On November 16, 2019, Eben Moglen spoke at a hackathon, co-hosted by the FreedomBox Foundation and the Software Freedom Law Center, to commemorate the upcoming ten-year anniversary of the FreedomBox. The video is © 2019 FreedomBox Foundation and licensed under the Creative Commons Attribution-ShareAlike 4.0 License (CC-BY-SA 4.0). The following transcription of Dr. Moglen’s speech and Q & A is © 2021 Paul Allen and licensed under the Creative Commons Attribution-ShareAlike 4.0 License (CC-BY-SA 4.0).
Eben Moglen: Hi. Welcome. It’s nice to see you. We’ve been waiting for you a long time.
We live in the middle of an Internet that we don’t want. We want much of what it contains which we can think of, for these purposes, as services. There are lots of different ways of describing the piping, and there are lots of different ways that you can map the protocols of the net, but one simple way of thinking about it is as a system of services that both applies to the way in which we think about the networking protocols themselves, one computer providing data or processing for another computer through ports that represent, in the actual configuration of these systems, services ready to be performed. And, above that, at the layer where machinery meets human beings, we experience the net as services: email, and calendaring, and notifications, and sharing of this and that, and pushing of our buttons.
So, if we think of the net as a system of services, then what has happened over the course of the history of net, most of which I have to admit to having lived through. I was in 1975 working for a time-sharing company called the Scientific Time-sharing Corporation. A couple of us – me, I was 16 then and a 19 year old runaway kid named John Gilmore who later made two different fortunes in the history of the net and helped to found the Electronic Frontier Foundation – Gilmore and I in the summer of 1975 were writing what we believe is the first networked email system in the world. It’s certainly the first one we know about for the delivery of email across multiple computers to yet more multiple people. That summer, another of my friends, Leigh Parks [sp?] was working in Champaign-Urbana, Illinois at the National Center for Supercomputing Research on a thing that was still then called the “Arpanet”. It wasn’t the “Internet” for a few more years. In living through all of that history, what I know and what you know – though for you it’s the Stone Age, of course – is that the net was being designed as a system of peers. It was being designed that way for engineering reasons. They’re odd engineering reasons. They’re Cold War military reasons. But the thing we call the Internet was being designed under the assumption that computer networks would come under attack with the intention of decapitating them in nuclear war. And what got the Pentagon interested in making the net in the first place was the idea of decentralized nets that could operate as systems of peers so that they couldn’t be decapitated by destroying the more privileged computers in the network. That basic set of ideas – that the purpose of the net was to enable peerage in order to have robustness – turns out to be important to the history of the human race. It’s why the net has all the survivability it has, which we experience all the time when society comes under stress. We think the first thing that will break is the net, but in fact it’s the last thing that breaks as the waters rise or natural disaster overwhelms a place. And it’s often one of the first things you can get back, with just some radio gear and some batteries inside a place that’s under significant social stress. So, the net works in that sense in the odd way that the engineers and the Pentagon who cooperated to fund its creation designed it to work.
But on top of that underlying structure of peerage – the way that TCP/IP and all the layer 2 stuff below it assumes that everybody’s entitled to a rough-equality of interconnection – the way that the net actually develops is as a series of highly centralized services. They largely replace or grow into spaces where federated services – that is, services in which lots of parties provide as well as consume the services, and do so on a basis of rough-equality and respectful interconnection – federated services have over time become centralized ones. In the beginning that happened in an era in which software was thought of as a product that you bought under shrink-wrap and put into one computer at a time. It’s getting a little archaic to describe how Microsoft became the most powerful and wealthy monopoly in the history of the world, but that was the way it did, with a model that software was a product sold in units for putting into computers. And in that world what gets centralized is the APIs that the most powerful operating system exposes to the world. And through that process the market for services is controlled.
The two most powerful governments on earth spent basically the larger part of 15 years each confronting Microsoft using anti-trust law over the monopolization of APIs and the resulting structuring of the services market in favour of itself. Those lawsuits didn’t work. They were lengthy. They were extremely expensive. They won some and lost some. But in the end what restructured the system of IT in the world was the free software movement, not Microsoft, not the anti-trust action against Microsoft. What ultimately changed things was free and open source software which, by eliminating the idea that we made software to sell to one another by the unit, and instead we cooperated to share software, and to allow ourselves to copy, modify, and redistribute – that process of changing how the world made software ultimately undermined the idea that services could be controlled by being the guy who made the in-box operating system and determining the APIs that ran the net.
That’s great. As generous as Danny was to me in explaining what I’ve been doing with my wasted life here at Columbia Law School as a law professor, what he didn’t say was I spent an awful lot of time doing that. The legal system of free and open source software which was meant to try and push back in favour of freedom by reducing the power of proprietary software was where I spent my midlife, where Richard Stallman and I came together, after Phil Zimmerman and I came together, and where my goal was to try and help people evolve social systems for sharing software that would ultimately undermine monopoly, proprietary control of software. We did that. It’s surprising in fact given how many of our fundamental ambitions remained unfulfilled, it’s surprising how well we did that, we changed the way people made software in the world.
It’s not just the copyleft and the GPL. It’s not just the systems of licensing. Of course, the lawyering is on one side of it. The other thing that changed was that we proved that we could make software that worked better, or at least was way more good at price-performance than proprietary software could make, because we could all get together and share stuff, and we could do write-once run-anywhere, and we had all kinds of advantages because we actually pursued the goals of the engineering instead of the goals of profit. And that turned out to make an enormous difference to the way that software worked. And then all the profit-makers piled in at the beginning of the twenty-first century. And IBM, and Hewlett-Packard, and all the other hardware manufacturers, realized that free software was good for them. And they got rid of AIX and they got rid of HPUX. And everybody decided the Linux kernel’s nice, why don’t we all make it better together? And the GNU utilities are nice, and why don’t we make those better together? And on top of that, we could take this other stuff the free software world does, like this HTTP thing, and maybe we could have a Web. And in effect of course what happened is that the world of commercial and enterprise IT moved in on us. And it started to use our kernel. And it started to use our compiler. And it started to use all the things we made that we called the World Wide Web that were invented in places like CERN. And of course it all worked very well. And it was good. And we had made it. And people got nice jobs helping to continue to make it.
And so an economy of free software came into existence. And because there was an economy of free software, there could be things like Google and Facebook which would not have been possible in a world of shrink-wrapped software, where you had to buy one unit for one computer. All of the scaling that they did to put power back in large clumps in the network depended upon free and open source software. And they built everything they did on top of it. And they played by the rules. Where they were supposed to give back, they gave back. And where they weren’t supposed to give back, they gave nothing back. And they created proprietary opportunities for themselves. And those opportunities fundamentally depended upon a new form of economics: Give people services quote for free – and spy the crap out of them. Use what you know about them, particularly what you know about them exactly now, in the last 45 minutes, to conduct hundreds of billions of online auctions a day, to present an advertisement to people based on what they know about what you might be wanting right this minute because they spy the crap out of you. That’s the model.
It’s extraordinarily profitable. The business of knowing what people want before you advertise to them is a very good business to be in. It’s like buying horse race tickets when you know which horse is going to win. And therefore it makes a ton of money. And auctioning off your power to do that billions of times a day makes a lot of money because you get all the surplus back by conducting those auctions, and getting people to compete to advertise to you for stuff that the wizard in the middle knows you already want. It seems improbable, and indeed without a whole bunch of free software it is improbable, but if you use that social strategy to deliver commodity services like email and file sharing and other such stuff, and you say to people, “Hey, I will give you this email and file sharing and calendaring and other such stuff. Please pay no attention to the fact that I’m spying the crap out of you, because I’m letting you have it for nothing, so what’s your complaint?” And I will sell you as many times per day as I can, and I will help optimize your possession of hardware which spies on you even better than basic hardware would, and I will get you to touch that thing 25, 26, 2700 times a day, and pretty much every time you touch that thing, somebody’s going to get an opportunity to address you about what the last 45 minutes of your life says you want, and I’m going to sell all those opportunities, and I’m going to keep all the money, but you get email for free. And I’m going to make you believe that that’s good; indeed, I’m going to make you believe that it’s impossible to be a human being in twenty-first century society without playing that game with me, and I’m going to win everything.
That’s a new form of the control over services. It depends no longer on the possession of a system of proprietary software which controls the APIs and makes everybody write to them. That’s unutterably primitive now. It depends instead upon changing the way computing works in a few fundamental ways. Maximize the surveillance utility of the endpoint computers that people use. Pack them with sensors. The contemporary smartphone is the densest collection of sensors gram-for-gram that the human race has ever made. We have stuff in orbit that is very, very capable. It can do all kinds of magic from way above the surface of the earth. Satellites have even more cameras than your smartphone now, and the glass is a little bigger, so the pictures are much better, but it weighs much more. The smartphone is a spy satellite in your pocket, maximized, optimized and arranged for you to want to carry it around, and make sure that it’s always fed. And it’s very good at sensing all kinds of things, and the things it senses are by definition about you, and every packet it emits is by definition about you, that is to say, is named “you” to the telecommunications service provider and to every platform that has access at one or another layer of the software to you. Everything the spy satellite senses is about you. And we produce these things relatively inexpensively because we want you to carry them, and we subsidize the naturally open source software in half of them and the proprietary jewellery-manufacturer software in the other part of them. And we use those things to know what you are wanting right now as well as possible, and to neck it all down through one place, and send it across some proprietary network, where we will split some value with the telecomms carrier. And then we will go to work on you, and that work doesn’t take very long, because we’re planning to do that work again and again all day long. We’re just going to keep selling you to people. That’s the new model of services delivery. Aren’t you glad you have free email, and isn’t it so convenient?
The problem is that it is very convenient. It’s good to live in digital society, and it’s good to have services, and people want the services, because they like the society. And I don’t blame them. That all makes total sense. It doesn’t make complete sense to me to carry the spy satellite in the pocket, so I don’t. I lead a rich and varied life in the net. I’m all over the place. I run a lot of computers. As Danny says, I’m a compulsive programmer. I like making software. I’ve been doing it for almost half a century, and I’m still doing it, because it’s still fun. But I don’t carry a smartphone, and you couldn’t make me. It would not occur to me to run a computer whose primary job is spying on me, and whose secondary job is making my phone calls. But it’s OK, I mean, we’re going to live like that.
So, let’s assume that the real goal of what we’re trying to do now is to ameliorate the really bad consequences of a net built around centralizing services and offering them cost-free in return for pervasive surveillance. Assume that problem with me, for a moment, because that’s the one that FreedomBox wants to solve.
Assume that the problem is, that the net is fundamentally misshapen in the way it offers services, and that services, which are really commodities, very easy to implement, very easy to stand up, simple to administer at the person-by-person or small-group-by-small-group level, those services are being sold to the human race at price apparently-zero, actuality of the tools of totalitarianism all around them all the time.
Assume that what the system does is to produce very, very, very high, very unequal profits, and that it does so by costing us the constant running of basic risk to human freedom.
Join me in that assumption, and it’s a very serious problem, and we should solve it. We shouldn’t tolerate it. We shouldn’t make deals with it. We’re going to have to, because the real world is real. But our goal should be to transform the technological structure of the net as fast as possible away from that, and never let it happen again. It’s a serious form of social pollution running rampant through the human race. It’s based upon a network which misshapes how it delivers services, and the problem that we have is that it passes off very simple services as magic you should be willing to be spied on for. So, we should undo that.
That was the point I was trying to make on that winter evening – a Friday night – in New York in 2010, when I was boring people with the idea that we shouldn’t do this, we should do it differently. It’s not a very complicated idea, so I don’t want to take a lot of credit for it. And particularly it’s not a very complicated idea if you grew up when I did in the infancy of the net, because it was then the reigning idea, and there was nothing very complicated about it. Everything we were doing, we were doing on that basis, so it was old news back then, at the beginning of time. The news is, we should federate the services. We should deliver them to one another. We should keep the scale of service delivery small. If you don’t pile up all the email in one place, then there’s nobody reading all the email, guaranteed. And if you don’t get all your services from people who offer you a subsidy in return for all your behaviour, every service you get from somebody who isn’t asking you for all your privacy in return, is one less thing destroying your privacy. Are you doing your sharing not through somebody who’s advertising to you and, therefore, selling you? Are you getting your email from someplace which isn’t reading all of it to figure out what you’re doing today? Is your calendar somewhere which is very convenient for you, but not really convenient for somebody who wants to data-mine all the calendars in the human race to figure out what’s going on? Etc. etc.
And it was perfectly clear by the beginning of the decade, when I started talking about all of this, that we had enough hardware around that would be inexpensive enough that we could provide services to one another and the people immediately adjacent to us very efficiently. That was true in 2010. Needless to say, it’s not less true in 2019, right? Cycles are cheap. Storage is cheap. SOCKS are wonderfully sophisticated computers on one chip, that you can put on a circuit board, that will conduct all their I/O to the outside world for them very inexpensively, and, by God, we have little micro-servers that can do everything that each of us actually wants.
In the course of all that time, two things also happened. First of all, the big companies that provided centralized services in return for spying on you all the time massively increased IT efficiency. They took the free software ideas that allowed us to treat operating systems as basically disposable commodities, and they created “clouds” – virtual machine systems with virtual operating systems, cheroots floating loose in the world that they called “containers” – all kinds of structures for putting processes and their associated memory at the right place which, for a great big, globe-girdling multinational company, means where electricity is cheap, where the costs of operating all of those processes, and all of their associated memory, add up to the smallest amount of money.
In other words, what they became was the most efficient businesses in the world at turning electrons into profits. It’s an achievement. There’s no question about it. It’s awe-inspiring, if you grew up in a world of mainframe computers, you know, water-cooled objects. It’s a stunning achievement to have become as efficient at directly turning electrons into profits as they are. It’s important to point out that one of the ways they do this is by reducing employment. In the global IT economy the cloud has been a neutron bomb which killed the people and left the services standing. That was its point, really. From the point of view of those people selling enterprise IT, to every kind of business and organization in the world right now, the great advantage is, make yourself the cloud. Do your utility computing through us. You can fire all those Sys Admins, because you don’t need any computers of your own anymore, and we’ll take care of it for you. Those of us who actually made our living, or planned to make our living, or pay for our educations, as I did, working as skilled-labour in the IT industry, might not think this is the ideal outcome. And we de-skill a lot when we concentrate a lot of knowledge in places, where only a few people need it, because only a few people can work.
One of the things that drew me to the free software movement, in the first place, was by sharing software we could teach everybody everything at no cost, because all the programs out there that you would need to do anything that computers could do, you could get by reading code that anybody could read and nobody had to pay for. So, free software, from my point of view, was the greatest technical reference library ever created in the world because it allowed you to go from zero to the state-of-the-art in everything there was, just by reading things we all shared, and nobody had to pay to look at. In the same way, of course, it would be really nice if, in the twenty-firstcentury, when billions of young people are coming to the net from the moment of their birth, skills in how to use the net, how to stand up services on the net, how to run computers attached to the net, how to help other people use computers, it would be better if those skills were proliferating throughout society than if they were being concentrated in an increasingly small workforce, located in an increasingly small number of data centres, belonging to a smaller and smaller number of employers.
There isn’t anything about this system of delivering commodity services using spying and the cloud that doesn’t have, from my point of view, more negative outcomes than it has positive ones. There’s no part of the way this whole system works, that doesn’t seem to me, like bad political-economy, bad social policy, bad human development policy. But it makes a ton of money and it works. And the services are convenient, and so everybody wants them. And so the human race is driving itself into a ditch in which it is confused about the idea that things are good because they work, under the assumption that that’s the only way they can have them. And that, without the system, we wouldn’t have the things that we consider convenient and, therefore, to the extent that it hurts our freedom, that’s just a price we have to pay.
So, what’s the solution? Provide people with the easiest way to manufacture as many services as possible using the cheapest hardware that there is, and do all of that in a way which is designed to prevent spying on the human being to the maximum extent possible, instead of maximizing the amount of spying that goes on. Those are the two technical objectives.
Nine years in, thanks to the people in this room and other extraordinarily self-sacrificing technologists who have been willing to work on the right thing, even if it didn’t pay them, we’re in pretty good shape.
What do you need? You need an integration of software effort that provides you a free software stack that you can copy and share without limitation, that will turn the smallest and cheapest hardware around into things that do real work, providing actual services to human beings, and which do so in a fashion designed to minimize instead of maximize spying. That’s what you need.
It turns out that, what that really means is, we need to take the best assemblage of free software that we know and optimize it, configure it, re-design it, re-engineer the little pieces of connection, alter UX, and try and imagine a different way of interacting with the machine. We need to do that, and make it available on the most inexpensive and powerful single board computers we can find. That’s the prescription.
Taken daily by the entire human race, it would eliminate the problem overnight, but that’s not a feasible outcome. You see that we’re not the entire human race in this room, not by a fairly long shot. We’re still, in that sense, pioneering technology, explaining to people, “Gee, we have this strange idea. [Holds up Pioneer FreedomBox] Here’s this tiny little board, inside one tiny little box, and you should use that to do things like synchronize the files between your laptop and your phone. And you should use things like that to have a Wiki that you and your friends can work on together. Whenever you want to put a project up you put up a Wiki page and start working on it together, the same way that my law practice works. And if you got a thing with a Transmission daemon in it, then everything around it can be doing BitTorrent communication through one central little spot, and we can share software and media and other things we use BitTorrents to catch. And so on and so on. We can have various kinds of services. They may not duplicate absolutely and precisely the services being used by the spying companies. It doesn’t matter whether they’re identical. The question is, do they help us to do all the things that we want to do, while costing us fewer of the things that we don’t want to do. And will we find ourselves in a world in which the data sits in places that we have chosen, and where you can’t break into it without a hammer or a credential nobody should have let out, instead of giving it away all the time to everybody and telling them, “Mind this for me. I’ll be back to collect it, whenever the GDPR says I might and I want to,” which is never for most people. “Please check here for Terms of Service that say, “We’re going to pillage your privacy completely, but we have an amazing new service we would like you to try. Check ‘Yes’ here.”
To get away from that world is worth a ton. As it is to have suddenly a privacy proxy between you and your browsing, so advertising starts dropping off your phone, and all kinds of zero-length GIFs that used to be following you around the net aren’t following you around anymore. As it is useful to be able to say, “Oh I would like to click here to have a Tor router, so that I could begin actually to help other people in the world to attain some working anonymity in their life on the net.” Etc., etc., etc., etc., right. All of it’s easy. We use it all the time. We use it all the time, if we are Sys Admins in business running free software and throwing up OpenVPN VPNs. And we use it all the time, if we are, you know, hyper-cool people with our own VPS under Newark Airport that we pay a couple dollars a month for, and then we have some other stuff behind it. I’m guilty of this. I use a lot of computing, and I hold it together with the stuff that people would hold it together with if they were IBM or CitiBank because it’s all free software, and it’s as easy for me as it is for them, right. So, industrial-strength computing security. Industrial-strength secure networking. All the stuff that we can give people because the software is zero dollars a unit, we don’t necessarily expect to see it gotten for people any other way.
If you look back at 2013, when I was giving the Snowden Lectures in this room, and trying to explain to people what I thought was going on because this young man had done this extraordinary act, I was talking about the companies’ positions. They’re platforms. I said, they could make more privacy. They could end-to-end encrypt their messaging. They could allow OpenPGP encryption of the email on their platforms, etc. And, over the course of time, people did fall into those ameliorations. WhatsApp is a, you know, a capably-encrypted system of communication used by billions of people around the planet. That’s not a tiny outcome. But why should it also come with pervasive traffic-analysis, and psycho-graphic marketing, and the opportunity to corrupt politics by giving political parties access to the inner lives of voters in return for, well, whatever favours it is that WhatsApp or its parent company, now in all capital letters, like a Trump tweet, might want to ask for later on? We can do that. In fact, we do do that. In fact, make a FreedomBox, run a Matrix server, give people the Riot client. Excellent messaging, very secure, end-to-end encrypted, very sweet. And nobody in the middle to attack you.
The government of France has gotten concerned about the fact that so many politicians, Ministers, bureaucrats, people within government, use WhatsApp for government business. And this makes the French government nervous about who might be listening. And so they say to themselves, “Maybe we should have a government-wide, secure messaging platform for the French Republic.” And so they start evaluating software and they decide that Matrix would be very good. And I think to myself, “Yeah we think that too at FreedomBox [holds up Pioneer Box]. In fact, maybe you guys just want to buy a bunch of boxes and make your own messaging network.” You can – that’s what this is about today, right.
The idea is, learn how to provide services so easily that the idea of the Sys Admin disappears, not because the job went into the cloud, but because administration of services isn’t complicated in that way anymore. That’s the software integration task my comrades have been pursuing now for years. Take an arbitrary or, not so arbitrary but at least supported piece of hardware, put an image in it, have it come up, spy around where it is in the network, decide what it is it needs to do in order to connect up for you. And say, OK, from now on, I can offer services over the net. I can offer them privately through onion routing. I can offer them publicly on the public net. I can do all sorts of things for you, and all you have to do is tell me which ones you want, one click at a time. And I will make installing the service, properly configured for both your network environment and privacy-conscious use, no harder than clicking once. And so, you can just go through the menu of all the various services that you might want to offer to yourself or to the world. Bang, bang, bang, bang, bang. Set them up. They are configured correctly. You are not leaking your information. You are not agreeing to give up everything. There were no Terms of Service, it was free software. You have the right to run every program for every purpose, fix it, modify, share, and you’re in a community. There’s support there. There’s help there. There’s like-minded people there, whose goal is to make the network work in a more privacy-respecting fashion, and to get themselves off platform services. In addition to everything else, therefore, you have what free software communities give you, which is friends, neighbours, countrymen. People who share a form of aspiration which is both technical and political at the same time. We want to have computers that work better for us, that work better for the people around us. We want to do things with computers that are good for society. And we also want to learn. We want to know more. We want to have more power. We want to be more effective in the world of the technology, because everything we’re doing we’re doing with software we can read ourselves and learn from, and understand, and contribute to. And all around us there are people who have gone through that very same process, and they will help us to become facile contributors to this code. We can have co-ownership of it. We don’t buy in. We live in. And under those circumstances, suddenly what is happening to the net is that great big objects that live by spying are competing against tiny little objects that live by teaching, learning, helping, and becoming better through common effort.
This is what we used to call a competitive free market, right? This is what we think the goal of a certain kind of liberal society is: to enable its citizens to go out into the world, and make things that better their condition, and to profit, morally as well as financially, from the fact that they have bettered themselves in life. FreedomBox is democratic technology. That’s my attitude about it. It is technology which makes better democrats – with a small “d” – because it enables people to do in society more, and it helps to confront the concentration of power in ways that are effective and that actually work.
If you want to have a movement for technological freedom in the twenty-first century you got to have deliverables [gestures to Pioneer FreedomBox], you have to say to people, “Here, use this. It makes life better.” So, that’s what we’re up to.
The purpose of a hackathon in that world is, well, to prove that we can hand out apple seeds and later there will be orchards, and to demonstrate that learning how to help liberate services in the network is neither boring nor complicated. And therefore, to say, and as we spread out and build our services, as we create what for us are the equivalent of all those App businesses out there, we create a different way of making technology. To be an application developer in the current world is to say, “I have a concept for things that people will engage with. Now let me go and get a bunch of toolkits, all of which spy on my users in different ways. Some of them know about Bluetooth beacons in stores, and some of them are really good at bugging near-field communications, all these. I will take a bunch of these toolkits and frameworks, and I will build my application, and I will get people to engage with whatever I am doing, and meantime bunches of other software will be siphoning off my users into their own little data receptacles, too. That’s being an app developer, and it’s great. You make lovely software, and people use it, and it makes funny noises, and everybody’s happy. But it’s also great to be a developer in our world, where you don’t do any of that. Where the APIs are basic public, open, always been there. The protocols are the ones we prefer because anybody can implement them on both sides. You start with a whole bunch, like say 53,000 Debian packages, and you make up a little bit more, and you put a few things together, and as we get the developer platform more and more simplified, and it becomes possible to pull in all kinds of componentry with mere declarative statements someplace that don’t even amount to code snippets, just Python data structures, and they say, “Here get me all this stuff, and give me a way of putting my application together, and now I will offer a service on the FreedomBox.” It wasn’t any harder to do the coding. It was just as simple, but there wasn’t any spying built inside. At no level. No framework. No toolkit. No me-myself. We make honest software here. It does what the user thinks it does, and it does it for the user, and it doesn’t sell the user to anybody else, and it isn’t engaged in covert operations against humankind. And yet it’s still simple and fun to program and works all the good ways, and can make your App overnight, and it will still be beautiful. And, really cheap hardware.
So, the other thing that happened during the course of this decade was we figured out that we were melting the planet – which is not a good idea. Converting electrons to profit very efficiently still produces heat because entropy is the rule of the universe. And even if you get a bunch of your electricity from, you know, renewable sources, that heat that IT throws off also expresses itself in carbon atoms going to random places.
Ask this question about the carbon footprint of cloud computing: Assume that there are some economies of scale – assume this, I’m not sure it’s true, we’ll come back to this in a moment – but assume there are some economies of scale, and that processing the personal lives of billions of people produces a little less in the way of carbon output per human service delivered, if you combine all of the stuff into the largest possible computers and run them in the most efficient places where electricity is cheapest. Assume that that’s true. It is still true that surveillance costs way more computing than the services themselves. It requires way more processing at the point where services are created. It requires way more storage to store all the output of surveillance. There is all the secondary processing connected with turning surveillance into advertising auctions. There are all the advertising auctions and there are all the advertisements and everything that happens because of them. In other words, even if you assume that services are efficiently delivered from an energy point of view in the cloud, the cloud itself is grossly energy inefficient because it spends an awful lot of energy, far more energy than it spends delivering services, surveilling, and processing, and selling advertising based on the people for whom services are performed. In other words, we’re not melting the planet less by doing it this way, we’re melting the planet way more by doing it this way. And if everybody has one spindle containing a terabyte drive, and there’s one single board computer, and they are delivering all the services that they need and much of the messaging and VoIP calling and all sorts of other stuff that the people around them need, we’re getting away for a tenth of a light bulb what Google and Facebook and the other platform companies are spending kilowatt hours on, because they’re delivering the services very efficiently, and they’re also surveilling and processing humankind very efficiently – but still, if you don’t do it, you don’t need the heat.
So, what we’re doing is in two senses ecologically relevant. We’re trying to change the information ecology of the network, and we’re trying to change the energy balance in how we account for the personal lives that we live in the net. I didn’t set out to think much about heat dissipation or the carbon footprint of cloud computing, but the more research we do around here the more interesting it becomes to me. And the point after all isn’t different from the point with which we started, at least as I see it, when we recognize that a network which is more distributed and that federates more services is more ecologically responsive than a network which consists of centralized services run for profit. It’s not that we don’t think that capitalism is efficient. Capitalism is very efficient. The problem is that when it doesn’t account for its externalities, it’s ruthlessly inefficient about creating them. So, it destroys physical environment. It messes up water. It pollutes atmosphere. It changes climate – not because it isn’t efficient, but because it isn’t forced to calculate those questions. If you’re running your own personal server, you know how much you paid for it. You know how much the electricity costs you. You know how hot it gets. You have feelings about that, which naturally cause you to think of computing as part of your personal environment. You do what you do because it’s an efficient way to do things, but you are also, of course, thinking about the externalities of your own life and the lives of the people around you. If you knew that the Raspberry Pi was puffing out little bits of poisonous gas, you’d stop using it. If you’re Google with a data centre, that is not what you do. You minimize it, and you reduce your risk as far as possible, and you cover your legal options, and you go on. This is the problem that we solve when we scale the net back toward human beings, and do so in a way that human beings can both understand and appreciate.
And the last thing I want to say is: and we teach people. These aren’t just boxes that provide services without spying, and they are easy and services are convenient, and Sunil and James and everybody have done such a great job putting the software together that you’ll never have to look under the hood, and it’s really easy to run. All of that’s true, but then you could look under the hood. That is to say, in there we have everything that you would need in order to understand an awful lot about twenty-first century computing: how services are delivered, how software environments are put together, how hardware support for cheap computers, like those you see around you in the world is performed, and also how to think about privacy. Again, free software is doing the job of teaching technologists, instead of removing technologists. Free software is doing the job of making more sophisticated technology more accessible to people.
That was all the happy part of my talk. The unhappy part of my talk is getting people to try this is really, really hard. We’ve spent the better part of a decade moving all the corners off the software, and James and Sunil and all our other comrades have in fact produced an extraordinary system. When Distro Watch, which is just a bunch of guys who are always installing distros and reviewing them, got around to reviewing FreedomBox as a GNU Linux distribution earlier this fall, the reviewer’s comment was, “I don’t get a chance to say this very much in my task of installing distributions and using them, but FreedomBox impressed me.” So, OK what I need to tell you, if you’re a geek connoisseur is, we make impressive software out of the usual parts. There’s nothing different about the parts. We just put them together in a creative, interesting, and potentially even impressive way, which means there’s some nice stuff in there for you to learn. But even so, you know, they have billions of users. We have hundreds, thousands, you know, little groups. And here you are. This is why I say, I’ve been waiting a long time. I didn’t mean you were late for the pizza. I mean, you know, lots of work went into having a place for you to come that we think is interesting and fun, and that has enormous possible value.
There’s lots of projects you can spend a weekend on, right. You have lots of options. I have lots of options. There’s lots of cute stuff, fun stuff, interesting stuff, mind-bending stuff to do with computers. But a couple of hours spent this way has all sorts of possibilities latent in it, because there really are technical options that are just as convincing as free software was in 1991. And there’s just as much danger to the human race that we get it wrong, and we have no privacy or freedom anymore, as there was when I was trying to keep the United States government from sending Phil Zimmerman to jail and making PGP illegal to export. In both cases, that is, in the time when I was young and in your now-time, the threats to human political freedom that follow from badly-configured or badly-controlled technology are real serious. But we thought, then, that it was serious, and we had no bloody idea how serious it was. After 2016, most of the world became aware that political integrity depends upon software freedom to a significant extent.
Back all those years ago that Danny was talking about, when I gave the “Freedom in the cloud” talk, in 2010, I said, “Mark Zuckerberg has done more harm to the human race than anybody else his age.” He was then 26. He’s now 35. If I had said that night back in 2010, in addition to my point about Mr. Zuckerberg, if I’d gone on to say, “… and 9 years from now, tens of millions around the world are going to agree with me unequivocally,” everybody in the room would have thought I had really lost my mind. But if you ask, “Who has done more harm to the human race at 35 than Mr. Zuckerberg?” I have a feeling that a lot of people around the world would scratch their heads and tell you, “Yeah, we can’t think of anybody else either.”
As improbable as that sounded, in 2010, it is equivalently improbable to say that, in 2029, we could substantially have repaired this problem, without using a whole lot of government force, and without using a whole lot of litigation, and without spending a ton of money, if instead what we did was learn how to provide services over cheap computers and help the people we know of in our lives to use them. That would grow and spread the way FOSS grew and spread. It wouldn’t necessarily be adopted by all of capitalism, but I promise you, at that point, people will make FreedomBoxes, and sell them as commercial products, and they’ll make money doing it, and they’ll be very happy to.
So, this is the part of the road where we say to you what free software projects always say: “Hey, we’ve got some stuff here. Proof of concept, running code. Actually works good. It makes product. We can sell you one, or you can make one for yourself – in fact, this afternoon, you could make two. We’ll share it with you. It costs nothing. It might help.”
I’m happy to take your questions. Thank you very much.
Question and Answer
Questioner #1: What might you see as a threat to centralizing and having FreedomBoxes?
Answer: Well, you know, the two most important obstacles that we face are ideas in our own minds. The first one of which is, “Privacy is over, get used to it” – which is very discouraging, indeed, and which demobilizes an awful lot of people. And as I tried to say before, it’s partly based on a rational calculation that happens to be based on factually incorrect premises, namely, I could not have the services I wanted, unless I sacrificed my privacy. So, therefore, I’m put to a terrible choice, and it’s not that hard. I choose my services, OK.
The second great problem that we face, which is also an idea in our own minds, is “I’m not doing anything wrong. I have nothing to hide. Why should I be so concerned about my privacy? Sharing is good.” This also is a set of rational calculations, based on incorrect premises, and once again it reflects, in some sense, a sensible idea, namely, I should be managing my risks according to what they are, and I think my risks aren’t very great, so why shouldn’t I do this? It’s again, it’s rationality and there’s nothing to be said against the people, except you haven’t been given the facts, so you don’t know. Those of us who lived in the twentieth century, we got the facts, because we lived at a time when governments behaved in ways of such unmitigated savagery and such determination to possess the totality of citizens’ lives, that many of us decided we would rather die than live that way. And some of us were compelled to take the short end of that choice – millions of my people, for example. So, we walked out of the twentieth century thinking, “Don’t trust your judgment about how much risk you’re in. Just because it’s 1934 and it feels OK. It could be 1939 in a moment. That’s the way the world is put together.” We were all adults and children together in a Golden Age in the history of the human race, which is ending before our eyes. People around us are going to learn the lessons of the twentieth century yet again, and they’re going to learn them in exactly the same way. Young people in Hong Kong, this afternoon, face a problem which is not different from the crucial problem of the twentieth century. “Do I leave and run away, and try and save myself somewhere else in the world, or do I attempt to believe in the possibility of my own freedom?” In Xinjang, our co-habitants on earth, the Chinese Communist Party, have created an image of perfected despotism, using the tools of smartphone surveillance, big data, social media, connection ultimately to genomic information which makes it impossible to evade control of your body generated inside. All the good stuff, every piece of what, in the twentieth century, I and my peers were afraid of, because science fiction taught us to fear it, we are seeing before our eyes. If we ignore that, it’s our own fault. And everywhere in the democracies people are now aware that commercial surveillance can threaten political integrity and undermine democracy itself. And that understanding is shared between government officials and young voters in the street, and the people manipulating the system, and the people manipulating the people manipulating the system. We all know it. It’s an open secret now. And everybody’s got one kind of incentive or another for not acting. So, the other big problem we have is we don’t feel the political temperature of the water in which we’re swimming. Therefore, “Privacy’s over, get used to it. My services are very convenient. I couldn’t give them up. I couldn’t live without them. I must continue to make the compromises I am making. I’m not doing anything wrong. Despotism is not really looking at me. All the data-leaking I am doing is not actually going to get me dragged off in the middle of the night, or my children harmed, or some vicious thing done to the people I love, that’s only going to happen to someone else.”
Those are the biggest problems that we have in defeating ourselves in changing the net, and they’re all ideas in our own mind, and the minds of the people around us.
The problem is, we need a lever and a place to stand; otherwise, the whole Archimedian thing doesn’t work, and we don’t change the world. So, this [holds up the Pioneer FreedomBox] is designed to be the fulcrum, right, it’s just a tool. And the abandonment of the unsafe ideas, that this isn’t really about me, and privacy is over, that’s the lever. We have to change people’s minds about that. I actually think the machinery is no more complicated than the world’s simplest machine. Lever plus fulcrum plus the desire to move the earth – that’s sufficient. You only need a place to stand. That’s what I actually think about this.
Of course, there are lots of technical problems. I can’t tell you how many thousands of problems my comrades have solved over the years, and we have many more to deal with, and they can get very irksome. Trying to figure out how to stand up an email server inside a FreedomBox is way worse than you think it is, because the software’s not the problem, but discovering how to send … Anyway, you know, it’s all like that. We have to solve problems. The more people we have the more all those bugs are shallow.
So, I don’t actually think of our problems are technological. I certainly don’t think our problems are economic. We run really nicely now on an $80 computer. I am convinced, after yesterday’s conversation with the Orange Pi Zero enthusiasts among us, we may be able to run OK on a $10 computer now. We will certainly be able to run on a $10 computer ten years from now. If we can do that, and if we can provide all the best-of-breed services, if we can give you WireGuard VPNs, and we can give you onion routing, and we can give you the best sync thing, and the best cloud storage, we don’t have to worry about the little difficulties. We have to worry about the big one, which is the problem in our own minds.
After Trinity, Albert Einstein said, “We have changed everything, except the way men think.” And if we you allow for the fact that he was, in fact, speaking about the way men think, he may even have been correct that we’ve changed everything except the way people think. The net is changing the way people think, and it’s not changing it in a good way. So, first we have to help people confront some beliefs which are incorrect. They’re rational, they’re not worthy of deprecation. These are people thinking and reaching logical conclusions. They don’t know the facts. Their conclusions are wrong because their predicates are wrong. Privacy is not over, don’t get used to it. You don’t have to give up personal information to have reasonably convenient services. No, it’s not the case that because you’re not doing anything wrong, right this minute, you have nothing to be afraid of forever. No, all of those premises are incorrect. Let’s help you to recalculate, and then let’s give you something to deal with to manage your risk. If we could get people to go through that thought process, I will buy all the other remaining problems that we have for a nickel, because I know that those problems will get solved. Other questions?
Questioner #2: Is this hardware, software, is it … can you use it with … would you have to just give up on Google, Gmail?
Answer: No, we’ll go as far as you go. We’re not going to take you any further than you want. A FreedomBox user in the Open Forum conducted over secure VoIP inside a FreedomBox – Mumble and Murmur – by the end of the day, if you haven’t fallen in love with Mumble, I’ll be very surprised. So, we’re holding our world-wide Q&A over the net through a FreedomBox yesterday, and one of the users says, he has almost completely de-Googled himself. And I thought, Wow, that’s really amazing. But he hasn’t fixed Gmail yet, because we haven’t a good email server. It’s still bothering me, right. Next year, maybe I can stand here and tell you, right.
But no, the goal isn’t to create sacrifices. The goal is to create alternatives.
Here’s the one that seems to me, from that point of view, the most important. So, you install a FreedomBox on a piece of cheap hardware, like a Lime 2 [points to Pioneer FreedomBox], and you put it in your apartment or your dorm room or somewhere where you have network service that you can sort of trust. I don’t mean trustworthy, I mean, it will stay up. And you tunnel from your smartphone with an OpenVPN client to that, and use it to pass all the traffic to the world. Now, when you are sitting in Starbucks, the untrusted Starbucks network is meaningless. It doesn’t have anything to see. It gets no real packets. It’s just watching encryption pass it by. When you have tunnelled your smartphone through a FreedomBox in some place where your primary network connection is, you’ve suddenly turned that smartphone into a much more secure object than it was before. I can’t say it’s a secure object, it’s a smartphone. But the way it’s being used is already way better, and from your point of view, the only thing that’s changed is your physical location has suddenly become unreadable from your network traffic, which is most of the time a really good thing. Right.
Questioner #2: So, it affects your devices, even if you’re not next to the box.
Answer: It can, yes, absolutely. If we didn’t have a VPN ability inside the FreedomBox, and if we weren’t offering it to you at your endpoint, then that would be a whole zone of privacy we weren’t even going after, so we go after it. But that doesn’t mean that you have to do it. If you want to run an Open VPN server, click. And if you don’t want to, don’t. We make you a Tor router, with one click, but we won’t force you to be a Tor router. There are parts of the world in which that would be a really bad idea. Want to run Radicale and do your own calendaring, fine. You want to use Google Calendar, go ahead and do it. But we’ll make it feasible for you not to. You like WhatsApp? OK you use WhatsApp. When you have the Riot client in the phone, and you have the Matrix server on the FreedomBox, and your question is, “Should I make my next Group on my own box, where nobody will know what’s going on, or should I make my next Group over at WhatsApp, where Facebook will get the traffic, even if they don’t get the content, maybe you’ll decide that the time has come to move your Group’s private to your own server. From my point of view, that’s the right answer, right. My law practice is held together by IRC, but I don’t run it through somebody else’s IRC server. It runs on our metal. And my telephone – that’s all Asterisk running on bare metal, and so I don’t interact with other people’s telephone companies. Similarly here, decide for yourself how hardcore you want to get. Decide over time. Learn at your own pace. The tool is there to make it possible for you to have those choices. It’s not making them for you. And every single time it makes an option available to you, the option it is making available to you, is configured the way the security guy would want it configured, if you had one. So, the default way that any service comes up is the way you would want it to occur, and we’re trying to make sure that you’re not leaking by accident, in the same way that Tails is trying to make that true of you, at the client end. Tails plus FreedomBox – that’s a pretty strong IT environment for any human being, no matter what her threat model might be.
Questioner #2: I guess, I’m sorry to….
Answer: No, please.
Questioner #2: I guess what I’m asking too, though, OK, so you can use Google or Facebook with the FreedomBox. Does the FreedomBox make it any safer to use Google or Facebook or Amazon or…? You can use like Facebook on FreedomBox, but does it make it safer?
Answer: I’m not sure that it would be fruitful for me to enter into debate about this, at the level of my own personal feelings, so I will only point out that if you’re routing your traffic through things that are trying to protect you, they will be a little bit of help to you in one or another way. But if you want my personal opinion, Facebook is unsafe at any speed, and always was, right. I never was there, because it was never attractive to me, for a moment, to turn my personal web traffic into an item in somebody else’s logs. I host my own web server, and I keep my own logs, for approximately 24 hours and sometimes less, and if you come knocking on my door, they’ll all go up in smoke immediately. To me, this is critically important. It’s true about my course Wikis, right. All my law school courses are taught off Wikis, and those Wikis are hosted on my hardware. And what happens to those logs is what I say happens to those logs, which is, I don’t believe in surveilled learning behaviour, so I dump it. Those of you who are Columbia students are acquainted with Canvas, an excellent proprietary software learning system, that spies the hell out of you. And there’s no way to say about this, one is a better kind of learning than another, until you make a qualitative inquiry, whether you think that surveilled learning is good. If you think surveilled learning is bad, if you think that’s not actually a good way to conduct law school, I would not have been capable of going to a surveilled law school, when I was a young man, that would not have occurred to me. If the books in the library were telling people I was reading them, I would have been worried about that. It seems odd to me, that students that I teach are not as worried about that as I would have been, but apparently the 1980s were, as Danny was saying, a very long time ago. 1984 was a long time ago. Absolutely true.
So, part of my difficulty in the question is that, I know that you and I are yearning for two different things. I’m yearning to tell you, “Yeah, you swim in a tank of poison all the time. Why don’t you stop doing it?”And you’re yearning for me to tell you, “FreedomBox is not going to pull you out of that tank of poison, but it’s going to remove some droplets for you from the tank,” which will make you feel happy and virtuous about participating in the hackathon. And I don’t want to disappoint you. But there’s a part of me which wants to say, “Really? Really? You’re going to do that?” But then, only one of us was in the Soviet Union. And so, Mr. Zuckerberg means two different things to the two of us. And in between, is our comrade, friend, and hero, Edward Snowden, living in the birdcage of the Czar Vladimir, in the post-Soviet Kremlin, trying to deliver to us a message that, no matter what you do, the most powerful people in the world are going to climb inside. And if you allow people to take this stuff from you on a commercial basis, then you have to remember who gets it in the end. The United States is a small country, from that point of view. My comrades who live in the world of Narendra Modhi and Amit Shah, the question is, “What is going to happen to 1.3 billion people?” And it should be understood that, if Reliance collects it, RSS will have it. This is not what we want, right. Learn the tool however you want. Make whatever compromises you want. Know a lot about the tool, so that the next person you teach, if she or he wants to do it, even more completely, you will have enabled them. That’s the deal I would make with you, OK? Make whatever compromises you want about this, that suit your life, but know how to equip somebody who just wants out of The Matrix, because being able to teach people how to do that is a really valuable skill in the twenty-first century. It will save lives, and I’d like you to have it. Is that a workable bargain, in a way?
Questioner #2: Yeah, I was more… I was just curious… you know, I’m not that well-versed in most of this stuff. I’m a mathematician by trade, and tech is fairly new to me. So, all of this is illuminating. I think that explains, sort of, what I was curious about.
Answer: You see, think of the beauty of mathematics, which has never informed on the people who [inaudible]. Think of the extraordinary value that comes from the possession of a language which is capable of describing everything that happens in the universe, except the mind of the mathematician thinking it. We, we [gesturing to the university] live in a very privileged environment in the human race. The thoughts within our own minds that we most value are not the ones that other people are most trying to tear from us. Consider, in that sense, the relationship between who you are as a mind in the human race, and technology as a gift you have. Most of the human race is now being adjusted to a world in which the ideas that they value are the ones that are spied the crap out of – which are about what to eat, and where to go, and how to live, and how to make a living, and who you love, and who you hang out with, and what your politics are. And so, at their minds is aimed a spotlight which you and I have always been able to evade, even before it existed. And the stuff that we’re learning about, is stuff which is designed to give them a space around their heads, which is just as real and just as well-protected, as the space around your thinking about your particular kind of math, or my particular kind of thinking about medieval English contract law, or how to write a Supreme Court brief in Oracle against Google. We are accustomed to having deep value in the parts of our minds which escape all this rot, even now. And we occupy, in that sense, two different kinds of power. If tech is new to you, it isn’t new to you in the same way it is new to the people who have just acquired connection to the world through a smartphone, and have no idea what its downside might be. And you understand what it is to have a part of your mental life which is off-limits to all of that, and that’s a luxury you would never give up. It’s how we all come to live in universities, because our thought-lives are so important to us.
Good luck with this. I think you’re going to love it.
If there’s nothing else, I should get out of the way, and let the technology begin.
Good. Have a wonderful afternoon. Thank you very much.