S15E09 The State of the Open Internet with Mallory Knodel
===

Charles (00:01) Hi everyone, I'm Charles Suggs, staff engineer at Smart Logic.

Emma Whamond (00:04) And I'm Emma a software developer at Smart we're your host for season 15, episode 9. We're joined by Mallory Knodel, executive director and founder of the Social Web Foundation. a special interest technologist working at the intersection of internet governance, human rights, open decentralized social platforms. Today we're discussing the state of the open why developers should care about governance, human rights, and who controls the digital spaces we all rely on. Hey.

Mallory Knodel (00:40) Hey, thanks for having me.

Charles (00:40) Mellon. Glad you could join us. Welcome to the podcast. so before we jump into the topic, ⁓ perhaps you could tell our listeners a little bit about yourself, your history, and ⁓ what you've been up to recently.

Mallory Knodel (00:53) Sure. So yeah, as you said, I've ⁓ I'm a public interest technologist. I have spent my whole career thinking about ways that we can leverage technology and science more broadly ⁓ for social movements in particular. ⁓ so I've been an activist also my whole life, prior prior to my even having a career. ⁓ and so it's been a real I think it's it's been a really interesting journey to see how Technology has served social movements over the years. it's I think I started my career really understanding that if I could just help social movement organizations build great websites and teach them how to use social media, that I would be really serving the greater good. And of course that's still true, right? We need ⁓ technology and the internet for communications. ⁓ but you know, as it As it turns out, our solidarity movements, our global solidarity movements around the world increasingly are pushing back on ⁓ corporatized, centralized tech use or the use of tech to ⁓ help authoritarian governments and to push you know these agendas. So yeah, I got deeply involved in that. And so it just evolved, right? It evolved from me ⁓ being a systems administrator to developing websites to working on the programmatic side of ⁓ nonprofit organizations that are trying to leverage tech for ⁓ women's rights and you know decolonization and have worked as well on advocacy in the United States because the US plays like a really particular role in you know global everything but in particular tech. And today I guess just zooming forward a a bunch of years ⁓ I spend a lot of time in global standards bodies where I feel that the impact of changing the way that low-layer protocols work in you know in service of ⁓ human rights can be really impactful. And just trying to narrow, because that's a huge field, ⁓ specifically working on ⁓ decentralized social media, ⁓ encrypted messaging, and things like ⁓ cybersecurity as it relates to, you know, child safety or more broadly the security of of social movements and activists and journalists. So that's kind of a in a nutshell, it's as brief as I can be when I talk about my work.

Charles (03:33) Would you say that that kind of explains what it means to be a a public interest technologist? or is there more that you you think is useful to know, especially to developers who are building software?

Mallory Knodel (03:45) You know, I think it's a really wide field, you know, term that can encompass so many different modes. I think I'm just one example and I've just had different I've I feel like I've had to both broaden and narrow my focus at different points just depending on the state of the world. But no, I mean I think I've seen folks who are who would consider themselves public interest technologists, ⁓ you know, because they're they're building you know, they're building software, they're right, they're creating ⁓ new apps or new services ⁓ for activists journalists, ⁓ et cetera. So yeah, I think there's just so many different ways to be public interest technologist. academics often are are doing a lot of research that may s maybe doesn't necessarily produce anything, but really produces a lot of knowledge that we need to understand the world, to understand the technical world that we're living in. So yeah, I just feel like it's a it's one of those ⁓ terms that you can apply to yourself if it feels like it fits, if you feel like the work you're doing is in the public interest, ⁓ then then that's what you are. so yeah, it's like being a radical. I we used to say the term we used before this was radical techie, right? Somebody with radical politics who knows a thing or two about tech. ⁓ so you know, the the terms evolve.

Emma Whamond (05:03) So as a dev, we build and deploy the internet to the internet. We might not always consider the governance behind it. ⁓ kind of coming into the public interest technologist, evolving that. Can you explain the concept of internet governance to those who may be unfamiliar?

Mallory Knodel (05:23) Yeah, this is something I love to talk about to audiences of all kinds. You know, I can riff about this with my next door neighbor for an endless amount of time. But you know, I think this idea of the internet being so decentralized and so distributed that there's really no one place you can go to file a complaint or, you know, ⁓ make a change to a protocol. It's just so distributed, ⁓ you know, physically, right? And then in terms of the companies and the organizations that steward it. but it's also really distributed in terms of how it is, ⁓ how the how the technology evolves, right? So I mean there is really one thing that is ⁓ that is true across the entire internet and that is the one protocol called Internet Protocol. It's you know got a pretty memorable name, right? And if you're if you are speaking Internet protocol, you're effectively on the internet, right? ⁓ But other than you the the Internet Engineering Task Force is the standards body that ⁓ standardized for the first time the Internet Protocol. ⁓ this is decades ago, right? there's so much else beyond the IETF. Although that's an example of a standards body that I go to regularly. They meet three times a year. They're still standardizing all kinds of things at at all layers of the internet for the most part. But then there's all these other mechanisms that are also kind of like bottlenecks, like the Well, maybe bottleneck sounds a little pejorative, but I I you can also think of them as like fronts of struggle for ⁓ you know, the equities involved in how the internet works. One is the domain name system, just the simple transference of a of a name to a number. You know, the website wikipedia.org to the IP address where that website is hosted. ⁓ that's a that's a send that's a system, it's a technical layer that's also then buttressed by a whole lot of governance. So how do those names get distributed? How do the numbers get distributed? How ⁓ do all those d equities play out as different kinds of ch you know changes are made to the domain name system. So I'm kind of giving a couple different examples of how internet governance works in practice. But then I think there's this really big world out there of internet governance Starting from like 2003 to 2005 when the United Nations started getting involved in the Internet. It took them a while. Like if you're tracking from when internet started to when the UN started getting involved, like that's kind of a big gap, right? So they finally started thinking like let's do something at a high level to really address or coordinate around ⁓ the dissemination, the extension of global networking. ⁓ And that's when the term internet governance really took off is this idea that it's very much a distributed governance system, just like the internet is itself a distributed system. No one entity, not even the United Nations themselves, really control it. ⁓ but it's more like what can you do as your institution in your particular role in making this global communication system work? What can you contribute to that? ⁓ and not like everybody you know who gets to control it, right? Like there's no more really ⁓ I guess wars or fights over, you know, who gets to control what. It's more like, yeah, how do we pitch in? How do we make this better? ⁓ and but there have been fights over the years, right? Like the International Telecommunication Union is one of the oldest UN bodies. It's probably the oldest standards organization as well. And they wanted to do like the naming and numbering that Now the DNS, which I mentioned before, is done by ⁓ an institution called the Internet Corporation for assigned names and numbers. ⁓ ITU really wanted to do that, right? They felt like they had telephony, they had telephone numbers, they they organized the telephone system. Why not do ⁓ the internet as well? ⁓ so there was like this sort of historical fight around like how the technology would be spec'd out. ⁓ would states be the ones to control it? Important things to know about the IETU is it's states only, which is what the UN system provides. ⁓ but ICANN isn't. ICANN is not states only, but it does include states. It includes also other organizations and other institutions and nonprofits and businesses. And anyway, ⁓ that's what we would call ICANN is an example of like multi-stakeholder internet governance because of all these different groups that get deeply embedded in it, whereas the UN is a little bit more like a multilateral institution. So there are different flavors of internet governance, but I'm gonna stop there because as you can tell, I think I'm really getting ⁓ ramped up and going on and on about it. But it's just such a wide and multi ⁓ modal way to engage in trying to figure out how to ⁓ make decisions on behalf of human rights or the public interest, ⁓ in changes to the technology, but also the governance of how internet and web actually work.

Charles (10:29) Yeah, it's a it's a pretty varied ⁓ topology of ⁓ different organizations and people. It do have you seen you s these standard bodies like IETF and W3C, but also governments, private companies that you're talking about and other kind of NGOs, non-governmental organizations all have some role in how the internet works and connects to various parts of it. ⁓ is that functioning in some kind of cooperation or is there some balance that's tilted somewhere? Is there a lot of tension within these organizations? How does that kind of play out over time?

Mallory Knodel (11:07) Yeah, this is a this is why I think it's so interesting to me is that it is kind of like a dramedy throughout, you know, the world. There it it varies, right? Like you can go really deep in one institution and kind of look at it and see its dynamics and understand where the imbalances are. I think it's hard to make generalizations though. I will say like some of the trends that I that have struck me over the years are

Charles (11:14) Yeah.

Mallory Knodel (11:31) You know, let's look at the UN system, for example, right? Like they're it's very difficult at the Security Council for Russia, China, the US to agree on anything, right? They can't get anything done because they're all enemies, or you know, they really they're really in opposition to each other most of the time. Those same com countries in a standards setting like the ITU or the IETF, they ⁓ or the or ICANN, right? They cooperate. They Managed to work together, they standardize technologies, you know, the major companies in China like Huawei, and then the major companies in the US and Canada like Cisco. I mean, they're all working together. The engineers know each other, they write specifications, they implement reference ⁓ implementations, they, you know, they they work on the spec, they get it out there, they build products. which is really Pretty staggering, right? So when the political system at the highest layers tends to fail, the sort of technical cooperation stays intact. There are certainly places where this does not work out, right? Like the internet infrastructure in like Palestine and Gaza, for example. ⁓ you know, for years the ITU has said things that support Palestinian internet and like connectivity in places and war zones. ⁓ But at the same time, like there's no way for them to politically recognize Palestine, ⁓ or or do really anything about it other than to maybe put some funding towards building the networks. You know, so as in like I think that the what I'm trying to illustrate in these two very vast ⁓ examples is like technology or technocracy alone is not enough in a lot of real material ways, but it is something. That I think still brings people together when they're just trying to solve problems and they're trying to solve ⁓ you know, interesting puzzles. ⁓ that there's a lot more cooperation than I think we realize. And that, but that can cut both ways, right? Like there can also be some ⁓ you know, too like too comforta we're too comfortable sometimes in this like purely technology-driven assignment, and we are not as good at bringing that politically forward. But that's not to say there aren't politics and standards. There certainly are. There's definitely actors who you know, and t when you when you asked about the power dynamic, I think the biggest imbalance, right, is that it is companies dominated. Even in like the ITU where it is states only that can be members, the delegations that the individual countries put together typically involve the ⁓ people from the companies of that country, right? So like France's delegation is gonna have a bunch of people from Orange Teleco or like, you know, just giving you an idea. So so by and large, like everywhere you look in internet governance, at least if it's a sort of technical mode, is going to be stacked with private companies. And so I feel like one of the important things about getting involved in these ⁓ you know, highly technical and sort of arcane spaces as a public interest technologist is trying to offset that balance a little bit with you know, representing nonprofits, representing human rights, 'cause you know, our voices are not as present. ⁓ and I think that if that we could tip that balance a little bit more, ⁓ I think the world would be a little bit of a different place.

Charles (15:07) envy anyone who has to go and sit through standards meetings at all. Mm-hmm.

Mallory Knodel (15:12) They're a little slow. I will say that, yeah.

Emma Whamond (15:18) So fascinating. I think this isn't spoken about enough, especially for developers and other people developing on the web. ⁓ A big theme of this season has been centralization. So, how did we get from the open protocol-based internet to a web where so much of our communication happens inside centralized platforms?

Mallory Knodel (15:44) I think there's a story that part of the story does happen in standards bodies. I do think that these are places where you know there's there's an obvious risk of anti-competitive behavior. So, you know, arcane spaces, ⁓ you do have to sort of pay to play in some ⁓ in some of them. They're all different in terms of their membership structure and participation or or what what have you, but you know, to be in a room with your competitors To talk about how you're gonna do things together sounds a lot like antitrust behavior to me, right? Or to anti- So I think that they they're very conscious of this on a legal level, right? They ⁓ have to keep their minutes open. ⁓ there is there have to be modes of of participation that are that don't lock out competitors. so there is some transparency mechanisms that you know are built in. And Those same mechanisms are kind of what helps people like me participate, right? Like they're a little bit more open. It's a bit frustrating, like the ITU operates like the UN and keeps a lot of their documents ⁓ behind you know, the they're they're inaccessible unless you're a state or unless you're part of the ITU formally. ⁓ so that that's a real barrier in that particular group. ⁓ so so. Even with best intentions, even with all the sort of legal infrastructure there to make sure that these companies are not colluding and price fixing and all of that, these things do tend to happen. I wouldn't say that it's necessarily the issue is cooperation between the different companies, but it does mean that ⁓ companies who become more powerful, they have bigger infrastructures, they impact more people, tend to be listened to a lot more. So, you know, if you arrive at the IETF or the W3C, we haven't talked enough about the W3C, so so let's do that. You know, the W3C World Wide Web Consortium is a membership organization. You apply to be members, you pay an annual fee. ⁓ if you just show up as an as a new person ⁓ and you have a specification and you want to put that forward as a document and get a lot of people to work on it, it's unlikely that folks will have the time or the interest in listening to you if you don't have a really large user base, if you don't already have kind of a reference implementation that's interoperable with another reference implementation and so on. So that's kind of like a baseline expectation. But then it gets sort of worse, right? Like coupled with the fact that, you know, if you are Google, you have a network like nobody else's network. You effectively have ⁓ you know, you're an you're on another planet. from everybody else who's running networks, right? So you can spot things, you can fill gaps that are new and novel that other people wouldn't have been able to spot or fill. You also are kind of like taking I've said this I in all seriousness, I think you're sometimes also taking a an easier road, right? Like you're so vertically integrated, you don't have a lot of really hard problems like the rest of us do with like only a few layers, you know, or whatever. So you know they they sometimes standardize things that benefit them. architecturally ⁓ because of where they sit. And so and also the things that they're standardizing or the things that they're reviewing that other people are trying to standardize come from a place of ⁓ power, right? They have a lot of ⁓ clout, they also have a lot of data, perspectives, trust, you know, etc. So these things over time are not necessarily nefarious or malicious, but they just amount to having a ton of influence in these bodies. ⁓ and I do think that that influence sometimes then translates into some political weight in different organizations. So it can mean that it's easier to kill or slow down work that your competitors are doing in a in a nuanced way that doesn't look like it's ⁓ an attack, but you know, these things absolutely do happen. And and in the case where, you know, standardization is so important for interoperability when you need like one protocol instead of two or in protocols. ⁓ It does feel sometimes like picking a winner ⁓ is what happens maybe on like a fine grained layer, like so only people who are really paying close attention to some of the changes will even notice that a winner is being picked. ⁓ but yeah, I think the story of centralization of the internet really comes from a lot of that tendency. and yeah, standards bodies definitely play a role. for sure. I also think that standards bodies, just one last element. Maybe there are more, but the one last one I can think of is that there's a lot, those are the places where a lot of deals maybe get done or relationships get built. Like I really, as an advocate, I really enjoy being able to go to standards bodies and talk to an engineer who can probably fix the problem much, much faster than anyone else in the company, right? It's just like a way faster way to advocate for what you need. And that's also true between companies. Like they can talk together, they can build relationships, they can get things fixed for each other, and they can also then go off and do other kinds of collaborations, go off and make other kinds of deals, you know, go off and in vertically integrate in new and novel ways, right? So that is also another way that centralization can be facilitated by ⁓ standards bodies.

Charles (21:11) How does this ⁓ kind of outsize influence that some of these companies have at these standards bodies and shaping the way these things are ⁓ formalized? How does that translate down to services that we might be building on top of? ⁓ whether that's you know email deliverability with a system that we're building, maybe we're not using Google or Microsoft for for email and we're using something else. ⁓ does it impact that? Does it tend to favor one service over another or does it keep the playing field pretty even?

Mallory Knodel (21:50) Yeah, I mean, I think that the goal is to try to keep it even. I do think the explicit aim is to just like, you know, come up with the best possible solution for all use cases. You know, what are the requirements? What are the constraints? Like, let's think of this from a an ideal perspective. At the end of the day, though, I really do think that these ⁓ what I've described in the previous response coupled with ⁓ the governance layer is what I think winds up making it not so. So here, like an example of what I mean, and we haven't talked about the standards body, and I'm excited to talk about it, which is ⁓ it's got the weirdest acronym of all time. It's called ⁓ the messaging, malware, and mobile anti-abuse working group. And the acronym is M3AAWG, and it's pronounced MOG, which is not so bad. So Getting from one to the other is important. Anyway, so MOG is a place they also meet an extraordinarily ⁓ multiple you know th three times a year. That's extraordinarily often. ⁓ it is about email and spam effectively, right? They are trying to solve the issue of spam. The email protocols are not enough. ⁓ even mobile telephony, like those protocols do nothing to to deal with the issue of spam. It's such a pervasive problem that they have this group that meets. Multiple times a year to come up with documents and specifications. Now, I think they make a good case for why they are completely nothing is public, everything is private. You have to be a member to get access to any of the documents or any of the meetings. And it's because, you know, if you're doing anti-malware specifications and you're somebody who deals in malware, like bo boy, wouldn't it be nice to know what kinds of ⁓ what kinds of ⁓ mechanisms they're coming up with. ⁓ it's a bit of a flimsy argument from a holistic cybersecurity perspective, but nonetheless, like we leave that aside. But what's happened over time is that you know effectively people won't really they don't really want to use ⁓ email services that are not part of the big few, right? That because the spam problem is so pervasive, because ⁓ they're choosing to use something like Google or Microsoft or Apple. Apple is a distant third, by the way. ⁓ That yeah, they they're that tends to centralize over time. And then I think what happens is they're, you know, in an effort to try to create protocol level signals that improve this for everyone. You don't just have to be part of this small group. You can go to the IETF, and the IETF is still innovating and extending ⁓ email and DNS and all that to to combat spam. ⁓ you have to then implement those protocols. So, like if you're a small email provider. You are ⁓ effectively at a disadvantage on all these levels, right? Like you will be actioned probably because somebody's account got hacked. ⁓ and then you're gray listed or you're block listed and getting back on is is really tough. So you have all that to deal with. And then you have these protocol changes that are happening at the standards body you've never heard of, and you need to implement them now because Gmail or Outlook or all of them have said if you don't implement this in 30 days like we're effectively not dealing with you anymore as like an email sender and receiver. So you have to then change, you know, your whole infrastructure to deal with like these new protocol signals. And so it can just feel like a lot. And I think those kinds of impacts, while yes, like they're meant to sort of evolve the space, they definitely disadvantage competitors. They definitely disadvantage people who are maybe running their own mail server because they do it for their friends and they're doing it effectively for free. and it's not necessarily a business, ⁓ that can just really like basically kick them off of the of the network. So hopefully that's like an illustration of how some of this works, but it also can be the same for I feel like podcasts went through a moment, right? Like podcasting is a surprisingly open system. and we can dive into, you know, why why that might be. But like ⁓ there have been attempts over time to kind of try to capture podcasts, to try to capture that space, and it's been ⁓ you know, an effort and trying to keep these things open so that, you know, anyone sort of in that network of providing podcasts and hosting them, you know, doesn't get yeah, doesn't get captured by like one winner who then tries to evolve the space in a way that you know builds a moat and makes them more successful than the rest.

Emma Whamond (26:32) So what happens to the users ⁓ like when the competitors are kicked off, maybe a single organization or maybe a couple few, Google, Microsoft, the big ones, take over the field. What do users lose when identity, audience, content, distribution are, they kind of become locked behind a platform, ⁓ and that platform takes over?

Mallory Knodel (26:58) Yeah. Well I think a lot of us have ⁓ you know, direct ex ⁓ experience with this sort of thing, right? ⁓ A platform makes a decision that we don't like, and you can post about it to your friends, and that's kind of your only recourse. Like you there's feels so and as you're posting it, as you're like typing out your rage, you can feel like the disempowerment happening because you know like you just can't you can't do anything about it. So that's one, right? Like we just lose agency ⁓ because there's no choice. We can't go, we can't go anywhere else. And so we really want to avoid that, right? Because that is like maybe the first issue, but then there's so many knock on effects that happen afterwards, right? We then kind of get locked into where we feel almost like exploited, like we're having to imbibe advertising or other kinds of you know, content being fed to us that we really don't want and we want to opt out of, but we, you know, again, we did we don't know how to do that. ⁓ so yeah, I think it's like the first of many escalating harms. And we're also ⁓ you know, I think that when we sort of get to content layer, and I think it sounds like you know we really we're we're moving in that direction, like away from these sort of lower layer protocols that just send bits and bytes back and forth. We're talking about email, we're talking about our private communications, we're talking about our social networks and how we communicate with our friends and ⁓ our cuss customers or ⁓ people who are interested in us on on a daily basis. ⁓ it it really becomes just this idea of ownership. Like we want to be able to own our our content or our personas or our networks, right? Like who follows us and who we follow. ⁓ and that just becomes really difficult when the forces are exactly against us, right? That the thing that we value, which is all of our friends, are actually kind of part of the business model of the platform that we chose to to build that network on. And and in a way, you know, sometimes companies will say this out loud. There have been a few instances in the media where it's almost like they they slipped and Said something they shouldn't have said, but it's like they consider that their property. They created that for you. You didn't create this, it not yours. This is something that that ⁓ you know we generated this value and it belongs to us. And so when they're in trouble in courts, or that's usually when it comes out, right? When in a court process, or you know, they're that that's you know, they they they will say that that you know it is something that you know, you might have posted this, like you know. two hundred and forty characters or something, but like everything else belongs to, you know, the company. That's I think a hard pill to swallow for a lot of us. And even if they even if they didn't believe that, right? Like let's take email, for example. Like ostensibly nobody's arguing that all of your emails are belong to the company, right? They're your emails. But it's really difficult to get them out, right? To get to get all your emails out and to put them in and on d on a different server, like is a huge, huge hurdle for a lot of people. And even if they're quote automated tools, right, there's all kinds of you know, dragons in that process, right? Like you think it's fine. You're like, okay, I've got the I've got the tarball. I'm ready. Like it just does not Yeah, exactly. Exactly. Or, God, some of it was encrypted with a key that I don't have anymore. Or you know, there's just like so many things that you it just is never it's never going to be easy because these platforms, like while they absolutely could make

Charles (30:23) None of the attachments or Yeah.

Mallory Knodel (30:39) migration tools and ⁓ these tools to to facilitate that, it is not in their best interest to do so. And so they have not made they've not got a whole team working on data portability, for example, because that's something for you, that's not something for them. That's not gonna make their revenue increase at all, right? A lot of questions, you know, in in corporate spaces are ⁓ you know, does your team make money? Well, none of those features would make that company any money. So they just don't get built. And that's some would say like the cynical view is that's what we would expect from, you know, big companies. Like that's how they operate. but I think we you know, there's lots and lots and lots of folks, I'm sure they're in your audience who make things for other reasons, right? And so obviously if you're making things for other reasons, you want people to have a good experience, you want them to use your stuff, ⁓ then you're prioritizing the user. But the problem is that you know th then those those tools or those tac those technologies or those services that value things other than profits don't have the same interoperability. They don't have you know, that goes back to all those ways they're disadvantaged in standards bodies, internet governance and so on. It's just like an uphill battle for them. So even though good ideas should prevail, that's just not always the case.

Charles (32:05) I think that's a great segue to talk about the Social Web Foundation. ⁓ can you tell us a little bit about the Social Web Foundation and the problems that you all are trying to solve?

Mallory Knodel (32:19) We are really focused on the open social web, like making social media a thing that is beyond the walled garden apps and services that we've all grown accustomed to, that we all kind of feel a bit trapped by. this is that, you know, this is you could download, you know, your social media cardball, right? And then actually take it somewhere, do something with it. And and maybe you won't ever have to do that again. Like some there's some notion of if we can really open up this space from a true interoperable protocol position, you'll never have to make a new account again. Like wouldn't that be so amazing? I I'm overwhelmed with the number of logins that I have to different platforms. And you know, but to to make that kind of it's a big bet to make a really big ecosystem shift like that. ⁓ so the Social Web Foundation is trying to steward that larger project, right? There's a lot of different folks building some cool stuff, and we just want to make sure everybody's kind of talking to each other. We're all kind of on the same page about things. We all feel like we're on the same side, right? We're all working towards the same goals. ⁓ but it just provides a little bit of that coordination. ⁓ so yeah, I can talk a little bit about how we think of the the the bigger change, right? years of nonprofits have me think in terms of the theory of change. ⁓ but you know if we want a really big, successful, joyful, fun, exciting social web that is not ⁓ a part of the traditional corporate webs or not not, I guess like walled up in in some of those, then we need to we need to have a few different objectives. One is there's got to be a protocol. Right, or a set of protocols that can be extended and evolved over time. So that's like one thing we work on. Both Evan and I are protocol people. He works a lot more in the W3C than I do these days, and I stay in my lane in the IETF. then we also work with the platform. So there's so many different platforms that are adopting this protocol called Activity Pub. You've also heard maybe of like AT Proto for Blue Sky and then DSNP for MiWe and a few other things. We're really focused on ActivityPub just because it seems like there are far more platforms using it right now and it's got a lot more decentralized weight behind it. ⁓ so the platforms that are implementing it, right? Like folks have heard of Mastodon, folks have heard maybe of ⁓ you know, Ghost or Automatics WordPress that uses it. anyway, there's more every day, and it's such a wonderful thing. So we try to work with the platforms then to Help them either implement the evolved protocol that we talked about in the first strategy or just kind of like interoperate together, know each other, get on the same page, meet each other at conferences, you know, whatever it is. So we work it with platforms, secondly. And then thirdly, like we're trying to really answer the question of, you know, what for, right? Like open web for what? And that I think is an important question to answer, especially if we want to work on human rights issues. We want to work with philanthropists to try to help them understand like how this is important for democracy, how this is important for, you know, ⁓ open information, et cetera. So the third piece is we really want to prioritize our time working with social movements, working with human rights organizations, ⁓ journalists, activists, and so on. So that's sort of the third thing that we do that when we when we are trying to answer that question of like, what are we building this for? Like for us, that's the answer. For others it might be arts and other, you know, Education, wonderful, wonderful things. We want to work with them too, but our priority is with the human rights community. and then yeah, we want to just create a strong institution for the long term. Like so, as any good, you know, theory of change for a nonprofit, you've got to think about the institution. So, you know, folks may have heard of like the well, the web foundation is a good example. We've talked about the W3C. There's a foundation that is that exists to preserve the web in the long term, and that's the Web Foundation. internet society is it plays a very similar role for for Internet. They've been around for a really long time and and steward that larger space, Internet for Internet's sake. And ⁓ so I think the Social Web Foundation tries to be that for the open social web. ⁓ you know, just really trying to be a strong institution in the long term as this evolves.

Charles (36:51) Something I've I've I've heard a number of people kind of in the last few years in the in the Elixir space for sure, but elsewhere, like, what is this Fediverse thing? I I don't understand. I wish someone could explain it to me in a way that it makes sense to me. ⁓ and so, you know, I'm hoping that you can provide provide us with an explanation of what the Fediverse is and and kind of what the difference is between building one better platform versus an ecosystem of interoperable platforms.

Mallory Knodel (37:23) Yeah. And I will say I every time I explain this I do it slightly differently. So let's hope it's one of the good days where it actually, you know, makes sense and lands. But you know, I I think we've all had this experience, right? Maybe I'll start with the one I I I talked a little bit about, my own personal feelings of like I just don't want to open another account again. Like I love TikTok. It's so great when people send me videos, it makes me happy. But I don't have an account because I just cannot. I just really I'm not gonna download download another app on my phone.

Charles (37:28) Ha ha ha.

Mallory Knodel (37:52) I I love, yeah, when people share things with me, but you know, I'm also pleased if they just like take a screenshot of ⁓ the the meme and like they they send that to me too, right? the having a Fediverse, having like this protocol, one or a few protocols underlie all of these different platforms would mean that I don't have to have a separate account. To engage with this. Like you can share it with me and then I can share it with somebody else and I can spread the love. It like, you know, all these emails die with me, or like mess you know, text messages say die with me, because I I don't have accounts on these platforms and so I can't, you know, like reshare or interact with it in any meaningful way. That would change, right? Like so I could have ⁓ you know, one account. Or maybe I have like an ultra ego. So I have two accounts, right? But largely, right, all of the content that exists on any other kind of social web, ⁓ I could interact with it, I could share it, I could like it, you know, there's just like a lot of things we can do. I also want to just talk about this for a second because the idea that, you know, I had my space, ⁓ I was an early adopter, my campus was an early adopter of Facebook back when it was like just for US colleges. This idea that we still can basically only do three things with any piece of content is like blows my mind. Like it has been twenty-five years, everyone, where you can like something, you can share it again, or you can make a comment on it. And like we have not evolved past that. It's like another example of like how these centralized platforms have dis disenfranchised us. Like there should be so many more ways that you can interact with with content beyond those those few things. but change has been very, very slow.

Charles (39:16) Yeah.

Mallory Knodel (39:36) So yeah, I mean I guess I what I want is ⁓ what the I think the the folks who are excited about the open social web is that you have a lot more agency because you have a lot more builders, you have a lot more ⁓ people trying to do cool stuff that helps present, create, consume content of all kinds within ⁓ many different conceptions of what a network can be. you know, you You don't have to necessarily f overthink the this is my professional life versus my personal life. I mean, those things of course you can do, but they don't have to be like physically separate and different services. another thing I really think would be great, we've already talked a little bit about is like the portability question. You know, stuff happens, right? Like companies go under, services expire, like something's no longer hip and cool anymore. And it's it sucks when it was like your thing and you invested a lot of time in it, right? Like it's hard to know what to do with all of that stuff that you put on a platform somewhere. You can download it and it basically just lives ⁓ on your hard drive, right? It would be so nice if that portability that a common protocol could unlock min means that you could take that and actually put it back up somewhere and and have your content live on. without the same context, but could still, you know, sort of approximate ⁓ your contribution ⁓ to the internet in some way. That's what I got. That was a long explanation, but I hope it helps like understand, you know, what's at stake. I think it's not just like, okay, download a new thing and log in, it's better than the rest. And then you do it, you're like, wow, this is not cool as or as shiny. And all of none of my friends are here. Cause I think that's sometimes what happens with the Fediverse for people. ⁓ but instead, like we're doing this slowly to build up you know, something that is much, much bigger. So I guess what I described, that sort of sad experience is like what we kind of tend to expect ⁓ when a new winner comes up or when there's like a new rockstar app, right? We expect to be excited like we did when there was TikTok, right? and that's just not what happens because I think right now we're trying to, you know, build the infrastructure that would allow something much bigger to to emerge.

Emma Whamond (42:00) I think that's great explanation. Yeah, that was a good one. ⁓ I read that your mission statement says that the Fediverse should be quote, growing, healthy, sustainable, and multipolar. I just wanted to talk about that word, ⁓ multipolar. It's doing a lot of work there. Why that specifically and why isn't decentralized enough?

Mallory Knodel (42:23) Right. that's a yeah, that's a good thing to dive into. I think it's because there's just a recognition that things are popular and that popularity can drive ⁓ more popularity, right? Like it it tends to be that we ⁓ as social networks huddle around certain kinds of things. So ⁓ I think the notion of multipolar is we recognize there are gonna be big services. And that's not a bad thing if they can interoperate with each other. And if they allow for things like you know, user agency to move around. ⁓ it's not trying to, you know, claim that your content is their content, you know, that sort of thing. So what I think folks were worried about in the beginning was that like some other open source protocols and things, that if you do get one big player who's an early adopter, that they can kind of eat the world a little bit, right? That they can kind of take over and then they dictate how that evolves. Folks in the Fediverse space are hyper vigilant against things like that. And so instead of saying we're we're we don't like big players, I think what we were trying to say with the multipolar idea is that we need multiple players and they and we want all of them to be successful and we want them. I think the idea of like choice amongst many means that there's a balance rather than ⁓ sort of this ⁓ resistance to something that almost feels like entropy, right? Like somebody's going to emerge. It would be bad if it were one. So like let's think about structuring this emergence so that many people win and that they are countervailing forces.

Emma Whamond (44:25) That's great. I like that. So along the same line of federation, could you describe for our audience at a high level what Activity Pub is, the protocol and how it fits into all of this?

Mallory Knodel (44:40) Yeah, so there are two things to know about ⁓ ActivityPub is ⁓ from a protocol layer, it's really two things. One is it's a data structure that's called Activity Streams. ⁓ and it just treats pieces of content ⁓ with an extensible framework. So there are some obvious ones, right? Like if I'm a if I'm a person versus so an account, right, versus like a post. So I Have a post, the post is an object, I am a user, right? And so it's all kind of like structuring all the data that's associated with that, right? ⁓ if I'm ⁓ a person, then I have ⁓ like a bio and I have a link that may, you know, verify that that I am associated with that link, and I follow people and they follow me, and et cetera, et cetera. And with a post, it has different features like. It's got a URL and it has a title, maybe. It might have ⁓ content. It might have been liked by someone else. It might have been reshared by someone else. Like it has all these different features associated with it. Activity streams ⁓ is a standard that codifies that. It can be extended. You can add all kinds of stuff to it if you want. You don't have to stick to what other people are using. Then there's Activity Pub, and that is the part that standardizes how. this content gets sent around and how users can interact with that content and how users can interact with each other. And this is the part where it's sort of standardizes like, okay, if I like something, then my server is contacting the server of the content to record that that like happened and that we have, you know, basically a mirrored image of that interaction. So it's more the protocol of like the activity, which is where Activity Pub comes from, right? ⁓ so I think that it fits in because it is really similar in ⁓ design to, you know, to the web, to this, it's not necessarily like you know, HTML and CSS for a website, right? Which kind of is a way of structuring that information. It's definitely using those protocols. It's building with things that already exist. but you can imagine, for example, activity streams being a really great way to ⁓ encapsulate other kinds of content off of social media. So like you could use activity streams to use as your like your instant messaging content structure, right? Instead of like XML, right? You could use activity streams and then you could ⁓ you know, maybe think about fitting ⁓ c messaging into social media. So direct messages is an example of that. Like we've all kind of come to expect that the social media applications that we're on, the platforms that we're on, they also have a messaging feature. ⁓ this also exists in the social web, right? Like ⁓ Mastodon and a few others have this idea of direct not all of them, but some of them do this notion of direct messaging. And in the sort of most basic way, they just treat the audience For a post that is a direct message ⁓ between two people. So the audience is not public, it's just like you and me, which is not exactly from a you know cybersecurity and encryption perspective, like I think an actual confidential conversation. So, you know, Activity Pub has been able to be integrated with ⁓ something called messaging layer security, which is another protocol, right, where we can treat treat those messages separately. So I'm just giving an example of like how you can literally fit, you know, different protocols together. ⁓ you know, so the activity streams content is enciphered. It's use it uses the activity pub protocol to send it between people on an activity pub server conversation. ⁓ and then right, it can be decrypted on either side. So like that's a way that it can just like fit in with other things. ⁓ But yeah, I don't know if that's helpful. I feel, yeah, like I'm sure there's a lot of different developers that listen to your podcast that are thinking about like, okay, how ⁓ you know, how does a content, how does content turn into an activity streams ⁓ object and and how does the activity pub server send it back and forth. So just trying to give some imagination there.

Emma Whamond (49:19) While doing research for this podcast, ⁓ this episode today, I learned that Threads, Meta's product, federates via ActivityPub. So it's the single biggest instance in the Fediverse by user count, and it's owned by Meta. Is that a win for openness or is it a kind of embrace, extend, extinguish risk? What do you think?

Mallory Knodel (49:45) Yes. I mean, yeah, you you're bringing up something that was worrying in the beginning, ⁓ and kind of why we have this piece about multipolar ⁓ in our mission statement and why the social web foundation was so critical in that time to to exist because folks were really worried about that ⁓ happening, that EEE. Right. so I think I my personal feeling, right, is that if this is an open protocol. We really cannot worry so much about how it gets used. I think what we need to worry about is the way that it gets extended and adopted more broadly. So, you know, things that have gone poorly in the past, ⁓ been like RSS, right, which was also, I think, you know, s maybe Google had something to do with that not going, ⁓ not not becoming as popular, staying as popular as it as it once was. ⁓ It it's important, I think, to focus on this governance aspect, to focus on like a stewardship that exists beyond ⁓ one one company, one player. ⁓ so that's sort of what the Social Web Foundation is trying to ⁓ provide. ⁓ at the same time, you know, like the other thing that we've talked about with respect to Meta ⁓ is you don't have to ⁓ federate with with threads. Like you can, as an instance operator, say, like we will not federate with this particular service or server. That's a choice that can be made. ⁓ the Fediverse, because it's an open ecosystem, because it's an open you know, protocol, it's it's also very much it's a it's an enthusiastic yes protocol, right? Like you have to follow someone to get their content. You have to interoperate with someone to be able for you the the users on either side to be able to talk to each other. ⁓ And if you don't want to, you don't have to, right? Like if I don't want to see anyone from threads, if I don't want anyone from threads to see me, like there are settings for that because again, like you have to sort of ⁓ opt in. So that's I think what buttresses is a little bit. But I do I not to say that there is no risk. I think there's an an absolute risk. I think we want another big player to sort of come and balance it out. I mean ideally like all of them, right? Like one mode is that decentralized social media becomes successful. Because all of the existing social media companies adopt it. Like that is one view of a successful future. And as long as you still get, you know, credible exit, portability, like all the things that we've been talking about that to be important, then we still win, right? so I I hope like we get more people joining the Fediverse at that size, not not fewer. and with the sort of infrastructures we have in place, the networks we have in place to talk amongst protocol developers and platforms, I feel confident that we can support those kinds of big players in the sense that we will not let them capture the space.

Charles (52:53) If you can't see content from someone before you if you haven't followed them, how do you discover people you don't already know? How do you discover information that's not in your your bubble in in a situation like this?

Mallory Knodel (53:11) Yeah, this discover the discoverability problem is a real one, for sure. It's if I could make a comparison just to the web is also had this problem, right? ⁓ it was not obvious how to find things at first and it was actually something like the internet was not the web in particular was not really popular right out the gate, right, because of this problem because you didn't even know what you didn't know. You didn't know what was out there until the mosaic browser and like then eventually, you know, Tim Berners-Lee and then other people started to figure out this problem of like, okay, you need to like index things. You need to like, here's a website of websites. Go to the website of websites and find some cool stuff. Web ring, yes, and you know, Google search. You know, we do come back around to these ideas of like, wow, and now, you know, there's a strong case to be made that Google search should just be like a public

Charles (53:54) Webring.

Mallory Knodel (54:09) Good because it you know effectively does something nobody else can do. ⁓ So, you know, we need to find solutions that are cross-cutting. ⁓ we need to think of this as not like different apps, not only that different apps can build on this, but like also intermediary cross cutting functions can be built. you know, another one that I think is a really hard problem ⁓ is ⁓ trust and safety, right? If you're a small platform provider, the idea that you ⁓ now need to like spin up not just like spam filters but like all kinds of other kinds of you know vigilance around certain content and so on is like a really heavy lift if you're if you're a small provider. It would be a lot more helpful if we had tools and intermediary services that can help with that and and things like that. So I think discoverability is just one of these hard problems we need to think about from an ecosystem level. And You know, again, something that I think Social Web Foundation or others kind of in this inter you know intermediary space like can start really thinking about and building. So we've tried. So tags tags.pub is a project you can look up. We talk about it on our website. ⁓ we're trying to think about the discoverability problem through hashtags and making it so that hashtags can be a discoverability vector across instances rather than per instance. Like I can use hash ch hashtags on like my instance, but that's obviously not the idea we want. I use it across across all instances. So you know, small small problems hopefully lead to bigger solutions is how we're starting out.

Charles (55:51) We have so many more questions we want to ask, but we are ⁓ coming up on time, unfortunately. ⁓ but we're trying to get a a a couple more in here. ⁓ I there's a couple of topics that I wanna hit on briefly, if we can, before we do wrap. One is about kind of ⁓ how the design of things affects society, but also getting a little bit into ⁓ Excuse me, AI getting onto social web and and and bots on the ⁓ on Activity Pub and and kind of does that magnify it? So let's let's let's start with ⁓ you know is there a relationship between the development of new technologies or the underpinning infrastructure and existing inequalities in our societies?

Mallory Knodel (56:24) Yeah. I would say enthusiastically, yes, unfortunately. ⁓ but you know, this is one of the reasons why we have engaged in standards bodies for so long, is just this idea that you know, f low level decisions can have big impacts, particularly on things like privacy, ⁓ free expression. But you know, other other stuff is wrapped up in that as well. And those rights are enabling rights and and so on. so yes, I think that you know we've talked a lot about how just even open architecture can make such a difference ⁓ for users' everyday experiences of the tools that they use. And it creates an environment in which folks can try to solve problems, ⁓ or folks can you know mitigate problems others created, right? And it's not just left up to, you know, spending so much time. I I my c having worked in human rights organizations during these crucial digital age years, right? I have so many colleagues, myself included, who have just spent so much public money, so much time, like begging Facebook or Google to like fix something. And they do, maybe eventually, and then it just takes one CEO level decision to undo all of it, right? ⁓ I just feel like we're trying to create something bigger, we're trying to create something h more hard, like more, more durable. so that yeah, we don't have to start from scratch every time a CEO buys something and then makes, you know, like yeah, this this will be this will endure, I think, beyond beyond that. And that is good for human rights in the long term.

Charles (58:33) So then, kind of extending from that briefly, you know, a lot of your recent work at the Social Web Foundation has been about AI and the social web colliding. Agentic AI is starting to interact on these open protocols. What are the governance questions you're worried about that aren't getting enough attention?

Mallory Knodel (58:51) Right. I think that there's been a lot of attention to the bot conversation, like bot crawlers. Maybe, maybe not in every community, but in from my vantage point, folks have talked a lot about how they don't want scrapers taking their data and throwing it in LLM somewhere. I think that's a totally valid concern, right? Like that's just how we got here, right? How we got to this place where there just so much intellectual theft went into building ⁓ you know, the biggest tech. sector, ⁓ et cetera, et cetera, right? so that's a that's a concern. I I think less is maybe appreciated when it comes to like the the how people will use it going forward. So like not so much bot crawling for the purposes of building an L L but like what happens when people are talking to an AI ⁓ agent that that agent is doing something for them and that might need ⁓ the the the AI agent to go out and find an answer or look for something. And let's say, you know, you're a journalist and you've built like your whole publication, or you're, you know, even a ⁓ some publications are on on the Fediverse, right? They've they've built all their ⁓ information on these platforms. and they, you know, that those agents need to be able to ground their answers in reality, right? They're They're stuck in whenever they were made, 2012 or something. They need to know what's happening today. That that bot is gonna crawl your site and it's gonna try to get the right answer. I don't know that we should necessarily be blocking all of that, right? I don't know that we should necessarily be walling off that kind of content. So it's something to think about. I think it's a gray area, right? I think it's something that we're not really interrogating very much. There's just sort of like a knee-jerk reaction, like. All bots are bad bots because they're building LLMs, and I think it's a little bit more nuanced than that. ⁓ I also just worry about like the mental model folks have around you know, what these agents could do. so yeah, there there's a lot of I guess there's a lot of business models being pitched out there. And so certainly w folks imagine that like you could I don't know, like build new AI tools for social media use, like algorithms. Like mostly we before the AI we were just talking about like the algorithms for your for your timelines, right? We've we've changed our language around that a little bit. But like, you know, algorithmic feeds is something also people really don't like because they feel like they it's manipulated to sell them stuff. But that doesn't mean that it has to be that way. You can also create algorithmic feeds for other things that are maybe pro social or that are ⁓ about like leveling the playing field in terms of voice and representation and diversity. Like there's a lot of different ways I think we can think about algorithmic feeds or curatorial ⁓ social web stuff that doesn't have to be ⁓ it they could be they could have different values driving it and they could have different outcomes for people. So I kind of am interested in exploring that and it doesn't I think all have to be negative. I mean it certainly could be, but You know, if we do cool stuff that isn't negative, maybe that can be prevailing for once. Charles (1:02:24) ⁓ I guess to kinda to kinda close here, are there I I'm curious this is two part. What so ⁓ when developers are building things for for the public, ⁓ what are the kinds of things that are worth thinking about and maybe a short list only 'cause we're short on time and and and what gives you hope about the open internet right now? Mallory Knodel (1:02:55) Right. I think folks need to really think through privacy implications. I'm like just I've spent so many years trying to undo privacy coming in very late in the game. So whatever you're building, make sure you are thinking about the closed confidential version of the thing that you're building. I think that people should also really, yeah, think about what users would find useful, right? It's not just what's easy to build or what benefits the service side of things. Sometimes you have to make a hard trade off to make this more usable for end users. usability is always huge, right? Like do a lot of testing. People will surprise you and how what their expectations are. And you just cannot short shrift it or just chalk it up to people not being power quote power users or whatever. Like you just have to accept that, you know, there's a lot of different people out there. And if you want them to use your stuff, you should ask them How they're gonna use it. ⁓ but yeah, I'm super hopeful. My gosh, I think this is the most exciting time to be building stuff. I know that there's a lot of fear out there, but I, you know, just talking about AI or whatever, you know, it's democratized, right? Like if we think that, you know, this this technology that dirt certainly existed before ChatGPT released it, if the thing that worries us is that like, everybody can use it now. Like that's something that authoritarians like to say, right? Like I like to think of this idea of we've democratized a really powerful tool and we should be excited to embrace it in all kinds of new ways that big corporates and governments haven't yet thought of because that's what I'm excited about. Emma Whamond (1:04:37) An exciting time, yeah. Well, I think we are gonna wrap. Thank you so much, Mallory, for being on, being able to talk to us. this is such an important topic, especially given all of the activity nowadays. and are there any final plugs, anything else that the user or the listener should know ⁓ that you're working on, anything about where they can find you and the work that you're doing? Mallory Knodel (1:05:03) Yeah. I'm I'm excited about the ⁓ you know int and encrypted DMs in the social web. So those that's gonna be work that r is released soon. one of our first reference implementations uses Elixir. So I know that there's overlap there. so I yeah, I hope that folks will follow that work. And you can always find us on ⁓ the web or the social web, social webfoundation.org. And yeah, I'm just really grateful that you thought my work would have overlap with your audience. I'm so excited to be invited to talk to you all about this stuff. And ⁓ yeah, here's to more conversations. Emma Whamond (1:05:45) Well that has been Mallory Nodal and for our Elixir community out there if you are interested in the Fediverse you can go in and explore. Pleuroma, Ecoma, Bonfire are all ⁓ projects in the in the Fetiverse that are in Elixir. So go and explore. Awesome. Thank you so much, Mallory. Charles (1:06:07) Thanks for joining us. Mallory Knodel (1:06:07) you. Yeah. Take care guys. Nice to see you. and I stay on for the thing. Okay, cool. Charles (1:06:10) Yes.