Brave’s Brendan Eich on Fixing Online Advertising


Peter McCormack: Hi Brendan, how are you?

Brendan Eich: I’m well thank you.

Peter McCormack: Okay great. So, there’s kind of three things I want to talk about with you today. I want to talk about privacy. I want to talk about Brave. And I want to talk about the BAT. I don’t want to do too much of a long background I think people know your rich history in the internet and that you created JavaScript. Was it like ten days?

Brendan Eich: In May 1995. And then you know, embedded in Netscape 2 in over the next two and half months and then it shipped.

Peter McCormack: Right. Okay. So what it feels like is when I look back at your history when you founded Mozilla or were working on Mozilla it was kind of a fight against Microsoft and now with Brave, this is a fight against Google. Do I see that right?

Brendan Eich: Google is certainly a big, a good example. People are also picking on Facebook a lot. Facebook’s making itself easy to pick on through, I think, sticking to its guns too much. And doing things like, you know, they took your phone number for securing your login and then they use that for ad targeting. That was a mistake. So it’s against the whole surveillance super powers of modern ad tech and social media. Ultimately also depends on advertising.

Peter McCormack: Right. And you obviously were kind of a key person during the early days of the internet. Are you seeing the growth in crypto currencies? Does that feel like a similar kind of time as when you’re working on the early browsers?

Brendan Eich: In some ways similar and in some ways it’s quite different because the early days of the web, people at first couldn’t do spirit transactions. Once they could, it quickly went to credit cards. And that made sense for lots of reasons but it also meant it was centralized, there was an interchange charge, fraud costs were put back on the merchants. So there were issues there that I think crypto currency’s promised to relieve. I think that’s the great thing about crypto, is its just censorship resistant and blockchains and things like that. The flip side I suppose is if you’re doing peer to peer messaging using block chain sort of based systems that are reaching consensus through some kind of protocol it’s harder. They tend to be slower to settle, there’s ways to get around that. New chains coming out are super-fast but not really widely adopted yet and the centralization has an intrinsic cost, you have to think differently, and most people still don’t know the difference between a browser and a search engine so it’s not clear when all this stuff will become as usable as the web became rather quickly in the nineties.

Brendan Eich: The browser was already out when Netscape made it even easier to use and then added security for e-commerce through encryption and people were used to some kind of a fat app that you click in the user interface of but the 10 blue links, and the search page, or the clicking on a link, or having some kind of a web form you fill out was not that far from the other metaphor that proceeded it. The great thing about the web was that it got anybody to be able to talk to anybody else, put up a sight, put up pictures of their cats, and even put up small shops that became big. To see that kind of explosion with BAT would be interesting but BAT is still very, very young.

Peter McCormack: Right okay. And having worked in the internet industry for something like 30 odd years now, just before we start talking about Brave, from your lens, how have you seen the evolution of the internet, both good and bad?

Brendan Eich: Well, so networks tend, in any system, in configuration space and in biology, biological system, you can get first place, second place winners, and they tend to take over and maybe they do really well for a long time. They use their advantages in the one field where they proper to take over others. Systems can get out of balance for a long time and stay out of equilibrium. People always assume equilibrium, and sorry, it’s a really stupid economics issue, everything’s at equilibrium and it’s not. In physics and biology, it’s really hard to study non-equilibrium systems so it’s understudied and that’s- life is a non-equilibrium system among other things, so there’s a bunch we don’t know and it’s hard to study and we don’t get necessarily fast rewards or 10 year pats as a researcher, this is true on computer networks.

Brendan Eich: Computer networks, easy money, and some of it’s not easy to do, let’s not minimize the effort but after what Friendster and MySpace, starting with the universities, had the right value prepositions at the right times and place and you’ve got a bunch of people on board Facebook, and you could say, some people argue that engagements dropping now but he used that great power with the main app to acquire other apps, Instagram, WhatsApp, so it’s just sort of like taking over adjacent systems or some sort of partial symbiosis in biology. We tend to have first and second place winner perpetuate themselves. This is not always for the best for society or for the whole ecosystem and I think that’s happening now, we’re living through it, you’re seeing regulators, in especially Europe and other places, India, Brazil, California. Our reaction to this, people are quite upset because a lot of people believe that Facebook is sort of a sci-op platform, had undo effects on elections, and it could have, it’s hard to prove a negative but it just has a dark side to it.

Brendan Eich: It’s not just about whether they tilted the election, it’s people get mobbed, the get trolled, they get ads they don’t like, they get malware distributed as fake ads, ransomware in fact so there’s a problem that’s not being solved rapidly enough by the big powers who are making tons of money. Cause they’re public companies they can’t really risk their profit, they can’t do something maturely adverse to their stock price and so they’re kind of stuck.

Brendan Eich: You see lots of talking to politicians, you’ll see some sort of cosmetic changes around the edges but real at form they have to come through innovations like Blockchain and other things. Like when I’m working on more than through software form, or self-regulation, or even the regulators who may not know what to do, they just know what they don’t want.

Peter McCormack: Okay. So I want to talk about Brave. I migrated to Brave around six weeks ago. I did an interview with Riccardo from Monero and I’d also read an article by Matthew Green about why he was quitting Chrome, so I ended up deleting all my data on Google, deleted Chrome, mobile and desktop and migrated to Brave and also to DuckDuckGo. So I’m just going to tell you a little bit about my experience and then obviously you can chip in and feedback. So the first download of Brave I struggled with the UX but then I migrated to the beta and I’ve used the beta since and I’ve found with the UX on that, because it’s a Chromium based build that I just love it. It crashes occasionally but it is beta so I accept that. So what was the idea on doing the chromium build?

Brendan Eich: We were always based on the chromium engine, except on iOS where Apple requires you to use their engine and I’ve built web engines for 20 years and I’m glad not to be in the business so we’re happy to use that back end engine of open source that renders the pixels from HTML and Java Script and CSS but what we did with the Beta that you started using is, the front end we were using a different approach than what the chromium open source uses and we switched from that different approach to essentially chromium’s front end but we had to remove a bunch of Google surveillance frankly. Things that tied in Google accounts, and sync but also things that signalled Google sites for various reasons. Not all was clear why. Some of it is an attempt that all big browsers make to do telemetry and report faults. Sometimes not with identifying the user but with reporting the fault so it can be fixed. Which is a good purpose but sometimes we couldn’t tell what these things were doing, and the code was fairly full with this stuff.

Brendan Eich: So part of this effort in switching over to the Chromium front end, besides adding the Brave Shields and the basic attention token and integrating them in a way that, that type of browser isn’t used, part of the effort was ripping out the Google stuff.

Peter McCormack: Right okay. And the background to Brave itself, like why did that project start? Where did that come from for you?

Brendan Eich: I really wanted to have a better experience myself and with my family too, the ads on, sort of the mobile web in particular, but all websites, all versions, all kinds of viewpoints, tablets and desktops as well, felt a drag on the network from all the tracking scripts and anybody who’s into this realizes that it’s not just invisible but it’s the invisible that costs you. Tracking scripts add up to such a hit in traffic, requesting response trying to measure you before your ad is placed and then afterwards, to confirm you saw the ad, that it can take up to 23 dollars a month, according to one study, half your data plan cost if you’re paying let’s say your data plan and you end up running the battery because the radio is number two battery consumer, or if it’s bursting on a lot, it could be number one. The screen is dimmer. All these direct costs on the user, and of course then there’s the risk of getting ransomware from an Ad on your PC, it’s not trivial this. So speed, battery life, the network bill, the data plan bill, and the clutter and annoyance and creepy feeling of being stalked by ads that would target you to remind you to buy something you’ve already bought, or that you decided not to buy, or that you’re still making your mind up about.

Brendan Eich: Only in the last case might targeting be considered productive. It might remind you to get something you were going to get, or drive you to get an Ad block and opt out of the entire system if it’s been ad blocked. So when I saw how the ecosystem had this problem and the users were feeling the pain, I thought, why not build something as a user that would be helpful and then having defended the user, what can we do for the creator? I mean the users are creators too and of course there are big publishers. Notionally ad blocking hurts the economics of the big publisher. Even if the user never clicks on an Ad a lot of ads are so called cost per impression or cost per million impressions, millions of impressions, thousands of impressions basis. So a fraction of a cent for you seeing the allegedly, so an ad blocker deprives a site of that revenue and it’s like white and you’d want to support it, how do you make up for that lost revenue? That’s not easy and that’s why we got into crypto-currency, we built a prototype like bitcoin and then built the Basic Attention Token.

Peter McCormack: Okay so we’ll come onto the Basic Attention Token, but I want to talk a little bit more about the ad blocking because obviously the internet has been a really tough ride for publishers. They’ve had to switch to entirely new business model. What do you think the publishers have got wrong?

Brendan Eich: I think what started out over 100 years ago with print media, was a direct sales model and you had like, the famous case of Jude Wanamaker, where he had a chain of department stores in Philadelphia, and he would get space in a local newspaper for his ads and the editor of the newspaper could make the content and the ads not clash, or be unsafe for Jude and they could do a deal over a handshake, or they could have a contract paid on amount of space that was given up in each edition of the paper. For a run of endless weeks, a catalogue or a store promotion, that model worked, and it still works on the web but it’s hard for most publishers to achieve on the web because the scales so much bigger. You’re not just getting local department stores to advertise in the local paper, you’re trying to get major brands, global brands to reach audiences who they want to market to across millions of sights.

Brendan Eich: Now you can’t do direct sales across millions of sites, it doesn’t work. So the big sites do direct sales and they make really good money. And they can often control the quality of the ads. They may use the Google Ad server and Ad exchange to put the ad in but they do the deal themselves. What everybody else, and even do this to fill the rest of their ad space uses is called programmatic advertising and this is the automated solution, it’s kind of a meaningless word, it means automated which means anything. Where ads are matched across many sites, maybe millions, based on tracking. So programmatic advertising depends on, in mass surveillance of your behaviours using third parties that set cookies on various sites and tracking.

Brendan Eich: Google had a search engine originally and they were calling ahead a search ad business, very clean, kind of a direct sold, first party ad experience because the little text ads in the search pages were separated from the organic content, they didn’t try to mix them in, or have people pay to put them in rank order where they didn’t belong and those ad results on the result page could be good. Google did well with that but if you didn’t convert them then they lost track of you and I think that’s why they started doing ad sense ads for bloggers, made an ad product that could be embedded, well if you paid 68% of the revenue, still boring, still kind of a big but not total part of their business but if they lost you from the search pages and the ad sense ads they had to get you another way and they bought double click. Double click was the dark ad server in 2000, I think it creates that, it was doing tracking, I know the double click in the 90’s.

Brendan Eich: I don’t know the whole history as well as I should but there was a history from the cookie on where very smart people, just working on their own figuring out how to do tracking across sites. So you have something called the cookie that is a little bit of storage in the browser for each domain. Let’s just simplify it to say domain name. Like part of the web address, but only a certain part of it, that storage can be updated from the server at that domain. The cookie is sent back after it’s been stored the first time and can be updated again and it’s sent for every request. Not just for the page you’re visiting but for embedded elements and that’s where the tracking comes in so you may have a cookie to remember that you’ve logged into your bank and you don’t have to re-log in every time you go back or start your browser. Or maybe it’s just a publisher site you pay for a paywall, access to get to their premium content.

Brendan Eich: Cookies are helpful. They cash for a short time, maybe for your bank it’s too short, some authentication token based on your log in credentials and that’s what they were designed for. But, let’s say in that publisher example with the premium content and the paywall. That can be a little one by one image, you don’t even see it. It could be a transparent … That imagine is requested, separately, over the hypertext transfer protocol and a cookie for the image servers domain is set. It’s sent back to the server and the server can update it and send it a response so every request gets a cookie and that includes these embedded elements and that’s how tracking works.

Brendan Eich: Figure that pixel and it’s little tiny image footprint, or something you don’t see but the server for it gets that pixel embedded in thousands of sites and you visit hundreds of those sites. Now that little image server can see you, if it doesn’t recognize you the first time, it makes up a number, user ID 1234, sets it in the foot key, sends it back, next time you go to a different site that embeds the same image, the cookie goes … there are various ways to do this but the simplest case is that it’s the same image embedded in a different way. The request can sort of tell the server how it was embedded so it can tell you’re on ESPN now and previously you were on New York Times, so user 1234 now owns a little [inaudible 00:16:13]. It started out at New York Times and now it’s at ESPN and you go to 99 other sites and it’s got you trapped and that is the basis for doing this automated ad system called programmatic advertising, which for most sites, is the only way to go they just don’t have the sales force or the scales to do direct ad deals with brands like, Procter & Gamble, or the car makers.

Peter McCormack: Right, but if they’re trying to monetize their platform why is this fundamentally bad?

Brendan Eich: So, first of all, that intermediation where through the tracker has cost you. It means you are paying more than just the sales force or the IT staff that you would pay for your direct sales. You’re paying the tracking partners, and there could be several of them, you’re paying for what’s called an ad exchange, Google runs the biggest, this is where after the tracker assembles some kind of profile, there’s a real time bid process that tries to match a collection of users based on their profile as an ensemble audience that the advertiser wants to reach against a set of ad spaces on sites that’s called the inventory and so this real time exchange has to take its fee.

Brendan Eich: Oh and don’t forget there’s also something called a demand side platform that the advertisers use, or their agencies use to run their ad campaigns, measure their … Try to make sure they’re working well, often publishers use what’s called a supply side platform. You’ve got an alphabet soup intermediaries and they take, according to one study, 45% of the gross spend but in reality it’s more like 68% and that’s just too much. It leaves the publishers on a defining revenue footing, especially with fraud. I haven’t talked about fraud yet but fraud is out of control in this system, there is a …The searcher has studied this and he estimates that at least 20% of all the 100 billion being sent this year in the United States, on digital ads is taking by fraud.

Brendan Eich: And what is fraud? Fraud is fake sites and fake humans, bots or fraud actors, if they’re cheap enough in some part of the world that seem to view ads or click on ads but do it to get the ad revenue that would be due to the publisher, the real publisher, they do it by faking the publisher. And they get paid, that’s the amazing part. The advertiser puts money into this system, weakly tries to make sure their money goes to the right places, goes to real people on real sites but at the end of the day the system is so lacking integrity that at least 20% of those 100 billion, 20 billion this year, could be more, could be twice that, is going to fraud and that means the real publishers did not get paid for that fraud. The real publisher is out that money.

Brendan Eich: So publishers are participating in a system that naively sort of like the barter theory of money, they think, oh this is just like the old days where Jude Wanamaker was selling his department stores wears through direct ads, ad spaces that he bought from local newspapers. It’s not like that, the direct model is great for the big publishers who can do it but almost everyone, even the bigs to some extent, everybody else, totally is depending on this automated system and there’s more to say about these ad exchanges. Not only do they get a fee, even if the fraudster is faking the site and faking the user, they get a fee if the malware vendor is faking the ad and putting it on real sites to get ransomware, that happens too.

Brendan Eich: So you have this traverse incentive system where the intermediaries get their fees whether or not the traffic is legitimate. Could be fake human traffic, could be fake ads going to real humans to take over their machines and ransomware and forbid for it so that’s why publishers are suffering. It’s kind of a long story, one of those systems where if you approach it like from the sort of kindergarten book, Jude Wanamaker department store, local Philadelphia newspaper model, you think, “Oh, it’s all good, I know how this works” but you have to look at the actual system and how it evolved and how it depends on massive tracking, massive ad exchanges, too many intermediaries how have conflicts of interests with their principles and who take fees no matter what and that’s a recipe for fraud and abuse and low pay to the person at the end of the payment chain which is the publisher.

Peter McCormack: Right so what you’re really saying is this is a highly inefficient system and there are different parts that effect different people so it’s poor for the users because they are ending up paying, you know their batteries are being drained and they are paying for data for all this tracking and all these adverts. It’s poor for the publishers because they’re not getting the right percentage of the revenue and it’s poor for advertisers because they are advertising to fake people, or not advertising to the right people. So it’s just a very inefficient marketplace.

Brendan Eich: I’d go further, it sounds … Inefficient is too kind, it is abusive, it is dangerous and it is misleading. The degree to which fraud is tolerated is shocking and some of these intermediaries go to trade shows and they talk among themselves on stage in panels and they joke about how many fake ads they profited from, how many viewable videos they took a key to liberate. I kid you not, it’s really corrupt. It reminds me of the big short. You’ve probably seen the movie version, Michael Lewis’ book. But it reminds me of the financial fraud that brought us into financial crisis 10 years ago so it’s no laughing matter, it’s not just a little inefficiency, it’s a big inefficiency, it’s moral hazard and it’s malware and fraud, criminal fraud.

Peter McCormack: Yeah, and I know about this industry, so I worked in it for 20 years. I had a digital advertising agency, and also I think you’ve read one of my articles I sent you a while ago, I’m not sure if you remember, I wrote something called Online Advertising Does Not Work, and we spoke about it a while back and there are-

Brendan Eich: Yes I do remember that yes.

Peter McCormack: Yeah, so I know about this industry and I also think there are plenty of issues relating to the problem that the advertising industry has had adapting to the internet and one of the things I notice from working in the industry is the volume of lies having to be told to justify the advertising. So the couple of things I had major issues with was firstly post impression tracking, which I felt was creating a network of coincidences because people are seeing millions and millions of ads are being distributed every month, and there may be an absolute coincidence that you’ve bought something.

Peter McCormack: The second is a lot of the ads were retargeting to people who had already purchased the items and then the last one that I also took massive issue with, especially with Facebook is with auto playing videos and auto playing adverts because there’s no choice and the engagement is faked. So I know the model is broken right but at the same time when you bring in an ad blocker I guess … So you’re completely cutting off the publisher so do you feel like you’re trying to force a new model onto the industry?

Brendan Eich: No, we’re too small to force it in the way that Google originally forces new API’s and Facebook forces new viewability, worse viewability standards on their audience. We’re small but what we’re trying to do, along with the rising tide of ad blocking usage, which is not going away, on mobile too, through new browsers because you cannot get an ad blocking extension for Chrome on Android so you have to get a new browser anyway so that’s we’re Brave actually has the most shares, on Android and second place OS for us is Windows, third place is iOS, which is surprising because Apple ties Safari in so well. Apple also makes an ad blocker app install model, content blockers available but we’re not here just to block ads.

Brendan Eich: We’re not trying to burn the system down, some of the purer site interact with occasionally think and in their best days make the case that the system is so rotten you cannot reform it by building a cleaner version that doesn’t have tracking, you have to burn it down so you have to block everything and then see who survives and then start over. I don’t agree but I understand when they make that case that we’re speculating about the future and we certainly see part errors where civilizations and technological systems like Ad Tech fail so to reform the system I think in the necessary parts that fund it, the user, the advertiser and the publisher, and transposing them into a cleaner model that doesn’t have tracking, doesn’t have these layers of agents that have conflicts against their principles, and that take fees no matter what’s going on, fraud or malware. That’s our approach.

Peter McCormack: Right okay. And is ad blocking like anti-virus, it’s a constantly moving goalpost and are publishers and advertisers trying to find new ways to trick essentially your shield so they can still have ads displayed?

Brendan Eich: Well it certainly is but all the big ad blockers and Brave collaborate on techniques and so there’s some cat and mouse there. But I would say it’s better than anti-virus in the sense that at the end, the browser determines the ground truth for what JavaScript means, what its semantics are, what it runs and that means that though the server might send you JavaScript that’s confiscated, they might use malware to try and hide what they’re doing from you so you can’t block ads. They can’t really do any better than that and that’s pretty weak. That means the browser wins in the end.

Brendan Eich: There’s a paper from Princeton from on this. It makes two cases with concrete work. One of them uses like in Chrome it’s called isolating worlds, so it deposits the publisher has put up an anti-ad blocker, we’ve seen these, they’ve started saying, “You’re using an ad blocker, you must either lower your ad blocker shields or subscribe, or go away” by the way, this is not productive for the publisher. Putting up these dialogues just reduces their Alexa score in my experience but they do try and in the paper from Arvin, the first technically uses, they put that entire page in an isolated world from the one that you see. In the one that you see they block everything and they feed the ad blocker. The anti-ad blocker.

Brendan Eich: We defeat the anti-ad blockers too so there’s an integration of cat and mouse cause it stops there with the browser winning but the other world contains the entire anti ad blocker thinking it won and it can’t tell any differently so the browser can not only control the horizontal and vertical it can put the remote content in the matrix and as you recall, unless the agents change it and Leo sees the black cat twice you don’t really know you’re in the Matrix. Or remember in Ready Player One, we’ll take a more recent movie, Sorento is in his super chair and he thinks he’s out of the oasis but he’s still in it. So you can play tricks on their real content, it cannot really play tricks on you. If you just try to hide, it can try to officiate.

Brendan Eich: There is a darker alternative called digital rights management, DRM where video, in particular, tries to be not only officiated or encrypted, but you’re not allowed to get the key unless you pay, and you’re not allowed to see the crypt assistance, hidden in hardware and software that if you even experiment with anyone they get a part to look at it, or try to fix the colour balance because you have colour blindness, you can go to jail. Now that shows that there are no technological solution here for ad blocking. Ad blocking can win, DRM can be defeated and it is routinely right. A movies straight up on the dark web before they’re even released to theatres, but because the Hollywood studios now have such cloud they try to bend past laws with fairly prognose side effects to criminalize hacking.

Brendan Eich: So we’re opposed to that but we do not see that working for ads because on the web, apart from HTML 5 video, which sadly DRM was imposed, by Netflix, Microsoft, and Google, and Apple going along with it, ultimately Mozilla had to go along and [inaudible 00:28:38] we’re all kind of complicit but there’s no alternative if people want their Netflix and their new movies, part from DRM HTML 5 video, the web is an open standards system and you cannot use these toxic laws to prevent ad blocking but ad blocking has good case law, it’s important. It’s part of the users right to use a reader mode, like we have in Safari and Brave has a super reader mode coming, or to use a screen reader if you’re visually impaired, to turn text into speech. These are all aspects of the web where you can take apart the content, you can master it up locally without filing and copyright, strip out the ads. You can strip out parts of the content you don’t want. That’s all within the users right and it’s part of the web standards by design and it’s most supported by case law.

Brendan Eich: So Ad blocking and anti-ad blocking are arms racing as you say but I see as Arvin does, I didn’t tell you the second thing within that paper. The second thing, you know how ads service is self-regulatory step have usually an opt out button in the morning in the corner, sometimes it’s ad choices or something like that. Second technique in their paper was to recognize using very trivial, sort of character recognition, or computer vision if you need it, is buttons and so therefore they can recognize ads visually. Now you have to load the ad and stuff for the tracking but you can still take the ad out so in the end the browser here again controls what the ground truth is in the sandbox where the JavaScript where the means the browser can win in the end and I think this leads to an armistices, it should lead to a better system that doesn’t require the toxic tracking that leads to all the problems. Venture mediation I mentioned, that’s better for users, that’s better for publishers, and better for advertisers and that’s what we’re trying to build and there may be other influencers just like ours, the pieces of what we do may go into the web standards, we’re collaborating already, there’ll be some seed, so we’re not out to corner the browser market so therefor I wouldn’t say we’re trying to force anybody.

Brendan Eich: We’re trying to lead by example and be part of a stubborn minority, you know transatlantic minority that can tilt whole markets, and does. For very practical reasons. The common example is salt. Which is halal and kosher and it’s only one SKU in every store. You don’t have three brands of salt for sale. That’s the minority. Certainly in the US, people who want Kosher salt or Halal salt are minorities but they insist on it and the salt makers don’t want to make three different kinds of salt and put them in three different places on the shelf so that shows minority effect and I would say Firefox is an example. It restarted the browser market, Firefox. We were growing it Internet Explorer share, Chrome didn’t exist, we had to show Google how to do it and even at small share, Firefox was punching above its weight with web developers, something Chrome did well with its damn tools. We were punching above our weight in the standards part.

Brendan Eich: So minority effect matter and it can tilt whole markets. Combined with regulatory action which is already upon us, Europe and elsewhere like I said, I think there is a better world we can get to.

Peter McCormack: And do you see a kind of irony that you’re … Part of what you’re doing is fighting against something you originally created?

Brendan Eich: Oh you know, we didn’t know what we were doing. I’ve said this before, we did not know what we were doing. In the 90s it was very fast, Marc Andreessen in 1993 with Eric Bina, he did all the heavy coding, in the mosaic engine, before they founded Netscape, Marc says, “Hey, we want to do images” and you know at this point when he’s announcing it to the WWW talk west, Eric Bina’s already coded it. It’s not like, oh this is a proposal. It’s like, oh this is shipping tomorrow on Mosaic and so they made the image element which can be embedded so that the bits for the images that are being coded and compressed pixel information, transformed into some other domain, come from a different server. That’s one of the legs of tracking, that’s why that one by one pixel I mentioned was the earliest kind of tracking element because the next year, 94, Netscape won, and that’s when they did the cookie.

Brendan Eich: We didn’t think through the consequences of these innovations. I did JavaScript in 95 and I made remote script. Sort of like the image that can load a script from a different server and fold that code into your own, first party, the including sites, sort of trust domain. I made that work just like you mentioned, I did it for the obvious reason, we need to include why we’re in code from a static server, an edge catch or something, after another party. Maybe even a safe stateless server for shared JavaScript providers.

Brendan Eich: So I did the obvious thing, we were all blindly going fast and we didn’t see until, maybe 96, I remember a conversation that was about tracking cookies, that the combination of the image with its remote loading, which carries cookies, and the cookie which was originally for the top page, recording your log in, could make tracking and then it was too late to change at that point. So I paid my dues for a lot of this in technical committees but I feel like this problem is bigger than any of us, we didn’t see it coming, typical unintended consequences and it’s a system problem so I think there’s an economic aspect to it that should motivate many parties to improve it.

Brendan Eich: But like I said the big superpowers who won, wasn’t Microsoft, it was Google, which didn’t exist in 95, that became sort of stuck in their high castles.

Peter McCormack: Right okay. And what is your relationship like with the Silicon Valley bohemians. The Facebook and the Google cause I notice the shield didn’t block the adverts on Facebook or the advert search results in Google, why is that?

Brendan Eich: Well, yeah we use … want that and they may do it for various reasons, it’s an option for those users, those are first party ads. I mentioned the Google search ads are in some ways like the Jude Wanamaker, direct sold department store goods ad in the newspaper. They’re placed on the content that Google owns, some Google servers. They don’t actually track given the other shields we have. We protect you from third party cookies and other tracking, we protect you from fingerprinting so those search ads are actually fairly benign and they’re first party ads so we try to leave first party ads alone anyway as a policy but it’s not a super laser bright line. The best example might be. Graphic has a sponsorship ad that’s just a big image you can click on to donate to the world wide founders, something like that. First party ad, hard to object to if it doesn’t have tracking and if it has tracking and you use a browser like Brave you’re protected from that tracking as you are everywhere. So those kinds of first party ads we try not to block. Those are the cleanest kinds of ads. Often they’re direct sold.

Brendan Eich: Unfortunately, some of the direct sold that do make money for instance or, the fashion site Elle,, while they’re direct sold they place using Google Tags, Google server and they have a click tracking partner to confirm that you watched the video of the carousel and that partner is tracking users across the web facility, creating a profiling premises so we end up blocking that anyway but the search ads are pretty clean. The Google search ads, Bing ads like guise, DuckDuckGo, or Quant, our partner in France and Germany, same deal.

Brendan Eich: Facebook ads in your feed are also first party. Now they have had malware occasionally in the past, it’s not been a huge problem. That’s one reason why we might block them. Twitter promoted tweets, everybody hates them, just on opinionated browser aesthetic grounds and based on a majority of our users we might block them but we don’t right now, because they’re first party. So I hope this first party, third party distinction helps, it doesn’t, like I said, it doesn’t have a super bright line around it because sometimes these direct sold first party ads are placed in some third party techniques and they get bought, they definitely get bought.

Peter McCormack: Okay. Well let’s talk about BAT. This is probably where you face most of your challenges, I remember a long debate with Francis Pouliot on Twitter that sparked me getting in touch to want to do this interview.

Brendan Eich: Oh it wasn’t that long-

Peter McCormack: It was a good one though. It was a nice fiery debate. So one of the common threads that comes across with people I talk to in crypto, I’m out at crypto con at the moment, I spoke to a few people here and there’s a common thread here is, I love Brave but I don’t like BAT. Why do you think this has become such a problem for so many people?

Brendan Eich: Well it’s not a problem for our users, we’re growing nicely because the crypto, a community that you’re dealing with there is tiny and not growing fast. That’s the gods truth, I have nothing against them, I run in the maximus not just … and Francois, who got a little nasty by trying to summon regulators against us for some such, calling the cops, from twitter which was pretty uncool and silly but also I know some bitcoin maximus that I respect and like but they tend to say, “You should be using bitcoin and you should make everyone take bitcoin and then you should not grow until that happens” and my answer is, “No, it’s your job as maximus to use bitcoin”, I’m making a better browser and I need to get some kind of tokens to flow, so don’t shoot the messenger, the world is not switching to bitcoin. We’re paying over 23 thousand publishers now every month, the biggest ones, the established ones. Nobody in the legal IT or Finance departments wants cryptic, doesn’t matter if its bitcoin or BAT, they don’t have anything against BAT, they don’t want any cryptic, they want fiat and so through our partner uphold who manages settlement and KYC and all the rest that Maximus also hate, we’re paying real businesses right now. We’re [inaudible 00:38:08] that means we have to deal with fiat because a lot of them want it, they do not want cryptic.

Brendan Eich: Now there are creators coming on to our platform from YouTube especially who like that and Uphold nicely lets them get that directly with no fee if they make [inaudible 00:38:25] it’s verified with us. So we can do BAT and we have people who like that. But if you’re at anything like a Bitcoin conference, you’re going to run into, what I would say is another intransigent minority and not a fast growing market, nothing generative for us that is against that because they want bitcoin and only bitcoin and that’s not in our users interest or in most of our publishers interests. We started with it right, we had bitcoin as prototype and it got very expensive, we had a base buy widget, people were buying small lots to give away every month out of their own good will, with the BAT we can actually give them grants from our growth, so there’s an advantage that with have with BAT it just doesn’t exist with bitcoin. But be that as it may if you believe that people should only pay themselves for the crypto, it was still hard to buy.

Brendan Eich: Now I’ve had Maximus say, “Oh you still use coin base” or “Oh, those users are idiots, they should have bought larger, lots of Bitcoin and then they would have advertised the bee better”, or “oh they should have used a smarter wallet to manage the pees”. That’s like saying average users, your grandparents, should become experts, it’s not realistic. Again it’s kind of cultic, Maximus tend, in the worst case, to be like Cargo Cult’s, none of them among the friends here, I’m not speaking of the ones I respect but the ones I don’t respect, could have invented bitcoin and now that it exists they make it into this wicker statue that they worship and they rally everybody else for not worshiping it.

Peter McCormack: Yeah, I mean I’ve been doing this podcast now for coming up to a year now and I’ve had a lot of interviews from Bitcoin Maximalist to people who run projects like yourself, and I sympathize with a lot of the Bitcoin Maximalism arguments but don’t agree entirely with it. But in reference to BAT for example, you’ve said publishers don’t want cryptocurrencies so why use the BATS token at all, why not take say the Patrion route and make everything payment in fiat?

Brendan Eich: First of all, we’re not requiring publishers to verify for users to start sending tokens toward their favourite creators. That’s important because when you try to climb both the scaling wall of acquiring publishers and of acquiring users in tandem, so that every step of the way you need to bring users with their publishers, and publishers with enough users to motivate the publisher, you will be in catch 22 the whole time. This is killed whole apps that tried to be like a better browser, like readable on iOS. Tried to get curated or custom premium content from publishers of note and they want to see big user numbers. Users want to see the custom content first. You’re stuck.

Brendan Eich: So we let users put tokens into and if we did this with fiat, I don’t think it would work first of all, I don’t think uphold would do it but I think we would end up having to be some kind of a money transmitter ourselves and we don’t want to do that. We want to use crypto for what it’s best at which is pier to pier where possible, escrow until then and decuple the problem of scaling creators and scaling users from one another. We want to get users first and we have enough users which we do now, we already have publishers coming on board, I mentioned Dow Jones media group in April, and many more since then, among the publishers who verified their main ownership with us in order to collect their tokens but did not otherwise do any kind of business development, didn’t require any pre sales or sales effort from us, like the Guardian, the Washington Post, vices verified quarts are coming in, there are a bunch of these publishers who have now seen our users scale and they’re coming on. So our model really is different from pier to pier payment model but again the maximus assume that that’s the future and then get angry at us for not building only that and either cajoling our users to adopt it, or failing getting enough users to back our business.

Brendan Eich: I’m not prepared to do that, I think that’s not productive for the world. I think we need different approaches that rescale and this pragmatism about getting to scale I think it actually offends some maximus, I think they would rather bitcoin stay small and kind of clubby and if the normies invaded it, it would be like they are welcome on the internet in the 90s, right it would be bad. And it could be bad, certainly for the technical culture but if you want crypto to be widely adopted it has to be usable by normal people, and normal people expect not only not to have to manage their private key to their wallet, which if they lose it they out their funds.

Brendan Eich: They have to be able to do a chargeback, oh you sent it to the wrong address, too bad, you’ve lost a crypto. With credit cards it’s trivial to charge back. It’s so easy for our clouds to get eaten by the merchants so … And on top of the interchange charts. So, there’s no free lunch here but the idea that everyone’s going to adopt pier to pier, then install their own wallets, especially with current tech, it’s ludicrous. We’re growing outside of the crypto bubble I would say and that’s why, I mentioned earlier, decentralization is hard, crypto’s exciting, attracted a lot of investment, that’s one of the other great functions besides censorship resistance, it’s liberated, capital toward projects that couldn’t get into the old boys clubs and venture in private equity. I’m in favour of that too, even though there’s a lot of nonsense projects and scam projects and the regulators are rightly clamping down on those but it’s still a small world. Crypto is not going to go mainstream until it has the importance as people use certain credit cards in my opinion.

Peter McCormack: Yeah, I wouldn’t agree with you when you say maximus want bitcoin to remain small, in a small club. I’m not sure that’s true but at the same time there seems to be some kind of clash here between … You have issues with crypto currencies related to bitcoin yet you are still using a cryptocurrency with BAT and I feel like there’s a conflict there. I would have thought a company like yourself if you felt like publishers just wanted fiat that you wouldn’t struggle to build some kind of fiat based system.

Brendan Eich: I noted earlier that some of our creators want that so there is an interest in that per say. Some of them may want bitcoin and they can get that too through uphold, which is just one exchange hop away and there’s a fee there of course. But it’s not that nobody wants bitcoin, obviously some people do, it’s just a small cohort, it’s not growing fast. Some people want BAT, I would say that’s not super growing but it’s a part of our publisher base that want that. But the big publishers, the extent ones that we’re trying to save from the deprivations of programmatic advertising, they want fiat and so we have a mix but if we only use fiat, we couldn’t satisfy the other people, and if we only use fiat we would have to be a money services business. We would have to be a bank or partner with one in a way that I don’t think banks would partner.

Brendan Eich: They won’t easily, or at all do this kind of lightweight, anonymous donation escrow service for instance that we rely on due to couple of publisher growth, from user growth. So I’m telling you a more complex story than the pier to pier story because the pier to pier story does not scale well. It has the usual that you have to get both ends of the connection, both piers have to be using bitcoin and a lot of these business that we’re already paying, I mentioned the Washington Post, the Guardian, they’re not going to do bitcoin and I need foreseeable structure. Certainly not because of our five million growing users, if we had 100 million users and we said you must use bitcoin, well we might get somewhere but you can’t get there from here except by doing what we’re doing so the coping is well rationalized, it allows us to use block chain for anonymity and pier to pier when we need it which we do need sometimes, when we can’t use the escrow ability and it also allows us to have microscopic audibility when ads start flowing the advertisers will pay retrospectively essentially so we’ll use out growth pool to prime the pump.

Brendan Eich: They’ll pay into a funding wallet and there’ll be a settlement wallet, there already is, which uphold manages, and you can see these, if you’re in block chain, you can see that we didn’t take more of a fee than we said. We’re not embezzling, which has been a problem in ad tech as you know, there’s been some scandals around the platform, so called, a bunch of them were lying about their gross versus their net and their publisher clients smoked them out and fired them because they were taking too much of a fee.

Brendan Eich: We have to be very out of the bullet, that’s another advantage of block chain, we’d like to make the usable custody for 30 days which allows them to change their mind about the anonymous donation mix and potentially repudiate ahead of time and then once that commits it’s gone, it’s not in their custody anymore and it’s a lightweight model because it’s only five dollars a month or 10 dollars, whatever the bad equivalent USC the user might prefer, and that can be done with crypto, it really can’t be done with fiat either.

Peter McCormack: And have you released any kind of figures on what kind of revenue numbers that publishers are starting to make by coming onto your platform?

Brendan Eich: Not yeah. I think we’re paying 23 thousand, we’re paying four figures is what I say, that’s what I’m going to go for, one of the biggest monthly pay outs to a publisher so it’s getting substantial and as we grow it’ll only get more so and this is a rolling stone that we hope to keep rolling and gathering.

Peter McCormack: I think four figures for a major publisher is probably way off what they would need to build a sustainable business.

Brendan Eich: Well they’re not. You can’t assume that we’re taking over their browser market, we’re not. What we’re doing is that we’re showing found money effect. With just ad blockers they get nothing. With us, and a small number of users, there are five million users but there are a property that a lot of our users go to so maybe it’s a few thousand of our users go there, maybe it’s only a few hundred.

Brendan Eich: If it’s a few thousand dollars from a few hundred users that’s a very high monthly pay out per user and that’s where some of our partners are coming on board and saying let’s do more because not just Dow Jones, we need … Many publishers we’re talking too and that we’ve been talking to for years have moved from, “I hate ad blocking” to “You know what I was thinking, we’re not in that old world category” by the way the biggest ad blockers take fees to let ads through. Which means they let tracking through. It’s totally corrupt in my opinion but we don’t do that.

Brendan Eich: We’re not an ad blocker we’re not a browser we’re a new platform and there for some of these publishers are saying, “I need a new platform. I’m dying on Google’s and Facebook’s and many lesser ad exchange, I’m just running out of business slowly, I need to get to a model of customers to pay me for premium products, or pay by the yard instead of me trying to pay well them for $260 a year, or $37 a month” those paywalls do not convert well. First of all the user has to be willing to risk their credit card to fraud and just do the credit card sign up which is a huge restriction, and rightly so. A huge point restriction and then they have to overpay, they have to cross subsidize content they will never read and don’t care to read. They can’t buy articles with tokens per article. Whereas with that, they can do that.

Brendan Eich: So we’re talking to publishers who are now saying we need to get two streams of revenue, one from ads which we’re dying on and one from a lightweight, low friction user pays model and that’s where they’re very interested in braving the bath because our lead users can, again the sovereign minority of that, certain publishers that can help the publisher prove the model and recover revenue that they would have pure ad blocker audience would have never had. They go up from zero to positive, they go up to multiple dollars per month per user and then they want to scale that to all their users so I’m going to tell a story about a site and you can guess which it is.

Brendan Eich: It has 80 million ad blocking unique visitors per month. It has many more non ad blocking but it has a high ad blocking instance, and they are ready to try something with us that would convert those ad blocker users, some of those whom are using the corrupt ad blockers I just mentioned to you, take [inaudible 00:50:39] to let ads through, they’re ready to convert those users to Brave users. That’s a big growth opportunity for us, but it’s no joke for them because if they do convert we can use the batch from user growth pool to pay them a referral award for every new user that comes and sticks around with us. We’re doing that already, you can get a referral code, We will also give the users ongoing grants so they can give back to this site and when the ads system which we’re in user trials right now on, starts performing we can replace the grants with revenue shares from the ads as we’ve been talking about from the beginning.

Peter McCormack: Yeah, so with those users making donations. It’s almost a socialist kind of system of opt in to choose to give money to content providers you like but realistically a lot of that has come from grants. Have you released any numbers on the amount of people who’ve actually optionally choosing to buy BAT tokens every month from their own card, through uphold and distribute them, versus those who’ve come from grant because it feels like-

Brendan Eich: No we haven’t but I can say that we had like 90 thousand users funding their wallets, we have over 413 thousand wallets funded, grants have made up the bulk of them and that’s because we do not expect the users, out of their own good will to open their wallet any more than these sites expect to have their macro paywalls convert you $37 a month. Why should you have to put in … like Zucco of ZCash phase was putting $20 a month into his Brave wallet, very good of him, thank you Zucco, no offense but I do not expect, and my business plans do not count on all users doing that, that would be another minority effect. Valuable effect.

Brendan Eich: League users who have discretionary income and send it toward their sites on purpose, that’s hugely valuable, I’m not knocking it, I’m just saying it was never our plan to require everyone to donate and it’s not socialist in any sense, it’s entirely voluntary but our model is more realistic in that currently everyone expects things to be free. Mark Zuckerberg even says so, this is why he rationalizes surveillance and ads because otherwise if he didn’t have those systems he’d have to charge for Facebook and only the 10% or whatever can afford it and isn’t he nice? So in some sense ads are more socialistic in that they certainly create a common that is despoiled because it’s alluded from the users without taking cents and that’s their dossiers, their attention, their ancient histories and then it can be used for all sorts of bad purposes so I would say what we’re building is more of a direct, user, reader, to publisher system, a direct advertiser to user system and the directness is what matters. The indirection through the ad exchanges and demand side platforms and so on and so forth, that’s been the source of much more.

Brendan Eich: So getting back the direct relations … By the way the brands that we’re talking to for ads that will pay the revenue that we share with the user, we give the user 70%, those advertisers want the user direct ad model because they are being commodified by Amazon mostly and similar marketplaces. So we’re trying to get back to more of a direct model without the overhead of the direct sales through the human handshake that I mentioned that happened 100 years ago and still happens, but only for the biggest publishers and brands.

Peter McCormack: And why did you implement a six month expiration on BAT tokens?

Brendan Eich: For the grants, we’re giving them out, but we know that’s a fraud magnet so we know that some people will claim them, or have bots claim them and then send them to a verified YouTube account that we only want Uphold to prove it’s a real, payable … Uphold does that for us, they do the bank like AML, KYC, anti-sanction list, stuff that you have to do or you go to federal jail. Uphold does that for us for the YouTube site but it still might be kind of a fake site with two videos that hardly anyone’s watched and yet we have found, as you might expect, because we’re giving grants to browser users without any KYC on the browser user, that there’s some fake browser users. Some bot browser instances. Of course some mechanical terms like browser [inaudible 00:54:40]. Instead of steered grants they’ve taken grants from us and they’ve steered them toward such a musty and under viewed YouTube account and we recognize that’s fraud and we clawed back.

Brendan Eich: I know there are people on Reddit saying, “Ha ha I got some.” We’re happy to grow by giving grants to legitimate users who want to give them back, so there’s always a risk somebody will get past our anti-fraud defences but we have to play the anti-fraud game, there’s no way round it when you’re giving out tokens, or giving out revenue shares of the ads, that’s also a fraud magnet.

Brendan Eich: Just like with ad tech today. The difference is, ad tech today, there’s no such thing as effective anti-fraud because it’s all a mix of JavaScript on the publisher’s page and the anti-fraud scripts cannot tell which way is up. They are in the matrix. Whereas we can use native code in the browser, we can even put it in a SDK in a secure enclave on intel, we can walk it down so we know it’s not been tampered with and because it’s native code it can read all the OS API’s to get all of the sensitive data on as real human using the browser without disclosing that to us.

Brendan Eich: We don’t track you but locally it can make judgements about this is a real human not a fraud bot. It’s not an android phone held in a vibrating rack with a robot finger tapping on it. Frauds just a systems but there’s a cost to fraud in doing that and if we raise the cost high enough we get a very low fraud rating and so does Google and Facebook. So do Google and Facebook on their native platform, it’s on the wild west of publisher pages, where Google is serving those double click ads, and direct sold ads and their thousand scripts competing for overriding each other’s idea of standard APIs.

Brendan Eich: That’s where you can’t tell which way is up. That’s where the anti-fraud scripts just fail. So we’re playing the FFR game and that’s … Long story short, that’s why we’re handing out these grants. We do not want to have permanent fraud magnets hanging up and we want our own anti-fraud system, so we can forget about some of those six month old grants and move on to newer grants, most of which, there the one users.

Peter McCormack: Okay. Well listen I’m conscious of time and I know you’ve got to go onto another call shortly, I just wanted to make a final point which you can comment on, is that, so I’m a content creator, I have a YouTube channel and a website and I’ve been through the Brave experience and the BAT token experience, my biggest issue was the huge complexity of the UX and the amount of friction there was from me trying to create any, or generate any revenue from my channels. Are you aware of that and is that something you guys are working on?

Brendan Eich: Yes, so I’d love to hear your experience in detail, now or later, I think we’re in this high customer care mode unlike the big companies, because we have to be and I think it’s the right way to evolve in general so I’d love to try and make it better. We’re working on here for everybody. So, when YouTubers come on and onboard with Uphold quickly but two factor off, I’ve heard of complaints. I don’t know what burned you in particular but I’d love to hear more details and fix it because we’re young. We’re just starting right, people say, “Oh, you’ve only done a few thousand a month to one of your top publishers, well then forget you” Rome wasn’t built in a day. We’re here for the long haul and we will go to scale we will make the friction as low as possible so I’d love to, perhaps after the show, hear more details and make sure you get satisfaction.

Peter McCormack: Yeah, I mean I can drop you a detailed email on my thoughts because look I like what you’re doing, I recognize a problem with the industry, I’ve written about it, I know there’s a lot that should and could be done. There are things I would do differently but I don’t have your full experience. So let me drop you an email on that. We probably should do a follow up at some point and also just to finish off, could you just tell people how to keep track of Brave, how to find out more about the BAT and who you want to hear from?

Brendan Eich: Sure. And by the way thanks for bearing with me here. A lot of this that I’ve talked about, as anybody who listens closely can tell, it’s really not about crypto block chain per say, it’s about a practical application in a certain domain and that’s where rubber meets the road. Doing what we’re doing has required thinking about the whole system so I appreciate the time to expand on that.

Brendan Eich: you can get it off the beta channel if it’s crashing, you should get on the stable channel now. Re-download from our main download page which will take you to the chromium front end for desktop, we’re bringing the batch to Android so that will take you to the app store, google play store where you can get that. We even hope to get the BAT on iOS because it’s an improved virtual currency in the app store believe it or not. Uphold app already uses that, and that got approved by Apple so we’re making headway where people thought it wasn’t even possible and if you want to read about Basic Attention Token, the Reddit is good but the site I’ve already told you about needs some new material but we do blog and we co-blog on both and but we blog every week. We have a bunch of interesting research projects coming and we do have a much delayed road map coming where we’ll project some ideas. Not try to pick the one direction to go over the horizon when there be dragons but rather we’ll try to lay out some ideas we think are worth investigating for the Apollo phase in our road map which are more than centralized and may use more, more advanced techniques and I’ll have more to say about that.

Brendan Eich: So stay tuned to our blog. Thanks.

Peter McCormack: Fantastic. Well look thanks for coming on Brendan and I’ll speak to you soon.

Brendan Eich: Thank you.