Yeah, Technology. What is it all

about? The Internet now I just see, is much more reflective of the real world. Then we we saw it or we imagined it 20 years ago when it was this very nascent, quirky little thing where it didn't have the problems of the real world because it didn't have all

the people of the world. That voice you just heard was EV Williams, billionaire cofounder of Twitter. I sat down with him last week to take this episode of Danny in the Valley, but our interview happened the day before. Twitter had some pretty big news, and I'm not talking about his decision to potentially up the limit for tweets from 142 280 characters. The company gave evidence in a closed door session of the Senate Intelligence Committee, which is looking into how Russian operatives use social media to influence the 2016 presidential election. Twitter said it has frozen 200 Russia linked accounts, But Mark Warner, the ranking Democrat on the committee, slammed the company, calling its presentation deeply disappointing. The quote, um Twitter showed an enormous lack of understanding about how serious this issue is the threat it poses to democratic institutions. That public dressing down perfectly tees up the conversation I had with Williams,

who's been working on this idea of online publishing, an idea sharing since the late nineties, when he launched his first company, Blogger, one of the first blogging platforms. Of course, he also started Twitter, where he is no longer involved in the day to day operations but remains on the board of directors and is its largest shareholder. His day job now is medium on online publisher that he launched five years ago to fix what he called the Broken Media. It was a mediums office just a few blocks from Twitter in downtown San Francisco, where I met with him last week. Way had a fascinating chat about where the Internet has gone wrong trolls, why it's so hard to turn a profit from content unless you're Facebook or Google, and how the future of journalism online may depend on a quirky idea. Clapping, yes,

digital applause. Anyhow, I will let him explain. So without further ado, here's f e. I'd be remiss if I didn't start with the obvious. These momentous times 280

characters Yeah, I'm stoked a long

time coming. I feel personally very excited about this.

That's good. A lot of people are upset, which I

understand, which is interesting. You have any sense of where that ISS I mean 140 characters always felt so limiting to me.

I think there's a natural sense that the brevity and the constraint is where the beauty is, which is certainly something that we always believed. There was a lot of power. There we did. The 148 character constraint was for the fairly arbitrary reason toe fit it into text messages. But then we were surprised that there was all this power to that constraint. And I think people associate the constraint with the power. Personally, I don't think 1 40 is a magic number. There's a lot of science and study that when that the team did that number and it's still fairly constrained, so obviously more expressive and more powerful.

One way to kind of frame all of this because you've been at this online publishing thing for a long time, longer than most since way back in 1999 longer. But I was just looking through the lens of Twitter, and then we can talk about what you're doing now at medium. I mean, there's idea that at the start of the Internet it was this great world changing thing that was gonna all kind of allows to share ideas, and it's done that. But it's also become this instrument to amplify the kind of darker side of things sitting here today. When you look at, say, Twitter, which has become like the megaphone of Donald Trump and lots of trolls, etcetera does that. Do you find it disconcerting?

The way I look at it now, as opposed to, say, 15 20 years ago, when we were starting down this journey and when when I say we, I mean those few of us who were excited about the potential of the Internet then and started building an enabling these ideas, it's just more complicated than we realized. There's a lot of emphasis on the negative side and the abuse and the misinformation, all kinds of things that are obviously not ideal. But it's important to have perspective and say that you know what? There's incredible, amazing things that that are still true. The positive side that we don't talk about because we take those for granted now. But it's not perfect. The Internet now I just see, is much more reflective of the real world. Then we we saw it or we imagined it 20 years ago when it was this very nascent, quirky little thing where it didn't have the problems of the real world because it didn't have a all the people of the world

and you're still a niche thing

was very niche. And when things they're small, they're manageable. And when there's bad actors that you know, they don't get the momentum when there's a person with terrible ideas that wouldn't be accepted in their community, they couldn't find the 1000 other people in the world who shared those ideas. And then they felt empowered by each other with the veil of anonymity, and that the networks were also less connected and lest immediate so there wasn't their potential for abuse and they didn't have the impact on what people read and believed in how they voted. And so now that everything is so intertwined, I think it's very obvious that while I still believe in the positive side of connecting people and enabling more people to share their thoughts. Oops. Also, there's all these problems. Let's

work on those problems. Did you ever envision that this could become such a kn instrument of presidential power?

It wasn't. Wasn't in our design, Doc.

That wasn't in your

decks. No, no, we weren't. Ah, on the other hand, I was working on Blogger in when with the little 2000 was a presidential election. But then 2002 is an election year. And that was when candidates started using blogging. It was a direct to the people medium. And then with 9 11 that was sort of the eruption of political blogging, the birth, many careers and tons of discourse and took that form from a what was seen his personal and or techie field to something more political, more mainstream Twitter, just one form of the same phenomenon of the the direct communication power of the Internet and all of the implications of that.

So before we get to medium, I think would be worth just cause a lot of our listeners are overseas to give ah kind of quick potted history of what you did before medium which he started 2012. So you started Blogger in 1999. Is that right? Yeah. After you dropped out

I university. Yeah, a few years after I dropped out of university and I kind of bounced around here in Silicon Valley. And then you started Blogger in 1999 worked on it for four years and then sold that to Google in 2003 and worked on Blogger there for a couple of years. And then I left there and I helped start coming called Rodeo, which was ill fated very early. Podcasting company. It was very early days of podcasting. And, you know, we imagined the future that we now have of this myriad options for new audio content. But none of it wasn't built in iTunes yet. Right? And so people still complain about the discovery of podcasts. But there was really nothing then, So we were trying to solve that problem as well as helping creators. A podcast.

So what we envisioned was not terribly unlike medium or lots of other content platforms, but for podcast, two things happened. One. It was really, really early. So is before the iPhone podcasts originally came from the iPod, And the idea was you would connect the iPod to your computer. You download some pod gas and then you'd walk around and listen to them. That was a lot of friction to actually get them onto the iPod E, although it was possible so part of Odio was there a software you installed on your computer that would connect iTunes And and so it was complicated. And then, actually, surprisingly early. Must our chagrin apple embrace podcasts? And that same year,

2005 they built podcast in the iTunes. They eliminated our major advantage that we were trying to solve, And so and it was still super early. No one was really listening to podcasts. Yeah, right. So that wasn't exactly a rocket ship, and that's when we started thinking about what else could we do? And we're thinking about new things, and that's where we started.

Twitter. The background of Twitter has been much written about lots of comings and goings, and who is CEO, etcetera? You're CEO for a while, Then you left and you started this in 2012. What's the big idea of medium?

So the big idea of medium is essentially very similar to the ideas of Blogger and Twitter. It was. It's similar in the sense that it's another platform for allowing people to put thoughts and ideas out into the world. There's a different motivation with blogger and then Twitter. The main problem we were trying to address was simply the Internet promised ability for everybody had to have a voice and put their thoughts out there. It was still hard, though, and so blogger made it easy to set up a website and update that website in the form of a blawg. Twitter made it an order of magnitude easier by eliminating the need to set anything up, eliminating the need to right much, you know, it was very is all about speed. At that point in 2012 after I left Twitter, I was thinking about the future and the state of the world. And one thing that was obvious was we had solved the problem of making and easy to put ideas out into the world. It was fairly momentous, and we take for granted again,

even back then, the idea if I ever thought I could put it out into the world if I have Internet access and all that. That's no easy, but it was clear at the time that that didn't make everything automatically better. We could build on top of that. And so the question with medium was Really how do we make things better?

I was reading before I came over here. When you set set it up, you had a very pithy and pointed way of saying kind of why you were doing this saying, Basically, it's is an alternative to what you called quote the highly optimized algorithms chum being slung by the truckload by low cost content purveyors.

Yeah, that's Ah, that's one way to put it. It was clear to me before anyone was talking about fake news and misinformation that the systems that were driving the creation and distribution of content online we're having a detrimental effect on the quality of the information we're consuming, at least on the commercials. Had I look at problems as systems and what, whether the incentives and feedback loops that are driving behavior, what the Internet had optimized for in and taught people who are putting the information out there, especially if they're trying to make money. But even if they were just trying to have influence and get attention, was that quantity trumped quality and attention was the goal and attention was rewarded in all attention was equal. It doesn't matter if it was You're making someone angry or or urine enlightening someone right. Attention was rewarded and you got more of it if you figured out how to do that in a high quantity and low cost.

Which is why cat videos do better than a 10,000 word investigative piece on Subject X, usually

at least for the effort. And so that seemed like that really undermined the promise of the Internet that it makes us all smarter and more enlightened and increases the understanding. And it didn't seem necessary. If we build better systems and change the incentives, we could change the direction of that trend. That was the motivation to start medium.

So initially it was just a platform kind of led. 1000 voices bloom, and then you tried to start with an ad base model or subscription and ads, or because you've had a couple pivot sense in terms of your trying to cut

a queen. We haven't We've had one pivot a business model. We've always been trying to do the same thing, which is we build a platform where anyone can publish as individuals, organizations for commercial and, um, commercial purposes. And the the goal has been in. What we've succeeded in doing is building a network. So if you publish on medium, unlike a standalone website, you can tap into readers on medium and you can build a following and you can reach an audience and is that network that is really the value. And this is where it's like Twitter and and other things where you can build a network and build a following. The difference is you can do it with more substantive contents, not social media. That was the initial plan. Then, as we started to move into professional content, we first were working with brands to do native content and and sponsored content in order to fund the creation of professional content. And that's what we decided. That's what we were doing a

year going on. Big brands like the ringer, et cetera. Well,

there are two types of rounds were bringing on commercial publishers like Bill Simmons, the ringer, and then we were we were working with them, as well as big brands like BMW and Intel on Samsung and G to do out advertorial do kind of, yes sponsored content, native native contents. It was never traditional Web ads, but it was brand based publishing. At the beginning of this year, in 2017 we decided that wasn't the best past. So we stopped doing. And we in March launched ah, consumer subscription that,

yes, I think it was in January. You had to lay off some people and close a couple offices in D. C and New York. Marchi Launch subscription March

We launched the beta of subscription, which we opened up in May. So it's been four months

since we really had five bucks a month. Yeah, and how many subscribers do you

have? That's not information we're sharing right now. How's it going? It's going great. It's growing growing every day, I think in the growth is accelerating. What we've launched about a month ago is what we call a medium partner program. The basic mechanism of the subscription, not unlike The New York Times or New Yorker, many other publications is a is a pay wall. So some content on medium is behind that Paywall. The majority is not.

But it is the garden of kind of cool stuff you get to.

There's a garden of cool stuff. I like that description of your Yeah, the vast majority of content of medium is still free, but so we're sort of transitioning into this world where more of it is behind the pay wall. And so our latest step in that was to enable writers in and publications to self publish behind our pay wall, right? And then we will pay them based on the engagement of their content. And the theory behind that is, I'm very confident that people will pay for high quality content. I think what's going to be hard is for every publisher to build their own subscription, let alone every author. So this is a way for people who want to put in the work to do this type of content, to actually get paid with the economics of a subscription. But essentially we all work together to get that subscription to critical

mass, right? And so you've been working on this for almost two decades? Do you think it's possible for the publisher of what's called quality content to have a viable business that's solely ad based?

Uh, no, not really. I hesitate to say it's not possible. There are people who do it now. He'll probably do it for a long time. In the world of publishing, I don't think there are any absolutes, especially for very niche publishers or for, you know, there's a lot of factors that play into it. Pretty much any model that you can think of will work for for someone because publishing its diverse says the food industry. But the general trends will be that that the market will buy for Kate, and there will be a free content focused on mass distribution. Advertising powered, low cost production, which generally but not always means low quality.

But there's certain type of content. You can produce a low cost and has mass appeal. And then there's higher production value, higher costs, content which I think will generally need to be paid for by a combination of consumer supported. I'm very confident there will be a market for high quality consumers supported content

as well, right? Do you think this is there's ah, parallel to be drawn between media or reading for lack of better words and the music industry and the transition it's gone through because it's only in the last year that's actually revenues have gone up after Napster, which came out in 1999.

Yes, absolutely. What's in particularly interesting music is that the content didn't really change. But convenience and packaging did in created incredible value proposition for consumers. For 10 Bucks Month access to everything, basically, and these these incredible tools for discovery that is way better than anything you could imagine before. It's just sort of a no brainer to pay for that, even though you could still go out and find music for free. What people say a lot about publishing is well, people are used to getting it for free, so you'll never convince them to pay. I just think that's untrue. People will pay for things, and they demonstrate that in all kinds of markets, media and otherwise, people pay for higher quality. People pay for convenience, people pay for identity. That's demonstrated across

the board, right? We talk about claps. Sure, so the system is on medium. So if I write something, you guys introduce the clap icon next to it. And depending on how many people click that button would determine how much I get paid. Is that right? Roughly. Because when this came out, I saw lots of the snark Army online came out Enforce. Yeah, was presumed having been a twitter so long you were not surprised.

Honestly, it caught me a bit off guard. The lack of even attempt to understand what we were doing. The need. Rick, this is a stupid idea. Yeah, yeah, yeah, I knew is unusual. I knew it wouldn't be obvious. The degree to which people still misunderstand it surprised me a little bit. I think certainly we could have done a better job of explaining it. But what we want to do is we want to measure value value, not according to us, but value according to the person consuming a piece of content right now on the Web, the way value is measured is pure attention.

How many people saw anything? Clicks, clicks and views that is value to an advertiser is not value to the people clicking and viewing. In the best possible scenario, you could read someone's brain and say after they consume something and if they looked at it and went away after three seconds because he realizes Clickbait or worthless, then it probably be no value or negative value. The Web doesn't register that If you saw an ad or a page views registered, it's as much value as if you read something for 10 minutes and change your view of the world. So that's obviously a terrible system and will not generate better stuff. A better system would be, if you could ask them. Was this good? How good was this? So was this good is actually a thing that we have on the Internet we've had for a few years. We have, like,

buttons and we we have hearts, and we have. We have these feedback mechanisms that say, Was this good on medium? The system We've had our equivalent of the like we call the recommend. Would you recommend that this is where someone's time? We've been studying this problem for a long time about what signals actually signify quality, and so the recommend is a good one, but in a sense it doesn't capture variability. So So what we've invented essentially, is the variable like,

because into one clap or 100 you can

do up to 50. There's actually a limit. And so the whole point is to say, How good was this to do that In a way that didn't make people think about it too much and was a natural sort of reaction? It correlates with how much someone appreciate something. It's a source of data that we didn't otherwise have.

Are you surprised how broken the media has become, or how much how thoroughly the Internet has broken the media

model? I try not to be overly negative on it because it's disappointing because so much is possible that we haven't yet done. There's no golden heyday were all newspaper articles were correct and you know it should be better by now I think we have and that was a theme even before we were focusing on the business model. That really motivated a lot of the the initial ideas and medium which is, you know, what, we really haven't changed that much and today, with with machine learning that there's so much more that's possible to make this all better. But it's held back by this business model on distribution problem,

and I guess that's leads to a kind of a different question, but related is that the medium? Right now it's five bucks a month and you get to the end of every month and you've just divvy up that pot. So that's temporary. Where you're just gonna give all that money to content producers? Yeah. Is that a viable business model? Long term to be it to actually. Then start, you know, taking your cut of that. Will there be enough left over for people to actually make money? We've

done the math on this, and we think it is. It's pretty good we're actually giving more than the $5 technically now, because we're seeding the system. But we feel good about that long term and subscription business. If you can get to critical Mass, it could be very healthy business. I think what the system will allow for is the very best stuff to rise to the top and be be rewarded as well as for us tow. Have a business on top that and it's not. It's not that different from other consumer media subscription businesses, which have built some very big companies and got in a lot of content creators paid very well.

You talk about stuff rising to the top, but if you're going back to the ad based model, if you look at, say, Facebook, that's an ad base model and it's a hugely powerful, but it's it kind of creates in perfect results. Talk about fake news and the kind of these filter bubbles. Why is it so hard to kind of make money? Is it the problem with the duopoly of Google and Facebook?

There's a distinction that is important, I think, in talking about ad base models in that there's ad driven content and there's ad driven distribution systems in the ladder. I would put Facebook, Google Twitter. The ads that drive those systems are not funding content. Their funding. Those systems, which provide all kinds of service, is for people. The problem is actually had driven content. So it's the most blatant fake news summits. Ideological lead driven. Ah, lot of it is profit driven. It wouldn't exist if people couldn't make money by getting someone to click on those headlines. But you could say their their complicity in their part of the same system.

But where relates is if you're a publisher a contra crater, and you choose to be driven by ads. Or if that's your model, you are competing with them because you're selling attention and they're selling attention, and they are not paying to create content. And they are have an incredibly efficient targeting and distribution method for attention. And so you're getting into a commodity business with the most massive lowest cost producers in the world. And so you're

probably in trouble speaking from a having been newspapers for 15 years, I can tell you firsthand, Yeah, I've been through many rounds of redundancies.

Another aspect that I don't think is super well understood is the programmatic buying, because ads alone could be a non commodity business, and they used to be more so so a London newspaper you have todo I want to buy these sophisticated readers who are based in London, and that's not a commodity that you're competing with people of the world on, and so you could charge a premium for accessing those people. It was natural to think that would apply online as well, So if you can build a a great brand B by doing high quality content and then you can sell advertising at a premium because it's associated with your brand, then you're in less of a commodity business, which is true, and a lot of people still do that programmatic buying is making that harder and harder. So people are buying less of the context that they're at is showing up in there, just buying the eyeballs that show up. And if

the automatic abs air effectively. But by algorithm, exactly. And so the big brands just say agency ex bias X amount of eyeballs. Yes, indiscriminate.

You're more likely You can buy him on any parameters that you want, including sites, but you're more likely to define the types of people that you want a target. Then you are going to say the place where I want to. So if you're some fancy magazine and you you used to sell those fancy magazine readers also use Facebook and Google, and they're probably much cheaper, and occasionally they visit some cheap website, and they're cheaper to buy over on those places and your your fancy magazine. Then you're selling the same product that they

are in that broader context. Just looking at Twitter for a second is do you think there's any hope that that what it will have become profitable? Twitter? Yeah.

Um, I, as a board member, can't comment on the future prospects

of Twitter. We'll throw you in jail. Yeah, Yeah, right. Um can talk about why you guys ended up not selling it earlier this year.

No. Didn't think then it didn't. It didn't seem like the best best outcome for

that company. Right? And justice going back to the kind of the broader context online and Twitter's place in it. And this idea of, you know, the Internet, it's an amplifier of everything. It's also amplifier of trolls. Is that something that worries you?

Yeah, I think the historically we underinvested in the defense mechanisms. But the company is is heavily invested in those mechanisms over the last 12 months, and I think they're substantial. Progress has been made. There's more to do. The whole industry's underestimated in those defense mechanisms, not entirely just about abuse, but misinformation. That's why the whole fake news thing happened.

You see, happening in the

Facebook exactly part of our ethos of Twitter was Twitter was about connecting people who didn't necessarily know people in that it wasn't a social. Know what people love about Twitter's connecting with people who they don't know? The architecture of the system, therefore, allows Maur interaction with strangers who may not be people you actually want to interact. And so the architecture of the system caused it to be Maur likely place for bad actors to thrive, but

especially with the possibility of being anonymous.

Yes, exactly anonymity. Again, while other systems may say, Well, that's that's a feature we don't wanna have. For Twitter, it's. It's been an important factor for many use cases, including people in oppressive regimes who are using Twitter as a way to coordinate and protests and all kinds of other things. It's a lot more complicated, I think, than other people. A lot of people, if they haven't thought about this deeply, very delicate line, right.

It's a very delicate line, and it's complicated, and it's we can definitely do a better job. The solutions are often, ah, lot thornier than people realize.

Yeah, because to your point around anonymity, it kind of it allows certain ideas and movements toe happen that wouldn't otherwise yet, um, to do you think anonymity on Twitter is safe

you mean All right. Is it gonna be,

in other words, like Facebook, It's very hard to actually set up a kind of a fake or earn an alias, if you will. Um, Twitter, it's pretty easy.

Yeah, um, I think Soudan Nimet e I I don't know. I don't think I'm not aware of any plans have changed the ability to be sued Anonymous, But it's slightly different than anonymity because, um, you can still build a reputation over time, and you still need to earn attention. So anonymous systems secret. And yet Jack and some of these other things that have grown very fastly have gone down in flames because they're not, um, anonymity has much bigger challenges. It also has as some of benefits. I think the s so far the cons of far outweigh the pros of truly anonymous systems.

Right? Right. And is it Do you have any sense of why? Um and I don't know how answerable this is, but why the big check? The big platforms Facebook, twitter, etcetera were kind of un prepared for that kind of the darker side and the the ability of those tools to actually just really amplify those voices, I mean, because having come from back from London, this is very distinct bubble from the rest of the world, but enough that has something to do with it.

But, um, I don't know either. I I think, um, I don't know either. Can't really say I think the that Yeah, yeah, I don't

know, Donna. Uh, but you think that that has the worm turned

the in terms of attention to the business? It absolutely Yeah, I think it's It's a vibrant conversation in Silicon Valley and it's tied with all kinds of other things, like awareness in general, of diversity issues and equality issues and abuse issues, both online and off is much, much higher, even in the last year or two then than ever. And I think that's a super positive thing, and it's gonna take a while to get there. All systemic issues that are not easy is solved, but I think the awareness is trending very positively in there. A lot of these things air un ignore a ble in the intensity of the conversation. It's growing much faster than the problem probably grew gradually and then suddenly, especially after the election, everyone's talking about it.

It does feel. I mean, I've only been here since January, but there does seem to be a real kind of like light bulbs were going on. Well, over half of

Silicon Valley is still a place where people come and are generally very idealistic. And I think more than other industries. And I think a lot of the world media and otherwise is saying a Silicon Valley. And like these guys with all this power and money, you need to be responsible actors and because of this money power. But compared to many other industries, I think the intentions and the ability are very, very high in my history in Silicon Valley, and 20 years of working on these systems haven't paid enough attention to many of the problematic aspects. But I do believe in the people and their their ability to solve problems. And so I'm

optimistic, so yeah, because it's it does seem like the kind of this tool was created and then kind of grew beyond everyone's imagination. And now there's a kind of oh, moment,

Yeah, yeah, but still combined. If there wasn't the optimism, we wouldn't have tried to do these things. These things are incredibly hard to do. It's not like, Oh, we just tried some things and this thing happened. I mean, we there's incredible thought in effort that goes into building any of these systems, and so I sort of look at it the same problems, the Internet as a whole. We we created these things. We thought lots of good things would happen. Lots of good things happened. Also,

some bad things happen. Let's look at those bad things and see how we can what we can do next. I mean, in the grand scheme of things, all of this stuff is very immature, and it's very early. Still, it's changing incredibly quickly. What was your worst day of work? My worst day of work. Uh, my worst day work was probably when I stepped

down from Twitter 2010. Yeah. Did you resign or be forced out, or what was the What was that so about?

I was asked to resign. Yeah, so I did.

Right. And at that time, Twitter was It's still private, But you guys were kind of getting ready to go public.

Yeah, it was private. For a couple of years after that. And, uh, Dick Oslo took over, as was my CEO took over CEO took a public and did a great job at that. But yeah, I was somewhat surprised by that move on

my board. But you and Jack er on good terms.

Yeah, we're good. I'm still on the board. They're doing

good work. Good. Do you think Facebook is good or bad? Influence up generally? No. Come. And that's it for another episode. I want to thank f for carving out a bit of time and sharing some of his highs and lows over the years. And it struck me after we spoke just how hard it is to make money online. If you think about it, Twitter's still hugely lost making. Odio was a bust medium. I'm gonna guess still deeply in the red. It makes you wonder where it's all going. Maybe Facebook and Google. They're just gonna eat the world. But we soldier on one podcast review at a time.

That's right. You know, I wasn't going to say goodbye without the usual plea. So please stop in an apple podcasts and give a review. It helps. And if you're not already subscribe, you can find us on Apple Android Tune in Spotify where, everywhere and of course, I'm also in The Sunday Times every weekend in the newspaper online at the times dot co dot UK and on Twitter at Danny Forts in Thanks again for listening, Doctor next week.