Welcome to Revolution.social, a podcast about the history and future of social media. And here we are, episode one. And you were able to get Jack Dorsey, a very notoriously reclusive figure in tech, but a very important one, somebody who was there with you at the very beginning of Twitter, to sit down for a long intimate conversation. So, why did you want to start this podcast with Jack?
Jack is an incredibly important figure in shaping social media. He's incredibly controversial. And I wanted to have a conversation where we went from the creation of Twitter through its growth, the problems it had, what happened when Elon Musk took over and bought Twitter, and then a bit about the creation of Blue Sky. How did it come about? Where did Jack play a role? And why has he decided to focus on permissionless protocols like Noster? So, we get a whole history of web 2.0, of social media, where we came from and where we're going in this conversation. And then these are questions so many of us have. So, let's not waste any time. Let's listen to the interview.
Welcome, Jack. This is the first episode of my podcast, Revolution at Social. Jack Dorsey obviously is famous as the founding CEO of Twitter and Block and Square, but I want to dig into where we came from and what are the real dynamics that shape power in social media protocols and apps and why do we need to go from social media apps, social media platforms and into protocols. So, welcome.
Thank you. Yeah, just for the context of folks, Jack and I met back at Odeo, which was the podcasting company that eventually became Twitter. And now we're doing a podcast. It's kind of funny, 20 years after that, then we're like, 'Oh yeah, maybe I actually want to do one of these podcasts.' I remember in Odeo, we tried to convince NPR that they should do podcasting. And they were like, 'No, no, podcasting will never work. We've got our local radio affiliates. They'll work on it.' So much of what we've done, people say, 'Oh, that'll never work.' Or, 'It's a crazy idea.' We've adopted it. And then after it happens, it becomes sort of obvious.
Yeah. Always slightly too early. And the company's name was actually obvious.
That was the holding company for audio.
Not obvious at the time. None of it was obvious when we were working on it. No, that's the interesting thing. None of it was obvious that we were working on it. And you know, we'll get into the work we're doing today, but when we started talking about protocols as an alternative to centralized platforms, I remember I went in to talk to you when you were still CEO of Twitter, I think in 2018, talking about, 'Hey, we should do some protocol stuff.' And the Twitter executives were like, 'No, you shouldn't invest in it. Twitter shouldn't invest in it. Protocols undermine Twitter.' And it took a bunch of years for you to sort of build up the ability to try and move Twitter towards a protocol.
I think it was even earlier than that because Twitter was so like we used Rails. Initially it was C and Python, but then we got Florian and he was a Rails core member and we used Rails and Rails had this incentive to be API focused. And Twitter at the time in all of the web two cohort of apps like Flickr and all these other things was very much driven by the API first. And when we launched, we launched to a very small set of people and then they invited their friends and their friends and their friends. But eventually people discovered the API and the API was so simple that people kind of treated it like a protocol. That's what caused all of our downtime in the early days is we were just the database for these things and people were building all these clients. We got huge in Japan because the Japanese developers built a client for Twitter because our API was so open and it was so permissible in terms of what you could do with it to the point where it actually brought us down every single day. We had no conception of rate limits or anything that would protect the overall system from developers not necessarily doing bad things but just doing things that they wanted to do. And that's what was special about Twitter. You could come to it. You could even if you're a brand new engineer, brand new programmer, like you're just getting into the space, you could actually code something that would end up buzzing your pocket friend, which was just cool. Like you did something from a program that manifested in reality as a physical thing like buzzing a phone. And so there was this faction within the company that we should really invest more in that and being more of a protocol layer. And then there was another faction that no, we should close it down and be the business and remove the API because there'll be competitive dynamics and all this. And for me it took me a long time to realize this and I didn't really put it into words until I came back as CEO the second time but it's hard for something like that to be a company because you have corporate incentives when it wants to be a protocol. That was the pure expression of it from day one and it was never really allowed to be that because it was on this fast track to becoming a public company and then that set the requirement for a business model which was just a copy of what Facebook was doing with promoted tweets and ads. And then we were in this category of being the smallest competitor in these giants of Google and Facebook and basically you're at the whim of these advertisers and these brand advertisers and what they can do and it creates a single point of failure because they can just remove the money and your revenue goes down completely. So if it were an open protocol, if it were truly an open project, you could build a business on top of it and you could build a very healthy business.
And people were building businesses on top of Twitter in the early days and they were helping Twitter as well. They were helping Twitter Inc. as well. But unfortunately they were shut down on the path to IPO and I think that was the real miss. But it's you know all these are non-regrettable if you learn from them and you can build them into new energy and new direction and that's what I intend to do and what you intend to do and you know now there's a much broader community focused on that where the protocol doesn't have to be a corporation that doesn't need to own the entire stack it can own the presentation layer it can own the interface and build a phenomenal business. But the foundation, the protocol can be owned by everyone and be a lot freer as well.
Yeah, absolutely. I mean, I remember there was at one point this experiment to build the API that supported Jabber XMPP with the idea that you could run your own nodes and part of the thing and the tech just didn't work.
Yeah. Like what happened was the only way to make the tech work was this kind of centralized control and the amount of money that you needed to fund the building of it. And then everybody looks at you or they look at Twitter and they're like, 'Why doesn't Twitter make different decisions? Why doesn't Jack make different decisions?' And the thing that always surprised me is people didn't realize that just being the CEO didn't mean that you get to make all the decisions. You're accountable to the investors, to the board, to the market.
Yeah, we had that conversation early on. We were talking to tons of people about whether or not to ban Donald Trump early on in his presidency. And, you know, I am a far-left anarchist and I hate Donald Trump and everything else, but what I told you at the time was I don't think you can because you have a fiscal responsibility to the investors and to the market and there's private equity and you don't have the agency that everyone thought you had.
Yeah. Well, I think that move in particular, I think it was actually correct for the business at the time.
Yeah. Exactly. That's what I was saying was like we had advertisers who were threatening to leave and that would tank our revenue, tank our stock price and then people would leave the company because they'll find compensation elsewhere. But it was the wrong thing for the internet. It was the wrong decision for the internet. And we saw that play out when I forget the social network name that was a competitor but it had much more lax moderation policies and suddenly all Parlor and Gab and those...
Yeah. Yeah. So Parlor I think it was and all the internet foundation companies started taking away like AWS and all the things that actually serve this thing and then suddenly one by one I don't think they worked in concert but they looked at each other they're doing that they're doing that we should do that as well and then suddenly they had no place to go on the internet and their only option was serving themselves but even DNS providers were taking those actions and that's exactly wrong for the internet and it shows exactly why protocols are so important and it shows exactly why open internet standards are so important so that that can't happen. I became the single point of failure in that particular case because any company making these decisions is effectively taking away the potential and taking away the potential voice because of business needs because of business incentives and in our particular case it was advertisers, investors, and ultimately the market and our own employees. So it's just a difficult position to be in as a public company. It sort of works better as a private company, but even that doesn't work because you still have this single point of failure where one person is making the decisions for all the people using it. And it's seen more and more as a public utility, but it doesn't act like a public utility. It doesn't have the incentive of public utility. It's just super challenging. So the only path is to recognize that the protocol layer should be open. It should be a standard and there should be a standard for social media that's uncontrollable and permissionless in the same way that there is one for email and one for the web and one for all these services that we now depend on that people use. And that was the beautiful time of the internet like when we had Gopher and the web and Usenet and IRC. All these things were completely decentralized. They worked really well. But what didn't work was discovery and that's the thing that was centralized. That's what was centralized by Google. That's what was centralized by Facebook, by Twitter, by Slack. But you look at every one of those and it's just like a corporate evolution centralizing discovery on a very open internet protocol.
It's kind of enclosing the commons. Yes, we built this internet commons of these different kind of protocols, these activities of the social space and these companies discovered they could go in and control the discovery point, control that sort of single choke point as sort of Cory Doctorow talks about chokepoint capitalism and then make a tremendous amount of money there. But also exclude you from doing things that don't fit through their path.
Yeah. And it worked. But now we have the technology to decentralize discovery but at the same time that we now have this technology it's happening again with AI where we do have a centralizing force in the form of corporations with their incentives and their ability to change the models without people knowing about it and change the prompts without people knowing about it and inject things that make it look like the AI said it but actually it's a corporate incentive and it's all playing over and over again. So this is the importance to me of open source open protocols and then looking for ways and I think there are very tangible very compelling ways to build businesses on top where you really own the interface and the experience and you are valuable because you can execute the best.
You have the best design and you have the best experience and you have the best capabilities off this shared protocol and off this shared understanding. That innovation. You know those early days of Twitter, people thought of Twitter as an innovative company, but Twitter was a really innovative company in that it listened to its users. Hashtags and at replies and the at symbol for the username and trending topics and like everything we think of as Twitter was just you listening to your users.
Yeah. Yeah. Nothing about it was invented. It was all discovery. It was all just pulling the thread. Pulling the thread on like what people are actually doing with this and what they want to do with it. We launched something and we thought it was this and then they showed us no I want to use it this way and it's broken this way but if you fix it slightly then it works and the at symbol is an example of that like we launched it to be a way to share what you were doing in the world and then people started mentioning each other and the mentioning each other was like oh they want to talk they want to have conversations and just doing the simple thing of like anytime you see at symbol name, link it to the profile page, and then put a reply tab on the home screen and also on their profile page. And suddenly Twitter went from being status updates to conversation. And then you saw the hashtag and you link that and suddenly it went from being status and conversations to news and discovery about like what's happening in the world. And like again, that was never the intention. It was all discovered. It was all pulling the thread. And I think that every service has this. No matter what the founder says, it's not like this divine invention. It's like, I built this thing, people are using it in a weird way. I build it, I tweak it slightly for that, and it blows up. And like every single success I've seen has more or less followed that path, even though people won't necessarily admit to that.
Yeah, absolutely. I want to talk about sort of your first attempt at helping create a decentralized social protocol which is Blue Sky which you got funded out of Twitter and I was invited as part of the sort of the committee to talk about the internet protocols, evaluate the things and you ended up picking Jay as the CEO to go build it. And you know I'm on Blue Sky. I know that you aren't right now, but I think there's two things interesting. One, it felt like everybody talked about it as Jack Dorsey's Blue Sky and didn't give credit that it was a community project and it was a lot of people doing it and Jay made adjustments to make it work more like Twitter and make it less permissionless. So, I was wondering if you just want to talk about sort of the goals behind it and where it came from.
Yeah. About maybe a year or two years into me coming back as CEO of Twitter decided that we needed to have the protocol detached from the company. So, the goal was to build a protocol that everyone can contribute to. And there's a tweet thread that I put out because at the same time that I wanted to give up the protocol layer, we had to make it something that investors understood. And why would Twitter open up its protocol such that people can contribute into it and take off it as well and build competing interfaces. And the reason why is it's immensely valuable to Twitter if we have the most content in the world. And if we make it easy for people to inject the content and build interfaces on top of it to inject it and also benefit from that content, we have a higher probability of getting more content that we can place ads against or build commerce around or do a bunch of things that we had planned. So we put out this call for we want to build a protocol. We want to hire five people. We did this at my other company, Square and Block specifically for Bitcoin, we hired five open source engineers. We said we're not going to direct them at all. We're going to pay them in Bitcoin. We're not going to pay them in our stock. So, they're incentivized by Bitcoin itself. And it worked. And that organization is called Spiral. And we can't tell them what to do. They work on whatever they want. and they built the LDK, which is a lightning development kit, the Bitcoin development kit. They're up to 10 people right now. We just give them a yearly budget and they go. So, I wanted to replicate that. That had been running for a year. I wanted to replicate that for Twitter but for a social media protocol. We went about a different path than we did with Spiral, which is we created this community discussion on how best to approach this work and brought in a bunch of people such as yourself and even fiatjaf who created Noster to comment on it. And Jay organized a bunch of this work and did a bunch of the work and we eventually decided that she was the best candidate for the work. And the goal at that time and this is before the company was sold was it would again have a yearly budget. It would be this organization within Square like Spiral is actually a part of the company. And we were open to that with Twitter and Blue Sky. I actually called it Blue Sky because we wanted it to be like, you know, this open protocol. And we wanted to leave it up to Jay or whoever we chose like should we be in the company or should it be a separate entity?
I mean, I remember advocating for it being a separate entity.
Yeah. Because after leaving Odeo, I'd gone off and worked at Yahoo with Caterina Fake and Stewart Butterfield, some of the Flickr folks, and Flickr was sort of held back by the fact that it was inside Yahoo.
And so my hope was if Blue Sky was spun off as a funded but independent entity, it wouldn't face the sort of potential middle management frustration with seeing it as a competitor. Yeah, I think if Twitter wasn't sold that would have been the wrong decision. I think it should have stayed in the company just like Spiral has because we're not competing with it. Like it's something that we want to build on top of and it protects it more, but given it was sold, it was probably the right decision. But it also created a different perverse incentive of like, okay, well now the company is sold. We're probably not going to get more money from Twitter proper. So, we need to find money for ourselves and that means we need to get VCs and it's a public benefit corporation. And to me, that was not the intention at all. That's when I left. Because I don't want to be on another board. I don't want to help another company build this thing. Even if it is purely building a protocol, it still has incentives to a VC base and that's not it for me. Like this has to be more of a public good completely and have none of those incentives. And around that time I found fiatjaf's markdown on Noster and I just read it. I'm like he's diagnosed every single problem Twitter has. He's got a very simple effective solution. And it's not a corporation. It's not VC funded. I'm just going to grant him Bitcoin and let's see where this thing goes. And then I just became all in because I love the community. I love the energy around it and I love the principles more importantly and they were more tamperproof than what Blue Sky is now in a situation around. Like for me it's not even about the permissionless aspect of it. It's just the structure behind it is going to create incentives that I think are going to be challenging independent of how great Jay is. I think Jay is great. I think the team is great, but the structure is what I disagree with and I want to push the energy in a different direction which is more like Bitcoin which is completely open and not owned by anyone from a protocol layer and that's what I see in Noster as well. But that's where I want to push my energy is stuff like that rather into the more corporate direction even if it is a public benefit corporation.
Which by the way, all the AI companies right now.
Yeah. They're all public benefit corporations, but public benefit corporations are still, you know, you can take account for workers and the environment and society, but you still have stakeholders in a bottom line. It's still a for-profit company.
And you're determining what public benefit means.
Yeah. As an organization, you know, I very much agree. Like I like Jay, I like Paul, I like Brian and you know Y and a bunch of the other folks at Blue Sky. I want them to succeed. But when I looked at what they were doing and the work, I looked at the Noster protocol and the app protocol, they're very similar. But the Noster protocol I could participate in. I could run my part of it and I didn't have to ask for permission and it is developed like the open source bazaar. There's no company at the center of it. Whereas Blue Sky is very much driven as an open protocol but a company at the center of it. And I'm worried that the market incentives around it are going to push them to have to make some tough decisions. And it's part of...
They already have. Yeah, they already have because they've had to push more on moderation for instance and even bow to some governments in particular. So again it's already played out in the way that I feared given what it is and Noster is more tamperproof and protocols like that are more tamperproof because they don't have these single points of failure of organizational structure that control the protocol layer. And part of the reason I'm doing this podcast and I'm launching this social media bill of rights is so that we can give people a clear idea on how to hold Blue Sky accountable in a way that they want, but they can't just do it internally. We need a community on the outside. We need both alternative protocols that are viable alternatives that show that you can build this in this open permissionless way. We need clear ways in which we can say you know Blue Sky can and can't do these things going forward because users need accountability on their trust and safety system. You get to decide what your governance system is. You get to decide you own your name. You own your content. You own your relationships. You have the right to exit. You know adversarial interoperability like Cory Doctorow talks about like you have the right to choose your algorithms portability like all of these things. Blue Sky tries to do right now, but they could walk away from.
Yeah, I think they could walk away from them. They could also try to do them and do them well, but they may be compelled to walk away from them. So, like I just...
That's the worry. Not that they aren't well-meaning and trying to do the right thing.
Exactly. A lot of these CEOs and leaders and these structures are all well-meaning, but it's what the pressure points on them are. And that's what worries me. I would take the other side of that and I would say like it's just not going to be possible for these orgs to live up to the rights that you mentioned because they have these single points of failure and you can't get around them.
That brings up a really interesting point. A lot of people are angry at you for advocating for and publicly talking about Elon buying Twitter and what happened to it, but they don't understand that he wasn't the first private equity attempt at taking over Twitter. It was Elliott Capital and Silver Lake who had a seat on the board and were concerned about it. And...
