Lisa Rein: Alright. Next up is Brewster Kahle. Brewster Kahle found­ed Aaron Swartz Day with me sev­er­al years ago. We decid­ed after the memo­r­i­al that some­thing had hap­pened there. The ener­gy there was incred­i­ble, and peo­ple kept say­ing we need to do some­thing like this again, not so sad, but hav­ing the ener­gy that’s here, the ded­i­ca­tion that’s here, and do some­thing with it. So that’s how things start­ed, and the rest is kind of his­to­ry. It just gets big­ger and big­ger every year. This is def­i­nite­ly the largest one, and Brewster’s talks every year are always very amaz­ing, so I’m look­ing for­ward to this one, too. 

Ladies and gen­tle­men, Brewster Kahle.

Brewster Kahle: Thank you Lisa very much for orga­niz­ing this for anoth­er year. And a bunch of Intenet Archive staff, admin staff that’ve been doing a bunch of the stuff behind the scenes. Thank you.

And wel­come to the Internet Archive. Those are some petabytes of servers behind you, serv­ing the world as we speak. The lit­tle stat­ues are made for the peo­ple, of the peo­ple that’ve been ded­i­cat­ing sev­er­al years of their lives at least to bring uni­ver­sal access to all knowl­edge.

Locking the Web open. We’ve spent the last 25 years pour­ing our lives into this open Internet. Our expres­sion of who we are, what we’ve done, all sorts of stuff has been going into the World Wide Web. It’s been fan­tas­tic. Can we lock the web open?

The way the Web is cod­ed deter­mines a lot of the way we live our lives online. As Larry Lessig, my friend and hero says, code is law.” So the way that we code the Web makes a huge dif­fer­ence. So does it have our val­ues in it? Freedom of expres­sion, uni­ver­sal access to all knowl­edge, pri­va­cy. Does it have that in it? Turns out, no. It’s actu­al­ly real­ly quite frag­ile, this World Wide Web that we’ve got. 

But it is huge. We know this because the Internet Archive col­lects about a bil­lion pages every week (we oper­ate the Wayback Machine) and the life of a web page is only a hun­dred days between the times it’s either changed or delet­ed. So they flip on and off servers, all the time.

Also, the World Wide Web isn’t reli­able. It’s mas­sive­ly avail­able, unless you hap­pen to live in China. Or cur­rent­ly in Russia you can’t get to the Internet Archive. Or in many of these coun­tries you can’t get to the New York Times or all sorts of oth­er things. So the Internet itself and the World Wide Web is not reli­ably acces­si­ble from places.

It also isn’t pri­vate. Something that I attribute to Ed Snowden was…not attribute, it was because of Ed Snowden’s revala­tions that there was an arti­cle in The Guardian that I found chill­ing. People that were going to WikiLeaks dur­ing the Cablegate times were being watched by GCHQ. And with IP address­es, that’s personally-identifiable in a lot of cas­es. They turned these over to the National Security Agency, that then used that to be able to spy on all of those peo­ple (even American cit­i­zens) because they were of height­ened inter­est.

Reader pri­va­cy, for the library world, is a very very impor­tant thing because we have a dread­ful his­to­ry of hav­ing peo­ple being round­ed up for what it is they’ve read and hav­ing bad things hap­pen to them. And we have this struc­ture now, on the World Wide Web.

But it is fun. So the World Wide Web is not very reli­able, it’s not very pri­vate, but it’s fun. We have all these wacko things start­ing to hap­pen, and you can go and change it and play with it and do all sorts of inter­est­ing things. So I think we’ve done one out of three of the big issues. It’s not reli­able, it’s not pri­vate, but it’s fun.

I’d sug­gest it’s time to fix the World Wide Web. I think we can get all three of these char­ac­ter­is­tics, and I’m going to sug­gest the way to do this is by build­ing a dis­trib­uted Web. This is a call to build a dis­trib­uted Web, to lock the Web open.

What do I mean by a dis­trib­uted Web? Isn’t the Web already dis­trib­uted? You know, there are servers all over the place. But it isn’t, in the sense that if any par­tic­u­lar piece of hard­ware goes out, that thing blips out. Or, if you stand in front of it, you can watch all traf­fic to it. Or if you stand in front of you, you can go and make sure that you can’t get to any­thing, or you can lim­it what you can have access to.

It’s not dis­trib­uted in the sense like the Internet is a dis­trib­uted sys­tem. So dis­trib­uted in the sense that any par­tic­u­lar pieces of hard­ware can be nuked, lit­er­al­ly, and the thing will con­tin­ue to work. It’s hard­er to build dis­trib­uted sys­tems than it is to build cen­tral­ized sys­tems. But they’re very use­ful if you do spend the time and effort to do it.

Let’s take anoth­er exam­ple of a dis­trib­uted sys­tem. Amazon​.com oper­ates data cen­ters all over the world, and you can go and use them to go and host your web sites or what­ev­er, and they move the ser­vices from machine to machine to get around hard­ware prob­lems. Or if they’re start­ing to be used more in a cer­tain area of the world, they migrate these ser­vices clos­er to peo­ple. So they repli­cate more and they can move around. What if we could make the World Wide Web on the open Internet oper­ate some­thing like an ama​zon​.com that is for every­one and for free?

And what if we could try to bring read­er pri­va­cy into this in a real sin­cere way? Reader pri­va­cy turns out to be hard­er than writer pri­va­cy. Going and mak­ing it so that it’s hard to go and find who’s read­ing what, it’s inter­est­ing because it’s a flip oppo­site in the phys­i­cal world. It used to be that pub­lish­ing a book was real­ly hard to do anony­mous­ly, but you can kind of sneak it around, put it in a [mum­bles, mimes shuf­fling objects around]. You could spread knowl­edge anony­mous­ly much more eas­i­ly in the phys­i­cal world, and it’s some­what flipped oppo­site now in the dig­i­tal world.

And this time around, if we’re going to rein­vent this World Wide Web, let’s put a time axis into it in the begin­ning. The Wayback Machine is real­ly kind of a kludge gone and put onto the World Wide Web after the fact. It’s basi­cal­ly a time axis for the World Wide Web, and it’s not even bun­dled into the browsers yet. So it’s not real­ly part of the World Wide Web yet, even though it’s been around for twen­ty years. So how do we go and build the next-generation World Wide Web to fix some of these prob­lems so that it’s more like a Git, if you will. So that there would be ver­sions, and that they would be dis­trib­uted in ways. 

I would also like to see if we can actu­al­ly fix some of the finan­cial prob­lems, We didn’t put in any real mech­a­nism for mak­ing any mon­ey on the web. Even if peo­ple want­ed to pay, it’s so clunky and hard that we couldn’t do this.

So, how? Okay, that sounds all very good. Well, there are a bunch of tools that Tim Berners-Lee didn’t have twenty-five years ago when he was build­ing the World Wide Web. For instance, the Javascript lay­er of the World Wide Web now is real­ly pret­ty impres­sive. I had my mind blown when Jason Scott and a bunch of col­lab­o­ra­tors made it so that you could emu­late IBM PCs and Macintoshes in Javascript in your brows­er, run­ning at speed. The idea that you could actu­al­ly take fif­teen, twen­ty year-old tech­nol­o­gy and run it in Javascript and dis­trib­ute it with Javascript was pret­ty amaz­ing to me. That means that we have a lay­er that we can dis­trib­ute smooth­ly and eas­i­ly that’s as pow­er­ful as, in a dis­trib­uted way, we could build and oper­at­ing sys­tem on top of it.

We also have strong cryp­to that wasn’t legal at the time to dis­trib­ute. We won that war, let’s use it. This blockchain idea of a dis­trib­uted sys­tem that can be used for nam­ing sys­tems and the like. And I’d sug­gest that Bitcoin is real­ly help­ful new tech­nol­o­gy that we can weave into our future sys­tems to make it so that if you author a par­tic­u­lar page, you’re going to have to sign that page, you might as well sign it with a Bitcoin address. 

So can we go and make it so that we can basi­cal­ly fix these prob­lems? I’d like to pose a goal. I would sug­gest we want WordPress-functionality web sites, but dis­trib­uted. Can we make it so that it’s real­ly easy and fun to go and put up a web site that has dif­fer­ent themes and maybe dif­fer­ent mod­ules in it, that has peo­ple with accounts that have dif­fer­ent roles? So there are admin­is­tra­tors and there’s edi­tors and com­menters, but it has these char­ac­ter­is­tics of being pri­vate and reli­able, served from many places in a peer-to-peer struc­ture such that the web site moves, the liv­ing web site moves. 

And I’d say it’s actu­al­ly pos­si­ble. There are peo­ple that are doing inter­est­ing parts of the prob­lem. There’s IPFS, there’s Maelstrom by Bittorent, there’s MaidSafe, Namecoin. There’s Bitcoin, of course, there’s proof of stor­age sys­tems that are start­ing to come out, Oceanstore, PeerJS/WebRTC.

All of these tech­nolo­gies are com­ing about. I’ve seen all of the pieces need­ed to build a dis­trib­uted Web. So I’d say it’s pos­si­ble to do. It takes a bunch of work. But to go and make it so that it’s not just wish­ful think­ing, I had a pro­gram­mer in Amsterdam take my blog from WordPress and dump it, and then try to make it into a peer-to-peer-served sys­tem. And I’m going to try to do some­thing that you’re real­ly not sup­posed to do. I’m going to try to do a live demo. 

So bear with me, and you’re going to have to applaud if this works. This is a gate­way address to IPFS. IPFS is a Bittorrent-like sys­tem. It took my web blog, all of the arti­cles and indices and also even a lit­tle search engine, and put it in… Let’s see if we can get it. So I double-click and it launch­es a web brows­er. [Kahle’s blog loads; audi­ence applauds]

So that wasn’t quite how I thought it was going to come up, but here is Locking the Web Open”. Here’s a blog post that’s work­ing. There’s recent posts, Tower of Babel Story,” and it’s being served off of a peer-to-peer sys­tem. So there is no one serv­er, there’s actu­al­ly a lot of servers that are serv­ing this par­tic­u­lar con­tent. Another thing that’s kind of cool is it has a search engine in it. So if I search on Stallman” it has search results that are dynam­i­cal­ly hap­pen­ing. So how is that hap­pen­ing? It’s a client-side app. It’s basi­cal­ly a lit­tle search engine and index that’s in Javascript that’s been down­loaded. So basi­cal­ly you can do active web sites, but you’re mov­ing them around. They’re basi­cal­ly self-contained to have all of the com­po­nents.

We’ve basi­cal­ly been able to fig­ure out how to at least show it’s pos­si­ble to go and take a liv­ing, dynam­ic web site. It doesn’t have the authen­ti­ca­tion sys­tems, it doesn’t have some of the oth­er com­po­nents to it, but it’s got sort of a Hello world” idea that you could go and make a dis­trib­uted Web and have it work.

So we’ve seen these pieces. The chal­lenge is now, can we go and weave this all togeth­er? Can we go and as a com­mu­ni­ty do some­thing actu­al­ly pret­ty hard? Can we go and fix the World Wide Web? Can we make it so that it’s archiv­able? Can we make it so that it’s privacy-enhanced? And still always keep atten­tion to mak­ing it fun. Can we lock the web open? Can we bake the First Amendment into the code itself? Can we make open­ness irrev­o­ca­ble? This I think is some­thing we can do, and we must do.

Thank you very much.

Further Reference

The Aaron Swartz Day web site.

Help Support Open Transcripts

If you found this useful or interesting, please consider supporting the project monthly at Patreon or once via Square Cash, or even just sharing the link. Thanks.