Why web3 is failing - Malisa Pusonja | Attic42

foreign no yes okay so yeah optimistic topic for some time a couple of years now it's more than six I guess I have been in this space and the core of this space is decentralization and I sensed from the beginning up until now that um I cannot say nobody but the feeling is uh that people are not happy with the decentralization as basically the percentage of the centralization of the ecosystem and it goes for the whole web 3 world but mainly I will focus on ethereum because it is mostly focused on decentralization in not and that is its core value this argument is not about ux and UI and stuff like that but I think it is relevant for builders in this space because uh if it is right then something is maybe conceptually wrong it is brand new argument so maybe the argument does not hold but let's see together so I want to expose you to the argument and then we will go through it so it follows like this first part of the argument is the meta it's fundamentals I will explain what does it mean then I will ask the question does this hold for the web tree and it's uh if it holds for red tree it calls for ethereum so so yeah and then the third part of the argument is centralization goes up the stack from protocol to applications and then the fourth argument is basically conclusion that because centralization and profits move in same direction that may be the reason why we feel that web3 is failing it may be objective or subjective thing that is up to discussion now so yeah let's see first part of the argument is meta is fundamentals what does it mean uh the best way to tackle this topic is we are explaining in short something called law of conservation of attractive profits which was the written how many of you have heard about this law it's non-scientific law but nonetheless Clayton Christians in the author of innovators dilemma and a couple of other famous books basically created this law which states that attractive profits will move to a next level of of some chain he calls it a value chain it can be a tech stack or whatever once the modularity and commodization cause attractive profits to disappear so basically if commodization and modularity cause to disappear cause attractive profits to disappear attractive profits will be basically uh moved to the next stage to the next stage to the next stage to the next stage so once we know that one additional thought here is basically thought of Douglas horoscope I highly recommend that you read his books is meta is eating the world in a way not matter as a Facebook but meta is a level that is derived from the fundamental level so basically you have for example markets real economy and on top of that you have a derivative which is stocks exchange for example and on top of that you have a derivative as a derivatives uh market and his claim is basically for some reason nobody knows why these meta things are hitting the fundamental things and one of the examples he is giving is that New York Stock Exchange for example was purchased by its derivative exchange so basically The Exchange which is derivative of the New York stage had a greater value so they could buy an exchange on top of which they exist in a way so so how would that basically function if we have a value captured let's see where am I this is value captured the larger the Box the greater the value capture these so if you have level 0 which is fundamental level then level one which is derivative of level one meta level and the level two is derivative of level one which is also at the level the claim by Douglas roskov and couple of other authors is basically that level one will capture the most value and for example Facebook in the case of as a this is the case for example for the Marcus derivatives Market will capture the the greatest value and for example in the case of the web tool let's call it protocol is TCP for example Facebook will capture the most value TCP as a practical capture last value and that may be the problem because at some point Facebook will buy TCP so the question is if this is true and we agree and we may not agree then does this apply to web3 and that is in 2016 an article drone the tire the title of the article is fed protocols basically the authors the author claim that ethereum or web3 are basically reversing this kind of behavior and protocols will become fat applications will become thin concerning the value capturing so basically he stated that the market cap of a protocol always grows faster than the combined value of all the application built on top of it so that would look like this so I will have a third protocol which is ethereum and I will help and its value will be this largest than the all apps built on top of it so my personal Excel of question back then like during 2016 is this really true and so far I'm trying to look at the web 3 world via that plans and to try and to find the examples of this but basically I will not find more examples that are counter to this so basically I will find more counter examples and I just created a couple of counter examples that basically does this still does not hold in web3 and this is so for example coinbase which is an app bought bison Trails which is a validator company cells use both mining jump trading which is a trading company called bought certo's one which is a staking or for example arbitrum which is L2 both Prismatic Labs which built ethereum to all so and additionally for example there is a large issue with the foundations like ethereum Foundation having people go away to form startups which are on app so you could be a part of protocol development and you you basically have larger incentive to go to build a startup than to basically be part of of a protocol development so yeah so so far you need to understand that law attract of uh uh uh attractive profits of or basically holds in in the same way for web 3 as for the rest of the world second part of the argument is centralization goes up the stack so if we have some protocol which we have decentralized and that is the centralized protocol number one that enables sorry that enables other startups or organization to build protocols on top of it which may become decentralized with and that is decentralized sorry which enable basically these companies to build on top of it some decentralized protocol and this number two protocols enable the centralized protocols and Tech stack to build another protocol on top of it so you for example Maybe ethereum which enables uni swap which enables another protocol which enables another protocol my basically claim here is when you first of all it only makes sense to decentralize bottom from bottom to up because if you decentralize decentralized if you basically decentralize this protocol under four if everything below is a centralized everything is still decentralized so it only makes sense to decentralize from bottom and my claim is basically when you decentralize for example ethereum suddenly centralization goes here if you decentralize VR unit slope centralization goes up so for example in the beginning there was no uni swap just ethereum and you basically needed to use centralized exchanges to do stuff on ethereum and after a while first axis which were expensive came to be so this is the second part of the argument and finally coming together these two things and that is we need to ask ourselves what centralization decentralization and profit slash value capturing mechanism have in common and my claim is that centralization go in some direction as Province and that may be the problem why we feel that battery is failing so we have three two we have a picture in picture so on the upper side there is this this is one picture image and on the bottom side this is the other so the question I posed myself is basically because profits and decentralization are decoupled so profits can behave in some way and the centralization can behave in some other way so they are not mutually dependent let's call it like that so profits in the this if we know that zero is a fundamental part of Texas Tech one is the the next level of text text so you can imagine for your convenience then zero is for example evm and third the thread box is application on top of it and the the question is how does the centralization uh what is the what is the direction of centralization of centralization and basically my claim is it goes up so as I said but the question is does the profit go up or it could it can go down and basically as I said previously it doesn't make sense for profit to go down because meta is fundamentals and also from the common sense perspective it doesn't doesn't make sense for profits to be on three if zero is not decentralized in any way so so my answer to this is basically for you in short profits and decentralization are moving in the same direction so it's at the beginning for example before ethereum profits would basically value capturing and value creating mechanisms were there and then ethereum started functioning then basically attractive profits for startups disappeared on that level this kind of things started to exist so a level by level decentralization goes up and up and profits goes up and up but the question is which comes first does decentralization first come and then after that profits or first there are profits and then the centralization and my claim is basically the first prophets appear and then after that that that level of tech stack can be decentralized so and they totally had make sense in a way it's also common sense because uh basically the law of attractive conservation of attractive attractive profits says precisely that if you make modularity and commoditization on top of you basically ethereum Tech stack the DVM and stuff like that attractive profits for startup will basically disappear on that level and suddenly a bunch of startups will be attracted by the profits on the next level and they will very basically try to innovate their make profits and decentralized stuff so so the question is who will on that level start building first or create faster a solution will it be when then for example on third level attractive profits came to be will you basically have a startups which will be totally decentralized or semi-decentralized or centralized and we can all agree there's basically centralized solution are more efficient in time so you can build centralized solution more fast so if you can build a centralized solution more fast and you can capture more profits faster than somebody who is building the centralized solution then you will build it or maybe not you but somebody else will be build that solution and why is that important because their new attractive profits appear on a new tax deck new solutions for starts will always be centralized and why is that important for us because at that point users will always use that solution which is centralized and they will feel that web3 is failing because they are always using on the the top level some centralized solution and yeah so but the question is if we have a finite number of these levels so for example we we see that it is a final step of tax Tech then maybe we can hope that someday we will decentralize everything so it's just a matter of time so we need to take five years ten years and everything will be decentralized and yeah we will have a solution and the other possibility is that there is no finite number of these levels and some levels disappear and you alert and then basically we have a problem because on the on the final level which appeared now all solutions will be centralized and why is that important because of the users so so before that this so this may be a subjective thing and maybe just user-filled things are centralized still but one thing we need to know as Builders is users want 100 of the centralization they want not for it to move somewhere else up the stack they are not interested in like I have decentralized text deck number three and I will explain to you why is that important and then they start to use it and because of traffic profits appearing on number four they're using a centralized solution so basically we as a builder need to understand that users want if they want the centralization there is a possibility they don't care but if they want they will not be satisfied if first three levels are decentralized and number four is centralized so they want that and because of law of attractive profits matter anything fundamentals and profits moving in the same direction as the centralization Slash centralization this if it is not infinite if it is not finite world they will always basically suffer so yeah so in short that is my argument so this is insert the argument meta is fundamentals we can discuss if it is true but so far we are counter examples this also holds with for web3 centralization goes up the stack when we decentralize something as a startup which enables other startups to build on top of it these startups will be basically centralized for first part while attractive profits or there may be commoditized and created the centralization after some time after that they will move centralization will go up the stack and because these things move into the same direction and they are not finite set of things that tree is failing yeah so so yeah I was shorter than I wanted to be but that's it so maybe we can discuss do you agree do you have any questions about this please just wait for the mic please hey so um I think it's I think it's very interesting but um if we follow this logic wouldn't it then be true today that all the Roll-Ups on top of ethereum combined value would be quite like larger than the market cap of ethereum which is not the case but if you yeah I understand but I'm not sure do attractive profits appear on on L2 that are larger currently than ethereum I'm not sure and I'm not sure in the facts how much do they have a value of so maybe you can help me with is it like polygon is in top ten and then other ones aren't I I so basically first of all maybe my argument is wrong but uh what I uh what you claim is uh ethereum still has more of a value than everything on top of it combined yes combined yeah so I am not sure does that say anything about ethereum or everything about it currently so maybe there are no uh not enough users or not enough value on top of ethereum for it to hold any let me see so the premise is the real question here is does that anything have to do that kind of reasoning does that anything have to do with fat protocols and team layers in a way let me go here so the claim with fat protocols which would hold which basically are in agree with what you are saying is that the market cap of the protocol always grows faster than the combined value of the applications built on top of it but L2 is just not only what is built on top of ethereum you agree yeah so you just took all just the L2 maybe that is my answer but you need to call take all the applications which are centralized also so all the exchanges and whatever all the wallets all the derivatives all all the D5 you know so I'm not sure if if you take like for example all the D5 on top of ethereum does that premise still holds so let me let me let me then start from a different direction like um you're you're saying that the centralization at the top level is problematic for the end users right or I mean if they want decentralization if we if we want yeah maybe there is a claim that users don't care about exactly like if they care they have a problem with interacting on a upper level with the totally centralized parties sure and if if we assume like today that um gaming is one of the verticals that is like very popular and is um uh transforming into a web3 industry quite quickly so if we have gaming for example and and it is produced the games are produced by centralized game studios right but as long as the underlying protocol that is used to um like exchange and keep and sell and buy certain items in the game is decentralized and allows basically is part of my argument if it is finite then we don't have a problem they don't care about whole stack being decentralized they just care about first three things and that's it if then that means that in decentralization is a finite thing so you need to decentralize concrete concretely like protocol level maybe middleware that's it this we saw the centralization problem then let's move on we don't need battery but okay but maybe maybe your definition of the three is different but like yeah basically any startup which is trying to build some kind of protocol on top of the centralized protocol is basically if they are protocol they will enable centralization on top of it but as long as they are infrastructure they will never have the the huge user base the user base is yeah yeah and that's precisely maybe maybe there is no general purpose decentralization and no infinite sets of the centralization but then we need to maybe rethink uh decentralized supercomputer stuff and stuff like that because they enable infinite sets of odd possibilities and if you enable infinite sets of possibilities you will suffer the problem with infinite number of centralization points sure yeah that that makes sense that totally makes sense but um I agree that if you have a finite number of things you need to decentralized then you may not have a problem because you solve this internally as long as the infrastructure on which the web tree is built is I would not agree totally because take an example of email email was totally decentralized from from the beginning SMTP part 3 IMAP you could spin up your decentralized even server you could communicate with other parties via protocol but fast forward a couple of Dozen Years you have totally centralized entity because of mainly spam I think so when you had imagine imagine billion a billion dollars billion people or zillion people using email using uh uh ethereum for example and I will just walk you through the email example and then we can switch to this How would how would it function ethereum we are spamming that of the tokens and stuff like that so at some point in time for example gmail which was using uh decentralized protocol started having labels which are totally centralized as they were not part of email standards so they were convenient so you couldn't use for example in map with them you could use just the front end of Gmail and after a while basically they want a bunch of users and at some point they have good spamming filters and you could spin up your own email but you needed now to spin up some open source spamming field 13 which was not good enough and at some point they added social stuff and uh promotions and the tabs which were also not part of the decentralized protocol and at some point they started blacklisting for example Microsoft any email which is not uh recognizes top 10 email services so you can still in theory bootstrap your own email service but Microsoft will never get your email my my email sent to your vice versa because they they don't care about the centralized part of the email and it all happened because more or less spam and couple convenient things you could get via email and that is precisely I think the problem of not thinking about the whole stack of decentralization and just focusing on three things because you can decentralize infrastructure and maybe decentralized middleware but suddenly spamming with tokens becomes such a problem that people start using just one wallet for example because they do not want each day to have zillion different spend tokens on their wallet to filter them out and stuff like that and suddenly you get the label on the wallet which you don't have as a standard and suddenly you have a couple of convenient things and then you end up with centralized solution which is totally centralized which is using irrelevant decentralized protocol so that was the reasoning why I thought that talking about the infinite number of possibilities in decentralization is really important now not 10 years from now and by saying it's just enough to have treating salt because it it's gripping up you know it's a continuous thing decentralization evolves all the time yeah yeah and you think you need to think about the whole stack if that is important for your user because it will creep up from somewhere which you never thought about because it is profit centralization point as I said a good example is email is it started totally the centralized now is totally centralized so I I think it's not just enough to think about first three things and say everything else will be good because it will not when we have experience in other decentralization uh tryouts like email is important yeah you have a yeah for example I'm really afraid for example as I said with Spam imagine onboarding now two billion people imagine how much of a spam you will get on your wallet so so somebody would offer some kind of convenience solution and you end up in centralized thing there is a possibility no one cares but if you care yeah and I think that maybe the the fundamentals of web3 are based in open source and I think that is something that will kind of help us um continuously sufficiently decentralized yeah they are but at some point they may not be I have an experience of a zillion different teams of open sourcing by Design not by default so open source in just the stuff they feel that they can open source again a bunch of proprietary software which is not open source which is basically which is okay but let's not talk about open sources being in web default thing it's not it's it's a bunch of to open source something while we're working there but if you know but I I don't know from Stark where to for example we thought we were talking with looping they claimed that I hope that helping will not hear this uh but they basically claim they are a roll-up ZK role for SDK roll-up protocol but they didn't open source the sequencer or how to call it I was like how are you protocol we cannot use you you know so yeah cool thanks thanks thank you okay those were like 15 questions do we have some more hello thank you for the presentation uh can you just switch to closing thoughts yeah I just wanted to ask about the second closing thought can you elaborate a bit on that yeah basically what I said is if you have uh if you don't have that statement that that protocols will run the world in web 3 but you still have the web 2 premise what if your application or if your organization is creating some decentralized part of the stack you are not solving users problem with that you're just moving users problem on the next level so you're not changing the game you're just changing the player because suddenly somebody else has a centralized entity which someday may buy the whole thing yes as derivatives Stock Exchange yes thank you so I think this is it this may be I was thinking about is this a builder's stage lecture and I think it is for any kind of Builder which is thinking about decentralization and providing users with decentralization centralization points May creep up I don't lie to users that use all their problems if you haven't thought about the whole stack maybe that that's okay but on I need you to be aware of this do we have some more questions for him okay thank you um I'm I'm interested what do you think about the example of uniswap so uniswap is kind of a non-profit but it has a token and the team has a lot of their own token so you could say that kind of they they do have they have some economic value from the protocol but they are not charging users for using the protocols so is that kind of like maybe a midpoint between decentralization and capture of profit or uh yeah good question which means I'm not sure so so let me try to rephrase and see did I understand uh you think that by having a token maybe they are not uh part of this reasoning let's call it like that they had a workaround something like that yeah but I'm I'm really not sure because if you are to say that uniswap basically has less of a value than ethereum that is that is true obviously but this is just one app but the question is if uniswap is uh that kind of protocol which enables a bunch of successful startups to build on top of it and at some point there may be a good enough startup which will have more value than a unit swap itself you know or something like that I'm not sure I cannot answer that question maybe I up because I I sat here uniswap and I was not sure to put it because I currently do not know enough about that ecosystem so sorry I cannot answer but having a token one thing I can say is by having a token that doesn't mean that you are not part of this game I think you are part of this game because you have a token because you can have a derivative which can basically follow you so if somebody is smart enough to create some kind of derivative on top of uni swap this is a threat to unison opportunity whatever so I don't think by having a token you are not part of this game I think precisely because of it you are part of this game because you follow the same rules do we have more questions no okay see I told you he's the best speaker of the day not because he's my boss um thank you Malaysia very much thank you hope it was good