The limits of decentralization in knowledge sharing

Transcript

[Music] foreign [Music] Audible right welcome everyone so we all know some wannabe Trailblazer who thinks that putting everything onto the blockchain will solve all the world's problems we've probably met quite a few people at this event these past few days but does such a strategy always work my contention here is no it actually doesn't and I'll be using the example of knowledge sharing processes let's cut through the buzzwords first

because you know when it comes to blockchain you can slap a token or smart contract on everything and call it decentralized I I use the definition of an instance when the production curation and management of knowledge is achieved using crypto economic incentive mechanisms without relying on a central coordinator or in words that would make more sense to an HCC crowd I mean staking and slashing mechanisms which I would argue is the only reliable way to ensure coordination at least above the base protocol level so in case there are any Bitcoin Maxes um you know I feel sorry for you but I'm not necessarily arguing against you right now um so what exactly do I mean by

knowledge in this instance because here I would distinguish Knowledge from information or data data is what I would call Raw unprocessed facts or details without context you know think of just numbers 120 110 100. information is when you add context to that data or structure in a way to make it meaningful or useful such as associating those numbers with stock names to denote stock prices knowledge on the other hand is when you apply that information for instance you're drawing conclusions on what to do with your stock portfolio once you get those weekly reports on your stock prices so a knowledge sharing process would be slightly different from

a data sharing or an information sharing process let me just draw this caveat there is a definitely overlapped among these examples and they're not mutually exclusive but it works as a heuristic I would refer to a public use data set as a data sharing process a curated list or even a library as an information sharing process and a scientific publishing journals encyclopedias or communities of practice as knowledge sharing processes and because of time constraints I will be focusing on my discussion today on encyclopedias and by encyclopedia I mean the only encyclopedia we've ever used since 2004. now let's have a show of hands how many of you have used

Wikipedia for serious academic research even though you knew you were not supposed to how many of you haven't I I guess you guys are liars but whatever it doesn't matter right there there's three reasons for this um because uh for Wikipedia success in this regards it's because of its one it's strong Community Spirit two its lack of Entry barriers and three it's emphasis on the collective ownership of content however despite all its successes being a top 10 website being that's primarily volunteer run and funded from donations Wikipedia has had its fair share of detractors over the

years and this criticism comes from three largely interrelated reasons firstly there's a really complex hierarchy of Journeys in Wikipedia you know there's users who are called editors there's users that are called admins oversights arbitrators another problem is it's ruled by one percent that you know in fact most um online communities where one percent of the editors in Wikipedia contribute 80 of the content and thirdly Wikipedia is accused of harboring editorial biases it's not too simple like Japanese Wikipedia is accused of being way too right-wing English Wikipedia is accused of being way too left wing so in

response to these there were some Geniuses who decided that we are going to bring decentralization to Wikipedia and solve all these problems and decentralize not only its governance but also its core um as core process of disbursing knowledge to the masses enter every pedia now they sought to bring staking and slashing to the editorial process of Wikipedia itself the way this works is you take IQ tokens you lock them up for 21 days you get something known as brain power points now you as an IQ editor you make a new page and then it's voted on by other people who have staked that token now

there's a challenge period and then if it's challenged um the other stakers vote on whether to accept your edit or not if you are successful you get curation rewards the stakers who voted to accept your article also get curation rewards these come from token inflation proposed platform fees now if your edit is rejected or if you were a voter in the staking pool who was in the minority you get slashed what I found interesting about their slashing model is the slash tokens are not burned or redistributed it's just that the lockup period is extended by how contentious the vote was for instance if it was for a 4951 split your slash would be you know not too long it was a 9010

split then your your slash would it be I don't know in the course of weeks or months it sounds pretty straightforward and on the service level it looks like a genius idea a way of replacing Wikipedia the question is do you think it actually worked it did not now what happened with everypedia the majority of submissions ended up being copy paste jobs of Wikipedia ultimately every period deployed a bot that would monitor Wikipedia for changes and scrape it and update it accordingly any new Breaking News topics were full of fake news misinformation original articles were mainly self-promotional materials and coin spam it's a so-called less

sensorial policy just resulted in a bunch of non-notable Articles and defamatory content ultimately the progress of every pedia ground to a halt most IQ token holders were more interested in speculating with the token rather than actually using it and now every pedia is being rebranded as IQ Wiki I'm not exactly sure what they're doing last time I checked them they've now decided to jump on the AI hype cycle honestly I Harbor no ill will towards them I wish them all the best but what exactly can we learn here what went wrong with everypedia it was the lack of community spirit that made Wikipedia so successful every PDF chose the route of financialization of all its

Community interactions the well-being of the individual contributor was given explicit Center Stage through the celebration of self-interest and as you can see by the IQ token price I you know that didn't really work out well although I'm not much of a price discussion person especially for reasons I'd rather not go into now I'm not going to be pretend to be an evolutionary Anthropologist here but based on the first page of Google which I searched for five minutes there seems to be research backing up what I'm saying here right a desire to volunteer and a strong Community Spirit they are related linked in a virtuous circle with each reinforcing the other and secondly one of the primary roles of

money is to facilitate temporary cooperation between strangers who will never meet again but if that relationship is is by definition temporary you can't really emphasize that Community ownership and the community identity if it's all based on money and thirdly the one of the theories is that money makes cooperation possible when you cannot rely on reputation and kinship alone so how do Wikipedia and every pedia square off with each other well Wikipedia effectively depersonalized the editing process by emphasizing Community ownership of editorial output you open a Wikipedia page there are no authors there are no attributions you can maybe go into the talk page in the background and you'll see arguments but they're not

never publicly attributed to the face of you this policy together with its low transaction costs to participation it fostered collaboration editors were encouraged to contribute even if they thought they were not experts or could not put in much effort okay maybe I'll change this here I'm not entirely sure this is true or not but I'll make that edit anyway maybe someone later is going to change it every pedia went the other way they explicitly Associated edits with contributors and directly financialized it this introduced an element of risk and transaction cost to participation and ultimately risk aversion deterred potential contributors from participating in that process because and some of the psychology literature backs it up the average rich

person tends to be more risk adverse than perfectly rational now risk aversion is something that you specifically don't want when it comes to sharing knowledge as opposed to sharing information or data because it it forgets about the importance of dissent staking and slashing crypto economic mechanisms they're perfect at emphasizing coordination not necessarily cooperation and those two are slightly different things because when you have enforced coordination there's no room for good faith disagreement from the majority uniformity and forced uniformity is useful when you want strict compliance with Norms or values such as in accounting practices or

building construction codes but where you want to safeguard diverse perspectives and open debates you might just end up stifling intellectual diversity and the expression of alternative viewpoints when you directly penalize dissent you promote Conformity group think and you kill heterogeneity and you know let's just look at it historically what is one of the reasons that we have made such strides in science since the 18th century or so one of the reasons is because we start burning scientists at the stake for saying things we don't like right so ultimately having direct uh direct penalties whether Financial or otherwise it does deter people from speaking their

mind so what did this culminate into right Quantified Financial gains were the only tangible benefit for every pedia editors and then on top of that there was a risk of penalty for not appeasing the majority ultimately nobody wanted to rock the boat a natural shelling Point ended up being formed because you know with crypto economic incentives and all we're all about shelling points what was the shelling point for every pedia it was Wikipedia now if you're trying to kill something as you advertise yourself but that ends up being yours the sole reference for your existence you're just a failed project all around so why am I telling you all of this well

part of the reason is nothing gives me more morbid Delight than walking into a crypto conference look in crypto Bros in their eyes and telling them that some of the things they've done is downright stupid but beyond that there's another message I want to emphasize here sensible system design entails knowing the fundamental culture and interactions a community organization or process requires if you're sitting here and you're thinking about all your fancy Game Theory variables all your cool smart contracts first of all sit back and think about what is that user on the other side behind that keyboard actually doing it and most importantly why is he doing what he's doing because I'm assuming all

these systems and mechanisms You're Building are for humans and they're not for rational self-interested autonomous agents or however you like to model them in the white papers ultimately you have to cater to the human element at play which which requires you to think about people as people not as autonomous agents the pursuit of Quantified tokenized incentives to coordinate human behavior works sometimes in other cases it compromises fluid and natural interactions we have to understand that enforced coordination is not always ideal there are situations where you want to facilitate descent or allow people to be wrong because you never know maybe that person being wrong might

end up being right later and it's it's just not you know it's just it's not efficient if you say that ultimately he might be able to convince someone we don't know how but he will let's keep these crypto economic designs because there's a bunch of buzzwords the blockchain hype emphasizes stuff such as transparency immutability trustlessness but people don't necessarily value these Concepts in their day-to-day lives just look at all the that's going on with the arkhams docs to earn sort of model that they've just released right people are thinking hey why are you doing this they're saying oh but it's all about the blockchain space no it's not it's about humans and what humans want we and you know we're not autonomous agents so I guess my message

here for the crypto crowd is stop being a bunch of goddamn artists go out touch grass talk to people thank you very much [Applause] if you want to trust me go ahead great I think I'm perfectly right here wonderful [Applause]