And I went back and looked at my own Medium account and I have two articles and each one is five years apart. I went back and read them and I thought they were really good.
The persona we have for you is a cicada, once every five years. But that's said with so much love because it means you're too busy living to be on the internet grind. Isn't that who you want to hear from? So my boss, the founder and CEO at the time, Ev, had this realization that the root of everything awful on the internet was the business model. As soon as you decide to have an ad-driven business model, you are incentivized to care about attention above all else.
It's great.
This week, we have Tony Stubblebine, the CEO of Medium. So recently you talked of course to Chris Best, the CEO of Substack and I'm just curious on how you see Substack and Medium's place in the world in relation to each other or not in relation to each other. Obviously Medium's been around a lot longer. It was founded in 2012. You know more of the history, so take it away. Substack and Medium are both platforms that are attempting to solve both the problem of giving substantive content online, getting away from tweet-length posts and putting into long-form content and addressing the issue of sustainable creation. How do you sustain authors? How do you sustain creators? How do you sustain journalism? And they have come at it from different directions. Substack is all the individual newsletter, the creator, the individual organization that empowers them as a platform. And Medium is more like a site that has sort of explored what is a collection of publication, what is the less frequent publisher, what is the way in which you can build stuff? And they're not quite competition, but they're so interesting because they're two models of how you can solve this problem. One of the things I really like as a consumer of Medium is that there's a small monthly fee that gives me access to everything. They obviously differ from Substack in other ways. I think content moderation is something that you've really honed in on.
Absolutely. Content moderation is something that's been in the news a lot with Substack about which publications they allow and who they're removing and who they're monetizing. But people don't realize that there's a really interesting model for content moderation that exists inside Medium. They don't realize that Medium isn't just a one-size-fits-all thing, but there are actually publications inside Medium that are similar to the subreddits. And those publications have people who act as moderators, act as editors, and they're deciding what content gets removed, what content gets highlighted. And so Medium handles it by empowering communities of people to do that moderation work rather than trying to provide a one-size-fits-all platform. And that's a really interesting model for how do we go forward and solve these complicated content moderation issues.
Well, in fact, you and I are editors of a publication on Medium called Revolution.social. So we invite our listeners to submit articles to that. So absolutely. Well, let's listen in.
Welcome to Revolution Social, where I am looking at the history of social media with the people who built it, understanding what's happened, and what we can do going forward. I am very excited to have Tony Stubblebine who is the CEO of Medium but who I first met at O'Reilly Media when he was an engineer there and we brought into Odeo as our engineering manager. Tony was my boss and was one of the most incredible, caring, engaged people to work for. And someone who cares a lot about humans, cares a lot about the people he is working with. And after leading the sort of engineering management of the pivot from Odeo to create Twitter, Tony left to work on his own startups. First a social media startup around conferences and connecting people, Crowdvine, and then looking at coaching. Took that and built a coaching business and then started writing about it and created Better Humans which was sort of one of the most successful publications on Medium and through that work ended up stepping up to become the CEO of Medium. I'm incredibly excited to have Tony here because not just is he running a platform Medium that is very important to social media but Tony has seen this whole history of where it's come from and how it's evolved. So welcome Tony.
Well, I mean, thank you for having me. I feel like, you know, when you work on a project and it doesn't work out, it's incredibly frustrating. And so, we had worked on podcasting in 2005. I was there only two or three months before we decided it wasn't going to work out and we started trying other things. So, I'm hopeful that this foray into podcasting that we're doing today is going to be finally we're going to do something in podcasting that works.
Well, you know, it is funny because I'm doing this podcast now 20 years after we were working on the first VC-backed commercial company in the podcasting space. And it's also interesting that podcasting is the one aspect of web 2.0, the one web 2.0 medium that stayed really open, stayed really diverse, and didn't end up with a company kind of dominating it in the same way that happened with so many other things. Photos and text like microblogging and video.
Yeah. Why do you think that is? Because for all creators, distribution's the hardest thing to get, and it would feel like every other media format has really been susceptible to, and I say this in a good way because this is essentially the Medium offer to authors: we have distribution to share, but essentially every other media format has ended up partnering with platforms in order to get distribution. And so, yeah, it is very notable that podcasting hasn't done that. And I already pay for Spotify. I would never notice if a podcast was Spotify only.
Yeah, that's fascinating. So, I think part of it is Apple's benign neglect. Like Apple came in and they added podcasting to iTunes. They added the directory. They created the podcasting app. And for us in Odeo as a teeny startup, it drove us out of the space. And then Apple didn't do much with it. They just left it as this sort of open thing and it wasn't very important to Apple.
Right. That's it. It's because it's outside of Apple's business model. That's where that benign neglect comes from and still is. It's like too small to matter probably. And you would never give up the Apple ecosystem just to be on some walled garden.
And we saw Spotify try and do it. But what we see now is like the reason we're doing a video podcast in part is because we upload the whole thing to YouTube and we then clip it to find the interesting moments and segments and bits that can go viral and then lead to people discovering the podcast and engaging in a deeper way.
Yeah. So, okay. I mean, right. I guess that probably happens everywhere. I mean, even I think about text, right, that it does show up in one place, but people are always trying to repurpose it. And I wonder if that'll actually get stronger in the era of AI. Because now you could write something and have it get turned into a video and get auto clipped for you. It's kind of we're entering a weird moment where the volume of content is going to go way up. It's not just more creators, but like that each creation is going to get amplified. I think I've said this publicly. I think Medium saw a 10x increase in the number of things posted after AI. Most of it was slop. I mean, so essentially our job got a little bit harder on the filtering side. Actually, a lot harder on the filtering side.
But yeah, this could be true. That's true on every platform right now.
I wonder if you could talk about that in a way because I know that Medium has gone through many iterations and there have been at one point paid editors and now there are publications on Medium that some people don't know about but function similar to subreddits where there's an editorial voice, there's people, there's curation and so you have algorithms to discover things but they are very driven by human curation. And I wonder if you could talk about the way in which that human curation and the people reading it deals with AI slop and deals with finding what is worth reading.
First of all, I love that this phrase AI slop has so much virality to it. I kind of think it was coined by a reporter writing about Medium first and then other places. And I was like I was really annoyed by the piece because she found the existence of AI slop but then didn't seem to get that it doesn't matter too much whether it exists. It matters whether it gets read and so I thought the whole piece was sort of nonsensical. It was a piece about the extra work we have to do on recommendation systems not about like oh Medium is getting flooded. Medium readers are not getting flooded. We and I think for this reason that you're getting at that we were kind of lucky we'd already built human systems and you know people like a lot of things in technology they sound new but they're actually old right and I think like AI content is sure it's AI generated but the issue is old it's just spam right or it's spam or it's low quality that's the actual issue so we already had systems in place to spot and filter and reject that stuff And it just was a matter of, oh, we're getting a lot more. And so when people think of it as a AI problem, they get really wrapped up in, you know, can you be 100% sure that you've detected whether this was AI generated? And it's like, well, that's not even the issue. The problem is that it sucks. And that's a problem we've been dealing with on like every social media platform since, you know, day one. And even though I think there will be a lot more AI generated content, we still think like writing has legs because humans are not going to stop being human. Like we're always going to be curious. I think people who write write first of all to examine their own thoughts, right? Can't have someone else do that for you right now. And I think they write to connect with other people. It's a way to connect with other people. So there's just a ton of pure human activity that doesn't get like AI does not replace that and because of that because of we have human curators we have publications which play a curatorial role essentially almost everything that ever gets recommended to you got read by a human that we trust beforehand right and they're just filtering out not all of it but there or maybe better to say filtering in like hey like this is a piece that matters right and before they say something matters they're judging is there a human story here which again goes back to this idea like we're often I'm like shaking my head we're often saying we're a humans first company which is like to me mind-blowing right like if we could just like back up and examine what the hell are we doing as an industry like we all grew up with like the movie Terminator is essentially mythology, right? We don't want the machines to replace us and then as soon as AI gets any good, people just get obsessed with replacing each other, right? And I just that's not happening. That's not what we're about. And at same like the idea that AI could write as well as a human is like this complete misunderstanding of what writing is, right? Like I think maybe naively some people think writing is about efficient communication and about facts and information but it's not. It's about a story and we already again what's new is old is we've already been through this idea that you could make stories more efficient with cliffnotes like before you know like we had summaries right that didn't the cliffnotes did not replace the book like you know your humanities professor did not say oh the Odyssey yeah the story is Odysseus had a hard time getting home end of story right like sure you could condense that book into one sentence and it would do nothing for you. And that's true for every story. And I you know, I think because I'd come from coaching, I'd come I'd seen a lot of the greater self-help industry. And so if you read a self-help book, I think it's really interesting how inefficient the information delivery is because a lot of times the whole first half of the book is going to be social proof like success stories. Oh, if you do this advice, your life will be so much better. Blah blah blah blah blah. And you're like, but what's the advice? Tell me what to do. Well, the advice is always really simple. But before you get the advice, you have to build up the confidence that it's going to work. And I think that's true at all sorts of like micro levels. like it's if you're getting technical advice, you want to know who's giving it to you because that gives you some faith that oh I'm in the same situation like you know and therefore it's worth my time to pursue this. So like everything that you're reading, it needs this human context around it. And unless the AI is out living a similar life to how you live, until we're to that point, it's just never going to have the right context and emotional valence required. And I mean, I'm not trying to even be like woo woo about this. I mean, this is like literally how our brains work. Like we are emotional creatures. like this is there's neuroscience about this.
When you talk about the way in which our brains work and how we're wired for stories and how even as the technology changes, we're the same people. It reminds me of a conversation we had 20 years ago during the hackathons that created Twitter. It was a month or so that we were doing these weekly hackathons. We had a conversation about the history of media and the way in which people are the same throughout like we evolved hundreds of thousands of years ago. And so we had a thing we did around the campfire. We had a thing we did at the coffee shop. We had a thing you did at the university when you were connected to the Unix computers and sharing it. And you then were talking to us about how do we do that thing with mobile phones that became Twitter and now it's how do we do that thing in an era where we have ubiquitous computing where we have all of these smartphones and now all of these large language models but we're still working on the same problems. There was a talk that my boss, who was our boss at Odeo, Ev Williams, he gave a talk at XOXO maybe the first one and the talk was about entrepreneurship and it was essentially make something that people already want to do easier. That's what we're trying to do. So yeah, I do maybe we'll talk about hype waves a little bit because I thought that was one of the big lessons of Odeo. But one of the things I've learned about a hype wave is a lot of times it's really confusing if it's being driven by technologists because they'll forget that we're still human, right? They'll think they'll get like what could the technology do and then there's often a product wave after that which is I think the AI wave just shifted from like really being driven by technologists to being more driven by product people the product people will think in human terms what is a human trying to do that this will help them achieve and I say this like this was such a harsh lesson from Odeo which I think and like I think every person from startups like every curious person wonders why do all of my projects fail, right? This is yeah it's hard not to spend a huge amount of time wondering that stuff. And so what I thought about Odeo was not just that Apple put us out of business, but that we were too early. Like even if Apple hadn't put us out of business, what were we going to do with podcasting in 2005? It just wasn't ready. But you wait 10 years later and podcasting is really taking off, right? And so that was a pretty harsh lesson that let me be patient through a whole set of waves. And I think one more thing about podcasting is I thought media was too late to blogging and so they felt burned and so when they saw podcasting they jumped on it too early as sort of a pendulum shift and so I'm just like always really careful when a hype wave comes up and so like I got to skip messaging systems like that was a whole there was a year where you could raise money just on like chat interfaces like way before AI. I skipped crypto which you know you could debate like people have made money. I still don't get the full utility of it.
Turns out you can make a lot of money inventing currencies.
And you know I think we mostly skipped the first wave of AI but I also knew what I was waiting for and you know like actually I've suddenly gotten a lot more interested in the last couple months. But it all comes from being burned on audio. What I remember from Odeo is like the two podcasts that I thought about the most were the super produced one from Adam who was the MTV VJ Adam Curry. Adam Curry, right? Super produced. Like this is a guy that like was deep in media. But then the one that I thought that was the most interesting was Venture Voice, which was just two venture capitalist investors essentially driving down 280 on the way to a meeting. And you know, the sound quality is terrible. The narrative arc terrible. The personality actually quite good because this is Dave Hornick, right? Like that guy has a lot of personality, maybe too much. But we heard things that we'd never heard before. And that's what I you know it's the sort of the idea of user generated content to me is the idea that my wisdom can be your wisdom and that can be done at scale. So anyone like everything that I want to do other people have done before me and I wish they could they had shared it and I could learn from them and that's I think what got me into blogging originally and that was the kind of the thing that I got most excited about in podcasting was not oh like here's a new distribution format for professional media but here's a way where we can hear people that you know you wouldn't normally have heard from. You know, I think that tension is always going is always happening. And I think podcasting to me feels like it's stayed more on the professional end. Like even people that even industry people tend to have a professional goal around it. Whereas writing because it's so much more accessible still tends to get a lot more amateur content creators who like this word amateur is tricky because it sounds like low quality but it's not because they are professionals at something else and that's the thing you would rather hear directly from the source. To me, it gets back to the really most powerful thing which is this is a profoundly democratizing transformation and that's not new in the history of media like you know the all the way back to the printing press it was a democratizing effort even if it was expensive to have a printing press and very expensive to print books each time we're doing this we are going through a wave of saying more people can participate and have a voice and now podcasting you know we have all these fancy microphones and things like that. We care about audio quality and editing, but it is still much more accessible than it was before and people who have a voice can engage and build in it. And Medium is a place where people can write.
And I went back and looked at my own Medium account and I have two articles and each one is five years apart. And I went back and I read them and I'm like, 'These are really good.' Like these were moments in, you know, one of them was an essay on Pepe Mujica winning the election in Uruguay and like a comparison of Uruguayan politics and American politics and how did Uruguayan politics produce this person? And I am a uniquely I'm one of the few people who could have produced that essay because I existed in these two worlds and on its own like I didn't want to create an entire newsletter. I didn't want it to be a professional activity. It was like I had this idea and I wanted to share it and I had this space to do it.
The persona that the name we have for your persona is a cicada like once every five years. But the thing is like that's said with so much love because what it means is you're too busy living to be on the internet grind. Like isn't that who you want to hear from? Like I think we really put a lot of effort into designing Medium so that a person that doesn't write that often could show up and get an audience that they wouldn't have to build an audience on their own. And it's for this reason because like if you have to write, if you have to manufacture something to write every single day, you run out of material. You start kind of making it up. And you see that in a lot of the punditry in the world. It's like they don't really believe what they're saying. They just believe it'll play, you know? And I just like much rather that people go live an interesting life and show up every now and then and share it. So that pattern of writing every five years is perfect. Actually, if we could get everyone to do that, we would have plenty of content, right?
The periodic writer or the writer that has something to say or someone else.
It's something that I think is very interesting in what you're doing with Medium because you're finding a very different way to deal with content discovery, audience creation, audience curation that isn't an algorithm that is driven by clicks. I know that Medium at some point has experimented with a financial model that was the more likes, the more claps, the more views that this particular article got, the more returned on it was. And it chased a BuzzFeed style like my goal is to get sort of maximum attention around this and I want to get exactly the right headline and I want to get everything else as opposed to satisfaction and overall quality of both the creator and the consumer of it. And I know that you've done a lot of work in Medium around changing what you're optimizing for. I was wondering if you could talk about that.
Yeah. And it's great because we've had a couple different eras. So we can really say what works and what doesn't work. So originally the main pivot essentially in 2017 for Medium is this experiment in business models. And this is before I was CEO but I was closely partnered with Medium at that point. And so my boss founder and CEO at the time Ev had this realization that the root of everything awful on the internet was the business model. As soon as you decide to have an ad-driven business model, you are incentivized to care about attention above all else. And it's like that in content that shows up in two really gross ways. Like one is that you're kind of unincentivized to write something great because if you think about it, it's like I get you onto the page. If you're too engrossed in the article, you're going to miss the ads, right? It's like I just has to be good enough for you to click on. It does no better or good enough for you to click on and scroll to the bottom but no better because it's all just a vehicle for the ads and then it really incentivizes the stuff that I think most people would say is the worst of the internet. Misinformation, division, hate, anger, it's just all of these things grab attention without offering any real substance back. But because it's grabbed your attention, you get to have a successful ad business. So Ev switched us to a single bundled subscription that any writer on Medium can participate in. And so this is an experiment. No one has ever done that with the UGC platform before, but he did it because he thought if we charged the customer directly, then our incentives would be aligned. And ever since it's been this Goldilocks quest to take this experiment and make it perfect, right? And so there's two experiments that predate me and then I'll tell you the one that I've liked. So the first is he's like, 'Uh, okay, now that I've done this, like it would stand to reason that people really want to pay for the highest quality. So why don't I on top of everything else go hire like a hundred of the top media professionals and start generating new journalism to put onto the platform' and he did this and the journalism was fantastic and it lost like let's say $50 million a year. And it's not just that it turned out to be a bad business. It's like I, as someone who was a partner at the time, I was like, what on I couldn't quite articulate it, but what I would say now is it betrayed the platform. All of the traffic is coming from amateur writers, right? Like the whole reason that you have this platform to launch professional writing into is because of all of this user-generated content. And so you're going to rob them of the traffic that would normally go to them in order to give them to professional media. Like why did you who you're paying at a loss at a loss, right? But it's like why have a blogging platform at all? Why not just start a media business? Why not start a new Huffington Post or something at that point? It's like if you're going to do it on a blogging platform, you can't rob the foundation in order to pay traditional media. That's one. And then also just as an entrepreneur, I feel like you only live once. The point of entrepreneurship is to make something that wouldn't otherwise exist in the world. And I was like I felt like that path was okay, we're going to do the Atlantic or like we're going, or like all the stuff that we were writing. Someone else for the most part probably would cover that somewhere. And you know I think the media industry is not that profitable but those people are at least a little bit profitable right so it just like it really didn't make any sense even though the people were great the execution was great the content was great but because of that we were able to build up a pretty big subscriber base and so that's where I can tell you that the next experiment which is what you were referencing did not work is then we switched to okay get rid of that experiment and instead just pay the community, but our what we paid on was attention. Even though like we built the business model to create an incentive for us not to care about that, our sort of everyone's pre-existing idea of how to do recommendation systems is very attention-based. And so it's like we're optimizing for clicks and reads and whatnot. And even as we got slightly better than that standard approach, it was still just attention. The people that were winning were like attention grabbers. And so there was like three playbooks essentially, none of which were on mission for us. So one was just get rich quick. Like Medium was swamped with get-rich-quick stuff because it's like, 'Oh, we're giving money away.' Oh shit, you know, like the whole like every hustler wants that, right?
Yeah. But which is exactly the dynamic we're seeing in Twitter today where you see all of these people who go and they collect a bunch of things and they make these tweet streams and they are like this is the you know it is super exciting and interesting and they don't give you all of the information. They give you the hook before the first tweet. So they get you the next one, and then in the end they're either trying to get the money out of Twitter for the drive.
There's this great phrase for that style of writing broetry, which I feel like I'm able to use it without ever defining it because people immediately are like, 'Yeah, that.' And it started on LinkedIn and then is definitely taking over Twitter right now. Twitter as we call it. I'm really resistant to
Yeah, I'm going to just dead name it.
Well, like Yeah, totally. If he's going to be transphobic, we get to deadname Twitter.
This is like I look I didn't have a problem like you know adopting like pronouns and whatnot, but like I'm I was on a meeting with some Meta people and I just could not stop referring to their company as Facebook. I'm just like I'm sorry like and same it's always going to be Twitter to me. until yeah people stop understanding what I mean.
