seems to be proliferating right now, which is, you know, it's kind of defined by fear and anger and, you know, a magnification of stakes, so everything is existential and kind of fabrication of novelty, so everything is super novel.

105
00:18:06.368 --> 00:18:25.278
And so there's this ti- type of content, and I think understanding the viral dynamics and how that content works will help us, you know, f- kind of fight against that content and make some content that still has a sh- chance at breaking through that is, is more positive and, and helps connect people together instead of just, like, an endless war of every tribe against every other tribe.

106
00:18:25.278 --> 00:18:32.218
Yeah, and if anyone wants to see SNARF in action, just open up X, like, literally anytime, and you will see it, because it's a very different...

107
00:18:32.258 --> 00:18:39.018
To me, it's like, it's a very different place, obviously, than it u- it used to be, and it's not about political persuasion.

108
00:18:39.088 --> 00:18:43.808
It's just, like, you're like, "Whoa, this is a very, very different product," and it's, it's very intense.

109
00:18:43.968 --> 00:18:56.160
I'd, I-I compare it to menthol cigarettes, [laughs] but I mean, that's maybe I'm being too kind not saying fentanyl 'cause it's like anyone who's ever smoked a menthol cigarette never got up the next day and was like, "Wow, I should've had more menthol cigarettes."

110
00:18:56.700 --> 00:19:10.060
I mean, it raises these questions about our capitalist system when a higher and higher percentage of GDP is things that are taking advantage of addictive cycles as opposed to things- Yeah... that are giving utility.

111
00:19:10.100 --> 00:19:20.290
And, you know, what does that mean if you have an economy where, you know, there's a huge portion of it that's, you know, driven by people who are in compulsive cycles. You know? Yeah.

112
00:19:20.300 --> 00:19:25.370
Like, the majority of alcohol sales are from people who are alcoholics- Well, look at the gambling... the majority of... Yeah.

113
00:19:25.370 --> 00:19:33.520
I mean, most of the on- I mean, to me, like, legal- this is a total aside, but, like, legalizing, like, sports betting, like, is gonna be looked at as, like, kind of a disaster.

114
00:19:33.600 --> 00:19:41.470
Maybe it was inevitable in our system, but it's preying on people. I mean, my God, thank God they didn't have that back when I was in college. Like, if [laughs] somebody- [laughs]...

115
00:19:41.520 --> 00:19:50.200
there would've been a lot of degenerate gambling- gamblers in McDermott Hall. You know, but you had to go to a bookie, like some sketchy guy dow- down, like, like... No.

116
00:19:50.620 --> 00:19:58.840
And sometimes you need friction for things because these things are, are very addictive. Let me ask you this, like, before we get into the SNARF stuff. How is...

117
00:19:58.880 --> 00:20:09.320
How much of this is just, don't take this the wrong way, but I think of, like, of BuzzFeed as, like, part of the millennial internet, and this is, like, now a different internet. Maybe it's the Gen Z internet.

118
00:20:09.780 --> 00:20:13.540
I think, you know, Facebook was the network of choice then.

119
00:20:13.900 --> 00:20:38.420
Now it's TikTok, and because TikTok's algorithm is all the things that you're describing, 'cause to me, like, if you were to look at the main sort of things that changed, the introduction of the newsfeed, you know, then when you combine that with the most sophisticated algorithm around of TikTok, and everyone seems to me, everyone reverse engineered the TikTok algorithm 'cause it was the best, and they just said, "Ah, forget it.

120
00:20:38.460 --> 00:20:45.960
Forget about followers or any of that stuff." But this, this works best for, you know, grabbing attention. It's, it's really good.

121
00:20:46.030 --> 00:20:57.730
But, like, i- how much of it is just, like, a cultural change of the, of the different internets? Well, I mean, where did, where do all the millennials go if they, if they... There's a cultural change- [laughs] I know.

122
00:20:57.730 --> 00:21:01.600
Where did they go?... and now the internet that they like doesn't exist anymore. Pinterest? I don't know.

123
00:21:02.120 --> 00:21:15.740
[laughs] So I, I, I mean, I think it's more than just a, a cultural change, otherwise you, you know, like in the past you'd have a generation grows up reading a magazine and they just age up with that magazine and they're still reading it, you know, when it's no longer the kids' magazine.

124
00:21:15.809 --> 00:21:27.550
Now it's a grown-ups magazine. I think this is a little bit different, where all the platforms are changing, and to me it feels like part of this larger problem of, of AI taking away people's agency. Yeah.

125
00:21:27.640 --> 00:21:40.020
I, I like AI, but I like it as something that extends our agency, where you're using AI tools to make things you couldn't have made otherwise to, you know, extend your, your will and your ideas and your creativity in the world.

126
00:21:40.060 --> 00:21:44.340
Assistive AI, basically. Yeah. It can do a lot for you.

127
00:21:44.400 --> 00:21:55.499
It can be super powerful, but you're directing it, and I think when you look at the AI, you know, like your example of, of, well, you used to, it used to be about followers and things like that.

128
00:21:55.600 --> 00:22:08.220
Well, y- you know, there's problems with the follower-following model of social media. But at least you're picking who you wanna follow. You're like, "Oh, this person's... has some pretty good tweets.

129
00:22:08.260 --> 00:22:15.660
I'm gonna follow this person," or, "This person's Instagram is interesting. I'm gonna follow this person, and then I'm gonna get some content from them." There's some agency there.

130
00:22:15.720 --> 00:22:21.120
You're choosing who to follow, and then that's influencing what, what kinds of content you get.

131
00:22:21.740 --> 00:22:51.480
The TikTok model, because AI is matching content to people based on, you know, these deep learning recommendation that is, is sort of figuring out how to make the most addictive content possible appear in your For You page, you end up losing a lot of agency, both because you're not picking who you're following and you're not picking which types of content reach you, and because the algorithm is so good at commanding your attention that you just spend hours and hours on this feed.

132
00:22:52.040 --> 00:23:04.840
And at the end of it, you often, you know, like the, the study that I, I linked in my memo where people were, you know, a large percentage of people were saying they would pay money to have TikTok and Instagram not exist, you know?

133
00:23:04.860 --> 00:23:15.480
So they're like, "I use this product all the time, but I would pay to not have this product in the world." It shows it's not giving positive utility- Yeah... but it's giving, it, it's taking away their agency.

134
00:23:15.560 --> 00:23:23.610
It's why people don't like it is because the, the AI is taking their agency away. Yeah. And so that's what we're fearing in the future is AI's gonna take our agency- Yeah...

135
00:23:23.640 --> 00:23:33.430
the robots are gonna control us, but it's already controlling a big portion of our, our, of people's days, and they don't like it. Yeah, and, and so the question is- Do you, do you... Well, let me ask you this.

136
00:23:33.480 --> 00:23:43.160
Do you include YouTube in this? I think that- Is it the last good platform? I think that YouTube, you know, c- c- is kind of like right there on the l- line.

137
00:23:43.300 --> 00:23:49.560
I think that, that YouTube Shorts have made it a little more TikTok-like- Yeah... which creates some of these kind of problems.

138
00:23:50.020 --> 00:24:02.590
But YouTube also, with a longer form video, you know, you're, you're, there's fewer points at which an algorithm's determining what you're seeing, and more points where you're watching something longer- Yeah...

139
00:24:02.600 --> 00:24:13.780
and spending time, time with it. So I do think YouTube has, you know, mitigated some of these things, and also it has such a deep archive of things from the past that it also, that also I think has, has helped YouTube.

140
00:24:14.020 --> 00:24:19.300
Yeah, I mean, there's junk food on YouTube, but there, there's a lot of nutrient-dense food on offer there.

141
00:24:19.330 --> 00:24:27.339
And, and I think that's the challenge of a lot of this stuff is like I, I'm totally with you, where I, I'm just like, oh my God, and maybe it's getting older where I'm like, I can't...

142
00:24:27.380 --> 00:24:33.960
This, this stuff is breaking my brain. It's gotta be breaking other people's brains. I don't feel good after. I, I feel like I went to the movie...

143
00:24:34.020 --> 00:24:40.140
Uh, if you go to a movie and you're just like, "Oh yeah, I'll take the..." They try to upsell you on, like, the large popcorn, never feel good afterwards.

144
00:24:40.620 --> 00:24:42.960
[laughs] Kind of feel that a little bit with social media these days.

145
00:24:43.460 --> 00:24:54.420
But at the same time- Yeah, the lar- the large popcorn is a good example, 'cause it's essentially-When behavioral economics is, is, is becoming more important than economics, right? Yeah, exactly.

146
00:24:54.440 --> 00:25:05.880
Where it's like, where it's like it's, it's a lot harder to create a product that creates lots of utility for a consumer than it is to figure out a man- a way to manipulate people into extracting more value from them.

147
00:25:06.280 --> 00:25:16.630
And so that's the problem, is too much of the economy becomes, "Oh, we're gonna have the right different sizes of popcorn, and we're gonna only charge a little bit more for the bigger one," but then we're sucking more money from- Right...

148
00:25:16.660 --> 00:25:26.000
out of them, and then they're eating more than they really wanna eat, and it's like nobody wins except the, the, the, the seller of the product is, is essentially extracting- Right...

149
00:25:26.020 --> 00:25:32.890
money from people, but without giving them additional utility. It- it's, it's... You know, that's, that's essentially the, the problem. Yeah.

150
00:25:32.900 --> 00:25:38.780
And I think social media is one example, but it's popped up in lots of other places in our economy as well. Yeah, it's funny 'cause that, that's the sort of line.

151
00:25:38.820 --> 00:25:54.580
Because, like, I think the easy sort of objection to, to you doing the anti-snarf, you know, memo is that, well, wait a second, you guys were manipulating people back when it was, like, not very sophis- not as sophisticated as [chuckles] it is now.

152
00:25:55.000 --> 00:26:08.260
And now, you know, machines a- as it turns out, are way better at doing manipulation of people. But the degree of manipulation is completely different. I, I, I can see that sort of...

153
00:26:08.300 --> 00:26:12.540
You know, I think it's, it's a nuanced thing, at least how I see it. Yeah, for sure.

154
00:26:12.700 --> 00:26:40.500
I mean, uh, look, I, I feel like BuzzFeed and myself personally wa- you know, we, we were pioneers of thinking about content and how it spreads on the internet and how to make, make viral content, and, you know, you asked earlier when did we hit the point where we, where we saw it was tipping over to this other place, and I think it was when Facebook launched what they called meaningful social interactions, this new thing that was supposed to improve their algorithm.

155
00:26:40.520 --> 00:26:50.600
Sounds great. Yeah, it sounds great, and looking back on it, that was a real tipping point for the internet because they looked at comments more than they looked at o- other signals.

156
00:26:50.640 --> 00:26:55.300
To me, I always thought sharing, people sharing with someone in their life- Mm... was more meaningful.

157
00:26:55.360 --> 00:27:06.120
Like, you're, y- you know, you- you're, you're not gonna share a piece of content with, you know, someone you love or a friend or a family member if you think the content is ter- terrible junk food.

158
00:27:06.600 --> 00:27:15.480
You're gonna share it- Yeah... if y- if it helps you connect to that person. And so that was earlier. Facebook was more about sharing. And then when this meaningful social interaction thing, it was more about comments.

159
00:27:15.540 --> 00:27:21.080
And, and so eliciting a comment, the fear and outrage thing really drove a lot of comments.

160
00:27:21.100 --> 00:27:32.180
And so if you put something that was controversial or even a piece of content that was broken, like there was a factual error in it, you get all these comments of people saying, "That's not true," and, "I hate that," and, "This is wrong," and, uh, you know.

161
00:27:32.540 --> 00:27:37.960
And there's all this fighting. Yeah. Especially if it's, like, a two-sided issue, and then that would drive adoption in the newsfeed.

162
00:27:37.980 --> 00:27:48.520
And then Facebook just slowly became a place for people to fight because the algorithm that they s- they sort of switched to, what, you know, was, was favoring things that elicited comments.

163
00:27:48.580 --> 00:27:57.080
And that was because Facebook was worried that, that, that less people were s- contributing data to Facebook and more people were becoming passive and just consuming media.

164
00:27:57.520 --> 00:28:04.580
But that meaningful social interactions change, I think, was the first step of kinda ruining, ruining the, the internet.

165
00:28:04.660 --> 00:28:15.280
And, you know, we had a post that someone wrote called, "Olives Are Terrible," and it was all about how olives are just these nasty salt bombs that do- taste terrible, right?

166
00:28:15.640 --> 00:28:25.040
And of course, a lot of people in the comments, you know, were like- And they break teeth. Yes. Yeah, there's a danger. That's- There's a safety issue... so it led to this huge fight, and then that went viral.

167
00:28:25.130 --> 00:28:29.750
And then really good content we created didn't go viral, and so I, like, sent that off- Wow...

168
00:28:29.750 --> 00:28:39.720
to Face- uh, Facebook's head of newsfeed, and I was like, "Why is this terrible article about olives being terrible getting all this distribution be- and it's just because it's a fight?"

169
00:28:39.800 --> 00:28:53.620
And, and, you know, we d- we were not in a position to make the, the, the content about the things that people fight even more about, like, things like abortion and, and race and- Yeah... you know, politics.

170
00:28:53.640 --> 00:28:57.400
And, like, we were, we didn't wanna make that kind of, kind of toxic content.

171
00:28:57.900 --> 00:29:10.300
So we're, we became at a disadvantage because, you know, the Breitbarts of the world could, could create their, like, you know, immigrant watch content and lead, which led to these massive fights online that got lots of distribution.

172
00:29:10.320 --> 00:29:21.880
And, and so, you know, I mostly was, like, trying to lobby the platforms of like, "Hey, you really made a huge mistake here. This is gonna lead to lots of tox- toxicity and fights and polarization."

173
00:29:21.940 --> 00:29:29.880
But, you know, it turned out that that, that was kinda okay for their business. Okay. So- So what is the, what, what is the AI role in this?

174
00:29:29.960 --> 00:29:37.820
Obviously, I mean, AI, it's thr- thrown around a lot, and obviously these, these algorithms have gotten incredibly sophisticated. They're using AI within them and, and whatnot.

175
00:29:37.840 --> 00:29:42.770
But now AI is also moving to, like, the creation sphere. And, like, you know, you...

176
00:29:43.120 --> 00:29:55.200
I think you have a somewhat nuanced take on, like, AI's role, particularly in creation, because, I mean, a year ago you talked about using more AI, AI for quizzes, and I've seen... I assume that it'll be true.

177
00:29:55.240 --> 00:30:03.740
I've seen studies, you know, 98% of the internet's content will be AI-generated, i- if it isn't already. I mean, there's a lot of [chuckles] AI stuff out there.

178
00:30:03.780 --> 00:30:14.060
Facebook, and from what I'm told, I've not been on Facebook in, like, seven years, is, like, filled with AI slop. Just the fact that there is the term AI slop out there tells you that it's, it's pretty prevalent.

179
00:30:14.080 --> 00:30:27.320
But what is your view on how AI can be, like, a force for good? Because I think it's got an interesting role in that it's on both sides [chuckles] of this. Like, it... The only way is to, like, you can't...

180
00:30:27.440 --> 00:30:39.000
I'd love to sort of do the sort of Brunello Cucinelli, like, Solomeo kind [chuckles] of thing, but, like, I think the only way out is to, is that it's gonna be AI battling AI in some ways.

181
00:30:39.560 --> 00:31:05.312
Yeah, I'm definitely not anti-AI, but I think if you have a really powerful deep-learning model and you say to it, "Maximize the amount of time people spend on this, you know, s- on this app," or just, "Maximize the amount of sales"-You, it leads to a fairly dystopian thing, you know, type of situation where you're, the AI is at odds with human agency.

182
00:31:05.672 --> 00:31:15.832
You know, you're basically telling it, "Control these people and get the behavior out of them that we want." That's different than saying, "Here's a bunch of different creative AI tools, use those tools."

183
00:31:15.912 --> 00:31:25.172
So you know, BuzzFeed created a emoji generator where if you had an idea for emoji that didn't exist, you could- Hmm... you could have our AI create that for you.

184
00:31:25.212 --> 00:31:30.432
You know, and then like a couple of months later, Apple came out with a emoji generator that was not nearly as cool as ours.

185
00:31:30.612 --> 00:31:42.552
But, but that kind of flip where this isn't just, you know, s- the AI is trying to get you to spend all your time, you know, whatever, buying products or consuming con- Starf content.

186
00:31:43.032 --> 00:31:54.962
The, your, y- the AI is at your service, and you can use it to, to do fun things or c- creative things and, you know, play games. We, we created a nepo baby simulator where you could- Yeah. I saw that...

187
00:31:54.962 --> 00:32:06.812
try to raise your nepo baby. We, you know, we've, we've built a lot of things that, that give the audience the, the agency to use the AI- Yeah... to, to make things, create things, play, connect with other people.

188
00:32:07.412 --> 00:32:08.752
Is there, is there a market for it?

189
00:32:09.012 --> 00:32:21.582
Like, I mean, honestly, I'm, I'm just like these platforms are incredibly popular for a reason, and it goes without like, you know, yes, they have addictive, uh, qualities, but, like, to me, a lot of these...

190
00:32:21.712 --> 00:32:31.412
Like, you mentioned, like, fentanyl, unfortunately we're still in, in the midst of it, but it is, it is, I think, declining. Like, y- the crack epidemic, you know, went mostly away, right?

191
00:32:31.492 --> 00:32:40.012
And I think a big, I think studies have shown it, like one of the big reasons is it was totally uncool to do crack. Like [laughs] it was just like you do not wanna grow up and be like...

192
00:32:40.052 --> 00:32:51.112
You know, crackhead was not a, a, a term of, you know, a compliment at all. And so a generation grows up and sees, you know, crack, you know, addicts around. They're like, "I don't wanna do that."

193
00:32:51.172 --> 00:32:59.932
Like, that's, see, that's- Yeah. No, I, I-... crack is what-... I don't wanna overstrain the analogy, but you are starting to see kids who feel that way about- Yeah... being on Instagram all the time.

194
00:33:00.232 --> 00:33:10.132
You know, where they, it's not cool. You know, it's like, "Oh, you're one of those people who doesn't do anything," and you know, "We're gonna go camping," or, "We're gonna g- you know, put our-" Yeah...

195
00:33:10.141 --> 00:33:15.592
"phone down and do other things." And look at all the run clubs in New York. Why else would that be? I mean- Yeah... people are good of revolting against, like, dating apps.

196
00:33:15.812 --> 00:33:26.772
[laughs] So but let's, let's talk about BF Island 'cause, like, I, I read through it. It's, it's hard to... I, I don't know. I, until I see the product, I don't, I don't really know. But what, what is the...

197
00:33:26.812 --> 00:33:39.332
The idea is to create a, a different place where people can use a lot of AI tools to express themselves, but it's not like, it's not a place for content. Ex- well, content is, is, is a different term.

198
00:33:39.352 --> 00:33:51.412
But, like, this is not like a place to bring together, you know, the Tasty content, the Huff- HuffPo content, like, all of that content. It's not a place for professional editorial content. Okay. You know?

199
00:33:51.472 --> 00:34:03.692
So our f- our, our approach will be Tasty, BuzzFeed, HuffPost should be the best possible editorial businesses, publishing businesses that they can be. We should run them lean.

200
00:34:03.712 --> 00:34:15.042
We should make great content that is trusted. We should hire great writers. And- Okay. This is building the new business... and maybe, maybe there's ways AI could help make the team- Okay...

201
00:34:15.042 --> 00:34:24.452
more efficient, but they're not gonna write the content. Now I get it. You know, they're, you know... And so, so that's, that's our core business, and we wanna keep innovating and evolving that core business.

202
00:34:24.942 --> 00:34:33.612
But this is something new, which is user-generated content, but imagining a social media platform that is AI native.

203
00:34:33.792 --> 00:34:50.432
So all the big social media platforms were built a decade plus ago, t- 20 years ago, something like that, and they were all built with the idea that users submit content, and then the content is somehow organized and spit back to, to, you know, in, in some ranked way.

204
00:34:50.492 --> 00:34:56.012
And, and you try to make, you know, an addictive consumption behavior and have enough people cons...

205
00:34:56.352 --> 00:35:03.692
you know, contributing content that there's enough supply of content for everyone consuming it, and AI was really not part of it.

206
00:35:04.252 --> 00:35:13.682
And one of the things that you saw happen over time is that all of these platforms ended up with a small creator class making all the content, and then everyone else very passive.

207
00:35:14.221 --> 00:35:25.812
And this is something that Facebook fought for years because they, you know, the very early days of Facebook, when it was really just about building your network and connecting with your group of friends, every, like, 80% of people contributed data to the service.

208
00:35:25.892 --> 00:35:31.812
Like, they would post things and do updates. And, and that, you know, they were naive back then. They thought that was normal.

209
00:35:31.832 --> 00:35:38.432
But that was, like, this ridiculous level of participation in a social platform that, you know, was pretty unprecedented.

210
00:35:38.752 --> 00:35:52.532
Almost every other platform, it, it, you know, it, it trends towards having, you know, 10% of people submit content, maybe 1% submit the majority of the content, and then the 99% of people are consuming content that that small group of people makes.

211
00:35:52.552 --> 00:36:03.792
You know, and so Reddit, it might be super active posters or, or mods, and, you know, on Twitter, it's the big accounts that are constantly tweeting and, a- and the average person doesn't tweet at all, or the average Facebook...

212
00:36:03.832 --> 00:36:07.992
You know, and, and, and, and so what happened with Facebook is over time, the average person stopped posting.

213
00:36:08.552 --> 00:36:15.492
And so people were just going and then, and then they, you know, they'd consume BuzzFeed content or creator, and then after that, creator content or things like that.

214
00:36:16.052 --> 00:36:31.992
And you, you ended up with, with something that wasn't really a creative social network where everyone's participating, and you ended up with a new kind of media where people are just consuming media made by other people, and the other people are creators, or the other people are media companies, but it doesn't really matter.

215
00:36:32.032 --> 00:36:37.772
It's, it, it becomes, like, a small number of people posting content and a large number of people consuming it.

216
00:36:37.892 --> 00:36:45.292
And despite what Zuckerberg and, and others tried to do, they ended up with media platforms, not really social media platforms. Hmm.

217
00:36:45.632 --> 00:36:55.812
And so what I think AI can unlock is the ability for more people to create content because it's a lot easier, and the barrier to entry is, is much lower.

218
00:36:56.252 --> 00:37:04.912
And so if it's, you know, like the emoji generator example, and you can l- just think of a funny idea for an emoji-And then the AI makes it.

219
00:37:05.372 --> 00:37:14.642
That's a lot lower bar, you know, for participation than if you have to, like, design a new emoji and make something that's visually cool. Yeah.

220
00:37:14.652 --> 00:37:30.152
And so I think there's a lot of examples like that, where people can take quizzes or use little toys or generators, can play games, can do things where a little bit of action leads to a lot of output because the AI is helping you do it.

221
00:37:30.192 --> 00:37:39.392
But, like, is, is the social graph, as you said... I haven't heard that in a while. Is that still, like, important? Is it, like, mapping people to each other? 'Cause, like, when I first... I was like, "Oh my God, wait.

222
00:37:39.532 --> 00:37:46.481
Isn't this like 10 years too late?" Like, I mean, this sounds [laughs] like, I mean, th- th- maybe this should've been, like, an alternative to, to Facebook.

223
00:37:46.532 --> 00:37:54.772
I mean, look, there was Path, there was, there was some other, like, alternatives that were taking slightly different approaches. They, they... N- None of them was able to compete with, with Facebook.

224
00:37:54.822 --> 00:38:01.892
And, you know, this, these, these platforms have shown that- Maybe they, maybe they were 10 years too early. [laughs] That could be.

225
00:38:02.052 --> 00:38:15.032
They, they- But these platforms have shown, once, once they get going and once they have distribution power, and the internet has again and again, uh, taught us this lesson, that the commanding heights are being able to control distribution 'cause you can do whatever you want.

226
00:38:15.492 --> 00:38:24.971
You can do whatever you want. So you can make every error in the book and you're still gonna win. There's a reason that Google never has given guidance, 'cause they don't have to. [laughs] They just don't.

227
00:38:25.252 --> 00:38:30.102
They'll just put another ad on a methylazima, like, results and... So is, is that...

228
00:38:30.372 --> 00:38:40.372
Is, is it to be, like, a destination and an alternative to, you know, these platforms that, depending on point of view, are, are fairly toxic? Yeah, it will be. The...

229
00:38:40.652 --> 00:38:50.332
I, I, I guess I would just say that new things, new apps usually break out when new technology enables things that weren't possible before.

230
00:38:50.342 --> 00:38:57.512
And so I think part of the issue is, you know, when Facebook was new, building something like Facebook was hard. I don't know.

231
00:38:57.552 --> 00:39:05.292
I mean, we're, we're old enough to remember when Twitter would go down all the time and- Oh, yeah. Fail whale... there were blog- Yeah. And there were blog posts that- Look at a T-shirt. Yeah, exactly.

232
00:39:05.352 --> 00:39:16.621
[laughs] The fail whale. There were blog posts that people would write, technical people, that would say, "It may not be possible to build an architecture with real-time simultaneous connections like Twitter- Damn...

233
00:39:16.672 --> 00:39:28.392
because no database architecture would be able to handle it." And I mean, I never thought those... Th- That, that line of thinking made, made sense, but- Now we got Three Mile Island o- reopening, so. [laughs] Yes.

234
00:39:28.492 --> 00:39:36.332
But it was, it was, it was hard to build these social platforms when they emerged because... And new technologies enabled them to exist, but they were... It was still hard.

235
00:39:36.852 --> 00:39:43.072
Now, it's trivial to build all these kinds of things, but the, the big platforms already have all the audience and all the distribution.

236
00:39:43.092 --> 00:39:59.192
But there's new technology, which is this gen AI technology, and, you know, if we can build something that is really playful and social, where there's a higher number of people participating and not just this, like, creator class that's making all the content and competing to make the most snarky content for everyone else to consume.

237
00:39:59.772 --> 00:40:08.312
If we can, if we can succeed at that, it's hard for these big platforms to copy what we're doing because if they did, they would make so much less money.

238
00:40:08.372 --> 00:40:15.512
Because it's so profitable to have people spend three hours a day staring at a feed of snarf that has a bunch of ads that are super targeted.

239
00:40:16.012 --> 00:40:25.292
A- And, and so we don't need to build something that is as valuable, um, as, as, you know, Facebook's newsfeed or Instagram speed.

240
00:40:25.752 --> 00:40:37.412
We, we can build something that is more playful, more consumer-centric, more fun, with a higher level of participation for all the people using the, using the app, that uses new technology that didn't exist before.

241
00:40:37.422 --> 00:40:39.292
And that's why now is a good time to build it.

242
00:40:39.372 --> 00:40:52.112
I, I, you know, I never wanted to build a social platform, you know, 10 years ago because there were so many of them and w- there wasn't social content to spread across these platforms.

243
00:40:52.152 --> 00:40:54.872
And so that's why we focused on, on the content- Yeah... layer.

244
00:40:55.392 --> 00:41:06.932
Um, now there's less places for social content to go, and the apps have basically become content companies where there's a small number of creators and a large number of consumers who passively consume snarf.

245
00:41:07.512 --> 00:41:25.292
And so building a social platform that actually is truly social and creative and playful feels like a big opportunity because technology exists now in the form of gen AI that makes the, makes it easier for everyone to participate without the heavy lift of, you know, being, like, a Instagram influencer or something like that.

246
00:41:25.371 --> 00:41:32.752
Yeah. And you're sort of going back to when, like, BuzzFeed was like... I mean, you used to have tools for people to create their own. Like, I mean- Yeah... I guess you still do- Yeah, we still do...

247
00:41:32.802 --> 00:41:41.852
but I mean, it was, there was more... There was a platform part of it that, you know, wasn't just professional, the, the professional creators, I guess, that are part of BuzzFeed.

248
00:41:42.022 --> 00:41:53.452
Um- Yeah, most BuzzFeed quizzes now are made by the community. Okay. Is this, like, an off-ramp from publishing? I mean, I don't blame you, but let's be clear. [laughs] I think, I- Well, it's different. This is, like...

249
00:41:53.512 --> 00:42:01.052
I mean, you're talking about being basically a, a tech company. Like, BuzzFeed was always, like, a, a tech-enabled media company. I mean, I don't know, I told investors.

250
00:42:01.132 --> 00:42:05.992
But, like, I always think, [laughs] you know, it, it... You, you built technology, right?

251
00:42:06.132 --> 00:42:17.812
But, like, you know, you, you were heavily, and you still are, obviously, heavily into the professional- professionally created content game. I mean, I think user-generated content won the internet. Like, it won.

252
00:42:18.012 --> 00:42:30.572
It wasn't very close and that's just, that's, that's reality. Yeah, I think you can build a good publishing business, but you... I, I think there's, there's upside in this BuzzFeed Island- Define good...

253
00:42:30.602 --> 00:42:33.452
uh, social media. Small and good. Yeah, you can.

254
00:42:33.672 --> 00:42:46.672
I, I, I guess, like, definitely the ambition has been knocked down a bit on the publishing digital media side, where it's not like we're gonna put a bunch of R&D into continually evolving to make a better digital media, uh, property.

255
00:42:46.712 --> 00:42:59.071
We'll, we'll, we'll do some innovation. We'll run things lean and ef- and efficiently. We'll try to be smart about how we distribute our content, and I think you can build a good, profitable business. But I think that...

256
00:42:59.112 --> 00:43:14.988
But I think what BuzzFeed Islands represents is a asymmetric bet-Inside of, of a more, you know, steady state business where it has the potential to really blow up and become something big Mm-hmm And it also is high risk, and maybe it won't, maybe it won't work.

257
00:43:15.368 --> 00:43:27.428
But it has the scalability that you can really only get with a tech-powered type of business, where it could continue to scale as new users join, and those new users use these AI tools for content creation.

258
00:43:27.748 --> 00:43:37.658
You can, you know, sc- scale up and grow to... in a, in a way that digital media companies can't, can't because of the limitations of- Yeah... of the model.

259
00:43:38.188 --> 00:43:44.688
So is announcing it ahead of time, like, a way to get, like, momentum to get, like, the, the resources to be able to...

260
00:43:44.748 --> 00:43:54.588
Or, I mean, obviously the costs of a lot of things have come down with, with AI and coding and whatnot, but I'm wondering what... You know, 'cause I, I always like to see things.

261
00:43:54.628 --> 00:44:01.308
Like, and I'm like, I don't- I mean, it sounds good, but I don't, I don't know because I haven't seen it or used it. I'm sure a lot of people feel that way. Yeah, totally.

262
00:44:01.428 --> 00:44:11.388
I mean, I, I think, I think it's easier to work on things when you can talk about them, you know, internally and externally. If we're gonna talk about something internally, you know- Yeah, code name...

263
00:44:11.968 --> 00:44:18.448
there could be a leak, there could be leaks, or it could be- Yeah... you know. So, so just, just, yeah, talking about our ideas.

264
00:44:18.468 --> 00:44:25.998
And then also we've had a, a lot of people reach out, you know, with interesting ideas who wanna help and participate 'cause they've been looking for something like this.

265
00:44:26.028 --> 00:44:38.008
So that also it's, it's a talent recruitment approach as well. Yeah. So w- a couple qu- quick questions on, on the business itself, is the, the SPAC, right? Could that have worked? I mean, like, y- you...

266
00:44:38.038 --> 00:44:43.908
'Cause I, I, I got it, like, at the time, like, of saying, okay, there, there should be, like, a... And a lot of people...

267
00:44:44.548 --> 00:44:58.468
Not a lot of people, but there were a lot of people, but others were talking about this absolutely at the time, right? About, you know, combining a lot of these digitally native media companies to have scale, right?

268
00:44:58.728 --> 00:45:10.208
And obviously when you guys did the SPAC, it was a, it was a different era, and it was at the end, I guess, of the SPAC. Chamath had already gotten out, so like, it, it... But, like, could that strategy have worked?

269
00:45:10.888 --> 00:45:18.568
Because, you know, you always go back, and sometimes it's just, it's timing, sometimes it's execution. You know, sometimes it's just like, this could never have worked anyway.

270
00:45:18.608 --> 00:45:27.698
Like, you know, you, you make decisions based on, you know, the, a lot of information that you have at the time and, and that's why, like, going back in time sometimes is, it's so artificial.

271
00:45:27.868 --> 00:45:41.488
But, like, could that have worked? I mean, the timing was really terrible. And even with the bad timing, we ended up, you know, selling Complex and First We Feast for around the amount of cash consideration that we paid.

272
00:45:42.028 --> 00:45:51.368
So despite terrible timing, we, we, we, we were able to unwind it. But I think if we had, you know, j- I mean, timing is always easy to play with.

273
00:45:51.428 --> 00:45:55.528
If you, you know, if you had perfect timing, obviously you'd be the best- Yeah... investor in the world.

274
00:45:55.628 --> 00:46:07.268
If we had gone out in the SPAC market six months earlier when the SPAC market was hot, and we just did it ourselves without the Complex acquisition, we would've raised a whole bunch of cash, we would've had a bunch of cash on our balance sheet.

275
00:46:07.308 --> 00:46:13.128
And then when the market kind of crashed and digital media values dropped, we could've bought things for really cheap- Yeah...

276
00:46:13.138 --> 00:46:23.497
and we'd have cash on our balance sheet and a bunch of assets that we could've gotten for very cheap. So if, if, if we had perfect timing, you know, we would've done the SPAC sooner. I know. That's the timing thing.

277
00:46:23.518 --> 00:46:26.458
But- We would've raised a lot of money, and then we would've bought things for- Right... for much cheaper.

278
00:46:26.458 --> 00:46:36.398
But, but- But w- but w- the market still could've, like, hated, like, the idea of, like, a digital media, like, roll-up basically of a lot of these assets. I guess it depends on the assets, I'm not sure.

279
00:46:36.398 --> 00:46:43.837
But nobody pulled it off. Like, there isn't... I mean, I guess Dotdash Meredith, but they, they, but they bought different assets, I mean, from what people were talking about.

280
00:46:44.358 --> 00:46:51.088
'Cause I just like, I guess it's just the, it's, you know, it's like... And I think you're talking about this model not being like completely ad-driven.

281
00:46:51.148 --> 00:47:06.448
It's like, I mean, if you could get out of the ad business, like [laughs] being a publisher and relying on the ad business has turned out to be, like, not a great business, you know, for many people, unless you're in specialty areas, you're doing some influence ads for lobbyists or whatever.

282
00:47:07.208 --> 00:47:18.208
Yeah. But part of it, it's hard to unwind the, the, the timing from whether is it a good business or not, because, you know, a lot of the consolidation happened at the top of the market, and then the market dropped.

283
00:47:18.328 --> 00:47:29.628
And- Mm-hmm... and if you're using debt to do the consolidation or, you know, you, you start to have overhang that makes these things challenging. I think value and price matters a lot.

284
00:47:29.688 --> 00:47:43.207
And so the digital media market was inflated and was... Companies were, were overvalued, you know, BuzzFeed included, where, where the market, you know, was, was giving very, giving a premium.

285
00:47:43.248 --> 00:47:48.028
And now it's undervalued, where people just are staying away from digital, digital media. Mm-hmm.

286
00:47:48.178 --> 00:47:58.588
And I think there is something in the middle where you could, you know, like someone could, could, could do some consolidation now and, and come out ahead because the values are, things are undervalued now.

287
00:47:58.618 --> 00:48:06.267
And I think when the market corrects a little bit, you're gonna see some appreciation. So some of it is, is, are these larger timing issues.

288
00:48:06.428 --> 00:48:23.628
But I think overall, you know, the, the decision that the big platforms made to not differentiate between different types of content and to favor creator content over publisher content because it's easier to negotiate, you know, you don't have to negotiate with creators, you don't have, you know, pressure- Oh, they love creators.

289
00:48:24.228 --> 00:48:28.888
Yeah, there's always- There's no unions, there's nothing. I mean, there's- Yes... they're [laughs] not gonna write bad stories about 'em.

290
00:48:29.128 --> 00:48:46.638
Like, they're- Yeah, so that made it harder for digital publishing, but I think it also was short-sighted by the platforms because, you know, Meta could have had a whole bunch of great news and a whole bunch of creator content and a whole bunch of professional entertainment, and been the ultimate bundle of interesting different types of content- Yeah...

291
00:48:46.638 --> 00:48:53.968
and paid different rates depending on the economics of those different industries. And instead we ended up with just creator SNARF and that's it.

292
00:48:54.028 --> 00:49:08.038
And, and I think that's, that's, in the long term, gonna be the undoing of the big platforms. Yeah. So you, you, you own a n- a, a news asset, HuffPost. Yep. And you had BuzzFeed News, but you closed that down.

293
00:49:08.068 --> 00:49:09.488
Is there like... I mean, you've been in the...

294
00:49:09.608 --> 00:49:26.832
I remember a- actually, actually, I think it was our last podcast, I asked you why, wh- why, why even have BuzzFeed News, and you said it was, you know, it, it-It gives, it gives heft to the brand, and, like, there's a lot of things beyond the sort of direct, okay, we're gonna monetize it directly, but it gives...

295
00:49:26.872 --> 00:49:36.732
You know, it, it ties things together, it gives you prominence, and I think that's, that's fair. But you do have to pay- But I, I, I took your advice and shut it down. Okay, yeah [laughs]. Well- So it's, it's all your...

296
00:49:36.742 --> 00:49:43.172
It's really all your fault... if anyone is listening to this, I did not. I, I did- It's really all your fault... that's not what I meant one, one bit whatsoever. [laughs] But look, news is a, is...

297
00:49:43.312 --> 00:49:55.732
I mean, look, as a business, it, it is... It's not, like, the CFO's dream of a business, I don't think, and a lot of that comes from, you know, the ad market and advertisers not wanting to be near news.

298
00:49:55.912 --> 00:50:05.352
And can it work as... 'Cause can it work as an ad-funded business? Like, I, I know you're, you're getting, you're getting, like...

299
00:50:05.412 --> 00:50:14.112
You get it lean, and you gotta get it, like, super lean, but then, like, how lean can you get before you can't have much impact? And that's, that's always the, the struggle.

300
00:50:14.492 --> 00:50:18.612
And yes, you can make the information work in various other, you know, really niche areas.

301
00:50:18.712 --> 00:50:31.712
But, you know, I think we had talked about this, is, like, everyone going behind paywalls is not necessarily great for democracy and our society, particularly at a time when there's tons of SNARF and misinformation out there.

302
00:50:32.312 --> 00:50:48.332
But I also struggle with, like, I don't really see necessarily a pathway where you can have a really ambitious news organization that is funded solely by, or, or, or mostly by advertising. Yeah. I think HuffPost...

303
00:50:48.372 --> 00:51:03.722
You know, when, when we looked at HuffPost compared to BuzzFeed News, you know, BuzzFeed News was getting most of its audience from Facebook and social, which was slowly going away, and HuffPost had a, a lot of front page audience, where people were coming every day to the front page- Mm-hmm...

304
00:51:03.792 --> 00:51:05.592
to see what was going on. And so when we were trying...

305
00:51:05.712 --> 00:51:17.172
Back then, we were trying to make a decision about how to move forward with news, it was clear that in, in a world where the platforms were pulling away from news, that HuffPost was a much stronger model.

306
00:51:17.212 --> 00:51:29.912
And HuffPost continues to be a strong model. I mean, it's not, you know, without its challenges, but the front page of HuffPost creates a, a, an opportunity to control our distribution.

307
00:51:30.212 --> 00:51:40.412
The programmatic advertising has been very strong there. The affiliate commerce is another area where the scale that we have can drive, you know, transactions.

308
00:51:41.032 --> 00:51:49.862
And so I think you can build a strong news brand, but it is, it is, you know, as you point out, not a cakewalk. Yeah. I mean, I just wonder with...

309
00:51:49.952 --> 00:52:03.162
when it comes to the advertising, 'cause I, I get it with commerce, but then it's like, okay, well, Google decides that, like, you know, the way you're doing, you know, commerce is, is not the way they want you to do commerce, and they basically shut your business down.

310
00:52:03.192 --> 00:52:12.812
I mean, they, they didn't like people running coupons, and then all of a sudden, hey, maybe Google's working on a coupons product. Weird coincidence. And, you know, when I...

311
00:52:12.892 --> 00:52:23.972
When it gets to the advertising side, I just want... I like... You know, there's a lot of groups out there, you know, trying to browbeat advertisers into, you know, supporting news, and I'm like, "Mm, I don't know."

312
00:52:24.032 --> 00:52:34.312
I mean, I like the, I like the idea of it, and that sounds, you- you know, right, but I just don't think that that's, like, a really great message in a hyper-capitalistic [laughs] system that we have.

313
00:52:34.372 --> 00:52:35.642
It's like, I think it's better...

314
00:52:35.672 --> 00:52:44.632
It's like, hey, you know, you gotta advertise here because, like, you're gonna sell more stuff, or you're gonna reach people you can't reach through platforms, so we're gonna give you something unique.

315
00:52:44.712 --> 00:52:54.712
I just don't know if- I mean, just- Go ahead... just so long as we have Jeff Bezos and Elon Musk and the biggest- Don't forget... the biggest companies and corporations in the world all supporting news, we'll be fine.

316
00:52:54.812 --> 00:53:01.752
[laughs] Yeah. I mean, maybe. But yeah, I just, I wonder whether the, the, the ad model can, can truly, can truly work.

317
00:53:01.772 --> 00:53:10.972
I mean, you have to just bolt on a lot of different, a lot of different, you know, ancillary businesses. I mean, the best news businesses are fronts for different businesses, I think, at the end of the day. Okay.

318
00:53:11.072 --> 00:53:25.052
So that was great. One final thing, like, I wanna... Oh, I wanted to ask you about websites. Is there a future for websites? I mean, if you look at another open protocol-based product, which is email- Yep...

319
00:53:25.072 --> 00:53:29.032
people have been trying to kill email for- That's true... a, a million years, and email persists.

320
00:53:29.212 --> 00:53:41.472
And part of the reason it persists is that it's possible to actually own your email address, and it's not like, you know, maybe you spend a lot of time on Slack, but then you move to another company, and, you know, people can't Slack you anymore.

321
00:53:41.512 --> 00:53:52.172
You know, they need your, your email. And, and so I think, I think open protocols are very resilient, and, and often they talk... There's conversations about how, oh, it's gonna die.

322
00:53:52.222 --> 00:53:54.612
It's the end of email, or it's the end of websites, or...

323
00:53:55.192 --> 00:54:04.412
And, and they're surprisingly resilient because they're p- it's possible for individual people to own them, and they're outside of sort of co- corporate or government ownership in the same way.

324
00:54:05.072 --> 00:54:25.072
And so I think websites, you know, we, we're seeing, you know, a lot of, like I was mentioning before, people coming to the front page of HuffPost to get their, to get their news, and, you know, people who are sick of SNARF and sick of the social platforms can go to a website and have a, a kind of beer and wine era- Yeah...

325
00:54:25.092 --> 00:54:28.832
of the internet enjoyment and surf the web a little. They could become vinyl. I don't know, but they could become...

326
00:54:28.992 --> 00:54:39.612
And vinyl came back, but, like, you know, I mean, when you look at agentic AI, it seems like a lot of stuff is, like, you know, obviously, like, websites are basically just a UI, like, on top of a database.

327
00:54:39.712 --> 00:54:52.232
Like, and the whole reason of reopening Three Mile Island is to be able to have some machine go out and do a lot of this work for us because friction must be eliminated.

328
00:54:52.812 --> 00:55:04.702
Listen, nothing is better than pulling a vinyl record, putting it on, pouring yourself a glass of wine- Yeah... and then surfing the web on your- [laughs]... go, on, on your laptop, going to the best websites out there.

329
00:55:05.252 --> 00:55:15.712
That's the life. I love it. All right. So I gotta ease into that. It's already, it's already 5:00 here. It's time for me to, like, surf the web. All right. Well, happy, happy- Open a bottle of wine... happy surfing.

330
00:55:15.952 --> 00:55:27.952
All right. Thank you, Jen. This was fun. Really appreciate it. All right. Thanks. Thanks, Brian. [outro music]