Jack Dorsey: We’re Losing our Free Will to Algorithms

So you don't need much of an introduction, but for a lot of people who don't know, you were interested in open source early on in your career. And of course, now you're known as co-founding and running very large companies. But you've been re-emphasizing open source a lot lately across your companies. So I guess first, focus on communications and how that kind of affects human rights. What do you think the current state of social media is? And what could make it better, I think, or more open in the world? Yeah, I mean, just to touch on the importance of open source, the only reason I'm here is because of open source. The only reason I'm here is for the grace of people who chose not to take a job or had a job and decided to give a bunch of time to something that they loved. And they made something that they loved, and they gave it away for free. And not only did they give the end product away, they gave the source code how to make it. So that anyone could copy their work. Anyone could monetize it if they wished. And it's just like this amazing selfless act that I was so inspired by. But it's the only reason I was able to learn how to program and build anything of importance and value. And I worked since I was about 13 or 14 years old, building my own tools and learning from, namely, Linux and Linux Tor Valids. And he taught the world what it means to share something of real note and real importance with Linux. Every single one of your devices in your pockets or at home, this TV, this screen, is running Linux, is running open source software. We all benefit from it, but it's very rare that companies and corporations give back and really push it. And this has been the story of the internet. It started as something that was truly open. Anyone could participate. Anyone could add something. And then as we discover more problems, very specific problems, they tend to be taken by a corporation and then closed off. And that closing of it forces this issue where you become dependent upon that company, even though the genesis of it was open. And it just feels, well, it's this constant pendulum swing, but it just feels wrong. It feels incorrect. And the biggest problem with this issue, in my view, is it creates a single point of failure. It creates a chokehold. You have a company, a CEO, a board that represents one place where an entity can put pressure on, whether that entity be a government, it be advertisers, it be a market, to bias them into a particular direction and to have them take particular actions, to remove accounts or to ask for phone numbers or email addresses or identifying information for people that are using the platform. Instead of what the internet wanted to be, which is everyone owns their own instance, everyone owns their own aspect and owns their own data. And because it is so decentralized, because it's so spread out across everything, it's very hard to attack because you have to attack the whole network to get any one particular part. So I think that the state of communication and social media in particular is we have these big companies that have single CEOs, which they could be the best people in the world and some of them are, but they represent single points of failure. They represent chokeholds that people can bias towards or institutions or governments can bias. And it's very, very hard to resist that, especially at scale, especially when your entire company's worth and its direction is at stake. When you were running Twitter, you had to deal with activist investors. Advertiser demands, country takedown demands for various types of content. And you've had to navigate that all. And you've been emphasizing how open sourcing some of that can help. And for example, you helped launch Blue Sky. Recently, it was announced that you had left the board. You've been donating to Nostr development and active on Nostr. What made you kind of go in that direction? And what do you think will make that successful compared to the hardships of making this work previously? Twitter was very unique. It didn't start as a company. It started as a Hack Week project and it was so basic and pure. It felt more like a protocol, like HTTP and SMTP that runs email or FTP, which is a file transfer protocol. It felt more like a protocol for public conversation than it did a company. And it always felt that way. And we had this API, which people could do anything with. And the reason we went down all the time in the early days was because like people were using the API nonstop and we saw all this abuse of it as well. Just nothing was rate limited. And we had shut down more and more of those endpoints because we kept going down because of the just free access. But I think ultimately that's what it wanted to be. But at the same time, we had to pay the bills. We had to become a company. We took investment. We took investors. We had a CEO after me. But the path, once we gave equity employees, there was only two ways to go out. Well, three, one to fail, the second to be acquired or the third to go public and chose to go public. And with going public, you need an advertising model or a business model. And we chose advertising because it worked for Facebook and Google. And suddenly you feel these pressures because you need to sell this product to advertisers. And the product in this case was actually the people using it. And when I came back to the company, a lot of this realization came to bear. One, we were selling people's attention. Two, I was a single point of failure. Three, we had a team that was tasked with making decisions around content and moderating it. And we didn't always have the best information to do so. Four, we had a bunch of governments around the world, Turkey, Russia, India, US, who was asking for information, asking for accounts to be taken down, where we had to push back under threat of being shut down in that country. So we developed a simple thing, which is a per country takedown, where we might take it down in the country, but it'd be available to the rest of the world. So you can still access within that country through a VPN. But even that was a pressure point that just did not seem right and did not sit right with me. And we realized the value of Twitter wasn't necessarily the protocol. It was the presentation and the business that we put on top of it. So we started diversifying away from just advertising business into commerce and subscription, not to the scale that we wanted to before I left. We also wanted to invest in a protocol that could be separate from Twitter, that we could build on top of, but we don't own and we don't control, and we can't truly take it down. So that we could have a presentation layer on top of it that shows certain things, but the protocol could show everything. And it'd be very, very hard to take anything down, for any entity to hide it or to take it down. And that's what Blue Sky was intending to be, was to build the protocol that Twitter could use, but other companies could use or individuals could use as well. And for people that haven't tried Nostr, especially in countries where Twitter's blocked or otherwise kind of has issues with moderation, Nostr is an open source protocol that different clients can and do build on. And then they can do a mobile app, they can do a desktop app. And what do you think, right now it's still fairly small, what do you think would help that model succeed? Or what do you think that model needs to succeed in order to grow large and be more impactful? Yeah, so we gave this funding, we gave $14 million to Blue Sky, which was a separate entity from Twitter, hired a lead to run it and build it. And after I left the company and it was acquired, I started searching for more projects like it. And I asked the question, what open source initiatives should I be funding that would be helpful to the public internet? And people kept tweeting at me that I should be looking at Nostr. And I found the GitHub that described it, and it's just like 100% what we wanted from Blue Sky, but it wasn't developed from the company, it wasn't funded from the company. It was completely independent. This paper diagnosed every single problem that we had and that I saw that we were trying to address with this protocol, but was doing it in a very grassroots and simple, dead simple way that felt like the early Twitter to me where any developer could get on and start playing with Twitter and actually feel it, like feel it immediately. And then I downloaded a client called Damas, and I created an account and it gives you this public key and secret key. And I could take the secret key and I could log into a completely different client or a different system and my identity and all my data went with me. And that was a super powerful feeling for me because you just don't have that experience right now. Your identity is owned by a company, you're giving it to a company. They own your data, they can do whatever they want with it. And for you to export it, to move it, they make it very difficult. Sometimes because they want to, but sometimes it's just really hard to put into a format that can be usable everywhere. But to be able to create content and to create and build your identity and then be able to move it around under your agency, I think is the most powerful idea. And I think the only reason people haven't really valued Noster yet is because they haven't had that feeling yet. It's there, it's powerful, it's available for everyone, but people don't have the need as much for it because it's not as convenient. But I think as we go on and on with all these corporate manifestations of social media and all the tools that we're using, people will have that desire more and they'll see the value in being able to own my data, own my identity, and really permission other clients into it instead of the other way around. We were talking backstage about the importance of algorithms. So it's not just about what we're allowed to say publicly, it's also these algorithms can effectively control what we think. They can show us some things, they can show us not other things, they can show us the order. They tend to kind of concentrate toward outrage and negativity. And with Noster, for example, you could have different clients with different algorithms or no algorithm, just rank order. Do you wanna talk a little bit about kind of the importance of open sourcing algorithms or choice of algorithms? Yeah, this is gonna sound a little bit crazy, but I think the free speech debate is a complete distraction right now. I think the real debate should be about free will. And we feel it right now because we are being programmed. We're being programmed based on what we say we're interested in. And we are told through these discovery mechanisms what is interesting. And as we engage and interact with this content, the algorithm continues to build more and more of this bias. But the algorithm, even if it's open source, is effectively a black box. You cannot predict 100% of the time how it's going to work, what it's going to show you. And it can be moved and changed at any time. And because people become so dependent upon it, it's actually changing and impacting the agency we have, the free agency we have. And I think the only answer to this is not to work harder at open sourcing algorithms or making them more explainable about what they're doing and why they're doing it, but to give people choice. Give people choice of what algorithm they want to use from a party that they trust. Give people choice to build their own algorithm that they can plug in on top of these networks and see what they want and they can shift them out. And give people choice to have really a marketplace around an algorithm that you can choose. I want to use this for these reasons. I don't trust that party anymore, so I'm not going to use this or I'm not going to use anything at all. I want to be the discovery mechanism. And that's really the biggest problem and why these corporations became so large and so valuable is because they solved the discovery problem on the internet. We talk a lot about the public square, but the public square cannot be owned by one company. The public square, by default, is the internet. But the problem with the public square is it's very hard to discover and to be matched with the things that you're truly interested in. And that's where the value of a Google came in. It helps you discover. That's where the value of a Facebook, it helps you discover your friends. The value of a Twitter helps you discover news and interesting content of the day. But if we can solve the discovery problem in an open source way, in a free agency way, that you get to choose how you see the world and what algorithms you're using and you know more or less how they're working and that you can turn them off and see everything, that's really powerful and that's what we need. And we just haven't seen a lot of motion there. Twitter took the first step some time ago when we enabled you to turn off the algorithm and just see who you're following. But the problem with that is you miss tons and tons of content because there's just millions and billions of tweets going by and you need some help. But to be able to trust the help, I think you need to be able to choose it and have agency over that. Otherwise, it really is attacking free will. It's programming how we think and we can resist it all we want. It knows us better than we know us because we tell it our preferences implicitly and explicitly all the time. It just feels super dangerous to continue to rely upon that without choice. On the topic of AI, which is pretty hyped up lately, open sourcing that has also been a topic of discussion versus regulatory control around safety regarding AI versus open sourcing it and letting it flourish that way. What do you think the development path is right now? Do you think that the open source model is going to win or do you think this will largely be a regulated activity in the future? I think open source always wins. I think the public will always win. But the same thing that happened to the internet and the discovery problem is happening with AI right now. You have something that was inherently based on sharing information, sharing research, sharing science. They're completely open and now being closed into five companies. And these five companies are building tools that we will all become entirely dependent upon. And because they're so complicated, we have no idea how to verify the correctness. We have no idea to verify how they work, what they're actually doing. And because we become so dependent on them, if you let just a few months or a few years go by, it becomes too late because those dependencies are so strong. So I think it really important that we have an open source alternative to these closed companies. And again, back to the point of these single points of failure and these chokeholds, that's the biggest problem. If we have open source, we have multiple people, millions of people around the world that can actually build these systems instead of being dependent upon a Sam Altman or a Nilan Mas or a Satya or anyone who's running these companies. It's their decisions that are guiding these tools that become the underlying fundamentals for all the experiences that we have on the internet and more and more off the internet as well. Because these systems are controlling every single aspect of our life. Every single day, someone will encounter some sort of intelligence that is interacting with them or dictating what they do or what they don't do with their day. And that's really, really scary when you realize that only five companies are building these tools and they're building them in a very closed way. Fortunately, there is an open source AI movement. As with any open source project, it's very slow, but it's very deliberate. And that deliberateness is very important. The slowness is important too because we take the time to learn what's truly important. And to build it in. And to make sure that we're pushing in the right way instead of just profit seeking and rent seeking, which is the incentive of a lot of these corporations. Well, we're out of time, but I would encourage everyone to take the first step and try some of this. And I think trying Nostra, for example, seeing how you can have more control and choice over the way you consume social media is a good step. Thank you, everyone. Thank you all.