[00:58.120 --> 00:58.890]  Hello, everyone.
[00:59.220 --> 01:01.580]  Well, welcome to the last talk of the night.
[01:01.960 --> 01:03.420]  I hope you enjoy it.
[01:04.200 --> 01:05.840]  We're super excited to have it here.
[01:06.090 --> 01:07.820]  Thank you so much for coming to the show.
[01:08.170 --> 01:09.900]  You know, it's been great to have you here.
[01:10.000 --> 01:12.950]  It's been so wonderful to have this as a live show again this time.
[01:13.540 --> 01:16.100]  So, again, thank you.
[01:16.200 --> 01:18.340]  We couldn't do this without your attendance, without your participation.
[01:18.820 --> 01:20.760]  Thank you to all the volunteers who are here.
[01:20.980 --> 01:22.720]  Thank you so much for all your time and all your assistance.
[01:23.000 --> 01:24.720]  You're the ones who helped make this work.
[01:25.700 --> 01:30.400]  Thank you for all the staff and for Manuel for making all of this happen.
[01:31.060 --> 01:33.620]  And with that, a couple of quick announcements.
[01:33.800 --> 01:39.430]  We have the simulcast of the closing ceremonies that are from here at 6 p.m.
[01:40.160 --> 01:44.060]  They'll be simulcast in room 206, not in a little theater like it's programmed in the schedule.
[01:44.300 --> 01:51.740]  So you can either listen to it here or you can go down to 206 on the second floor in the back room and watch the closing ceremonies there.
[01:52.840 --> 01:55.900]  The 2600 store will be open during the closing ceremonies.
[01:56.120 --> 02:00.900]  So it's your last chance to get swag and gifts for your families, friends, and loved ones.
[02:03.220 --> 02:07.600]  We will need some help with loadout for anyone who is around tomorrow morning.
[02:07.840 --> 02:12.440]  Talk to the info desk or talk to any of the crew staff and let them know that you're interested in helping.
[02:13.060 --> 02:15.650]  And with that, let's jump into our last talk.
[02:16.210 --> 02:22.270]  It is from Cory Doctorow, renowned author, activist, and journalist who cares deeply about how technology and impacts people.
[02:22.680 --> 02:33.560]  Some of his recent works include Little Brother, how to destroy surveillance capitalism, the graphic novel in real life, and the soon to be released Checkpoint Capitalism coming out in September.
[02:34.050 --> 02:40.660]  His talk is titled Seize the Means of Computation, How Interoperability Can Take the Internet Back from Big Tech.
[02:40.860 --> 02:47.480]  And it focuses on what big tech fears the most, technology operated by and for the people who use it.
[02:47.900 --> 02:52.900]  If you have any questions during Cory's chat talk, you can submit them to the dedicated Matrix chat channel.
[02:53.100 --> 02:57.480]  We will ask as many of them as time will allow during the QA slot at the end of the session.
[02:57.720 --> 02:59.140]  And now over to you, Cory.
[03:00.680 --> 03:02.060]  Well, thank you very much.
[03:02.280 --> 03:03.800]  What a treat to be here.
[03:04.040 --> 03:10.600]  I was virtually present for the last one of these hopes, and I had very much hoped to be there in person this time.
[03:10.950 --> 03:19.060]  However, it overlapped with the first family vacation I've had with my distant relatives, the first chance to see my niece and nephew in three years.
[03:19.180 --> 03:21.600]  So I hope you'll agree that that was a good reason for the first time I've had.
[03:21.600 --> 03:23.040]  To miss this.
[03:23.420 --> 03:26.740]  And I did spend the last two weeks on a beach preparing.
[03:26.960 --> 03:32.840]  I read my entire backlog of 2600 magazines that had piled up, but that I hadn't had a chance to read.
[03:33.010 --> 03:39.040]  So I'm fresh off reading literally two years worth of 2600s back to back and ready to go.
[03:39.200 --> 03:43.620]  And looking forward to seeing as many of you as I can in person the next time one of these happens.
[03:43.620 --> 03:46.510]  Because I really do enjoy hope.
[03:46.510 --> 03:52.360]  And I'm jealous of you all there in person, exchanging microbes with one another.
[03:52.980 --> 04:01.420]  So this is not a talk for people who want to, or rather, this is a talk for people who want to destroy big tech.
[04:01.670 --> 04:05.510]  This is not a talk for people who want to tame big tech.
[04:05.510 --> 04:08.300]  There is no fixing big tech.
[04:08.730 --> 04:13.900]  It's also not a talk for people who want to get rid of technology itself.
[04:14.420 --> 04:16.460]  Technology is not the problem.
[04:16.860 --> 04:26.840]  Stop thinking about what technology does and start thinking about what technology, about who technology does it to and who technology does it for.
[04:26.840 --> 04:31.900]  This is a talk, as you heard, about the thing that big tech fears the most.
[04:32.300 --> 04:37.020]  Technology operated by and for the people who use it.
[04:38.180 --> 04:40.980]  Today's tech barons are not evil geniuses.
[04:41.120 --> 04:42.560]  They're not evil.
[04:42.900 --> 04:48.120]  Their ambitions to world domination are not novel, nor are they especially grandiose.
[04:48.120 --> 04:59.120]  The founders of AltaVista and DEC and Sun Microsystems and Commodore all talk the same line about taking over the world, about being in charge of our future.
[04:59.460 --> 05:02.560]  The difference is, all of those guys, they didn't get it.
[05:03.000 --> 05:06.060]  But today's tech leaders, they're also not geniuses.
[05:06.440 --> 05:18.460]  The reason the tech industry spent generations churning as new companies and new systems supplanted the old was that in the olden days, we enforced competition rules, and that made sure that this could happen.
[05:18.820 --> 05:22.720]  Like, we used to ban companies from buying their competitors.
[05:22.800 --> 05:26.020]  We used to ban them from creating vertical monopolies.
[05:26.440 --> 05:28.980]  The new crop of leaders aren't being displaced.
[05:28.980 --> 05:30.160]  It's true, right?
[05:30.320 --> 05:36.000]  Mark Zuckerberg has been on the top of the dog pile for a lot longer than the founders of Silicon Graphics.
[05:36.140 --> 05:40.620]  But that's not because of his incredible leadership and vision.
[05:40.620 --> 05:53.320]  It's because 40 years ago, we shot antitrust law in the guts, and we let companies led by mediocre idiots, no better than their forebears, establish monopolies.
[05:53.620 --> 06:03.780]  These donkeys are able to parlay those monopoly winnings into policies that prevent new technologies and new companies from supplanting theirs.
[06:04.260 --> 06:08.760]  They get to decide who can compete with them and how they can compete.
[06:10.320 --> 06:14.760]  Notably, today's tech giants are able to wield the law against interoperators.
[06:15.600 --> 06:21.380]  That's the new technologies that plug into their existing services, systems, and platforms.
[06:21.660 --> 06:27.040]  And that is a privilege that none of yesterday's easily toppled giants had.
[06:27.520 --> 06:34.200]  You know, IBM did not like the fact that its competitors made software and printers and keyboards and storage for its mainframes.
[06:34.200 --> 06:41.980]  But to stop them from doing that, it had to try to build a computer that no one else could reverse engineer and improve on.
[06:42.300 --> 06:47.920]  And for complex reasons that I think this audience probably understands pretty well, this is impossible.
[06:47.920 --> 07:04.940]  The very bedrock of computer science, ideas named for mid-century computing demigods like Turing completeness and von Neumann machines, dictates that the creation of non-interoperable computers is a fool's errand.
[07:04.940 --> 07:06.480]  It's fantasy.
[07:06.800 --> 07:07.120]  It's fantasy.
[07:07.120 --> 07:07.940]  It's not science fiction.
[07:08.160 --> 07:11.380]  It's like a time machine or a faster-than-light drive.
[07:11.600 --> 07:12.620]  A thought experiment.
[07:12.840 --> 07:15.660]  Not something anyone's ever going to be able to build.
[07:16.420 --> 07:21.440]  Today's tech giants have not invented interoperability-proof computers.
[07:21.440 --> 07:29.700]  Instead, they've invented laws that make interoperability illegal, unless they give permission for it.
[07:30.080 --> 07:43.480]  They've created a new complex thicket of copyright, patent, trade secret, non-compete, and other IP rights that combined conjure up a new offense that we could think of as felony contempt of business model.
[07:43.480 --> 07:51.120]  The right of large companies to dictate how their customers, their competitors, even their critics can use their products.
[07:51.880 --> 07:56.200]  Now, why have tech firms gone to war against interoperability?
[07:56.380 --> 07:59.560]  Why do they fear interoperability so much?
[07:59.940 --> 08:01.300]  The answer is simple.
[08:01.560 --> 08:04.060]  It's all about switching costs.
[08:05.020 --> 08:14.240]  Whenever economists dather to hand-wave away the rise of big tech as historically inevitable, they make frequent reference to the idea of network effects.
[08:14.800 --> 08:19.920]  This is an economist term of art for a product that gets more valuable every time it wins a new customer.
[08:20.120 --> 08:23.340]  Like, you joined Facebook to chat with the people who are already there.
[08:23.340 --> 08:25.680]  They had made it valuable for you to join.
[08:25.880 --> 08:30.700]  And then once you joined, Facebook got more valuable because other people joined because you were there.
[08:31.300 --> 08:32.540]  Network effects are there.
[08:32.760 --> 08:36.400]  Network effects are indeed how big tech gets big.
[08:37.170 --> 08:41.460]  But switching costs are how big tech stays big.
[08:42.570 --> 08:43.480]  Switching costs.
[08:43.710 --> 08:48.600]  That's the stuff that you have to give up when you stop using a product or service.
[08:48.830 --> 08:52.760]  Like, if you quit Facebook, you lose the family photos you've uploaded.
[08:52.760 --> 08:58.140]  You lose access to the friends, the family, the community, the customers you go there to hang out with and talk to.
[08:58.980 --> 09:07.070]  When switching costs are high enough, people keep using products and services, even when they hate the products and services.
[09:07.380 --> 09:12.900]  So long as the pain of staying is less than the pain of leaving, the users stay.
[09:14.200 --> 09:26.960]  So this is not actually in my speech notes, but it's just occurred to me that this is a way to think about the shibboleth or the idea that our technology is addictive.
[09:27.310 --> 09:29.840]  I understand why people say technology is addictive.
[09:30.200 --> 09:40.980]  After all, if someone is using Facebook every day and talking about how much they hate Facebook, they kind of sound like a smoker who's talking about how much they hate cigarettes while lighting another one.
[09:40.980 --> 09:45.680]  And so maybe you could call this addictive, but I don't think addiction is the right model here.
[09:45.800 --> 09:47.840]  I think you can understand this as switching costs.
[09:48.180 --> 09:50.540]  Like, addiction is a kind of switching cost.
[09:50.720 --> 09:56.900]  You know, speaking as someone who used to smoke a lot and who gave it up, there's a really high switching cost to giving up cigarettes, right?
[09:56.920 --> 10:04.420]  It's not just, you know, losing the instant ability to meet friends out by the dumpster where they banish you to have a cigarette.
[10:04.540 --> 10:08.340]  It's also, you know, the physical withdrawal from not smoking.
[10:08.340 --> 10:10.360]  Those switching costs are really high.
[10:10.660 --> 10:13.400]  And it hurts to quit smoking.
[10:13.400 --> 10:15.020]  And it hurts to quit Facebook.
[10:15.020 --> 10:16.660]  Not because Facebook is great.
[10:16.900 --> 10:20.400]  And, you know, quitting cigarettes doesn't hurt because cigarettes are great.
[10:20.560 --> 10:27.520]  It hurts because the people you leave behind are stuck there and you lose some of your social relation to them.
[10:27.660 --> 10:29.360]  It's that high switching cost.
[10:29.360 --> 10:36.180]  I think that the narrative that technology is addictive serves technology companies, right?
[10:36.200 --> 10:44.680]  If you're going out there and pitching a venture capitalist or private equity fund and you can say, look, I'm going to invent something as addictive as a jewel pod.
[10:45.360 --> 10:47.320]  It's going to be in social media form.
[10:47.780 --> 10:51.960]  Then, yeah, they're going to want to invest in you because they like that idea.
[10:52.460 --> 10:55.340]  But it's not really that they're making it addictive.
[10:55.500 --> 10:59.880]  It's that they're trapping the people you want to talk to inside a walled garden.
[11:00.460 --> 11:01.940]  So that was all ad lib.
[11:02.080 --> 11:03.180]  It wasn't in my talk notes.
[11:03.260 --> 11:04.100]  I hope it makes sense.
[11:04.220 --> 11:07.200]  Normally, I practice this stuff, but I just had this idea as I was talking to you.
[11:07.200 --> 11:09.720]  I'm going to go back to the regularly scheduled talk here.
[11:09.880 --> 11:11.920]  They told me I had longer than I thought I would.
[11:12.040 --> 11:15.020]  So I'm a little giddy with the power of being able to run along here.
[11:15.360 --> 11:22.060]  So when switching costs are high enough, people keep using products and services, even though they hate the products and services.
[11:22.320 --> 11:27.740]  So long as the pain of staying is less than the pain of leaving, the users will stay.
[11:28.300 --> 11:36.120]  That means that when companies can raise their switching costs, they also get to treat their customers worse and still keep their business.
[11:36.120 --> 11:49.400]  That is to say, they can shift value from the customers to their shareholders without worrying about the customers leaving because the customers will have to pay such a high price to go that they'll stay even when they think it's unfair.
[11:50.240 --> 11:59.080]  And the corollary of this is that if you can make switching costs lower, you force companies to treat you better if they want to keep your business.
[11:59.080 --> 12:12.680]  You know, if the users who quit Facebook, who resigned their account and deleted the app could still talk to their Facebook friends through another service, a rival service that could interchange messages with Facebook, Facebook would be in serious trouble.
[12:13.300 --> 12:17.520]  And interoperability is the way that we lower switching costs.
[12:18.320 --> 12:25.340]  Interoperability is what allows us, the users of technology, to set the terms on which we use that technology.
[12:25.800 --> 12:34.000]  It allows us to use the parts of the products and services that benefit us and also block the parts that hurt us, that don't benefit us, that we don't like.
[12:34.560 --> 12:38.240]  There are a lot of things we should do to fix big tech.
[12:38.480 --> 12:40.260]  We should change the rules for mergers.
[12:40.420 --> 12:41.960]  And actually, the FTC is doing that.
[12:42.180 --> 12:44.700]  We should pass comprehensive privacy legislation.
[12:44.700 --> 12:46.460]  That looks like it might be dead in Congress.
[12:46.600 --> 12:49.040]  We should ban deceptive dark practices.
[12:49.260 --> 12:53.600]  And of course, we should break up these giant companies into smaller competing firms.
[12:53.760 --> 12:56.720]  But that stuff, it takes a long time.
[12:56.880 --> 13:00.000]  Like, I think we underappreciate just how long it takes.
[13:00.000 --> 13:04.380]  Like, we think about AT&T as being broken up after a seven-year lawsuit.
[13:04.560 --> 13:08.700]  That was just the latest attempt by the DOJ to break up AT&T.
[13:08.820 --> 13:19.520]  If you start from the first antitrust action filed against the Bell system to the day it was broken up in 1982, that was a process that took 69 years.
[13:20.100 --> 13:23.760]  We don't have 69 years to fix the Internet.
[13:24.280 --> 13:28.000]  By contrast, interoperability is immediate.
[13:28.000 --> 13:41.920]  If you make it legal for new technologies to plug into existing ones, that is, if you make it legal to blast holes in every walled garden, the users, that's us, we get immediate profound relief.
[13:42.240 --> 13:49.740]  Relief from manipulation, from high-handed moderation, from surveillance, from price gouging, from disgusting or misleading algorithmic suggestions.
[13:49.740 --> 13:52.780]  That whole panoply of technology sins.
[13:53.640 --> 14:03.460]  Now, I'm going to go into what I mean by interoperability, the different kinds of interoperability, from widely adopted formal standards to the guerrilla warfare of reverse engineering.
[14:03.640 --> 14:07.920]  And I'll touch on the legislative and judicial history of the war on interop.
[14:08.040 --> 14:16.360]  And I'll describe a diversity of tactics, commercial, legal, technological, and social, that we can use to foster interoperability.
[14:16.360 --> 14:26.020]  I'll explain how a well-constructed interoperability policy can be sturdy enough to beat the attacks against that, whether they are legal or technical or social or commercial.
[14:26.680 --> 14:28.560]  These are shovel-ready ideas.
[14:28.680 --> 14:38.740]  These are a means to dismantle big tech's control over our digital lives and devolve control to the people who suffer the most under the hegemony of big tech.
[14:38.740 --> 14:45.940]  The marginalized users, the low-level tech workers, and the people who live downstream of tech's exhaust plume.
[14:46.060 --> 14:58.580]  The people choking on toxic waste from the tech industry, and the people living under dictatorships, where control is maintained with off-the-shelf cyber weapons used to hunt and neutralize opposition figures.
[14:59.400 --> 15:03.260]  So, let's start with how big tech got big.
[15:03.480 --> 15:09.080]  And to understand how big tech got big, we have to understand the history of competition law in the U.S.
[15:09.780 --> 15:20.820]  So, in the U.S., the first competition laws to get passed, the Sherman Antitrust Act in 1890, they were passed on the idea that there was something dangerous about companies being big.
[15:21.160 --> 15:28.800]  That once a company attained a certain scale, a certain degree of profitability, it became effectively impossible to regulate them.
[15:29.020 --> 15:31.280]  That they became something like a king.
[15:32.020 --> 15:47.000]  And, you know, when John Sherman, the senator whose name is on the Sherman Act, the first antitrust law, when he was trying to convince his fellow senators to pass the law, he gave a speech where he said, we wouldn't suffer a king to tell us how to live our lives.
[15:47.060 --> 15:51.920]  We should not suffer a king of industry who tells us how to live our lives.
[15:51.920 --> 15:58.020]  And so this harmful dominance idea became the pole star of American competition law.
[15:58.340 --> 16:02.340]  Something that was used to break up the standard oil company, John D.
[16:02.500 --> 16:15.660]  Rockefeller's empire of both oil refineries and oil pipelines and oil wells, but also the banks and the railroads and the shipping lines that did all of the oil marketing.
[16:15.660 --> 16:18.960]  All of that stuff was under one company's control.
[16:18.960 --> 16:24.500]  And he used that to structure really control over what amounted to the destiny of the nation.
[16:24.820 --> 16:36.240]  And so it was the Sherman Act and its successors like the Clayton Act that were used to break up the standard oil company, not because it was bad at making oil, not because it didn't make oil cheap.
[16:36.240 --> 16:43.960]  There was lots of evidence that in places where the standard oil company reigned supreme, they could create all kinds of efficiencies and lower the cost of oil.
[16:44.160 --> 16:50.980]  But because it was dangerous to let one unelected, unaccountable man, this king of industry, John D.
[16:51.120 --> 16:59.860]  Rockefeller, often drawn in the editorial cartoons of the day, wearing a crown, it was dangerous to let him rule, that it was antithetical to democracy.
[16:59.860 --> 17:06.620]  And that was the basis for American antitrust law until the late 1970s.
[17:06.840 --> 17:10.750]  It actually became the basis for competition law all around the world.
[17:11.320 --> 17:14.340]  You may know that after World War II, the U.S.
[17:14.440 --> 17:27.600]  went to Europe and under the guidance of the Marshall Plan, helped draft the new laws for postwar Europe, which included laws that looked a lot like American antitrust laws and held to the same standard.
[17:27.600 --> 17:31.340]  This idea that dominance itself was harmful.
[17:31.820 --> 17:42.140]  And the thing about harmful dominance as a standard is it really gave standing to everybody who might want to talk about how a company's behavior was hurting it, right?
[17:42.160 --> 17:54.800]  If you were a worker at that company where it was the only employer in town and you had to accept bad working conditions or low pay because there was nowhere else to go, you had standing under this harmful dominance standard to ask a regulator to intervene.
[17:54.800 --> 18:06.180]  Same if you lived in the town and its pollution was harming you and you couldn't get the town council to do something about it because as the only company in town, the town council was under its thumb.
[18:06.380 --> 18:07.560]  You also had standing.
[18:07.680 --> 18:09.060]  You had standing as its customers.
[18:09.320 --> 18:10.880]  You had standing as its suppliers.
[18:10.880 --> 18:18.000]  If they said, we're going to pay you less for your inputs to our manufacturing process than it costs you to make them.
[18:18.160 --> 18:25.180]  And since we're the only game in town, you're going to have to suck it up and find some way to make a go of it.
[18:25.300 --> 18:28.900]  You also had standing to seek redress for this.
[18:29.160 --> 18:31.820]  And so this became the model for a long time.
[18:32.000 --> 18:40.460]  It meant that any time a company got too big for its britches, any of the people who its dominance harmed was able to show up and ask a regulator to do something about it.
[18:40.660 --> 18:42.790]  And then came the 1970s.
[18:42.820 --> 18:50.600]  And in the 1970s, we saw the rise of what had been a very fringe theory of antitrust suddenly come to the fore.
[18:51.030 --> 18:54.320]  This theory was promulgated by a jurist called Robert Bork.
[18:54.880 --> 18:56.820]  A lot of people don't remember him these days.
[18:56.980 --> 19:08.440]  He was Nixon's solicitor general and was so complicit in Nixon's crimes that when Reagan tried to make him a Supreme Court judge, he completely blew the confirmation hearing.
[19:08.600 --> 19:12.000]  If you've ever heard the term to be Borked, that's what it's about.
[19:12.200 --> 19:19.790]  It's about Robert Bork showing up in front of the Senate and just like covering himself in shame and trying to defend his actions under Nixon.
[19:20.420 --> 19:29.580]  Robert Bork didn't make it onto the Supreme Court, but he actually had a much bigger effect on our society than he would have if he'd merely served in the highest court for the rest of his life.
[19:29.740 --> 19:43.530]  And that was because he was able to promulgate a weird legal theory that he had come up with at the University of Chicago School of Economics, which is really the kind of cradle of modern hypercapitalism, what we call neoliberalism.
[19:43.860 --> 19:48.100]  And this theory, it's called the consumer welfare theory of antitrust.
[19:48.100 --> 19:53.100]  And it's both an economic theory and kind of a conspiracy theory.
[19:53.340 --> 20:09.440]  He said that although the text of the Sherman Act and the Clayton Act and the FTC Act and the other antitrust laws, although these were really like obviously about harmful dominance, and although the people whose names were on them, like Senator Sherman,
[20:10.800 --> 20:13.060]  stood in front of the entire U.S.
[20:13.180 --> 20:20.320]  Senate and gave speeches explaining what he meant by his law, that actually we'd misread the law all this time.
[20:20.540 --> 20:29.760]  That the true meaning of antitrust, like the true meaning of Christmas, the true meaning of antitrust was to promote consumer welfare.
[20:29.760 --> 20:45.140]  And what he meant by consumer welfare was that he felt that it was important that America make policy that fought monopolies only if they would raise prices or lower quality.
[20:45.380 --> 20:59.220]  That unless a monopoly was raising prices or lowering quality, that not only should we keep our mitts off of it, not bring antitrust action against it, but we should actually celebrate it as a kind of efficient system.
[20:59.960 --> 21:02.240]  You may recognize that rhetoric, right?
[21:02.400 --> 21:04.340]  Why do people say Amazon is good?
[21:04.420 --> 21:06.360]  Well, look at how efficiently it ships things to you.
[21:06.420 --> 21:08.900]  Why do people say Apple's walled garden is good?
[21:08.940 --> 21:11.120]  Look at how efficiently everything just works together.
[21:11.320 --> 21:25.240]  That was the idea, that if we allowed monopolies free reign, that the people who ran them, these extraordinary geniuses who had the great foresight to build these empires of the sort that John D.
[21:25.360 --> 21:39.530]  Rockefeller had been unjustly deprived of with the breakup of the Standard Oil Company, that we could reap the benefits as consumers in the form of cheaper goods that were of higher quality.
[21:39.900 --> 21:45.840]  And that it was far more common that a monopoly was good than it was bad.
[21:46.000 --> 21:53.680]  And if we ever tried to do something about a monopoly, we would probably end up shooting ourselves in the foot by undermining this consumer welfare.
[21:54.500 --> 21:57.860]  Now, on its face, consumer welfare makes a certain kind of sense.
[21:57.980 --> 22:03.520]  You can kind of see how people would say, all right, well, yeah, it's an empirical standard.
[22:03.760 --> 22:07.100]  If so long as like a company isn't raising prices, we'll leave it alone.
[22:07.200 --> 22:08.860]  And when it does raise prices, we will.
[22:08.860 --> 22:30.100]  But not so fast, because when Bork and his pals at the University of Chicago School of Economics talked about raising prices following a merger or following an acquisition or after some anti-competitive conduct, they didn't mean, you know, one company buys its major rivals and the price goes up.
[22:30.100 --> 22:42.380]  They said, well, you know, sometimes a company buys its major rival and the price goes up because oil went up or labor costs went up or other key inputs went up or, you know, like the moon is in Venus or whatever.
[22:42.580 --> 22:46.860]  Like there could be some exogenous reason, some external reason that drove up the prices.
[22:46.860 --> 22:57.500]  He said, you need to make sure that if prices go up following a merger or some anti-competitive conduct or an acquisition, you need to make sure that it was because of that.
[22:57.680 --> 23:12.480]  And the only way to be sure is to build these very complex, ornate, very abstract mathematical models that were invented at the University of Chicago and that only economists from the University of Chicago knew how to make and that they would make for money,
[23:12.700 --> 23:13.100]  right?
[23:13.100 --> 23:23.100]  If you were going to buy your major competitor, you could hire a University of Chicago economist to build a model that proved that it would make prices go up.
[23:23.380 --> 23:36.340]  And then if after buying your major competitor, you doubled your prices, you could hire that economist again to build another model to prove that the prices went up for some other reason and that you should be left alone.
[23:36.640 --> 23:40.960]  So this was kind of a kind of learned helplessness for antitrust regulators.
[23:40.960 --> 23:53.500]  That really what they did was they kind of transformed themselves, these University of Chicago economists, transformed themselves into kind of court sorcerers for the American business lobby.
[23:53.900 --> 24:13.080]  Where if you were upset that a merger was taking place or that a company was doing something anti-competitive or that after a merger prices went up or the company started abusing its suppliers or its workers, you could go and petition for redress in front of the regulator,
[24:13.240 --> 24:17.100]  but the court sorcerer would be there, this University of Chicago economist.
[24:17.420 --> 24:33.780]  And you would give your case, you'd make your case, and then he would drag a goat into the court and he would slice it open, spread its guts out on the floor of the court before the king and the courtiers, and he would stare at the guts and he would stroke his beard and he would say,
[24:33.900 --> 24:38.680]  you know, I've looked in these guts and I have to tell you that you're wrong.
[24:39.900 --> 24:41.060]  Everything is fine.
[24:41.620 --> 24:46.940]  And the thing you're seeking, this neutering of this corporate behemoth, it wouldn't help you.
[24:47.340 --> 24:55.780]  And if you had the temerity to peer over this court sorcerer's shoulders and say, you know, I'm looking at these guts and I just don't see that.
[24:55.780 --> 25:00.800]  He could go, ah, look who thinks he knows how to read the guts of a goat.
[25:00.960 --> 25:03.860]  Did you go to the University of Chicago School of Economics?
[25:04.500 --> 25:09.740]  So as a result, you saw 40 years of antitrust forbearance.
[25:10.140 --> 25:12.420]  Ronald Reagan was in love with this theory.
[25:12.660 --> 25:13.920]  So was Margaret Thatcher.
[25:14.160 --> 25:20.800]  So were European neoliberals of the day, like Helmut Kohl, the Canadian leader, Brian Mulroney.
[25:20.960 --> 25:29.200]  They all adopted this idea that the true meaning of antitrust was preserving consumer welfare, which is to say, doing nothing.
[25:29.820 --> 25:36.860]  And doing nothing for 40 years about monopoly formation and maintenance led to monopolies.
[25:37.120 --> 25:38.080]  Surprise, surprise.
[25:38.660 --> 25:45.820]  Today, practically every industry you care to name is concentrated into four or five hands.
[25:46.140 --> 25:49.000]  You know, all the beer you drink is made by one of two companies.
[25:49.140 --> 25:51.980]  All the spirits you drink is made by one of two companies.
[25:51.980 --> 26:10.800]  All the wrestling you watch is run by one company, you know, founded and run by this kind of rapey weirdo, just had to resign, who bought all of his competitors and then reclassified his employees, the performers who do the work of wrestling, as contractors.
[26:10.800 --> 26:23.040]  So he could take away their medical insurance, which is why if you go on GoFundMe, all the wrestlers you grew up watching are now they're begging for pennies to help cover their medical bills so they can die with dignity of their work-related injuries.
[26:23.220 --> 26:27.560]  In their 50s, every industry looks like the tech industry.
[26:27.840 --> 26:35.420]  There's one company that makes all the eyeglasses, a company called Luxottica Essilor, a merger of a giant French company and a giant Italian company.
[26:36.080 --> 26:40.880]  Every lens you've ever had in your glasses, chances are it was made in an Essilor lab.
[26:40.960 --> 26:43.060]  They make more than 50% of the lenses in the world.
[26:43.460 --> 26:47.420]  Every frame you've ever had probably came from Luxottica.
[26:47.600 --> 26:51.620]  They own every brand from Coach to Dolce & Gabbana to Bausch & Loam.
[26:52.040 --> 26:56.260]  They own every retailer you might have shopped at.
[26:56.260 --> 26:57.560]  They own Sunglass Hut.
[26:57.720 --> 26:59.600]  They own LensCrafters.
[27:00.000 --> 27:02.620]  And also, they own the largest insurer.
[27:02.840 --> 27:14.440]  So even if you went to like a hipster indie optometrist who like gnawed your glasses frames out of a log with his own teeth, the insurance bill was covered by Luxottica Essilor.
[27:14.560 --> 27:15.680]  You can't escape them.
[27:15.760 --> 27:19.620]  And of course, they've raised the prices on glasses 1000% in the last 10 years.
[27:19.800 --> 27:21.000]  They stole your eyes.
[27:21.240 --> 27:22.520]  And it's not just them.
[27:22.900 --> 27:24.240]  Finance industry looks like that.
[27:24.240 --> 27:26.940]  Shipping is run by four giant cartels.
[27:27.180 --> 27:32.000]  And you know, when a company is that big, harmful dominance, no one can tell them what to do.
[27:33.520 --> 27:35.980]  Because they're so big that they just ignore you.
[27:36.280 --> 27:46.800]  And so for like 15 years, regulators and experts have been saying, you know, you keep building these ships bigger and bigger because you get these economies of scale on diesel fuel when you make the container ship bigger.
[27:47.080 --> 27:50.380]  But eventually, one of these is going to get stuck in the Suez Canal.
[27:50.760 --> 27:53.500]  And they were like, what do you know about shipping?
[27:53.500 --> 28:07.580]  So much of the dysfunction in our lives boils down to four incredibly powerful people, usually dudes, run the entire industry, and they don't listen to anyone.
[28:07.580 --> 28:09.420]  They just do what's good for them.
[28:10.280 --> 28:15.040]  So yeah, on the one hand, tech is not exceptional.
[28:15.040 --> 28:22.820]  Tech, the structure of tech, the economics of tech, this winner-take-all function of tech.
[28:22.820 --> 28:24.840]  It looks like the running shoe market.
[28:24.840 --> 28:26.560]  It looks like the beer market.
[28:26.720 --> 28:28.580]  It looks like the cheerleader uniform market.
[28:28.580 --> 28:30.280]  It looks like the horse show market.
[28:30.460 --> 28:32.940]  It looks like every other industry that has been cornered.
[28:33.060 --> 28:36.640]  But tech is different in some really important ways.
[28:37.100 --> 28:41.460]  Tech is different because it has network effects.
[28:41.880 --> 28:44.360]  Now, lots of industries have network effects, but not like tech.
[28:44.660 --> 28:46.520]  Tech has network effects on steroids.
[28:46.520 --> 28:54.100]  Like if you're old, you might remember like one person sent a Hotmail message to India that said like, get a free email account.
[28:54.320 --> 28:59.780]  And then like within a couple of weeks, some gigantic number of people in India were on Hotmail.
[28:59.940 --> 29:01.080]  Tens of millions of people.
[29:01.240 --> 29:05.080]  It was basically everyone had an Internet account in India had Hotmail within a very short time.
[29:05.080 --> 29:07.160]  That was the network effect in action, right?
[29:07.360 --> 29:10.200]  And, you know, it's kind of amazing.
[29:10.460 --> 29:17.120]  It's great when the people you love show up in the services you use and make it better and you make it better for them.
[29:17.240 --> 29:20.680]  It's like a big pot of stone soup where we're all contributing something.
[29:20.860 --> 29:22.220]  We're all getting something back out of it.
[29:22.340 --> 29:26.380]  So tech has network effects in a way that like cheerleading uniforms don't.
[29:26.740 --> 29:33.140]  But tech also has intrinsically low switching costs.
[29:33.140 --> 29:46.260]  It has low switching costs because of Turing completeness, because of von Neumann machines, because the only computer we know how to make is the computer that can run every program we know how to write.
[29:46.760 --> 30:03.920]  And that means that every walled garden is built on sand, that given enough time, technologists can figure out how to tunnel through the walled garden, how to build something that plugs into it, how to make a word processing program that can read the old word processors files,
[30:04.180 --> 30:07.800]  how to build a network protocol that can plug into the existing network protocol.
[30:08.280 --> 30:10.860]  And, you know, tech companies, they get this.
[30:11.100 --> 30:16.900]  They understand that raising switching costs is their best way of growing.
[30:17.820 --> 30:20.040]  The FTC is suing Facebook.
[30:20.360 --> 30:21.220]  Thank goodness.
[30:21.580 --> 30:25.260]  And they just released a big tranche of documents about seven, eight months ago.
[30:25.260 --> 30:30.180]  They had released these documents before last summer, but then they re-released them without the redactions.
[30:30.440 --> 30:35.040]  And it has all these internal emails between Mark Zuckerberg and his lieutenants.
[30:35.600 --> 30:44.580]  And one of them is the product manager designing Google photos, who sends this email saying, we are going to design... or not Google photos, I beg your pardon, Facebook photos, of course.
[30:44.740 --> 30:52.980]  He says, we're going to design Facebook photos to make it a really good product, but not because he thinks people want a good product.
[30:53.140 --> 31:07.700]  He says, if we can lure people into putting their family photos on Facebook, but not give them any way to export them, we will make the switching costs so high that they'll stay on Facebook, even if, say, Google Plus is a better service.
[31:08.080 --> 31:17.840]  So the tech companies really understand that making switching costs high is the way that you shift value from users to shareholders.
[31:18.440 --> 31:23.560]  Now, there are lots of people who say, all right, then what we need to do is just make these companies behave.
[31:23.780 --> 31:27.520]  We need to create some rules about how their products can work.
[31:27.680 --> 31:29.440]  And yeah, all right, I get that.
[31:29.600 --> 31:30.940]  And there's places where that can work.
[31:31.040 --> 31:34.300]  But honestly, the answer to the machine is almost never the machine.
[31:34.520 --> 31:38.980]  And, you know, every attempt we've made to make big tech behave, just handed it more power.
[31:39.160 --> 31:48.140]  Like, when you deputize tech companies to solve the problems of tech companies, then you make it prohibitively expensive to compete with tech companies.
[31:48.320 --> 32:04.100]  If you say, for example, in order to fight the harassment that your platform enables, you have to hire a million moderators, then what you say is that if you want to make a new platform, one that is maybe more selective about who can use it, or that has different moderation guidelines,
[32:04.320 --> 32:10.220]  or that lets communities moderate yourself, you have to come up with enough money to hire a million moderators.
[32:10.220 --> 32:13.760]  You know, Facebook couldn't have started if that was the rule when it started.
[32:13.940 --> 32:15.880]  Twitter couldn't have started if that was the rule when it started.
[32:16.120 --> 32:18.900]  So all you do is you hand them perpetual dominance.
[32:20.080 --> 32:24.360]  And you give the tech giants a great excuse to block interoperability.
[32:24.620 --> 32:35.580]  Because if you say, look, you're on the hook for fraud on your platform, harassment on your platform, hate speech on your platform, any other kind of the host of evils, which are, you know, don't get me wrong, actual evils.
[32:35.580 --> 32:48.960]  Then what they'll say is, how are we going to fight terrorist content on our platform if third parties can connect to our platform and exchange messages with our users, and we can't block them?
[32:49.500 --> 33:03.320]  So every one of these rules, the terror regulation and NETS-DG and the European Union that are about stopping harmful content and filtering them, the European Union's copyright directive, which is supposed to impose copyright filters on all user-generated content platforms,
[33:03.320 --> 33:13.280]  Canada's lawful but awful proposals where platforms are going to be responsible for policing the speech of their users and taking down bad speech within a very short timeframe.
[33:13.680 --> 33:19.620]  All of these are destined to fail because there really aren't enough human moderators to make those calls.
[33:19.840 --> 33:25.740]  They're also destined to fail because we're not going to be able to solve this with machine learning and so-called artificial intelligence.
[33:25.740 --> 33:29.680]  But what they will do is reinforce the dominance of these firms.
[33:29.680 --> 33:30.860]  It's not going to weaken them.
[33:30.900 --> 33:44.600]  It's going to create a capital moat around them where you can't enter the industry unless you can afford to undertake these kind of cosmetic but very expensive measures that are supposed to weaken tech, but in fact, strengthen it.
[33:45.100 --> 33:48.160]  So what does weaken big tech?
[33:49.520 --> 33:50.160]  Interoperability.
[33:51.200 --> 33:55.980]  Interoperability, that feature that is intrinsic to computers because they are universal machines.
[33:56.780 --> 33:59.840]  So here's an example of how big tech got weakened.
[33:59.980 --> 34:00.900]  A historic example.
[34:00.940 --> 34:04.680]  I think a lot of us probably remember, not just people who are as old as me.
[34:05.960 --> 34:12.760]  If you remember in the early 2000s, there was a moment when it really looked like Apple was about to go extinct.
[34:12.980 --> 34:14.890]  The Mac was going to be a dead platform.
[34:15.240 --> 34:17.140]  I was a CIO back then.
[34:17.410 --> 34:22.140]  I basically ran the technology for a lot of small and medium-sized companies.
[34:23.040 --> 34:25.140]  And, you know, they would have LANs.
[34:25.670 --> 34:28.720]  And on the LAN would be a mix of Windows and Mac machines.
[34:28.950 --> 34:30.000]  Mostly Windows machines.
[34:30.160 --> 34:31.930]  Usually we'd have a Mac for the designer.
[34:32.180 --> 34:35.910]  Maybe the CIO could say, well, I like using my Mac and I'm going to keep it.
[34:36.080 --> 34:41.870]  But for the most part, you know, the accountants or the engineers, the other people in the firm, they would end up using Windows machines.
[34:42.080 --> 34:43.640]  That was the way that it worked.
[34:43.680 --> 34:45.580]  It was what they were familiar with and it's what they liked.
[34:45.580 --> 34:55.220]  And there was a real problem with Windows machines and a Mac network, or rather Macs and Windows machines sharing the same network in the same work environment.
[34:55.390 --> 34:56.950]  And it was Microsoft Office.
[34:57.700 --> 35:04.260]  Microsoft Office did exist for the Mac in the early 2000s, but it was like the most cursed piece of software Microsoft ever released.
[35:04.560 --> 35:11.640]  If you were a Mac user and a Windows colleague sent you a Word file or an Excel file or a PowerPoint file, chances were you couldn't open it.
[35:11.740 --> 35:14.720]  But if you could and you saved it again, they wouldn't be able to open it.
[35:14.720 --> 35:16.320]  In fact, maybe you couldn't open it again.
[35:16.480 --> 35:20.500]  Like that file would be forever cursed and off limits.
[35:21.060 --> 35:24.120]  Abandon hope all ye who hit the floppy disk icon.
[35:24.960 --> 35:32.320]  And, you know, it got to the point where we just said like, we're going to have to buy the designers a Windows machine, just some low powered system.
[35:32.500 --> 35:34.740]  We're going to stick on their desk next to their Mac.
[35:34.910 --> 35:39.560]  So the Mac will be what they do their work with, but the Windows machine will be how they exchange documents with their colleagues.
[35:39.930 --> 35:43.200]  And that got very unwieldy and it became hard to administer.
[35:43.200 --> 35:52.200]  And so we just started beefing up the graphics cards in those Windows machines and buying like Adobe for Windows and throwing away the Macs.
[35:52.450 --> 35:54.220]  Now, Apple knew this was happening.
[35:54.360 --> 35:57.100]  This was an existential crisis for Apple.
[35:57.820 --> 35:59.880]  And so Steve Jobs had an idea.
[36:00.180 --> 36:04.430]  He took some of his technologists and he said, go reverse engineer Microsoft Office.
[36:04.820 --> 36:08.060]  Figure out how their file formats work and make me a new product.
[36:08.060 --> 36:15.620]  A product called the iWork suite, which numbers, pages, keynote, read and write, Word, Excel and PowerPoint files perfectly.
[36:15.950 --> 36:18.120]  They just they just tasked a team of engineers.
[36:18.360 --> 36:26.220]  Every time Microsoft shift shift an update, Apple would go and reverse engineer the new file formats and maintain file format compatibility.
[36:27.310 --> 36:29.020]  And that's what saved them.
[36:29.020 --> 36:37.060]  Because now you just didn't have to be on Windows to communicate with people who were using Windows.
[36:38.000 --> 36:43.720]  And right after the iWork suite came out, Apple conceived of quite an amazing ad campaign.
[36:43.960 --> 36:45.340]  The switch campaign.
[36:45.560 --> 36:48.560]  Which just advertised that switching costs were now really low.
[36:48.800 --> 36:51.880]  That the cost of switching to a Mac was just like buying a Mac.
[36:51.880 --> 36:53.880]  It wasn't giving up all your files.
[36:54.040 --> 37:00.260]  You just, you know, put your files on a zip cartridge or whatever, and plugged it into your Mac and your files would just work.
[37:00.450 --> 37:02.660]  They just work thanks to the iWork suite.
[37:03.020 --> 37:07.040]  So that is how Apple got saved.
[37:07.240 --> 37:13.540]  That was how Apple, which was then medium tech, really pushed back against Microsoft, which was big tech.
[37:14.080 --> 37:24.720]  And, you know, regulators, they have some understanding that interoperability is special and that interoperability is a remedy available in technology that is not available in other domains.
[37:24.950 --> 37:31.160]  And so we are seeing some interoperability mandates showing up in lots of different contexts.
[37:31.380 --> 37:35.100]  So the Chinese cyberspace regulation has some interoperability mandates.
[37:35.340 --> 37:37.680]  In the European Union, there's the Digital Markets Act.
[37:37.860 --> 37:38.500]  In the U.S.
[37:38.600 --> 37:39.740]  there's the Access Act.
[37:39.780 --> 37:51.480]  And in the U.K., the Competition and Markets Authority, which is their antitrust regulator, has recommended, uh, interoperability mandates And these mandates, they're kind of a mixed bag.
[37:51.640 --> 38:01.430]  Like in Europe, the Digital Markets Act, they decided that the first interoperability mandate would be on end-to-end encrypted messaging services, which, you know, don't get me wrong, they should be interoperable.
[38:01.540 --> 38:10.080]  The idea that you have to stay on WhatsApp or convince everyone you like to leave WhatsApp sucks because, you know, they spy on you when you use WhatsApp.
[38:10.080 --> 38:17.360]  I would like people to be able to leave WhatsApp as easily as possible without getting spied on by Mark Zuckerberg and his creepy cohort.
[38:17.980 --> 38:20.740]  And, you know, interoperability is what's going to do that.
[38:20.860 --> 38:24.950]  The problem is that, you know, the stakes in end-to-end encrypted messaging are really high.
[38:25.500 --> 38:26.880]  And it's easy to screw it up.
[38:27.020 --> 38:29.780]  And the European Union wants action on this like yesterday.
[38:30.300 --> 38:34.340]  And if they rush this, they're going to blow up the security model, right?
[38:34.340 --> 38:50.220]  It's not an area where you can afford to make a mistake because, you know, as we see with the NSO Group and their Pegasus malware, the most repressive, awful governments in the world go to enormous lengths to break end-to-end encrypted messaging services so they can figure out who to round up and torture and murder.
[38:50.220 --> 39:06.580]  You know, Jamal Khashoggi was lured to his death after the NSO Group's Pegasus software was wielded by the Saudi authorities to break into the WhatsApp accounts of his social media contacts and figure out his movements and trick him.
[39:06.580 --> 39:16.220]  And so, you know, we really, as much as it's important to get end-to-end encrypted messaging going, making it the first one is kind of a mistake.
[39:16.600 --> 39:20.720]  And I lament that the European Union has decided to start there.
[39:20.860 --> 39:22.280]  But nevertheless, it's a good idea.
[39:22.820 --> 39:27.280]  And I think, you know, the best case scenario right now is that maybe they change what they start with.
[39:27.460 --> 39:35.260]  The second best case scenario is that they extend the timeline so that we've got time to do it right and really standardize an interoperable protocol.
[39:35.260 --> 39:44.410]  And test it and make sure that the implementations are robust so that we don't expose billions of ED messenger users to risk.
[39:44.560 --> 40:00.340]  But nevertheless, it's pretty exciting that regulators in Europe, in the UK, in Australia, in Canada, in the U.S., and even in China are all thinking about interoperability as a specific remedy to weaken big tech platforms.
[40:00.340 --> 40:02.260]  I like that.
[40:02.500 --> 40:07.360]  But administering these interop mandates is going to be really hard.
[40:07.360 --> 40:22.080]  You know, if you say to Facebook, hey, you've got to stand up and expose an API that will allow like Mastodon and Diaspora instances to exchange messages with Facebook users, then you're going to have this problem that periodically Facebook is going to say,
[40:22.240 --> 40:26.880]  we took down the API because someone was doing something bad, right?
[40:26.880 --> 40:29.100]  Our intrusion detection system tripped.
[40:29.280 --> 40:32.280]  We thought someone was exfiltrating user data.
[40:32.740 --> 40:37.240]  And so out of an abundance of caution, we shut it down to do some forensics and figure out if this was right.
[40:37.240 --> 40:39.320]  And like, we want them to do that.
[40:39.500 --> 40:50.640]  We absolutely want companies that are guardians of user data to have good tripwires on their interfaces to make sure that those interfaces aren't being abused to invade their users privacy.
[40:50.640 --> 40:57.020]  But also, Facebook just wants to f*ck with companies that are interoperating with it.
[40:57.140 --> 41:07.930]  If you force Facebook to open up interoperability in order to weaken it, it's going to do everything it can to ruin the businesses of the companies that interoperate with it.
[41:08.340 --> 41:10.720]  And, you know, Facebook cheats all the time.
[41:10.840 --> 41:13.220]  It would be pretty naive to think that they weren't cheating here.
[41:13.220 --> 41:30.600]  And distinguishing the instance in which Facebook shuts down its API on the pretext of blocking data exfiltration and the instance in which Facebook shuts down its API because there was real data exfiltration or it looked like there was real data exfiltration.
[41:30.600 --> 41:35.700]  That is a really hard problem because Facebook runs this, like, bespoke network infrastructure.
[41:36.020 --> 41:41.500]  And to a first approximation, everybody who understands how that network infrastructure works is a Facebook employee.
[41:41.500 --> 41:45.300]  And so who do you ask when Facebook says, oh, we shut this down?
[41:45.780 --> 41:53.520]  Who do you ask to go and have a look and decide whether or not Facebook was lying to you or whether something really bad was happening?
[41:53.600 --> 42:04.780]  It is a long, fact-intensive process, and it will create long, intensive delays in figuring out whether Facebook's outages are pretextual or legitimate.
[42:05.120 --> 42:09.220]  And those delays will have real consequences for establishing competition.
[42:09.220 --> 42:24.000]  You know, if you're the founder of a startup that has lured users to your platform with the promise of being able to leave Facebook, have more privacy, be in charge of your own moderation decisions, but still be in communities on Facebook by interoperating with them,
[42:24.160 --> 42:32.960]  still communicate with users on Facebook, and like every other day, your service goes down and you don't even know when it's going to go back up, you're going to lose users.
[42:32.960 --> 42:38.740]  And if you're an investor who invested in that startup, you're going to lose confidence in that kind of startup.
[42:38.980 --> 42:42.300]  And, you know, if you're the user, you're going to go back to Facebook.
[42:42.820 --> 42:46.320]  And so we need to do something about this.
[42:46.360 --> 42:51.520]  And we don't have to throw up our hands and say, well, tech mandates can't be administered, so we shouldn't bother it.
[42:51.960 --> 42:52.580]  Far from it.
[42:52.580 --> 43:00.620]  I think we can devise a mechanism that will make big companies want to cooperate rather than defect, want to do the right thing.
[43:00.820 --> 43:05.740]  And that mechanism is something called adversarial interoperability.
[43:06.080 --> 43:11.660]  At Electronic Frontier Foundation, we called it this for a long time, but it's very hard to pronounce.
[43:11.660 --> 43:16.410]  And the acronym for adversarial interoperability is AI, which is kind of confusing.
[43:16.640 --> 43:19.660]  So we kind of brainstormed internally about another name for this.
[43:19.720 --> 43:26.100]  We came up with competitive compatibility, which we really like because it shortens to COMCOM, which is fun to say.
[43:26.240 --> 43:27.500]  So I'm going to call this COMCOM.
[43:27.640 --> 43:30.740]  When you hear COMCOM, think adversarial interoperability.
[43:31.120 --> 43:31.860]  So what's COMCOM?
[43:32.020 --> 43:38.580]  COMCOM is a catch-all term for the guerrilla tactics that kept the tech industry dynamic and pluralistic for decades.
[43:38.580 --> 43:49.220]  The stuff that every issue of 2600 magazine is filled with in tiny, impossible-to-read type if you have old eyes like me and Luxottica Essilor has stolen your eyes.
[43:49.280 --> 43:55.680]  I say this having spent the last two weeks on a beach reading a year and a half worth of issues cover to cover.
[43:57.120 --> 44:04.140]  Those tactics, reverse engineering, scraping, bots, the hacker tactics, those are how we save the Internet.
[44:04.520 --> 44:07.180]  Those are how the Internet was saved over and over again.
[44:07.180 --> 44:10.200]  It was COMCOM that created IBM PC clones.
[44:10.340 --> 44:14.380]  It was COMCOM that let Windows users take their Microsoft Office documents with them to the Mac.
[44:14.520 --> 44:17.180]  It was COMCOM that let Google index the web.
[44:17.320 --> 44:20.300]  It was COMCOM that let Facebook enclose the web.
[44:20.600 --> 44:34.420]  And COMCOM is built on that foundation of computer science, the universality, the bedrock principle that we know only how to make one kind of computer, the Turing-complete von Neumann machine that can run every program we know how to write.
[44:34.420 --> 44:38.360]  And that means that interoperators always have the advantage.
[44:39.000 --> 44:44.360]  To block reverse engineering, you have to build a technology and make no mistakes.
[44:44.680 --> 44:52.860]  To successfully reverse engineer that technology, you only need to find and exploit one mistake, one error made by the incumbent firm.
[44:53.520 --> 44:56.340]  COMCOM and mandates are like two-part epoxy.
[44:56.620 --> 44:59.020]  The mandate is strong, but it's rigid.
[44:59.020 --> 45:04.280]  The adversarial interop, the COMCOM, it's flexible, but it's chaotic, right?
[45:04.340 --> 45:13.380]  If you're using bots to maintain interoperability, to scrape a community on Facebook and export it to Diaspora, it's going to break all the time, right?
[45:13.440 --> 45:17.520]  They're going to change their HTML DOM model and you're going to have to rewrite your bot and whatever.
[45:17.520 --> 45:22.340]  So it's, it's chaotic, but it's flexible and it can handle what you can throw at it.
[45:23.740 --> 45:29.420]  Together, both COMCOM and mandates, they're stronger and more robust than either of them are on their own.
[45:29.600 --> 45:31.380]  So let me give you a real-world example.
[45:31.800 --> 45:39.620]  Back in 2012, the voters of Massachusetts passed an interoperability mandate via a ballot initiative with an 80% majority.
[45:39.620 --> 45:50.760]  The mandate was an automotive right to repair rule, and it required automakers to allow independent mechanics to access diagnostic information that was traversing the wired network within the car.
[45:50.900 --> 45:51.860]  It's called the CAN bus.
[45:51.980 --> 45:52.760]  You've probably heard of it.
[45:53.260 --> 45:56.000]  Now, the automakers had an obvious counter move.
[45:56.120 --> 46:01.660]  They switched to relaying diagnostic information over wireless networks, not the CAN bus.
[46:02.040 --> 46:03.860]  Wireless networks were exempted from the mandate.
[46:04.040 --> 46:08.840]  And this was back in 2012, before the supply chain shocks of the pandemic, when chips were really cheap.
[46:08.840 --> 46:14.520]  And they could buy, like, a system on a chip for $0.26 with a little Wi-Fi emitter in it.
[46:14.580 --> 46:20.300]  And so they could build a little YLAN inside every one of our cars and pass the error codes around there.
[46:20.360 --> 46:23.860]  And that way, the mechanics couldn't access that data.
[46:24.660 --> 46:32.240]  In 2020, eight years later, voters returned to the ballot box, and they passed another nearly identical automotive repair mandate.
[46:32.480 --> 46:34.280]  This one patched that hole.
[46:34.280 --> 46:38.940]  It said that independent mechanics needed to be furnished with access to wireless error codes, too.
[46:39.080 --> 46:45.400]  That the point of the mandate was the error code, not the underlying substrate of the network that it transmitted on.
[46:45.580 --> 46:48.920]  And that mandate also passed with an overwhelming majority, again, about 80%.
[46:48.920 --> 46:51.960]  But the automakers have since gone to court, and it's two years later.
[46:52.100 --> 46:54.000]  And that mandate is still not in effect.
[46:54.000 --> 47:03.180]  That means that it's been a decade since the automaker's monopolistic capture of the repair market was banned by law, but they still get to run that racket.
[47:03.620 --> 47:09.480]  Now, that is a decade in which mechanics closed their shop and exited the industry, or they went to work for big car.
[47:09.480 --> 47:22.920]  Where drivers learned that if you go to an independent mechanic, it's a roll of the dice, that they may put it up on the lift, take a look at it, suck their teeth, and say, Sorry, only the manufacturer can fix this one, because I can't get at the diagnostic codes.
[47:23.260 --> 47:30.880]  This was a decade where investors learned that big car has a kill zone around its products, and only fools would risk entering it.
[47:31.740 --> 47:35.520]  But imagine if the mandate had been bolstered by ComCon.
[47:36.140 --> 47:43.980]  Imagine if, say, three bright MIT kids could have used their interoperator's advantage to reverse engineer the error codes in those new cars.
[47:44.100 --> 47:53.040]  They could have designed a gadget with a $3 bill of materials, ordered a container load of them from Guangzhou, had them shipped to the port of Los Angeles, and sold them to mechanics across the state.
[47:53.040 --> 48:00.680]  And because those markets are super leaky, they could have sold them to mechanics across the country, and across the world, at a cost of $50 or $100.
[48:01.640 --> 48:05.820]  Imagine if they could have found investors to help them build ancillary services on top of that.
[48:06.140 --> 48:12.520]  Warranty services, conduits for third-party parts, all kinds of stuff that's super high margin that the automakers love.
[48:13.400 --> 48:15.180]  Well, that would have been a different story.
[48:15.820 --> 48:19.960]  You'd think that in a situation like that, the automakers would have been incentivized to do the right thing.
[48:19.960 --> 48:26.320]  To not create the space in which these $3 gadgets flood the market and eat their lunch.
[48:26.860 --> 48:35.160]  After all, the one thing that a publicly listed company hates more than complying with regulations is facing unquantifiable risks.
[48:35.600 --> 48:40.980]  The public markets, they are notoriously vengeful when they are surprised with bad news.
[48:40.980 --> 48:49.540]  Like, think back to last January, when Facebook revealed its first quarter number or last quarter numbers for 2021.
[48:50.000 --> 48:55.000]  And showed that their users had declined slightly more than they had anticipated.
[48:55.280 --> 48:59.640]  And suffered the single largest one-day loss of any company in the history of the world.
[48:59.860 --> 49:04.120]  Losing $280 billion off their market cap in one day.
[49:04.200 --> 49:07.180]  About 25% of the company's value.
[49:07.180 --> 49:19.720]  So, you'd hope that when confronted with either the chaos of ComCom, techno guerrilla warfare, or the predictable world of just, like, doing what the law says they have to do, they would comply with the law.
[49:20.280 --> 49:24.680]  But, of course, no one ever lost money betting on the hubris of corporate monopolies.
[49:24.800 --> 49:26.200]  So, maybe they would cheat.
[49:26.440 --> 49:30.000]  And if they did, well, the MIT kids could have come to the rescue.
[49:30.000 --> 49:32.900]  ComCom and mandates are complementary.
[49:33.520 --> 49:34.840]  They're strength and flexibility.
[49:34.840 --> 49:36.340]  They're the carrot and the stick.
[49:36.700 --> 49:52.940]  And the risks to public firms from the surprising bad news of ComCom, the unquantifiable risk of ComCom, it falls disproportionately on senior managers who are the ones who make the decisions about whether to cheat or not to cheat.
[49:52.940 --> 49:54.260]  I mean, think about it.
[49:54.360 --> 50:05.260]  If you're top management at Facebook, the major part of your compensation comes from stock, which means that when the stock of Facebook falls, it wipes out your personal net worth.
[50:05.600 --> 50:12.060]  So, wouldn't you rather comply with the law than risk the surprising bad news that you have to confront your investors with?
[50:12.060 --> 50:20.300]  Not to mention that if you want to get involved in guerrilla warfare against ComCom, against interoperators, you need to hire some engineers.
[50:20.600 --> 50:24.100]  And if your stock price keeps falling off a cliff, you're going to have to pay them cash.
[50:24.300 --> 50:29.920]  You're not going to be able to pay them in stock, which is one of the ways that monopolies get huge advantages over their rivals.
[50:30.140 --> 50:38.720]  They can pay their employees in money they print themselves in stock instead of having to pay them with the cash that their shareholders prefer.
[50:38.720 --> 50:50.300]  Now, if they lose that advantage because their stock has become really volatile, because they keep cheating, and every time they cheat, it tempts people into ComCom, well, then they're going to find themselves getting weaker and weaker.
[50:51.160 --> 50:55.180]  But ComCom is hard, and not for technological reasons.
[50:55.860 --> 51:05.020]  Large firms have promulgated new legislation and new regulations and new interpretations of existing rules that make ComCom illegal, even sometimes a felony.
[51:05.020 --> 51:10.620]  So, you know, if you're a 2600 reader, you probably know about Section 1201 of the DMCA.
[51:11.260 --> 51:18.240]  In fact, the 2600 case is the case that established that Section 1201 was just as bad as we feared it would be.
[51:18.600 --> 51:28.620]  Section 1201 is the law that says that it's a felony punishable by up to five-year prison sentence to reveal defects in DRM systems and access control systems.
[51:30.040 --> 51:39.220]  And 2600 published the DeCSS code, and we at EFF unsuccessfully went to court to defend them, because it turns out that the DMCA means exactly what it says it does.
[51:39.220 --> 51:56.240]  I'll say that we are today representing two other clients, Matthew Green from Johns Hopkins and Bunny Huang, ex of MIT, who's the guy who hacked the Xbox and is currently building the precursor, a great little mobile platform, in a constitutional bid to overturn Section 1201.
[51:56.240 --> 52:05.300]  But Section 1201, combined with its European equivalent Article 6 of the European Copyright Directive, is a very powerful tool for making ComCom illegal.
[52:05.620 --> 52:08.680]  So are the cybersecurity laws like the Computer Fraud and Abuse Act.
[52:08.800 --> 52:12.140]  That's the law that was used to hound poor Aaron Swartz into his grave.
[52:12.300 --> 52:18.440]  And while it was narrowed significantly by the Supreme Court last summer with the Van Buren case, it is still quite a danger.
[52:18.440 --> 52:23.620]  And then there's, on top of that, software patents and exotic contract theories like tortious interference.
[52:23.740 --> 52:29.320]  They all come together to make it very legally fraught to do ComCom.
[52:29.600 --> 52:35.560]  And reforming all these laws, which again, that's something we should absolutely do, it is a big lift and a multi-year project.
[52:35.800 --> 52:41.560]  And just like corporate breakups, we don't have time to wait for that to play out before we fix tech.
[52:41.980 --> 52:46.980]  Luckily, we don't have to wait for those laws to be reformed in order to make the world safe for ComCom.
[52:46.980 --> 52:53.500]  For example, we could pass the Access Act and or the DMA, and just wait for the big companies to cheat.
[52:54.040 --> 52:55.380]  Because they cheat.
[52:55.780 --> 53:03.180]  And when we catch them cheating, and we drag them into court, and they offer a settlement, because that's what they would prefer.
[53:03.360 --> 53:05.180]  They don't want to spend 15 years in court.
[53:05.560 --> 53:09.460]  We could, as part of that settlement, impose a special master on them.
[53:09.560 --> 53:12.900]  This is one of the remedies that's available to antitrust prosecutors.
[53:12.900 --> 53:26.800]  We could say to them, you're going to have an adult supervisor, a lawyer, independent lawyer, who is going to have a first look at any legal threat you send to any competitor that is doing interoperability.
[53:26.800 --> 53:33.800]  To make sure that when you threaten to sue someone for hacking your systems, that they're actually hacking your systems.
[53:33.980 --> 53:36.280]  That they're doing something that's adverse to your users' interests.
[53:36.460 --> 53:40.760]  And not, say, adding a repair facility to your system.
[53:40.760 --> 53:41.880]  Or making it more usable.
[53:42.120 --> 53:43.380]  Or making it more accessible.
[53:43.720 --> 53:45.360]  Or making it more secure or private.
[53:45.560 --> 53:47.200]  Or just giving users more autonomy.
[53:47.200 --> 53:55.560]  If that's the case, if it's a pretextual lawsuit to block competition, your special master is going to just say, sorry, you can't do it.
[53:56.100 --> 53:58.180]  And there are other means at our disposal, too.
[53:58.440 --> 54:02.360]  One that I really like the idea of is reforming our procurement rules.
[54:02.580 --> 54:05.240]  It's amazing how powerful procurement could be.
[54:05.420 --> 54:12.060]  After all, Uncle Sam and state governments and local governments, they spend a lot of money on technology.
[54:12.060 --> 54:18.900]  And as a matter of good, prudent administration, no public body should ever procure any product.
[54:19.100 --> 54:27.380]  Whether it's a car or Google Classroom or mobile phones for its employees.
[54:27.660 --> 54:34.180]  Without first extracting a binding promise from the vendor not to legally attack interoperators.
[54:34.380 --> 54:36.240]  This is a very old principle.
[54:36.540 --> 54:39.340]  A bedrock that we have somehow forgotten.
[54:39.340 --> 54:47.800]  You know, Abe Lincoln told the armorers who supplied the rifles to the Union Army that they must use interoperable tooling and ammunition.
[54:47.960 --> 54:49.060]  I mean, like, obviously.
[54:49.460 --> 54:55.480]  Like, what kind of commander-in-chief would say, yeah, give us proprietary ammo.
[54:55.480 --> 54:58.480]  And if you decide not to make it anymore, we'll just, like, wait.
[54:58.660 --> 55:00.720]  Or buy different guns for the army.
[55:00.900 --> 55:03.060]  The enemy will understand, right?
[55:03.200 --> 55:05.440]  Well, I'll tell you what kind of commander-in-chief would say that.
[55:05.600 --> 55:07.700]  All of the current commanders-in-chief.
[55:07.700 --> 55:10.000]  Because today, the entire U.S.
[55:10.360 --> 55:13.640]  military budget, which, like, again, don't get me wrong, we should be zeroing that out.
[55:13.760 --> 55:14.640]  But the entire U.S.
[55:14.980 --> 55:19.700]  military budget is filled with price gouging from firms that have blocked interoperability.
[55:19.840 --> 55:23.080]  For example, in aerospace, you have these private equity roll-ups.
[55:23.240 --> 55:30.220]  Where these private equity weirdos go out and they identify single source suppliers for parts that are in the aerospace supply chain.
[55:30.220 --> 55:33.100]  So a little widget that's used in a plane or a bomber or a drone.
[55:33.280 --> 55:35.300]  And they go and they buy that company up.
[55:35.460 --> 55:37.100]  It's the only supplier for that widget.
[55:37.520 --> 55:39.780]  And then here's the crazy part.
[55:39.900 --> 55:41.640]  They lower the price of the widget.
[55:41.780 --> 55:47.100]  They make it basically free for the primary military contractors like Boeing or Lockheed.
[55:47.100 --> 55:55.240]  So Boeing and Lockheed, they're like, yeah, we'll just design our new stuff around this because we don't even pay to source these parts in our manufacturer.
[55:55.700 --> 56:05.800]  But then, once they transfer this to the Air Force, the Army, the Navy, and they need to fix it, and they order replacement parts, the parts come at a 10,000% markup.
[56:05.800 --> 56:26.120]  So, you know, again, just as a matter of good, prudent administration, nobody should be buying anything with public money unless they also have a promise from the manufacturer that they're not going to go after the firms and say to them, or they're not going to go after interoperators and say to them that they're not allowed to make replacement parts,
[56:26.280 --> 56:32.760]  not allowed to make software, not allowed to make plugins, not allowed to take over the management of the tool, and so on.
[56:32.760 --> 56:47.520]  You know, every time a government motor pool buys a car, every time a school district licenses Google Classroom, every time a ministry or a department licenses Microsoft Teams for its video conferencing, they should be extracting binding non-aggression covenants from those suppliers.
[56:48.020 --> 56:53.340]  These aren't the only means to restore the legal basis for the ancient and honorable tradition of ComCom.
[56:53.780 --> 56:57.720]  Like, we could enact an interoperator's defense in law.
[56:57.720 --> 57:05.660]  We could just say, notwithstanding all these other laws, if you can show that you are doing something bona fide and in the public interest, that you have a defense in law.
[57:06.580 --> 57:09.240]  Irrespective of how we restore ComCom, we should do it.
[57:09.520 --> 57:16.060]  We should restore the mechanism that stabilizes our interop mandates and makes them administratable.
[57:16.160 --> 57:19.440]  Because that is how we can fight for the users.
[57:19.780 --> 57:22.720]  That is how we can put people in charge of their technology.
[57:22.720 --> 57:30.960]  That is how we can let them pick the part of the tool that provides them with value and leave the part of the tool that sucks value out of them.
[57:31.180 --> 57:38.900]  That is how we can unleash the hackers who've been reading 2600 ever since the DeCSS case or even before.
[57:39.100 --> 57:45.000]  We can unleash them to make the tools that help the users live a dignified technological existence.
[57:45.000 --> 57:52.900]  And it is how we build the vanguard of the antitrust movement that will take back control from all kinds of corporations.
[57:52.900 --> 58:01.000]  Because once we tame big tech with antitrust, we will have built an antitrust movement that can go after the people who make your glasses.
[58:01.000 --> 58:06.800]  And that scumbag who ran the wrestling league and the people who own all the cheerleader uniform companies.
[58:06.980 --> 58:10.560]  And the idiots who keep building bigger ships that get stuck in the Suez Canal.
[58:10.620 --> 58:12.460]  We can break them all up.
[58:12.620 --> 58:14.700]  We can usher in a new era.
[58:14.700 --> 58:21.240]  So every time I talk to 2600, the last time I talked to 2600, I end my talk with the same thing.
[58:22.540 --> 58:27.700]  And it's a quote or a story about the scholar James Boyle.
[58:27.880 --> 58:32.340]  Jamie is a copyright scholar at Duke University.
[58:32.780 --> 58:35.580]  He runs the Center for the Public Domain with Jennifer Jenkins.
[58:36.120 --> 58:38.900]  Jamie is a great...
[58:38.900 --> 58:41.800]  He was part of the Creative Commons founding team.
[58:41.800 --> 58:43.880]  And he's just a great all-rounder.
[58:43.880 --> 58:56.340]  And he tells the story about the coining of the term ecology and what it did to the people who had issues with the way that we treated our planet.
[58:56.740 --> 59:07.700]  And he says that before the term ecology was coined, it wasn't obvious that if I cared about owls and you cared about the ozone layer, that we were fighting the same fight.
[59:07.700 --> 59:13.900]  Like, on its face, what do charismatic nocturnal avians have to do with the gaseous composition of the upper atmosphere?
[59:14.300 --> 59:21.640]  But Jamie says that the term ecology took a thousand movements, a thousand issues, and welded them together into one movement.
[59:21.640 --> 59:25.560]  That it created a kind of coherence between these different struggles.
[59:26.240 --> 59:32.040]  And today, you know, if you use technology, you understand what monopolies have done to the wild and woolly Internet.
[59:32.320 --> 59:40.860]  That, you know, the transformation of the web into five giant websites filled with screenshots of text from the other four was not an accident, it was a plan.
[59:40.860 --> 59:43.700]  And it happened because regulators let it happen.
[59:43.700 --> 59:45.480]  And we could take our Internet back.
[59:46.040 --> 59:59.060]  But, you know, if you're involved in, I don't care what it is, cheerleader uniforms or sustainable energy or professional wrestling or eyeglasses, you also understand that something rotten is going on.
[59:59.060 --> 01:00:08.720]  And as we make the connection between, & between the fact that there's only 4 publishers left and one bookstore online and one national chain bookstore.
[01:00:09.440 --> 01:00:14.180]  & between the fact that there's only 3 record labels left and 2 audio streaming platforms.
[01:00:14.980 --> 01:00:21.580]  Between the fact that there's 5 giant banks and that there's just a couple of giant hydrocarbon companies that run all the gas lines.
[01:00:21.580 --> 01:00:34.820]  Once we make those connections and we realize that all of these arose out of the same regulatory reform, that Robert Bork secret history where we replaced the harmful dominance standard with the consumer welfare standard.
[01:00:35.020 --> 01:00:40.940]  And then we said the consumer welfare standard is a needle that's impossible to thread and all monopolies should be treated as presumptively good.
[01:00:41.060 --> 01:00:45.020]  Once we understand that, we will raise an army that will fight.
[01:00:45.180 --> 01:00:51.560]  And just like the people who cared about owls fought for the ozone layer and the people who cared about the ozone layer fought for the owls, the people who are worried about cheerleaders.
[01:00:51.580 --> 01:00:57.900]  The other uniforms will be there to bust Facebook and the people who want to bust Facebook will be there to bust the eyeglasses monopoly too.
[01:00:58.340 --> 01:00:59.600]  That is how we're going to win.
[01:00:59.940 --> 01:01:05.100]  And it's going to start with this because tech is not exceptional, except to the extent that it is exceptional.
[01:01:05.280 --> 01:01:09.960]  And the thing that is exceptional about tech is it's the battlefield that we're going to fight all these other battles on.
[01:01:10.100 --> 01:01:14.460]  It's how we're going to find each other, it's how we're going to coordinate our work, and it's how we're going to win.
[01:01:15.380 --> 01:01:16.520]  So that's my talk.
[01:01:16.820 --> 01:01:18.460]  I really appreciate you listening.
[01:01:18.460 --> 01:01:23.980]  This is the part where I tab back to my Zoom window and make sure that I haven't been disconnected for the last hour.
[01:01:24.180 --> 01:01:25.320]  Looks like I'm still good.
[01:01:25.700 --> 01:01:26.340]  That's good.
[01:01:26.520 --> 01:01:27.600]  I would have been embarrassing.
[01:01:27.940 --> 01:01:29.540]  And now I take your questions.
[01:01:29.560 --> 01:01:32.200]  So I hope you have some good ones.
[01:01:45.350 --> 01:01:46.130]  All right, Corey.
[01:01:46.270 --> 01:01:49.870]  Well, you have not been tapped out on your Zoom window for the last hour.
[01:01:49.990 --> 01:01:53.170]  Everyone has been rapidly listening to your talk.
[01:01:53.430 --> 01:01:55.690]  So with that, we'll start doing Q&A.
[01:01:55.910 --> 01:01:57.170]  I don't know if anyone has any questions.
[01:01:57.170 --> 01:02:02.190]  We have a number of questions in the Matrix chat, so I'll try and alternate back and forth if we have some live audience members.
[01:02:07.460 --> 01:02:08.100]  Hi, Corey.
[01:02:08.360 --> 01:02:08.840]  It's home.
[01:02:09.020 --> 01:02:09.100]  Hi.
[01:02:09.500 --> 01:02:09.880]  Hey.
[01:02:11.440 --> 01:02:21.280]  I'm wondering how...so when I think about how ComCom plays out in mobile apps, which is where a lot of the action is now, is software.
[01:02:22.800 --> 01:02:30.720]  You have the iOS app store, and then you have underneath that, you know, they have control over what apps people can install.
[01:02:30.740 --> 01:02:41.240]  And so, you know, say someone made a ComCom Instagram that didn't have ads in it, which is the main...that's the main way Instagram puts the screws to people, is it makes them watch a ton of ads in order to see pictures of whatever they're interested in.
[01:02:42.620 --> 01:02:49.100]  Can't...I mean, unless you unravel the app store with ComCom, Facebook will just ask the app store to block the thing.
[01:02:49.100 --> 01:02:58.660]  And then, if unraveling the app store sounds like a harder problem, like, pretty akin to the one you were talking about of the situation where someone just disrupts you.
[01:02:59.200 --> 01:03:04.920]  You know, if you made an alternate app store by hacking the iPhone, every single software update, something would fail.
[01:03:06.500 --> 01:03:10.760]  And, you know, it throws things back in the territory of, like, okay, we need some legal intervention.
[01:03:11.500 --> 01:03:13.900]  So, I'm just thinking about how that plays out with apps.
[01:03:14.020 --> 01:03:16.620]  And do you...how do you think that plays out?
[01:03:16.700 --> 01:03:18.500]  Do you see something more optimistic there?
[01:03:18.660 --> 01:03:28.060]  Or is it...is it pretty much that we won't win that much until we have the big fight about competition policy and win?
[01:03:28.060 --> 01:03:33.400]  I think that you're right that app stores are a real locus of control.
[01:03:33.700 --> 01:03:39.660]  But it's also...it also means that there's a real constituency of people who want to see them reformed.
[01:03:39.840 --> 01:03:56.300]  So, we just at EFF submitted comments that I drafted on the Competition and Markets Authority in the UK inquiry on whether they should force Apple to support browser engines other than WebKit, specifically browser engines that would allow for web apps to be deployed.
[01:03:56.560 --> 01:03:59.020]  Because that's how you get around it, right?
[01:03:59.240 --> 01:04:11.940]  Is if you have full HTML5 web app support, you can just basically...in fact, you know, most apps are just HTML5 web apps that have been, you know, wrapped up with some iOS gubbins.
[01:04:13.580 --> 01:04:16.340]  And you're right that you need...you probably need a mandate.
[01:04:16.340 --> 01:04:45.460]  But again, I think that if we said to Apple, not only are we going to force you to support non-WebKit browser engines on iOS, but we are also going to immunize people who do things like exploit Checkmate, which is the defect in the secure computing environment for iOS that affects eight generations of iOS devices spanning eight years,
[01:04:45.720 --> 01:04:52.380]  we're going to immunize them from liability under anti-circumvention and contract and, you know, software patent or whatever.
[01:04:52.380 --> 01:05:16.420]  If they use that to add another browser engine to iOS or add, you know, browser engines or extend the WebKit browser engine support so that it can be a substrate for full-fledged web apps, then Apple will have a strong incentive to, you know, support WebKit or,
[01:05:16.440 --> 01:05:33.760]  you know, WebKit alternatives that support real web apps, not because ComCom will be like a stable remedy if it cheats, but because Apple will hurt as badly in a world in which ComCom is legal as the people who have to throw engineering at making ComCom work,
[01:05:33.760 --> 01:05:45.520]  that it will just be a nightmare for Apple to not be able to go to court and stop people from exploiting Checkmate to add browser engines.
[01:05:45.960 --> 01:05:47.980]  It's just like their worst-case scenario.
[01:05:48.520 --> 01:05:50.300]  It makes them hire a bunch of engineers.
[01:05:50.300 --> 01:05:52.600]  It makes them divert engineers from products that are existing.
[01:05:52.820 --> 01:05:56.960]  It makes them, you know, do all kinds of stuff they just really don't want to do.
[01:05:56.960 --> 01:06:18.040]  And so, as between the marginal losses from having to operate a web store or an app store that people use because it's better than the other app stores and not because it's illegal to make another app store, and, you know, having to only fight with guerrilla warfare and never with lawyers,
[01:06:18.240 --> 01:06:18.340]  right?
[01:06:18.440 --> 01:06:27.700]  You can think about, like, these tech companies as having a building full of lawyers and a building full of engineers, and whenever anyone starts to interoperate with them, they don't even stop at the building full of engineers.
[01:06:27.840 --> 01:06:38.040]  They go straight to the building full of lawyers because they understand that if the lawyer drags you into court and, like, destroys you, that, you know, no one's going to invest in another company that wants to do what you're going to do.
[01:06:38.100 --> 01:06:46.000]  And anyone who's, like, thinking of doing it will have a hard time recruiting participants and, you know, if they're doing it under a university, the university won't approve it and so on.
[01:06:46.100 --> 01:06:48.800]  So they really want to just, like, win the fight once with a lawyer.
[01:06:48.940 --> 01:06:55.160]  And if we take away their right to win the fight with a lawyer, and we force them to fight with engineers, they might decide not to fight at all.
[01:06:55.160 --> 01:06:59.420]  And if they do decide to fight, well, then we'll at least be able to, like, patch in a remedy.
[01:06:59.660 --> 01:07:02.000]  It won't be as good as a stable mandate.
[01:07:02.340 --> 01:07:24.460]  But, you know, one of the things that I didn't mention when I told the story of iWorkSuite and Microsoft and Pages and Keynote and Numbers is Microsoft, after Pages was a success, after a couple of rounds of combat where Microsoft changed the file format and Pages was able to match those changes quickly so they could maintain compatibility,
[01:07:25.090 --> 01:07:29.980]  Microsoft took its Word files, its Office files, to a standards body and standardized them.
[01:07:30.120 --> 01:07:34.060]  That's that X and XLSX and DocX and PPTX.
[01:07:34.200 --> 01:07:36.820]  That's the XML standard version of those documents.
[01:07:37.060 --> 01:07:43.560]  And it's why we now have, like, Google Docs and LibreOffice and all those other things that read and write Office files.
[01:07:43.620 --> 01:07:53.740]  It's why you can paste an Office file into a browser window and have it maintain, like, formatting and links and data structures and so on, because it's been standardized.
[01:07:53.740 --> 01:08:11.280]  Like, there just came a point where breaking interoperability became a pure cost center with no profits and if we make it into a pure cost center with no profits, then the company will just stop trying to do it or is likely to stop trying to do it.
[01:08:11.280 --> 01:08:13.620]  Because they're not doing it just because they're ornery.
[01:08:13.780 --> 01:08:15.880]  They're doing it because it makes them a shit ton of money.
[01:08:16.240 --> 01:08:19.860]  And if we... it doesn't have to be the case that anyone else makes money.
[01:08:19.960 --> 01:08:27.080]  We just have to make it so they have such high cost that they no longer make money by blocking interoperability by blocking interoperability and then they'll change their minds.
[01:08:28.820 --> 01:08:29.260]  Thanks.
[01:08:32.660 --> 01:08:33.560]  Hi, thank you.
[01:08:33.680 --> 01:08:35.960]  This was really interesting and I'm a fan of your novels.
[01:08:36.600 --> 01:08:47.540]  I wanted to ask... you talked a bit about fighting abuses such as terrorism, harassment, these things that people do badly with with every platform they can find to do it on.
[01:08:48.020 --> 01:08:59.480]  And one of the arguments that you hear, I think, often even from people who work on things like content moderation within, the big social media companies, is that if you break up the company, it makes content moderation harder.
[01:08:59.900 --> 01:09:09.120]  And I think you alluded to earlier, almost saying that it's like a losing battle even within the company, which I doubt you would say that we should just give up on it completely.
[01:09:09.120 --> 01:09:10.820]  So I guess I'm wondering, what's your solution?
[01:09:10.980 --> 01:09:18.280]  How do we have effective, like, non-abusive Internet in a world where there's sort of more smaller players?
[01:09:19.100 --> 01:09:23.240]  So I guess it depends on what your non-abusive Internet looks like, right?
[01:09:23.240 --> 01:09:36.320]  Is your non-abusive Internet, and I don't mean to single out you, I mean, like, is the hypothetical non-abusive Internet that we're trying to build, is it an Internet in which odious people never say disgusting things, not even to each other?
[01:09:36.320 --> 01:09:46.420]  Or is it one in which you can be a full-fledged participant in the Internet without having to be subjected to the disgusting views of odious people?
[01:09:46.840 --> 01:09:49.240]  And I'm in favor of the second one.
[01:09:49.380 --> 01:09:51.640]  I think the first one really is a fool's errand.
[01:09:52.260 --> 01:09:59.280]  You know, I deplore racism as much as the next person, and I would love to live in a world free from racism.
[01:09:59.840 --> 01:10:08.320]  I believe that people who operate public spaces should have the right to kick out others for making racist statements.
[01:10:08.320 --> 01:10:17.640]  But I also think that it's probably a bridge too far to say that we are going to try and engineer a system in which no one is racist, not even in the privacy of their own living room, right?
[01:10:17.760 --> 01:10:23.400]  Or in the privacy of their own campfire discussion with their co-racists, that that's probably like...
[01:10:23.400 --> 01:10:25.340]  And to me, that's what 4chan is, right?
[01:10:25.520 --> 01:10:31.640]  It's probably too much to hope that there will never be a conversation space like that.
[01:10:31.640 --> 01:10:34.380]  And if what we try to do, if the argument...
[01:10:35.900 --> 01:10:40.460]  There's more than one way that people involved in content moderation express this, and you alluded to one.
[01:10:41.680 --> 01:10:57.280]  One of the things that people involved in content moderation battles say or imply is, if we keep all the conversation on Facebook, if we make it effectively impossible to have a conversational space that's not Facebook, all we need to do is make Facebook good,
[01:10:57.280 --> 01:10:59.780]  and then all the bad conversation disappears.
[01:10:59.780 --> 01:11:03.940]  And I think that that is both wrong and not good.
[01:11:04.200 --> 01:11:06.400]  Even if it were right, it would be bad.
[01:11:07.380 --> 01:11:20.840]  I don't think that it is a net positive, no matter how good Facebook is at moderating our conversations, I don't think it's a net positive to vest all authority over how we conduct discourse in one firm.
[01:11:22.200 --> 01:11:24.540]  And I don't want to imply that you do.
[01:11:24.680 --> 01:11:40.520]  I'm just trying to address one of the arguments from the centralized content moderation side, I think that there's a softer version of this, which is, well, if most conversation happens on Facebook, then we accomplish much the same things.
[01:11:40.640 --> 01:11:44.420]  We chase the odious people with the hateful views to the periphery.
[01:11:44.420 --> 01:12:03.600]  And the problem with that is that content moderation at scale has a lot of failure modes that disproportionately affect the people that we worry about when we worry about hateful conduct online, harassment and so on online, marginalized users.
[01:12:03.600 --> 01:12:20.440]  That making a three ring binder thick enough to contain the policy that states what is and isn't conversationally acceptable among three billion Facebook users and a hundred companies speaking a thousand languages is impossible.
[01:12:20.840 --> 01:12:35.020]  And so that means that Facebook is always going to make errors and that those errors will be made by a kind of civil bureaucratic apparatus, which is what we can think of Facebook moderation as.
[01:12:35.220 --> 01:12:48.680]  And for Facebook, think YouTube or Twitter or any of these other big firms that, you know, if you've ever been in social media jail or had a post taken down or whatever and had to argue with them, you understand that there's a kind of weird kangaroo court that they all run.
[01:12:48.680 --> 01:13:05.540]  And the reason that they're running a kangaroo court is if they ran an actual formal justice system for this, if every time they tagged or blocked or took down a piece of content, you had the right to an attorney who could argue for that content, you know,
[01:13:05.980 --> 01:13:09.180]  they would... it would be full employment for Internet lawyers, right?
[01:13:09.240 --> 01:13:15.080]  There would just be like an infinitude of people whose job it would be to argue about their content moderation choices.
[01:13:15.080 --> 01:13:17.000]  So it's this very impossible situation.
[01:13:17.740 --> 01:13:31.040]  Meanwhile, the people who are like at the coalface, the people who are having the discussion, who are in the social space, are capable of understanding the nuance and also understanding when exceptions should and shouldn't be made.
[01:13:31.320 --> 01:14:01.180]  And so, you know, the classic example would be survivors of terrorist violence who collect evidence of the terrorist violence for use in human rights tribunals and other important reckonings for the perpetrators of the violence, who are capable of understanding what elements of that video or audio or text or still imagery is important to bearing witness.
[01:14:01.180 --> 01:14:11.560]  And that is a thing that an external moderator who is far removed from a context that thankfully most of us don't have to contend with, will struggle to get right.
[01:14:12.260 --> 01:14:15.040]  And those failures are very consequential.
[01:14:15.980 --> 01:14:37.380]  Likewise, if you're in a marginalized community in which you discuss and commiserate and give strength to one another over the racist or xenophobic attacks that you've been subject to online, and you repeat those slurs, and you talk about why they were hurtful,
[01:14:37.520 --> 01:15:01.380]  and a moderator has to decide whether your use of a slur is impermissible, they lack the context often to understand what the difference is between a group of people who are against whom slurs are directed using it is, versus the people who use those slurs to harm those marginalized users.
[01:15:01.680 --> 01:15:07.120]  And so I think the right answer is to give the communities themselves more control.
[01:15:07.120 --> 01:15:30.420]  And so, you know, I like the model of the Fediverse, and I think the idea that you can stand up a server that connects to other servers, whose users can exchange messages with one another, who can block other servers or users if they think that those servers' policies are bad or the user is bad,
[01:15:30.420 --> 01:15:41.400]  and where it's very easy to move from one server to another if the moderation policies don't suit you, and in which you have conversations that become, you know, potentially fragmented, either within a conversation.
[01:15:41.400 --> 01:15:44.140]  So you might have, say, a group of...
[01:15:44.660 --> 01:15:49.780]  Well, we actually represented at EFF, we represented a group of breast cancer previvors.
[01:15:49.780 --> 01:16:01.800]  So these are women who carry a gene that makes it likely that they could develop breast cancer, and whose female relatives have therefore often contended with or even died of breast cancer.
[01:16:02.360 --> 01:16:04.640]  And so that group might split, right?
[01:16:04.740 --> 01:16:09.460]  You might say there might be different views about what is and is an acceptable discourse within that group.
[01:16:09.640 --> 01:16:16.080]  And rather than, like, Facebook making the final judgment call about what that is, the users themselves could decide to split.
[01:16:16.080 --> 01:16:25.380]  Or a user or a substantial portion of those users might migrate to a server where certain words, phrases, or conduct is blocked.
[01:16:25.860 --> 01:16:31.580]  And so, you know, they would effectively be shadow banning other members of the group without the shadow, right?
[01:16:31.660 --> 01:16:43.900]  The other members of the group would know that somewhere out there there's a server where if they use certain phrases or certain words or whatever, or if they act in certain ways, their messages will disappear and maybe their account will be blocked from those users.
[01:16:43.900 --> 01:16:47.020]  And that'll just be, like, something they have to contend with.
[01:16:47.480 --> 01:16:53.000]  And I think that that is a wildly imperfect system whose virtue is that it's better than all the other systems.
[01:16:53.000 --> 01:17:00.020]  That the people who are in the community understand its norms in a way that just can't be adequately captured.
[01:17:00.020 --> 01:17:10.320]  And that what you lose from having the resources of a company with the size and the monopoly control of Facebook that can extract huge margins, right?
[01:17:10.820 --> 01:17:15.360]  Facebook's cost of capital is, like, 9%, and its return on capital is, like, 50%.
[01:17:15.360 --> 01:17:20.900]  So basically, every time they raise a dollar, they stick it into operations, and they get $5 out.
[01:17:21.020 --> 01:17:22.860]  It's just like a machine for printing money.
[01:17:23.240 --> 01:17:36.320]  And, you know, most of that money ends up in their shareholders' pockets, but some of it goes for, you know, giant boiler rooms full of traumatized Filipino content moderators who watch beheading videos all day and are denied benefits.
[01:17:37.000 --> 01:17:48.800]  And, you know, the argument that if you can just... if you have those windfall profits, you can spend them on those boiler rooms, and that that won't be available to community groups is true.
[01:17:49.480 --> 01:17:55.300]  But on the other hand, this complexity of the problem is not linear.
[01:17:56.480 --> 01:18:13.480]  That trying to solve all the community's problems and trying to come up with a single standard for all the communities and all the spaces and all the contingencies, and to understand them and to manage that well, is a problem that gets harder as the groups get bigger.
[01:18:14.340 --> 01:18:23.300]  And once the problem is very large, once you're trying to come up with a coherent standard for billions of people, it becomes impossible.
[01:18:23.300 --> 01:18:26.360]  And so I think it erases the advantage of lots of money.
[01:18:26.560 --> 01:18:32.880]  And, of course, anyone who stays on Facebook will be subject to Facebook's moderation.
[01:18:32.880 --> 01:18:54.060]  And so if you find yourself somewhere else on the Internet, having left Facebook, and not liking the fact that this less resourced but also less complicated content moderation team is not protecting you adequately and you long for the tender embrace of Facebook's moderation policies,
[01:18:55.160 --> 01:18:56.420]  you know, nothing stops you.
[01:18:56.480 --> 01:18:57.640]  In fact, this makes it really easy.
[01:18:57.640 --> 01:19:03.700]  Like, one of the things that... the hegemony of Facebook and the blocking of interoperability...
[01:19:03.700 --> 01:19:06.100]  And, you know, for Facebook, read all the other social spaces as well.
[01:19:06.420 --> 01:19:11.320]  One of the things that it does is it means that we can't even figure out whether we might like something else better.
[01:19:11.580 --> 01:19:19.760]  We end up having these conversations entirely in the hypothetical, where we're not able to really, you know, give people the autonomy.
[01:19:20.560 --> 01:19:27.340]  And, you know, Facebook doesn't make it very hard to leave Facebook because they want to protect you.
[01:19:27.780 --> 01:19:37.720]  They make it very hard to leave Facebook because they understand that the harder it is to leave, the less they have to protect you and the more you'll put up with before you leave.
[01:19:37.720 --> 01:19:53.620]  And so, I do think that adding interoperability makes the content moderation both more possible, because we can make it more local, and for the large platforms, it aligns their incentives with making content moderation better.
[01:19:53.980 --> 01:19:58.420]  And the last thing I'll say about Como is that the...
[01:19:58.420 --> 01:19:59.660]  Oh, no.
[01:19:59.800 --> 01:20:00.700]  I just blew my buffer.
[01:20:00.880 --> 01:20:02.400]  What was I going to say about content moderation?
[01:20:02.400 --> 01:20:03.480]  Oh, yeah.
[01:20:04.140 --> 01:20:16.540]  It's that there's a failure mode of asking sort of content moderation cops to block harassment or other odious acts online, which is that you have to define them, right?
[01:20:16.620 --> 01:20:25.700]  You have to have... somewhere you have to tell your content moderators what constitutes bad acts online, what is beyond the pale, and what they have to block.
[01:20:25.900 --> 01:20:45.220]  And if you are a kind of professional harasser, you've got like every hour that God sent to figure out what is and is not harassment according to the three-ring binder or the consensus of the moderators, the policy of Facebook or whatever platform you're on.
[01:20:45.740 --> 01:20:54.540]  And you can devise a range of conduct that is almost but not quite harassment or almost but not quite a slur.
[01:20:54.540 --> 01:21:01.380]  And for the person receiving that conduct, it is indistinguishable from harassment or a slur.
[01:21:01.840 --> 01:21:05.100]  But for the content moderators, it's not, right?
[01:21:05.180 --> 01:21:11.580]  For the content moderators, you, as like an Internet lawyer, have figured out how to engage in conduct that they will permit.
[01:21:12.040 --> 01:21:20.420]  And then even worse is that because you know where the line is, you can goad your opponents into crossing the line.
[01:21:20.680 --> 01:21:27.960]  So you may remember that there is this moment where Facebook said, what was it?
[01:21:28.020 --> 01:21:33.440]  Kill all white men is not permissible, but kill all black children was.
[01:21:33.560 --> 01:21:39.420]  And there was something where it was like, it's because their age was not a protected category, but sex was.
[01:21:39.420 --> 01:21:41.960]  Some crazy thing like this.
[01:21:42.340 --> 01:21:54.220]  And if you know this weird kind of loophole in Facebook moderation policy, you can exploit it to goad your enemy into crossing the line.
[01:21:54.380 --> 01:21:59.720]  And then you can go to the moderator and you can say, look, this person is engaged in harassment.
[01:22:00.280 --> 01:22:01.100]  Suspend their account.
[01:22:01.280 --> 01:22:05.920]  And this is, in fact, the thing we see all the time on Twitter and Facebook where you have professional trolls.
[01:22:05.920 --> 01:22:16.220]  And so I think that like just structurally throwing more moderators at content moderation fails for like really important deep rules.
[01:22:16.280 --> 01:22:20.780]  And it starts to look like doing the same thing over and over again and expecting a different result.
[01:22:21.040 --> 01:22:30.220]  And I think we need to try different structures for moderation rather than trying to like moderate harder, right?
[01:22:30.340 --> 01:22:31.680]  Trying to spend more money on moderation.
[01:22:31.900 --> 01:22:33.320]  I hope that answers your question.
[01:22:33.460 --> 01:22:34.900]  It's a really good question.
[01:22:34.900 --> 01:22:46.460]  I really want to stress that because, you know, the life of people on the Internet who are on the wrong end of harassment campaigns is miserable in a way that really beggars description.
[01:22:46.460 --> 01:22:50.700]  And if you've never been subjected to it, you know, you can't really understand it.
[01:22:50.760 --> 01:22:53.420]  And it's not a consideration I take lightly.
[01:22:53.580 --> 01:22:58.640]  We really do need to keep that top of mind as we think about this interop stuff.
[01:23:00.920 --> 01:23:01.520]  All right.
[01:23:01.720 --> 01:23:04.780]  One question from our matrix chat, and then we'll get back into the live audience.
[01:23:06.600 --> 01:23:07.560]  From IDKI2P.
[01:23:07.820 --> 01:23:15.040]  What do you think will be the outcome of EU enforcement of the right of portability on the business viability of walled gardens and social media?
[01:23:15.340 --> 01:23:18.320]  Can walled gardens be legislated open by enforcing this right?
[01:23:19.420 --> 01:23:21.960]  So I think the portability writes not much.
[01:23:22.480 --> 01:23:26.940]  I think interoperability is, but portability is like really just...
[01:23:26.940 --> 01:23:31.120]  Like portability is super useful for static services, right?
[01:23:31.180 --> 01:23:40.160]  If you're like a Google Photos user and there's like a one-click button to become a Flickr user, say, that's pretty useful.
[01:23:41.500 --> 01:23:48.120]  Or, you know, if you're like trying to take care of the digital assets of, you know, a parent who has died.
[01:23:48.340 --> 01:23:52.260]  And you want to export all their posts and stick them on a thumb drive so you can look at them later.
[01:23:52.380 --> 01:23:53.520]  Portability is really useful.
[01:23:54.200 --> 01:23:57.860]  But most of our networked activity is interactive.
[01:23:58.500 --> 01:24:07.800]  You know, portability is nice if you're talking about things like, I can take my containerized cloud server and move it to a different host with minimal setup.
[01:24:07.800 --> 01:24:11.800]  But portability for an individual user gets you very, very little.
[01:24:12.220 --> 01:24:14.940]  You want to really maintain the relationship, right?
[01:24:15.020 --> 01:24:20.160]  Like if the reason you're on Facebook is that your customers are there, your community is there.
[01:24:20.620 --> 01:24:27.800]  Say you've moved across the country or across the world for work or university and the way you stay in touch with your family is through Facebook.
[01:24:28.100 --> 01:24:33.320]  I actually have some family in South America who are going through tough times right now and I'm trying to help them out.
[01:24:33.700 --> 01:24:37.740]  And everything there is done on WhatsApp and I'm a Zucker vegan and I don't have WhatsApp.
[01:24:37.740 --> 01:24:39.460]  Because I don't use any Facebook products.
[01:24:39.640 --> 01:24:40.600]  And it's really hard, right?
[01:24:40.720 --> 01:24:44.240]  Like that's the thing you lose when you leave a platform.
[01:24:44.460 --> 01:24:47.640]  Not just the data, but the ongoing interaction.
[01:24:48.280 --> 01:24:56.880]  And so that's why I think that interoperability and exposing APIs and allowing for real-time exchange of data is far more important than interop.
[01:24:57.000 --> 01:24:59.040]  Interop is good... or within data portability.
[01:24:59.200 --> 01:25:08.780]  Data portability would be a really good way if we had interop to leave Facebook and get set up on a Diaspora or a Mastodon instance and have an environment that has all your archives and stuff in it.
[01:25:08.860 --> 01:25:09.480]  That'd be great.
[01:25:11.300 --> 01:25:17.240]  But portability beyond that is... it's really like... it's data at rest, right?
[01:25:17.360 --> 01:25:19.560]  And what you need is the data in motion.
[01:25:21.220 --> 01:25:21.760]  All right.
[01:25:21.880 --> 01:25:22.940]  Next question from the audience.
[01:25:26.730 --> 01:25:29.870]  Corey, it's an honor to hear your talk today and super fan.
[01:25:29.870 --> 01:25:31.530]  So I'm going to try not to do that.
[01:25:32.610 --> 01:25:35.350]  You did a great job and you've been talking for a really long time.
[01:25:35.490 --> 01:25:36.690]  So here's my softball question.
[01:25:36.890 --> 01:25:38.250]  I've been staring at your chair.
[01:25:38.710 --> 01:25:40.450]  I have severe spinal issues.
[01:25:40.750 --> 01:25:42.750]  Can you tell me your kind of chair?
[01:25:45.090 --> 01:25:45.570]  Sure.
[01:25:45.970 --> 01:25:48.930]  It's a leap chair from Steelcase.
[01:25:49.750 --> 01:25:51.410]  It's about five or six years old.
[01:25:51.950 --> 01:25:54.510]  This is my unpaid endorsement for leap chairs.
[01:25:54.510 --> 01:25:59.510]  I had like fetishized leap chairs in magazines for a long time.
[01:25:59.910 --> 01:26:02.330]  And I tried them in showrooms, but I couldn't afford them.
[01:26:02.510 --> 01:26:07.390]  And in 2001, after the dot-com crash, I was walking through the mission in San Francisco.
[01:26:07.670 --> 01:26:14.390]  And there was a guy on the street with 50 leap chairs that he was selling for 50 bucks each, still in their plastic wrap.
[01:26:14.550 --> 01:26:18.450]  And he also had like 11 boxes of t-shirts for his failed dot-com.
[01:26:18.550 --> 01:26:19.250]  It was very weird.
[01:26:19.330 --> 01:26:25.610]  It was like millionaires selling apples from their Rolls Royces during the depression, except for really good ergonomic chairs.
[01:26:25.950 --> 01:26:31.150]  So I bought four of those and used them as my dining room chairs for like several years.
[01:26:31.310 --> 01:26:32.370]  Left San Francisco.
[01:26:32.970 --> 01:26:37.130]  Flew one of them to London and used it in London for the next like 13 years.
[01:26:37.790 --> 01:26:41.170]  Moved back here and had another one and moved that one back here.
[01:26:41.870 --> 01:26:43.130]  I forget where the other one went.
[01:26:43.230 --> 01:26:43.890]  Maybe it was in storage.
[01:26:44.070 --> 01:26:47.090]  Anyway, eventually, like after like literally 20 years, they fell apart.
[01:26:47.270 --> 01:26:48.570]  And then I bought this one.
[01:26:48.750 --> 01:26:50.310]  And I've had it for, you know, four or five years.
[01:26:50.490 --> 01:26:51.030]  And it's great.
[01:26:51.910 --> 01:26:55.450]  That is my unpaid endorsement of Steelcase Leap Chairs.
[01:26:55.610 --> 01:26:58.110]  And I sympathize with you for your spinal problems.
[01:26:58.310 --> 01:27:01.750]  I had both of my hips replaced in the last year.
[01:27:02.230 --> 01:27:05.310]  And if you go to the Internet Archive and search for Dr. O.
[01:27:05.430 --> 01:27:12.790]  Femur, you can get a high-res laser scan, Creative Commons Zero, of my diseased arthritic femur that you can print yourself.
[01:27:13.590 --> 01:27:14.050]  Awesome.
[01:27:14.670 --> 01:27:15.290]  Thank you.
[01:27:15.370 --> 01:27:16.230]  I am also a robot.
[01:27:16.230 --> 01:27:17.570]  I have an implanted device.
[01:27:17.810 --> 01:27:18.750]  Oh, very good.
[01:27:18.750 --> 01:27:19.810]  I have a lot of my issues.
[01:27:20.090 --> 01:27:21.470]  Thank you so much for your answer.
[01:27:21.630 --> 01:27:22.570]  Well, I'm glad to hear that.
[01:27:22.850 --> 01:27:23.010]  Yeah.
[01:27:23.010 --> 01:27:23.490]  Thank you.
[01:27:23.670 --> 01:27:23.870]  Cheers.
[01:27:24.730 --> 01:27:25.050]  All right.
[01:27:25.130 --> 01:27:25.990]  Another audience question.
[01:27:28.790 --> 01:27:29.430]  Hey, Corey.
[01:27:29.730 --> 01:27:30.510]  Thanks for the good talk.
[01:27:30.690 --> 01:27:34.670]  It's been great to see your ideas develop over the past couple of years on Interop.
[01:27:36.670 --> 01:27:44.410]  I, with Interop, I'm thinking a lot about the way we produce data on the web and the relationship between data production and interoperability.
[01:27:44.710 --> 01:27:56.290]  How do you view interoperability, how do you view that interoperability sort of will impact, or would impact, though we produce data and how we might gain access and control over it?
[01:27:56.410 --> 01:28:09.310]  I can see how Interop might do really well at increasing access to data production, the data we all produce, and it produces value for advertising industries and for capital as we know it right now.
[01:28:09.530 --> 01:28:15.390]  But I don't see a solution inside of Interop for increased control over my data production.
[01:28:15.850 --> 01:28:16.370]  Yeah.
[01:28:16.430 --> 01:28:18.710]  And relatedly, how it relates to the advertising industry.
[01:28:19.490 --> 01:28:22.450]  You're absolutely right, and I'm glad you brought that up.
[01:28:22.650 --> 01:28:29.750]  I wrote a paper a year and a half ago that has been updated once, and we're about to update it again, with my colleague at EFF, Bennett Cyphers.
[01:28:29.870 --> 01:28:34.370]  And we actually do have a technologist at EFF whose surname is Cyphers, which is just awesome.
[01:28:34.830 --> 01:28:37.010]  And it's called Privacy Without Monopoly.
[01:28:37.430 --> 01:28:51.250]  And it specifically addresses the relationship of interoperability and privacy, and the argument from large firms that interoperability presents data protection risks, which it absolutely does.
[01:28:52.010 --> 01:29:01.810]  And, you know, just as with the content moderation question, it's not wrong to say that the firm sometimes do things that benefit their users.
[01:29:02.170 --> 01:29:06.990]  The problem is that it's entirely at their discretion.
[01:29:07.430 --> 01:29:12.550]  So the data protection we get from Facebook, and we get a lot of data protection from Facebook.
[01:29:12.550 --> 01:29:23.330]  Facebook really does stop a lot of people from exfiltrating data from Facebook, is entirely dependent on whether Facebook thinks that that is a kind of data protection that you deserve.
[01:29:23.770 --> 01:29:29.450]  And the one kind of data protection that Facebook absolutely doesn't think you deserve is protection from its own data mining.
[01:29:31.130 --> 01:29:36.850]  They may stop the next Cambridge Analytica, but they won't stop the next internal project that exploits your data.
[01:29:37.250 --> 01:29:50.510]  And moreover, all of the firms, even the ones that have much better track records on data protection like Apple, routinely engage in bad data handling conduct when it is in their own interests.
[01:29:50.910 --> 01:30:03.930]  So, you know, Apple talks a big game about N10 encryption, and they did go to, you know, the court to fight the FBI when they were trying to backdoor iOS to get access to the San Bernardino shooters devices.
[01:30:04.610 --> 01:30:18.690]  But when the Chinese state said that they had to remove all the working VPNs from the App Store as a condition of remaining in the country so that the state could spy on Chinese users and, you know, put them in gulags.
[01:30:18.750 --> 01:30:24.810]  This was when a million Uyghurs and Turkic minorities in Xinjiang province were being imprisoned.
[01:30:25.790 --> 01:30:28.830]  You know, that Apple would go along with it, right?
[01:30:28.950 --> 01:30:43.350]  They weren't going to let people mine your data to sell you a fidget spinner, but they were happy to let the Chinese government mine your data to figure out whether to torture you if the alternative was losing access to Chinese manufacturing and Chinese consumers.
[01:30:43.350 --> 01:30:52.490]  And so we need something beyond the firm's goodwill and unilateral choices about when to protect our privacy.
[01:30:52.830 --> 01:31:00.690]  And, you know, ultimately, you safeguard your rights, not just with technology, but with technology backstopped by law.
[01:31:00.950 --> 01:31:25.690]  You know, as far back as the 1990s, cryptographers were recognizing the rubber hose cryptanalysis problem that it didn't really matter how long your passphrase was or how many bits your key was or how robust your cipher was if it was lawful for the state to tie you to a chair and hit you with a rubber hose until you revealed that passphrase and that adding more bits would not help.
[01:31:25.850 --> 01:31:40.870]  And so that ultimately the role of encryption in human rights would be to create a space that that was temporary at best where you could organize to demand human rights from the state.
[01:31:41.110 --> 01:32:05.350]  And that ultimately is how we protect privacy and data rights online is to create good free-standing digital human rights laws, a federal privacy law with a private right of action that applies equally to Facebook or Google or Apple or Microsoft or Salesforce or any other big tech companies and the interoperators that interoperate with them.
[01:32:05.850 --> 01:32:23.730]  You know, if Facebook says we're not going to allow this company to plug into the API that we've been mandated because we think they have bad data handling practices, we should be able to compare Facebook's claims to an empirical standard that says what is and is not permissible data handling.
[01:32:24.370 --> 01:32:30.270]  And the floor underneath it shouldn't be set by market forces, but rather by democratically accountable law.
[01:32:30.790 --> 01:32:53.370]  So, as I mentioned, we wrote this paper, Privacy Without Monopoly, and we wrote an appendix to it about how this plays out in the European Union with the GDPR and how upcoming European interoperability mandates in the Digital Markets Act and the Digital Services Act can interact or lean on the GDPR,
[01:32:53.430 --> 01:33:08.510]  the General Data Privacy Regulation, to create a regime in which we don't have to take the tech companies' word for it when they say that they are defending our privacy or that someone else shouldn't be allowed in because they're not going to defend our privacy.
[01:33:09.070 --> 01:33:24.790]  And I'm just finishing a draft, a second appendix to this for the Indian Law Review that makes reference to Indian privacy law, and it will talk about how this plays out in the Global South.
[01:33:25.150 --> 01:33:37.050]  So, I think you're right that, you know, this does create new data flows, and wherever there's new data flows, there's new opportunities for bad data handling practices and infringements on our data rights.
[01:33:37.250 --> 01:33:39.950]  And I think that we don't fix that with technology.
[01:33:40.090 --> 01:33:52.210]  We fix that by having democratically accountable entities that are stronger than the companies, that can force the companies to behave themselves, and by having remedies that are available to everyone through private rights of action.
[01:33:52.390 --> 01:34:00.510]  This is one of the things that's really stalled the federal privacy bill that has been on again, off again all summer in DC, is whether it will have a private right of action.
[01:34:00.630 --> 01:34:09.430]  And a private right of action is when, if your rights are violated under a law, you yourself can sue to have that law enforced.
[01:34:09.910 --> 01:34:31.950]  And, you know, more to the point, EFF can sue on your behalf, or the ACLU can, or, you know, Public Citizen can, or some other activist group can, as opposed to laws that don't have a private right of action, which are only enforced by prosecutors, where you have to hope that someone who works for the government decides that your rights are worth defending.
[01:34:32.250 --> 01:34:35.110]  And if they don't, then you're out of luck.
[01:34:35.570 --> 01:34:43.590]  And that is, I think, has been a major defect in a lot of privacy laws, and a lot of human rights laws more generally.
[01:34:43.750 --> 01:34:49.130]  And when you hear about human rights laws that have gone nowhere, it is often because they didn't have a private right of action.
[01:34:49.490 --> 01:35:02.530]  And, you know, the people who oppose private rights of action tend to make a lot of kind of tort reform noises about nuisance lawsuits, and how companies will attract all kinds of bottom-feeder 1-800-lawyer types.
[01:35:02.870 --> 01:35:14.850]  And, you know, there's ways to devise laws that prevent that from happening, you know, where you can have early showings of fact, like the slap laws that protect journalists and public speakers.
[01:35:15.270 --> 01:35:31.490]  You can have a rule that says, if you can show facts early on in the court case that say that you were doing something in the public interest, and that this is about shutting you up, then, you know, you are immunized from further litigation and you can get your fees back.
[01:35:31.850 --> 01:35:39.290]  That is a very powerful legal tool that stops even deep-pocketed adversaries from ruining you in court.
[01:35:39.430 --> 01:35:57.390]  I was recently... had a fight picked with me on Twitter by a thin-skinned billionaire who runs a credit card company in which I, you know, sort of made him look like a fool, and then his lawyers threatened to sue me unless I retracted my statements.
[01:35:57.390 --> 01:36:14.290]  And I was able to get a First Amendment lawyer to send them a very stiff letter reminding them that we all live in California, which is a very powerful anti-slap law, and discussing exactly how much money he could extract from them in fees if they were dumb enough to go ahead with this and how stupid they would look and what the Streisand effect was,
[01:36:14.290 --> 01:36:16.290]  and they just went away with their tail between their legs.
[01:36:16.430 --> 01:36:22.050]  So you can design laws that don't allow for this kind of abuse.
[01:36:22.430 --> 01:36:24.130]  It just takes careful design.
[01:36:24.350 --> 01:36:29.670]  And so I think that is ultimately how we defend our privacy rights, is by creating legal rights to privacy.
[01:36:29.810 --> 01:36:37.670]  Not just privacy technology, but legal rights that give you an area of redress if your privacy is violated.
[01:36:39.630 --> 01:36:40.230]  Thanks.
[01:36:40.910 --> 01:36:41.470]  All right.
[01:36:41.610 --> 01:36:43.230]  We have another Matrix question.
[01:36:43.750 --> 01:36:51.210]  Can you speak more to any ComCom implications for long-term digital asset preservation in terms of either public good, i.e.
[01:36:51.210 --> 01:36:53.250]  libraries, and the private good, i.e.
[01:36:53.290 --> 01:36:59.270]  families and her friends wishing to preserve the memories and digital artifacts of deceased individuals who've authorized their access?
[01:36:59.910 --> 01:37:01.230]  Yeah, 100%.
[01:37:01.230 --> 01:37:07.110]  You know, I got my start in the tech industry programming CD-ROMs for Voyager and HyperTalk.
[01:37:07.110 --> 01:37:07.810]  Right?
[01:37:08.050 --> 01:37:10.050]  Like, talk about dead platforms.
[01:37:10.330 --> 01:37:17.310]  But you can run all of those programs in a browser that can simulate a 68K Mac in JavaScript.
[01:37:17.950 --> 01:37:18.270]  Right?
[01:37:18.630 --> 01:37:25.030]  The implication of the universal Turing machine is that you can nest them as deep as you need to go.
[01:37:25.030 --> 01:37:27.170]  You can emulate any technology that you need.
[01:37:27.290 --> 01:37:31.830]  There's this weird story that we tell ourselves about obsolete data.
[01:37:32.590 --> 01:37:39.910]  And when you hear about obsolete data and like, oh, they couldn't recover this file anymore and all this data they created is lost.
[01:37:39.910 --> 01:37:44.190]  And you drill into the story, it's almost never about obsolete data.
[01:37:44.450 --> 01:37:51.850]  It's usually about obsolete storage and specifically about storage media that is fragile and has decomposed over the years.
[01:37:51.850 --> 01:38:05.550]  Because generally, if you've got data in a, you know, failure-resistant SSD that, you know, is part of an array that makes backups and so on, the data itself is pretty safe.
[01:38:05.550 --> 01:38:10.290]  And understanding the data structures, you know, if they're documented, it's very straightforward.
[01:38:10.430 --> 01:38:17.430]  But even if they're not documented, especially older data structures, because the computers themselves were so limited, those data structures are...
[01:38:17.430 --> 01:38:23.670]  I don't know, trivialize the work of the people who do it, but they're tractable to reverse engineer and build viewers for.
[01:38:23.950 --> 01:38:33.690]  And even when they're not, you can just fire up VMs that, you know, emulate the processor, run the operating system, and then launch the application.
[01:38:33.690 --> 01:38:49.510]  And, you know, the story of our data going bad is really a story about the fragility of media and the use of the law to block interoperability and not about the data itself.
[01:38:49.510 --> 01:38:51.330]  The data is effectively immortal.
[01:38:51.550 --> 01:39:01.730]  The logo programs I wrote on an Apple II Plus in 1979 run on the Ubuntu computer that I'm speaking to you from now, decades later.
[01:39:01.730 --> 01:39:03.810]  And they run very well.
[01:39:03.970 --> 01:39:04.810]  There's just no problem.
[01:39:04.930 --> 01:39:08.390]  I actually just had to do for a book cover...
[01:39:08.390 --> 01:39:11.070]  I was having a book cover discussion with my publisher.
[01:39:11.270 --> 01:39:13.870]  And I said, what if we replicate an old browser window?
[01:39:13.870 --> 01:39:26.850]  And I went and I launched in my own browser window, I launched a version of NCSA Mosaic from the early 1990s.
[01:39:26.910 --> 01:39:34.370]  And then it retrieved a web page from the Internet Archive that was contemporaneous with it and rendered it.
[01:39:34.370 --> 01:39:44.390]  And then I could like toggle back and forth between different browser renderers, different VMs, different captures of this file and see how it rendered in all these different ways.
[01:39:44.650 --> 01:39:50.990]  And it was all being done without pegging even one of the cores on my little consumer laptop.
[01:39:51.950 --> 01:40:01.130]  You know, the marvelous thing about the incredible increasing power of computation is that simulating older computers, it doesn't have to be done very efficiently.
[01:40:01.470 --> 01:40:03.950]  That you have lots and lots and lots of headroom.
[01:40:04.150 --> 01:40:10.510]  You know, I just saw a headline that was so weird and good that I couldn't bring myself to go and find out what it meant.
[01:40:10.610 --> 01:40:14.210]  But it was, man figures out how to play Doom inside of Doom.
[01:40:14.210 --> 01:40:19.330]  Which I assume means that, like, they implemented Doom on a VM inside of Doom or something.
[01:40:19.670 --> 01:40:27.030]  And that's the amazing and wonderful and magical thing about living in an era of exponential increase in computing power.
[01:40:27.750 --> 01:40:33.170]  Is that, you know, you can kind of suck at making emulators and still emulate very well.
[01:40:33.410 --> 01:40:37.750]  So yeah, I think that it's really interoperability is key to our preservation.
[01:40:37.750 --> 01:40:42.290]  Because it's not interoperability to open the file and save it again in a new file format.
[01:40:42.290 --> 01:40:45.910]  Or it's not preservation, rather, to open the file and save it again in a new file format.
[01:40:45.990 --> 01:40:49.250]  That's like saying, well, we saved the Mona Lisa by photocopying it, right?
[01:40:49.650 --> 01:40:54.390]  Preservation is seeing things in their original context as well as in modern ones.
[01:40:54.450 --> 01:41:01.330]  And, you know, the best chance we have of recovering those original contexts and sharing them widely is through emulation.
[01:41:02.730 --> 01:41:03.690]  All right, thank you.
[01:41:03.910 --> 01:41:05.010]  To another live question.
[01:41:08.160 --> 01:41:09.860]  Thank you very much for the great talk.
[01:41:09.860 --> 01:41:26.800]  So if somebody 100% understands and believes in your vision, and they had a lot of money to be able to spend, whether through grants to open source projects or investments in alternative platforms, where should they be putting their money so that they could see your vision come to fruition?
[01:41:26.800 --> 01:41:34.840]  Because I'm a director at an engineering and venture capital and philanthropy organization that has helped my organization in the past.
[01:41:34.860 --> 01:41:42.600]  And we want to be able to see this kind of future of more people having freedom over their devices, freedom over where they post their content and platforms into the future.
[01:41:42.600 --> 01:41:52.620]  And they really want to know where would their money best be spent to see a positive outcome here, if their concern is a positive outcome and a freedom-loving outcome, more so than just making their money back.
[01:41:52.620 --> 01:41:54.720]  Yeah, that's a great question.
[01:41:55.140 --> 01:42:01.020]  And those theory of change questions, they're really gnarly and they keep me up at night.
[01:42:03.820 --> 01:42:09.880]  So Lawrence Lessig has this idea that our world is governed by four forces.
[01:42:10.260 --> 01:42:13.260]  Code, which is like what's technologically possible.
[01:42:13.600 --> 01:42:15.800]  Law, what's lawfully permissible.
[01:42:16.300 --> 01:42:18.060]  Norms, what's socially acceptable.
[01:42:18.600 --> 01:42:20.500]  And markets, what's profitable.
[01:42:21.940 --> 01:42:24.520]  And you're kind of saying, well, what if we took the markets out of it?
[01:42:24.640 --> 01:42:26.220]  What are the other interventions we could make?
[01:42:26.320 --> 01:42:28.720]  Although, maybe markets too, depending.
[01:42:30.600 --> 01:42:45.600]  And, you know, one of the implications for me about this idea that we have these sort of four cardinal directions that we can move in, is that if you find yourself unable to make an intervention intervention in the domain that you're most comfortable with,
[01:42:45.700 --> 01:42:45.740]  right?
[01:42:45.800 --> 01:42:46.520]  Like if you just...
[01:42:46.520 --> 01:43:06.440]  If every program you want to write that would make things better, you know, interoperability-wise, would just land you in jail, then maybe what you need to do is instead write, you know, write checks to lawmakers or to activist groups that are lobbying to reform those laws.
[01:43:06.440 --> 01:43:12.380]  Or maybe what you need to do is launch a publicity campaign that makes people aware of what they're losing.
[01:43:12.520 --> 01:43:16.420]  Because, you know, people just don't understand what an interoperable Internet would look like.
[01:43:16.540 --> 01:43:19.880]  There's a whole generation that has just grown up within these walled gardens.
[01:43:22.780 --> 01:43:31.380]  I spent a lot of time when the European Union announced that they were going to start their interoperability mandates in the Digital Markets Act with end-to-end encrypted messengers.
[01:43:31.640 --> 01:43:36.660]  I spent a lot of time asking myself, like, why are they doing this with E2E messengers?
[01:43:37.540 --> 01:43:46.380]  And I concluded, after talking to some people in Brussels, that it's because, like, none of them have ever experienced federated conversation.
[01:43:46.580 --> 01:43:49.340]  Like, none of them were ever, you know, on USENET, say, or FidoNet.
[01:43:49.340 --> 01:43:51.480]  None of them, like, they literally just can't imagine it.
[01:43:51.800 --> 01:43:54.340]  Whereas, they've all got SMS, right?
[01:43:54.380 --> 01:44:03.140]  They all know what it's like to be, like, a German lawmaker in Brussels who takes out your phone and sends a text to, you know, a Dutch colleague on holiday in Spain.
[01:44:03.760 --> 01:44:05.780]  And, you know, it just works.
[01:44:05.940 --> 01:44:08.760]  And they're like, how hard can it be to build an interoperable messenger?
[01:44:09.100 --> 01:44:13.500]  What they're missing is that, you know, the reason we want to replace SMS is because it's a dumpster fire.
[01:44:13.860 --> 01:44:17.200]  And, you know, SMS works great if you don't care about how it fails.
[01:44:17.200 --> 01:44:19.680]  If you do care about how it fails, SMS is terrible.
[01:44:19.960 --> 01:44:34.580]  And so, you know, you might want to make a public information campaign to try and make people aware of what they've lost through a couple of lost decades of large firms being able to foreclose on new market entrants and interoperability.
[01:44:34.580 --> 01:44:45.160]  And so, you know, you also might want to make grants to tech nonprofits that are doing interoperability work.
[01:44:45.240 --> 01:44:57.820]  So, for example, the main core developer on Pigeon, which is, you know, the latest incarnation of XMPP, which holds together a bunch of other protocols and works with Matrix, which is a great platform.
[01:44:57.820 --> 01:45:02.480]  I understand we're using it here today, but Pigeon's a really interesting example because there's one developer.
[01:45:02.800 --> 01:45:07.260]  He was, like, working for big tech companies and he saved up a bunch of money.
[01:45:07.300 --> 01:45:11.980]  And he was like, if I move back to the Midwest, I can live on, like, $18,000 a year.
[01:45:12.100 --> 01:45:18.960]  And I can spend, like, the next three years just hacking on XMPP, just hacking on Pigeon, making it better, which is what he's doing.
[01:45:20.160 --> 01:45:21.840]  And he's running out of money.
[01:45:22.080 --> 01:45:41.780]  And the thing that really got to him was that one of the companies that makes cyber weapons, you know, a company that makes tools for lawful interception like the NSO Group, posted a $100,000 bounty for bugs in his code so that they could write exploits for it and spy on him,
[01:45:42.440 --> 01:45:44.460]  and spy on his millions of users.
[01:45:44.600 --> 01:45:46.620]  So this one guy is supporting millions of users.
[01:45:47.060 --> 01:46:02.260]  And so, you know, if you wanted to spread your money around, you could go and you could find these individuals who are maintaining this public interest Internet code and just give them enough money to hire a couple of colleagues and, you know, pay their bills and not have to worry about these $100,000
[01:46:02.260 --> 01:46:06.520]  bounties, you know, not being horribly outmatched by these $100,000 bounties.
[01:46:06.660 --> 01:46:08.520]  And there are a bunch of these little projects.
[01:46:08.600 --> 01:46:30.340]  You may remember after Snowden came in from the cold, and it was revealed that he'd used Enigma as the front-end for GPG for his email, that there was like one guy in Germany who worked on it as a hobby and that everyone who was using GPG with Enigma in the world was reliant on this one person's spare time to make it work.
[01:46:30.520 --> 01:46:38.220]  And, you know, to their credit, a bunch of firms that use those tools then threw some money at him so that he's got like... I think he's got a helper now.
[01:46:38.380 --> 01:46:39.880]  But still, it's just a couple of people.
[01:46:40.060 --> 01:46:43.720]  So there are a bunch of those projects you could track down and throw some money at.
[01:46:45.840 --> 01:46:50.600]  On the public information side, there's a lot more of this stuff going on now.
[01:46:50.780 --> 01:46:54.300]  And certainly, like, there are big tech-funded PACs.
[01:46:54.440 --> 01:47:12.360]  Like, Facebook just ran these incredible scare ads about the Access Act and other interoperability and antitrust laws in Congress, where it's just like, if we allow, you know, the American Congress to weaken Facebook, China will invade America, right?
[01:47:12.480 --> 01:47:14.540]  And that was like... that was their whole pitch.
[01:47:14.760 --> 01:47:24.140]  And, you know, certainly, like, finding organizations that want to run counter-ads to that would be really good, especially if you focus them in D.C.
[01:47:24.280 --> 01:47:32.680]  and, you know, in swing states and in races where it mattered and in races for the chairmanship of the committee, relevant committees and so on.
[01:47:32.840 --> 01:47:35.240]  And that would be a good way to affect norms.
[01:47:36.320 --> 01:47:42.420]  If you wanted to affect law, you know, there are... there's...
[01:47:43.680 --> 01:47:56.220]  there's going to be a lot of need for funding public interest individuals and groups to participate in standards-making around the Access Act and the Digital Markets Act.
[01:47:56.380 --> 01:47:58.640]  So, you know, all of these... these tech mandates are going to...
[01:47:58.640 --> 01:48:04.180]  are going to result in some kind of standard-setting organization, some kind of standard-setting effort.
[01:48:04.460 --> 01:48:16.940]  And, you know, this is the kind of incredibly boring, off-screen, super important, deep structural work that, if it goes wrong, is going to be really hard to remediate.
[01:48:16.940 --> 01:48:36.500]  And so, you know, if... if the committee that writes the standard for interoperable Facebook is, like, 400 Facebook lawyers and engineers, one, you know, overseer from NIST, and, like, two academics, and one person from a startup, that standard is going to be completely useless.
[01:48:36.820 --> 01:48:43.240]  So creating grants to send people to those committees is... is an area where you could make a huge difference.
[01:48:43.800 --> 01:48:50.860]  And having done a lot of time as... in public interest standard setting, it is, like, the most thankless and impossible work.
[01:48:51.020 --> 01:48:52.460]  And you're so outgunned.
[01:48:52.560 --> 01:48:56.040]  And having more money to bring people in would have made a huge difference.
[01:48:56.280 --> 01:49:06.160]  You know, there were... there was often the case where I would have loved to have brought in an academic or an activist who works on disability rights to come and participate in standard settings.
[01:49:06.820 --> 01:49:08.760]  But, you know, I couldn't pay them.
[01:49:08.960 --> 01:49:12.700]  And they needed... they couldn't afford to volunteer their time.
[01:49:12.940 --> 01:49:17.320]  So there... that's an area where you could make a really big difference on the... on the tech side.
[01:49:17.800 --> 01:49:19.000]  So let's see, where are we at?
[01:49:19.180 --> 01:49:23.440]  Norms, laws, code, and markets.
[01:49:23.700 --> 01:49:25.020]  So... and market participants.
[01:49:25.220 --> 01:49:29.460]  So there are firms that are trying to do really interesting things with Interop.
[01:49:30.180 --> 01:49:33.620]  I've been talking with these folks at a company called Para.
[01:49:33.880 --> 01:49:34.420]  They're a startup.
[01:49:34.920 --> 01:49:43.000]  They're funded by Union Square Ventures, the folks who backed Kickstarter and a bunch of other startups that have gone on to, like, do sort of public interest work.
[01:49:43.420 --> 01:49:48.500]  And Para, they reverse engineered the DoorDash app that Dashers use.
[01:49:48.640 --> 01:49:51.180]  That's what DoorDash is cute named for their gig workers.
[01:49:51.920 --> 01:49:55.940]  And the app hid the tip from them.
[01:49:55.940 --> 01:50:04.340]  So if you booked a food delivery and you said, you know, I'll pay like a $10 tip, the driver wouldn't be able to see the tip until they made the delivery.
[01:50:04.600 --> 01:50:15.520]  And what that meant, because so much of the compensation for the driver comes out of the tip, they wouldn't know if the run was profitable until after they made the run, right?
[01:50:15.700 --> 01:50:24.420]  And that was a way for DoorDash to get their drivers to subsidize unprofitable runs.
[01:50:24.420 --> 01:50:31.280]  So DoorDash wanted to, you know, blitz scale, they wanted to grow their business by offering below cost deliveries to customers.
[01:50:31.460 --> 01:50:37.860]  Because, like, who wouldn't use a service that costs you less to use than it costs to actually deliver?
[01:50:38.220 --> 01:50:44.440]  And by hiding the tip amounts from the drivers, they were able to, like, effectively operate them like blind boxes, right?
[01:50:44.640 --> 01:50:48.560]  You know, and turn them into, like, make doing runs into a slot machine.
[01:50:48.720 --> 01:50:51.600]  Maybe this run will make you a lot of money or maybe you'll lose $4.
[01:50:51.600 --> 01:50:52.380]  Who knows?
[01:50:52.800 --> 01:51:04.140]  And it turned out that the way that they relayed these messages to the DoorDash app that the drivers had is they would send them, like, a JSON file or an XML file.
[01:51:04.340 --> 01:51:07.800]  And within it, it would have, like, the full tip amount field.
[01:51:08.080 --> 01:51:10.640]  And then it would have a flag that said, do not display.
[01:51:11.260 --> 01:51:19.680]  And so, Para wrote an app that would intercept that message and display the tip amount that was otherwise hidden.
[01:51:20.060 --> 01:51:22.840]  And DoorDash accused them of hacking the app.
[01:51:23.000 --> 01:51:26.080]  They said they were going to steal drivers' social security numbers with it.
[01:51:26.180 --> 01:51:28.280]  They said they were violating the DMCA.
[01:51:28.640 --> 01:51:30.720]  They just, like, went all out on them.
[01:51:30.900 --> 01:51:37.220]  And then they finally, like, did the one thing that actually would work, which is they stopped transmitting the full tip amount in the clear.
[01:51:37.220 --> 01:51:38.660]  And now they encrypt it.
[01:51:39.060 --> 01:51:48.080]  But Para and DoorDash are now actually, like, locked in this war with one another where every time DoorDash ships an update, Para ships an update to break it.
[01:51:48.240 --> 01:51:49.220]  And they just did...
[01:51:49.220 --> 01:51:55.640]  Para just did some public communications around this where they said, like, we are now facing, like, the full engineering might of DoorDash.
[01:51:55.780 --> 01:52:06.340]  They've now made their number one engineering priority, making it so that drivers can't figure out what they're getting paid and blocking our app, and they're throwing more and more engineering time at it.
[01:52:06.420 --> 01:52:10.180]  We, you know, we might reach the edge of our ability to do this.
[01:52:10.420 --> 01:52:24.860]  You know, if you wanted to throw some money at a firm, not for profit, but for just to help them do the right thing, you could give them enough money to hire more engineers because, you know, in general they have an interoperator's defense or an interoperator's advantage,
[01:52:25.380 --> 01:52:29.600]  right, which is that DoorDash has to make no mistakes and they just have to find one mistake.
[01:52:30.000 --> 01:52:32.880]  But, you know, DoorDash has billions of dollars in capital.
[01:52:34.000 --> 01:52:44.440]  They can just throw a lot of engineers at patching and repatching and repatching, and it becomes a war of attrition, where still, Para has the advantage if they can access capital as well.
[01:52:44.480 --> 01:52:47.060]  So that might be an area where you could work in markets to do things.
[01:52:47.300 --> 01:52:49.480]  So I don't think there's just, like, one way to get there.
[01:52:49.960 --> 01:52:55.140]  I think you go as far as you can with code, and then if you hit a wall, you're like, well, what can I do with norms?
[01:52:55.180 --> 01:52:56.160]  Or what can I do with markets?
[01:52:56.220 --> 01:52:57.320]  Or what can I do with law?
[01:52:57.320 --> 01:53:01.140]  And then if, you know, the legal process stalls out, you're like, well, what code can I write?
[01:53:01.680 --> 01:53:04.460]  And going around and around like that is how you make the difference.
[01:53:05.120 --> 01:53:05.400]  Thank you.
[01:53:06.940 --> 01:53:09.540]  All right, I know we're eight minutes before the hour, Corey.
[01:53:09.780 --> 01:53:11.060]  Do you want to take one more question?
[01:53:11.300 --> 01:53:12.160]  Yeah, that'd be fine.
[01:53:12.340 --> 01:53:13.880]  Okay, we got one last question from the audience.
[01:53:17.360 --> 01:53:17.760]  Hello.
[01:53:17.980 --> 01:53:24.400]  This is a question about the upstream incentives for interoperability and how technology affects those upstream incentives.
[01:53:24.400 --> 01:53:33.320]  So with distributed autonomous organizations, projects like Erbit, other things where you have notions of digital sovereignty being promoted very intensely.
[01:53:36.660 --> 01:53:41.280]  Do you see this as being a possible route forwards for interoperability?
[01:53:41.440 --> 01:53:47.800]  Or do you think that they really have to attack, not attack, really have to address it with the current super majors?
[01:53:47.800 --> 01:53:56.300]  Or can future technological opportunities dissolve the network effect monopoly, which current social media majors have?
[01:53:57.660 --> 01:54:02.660]  Well, I think that the nature of the institution is less important.
[01:54:02.820 --> 01:54:05.980]  I mean, you could say, well, what if it's a DAO?
[01:54:06.120 --> 01:54:09.080]  And I know that like DAO can encapsulate lots of different kinds of organizations.
[01:54:09.280 --> 01:54:12.940]  But like, let's say it's a good DAO full of public interested people with a good governance model.
[01:54:13.200 --> 01:54:18.520]  You could replace that with a cooperative or a nonprofit or, you know, a group of tinkers or whatever.
[01:54:18.740 --> 01:54:21.700]  They would still all have the same problem, which is switching costs.
[01:54:21.920 --> 01:54:32.240]  That until we address the interoperability question, the fact that everyone is already on these big platforms inside these walled gardens makes it a non-starter.
[01:54:32.540 --> 01:54:34.960]  You know, when Facebook started...
[01:54:35.740 --> 01:54:37.740]  Not when they started, but when they went public, right?
[01:54:37.740 --> 01:54:44.580]  When they went beyond universities and said anyone can join even if you don't have a .edu account, you can get a Facebook account.
[01:54:44.580 --> 01:54:56.100]  They had a really big problem, which was that everyone who wanted to use a general-purpose social media network already had an account on MySpace, which was owned by this rapacious billionaire, Rupert Murdoch, who was not going to let them go, right?
[01:54:56.140 --> 01:54:57.540]  He had them locked in pretty thoroughly.
[01:54:58.120 --> 01:55:00.120]  And, you know, what Facebook did is they...
[01:55:00.120 --> 01:55:10.380]  Rather than saying to people, we have a better service, all that it lacks is everyone you want to talk to, why don't you just come here and twiddle your thumbs until they decide to leave?
[01:55:10.380 --> 01:55:16.820]  Or trying to organize a, like, international everyone quits MySpace and goes to Facebook day next Wednesday at 2 p.m.
[01:55:17.940 --> 01:55:33.420]  What they did was they created, like, a bot that you could give your MySpace credentials to, switch to Facebook, and it would go log into MySpace's use, scrape your inbox, stick them in your Facebook inbox, scrape your communities, stick them in your Facebook communities,
[01:55:33.740 --> 01:55:38.240]  let you reply to them, and then it would, like, autopilot those responses back out again to MySpace.
[01:55:38.830 --> 01:55:43.960]  So it meant that you didn't have to choose between using the service that was run by people you liked more.
[01:55:44.000 --> 01:55:50.420]  And, you know, we forget this, but in the early days, Facebook's pitch was, we're the privacy-respecting alternative.
[01:55:50.840 --> 01:55:53.940]  And MySpace will sell your data, Facebook won't.
[01:55:54.080 --> 01:55:56.940]  That was their promise when they lured people in.
[01:55:58.410 --> 01:56:05.800]  And you didn't have to choose between the privacy-respecting community or the privacy-respecting platform and the communities you belong to.
[01:56:05.920 --> 01:56:09.240]  You could access the communities you belong to from the privacy-respecting platform.
[01:56:09.640 --> 01:56:15.500]  And, you know, as a technical matter, I think it's not hard to build interoperable bridges into these services.
[01:56:15.500 --> 01:56:21.740]  We can, you know, hack them in using bots, using scrapers, using reverse engineering.
[01:56:21.740 --> 01:56:32.400]  But to make that sustainable, we also have to make it legal, because otherwise they will skip past the building full of engineers, go to the building full of lawyers, and reduce you to a smoking crater.
[01:56:33.420 --> 01:56:35.880]  And if we can make the...
[01:56:36.560 --> 01:57:00.140]  If we can put you on a sound legal footing, then you can build the bridge, whether you're running a DAO or co-op or whatever, you can build a bridge that lets people have one foot in Facebook and one foot in your new service, and not have to choose between the people that matter to them and a service whose operating ethos is more preferable to them.
[01:57:00.620 --> 01:57:03.640]  You know, my grandmother was a Soviet refugee.
[01:57:03.840 --> 01:57:08.500]  She was a child civil defense worker, basically a child soldier in the Siege of Leningrad.
[01:57:08.720 --> 01:57:18.440]  And, you know, for three years hauled bodies and brought ammunition up to the front, and then went to Siberia over the winter ice, met my grandfather, got knocked up.
[01:57:18.440 --> 01:57:20.520]  They deserted and went to Azerbaijan.
[01:57:20.520 --> 01:57:21.680]  My dad was born.
[01:57:21.900 --> 01:57:26.220]  And then they crossed Europe, got to a displaced person boat and came to Canada.
[01:57:26.400 --> 01:57:27.780]  That's how I became a Canadian.
[01:57:27.980 --> 01:57:29.740]  That's why I wasn't raised in Russia.
[01:57:30.000 --> 01:57:31.580]  And my grandmother was the only one.
[01:57:31.720 --> 01:57:33.260]  Her family stayed behind in Leningrad.
[01:57:34.280 --> 01:57:37.120]  And she lost contact with them for 15 years.
[01:57:37.340 --> 01:57:47.520]  My dad has this incredible story about, you know, one day the phone rang in their house in Toronto, and my grandmother picked it up and said, mama, mama, and burst into tears.
[01:57:48.020 --> 01:57:50.260]  She didn't know if her mother was dead or alive.
[01:57:50.920 --> 01:57:54.780]  And, you know, I often asked her, like, why did you go?
[01:57:55.620 --> 01:57:59.320]  And I often wondered why the other ones stayed.
[01:58:00.160 --> 01:58:04.620]  And she said, you know, they stayed because leaving meant losing everything.
[01:58:04.620 --> 01:58:12.920]  Everything you owned, every social connection you had, maybe never making contact with everyone you loved in the world, losing them forever.
[01:58:13.220 --> 01:58:13.660]  Right?
[01:58:13.740 --> 01:58:16.480]  The switching costs of leaving the Soviet Union were really high.
[01:58:16.780 --> 01:58:22.000]  You know, and in the chaos after the war, a lot of people could have done what my grandparents did and didn't.
[01:58:22.000 --> 01:58:33.120]  They stayed put because even though there were enormous problems with staying where they were, the cost of leaving was so high that they couldn't bear it.
[01:58:33.400 --> 01:58:46.120]  And even if they had cause to regret that choice later on, it just got worse and worse, harder and harder to leave because they were in this mutual hostage taking situation where I couldn't leave because you wouldn't leave and you couldn't leave because I wouldn't leave.
[01:58:47.000 --> 01:58:50.400]  And giving people the tools to leave easily.
[01:58:50.680 --> 01:58:50.760]  Right?
[01:58:50.880 --> 01:59:04.100]  If you think about what the relatively open borders of the European Union mean today, where, you know, if you leave Berlin, like the former East Germany used to say that the reason the Berlin Wall was there was not to keep East Germans in.
[01:59:04.180 --> 01:59:06.540]  It was to keep everyone out of the workers' paradise.
[01:59:06.940 --> 01:59:07.040]  Right?
[01:59:07.720 --> 01:59:14.140]  Today, if you grew up in Berlin, you can like try going to Paris and getting a job in a cafe.
[01:59:14.800 --> 01:59:22.460]  And, you know, you can like zoom with your family every night and you can stay on SMS or, you know, WhatsApp and your family groups.
[01:59:22.480 --> 01:59:27.040]  And you can even tune into the radio station back home and just stream it wherever you are.
[01:59:27.180 --> 01:59:28.700]  And if you don't like it, you can go back.
[01:59:28.880 --> 01:59:28.960]  Right?
[01:59:29.020 --> 01:59:30.220]  You just get on the train, you go back.
[01:59:30.340 --> 01:59:31.000]  It's really easy.
[01:59:31.540 --> 01:59:34.780]  And you compare that to what it's like when the switching costs are high.
[01:59:35.440 --> 01:59:38.600]  And the extent to which people end up choosing things that they don't like.
[01:59:38.720 --> 01:59:38.780]  Right?
[01:59:38.840 --> 01:59:40.060]  That aren't optimal for them.
[01:59:40.600 --> 01:59:44.680]  It's like people continuing to smoke and saying, I wish I could quit these things.
[01:59:44.700 --> 01:59:46.620]  But the switching costs are so high.
[01:59:46.780 --> 01:59:54.280]  When the switching costs are high, people get stuck in situations that make them miserable, that bring them to harm every day.
[01:59:54.700 --> 01:59:58.580]  And building a better alternative somewhere else doesn't change the switching costs.
[01:59:58.860 --> 02:00:00.720]  You have to knock down the walls.
[02:00:00.720 --> 02:00:05.440]  And the way that we're going to knock down the walls is with a combination of technology and law.
[02:00:05.720 --> 02:00:11.380]  We're going to make the world safe for hackers who want to open things up and give users autonomy.
[02:00:11.380 --> 02:00:24.880]  And then the people like you at this conference, who have spent so many years and written so many words and done so many talks about how we can seize the means of computation.
[02:00:25.320 --> 02:00:28.320]  How we can take control of our technological destiny.
[02:00:29.550 --> 02:00:34.540]  You'll help those people get the technological freedom that they deserve.
[02:00:34.540 --> 02:00:36.340]  And that's, I think, how we get there.
[02:00:36.460 --> 02:00:41.600]  And it's not the organizational structure that is the thing that will make the difference.
[02:00:42.460 --> 02:00:46.220]  It's that meme of the two clasped hands.
[02:00:46.400 --> 02:00:52.820]  And it's nerds and legal activists working together to make the world safe for a better Internet.
[02:00:55.980 --> 02:00:56.500]  All right.
[02:00:56.580 --> 02:00:58.480]  Thank you so much for your fascinating talk.
[02:01:00.400 --> 02:01:00.800]  Thank you.
[02:01:00.800 --> 02:01:01.700]  Thank you so much.
[02:01:02.180 --> 02:01:05.220]  I hope to see you all in a couple of years out there in Long Island.
[02:01:15.280 --> 02:01:19.640]  And a round of applause for you, the audience, for your participation, for your questions, for your interests.
[02:01:19.960 --> 02:01:20.440]  Thank you.
[02:01:25.750 --> 02:01:26.270]  All right.
[02:01:26.330 --> 02:01:27.710]  So this was the last talk of the day.
[02:01:27.930 --> 02:01:28.410]  Stick around.
[02:01:28.590 --> 02:01:32.050]  Shortly, they'll be doing the closing remarks with everyone.
[02:01:32.610 --> 02:01:36.510]  And the Hacker Store, 2600 Store, will still be open during the closing remarks.
[02:01:36.530 --> 02:01:39.270]  So please stop by and last chance for swag.
[02:01:39.270 --> 02:01:40.670]  So thank you, everyone.