29. From Funding to Cost - How Movements Like DeSci are Working to Fix Science - Taipei 2023

[Music] all right thank you David um it's awesome to be here at funding the commons and uh see so much uh reception for decentralized Science and Science in general um today we would love to you know we have esteemed panelists here flying in from all across the world um and we have an opportunity and we thought it'd be great to uh share a little bit about uh what everyone's doing in D dii uh but before that what is even dii because I I I'm I'm sure that not everyone that's uh here today uh knows what that is so maybe we could uh quickly kind of go around and just tell a very very short what what you think thesei is um starting with julani julani last time I just gonna say when you defined it in Montreal two weeks ago I was like maybe I'll just let you define it uh so first of all I'm happy to be here thank you to funding in the Commons and everybody here um my I think dii has just like any other Nan web3 has its own definition everybody has their own interpretation for me dii is the opportunity to open up access to science make it more democratic um and de abstract it so bring people the opportunity to participate in science in a way that they haven't done so before not just for formally trained scientists like ourselves here but also for people who are scientific curious or just On The Fringe of science giving them everybody the opportunity to participate not bad uh so yeah for me uh open it kind of Rifts off of the open science movement that has been going on for a very long time at this point in time uh however operationally implementing the this the uh the philosophy has been very difficult and um so as of right now we're kind of suffering the science in general and the academy has been suffering from the over NE liberalization of um yeah the academy has been over neoliberalized so basically you have big uh universities and you have big publishing basically um uh becoming big business and they have monetized Prestige uh and the idea is it's been producing a lot of uh poor incentive structures and the idea is how can we use this technology now to uh red decentralize science cool uh hi everybody so this is I think like an open question for a while and still is uh I think that that um there have been two emerging definitions or categories of definitions for dii one is the larger picture egalitarian definition which is uh includes Open Access includes access uh Community tools it includes bringing in people that have typically been excluded from the conversation such as rare patient disease groups uh individuals that are perhaps performing citizen science um Etc and then the other side is the technical definition or technological definition which is a definition gear towards AI or autonomous agents so to back that up we collected some data a couple years ago and would love to collaborate with anyone who wants to continue collaborating on this uh building up the data set uh there was four four main uh issues or ideas that people identified as getting into dii one was uh access to global community and collaboration two was access to data sets and three the computational resources to uh to to deal with them and then lastly and I think most importantly is uh compensation and fair attribution for the work that you've performed so this was just for context this was uh fielded with those that have that are currently in Academia and had heard about dii and I asked them to complete the survey as they were exploring exploring so I bet you this has changed since then yeah uh thank thanks everybody uh so this is the only thing that I'm going to ask for everyone's opinion on because essentially I feel like it's one of those things that I I feel is is important to help Define is that there is actually a lot of different definitions of what but at the root of it is trying to address the issues that we're seeing in science and how do we make sure that the you know discoveries that we have in science can get distributed into the population how do we encourage those discoveries how do we fun those things so those are all parts of the equation that uh we we're you know here to working on so I think maybe it makes sense um for each person kind of to go through the project you're working on and the problem space because essentially there are way too many there's no point where we have 30 minutes there's no way we're going to even get through you know even part of the problems in the D side but uh in science but maybe we can just go from some of the main ones and everyone can bring up what they're working on and then what problem you're trying solving and maybe a little bit of context about it but again we're it's it's complicated so we'll try to keep it as as high level as possible yeah I I think the panelists here kind of fit each of the individual pillars or a good spectrum of each of the individual pillars dii world is sits really at a social layer so our whole goal is to help broaden the access points to come into decentralized science so I wish we had the graph of the decentralized science landscape um which is a slightly different graph but there are hundreds of project is building a variety of different tools that all address different aspects of scientific problems and so our goal is to kind of create that information Matrix that framework that allows people to come in if you're interested in D if you're interested in changing certain paradigms in science to learn about what projects and communities are already doing so and then how best to engage with them to then help further that mission statement alongside that we're also starting to build out technical tooling that helps to create that kind of fertile substrate for composability between these toolings so if you're developing a compute framework that is more Open Access you need to be in contact with or you want to be in contact with communities that are interested in funding that kind of aspect or publishing on that or leveraging identity Solutions well we can create this this composable framework where everything kind of comes together so you have an amalgamation of effort and hive mind to help push the needle more forward so that's what we do uh so science publishing Dow is uh like I had mentioned trying to redc centralize um uh or help with that that effort and the idea is you know you have these it's this uh community that has existed and has established certain methodologies for about you know 400 years or so um which is basically this very naturally decentralized uh knowledge generation and also dissemination process uh so the idea is uh but at the same time at the moment uh like I'd mentioned before the uh over neoliberalization of the economy has um made some pretty per perverse incentives within the ecosystem where now uh publishing is more easily manipulated and maybe people are more incentivized to lie as opposed to tell the truth so um the question is how do you actually recreate um uh the the circumstances in which uh the science scientists can uh self-manage their own um reputation infrastructures for example so that we no longer have to be as dependent on the current 23 billion uh industry that manages the uh the uh the that reputation from structure thanks Jonathan and Shady uh shy it's sh um yeah that's awesome I feel like D World kind of fits in as the NSF or you know this Foundation model we is like this aggregation layer for funders for uh experts and and so on and so forth uh and then science publishing down reminds me a lot of the like The Meta professional uh societies so that aggregation reputation aggregation layer uh as for me I uh started off dii basically getting into dii um with like hey I just I got my PhD I just want to do my research in the cloud like I don't want to go to a postto in the middle of nowhere and and drive an MRI truck in the ru I I mean it sounds fun but I would like to connect with other people as I do my research um and not be isolated as I typically have been so you know at the beginning um one of the things I was really interested in is distributed mentorship networks I think that's one of the most under tapped areas just in science in general because how far you get really depends on your Mentor your mentorship network and who you can work with closely uh a little bit after that I really uh got into a problem that was close to me in my own research which is uh sharing data and access to computational resources and we built out a tool called the open science Commons or opsi Commons uh with some help from file coin Foundation who helped fund that work uh as part of building that we realized there was a very outstanding problem in any sort of application that would be built on a web 3 infrastructure so why web 3 it's permissionless it's open access and it's a great base layer for knowledge as a computational Ed point maybe we can get into this a little bit later so at that point we started doubling down on identity and reputation systems focusing specifically on identity verification uh so that we could prove that those who were uploading data or accessing resources actually were verified and had the proper credentials and it wasn't say uh a bot posing as a researcher to get Monero compute time right which is a huge problem that we saw during covid people taking advantage of GitHub actions and web hooks to mine cryptocurrencies which actually kind of excluded more opens science practitioners uh from from uh from from the dze space uh in association with blockchain uh but that's a little bit so the project is called holim uh it's a zero knowledge identity uh and data custody protocol and could definitely talk more about that yeah uh yeah um so there's a we put up a graphic on the side and essentially uh to highlight essentially uh bunch of the axes of where the problems that we see uh they're all complicated problems they they stem the entire uh field of research uh into the funding into the way it's done into the entire you know system and uh I I would like to go into maybe a more tangible example because I I don't maybe not everyone is as familiar with science what does it actually look like uh you know to you know to use your product in a way of of where like eventually where it's going to be what how will the world look or how would science be improved and obviously there's going to be a lot of different points here but maybe the most important Point uh uh of how it would be different than it is right now what if you know when you achieve your goal of you know whatever your organization and Company's working on uh jalani first so I think for us it is the the ultimate end goal or the the the the most mature form of the technical product that we're building is what's going to allow anybody to build science and collaborate uh as easily as possible and also do this by leveraging tools like holim and like science publishing Dow to kind of push everything forward so within our framework and I think in the graph we have examples of AI we Implement AI as these non-biased entities to then score um ratings on on on on projects or on on on data being generated so I'm an individual I'm pulling together information on neuroimmunology and I'm posting this in an open- Source way such that anybody else can build on top of that now with that composability layer what also is an in Incorporated with there and what web 3 brings is the ability to track the provence of the data and everything that derives from that so if you for example build on the data that I'm doing there's a chain of custody that exists such that every new piece of information that builds off of that is attributed to that chain of custody going through and so when you think about that and you devise incentivization layers that can exist on top of that whereas currently in this current scientific p in certain scientific framework only the endpoint scientist gets rewarded if that for the contribution in this system and what web 3 allows is that trickle down effect where funding can distribute itself across the entire network of participants that have participated in the generation of this data and that's something that we haven't seen before and the intricacies and the little the nice Parts about that is that you start to be able to incentivize or to monetize small contributions of science where you don't need a massive amounts of individual you don't need one individual to generate Mass amounts of fund you can do this by into small people providing small bits of data that culminate to make a whole now and I think that makes science more applicable more engaging more distributed more accessible um than its current iteration where it's oftentimes held captive by either funding agencies that get to distribute a lot of sums of money to one individual um whereas the rest of the people kind of get excluded so to us that's that's kind of our end goal of what we see or we hope to see with science making it much more open um so part of that is uh that you mentioned with reputation uh goes back into why does the reputation even matter it's because reputation is one of the key pieces that you use to get funding or use to exchange for resources right so without uh good ways to attribute all the steps of the process for example right now in science we don't very uh really care about anyone that's replicating studies even though we should because replication is very important but as a professor if you're doing that no one's going to care right no one wants you to do that it's a waste and even if you're a student doing it you you don't really get any that much out of it so things like this giving us a a building blocks I think is really important um Jonathan to you in the on the publishing side uh you know Publishers with all their control and as well as kind of facilitating the system that is basically blocking anyone for Access with behind a pay wall right which is the opposite of what we want with science um how do you see you know you know from the founding of of science uh publishing Dow where where could you see this going what would be the biggest impact that you can see so I guess uh closer uh it's funny because I had mentioned uh that uh the current uh publishing Market is about $23 billion and you know I've talked to a number of uh funders about like my project and such and they're asking about you know what part of the market you're trying to capture Etc and I'm I'm and I basically have to tell that I'm actually trying to deflate the industry by about 1,000x um so that instead of having to you know pay to publish effectively um and to pay to read um which is roughly maybe uh anywhere between $2,000 to 11,000 to publish um and it's also maybe around $50 uh 35 to $50 to just read an article um the idea is to just reduce that entire cost structure down um by by a whole ton so I mean I guess in terms of more specific um ways in which you would interact with the the system though it is basically there are a lot of Concepts that we've been talking about uh over the conference that we're basically trying to integrate together um and a lot of it is just basically about uh attribution and citation structures using a uh our own reputation infrastructure um so that for example um you can have an entire Community that's building all sorts of basically potential energy um attributing uh information but then once for example uh you uh the work eventually reaches the level of interfacing with with um with industry that money and that those citation structures uh basically provide the means for uh the money and the reputation to basically trickle down into the ecosystem and the hope is that it trickles deep into the ecosystem so that uh the uh the basic Sciences actually have their own ways of being funded um not just taking directions from either industry or from the government uh so there's you know this combination of you know uh this is another way that we've been of approaching the retroactive funding problem as well as having each paper be its own you know you know non-tradable nft for example so you can't trade your physics you know reputation for for you know dance tokens um uh the uh and then you know there are these ideas of seasonal governance that we're also trying to implement so the idea is how do you kind of create a healthy knowledge decentralized knowledge form and the idea is how do you actually oscillate between setting incentives for review and then for Innovation and then the idea is if you have for example uh Innovation to the degree where you start to have too much uh negative uh review and negative replications then the incentive structures automatically shift towards incentivizing review in a predator pre model so there are a lot of ideas that we're trying to play with there to as an experience of what it would be like within the the Dow cool cool um and and and shy uh with uh the concept of of ZK uh I don't think is as um well known in d uh in D specifically right now so maybe you can I mean in general it's new so maybe you can talk us through how you're thinking about that and the work that you've done in the d space yeah right on I'd love to speak a little bit about D I did want to make one quick injection uh just to continue because I felt like we had a really good uh one after the other uh so I just wanted to bring us back to Juan benet's first talk uh at funding The Commons New York City when he talked about like why should we care about this so you know let's take a look at the ethereum marketz uh market capitalization uh back at the peak of the bull market when the total crypto market cap was around you know $1 trillion dollars are approaching it right um I think at that time ethereum had somewhere like around like a 250 billion correct me if I'm wrong if anyone knows the exact number uh that is comparable to the total amount of money put into research and development by the US government right so when we're talking about funding and what opportunities exist if we could just route some small percentage of the value that's uh exchanged for a pure Financial protocol that can make an a massive economic impact on basic science research that drives the rest of the economy forward in a virtuous positive feedback loop so there's a lot there to to unpack and I hope we'll get into it um but as for the role of of ZK and why that's something that I that I've focused on um all right so yeah so when we talk about science we often talk about uh open science we often talk about data being accessible readable something that you can access Etc however a lot of that open science requires uh a data manufacturing process that often begins under some siloed secret you know uh conditions right or some uh privileged uh conditions for medical data that might include uh uh sorry excuse me um personally identifying information uh for let's say physics data you might be involved in a research project where you're double or triple blinded because you want to make sure sure that the results that you Converge on aren't biased by some group think or some other team member influencing you know influencing how you're thinking um and so there's tons of examples where uh privacy secrecy and role-based condition um role-based access conditions are really really important so that's on the science procedure side on the identity and reputation side which is distinct and different there definitely is a huge opport Unity here to take Concepts behind reputation identity personas and Abstract them away into computational Primitives and what do I mean by that so currently in Academia your identity and reputation is your currency and that's counted in the number of papers the number of Grants the number of peer reviews etc etc so and that's a very kind of abstract ethereal sort of thing like if you try to put your finger on it each study section for each NSF Grant will will compute your reputation a little bit differently so we have an opportunity here to create oracles that allow us to take um data as endpoints say in a distributed science graph like with science publishing Dow um and to compute relative metrics that make sense for the context whether whether it's retroactive public goods funding whether it's pairing uh researchers together to work on projects across interdisciplinary uh fields so that's one side of it the other side of it is it's really important to give researchers control of their identity as a researcher I was guilty of projecting my identity as that of the institution's identity right everything that happened everything that I did everything in my career was related to which institution I was at I'm you know I'm a Harvard postdoc I'm a Columbia University you know tenured faculty Etc so bringing that sense of ownership back to academic researchers and empowering them is the best thing we could do to bring them closer to entrepreneurship where they can have direct owned impacts on the work that they're doing right and I think it totally changes the game to do that we need pseudonymity or the option to opt into pseudonymity in these digital systems and definitely more to say on that as well yeah thank you for that um that uh so actually that brings up another point which is uh currently it seems to be really hard to monetize anything uh of your science and and basically for the majority vast majority of scientists it's almost impossible to monetize uh even though you could be doing stuff that is very very valuable and you are doing stuff that's very very valuable as we can see from you know basically the Publications and the citations and and everything else um and like one of the current methods that we have right now is oh yeah start a company and then take it IPO yeah that's a that's a great idea to make a scientist do that to get their funding right so um there's a lot uh there I feel like that if there's more in the reputation side and more maybe there's more there than the like measured collaboration could that also change the way that resource distribution or uh you know allocation can be done I I think there is I'm I'm always still trying to figure out what is the process of doing that um and getting into that point so I I feel like this is something that I uh I've been thinking about a lot and it's one of like it's really uh yeah it's been a it's been a challenge to wrap my head around just how many of issues there are but um of every web 3 application I've found that you know uh science has one of the most direct like entry uh bar uh like direct uh entry point because so many things have this reputation along with it that's really really important and already is trying to be measured like it's it's something that's trying to do that already so I I felt there's a lot here that can be done and also maybe it's uh a little bit easier to get started because the group I know technically it's a it's a it's a smaller Market there's less people so I feel like there's a that more more things can can happen there um so we have five minutes left I would love to kind of go through just a little bit more about you know uh maybe one or two uh one or two minutes per person just a takeaway or something like that you're excited about in this space that you're working on um you know that's most promising just you know currently what from from what you're seeing right now uh galani so I'll take it I'll take it from the the the bird's eye view of the space um to your actual point Chris I think the everything that you hear about in web 3 about its ability to shift the Creator economy um it's oftentimes applied to like the artistic Renditions to the entertainment and all that kind of industry but it's it holds true to the science a scientist is just an artist that is painting the laws of the universe we're all investigating these kind of things and we create assets that often times are captured by political or institutional actors that is not within our our framework that we can't we can't leverage those things and I think the ultimate goal of what dii is is to and and as much as it is to bring more people into the space and give them opportunities to create the kind of science that they want to create do the kind of science that they want to do so you don't have to Pivot necessarily what it is also going to do is going to allow the creators of set economy or of those assets to govern and own those assets and be able to use them however they see fit and that could be through the lens of more better frames of collaboration better ways to monetize such that you're not captured by those entities that really are controlling how what science gets push forward and this happens at the level of the scientists but this also happens at the level of and we're talking because we all are scientists here but this also happens at the level of patient populations those endpoint users that are stakeholders in scientific development in therapies that come out of that them having a say them being part of the governance of particular therapies or assets or what gets to get developed and what gets to get funded I think that's what the moonshot goal of dii is and there's so many permutations of that and we don't have the opportunity to talk about it but I highly encourage anybody who do who does have questions to then come up and ask us these questions for like down the road so that's my 2 million in view yeah I think uh with dii it's it's just an interesting space and we're you know at this conference funding the commons and uh it uh it's just it's a pretty inherently challenging space because our the product of of science is effectively intellectual soil and it's like oh what are you doing I'm making mulch I don't know like it's I'm making fertilizer I'm making you know it's it's you don't use uh you can't tightly get an Roi on every little patch of soil right and it's not even it's kind of a a moot point but the idea is if you have unhealthy soil nothing else can grow um so the idea continues to be how do you kind of uh allow for this kind of funding and not H having to immediately measure even impact for example because you have no idea what the potential impact of any sort of research can can potentially hold somewhere down the line um and on you know the other front it's uh I think the the D Community does uh uh everyone gets along really well but we also need a little bit more uh cohesion in terms of actually being able to pull together our uh knowledge and what we're finding because it's a it's a big Hydra of a problem as you can see and um none of us are going to be able to do it alone uh and uh unless we actually come together and potentially as a professional society and test a lot of the infrastructures that we've been building that we're um uh predicting is going to work uh the scientific larger scientific Community are isn't probably going to take us very seriously right on yeah that's a good one I think uh really briefly what I'm super excited about is the maturation of lot a lot of these web 3 Technologies I mean right now gitcoin is running uh a grant on arbitrum with the American Cancer Society that's incredible you know the fact that you're raising funding for cancer research and patient outcomes you know through a web 3 stack that shows that there's value there that we're solving something important I think you know there's everything from the retroactive public goods funding round to the maturation of tech like IPL D and ipfs you know through D vertical labs and D nodes so you know once these things really are reaching and have reached uh maturation it's really right pickings for an engineer scientist to come in and begin solving commensurable problems in their space so we're starting to see that now I think uh with different groups and everything from for example uh intellectual property rights pooling with for example weeda uh which is which is uh uh architecture for how do you share the upside of Ip and a copy left uh infrastructure a copy left Paradigm uh we're seeing that for example in uh local context which is an an indigenous uh a research organization or I guess you could call them like a curator research data curator that attempts to establish Providence between indigenous people that perhaps produce data for anthropological research or for geolog GE traffic research uh and allowing them to basically claim rights on how people use that instead of you know to combat appropriation and uh exploitation of those indigenous communities uh and that's just like a few examples right there's there's I think many more that are on their way all right uh I think that's it for our time so thanks all our panelists for joining us today and uh thanks our audience for uh uh joining us thanks for having us thank you thank you for staying a round of the pause [Music]