'Technology Freedom Cooperative' To Build Infrastructure for Human Rights and Justice Groups

Audio of this conversation is available via your favorite podcast service.

How can human rights and social justice advocates and investigators protect their data and technology infrastructure in an age of rising authoritarianism in the United States and abroad? Earlier this month, a coalition of human rights, justice, and data organizations launched the Technology Freedom Cooperative, a project designed to help nonprofit organizations protect evidence, reduce their dependence on big tech platforms, and "build independent capacity to work with large language models."

I spoke to two of its members—Maheen Khan, director of technology at the Invisible Institute, and Patrick Ball, director of research at HRDAG to learn more about the initiative.

What follows is a lightly edited transcript of the discussion.

Maheen Khan:

My name is Maheen Khan. I'm the director of technology at the Invisible Institute.

Patrick Ball:

I'm Patrick Ball. I'm the director of research at the Human Rights Data Analysis Group.

Justin Hendrix:

I am pleased the two of you could join me today, and we're going to talk a little bit about this new project that you are working on together, the Technology Freedom Cooperative, which also, as I understand, involves a range of other partners. We'll hear a little more about that. But I think I might just first ask you each to describe what your own organizations do so that the listener has a little bit of grounding for who you are, how you come to these questions that we're going to discuss today. Maybe Maheen, I'll start with you.

Maheen Khan:

I work at the Invisible Institute. We are an investigative journalism nonprofit based out of the Woodlawn neighborhood in Chicago. And our main thing that we do is long-form investigations into policing misconduct. And part of the way that we do that is that we've gotten a ton of police misconduct data over the past decade through Freedom of Information Act requests, also known as FOIA requests. So yeah, our mission is kind of to use data and storytelling to help people hold their government institutions accountable, oftentimes that being the police. And yeah, I'm the director of technology, so I help with a lot of data analysis stuff that we do for stories and any kind of tech work that comes up. For example, we have a data tool called the Citizens Police Data Project where people can look up any Chicago police officer and see their misconduct history, that we've published on a tool.

Justin Hendrix:

And Patrick, how about you?

Patrick Ball:

So the Human Right Data Analysis Group provides assistance so that groups working on justice and accountability have the strongest possible scientific foundations for that argument. So we started our work abroad. My first project was in El Salvador in 1991, and we've worked in over 30 countries since then, trying to put together rigorous data analysis, inferential statistics, machine learning, and most recently AI analysis so that our partners can make arguments about violence by the state. So we work with truth commissions, with tribunals that are trying former officials. I've testified in the trials of three former heads of state, three former presidents, for example, and we work with lots and lots of partners. We've worked with the Invisible Institute for 10 years. And it's been an enormously fruitful collaboration and they are exactly the kind of partners that we have the most to offer because they go get all this amazing data and they have really clear things to do with it, arguments to make, stories to tell, communities to engage, and that's perfect. So this is a really happy collaboration for us.

Justin Hendrix:

We're going to talk about this new collaboration, which invites in other partners as well, this Technology Freedom Cooperative. I think before we get into the meat of why and what it's about, maybe since we're talking about what the orgs do, et cetera, who are your other partners in this cooperative?

Patrick Ball:

So one is a group in Mexico called Data Cívica. Data Cívica does a similar kind of collection of largely public records, although they do some investigation, some collaboration with groups on the ground to study how the government of Mexico uses funding. They work a lot on the transparency records. They look a lot at violence, both violence against women, el feminicidio, the killing of women because of their gender. They look a lot at disappearances and they look a lot at the recovery of the human remains of disappeared people. That's one of the projects we've collaborated with them most closely.

So our partner in New Orleans is Innocence and Justice Louisiana, which is a group that spun out of the Innocence Network. They largely work on getting people who have been unjustly convicted exonerated and released from prison. And we work with Kilimetro Cero in Puerto Rico that has had a series of just stunning successes in their work exposing and demanding reforms of violence from the Puerto Rican police, who have done a lot of pretty terrible stuff over the last decade. And Kilimetro Cero has managed to get a lot of reforms into place to reduce that violence.

Justin Hendrix:

This cooperative has a bunch of different purposes. You talk about bringing them together to protect critical evidence, to reduce dependence on big tech platforms, to build independent capacity, to work with large language models. I want to talk a little bit about each one of those things and what you plan to do around them. But I thought I might start by just asking what's the motivation for this cooperative to come together right now? Was there a specific set of events or I suppose a confluence of things that have happened that made you feel this was the moment to launch this new initiative?

Maheen Khan:

Well, I think a big part of the timing is, yeah, with this current administration, the Trump administration and their hostility towards journalists and just people trying to keep records. So I definitely think that was one of the catalysts of why now. Even though I know Patrick and the folks at HRDAG, and I think so many other groups that have accumulated a lot of data over however many years, I do think everyone's really been thinking about our reliance on these big tech companies, that they don't have any sort of allegiance to us or human rights abuse, uncovering human rights. If they decide on a day that they want to close our accounts, they can do that on a whim. So I think these are things that I know a lot of people have been thinking about for quite some time as we've been slowly descending into fascism. But I think this has felt like the right moment, where we're seeing how quickly we're descending into fascism, like speed up.

Patrick Ball:

Yeah, there was a particularly triggering event for us near the beginning of this administration when the Trump foreign policies team decided to use one of the most powerful levers in the foreign policy powers, which is to place a person on what's called the sanctioned persons list. And the sanctioned persons list was a creation of the 9/11 global war on terror era policies in which putting someone on this list means that they are basically no business that does any business in dollars is allowed to do business with them. This was a list created for the world's worst terrorists. It was created for Osama bin Laden. Yeah, well, the Trump administration came in and decided to use this against the international criminal court. They placed several of the senior officials and judges at the international criminal court on this list. And the effect of that is those people lose access to doing any business in dollars. So that means you can't use a bank, you can't have a cell phone account.

You can't do things that you have to do in dollars because every multinational company in the world does business in dollars, so you just can't do business with them. So people, the story that one judge told is that she came home and said, "Hey, Alexa," and nothing happened because Amazon turned her account off, Microsoft turned her accounts off, and Google turned her accounts off. And this just should be an absolute red flag for anybody who thinks that democracy is strengthened by vigorous debate, because people debating with the Trump administration are now literally in the same category as the world's worst terrorists. This is not just an analogy, indeed, we're using exactly the same tools against them. And I looked at that and said, "Well, what if they turn that against domestic groups?" Certainly, many administrations going back into the 1970s have used de-banking as a tool against domestic critics. So a domestic critic, someone that the administration doesn't like, is prohibited from being part of the banking system.

So if you were a cannabis retailer in a state where that was legal, you couldn't have a bank account. If you're a sex worker or an organization of sex workers, you can't have a bank account. If you were a supporter of Chelsea Manning, you couldn't have a bank account. This is all really well documented. There's a great new book about it called Transaction Denied, I recommend for folks. But what it does is it really should open our eyes to the kind of non-legal powers. I don't want to say illegal because it's not clear if they're illegal or not, but they're not inside the justice system. These powers that the administration has to manipulate private sector actors against the critics of the government. And those of us who are critics of the government should take that very seriously and ask, "Who are we going to do business with?"

Justin Hendrix:

Patrick, Maheen, what I understand you're going to do is try to build infrastructure for organizations who are trying to collect human rights violations, evidence, state violence evidence, perhaps attempt to take on wrongful convictions, other abuses of power. Maybe bring this home for us, Maheen, in the context of your work there in Chicago. We've seen of course a significant state intervention there and extraordinary spectacular ICE CBD raid there. We've seen a lot of activity around that in your city. I don't know, how can you imagine this infrastructure potentially helping you?

Maheen Khan:

Yeah. Well, to take a step back, Invisible Institute was founded in 2014. And one of the catalysts for that was that the founder, Jamie Kalven, had filed a lawsuit that resulted in establishing that in Illinois, police misconduct files are public information. You are allowed to FOIA for those. And that was the beginning of Invisible Institute. Because of that lawsuit, the org got a ton of police misconduct files in all these different CSVs. And the question was, okay, this is a lot of data and okay, now anyone can theoretically file a FOIA request to get this data. But FOIA requests can cost money, it can take a while to get that data back, and it can be kind of difficult to parse through that. So that was why we built a tool where people can just look up officers and see their misconduct history. But yeah, since 2014, we filed so many more FOIA requests and have been adding on to the different types of data that we get.

We also have another tool where you can see employment histories of officers within a state. So yeah, this has been an accumulation of data that we've collected for over a decade. And if one day we were to just lose access to all of that, we would just lose so much of the work that we've done for over a decade. And I think related to, like with Midway Blitz, which occurred in Chicago, which was when a lot of customs and border patrol agents, ICE agents were in Chicago, we were part of a collaborative with some other newsrooms in Chicago where we were collecting video and photo footage from people who were witnessing ICE agents detaining people. And that's also data that we wouldn't want to lose. So yeah, I think for us, it is really preserving all of this data that we've accumulated over a really long time that I think would be very hard to...

We couldn't just file a FOIA request to get it all back. So I think for us, it is really like we're not necessarily scared that the government will see these files that we've got from FOIA requests because we're getting them from a government agency anyways, we're nervous about us ever losing access if... A lot of our stuff is hosted on Google Drive or Dropbox, if they decide to close our account tomorrow, we would lose all of that. So for us having it in a server that's separate from all these big tech companies and it's hosted by us and by a really amazing infrastructure that Patrick and other folks at HRDAG have spent months building, for us, that's where it feels so important. We've been collecting so much data for so long and we want to make sure to preserve that and have access to it.

Justin Hendrix:

Patrick, walk me through the mechanics. How is this set up technically? Who owns what physically? Where are you keeping information? How are you splitting things up across jurisdictions? What's the kind of technical approach?

Patrick Ball:

So the idea is first, as you start out with the lead, is to distribute the information, to decentralize it across lots of different places, and lots of different places and lots of different owners and lots of different jurisdictions. Those things all go together in the notion of place. But I want to say it's not just different houses or offices or co-location facilities, it's different legal regimes and different countries and different disaster zones, because disaster is also one of the things you have to worry about when we're talking about data redundancy. So what we do is, we say, "Okay, each of the five organizations in the cooperative starts out with a server." And that server, they put whatever information they want to put into our archive into that server. They scrape it down from their Dropbox accounts or their G drive accounts or their Amazon S3 accounts or Microsoft Azure or SharePoint, or whatever the tool is that they're using in the cloud to store the data.

We make copies of all that data on their local server. Now, they can keep using the copies. There's nothing wrong with using those cloud copies as long as we have access to their accounts. So we're not trying to displace that usage necessarily, we want to be sure that we can respond if the companies cut us off for whatever non-legal reason they make that decision to do that. So okay, so now we have the data on this local server. Great. Let's bundle it up, bundle it up in little chunks, and the chunks are about a gigabyte each. And the reason we do that is it's just easier to handle things when they're in standard size pieces, to handle them and move them around. If we have hundreds of thousands of tiny little files, that becomes a big sort of catalog headache. You have to keep that proper.

So the first thing you do is you make a catalog of it all, put it in a database, and then bundle it up into these one gigabyte chunks. And each of these one gigabyte chunks has a bunch of important properties. First of those properties is that we know what's in it. So we have a catalog. The second is that we know what's in it in a rigorous way, which means we have a digital sort of fingerprint, called a hash, that represents the authenticity of each file and we store that separately. The next thing that we know is that we know who owns that data. Whose is it? It's the Invisible Institute's data. And we know that because that whole bundle carries a digital signature which can only be created by the Invisible Institute. So the Invisible Institute has signed that commit. We know it belongs to them. So we know the content of it, we know it belongs to them.

We do one more thing here, which is that we take the digital signature of the whole catalog, which, for nerds watching, is the Merkle root of the hash tree across all the files that we've put in this bundle, and we store that on the Bitcoin blockchain. And the reason we do that is that once that's on the Bitcoin blockchain, now we have a timestamp. So that's another guarantee that this data existed at or before the time that that timestamp was created. None of that is, the term in cryptography is repudiable. It cannot be repudiated. All of that is provable mathematically, that these are the files, this is who owns them, and this is when they were created. Then we encrypt all that data with a key that only the partner holds, and that's one commit. Now we repeat that, I think we're up to about 18,000 times across each of the five groups. Sorry, 18,000 times, that's just Invisible Institute. Sorry, it's 40,000 times for the whole group.

Invisible Institute's got the biggest store, but the rest of us contribute quite a bit to it. So now the commits, we have each commit, each commit lives in the server that org owns. Now the commit gets distributed across the entire network. So we have about 20 servers now, and they're distributed across Mexico, the United States, Canada, and a European country that I'm not going to name here. And each of those servers is independent. Nobody tells any server what to do. Every server has shared a rule set, but the idea of decentralization is nobody's saying, "Oh, you over there, make these copies." Instead, the one over there says, "Hey, I see that the Invisible Institute server has published a new commit and they say that they'd like to have five copies of that commit. Am I the right server to make that copy?" And it uses an algorithm that was developed way back, more than 20 years ago by a project called BitTorrent that decides where the files should go. And using that decentralization algorithm, it says, "Oh, yeah, I'm one of the servers that should take a copy of that."

So it reaches out to the Invisible Institute server and grabs that commit and pulls it down. For people who want to know, all the computers are inside what's called a tail scale network or a tail net, which we self-host on a headscale server that we manage in one of our co-location facilities. Each server exposes a specific read-only location to other machines that are doing this replication so they can pull files from each other. So now our other server has pulled that file across and at least three others have as well. So now there's five copies. So there's five copies that are distributed. There's some other rules about jurisdiction routing and location routing. Because my office has six different servers in it, so it doesn't do us any good if all the copies are in one office. So there's all sorts of these little more subtle rules to make sure that a commit gets to, well, there's one in Chicago and there's one in Puerto Rico and there's one in Canada and there's one in Europe.

Justin Hendrix:

So let's just... Maybe to make it real though, someone comes into your office there, maybe they knock down the door or maybe something more, I don't know, normal happens, like the apartment floods, what have you, those six servers you're running in the place are gone or otherwise disconnected. You're saying that effectively you have liberated this information to parts unknown and that won't matter?

Patrick Ball:

Well, it'll be a hassle, but it'll be a manageable hassle. It'll be a recoverable hassle. In fact, just this week I successfully... I did a drill in which I replicated the entire store of one of the organizations with just the key. So all you need is the decryption key from the organization who was lost, in this sense, and you need another place to put the data. And so I had those things and I put the key and the other place to put the data and I was able to recover from machines that are not the one that was, in the drill sense, lost. So if somebody comes to our office and all the computers there for whatever reason are gone or unavailable or destroyed, then I could rock up in another city and with a machine that was adequate to get all our data and with all our keys, which are very carefully protected and backed up in encrypted and distributed ways, recover all the information to this new server and restart our project.

Justin Hendrix:

One of the things that Corey Doctorow said about this project, he talked about it as a kind of preparation for a post-American internet, I think is the way he kind of characterized this. How do you think about that? You're effectively trying to make sure that you can work across borders, that you're not reliant on infrastructure that could obviously be subject to the long arm of the American government or perhaps other governments. Where do you situate this in that thinking about United States and internet freedom within it?

Maheen Khan:

I think for me, part of this project I think is also fighting against how technology and new tech tools or tech companies function under capitalism. I've been really interested in technology for a long time. I studied computer science in college and I think I've just thought the idea of coding and software and hardware has just been very interesting. But I think I have also been pretty turned off by I think the fact that it feels like if you're working in tech, the whole thing is like, "Oh, yes, what's the newest thing? What is the thing that's going to make us even more efficient? And what's going to help these companies make more money as we do these things with these tech tools?"

And I think that this cooperative is kind of pushing back against that, and I think instead is bringing together people who value technology and really see that it can help us create and build new things that we couldn't have thought of. But that it's not for the purpose of just scaling or getting big or making more money, instead, this is to push back against that and be like, "We can build these really interesting new technologies and these infrastructures, but for a bigger purpose, for preserving these human rights violations that we're all documenting. And it's not just about becoming more efficient and being able to scale up." So I really see it as kind of pushing back against that.

And I think especially in the current context with how AI is being pushed everywhere, I'm someone who is very interested in technology and these, but I'm also extremely critical and extremely skeptical. I switched from Google search to DuckDuckGo a couple months ago because I really didn't like that Google Search will not let me turn off AI search results. And I don't want to be forced to opt into all of these AI things. I want to be able to be critical and think about, "Oh, what are the use cases in a contained way where it actually makes sense to be using this specific tool?" And I feel like the TFC is a space to do that and is a space to show people technology doesn't just have to be, for these efficiencies, in order to make these companies richer, it can actually be for people and controlled by people. So that's how I've seen it as. And I feel like that's what I was looking for so long when I was taking computer science classes or trying to learn more about technology, but it's all been from such, oftentimes, an apolitical lens.

Patrick Ball:

I want to really agree with Maheen that I object to the conflation of technology and capitalism. They're not the same. Now, most technology gets built by big capitalist companies, for sure, but there's plenty of tech that is built in other ways. And I have always been a huge fan, going back to the mid '90s, of technology that's built for people who are trying to advance human freedom. That's the goal that I think we should be having. How can we increase human dignity and human freedom with our tools and with our work? And that's what the technology freedom cooperative is about. It's about using technology to advance freedom rather than... We're not being free of technology, to the contrary, we're embracing technology to build tools that'll lead us in this direction.

So to Corey's point that you asked about a post-American internet, I would ask us, "Why do we..." I'm not objecting to America or objecting to China or objecting to Europe, I'm objecting to government. I'm objecting to corporations. I'm saying, "Hey, why don't we build stuff for us, for people? Those other big structures are just problems to route around. They're not helping us now." And indeed, many are working, not maybe deliberately to harm us, but they're working in ways that other actors can turn to harm us. I don't think that, for example, big technology companies that built surveillance into every single piece of their tooling did so because they intended to destroy our ability to organize politically, but that's the effect.

That may not be the intention, but that's the effect because that's what is being weaponized against us by the people who say, "Hey, gosh, you've collected all this information on citizens. Why don't we use that to surveil them for political reasons?" Well, that's terrible, and we need to recognize the unintended consequences, or whether they're intended or not is irrelevant to me really, the consequences in any case of technology that, just like Maheen said, is driven by scaling and profit and surveillance. That's not good for us.

Justin Hendrix:

Part of the announcement of this initiative is that you intend to build independent capacity to work with large language models to help develop local AI infrastructure for organizations that are working on human rights issues. What does that entail? How are you planning to do that in a way that stays within the constraints of some of the things that you just said?

Patrick Ball:

Well, we certainly do not have the capacity to train our own models. That's too expensive. What we have is a bunch of GPUs, a bunch of machines that are capable of running large language models, and we're up to, I guess three machines, and I hope to be up to five by the end of the year, of machines that are in independent locations and jurisdictions, again, that provide us, with everybody in the cooperative, a way to put queries through. And those queries can do all kinds of really valuable and useful things. I think some of the ones that I've found most powerful scaling are when we get, for example, hundreds or thousands of hours of body-worn camera footage in court disclosed... Court orders the police to disclose this footage to different kinds of projects that we're participating in.

We get the footage and we can use LLMs to watch the footage and identify officers and specific kinds of events and link them across a whole set of videos, and then later across the underlying documents for use of force or for the stop records or whatever it is that is the useful information in order to enable defense lawyers to demonstrate that what the police are saying is untrue in crucial ways. And that's really important to be able to do, but that there's just not enough human attention on the planet to be able to read all that stuff or watch all that video and do all that classification by hand. So an LLM can enable us to do that. Like I said, we decentralize that, but we also then enable our partners, like Invisible Institute or the other partners, to be able to use those models in their workflow to enable their work. And Invisible has incredibly exciting projects doing this kind of stuff. So I would like to stop, and Maheen, I'd love to hear your descriptions of some of that stuff.

Maheen Khan:

Another example of, I think one of the services that I think will be really useful on these servers is the transcription service. Right now, we pay for transcription services for when we interview sources or something and need to transcribe those interviews. If we have our own closed server that can do that for us, that's huge because then also you're not uploading audio to this company that you pay a service for, instead, it's in this closed system. So I think that feels like a really interesting use case for me. And then we've previously used machine learning tools for really digging into complaints, like a trove of complaints that we have against police officers to really understand what is actually happening in those complaints. Because complaints get one category basically, but a lot of times that category doesn't tell the full picture at all of everything going on under the hood of that complaint.

So I think using LLMs have been really helpful in the past of actually figuring out what's actually going on in these complaints against CPD officers. It's actually not telling you the whole story when it says, "Oh, this was an illegal search case or operations violation." So I think being able to really extract stuff that's deep in the files of those types of complaints is really helpful to see what is actually the magnitude of what is going on with people feeling violated or hurt by the police. I think it's really exciting that we might be able to experiment with different ways that AI can support us in doing that work, but in very closed environments, where we're really deciding how we're using it and also just what types of models we're using.

Justin Hendrix:

There is so much discussion about artificial intelligence, particularly around the dangers and the fears that many groups have about ways in which the technology will negatively impact questions around social justice and equity, but it seems like you're really trying here to look at a way to take advantage, I assume, of open way models and other open source technologies to see if you can go the other direction.

Patrick Ball:

The cool thing is that for most kinds of technology, if big powerful players can use it to take away people's freedom and dignity, the rest of us can use that technology or a version of it that's maybe freer to reaffirm and to insist on justice and dignity. And I think database technology gives us a real historical memory of the kinds of usage that we can put to machine learning back until a few years ago, and now LLMs.

I think all of these technologies do both, and a lot of times we forget about the way that we, as people who want to advance justice and dignity, can use these tools. And this project is, it's not really our intent, but one of the things that will come out of it I think is these examples of what we're doing with these technologies, which certainly can be used to do horrible things and will be used to do horrible things. Let's not be naive. Palantir is nobody's friend. They're going to do horrible things. They're already doing horrible things. I have actually a list of horrible things they've already done, but we can use similar technology to do good things too.

Maheen Khan:

I think just to go off what Patrick is saying, like I said earlier, I'm super critical and skeptical of when we're all talking about AI. And I think even just saying AI, it's like even that is such a broad term that can mean so many different things. Are you talking about generative AI? Are you talking about the chatbots or are you talking about a transcription service? And I do think that bringing it back to capitalism, that under a capitalist society, I do think any sort of new technology, I do think the focus, for one, the companies that are the ones behind building, it's like, yeah, their priority is how can we keep growing it, scaling it out so then we can make more money and be more efficient and cut costs in these way? I think that's unavoidable under the type of world we live in at this time.

And I definitely think a lot of the ways that AI is being used currently is really detrimental to society. I think AI is being marketed as solutions for I think very real problems that exist. For example, I do think people are very lonely. Because I think we live in a very individualistic society where it's really hard to build community and just have a lot of people around you who live near you and can offer support to one another. I don't think the solution to that is for people to have a chatbot that becomes their best friend and the only person they talk to. So I definitely have a lot of issues with how AI is trying to be marketed as that solution to a lot of real problems that exist. And I think that's why this cooperative is really great because I think it is able to show these very concrete ways of using AI tools that are genuinely to help us get away from those big tech companies that are actually pushing it onto us in these harmful ways.

So yeah, I think I have huge problems with how it's being pushed on all of us. And also I feel like at the same time, it is really important to understand, here are these very explicit ways where it can be useful and you don't need to scale it out so much so that it's really harmful for the environment as well. Part of why AI... Everyone's talking about how harmful these hyperscale data centers are for the environment is because of how big they're getting and how much they're expanding, because these companies keep trying to make AI bigger and bigger. And it's like, that's not what we're trying to do. The whole point is we're trying to keep it very contained and very much fulfilling these certain needs that can help us do our work better, but we're not just trying to use AI to use AI because it's the new thing.

Justin Hendrix:

I understand the cooperative is intended to serve its partners, to start. That's the clear goal of it. So I guess on one level, why publicize it? I do know that you were also planning to, as you say, Patrick, release findings and share with the community, but what's the goal? What's the goal of even coming on a podcast like this or telling others about what you're up to?

Patrick Ball:

So we very self-consciously are organizing as a cooperative. And the cooperative has members. We're not intending, as Maheen has emphasized, we're not intending to get any bigger, but I sure would love it if other groups who all trust each other want to create their own cooperatives. That's the scaling model if there is a scaling model here. It's that if other groups want to put it together, well, gosh, we'd love to help. And I've written all this technology and it's all open source and I'd be really delighted to share it with them. And I'll even help them get it started and I'll help them figure out what hardware they need to buy and how to provision it and how to think through where you put it and stuff like that. That would be really fun. And I think the big reason to go public is to try to spur a conversation, a horizontal conversation among groups that are working to advance human freedom and dignity and justice to think about, "Well, okay, what can we do with other groups?"

So maybe there's groups working on climate change or there's groups working on reproductive justice or there's groups working on LGBT issues or there's groups working on, fill in the blank, whatever that groups are. And they're groups that have affinity, that trust each other, that are connected to each other, that say, "Okay, well, we're going to throw in together a little bit here and build a technology infrastructure for ourselves that we control. Oh, sure, those folks at TFC will help us out. They'll share some tech with us and we'll have meetings with them." And I'll be happy to maybe show up and help them assemble some computers. That's one of my favorite things, but it's theirs and they control it. And whether it succeeds or fails, is up to them. And their focus is... Ours is up to us. And so I think that's the big reason to make it public, is to inspire this idea.

I have a kind of higher order hope for it. I'm not sure. And that's kind of my hope for this podcast, is to say, "Why do we let big companies be the people who define what technology is and what it's for?" It's not. It's ours. We're human beings. We make tools. It's what we do. Tool making is one of the sort of fundamental things that makes us people. I'm not going to let some big company say that they're making tools and I'm just a consumer of those tools. No, to hell with them. I'm going to build the tools.

So that's the other piece of it is to give some sense that this is not just possible... It's not even that hard. This stuff's all out there, I'm just plugging it together. I'm writing a ton of software, but when I say writing software, I'm gluing together all this other stuff that all these other people have written. That's what the free software movement does is it puts these pieces out there and then we can tie them all together for other purposes. And so I think that's the reason that we want to go public.

Maheen Khan:

Yeah. I think people just need to really see that there are these ways of using technology. I feel like technology for a lot of people feels like this weird black box of just like, yes, the internet just exists, these things exist. And I think it can feel very much like we have absolutely no control of this just because it feels so technical, which it is. And I think it is really important to show people there are these creative and radical ways to be using this technology and taking control of it and deciding for ourselves what we want to do with it.

Justin Hendrix:

Maheen, Patrick, I wish you the best with this project. I look forward to checking in on it in a couple of years, finding out what some of the learnings are. And I hope you will hear from some in the Tech Policy Press community who may be keen to follow it more closely and perhaps learn how to replicate what you're doing. Thank you very much.

Patrick Ball:

It was fun. Thanks.

Maheen Khan:

Thank you.