Swarm Summer School ‘21, Day 1: Michiel de Jong - decentralised solutions

um then i'd like to introduce you to our next speaker which is michiel lyon hi with you um the heel is where is leon is founder of pondersource um co-founder of the project federated bookkeeping um co-developed the privacy awareness plugin terms of service didn't read for firefox and has worked on data handling at the solid project where you build data pods and ripple so he's now a new guy in the decentralized space as i would say and i asked michael and we asked him to present some of his experiences with these projects and how and why especially the why they were built in such a manner with data awareness in mind again feel free to ask questions during this workshop it's going to be interactive please use the you can use the q a option in below in zoom and otherwise the discord is also open if you're active there so take it away thank you yeah i um what i'm working at now is uh pondersource which is a small company that is a post-growth nonprofit uh because you know this year we went to the um uh blockchain incubator at this center utrecht and uh contrary to what melanie ryback said we didn't get the slide about the hockey stick curve going up to the right we got a talk from bellany rebecca telling us like uh you shouldn't think about it that way you know you this venture capital that's not the only way you know and then no obviously i got all of us thinking and i um kept that was i kept thinking about her presentation uh the rest of the weeks and i had another call with her where she explained some of the concepts a bit more and answered like a list of my questions and then so then i decided to make a ponder source a non-profit company and that's where we are now i'll um in this presentation i'll share i'll try to tell a bit a bit about what i've worked on in the past and and where i'm now if i can uh click play here so yeah what i wanted to start with is why do anything why do i why do anything in general so why would you um choose how do you choose what you will do in your in your work so uh one uh school of thought that i came across recently is called effective altruism which starts from the assumption that you want to improve the world you want to do something for other people you want to output something positive and a lot of people get very enthusiastic when they have an idea about okay this is what i want to build and i invented it and it's going to be great it's going to be so many people are going to be helped with it and that you need that enthusiasm of course but you also need some some science to it so effective altruism says like well when we do business or we do serious stuff that we're always measuring and we're always keeping statistics but when we do charity or we want to uh do something amazing to change the world then suddenly it's all enthusiasm and volunteers um part of it is when you volunteer uh then it's sort of an extra gift so nobody's gonna tell you that you shouldn't do it or you should do something else or that you should do it differently uh but this also hurts how altruism um can be effective so um one of the stories that this book about the factory authorism uh quotes is about a a project that was called playpump and the idea of play pub is great you um in in some african villages there's a lack of drinking water and uh usually the women would have to go get the drinking water at the well and they have to carry it on their heads and um it's a lot of work and um also there's not so much toys to play with for kids so this um uh this charity said well let's solve two problems within one go we pay we build a play pump where kids can just run around and and uh use it as a carousel but at the same time it pumps up the water and um they built this thing and then they got crowdfunding and they got donations and they got good reviews and they were on the newspapers and some politicians back them and they built all those play pumps and in lots of footages and then one or two years later they came back and they say hey why aren't you using the play pump to you know to play and to pump water and then in some villages they saw that um the mothers were actually walking around the bay pub and they would uh get dizzy and end up throwing up because of so many um rounds they had to walk and um in other versions it was broken and there would not be because it's such a special design there wouldn't be local tools to fix it and um the this contrast with the old pumps that they would use normally which uh were actually cheaper to build they could be um fixed locally uh with local tools and uh they were easier to use you could pump more water um as a grown-up using them and the reason the kids didn't want to play with these play pubs is that the reason of a carousel is fun is that because if you turn it and you jump on it then you go round and round and round around it's um it just keeps going that's what that's what i know you you got the g-forces that's what the fun of the carousel is but with these carousel you actually had to work because you have to pump the water and so the toy wasn't fun and they hadn't tried it before they installed it so um if and they contrast it with other charity projects who for instance hand out mosquito nets in malaria area so if you hand out a mosquito net for free then um hopefully most of the mosquito nets you you hand out will be hung up above children's beds and they will not get a bit by malaria mosquitoes and the mosquito nets are pretty cheap and um this is a this is what they call an effective form of altruism because uh it's cheap and the um chance of saving a children's life is high and the other thing that giving out mosquito net has is that it's easy to measure that it actually is effective so you know exactly um how many nets you're giving away and you know exactly um how many lives this is going to save and there they have a website where you can see if you're going to give to a charity uh what will be what will probably be most effective um and another uh thing that these effective altruism movement uh hits out as well how can you be most effective so one way is to do something very predictable like these mosquito nets um the other thing is always do your research don't just start building these pumps that and then after two years conclude that it doesn't work but the third thing is to say well if you can do something that will have an impact for a long time then that multiplies the impact of your work so if you um [Music] if you give somebody fish then they eat now but if you teach them to fish then they will eat for many many days so if a person has another uh 10 000 days to live then if you can do a small thing that will change every next day of their lives or you do a big thing that just changes today then the big thing would have to be 10 000 times bigger to have the same impact as a small long-term thing so in general anything you can do that's longer term usually trumps anything can do uh short term and um mostly what effective altruism is applied to is how when you give to charities so um but but the other thing uh that it looks at is well you as a person instead of working a day a random day job and then donating part of your salary to a charity maybe you can do something charitable yourself and you can also measure the trade-off in that so um there's this website called 80 thousand hours which uh states that you have 80 000 hours in your professional life and how can you measure how can you maximize the impact that your work has on the world assuming that that's what you want to do you can get to maximize how rich you will get yourself but i think for most people you know you want to have a comfortable salary and beyond that you just want to do something that's fun to do and that also has a big impact and if you combine that with the long-term uh uh factor then um my question to you is um what would your impact of the work you do now or will do in the next five years or so have in a hundred generations from now so if you think in the long long long term future and uh you'll notice how i'll slowly go towards the topic of open sourcing and decentralized solutions um you could for instance i don't know um sell ice cream that has a very high fun factor to the person getting the ice cream that lasts for five minutes you could also build cars um and a car probably lasts 10 years right so you make people happy for 10 years you could make buildings maybe that lasts for a hundred years um but if you think of um the if you think back a hundred generations um so around about the year zero what could you have done then that will still impact us now um i can think of one person who impacts us from 2000 years ago by teaching how to love each other and but so basically spreading an idea evangelism uh i guess that's actually the that's the evangelist that the word vegeta came from um but in general um writing a book or writing something that is freely available and permanent is i think the most long-term thing you can do and if you translate that to our world then i think there are two things we can do that would have a long lasting effect so one is writing software and the other one would be spreading consciousness about topics so spreading ideas or evangelism um so there are obviously many things you can do in your work and uh i i assume most of you are software engineers there's um um i forgot who it was but somebody who worked at facebook who said um uh the best minds of our generations are of our generation are lost to make people click on ads and i think that's true for a lot of uh big tech companies and probably if you're at the swarm uh summer school then you chose not to work at a big tech company so you're already uh in um sort of in the alternative field uh but i think it's very important to think of what you wanna as as a as a software engineer you cannot be politically neutral you cannot just say oh i just write the software that my boss tells me to write and then just not care about whether that software goes into into a bomb or into surveillance software uh you have by writing software you have a responsibility to think about how your work impacts the world um and um uh i think it's an important thing to think about so thinking about your own career and the effectiveness of the altruistic effectiveness of your work um i'd like you to fill in uh a poll on this website www.mentee.com and then you fill in this code um so i'd like you all to go to that website now and fill in that code and then i can uh have to uh now i have to yeah there so the question is how can you have the biggest impact and um i put three options here uh it's obviously a bit leading um so one way to have an impact is earn to give and sometimes that is the best uh option there was in this book i read about effective authors and there was a doctor who wanted to go to africa to you know heal kids there and um uh he decided he actually made a calculation uh mathematical calculation how many lives he would save by being there for moving there for a few years with this family and uh well the other option of just working in um uh wherever in chicago or where he was um and um i can you please share the code again yes i'll i have a problem clicking between zoom and my screenshot the code is right here there's uh the other option was to just work as a doctor at a hospital and um to earn a lot of money and then give that money to a charity so that other people could change the world in in the poorest countries and that is a that is a realistic conclusion and sometimes it's it's good to you know you can say i have it i pick it i choose a job or a project that maybe doesn't have such a big impact but i know it's going to make a lot of money and then i can um uh donate that money to for instance uh malarianis i'm seeing i'm not convincing anybody yet of the earned to give model i guess that's a bias because you're already in the in the swarm summer school so you already want to build stuff that makes an impact um it's interesting to see that so many of you think that uh um you can have the biggest impact by raising awareness um i'd expect more people to say right software um so it's it's uh it's mainly for you to think about like how am i choosing what i can do and um yeah to be aware of that why am i getting into this uh swarm based software development project so um i'll continue with the slice presentation here so there's combining these things together uh for me got to uh when i um when i thought about what i want to work on i get the conclusion that open source software is important because if i work at a company where i write closed source software then i'm helping this company to sell a product but as soon as this company disappears then the value that i produced is lost again whereas if i contribute to open source software there's only one pool global pool of open source software where project filled different uh niches and projects do um appear and disappear and they merge and they um succeed each other but this is one communal process where whatever you contribute will help one of the current open source projects to become better and the next generation of open source project will build on the on the current ones and so by contributing to open source software now you are contributing to the um the global uh software pool of humanity that can only grow and help us uh become more effective in in what we want and the second thing i wrote down is open networks uh of course when you write software there is a the software will be run under the control of users or under the control of enterprises as part of services they will also create networks network effects so um [Music] maybe the software is set up to only interact with itself um and that creates a network right away um or maybe there is just uh there's an intention to make the software uh compatible and federated with other pieces of software uh but there's just not enough um execution on that intention and that means that even um uh that you still create closed networks and i think the highest value can be achieved if all the networks are internet worked into a global network um so that's why i think if there are ways to build open network open networks that is an important thing um to work on maybe even more so than opens open source software itself which are just the implementations of the protocols of these networks so um the to create open networks you need to develop protocols or agree on protocols and you um yes sorry those are really the two things you need to develop the protocols and you need to agree on them and uh sometimes there's a i was going to write open protocols there but just writing protocols in itself is not useful you also need to agree on using them and they need to be flexible enough to allow different people to to want to use them but they also need to be strict enough so that they actually create interoperability um and open formats is maybe a higher layer of the same thing um so in open networks i'm thinking of who you interact with and open formats i'm thinking of which pieces of software you interact with so um you uh you could have an open network where you share files but then if each file is in a different format you need to convert it before you can use it that's still not useful so i think these three things are um important ways to uh to build good software to build good computer tools to uh to help humanity achieve its goals um it may be a bit cryptic um because i try to choose these high level words but maybe it'll become clearer if i um give the examples of what i worked on myself over the years um so i'd want to just uh go into sort of storytelling mode where i just tell you what i did and how it was and uh what i enjoyed and uh just as like a telling a anecdote at a campfire so um uh there's uh uh in um hi robert did you want to say interrupt something or are i was indeed curious if you could highlight in those stories uh where their their your philosophy you just described how it um materialized into these projects and why right yeah um so i think uh i hope it's obvious from the topics of the project uh so to me it was uh there's a there's the motivation so this password about the motivation or how can you impact the world and then um yeah the second part is how i uh did that in past projects and then the third part is about federated bookkeeping um so uh yeah i'll try so it started when i was uh working at a social network site in um in spain uh it was sort of a spain only version of facebook and um there was a colleague of mine who's uh the the chromebook had just been launched and a colleague of mine said well you know this new chromebook is pretty cool right but um it's also pretty dark you know because if you realize what google's trying to do to us and i had no idea because you know we were all there um hoping that one day we could work at google and that would be just the coolest uh career path that we could think of uh like you were allowed to move from madrid to uh to silicon valley and work at a big tech company that would be really the definition of success for us and um there was we yeah we still thought of google as the when they started with the don't be evil uh meme that they sort of left right right about that time um and he said well the reason that they're launching the chromebook is that it it doesn't have a hard disk so normally when you use a laptop it will have a hard disk where you store your um your files and um this is a laptop that you can use and it feels like you're using uh just your laptop with your local file system but instead of a file system you have the cloud and that's what it's designed to do it's an interface to the cloud and so it helps us to store our data in the cloud under google's control um and um that's obviously google wants to do that so that they can uh target more advertising to us uh making people click on ads and it's obviously not good for the user because uh you want to be in control um of your data and your uh your digital integrity as we saw in the first presentation and um it was sort of a wake-up call to me so as we uh started to sort of brainstorm like oh yeah but you know what what could we can we do something to change this could we just build a prototype of a a better internet and um then i uh started thinking about this more and i came across this presentation from eva moglin uh samuel this was 2010 and um in this presentation ian magnum is the uh important partner of uh richard stallman at the law free software law institute i think it's called something like that and um he gave this presentation freedom in the cloud where he explains exactly that yeah these companies want to they give you services but they really want your data and they want to take away your digital sovereignty and that is their business model they want to they build platforms which uh just lock you in and um he gave this talk at the debian conference to an audience of debian contributors and he said well let's take a plug computer like a small computer this raspberry pi didn't exist yet but you already had small computers that you could plug in uh just a wall socket and um that you could use as a local file store and you could run some services on there and some um firewall and peer-to-peer software and that's and that's when he proposed the freedom box project and there were more projects that year there was diaspora there was there were projects that about personal data stores and um i thought well let's try to uh with this colleague this thread i said well let's try to build something where uh people stay in control of the data they upload and the data that you upload is actually not sent to a central server but it's distributed peer-to-peer and i took some time off work to try to build that and i ended up building a the unhosted project which is a separation of applications and data and um there were some there was some web web platform technology that was just new where you could do cross-origin requests from one website to another and there could be a website that acts as an application but instead of creating an account in their database you would tell this application you would tell this website well my data store is there and connect my data store from the browser cross-origin so that was the idea of the unhosted project you take the application into the browser and you make a connection from the browser to the user's data store so bring your own data store to a website it's the idea of personal data stores implement using implemented in the in the browser basically and i got some uh some good feedback for that from a journalist wrote about it and it got into hacker news and then i went to the open source conference and people told me well you could really with with the twitter followers you have now you could really crowdfund this and quit your day job and i didn't believe them but somebody else said oh yeah definitely and then say okay well if two people say it then i'll just try it and i had six thousand euros left in my bank account and i moved to berlin and i said well i'll live off a thousand euros a month uh so i can run this project for six months but for every thousand euros that you donate i'll add one month to the project and that crowdfunding campaign raised we printed t-shirts with this logo on it and uh the crowdfunding campaign raised four thousand euros so that was for another four months and then the nlnet foundation uh the same foundation to which radically open security uh donates their uh ninety percent of their profits uh funded my project uh which means that i could work on it they funded it multiple times and i could work on it for five years uh before the funding ran out so that was great um it was a great time got lots of fun and i didn't have to think about uh how to make money i could just focus on how to uh improve the narrative about what we were trying to change and to write code and example apps and try to get people into um building on hosted web apps and uh we did it together so that's me on the viewers right in the middle is uh kenny the guy who made this remark about the chromebook to me who inspired me to uh to think about it in the first way and on the left is jan who works at next cloud now and this is on our the first anniversary of the project we went to uh we found out that the hills in the hills of bohemia there's a village called unhushed which is the first six liters of unhosted so we thought we should just go there uh so we went there a few years in a row for our conference tiny village but funny that it had the same name um another thing we did uh as you know because we didn't have to think about money uh the money came in from donations and we didn't have customers we just have the open source uh goals um we could just go anywhere which is digital nomads right just a laptop and a backpack and then at some point we said i would go to southeast asia in the winter and i met some other people who did the same thing but then the summer maybe you still want to be in berlin to meet some programmers again and then we said why don't we all go to the same tropical island yeah that would be great so we went to the same tropical islands we called it hacker beach so we um created a twitter account and um said well let's uh this year let's go to fuquak in vietnam and let's talk to the same island so it's totally unorganized just people with laptops uh sitting at beach uh restaurants and uh this january is gonna be the 10th edition of hacker beach so if you're a digital nomad that you can find them on um on irc or on twitter and uh say that you want to join and vote on where to go and it's all all the month of january it's just hackers on the beach it's a great uh great fun experience and so um what we worked on uh was really the idea of a personal data store so allow people to control their personal little file system in the cloud and uh the so the that first part is easy to give somebody a file server you just now run a server with a user database you say okay here people could have a an account the model we used was the the federated model so different services offer personal data store service and you can just switch but it's based on um the domain name so your your personal data has a url and um the freedom comes from the fact that you can switch and you could also self-host it could even just be if you have a static ip address it could even be in your house so it was not as decentralized as ethereum but still decentralized enough to decouple the data storage from the applications and um i worked on the on hosted project for five years and then i worked on firefox os which was mozilla's attempt to create a smartphone operating system uh or they did create this uh they stopped the project because not a lot not enough apps were html only uh to keep the phone attractive uh but it was a great project and we learned a lot from that and after that i also worked on the solid project with uh sir tim berners-lee which is very similar to the unhoused project in that it gives people a personal data store and it allows people to control to choose like my data is at mchilldiar.com or my data is at this service which i trust or it's my dropbox etc and the difference between solid and host is that solid is also uh aimed at linked data and uh from a semantic web perspective so the data that you store in a solid part would normally be a lot of linked data with uh uh subject predicate object triples and knowledge stored that way in a machine readable way um and in an announced personal data store you maybe uh store more just jason um so the problems we run into is really well one big problem is who cares so a lot of people who would so developers would try it out right they would get a personal data store um but it was hard to really interest a larger group of people to sign up especially if we would ask money for it so um there were not so many people who who cared that that much about having a personal data store i think that did change a little bit in the last 10 years there's much more awareness now in the general public about for instance entrant encryption [Music] but at this time uh it was it was pretty hard to find actual lots of enthusiasm but not so many real users and especially the fact that the service has to be the storage has to be paid for makes it hard um and i guess you'll get the same problem with uh with swarm that somebody has to pay and um uh there's or contribute uh to the hosting costs and um there's an added barrier just the payment the action of the payment is is a ui barrier uh and i guess with swarm that you yeah first you have to um uh get into the the right tools to actually be able to pay for it um and maybe our biggest uh problem that we underestimates who built who builds the apps and there was we built some apps but we focused mainly on providing the the sdk and making it easy for other people to build the apps and then we did hold uh hackathons a bit like this summer school we didn't have any uh grants to give away but um there and there were about 20 apps built for another web apps uh and um uh but really not enough to really attract a lot of users and the feedback we got from app developers was often that like yeah it's one thing to build an app and then it's harder to build it on top of personal data stores uh because uh you might not be able to search the database or change the schema like you could in a normal lamp or or ruby on rails or django stack um and um the third thing is that the sign up for their users becomes harder so they say yeah use my app oh first you need a personal data store otherwise you can't sign in and so that's confused and then you have to choose because you want to give the users the choice of where their data is stored so that makes the ux the sign up a lot harder and um anyway those are some of the questions we ran into uh the protocols that we so reebok storage is the uh the wire protocol of the unhosted project and um uh web dev is the one where people say well why don't you just use web dev which is used by for instance next cloud and own cloud and most personal cloud servers where um the difference between a personal data store and a personal cloud server is not that big but the personal cloud server has the uh it's mainly thought of as a file server with apps on there so the apps are on there and a personal data store is just the storage of other apps so it's really a back-end to other apps and a personal cloud server is more integrated but in both cases it's about uh choosing where your where your data is stored and solid has its own protocol and recently we wrote the test suite for solid and we also wrote a plugin for for nexcloud which makes a next cloud server solid compatible so that's sort of the next step of open network so we take these protocols these personal data store protocols and then make personal data stores interoperable with each other because i guess there will never be one single system for personal data stores or maybe yeah maybe one day there will be a dominant one but uh not anytime soon so a good thing we can do now is to make them interoperable for instance with server side plugins or with client-side polyfills or abstraction layers and it's uh it's good to know that the when i started in to look at personal data stores and tried to you know build a more data fair internet there was crowdfunding was just new then so there was um um the diaspora crowdfunding raised 200 000 dollars that was big news and then so there was some awareness about crowdfunding and there were some grants and so it's just possible to just say okay i'm going to run an open source project but now the landscape is actually much better especially in the eu as well thanks to the next generation internet initiative and i think it's there's it's partially due to the mainstream awareness that's really grown especially also after the snowden revelations and cambridge analytica and more and more mainstream press has been thinking about data aware applications and networks um but also the eu has realized that um yeah it can't just let all its citizens data sit in the hands of american companies and so the eu wants to stay a little bit sovereign in that sense and also um the eu wants the internet to stay human-centric um and uh i think it's worked worth taxpayers money to donate to next generation ngi central generation internet um funding so the way ngi works is it's split up into a number of um funds each of them are about six million euros one of them is dubsy where i'm in the virus advisory board and dobson stands for data portability solutions or systems i don't know some ass incubator and the goal of dobsy is to give money to projects that improve the portability of data so um it's uh it relates to the the personal data store interoperability that i mentioned earlier but you could think of it as a higher layer of data freedom so first of all you choose where your data is that is the personal data store then you say okay now i want personal data source to be interoperable because i don't want to be locked into one specific type of data store so that's the at the protocol layer um and then the third so yeah so and the third thing is when the data is in there i want the different applications that use that data to store the same data in the same way so this links back to my earlier slide of open source open networks and open format it's about you want your an open source server software that can host your data uh and you want uh so that you can choose where your data is and you want um the protocols to be interoperable so that there is an open network there's one open network of personal data stores and you want um open data formats so that is if you have for instance a photo app that stores a photo and then a photo album app that allows you to create a photo album that those two start uh actually well image format is a bad example maybe suppose you have contacts data so you have data about your friends in your address book and you store your address book on your personal data store and then you have a different app for instance the chat app where you want to be able to open a chat with somebody and then maybe uh maybe it can say see that you have 37 contacts in your address book and it can display the names but maybe the avatar is just broken in this app because the app you store the data with uses a field called avatar and the apps you're reading the data with uses a field called profile picture and it doesn't know that avatar and profile picture are the same thing that's the data format problem so um data format is a huge problem in any type of software interoperability um even between versions of the same piece of software you get this problem you want you always have data migrations where you um have to reshape the data to work with a new version of the software but between um pieces of software it's even harder because no everybody just goes and write their own thing um it's even happened to me that i wrote an application that was not even compatible with an early earlier application that i wrote myself just because you know i went from a different design principle and ended up storing the data in a different way that's um that makes personal data stores almost useless so if we can't get the data formats to be interoperable then it's almost useless that you control where your data is if you can't read it out with other applications and so one way one approach to solving that is with schemas and shapes and that's been tried since the 90s you know with xml um data uh what's it xsd and xslt et cetera um and a lot of research has been gone into how you can describe your ontologies that describe a data format and this restrict and validate a schema and then but the problem with that is that yeah sure it's good to formalize and to validate the data format but then what you get is two applications that just use different schemas so you still have the same problem at that level so what i think now and it's sort of a school of fault in data probability is that you should always let the data [Music] you shouldn't try to choose one best schema that everybody wants to use you just have to document in which schema the data is and then have conversions when you read the data so the app that reads the data should be able to read data from dozens of formats and this is uh there's a it's related to a trend in in data warehousing in general where uh uh for instance big enterprise software packages look at uh uh data logs from big systems they used to call it a data warehouse and then somebody said well let's not think of instead of thinking of a data warehouse let's think of a data lake where we don't store the converted curated data but we just thought of raw data in its source format and then um apply the schema on read so um try to have you can have that in many different data formats and the effort should be in making your application able to read data from as many other applications as possible and of course you want that reading part of your software that is obviously a a part of your software that you can reuse so then you can um have these transformations accessed in software libraries that you reuse uh that the risk there is that somebody tries to say well let's uh convert from 30 different formats to this one super new format and then um that's of course trying to propose a single formula um but uh so it's it's an almost it's it's necessarily a hard problem to solve but it's a problem that you will have to try to solve if you want to give people data sovereignty because your application will be storing data and you want other applications to re to be able to read it and your application should be able to allow people to switch to it from other uh formats and uh so it should also read other existing formats so if you do anything to do with beta portability you can apply for this grant uh there's another call open next month where you can get 150 000 euros if you have a good proposal um data portability incubator from next generation internet so um maybe i primed you a little bit but my next question is about uh what is important in um with regards to how data is stored so i wrote down four options and uh so the first one would be portability so if you want to uh so to a to an end user or to humanity as a whole the most important thing is it that data is portable or is it that data is confidential and this includes confidentiality includes uh protection from state surveillance so uh i guess that's a big motivator for people to use swarm would be that yeah and then the nsa can't get to it if it's encrypted um availability obviously you want data to stay availab available that's why otherwise you just put it you could keep it on your own laptop and you would have to do all these complicated things storing it in the cloud and this includes censorship resistance because if your data is censored then it's no longer available and um the fourth one i put it there is attribution and um attribution about when you create something that is of value to others you want to be um you want it to be known that you uh contributed it so it's like you wrote a song you want people doing it it's your song it's just nice you know it feels good to hear that people liked your song and part of attribution is monetization and i think when you create content it's important that if you create content as a full-time activity then you need some way to make money with of that and the way it works the classic web business model is to say i'm going to put my content online and i'm going to put ads on it that's the classic business model and then so that leads to user surveillance because then the ads get make more money and so that's a business model we don't want but how do you monetize um creations uh if not through ads so there's a another good grant that you could apply work uh for its grant for the web grant for the web does assume that you use their web monetization standard and it's which is linked to interledger which again is sort of linked to xrp but um the ideas behind it are very general like how can we create a different business model for the web um linked to brave browser obviously as well and flatter and yeah so in the way your app stores data which aspects do you want to improve most compared to classic big tech and compared to google drive what does your app do better is it data affordability data confidentiality including state citizenship res story state surveillance resistance is it availability including censorship resistance or is it attribution including monetization so we can go to the same mentee.com website um same code and i'll um go to the next poll here yeah data awareness what matters most the code is here oh sorry six seven four six nine seven three six let's see what people say soft um cool looks like uh confidentiality is the winner there which is uh yeah so it's uh it's interesting it's uh sort of to expect it i guess uh but it's interesting to me um as i'm not from this forum project and so maybe i haven't thought of it that much um but it's really cool to see uh and yeah i agree that with um uh actually with this first speaker uh emphasized this as well that uh it's not about the um the nothing to hide argument it's about um your safe home uh in the digital world uh should just be uh integral and should not be uh snooped on cool well great to see uh how you all think about that question uh i've got another open question uh coming up soon and then you'll just um i hope you can do that with a zoom or where people can just actually talk but i'll um continue with my um [Music] a little bit more of my projects that i worked on and then we'll get to uh how you want to structure your project so one project i quickly wanted to mention now is terms of service and read robert already mentioned an introduction and this is a raising awareness project that in the first meter scored highest higher than writing code raising awareness so that's actually all that terms of service and greed does we crowd read the terms of service and the privacy policy of big websites it's uh which is something nobody does you know it's the biggest lie on the web where you tick the box and say yes i have read and created terms of this website you just want to create the account and obviously you didn't read it so um we um as a pun on tldr we created to sdr and actually started reading them and then rating all the paragraphs so we had some labels where i said okay this is a class action waiver this is about how long they keep the logs and then from all these tags we created a rating and then it ends up with a label of class a for very good or class b c d e e is very bad and this started eight years ago as sort of a side project and then we kept doing it we took over the toss back project from the electronic frontier foundation uh toss back crawls the terms of service every day because uh lots of servers say we can change them uni laterally and you can uh you have to check back to this site uh that's toastback.

org and um i just heard this morning that we're on hacker news again with the uh terms of service didn't read projects so it's great to there's a new now i'm no longer directly involved in it as uh since last year there's a new team working on it and a new crowd for the third crowdfunding campaign um so that was anyway that was my awareness uh project that i worked on uh and then i um before i moved more to money systems if you will so um what i work on now is ponder source and as i said i had a startup incubator where melanie rubrick gave a talk and i decided to make it a non-profit um but then still there were some uh choices to make about the details and um one of the questions i asked like yeah should it be investor owned well i guess that's i already answered that with no but then what a lot of um innovative uh business models do is they are sort of co-ops cooperatives so they are worker owned and um a lot of um non-profit companies are for instance let's they say let's be like uber but owned by the drivers and um then i thought well why would it have to be owned by the drivers or even owned by the workers um there are some very good examples of factories that were owned by the factory workers for instance so they could vote about how it should work but in a software company i still think that worker owned is not good enough um because as workers we uh we got a salary and we got our bonus scheme and we got our pension scheme and that we should by that with that we should be have been rewarded we got our attribution importantly and then from that point on um we the exchange is done you know we've done our work and we got paid for it and if i um i shouldn't be able to then later say like oh i had this i created this ad blocker and uh it's open source but it's mine and now i'm gonna sell it and the ad blocker becomes a advertising plugin um you know even if i wrote it when you open source something and you actually you really donate your software to the commons of to the common good of humanity then i feel that you have a right to attribution so people should always know that it was you who wrote that software and they should praise you for it you get the applause um but if you really release your software into the public domain then you should no longer be the owner of it then it should just be owned by the public domain owner in the sense of uh who can shut it down or who can sell it uh et cetera and um that's why we went for a foundation and so and i've been i thought it was we were gonna make it a co-op but then in the end uh i decided i thought about the fundraising um uh institute construction um and that was a bit too altruistic for me i or with two not that's not the right words it was giving away too much control so i thought okay we're gonna do earn to give projects so we have a balance between projects and products in projects we get paid for the hours we work and that's the stable basis of the company and in products we develop something new which maybe in two years from now will make us some money or not and um sometimes we just do things that we don't that cost us money uh but we think they have to be done so within pondersource we balance those things so uh we have a few projects that we find important but that make us money and we have a few products that we also find important but that cost us money and um in this balance there's a uh yeah once we've done this work uh [Music] we released software into the public domain and then we should no longer be able to uh to control what happens with it um there's um okay so my next slide is uh sorry one slide out so we have the way we did it in the end so it's not a bonus uh it's not a uh sorry not a cooperative but we do have a bonus structure i think because um um yeah this is sort of proven to work well in uh vc funded startups uh the idea that you have a small team and you have 360 reviews of the um people you review your manager your manager reviews you and if it goes well you get a bonus so that motivates people and you don't want to give that away you could and if you could make a totally volunteer led project which you know the house project was and that's nice because it doesn't cost any money but it's also hard if you can never tell uh if you cannot coordinate what people work on and uh and you cannot align um people's goals as well uh with volunteers as with uh people who are committed to um and rewarded for uh reaching company goals um so there is a bonus uh a salary and a bonus if we reach the quarterly goals there's no ownership and everything we do is for the commons and um so my question to you as the audience uh and i don't know because i need robert to uh unmute people if they want to answer this i got two questions um so when you think of your project that you're going to build on top of swarm uh who will decide so um who will own the keys to the future of the project and who will profit uh you know are you gonna uh um i guess icos are uh i don't know if they're illegal anymore but if you maybe you want to have some sort of reward where uh the software development is paid for um so how will the profit structure work and how will the decision structure work in the governance of your uh swarm-based project yeah uh thank you you can uh ask or respond via the if you raise your hand in zoom then you'll be unmuted to otherwise we'll also keep in touch with the discord channel and the qa section if you raise your hand we'll give you the voice in the meantime again [Music] if you do intend to build your data aware decentralized application take a look at the next generation internet grants and the other one was the groundswell grant for the web some people don't know yet who will decide who a profit so did you by the way given the infrastructure we have these days uh the opportunities with decentralized solutions do you think of making your organization a doubt to um automate it that way um [Music] yeah i didn't uh i didn't really consider that i don't know really why maybe i should have considered it longer um so um yeah good question uh well we need um we need a legal structure anyway um yeah or we could be a dao with freelancers i guess yeah um i guess i sort of automatically started searching in uh the the traditional legal system for a an employee structure one of the reasons is i guess it's easier to hire people in traditional working contracts and i did what i did there were two things i did consider one of them was freelancers but i think that it's it's good to get good talent from freelancers but it's hard to bind them and to build up a team and to build up a team feeling and uh and a flywheel i think but we do work with um we do that a little bit we have one freelancer um which is just a friend i did previous project with and so yeah repeated freelancer which is sort of the uh one uh you could also call it the co-working model and co-working spaces are flexible offices but they are also places where freelancers temporarily work on a project or somebody find it like that to me once and um the other thing we considered was git bounty or git was get tip and get bounty i think and bounty hub so just to give out bounties and um uh see if people would just write part of our code for a bounty um but if it felt scary i got uh two questions here they're both uh open-ended so um first of all from gregor i really learned that among others you mentioned how apps data can become unbounded as today as users were convinced that it's all the same still this is quite abstract to end users how do you think we can increase awareness in your opinion what are the biggest obstacles to adoption of the proposed paradigm change um yeah that's uh that's a good question if you do it well then the user doesn't notice that the app is storing the data on swarm or on a personal data store and then it doesn't change the experience of the user i think but i think the aha moment for the end user is the second time they use an app with that data store that it's already pre-filled with the data they used from the first app so for that to work there have to be two or more good apps that people would both want to use um but if there are enough apps then i think that's a real incentive for end users to use it so that is um of the four things you want to improve about data awareness that we voted on earlier that's the portability one where your data is portable from one app to another um and for um [Music] yeah for availability you could think of a sort of a wikileaks app um you know that's um obviously when you have data um that you don't want to be censored that could uh that could be a pretty self-explanatory niche application um and i think for confidentiality it's hard because in practice people say oh that's so bad i'm never going to use facebook again but then they do keep using facebook because you know they're just using it and while you're using it you don't really care and when you think about it you do care um it's like well yeah when you think about how your chicken sandwich is produced then if you think about it too much you don't want the sandwich anymore but if you don't think about it you just eat it um so that's the harder one but i think portability is doable with two or more apps for confidentiality it's doable in a niche um and um the hard problem i think is the chicken and hack problem between apps and users because app developers want a lot of users who have a data store on a given platform and users want a lot of apps to be available before they're gonna sign up so that's uh and and the way to break that cycle is for instance with grants um as swarm is doing now so i think that's a great uh great approach thank you um the other question was um you mentioned the interoperability between these different platforms and where do you see synergies between swarm the unhosted and the solar project if any um well i think uh you could have uh you would have to write a polyglot uh client library so that that talks all the apis and we've been doing that a little bit we so f within the unhosted project we created a client library that can also allows the users to store their data on dropbox or on google drive and then you could have apps that are compatible with not just a swarm or not just solid with both and that would be really cool uh so i think that would be uh should be doable the um and then it's up to the user whether they want so then their data would be safer in swarm uh than in solid because in solid it's still hosted on a pod provider which is still run probably on aws which still has the nsa still has a backdoor um so uh that swarm is sort of the higher goal of the higher level of decentralization um but you could uh um yeah you could imagine writing an app that's compatible with multiple storage uh platforms and now through that there's this energy because you know one app helps multiple storage platforms forward um there are some technical hurdles i guess to get it to get that working because every you want to the start platform sort of dictates how the uh how you're going to access the data um but in theory that should be uh that should be a way to have this energy and awareness obviously um just to uh jointly uh create awareness i thought maybe there's somebody from the australians just want to talk just to say what their project will be or what they want to build or what their questions are somebody from new orleans wants to say anything yeah again if you raise your hand then a gospel can give you the word yes we have a question from petra so actually you know my project is kind of like developing tools to help new ways of deciding uh who and how to distribute profit in particular so um pretty much in all information systems that we're dealing with there are data sources and data syncs and it's about cl you know controlling or creating a system by which we share control of that transfer of information across the network um and then how you know how is it propagated but also how does it evolve and how is the attribution tracked by the system in a way that's anonymized or separated from the actual interaction with the system by the end user um so that you know after the process comes along um and there's a product from the process then uh as that's being consumed in the data sync that triggers back a feedback loop right uh a lot of dials are are functioning in that kind of architecture um and so i think uh i guess my my question to you would be um and it's related to this idea of uh intellectual uh philanthropy uh effective uh philanthropy right um one of the things i've run into in non-profit organizations that they tend to become dominated by uh the sources of funding um that are available to them um even if the best of intentions from those organizations that are trying to be socially responsible by contributing to those non-governmental or non-profit initiatives there's a sort of social endearment that occurs uh in in these ideologies so um i guess uh you know a lot of what's happening in swarm right now i think is this uh development of the idea of the difference between data and capital um but you know in cryptocurrency we understand data is money so i guess uh that's kind of my question is how can we separate data and money from capital uh with these new systems and these new architectures that we're developing here yeah i think um so personal data is uh primarily for your own use until you share it as a user as an end user so there that data is not money or at least it shouldn't be it becomes money when potential advertisers get to look at it but that's not the goal of why the application is storing your data and there's on the opposite end of the spectrum there's public data which we want to be published in a censorship proof way like you know the contents of wikipedia which i guess is the the prime example of what will still have an impact a hundred generations from now and there's when you contribute to the pool of public knowledge then um i think you probably want attribution it's nice to sign your work um so i'm interested in your idea of anonymizing that part of it so as i understand it's you could contribute something and get money for it but stay anonymous so not get applause for it well you i guess you could get applause under a pseudonym then right um so um uh yeah and there's a when you contribute to public data there's there could be just a um there's not a specific end user who profits from it uh so you're not there's no transaction there's no receipt there's no recipient because humanity is the recipient of the data you produce that the data sync is the commons and um yeah when you share contents with your audience maybe as a creator that's where you want to monetize it so i think there it's would be very useful to have just a way to subscribe to a podcast maybe on swarm uh where um the consumers of this podcast are just uh paying to access it and um there there is the um you could pay for access you could pay to um just you could donate to an activity that creates data uh you could also contribute to the hosting uh of the data to make sure it stays available um so yeah it sounds like a cool project where um you make these things possible and you help people to set up these structures i guess you want to give people all these options um you know contribute to the hosting uh make it public uh pay for access i guess when you say you have to pay to hear my podcast you don't really want to exclude the people who don't pay you just want to give them a reason to click pay before they consume so another almost equivalent model is there is that you can say you can consume it and then later donate and something that has been done a lot is flat rate so uh started with flatter but spotify does the same and so does uh coil with uh who are partially behind grad for the web where you say i just contribute five dollars a month and that gets distributed over whichever data uh the system says i have used uh thank you definitely um for your insights and the presentation thank you very much cool i'll um i've got a few more slides but i see i only have seven minutes left uh i wanted to uh robert can i you just use these um last minutes to talk about federated bookkeeping uh yeah this might even go into might relate even to petra's question in this case i have one final question for you later on oh okay uh so i'll get uh am i in my yeah um [Music] so there's hold on i have to i have a problem controlling uh keynote and controlling zoom is my point of dispute so federated bookkeeping the idea is that what we don't want is banks obviously um and um there's the idea that for alice to interact to send value to bob she has to go through a bank for an anonymous currency because the um the us dollar is backed by a war machine that will and and the police that will collect taxes that will back the us dollar and um this is an anonymized anonymized bubble where alice has a balance in dollars and bob has a balance in dollars and um you swap these for each other but really what you want to do is to keep track of what how alice and bob interact with each other and you don't want the world to be about how alice interacts just on her own with her bank and how bob interacts on his own with his bank where you know every you get a model where the the structure of the world is made up by this bank and every user is isolated in a little pod in a big matrix of little pods where we dream things that aren't real and we end up in the movie the metrics that is what banks are they separate us from each other and our only interaction is through this systems but a provocative thing for you ethereum users is i will argue that when you use cryptocurrency that is also an anonymized um currency a um any cryptocurrency uh bitcoin or ethereum or xrp or whatever is a bubble that is held up by speculators and when i want to interact with you i interact with the bubble and you interact with the bubble and we could have just interacted with each other peer-to-peer and so that's what i tried to to build for the last five years or so is a system where people can interact peer-to-peer so i i formulated the whispering merchants problem which are these merchants sitting around a tree and everybody is only allowed to talk with their neighbor direct neighbors but you want uh to exchange goods so that value goes all the way around the tree and everybody receives something and everybody gives something um and uh a solution i uh found for that was well actually i first invented a solution and i formulated the whispering merchant's problems to explain what my idea did it's called ledger loops uh where you send a conditional transaction around the tree in one direction and when it arrives back at you you say aha it made a circle so that's uh full circle that's good then you send back the pre-image of the hash condition goes around the tree in the other direction and then um you have a two-phase commit of a circular uh chain of transaction then somebody pointed me to that actually that's the same thing that inter-ledger does and i later found out that lightning networks also does it and there was a ethereum based uh projects for which i forgot the name that also uses um timelock contracts to create a two-phase commit where you do a multi-hub transaction in a peer-to-peer mutual credit network and if you look at uh ripple classic which was original ripple before xrp started um there was uh the idea was to create a credit network sort of like a local economic trade system but where everybody can issue currency and you can just send ious and um i've been researching that uh i ended up working for ripple in luxembourg on the in the ledger protocol and after that i researched it some more and i i renamed it to federated bookkeeping and it's the idea of let's not talk about payments because payments imply that we're settling against a bubble and um i want to talk about how we coordinate uh how we trade with each other and how we coordinate what we give and what we take to our environment and that's just bookkeeping a side effect is if you talk about payments then you have to you fall under payments regulation if you talk about bookkeeping you don't fall under as much regulation so um it's easier to have a federated bookkeeping node running um without uh being um falling on the money laundering uh law then for instance running an interledger node um and the idea of federated bookkeeping is that you are uh you have your bookkeeping system you send messages out you receive messages in these messages only have as much value as you assign to them based on how you trust your neighbors but with repeat by doing circular debt resolution in this network you can have the same effect as a money network as long as the network's big enough so we now have a weekly call there are uh you see there's five of us we're sort of a team that formed around federal bookkeeping that brainstorms once a week there's a getter channel um if you go to federatedbookkeeping.org you can get to get a channel from there you can get to our weekly ditsy call um and um this is our sort of novel idea where we say well yeah okay cryptocurrency is still bubble tech what if we could connect bookkeeping systems peer-to-peer into a global network uh call it bookkeeping and not payment and that way be connected but sovereign so you control your own data because it's your bookkeeping system whether you host that on a encrypted uh system in your pocket or you host it at amazon that's your choice your you can choose where you keep your bookkeeping you could even do it on paper but it's connected through messages and these messages can be used as well or better than payments to coordinate trade because if we pay each other then we give a price signal uh which says i need what you made because i'm paying for it so apparently i need it but these price signals aren't that good you know sometimes we're lured into buying something which we don't really need also it's just one number so maybe through a bookkeeping message i could tell you much more precisely what i need or i could tell you conditionally what i need so i could say i would buy this stamp from you if you also have that one because that would complete my series and so you could do more complex almost like smart expressions of uh price signals making the uh and then the hard part of course is to to implement the the um the invisible hand in this system so that uh the value flows through it um but the concept is that federal bookkeeping is a generalization of payments and if you look at uh how banks work and how money has worked since the departure from the gold standard correspondent banking is really just bookkeeping there's no um uh money move no money is moved um there's only uh records that keep track of what people owe each other it's only iou records that change um so the thing we talked about literally since last week is if we could do virtual organizations this is part of robert's earlier questions have you considered being a dao so we started thinking about could we do virtual organizations with attributions in a peer-to-peer mesh so could we use federal bookkeeping federated bookkeeping as a an alternative to daos and the answer is i don't know yet we're trying to work out what that would mean and another thing we're doing is to make fair great bookkeeping work is we looked at how electronic data exchange works and that's been worked on since the 80s people started saying paperless office they said well we should stop sending paper invoices and instead people are now sending pdf invoices but since a few years um the eu has backed an e-invoicing network that is uh becoming european there were some national ones already um it's called pebble and um it the architecture is uh very top down so in order to enter pebble you need to get certified as a gateway you need to and then you check that the participants are legal entities uh legal um yeah legal persons in a country uh so that's very top-down and government-oriented um but one project we're doing which we wanted to call um pebble direct but we got called back saying no you can't call it pebble because it goes against all the principles of purple if you remove the man in the middle um so we called it as4 direct as4 is the underlying protocol of purple one layer down um so the idea of as4 direct is to say we do e invoicing the way the eu proposes but instead of using a four corner model where you have to send to a sending gateway which is certified and that is allowed to talk to a receiving gateway that's also certified and that will deliver to the recipient uh so you see two men in the middle with every message that you send which is sort of the opposite what any um blockchain person would want to build we say well we just use the same format it's actually xml and soap which is fun and old-fashioned um but uh it works and we use it directly from bookkeeping system to bookkeeping system and in order to know that an invoice comes from the company you think it comes from we use self-hosted identities where on your on the website or the the endpoint of your bookkeeping system you host your certificate which shows who you are and we also want to link this to people so that sometimes you may send through if you said to a government you have to go through pebble so there will be men in the middle but if you send to a friend to maybe a freelancer who also works on the project um then you can just send bookkeeping to be keeping or send through a store message for this um and then the next thing we want to do also to and that's when this presentation comes through circle i was afraid i didn't have enough slides for an hour and a half but apparently i did so uh we're big fans of the let's encrypt project um which gives free ssl for everybody i'm sure you know it you can just run a script and you get a cell certificate for your website they have 285 million users who don't pay and they have 95 donators sponsors who pay for the costs of running one server and a team or one big database server one small cluster of servers and a team of i think 16 people um so a very small team that's changing the world of https and we want to do the same for pebble so as we're an earn to give company we have two types of activities projects and products the projects make sure that we keep routing and that we have enough money one thing that's maybe implied but i didn't mention explicitly is that we hot we our team grows organically so since we don't take venture capital we can only hire the next person if there's enough money in the bank uh we are uh 2.2 persons now and we want to move to 3.2 soon uh but first we wait for our uh project customers to pay some invoices and then we will hire the third full-time person and we also use our income from our projects to fund products and that's how it's possible that we can give away products for free and our main project that product that we're exciting about are excited about is well it doesn't have a name yet you could call it let's purple because it be let's encrypt for pebbles so right now it costs about 40 euros a month to get a pebble id and we've been talking to a number of certified pebble providers and asking them for price information and we're just going to mix the cheapest ones and then pay for that ourselves from our profits and then give the pebble connection away for free so that everybody can get onto pebble and we got electronic invoicing for everybody and the invoicing sending invoices is can be the basis of the federated bookkeeping so that uh you can keep track of for instance work you did in a project uh when you fix an issue you send a sort of invoice to the what would be the dao um and uh to automate all of that also between um uh machine operated services so you could have a automatic service that when you send a purchase order a server is started or a node in swarm has started and you could totally automate this bookkeeping process and create structures um that sort of do business with each other in an automated way but in order to do that business one way to try to achieve that is to say well let's do micro payments where services pay for each other that's what um princess cody is does and our approach is though well let's not use payments but let's use bookkeeping so just purchase orders and invoices that keep track of what these services need from each other and when a new node needs to be spun up etc so yeah that was it uh thank you for your uh for the opportunity to speak here it's been a lot of fun and uh uh thank you for listening good luck with all your projects uh i mentioned a bunch of projects but the one side i love to see you is federallybookkeeping.org and um pondosaurus.

org is our our non-profit uh company and um yeah wish you all the best of luck with your swarm projects and uh i don't know if we want to do questions or uh or wrap it up i'll leave that to robert yes to all participants don't leave yet we have some announcements to make and i have one closing question for you michael which was um or is from church as you mentioned or how what what impact can you make in 2000 years in 100 generations what would be the lifetime of the decentralized technologies we're working with today will they stick around for another 2000 years or will they be gone in the next 10 years what is your expectation on that and you're you're muted that way yes sir yeah it's interesting especially if you think of uh cryptocurrencies then it would be so easy to just fork uh to say i'm i'm um uh instead of bitcoin we'll start using uh schmidt coin and uh let's just uh start with a new um genesis blog um and if that happens that all the money that was ever paid for bitcoins uh will have been will be gone because the value of bitcoin would go to zero um but as long as there's no reason to do so that will just not happen and we'll stick with the same cryptocurrencies that we have now um so i don't think that the current um dlt networks will necessarily stay for a thousand years there'll be new ones uh that succeeded uh if you look at how many news new ones get created uh all the time um but the new ones will be inspired on the older ones so uh in that they they live on and then um all these inventions live in live on in the next generation so um i definitely think that um uh that smart contracts and uh um data aware apps and all the things that we're working on now are a change that in 100 years we'll say oh yeah that's when they invented data aware apps that's when they have ended swarm and it'll be in history books um and um we will use something like swarm it might not be cold swarm but um yeah i'm pretty sure that in 2002 if we don't uh bomb each other that we'll still be using something like swan by then thank you thank you for that