Science as a Public Good

Transcript

foreign okay hi hi everyone so i'm philip from the dsi foundation and from desalapse and i want to say a couple of words today about science as a public good and how impact certificates may actually help us to finance science and research more effectively than we're currently doing it and

just as a as a warning what i'm going to talk about are not final solutions so you will see that there's a lot of opportunities but there's also a really huge design space so there's a lot of things that we need to figure out so rather than presenting final solutions i really just want to get a discussion started on this and i want to start by giving a huge shout out to our friends at protocol labs and in particular to juan binet who actually made us start thinking about using impact certificates when we first talked to them about desai and the system that we're building so let's start with uh some basics so just as a brief reminder in economics we actually define a public good as a good that is non-excludable and non-rivalrous and science and research are primary

examples of public goods so right so if you think of something like calculus i can stop you from from using calculus and if i use calculus it doesn't reduce your value of using calculus in any way but that actually brings a problem with it because traditional market mechanisms where you actually want to charge people for the usage of a good don't really work well and we end up providing fewer public goods than what is socially optimal and as a result public goods they often depend on government interventions and are and on altruism to being to be provided so of course this is actually what's going on at the moment so we have mostly government sponsorship and altruistic sponsors and foundations actually

funding a lot of the basic research that is going on but the problem is that the current funding systems are far from optimal right so there's uh there's a large degree of inefficiency and this is actually where impact certificates can come in so the basic idea of an impact certificate is of course that it is much easier to determine what was impactful and what should be rewarded after it has been carried out right so in in retrospect you have a lot more information that enables you to make much smarter decisions about where to put the money and so technically speaking an intel impact certificate is a fungible transferable tokenized certificate that facilitates allocating retrospective rewards

and it supports hierarchies of credit assignment and should be open to various mechanisms and then there's also this idea that as the owner of an impact certificate you can actually burn your claims for financial future rewards in order to get like the ultimate reputation gain right and science is a fantastic application for impact certificates so in contrast to a lot of other public goods that we that we want to finance science actually doesn't have some of the the design problems that we have with other public goods so for example research projects and publications they typically claim a very well-defined contribution space so there's no discussion about

what is actually being claimed and we have a very well established auditing mechanism for basically proving the uniqueness and the quality of science and that is of course peer review and peer review is of course far from being perfect but at least it's an established system that we can build on and that we can improve upon we also have well-established norms regarding authorship claims and credit attribution to funders and and prior work and finally an important issue here is also the question so how do we measure impact and of course in science we have various measures to do that the most important currency is of course you know authorship and and

citations and although the all these quantitative proxies of impact are far from perfect at least we're not starting from zero right so we can build up on these things that already exist okay so looking at the design space the first question is who is actually going to put up the money for that and i would say that we can broadly differentiate between two categories so we have on the one hand we have the retroactive funders so these are going to be people and institutions that are putting money up that will be retroactively distributed to the holders of impact certificates and i would call these people and institutions primarily impact funders because they're going into this not with a profit incentive but

primarily with the incentive to just sponsor science and to get impactful science right and then you have the buyers of the impact certificates and that's not necessarily a completely disjunct group of people or institutions right so it's totally possible that you have some institutions or some funders that are actually acting on both sides of the market but as impact certificate buyers you can also think of normal people like you and i who can actually make a statement that this is a particular piece of research that we care about and we can support it right and you can also think about institutional investors who come in and who buy these impact certificates with a

profit incentive basically trying to to buy them at a lower price than they could ultimately sell them for or be rewarded for so who gets what the impact certificate buyers they obviously get reputation and the potential for financial returns in the future the researchers they actually we can make them participate in the financial returns and like the researchers by themselves we don't need the impact certificate for them to claim credit for the work that they're doing because that is already done through the authorship on the paper or whatever they're um they're doing so that is not really a concern but there is still a very good reason to give them sort of partial co-ownership in the impact certificates and that is

because we want to align the incentives of the scientists and the funders to alleviate principal agent problems right and that actually opens up a lot of possibilities for reducing auditing and you know overheads that comes with with traditional financing mechanisms for uh for science and it also opens up the possibility for scientists to actually earn additional money and some of them they may actually make enough from these impact certificates that they can be independent and that brings up the idea of having a flexible cap table with tradable shares for impact certificates so this would be very similar to a

company where you basically have shares to quantify how much of an impact certificate or a company is being owned by whom and then the owners they can sell their shares and you may also have a mechanism where you do capital raises by just issuing new shares that lead to a delusion of the the ownership of the previous owners and that would actually create the possibility to raise additional capital later and options to opt out for financial rewards for ultimate reputation gains and that could be done very flexibly by by each owner of that impact certificate individually so that's actually a super super powerful mechanism that is very well

proven uh in in markets and it would just be super cool if you could bring something like that to find in science as well and then there is a question so how would you actually have a fair distribution of share ownership among the scientists and the funders right and the basic principle here should probably be that marginal rewards should be equal to marginal contribution and if you take that starting point then you come up with a number of conclusions that you arrive to so for example you would conclude that early stage investors who have a much higher risk by investing into the project early they should get higher equity shares per dollar invested than later stage

investors right and so on and so forth and then there's this big question so how do we actually measure impact and when do we do that so my sense is that there is probably not going to be one approach that is going to be optimal for everything probably we need pluralistic approaches where we experiment with different ways of actually quantifying impact and ranking the impact of different projects but besides from algorithmic approaches there's also an alternative approach where you say well we should just leave it up to the preferences of the funders and it should be them who ultimately decide what they think was impactful and important right so you can think of uh of a funding mechanism that uses

quadratic retroactive funding here so what are the incentives so once you have the impact certificates and you have a large pre-committed fund for retroactive rewards you actually create incentives for other people to provide grants and prices and other forms of investments into future research and the impact funders those that actually put up the retroactive rewards they have a pretty big incentive to use this mechanism because they ultimately will get a much bigger bang for the buck right so they can actually spend their money much more efficiently and that's because they can actually have co-investments from other parties

so through crowdfunding they sort of also have this guarantee that a part of their funds at least will be allocated very efficiently based on information about what actually worked and because we can actually solve this principle agent problem to some extent it reduces the need for overhead and monitoring and finally because everything is going to be on the blockchain we have this this beautiful and open data about who did what and how we can actually have very detailed information about the impact of these investments which is what these funders ultimately care about so the investors in the in the impact certificates of course they would also get reputation and potential financial

rewards and for the scientists they would get new sources of funding for their research and again reputation and actually this is also quite important rewards for for doing open high impact science so now i'm just going to give you two examples of how this may work in practice so the first one is what we call open science certificates so the idea of an osc is not that you finance the underlying research but you rather finance the process of making the results of the research openly and publicly available and for evaluating the quality of it so basically an osc verifies that a scientist has created a valuable fully open scientific contribution

so the purchase price of an osc could cover things like the costs of peer review so rewarding peer reviewers for the work that they're doing it could also cover storage and gas fees and it may also contribute to an open source development pool that could be used for financing the the ecosystem and this is actually a pretty cool mechanism how we can bootstrap a truly open science economy and these open science certificates they would actually be really valuable to researchers because scientists from around the globe they're actually being put under quite a lot of pressure to publish their work in open science journals and also so that there is a lot of pressure like

from from funders from the public but also from from their employers for very very good reasons and by having this mechanism we could actually if uh like we could go a far step beyond just putting your work in an open access journal and paying a publication fee you would literally not need a publication fee we would make everything open from the beginning so what is the process so people would just go on our platform and they would create what we call a design note so that's a research object where you can store the manuscript the data the code where you can also have videos annotations all different other sorts of supporting information you make it fully publicly

available and it's going to be stored decentrally on ipfs and filecoin and then a qualified community of independent scientists they conduct in-depth peer review and they reward their referees for the work that they're doing and in our system this would be it done in what we call arcs or autonomous research communities as a third step the original nodes that meet open sign standards they would then receive an open sign certificate and then investors sponsors or collectors they can buy these things and those of the owners of the oscs that are potentially not interested in financial rewards they can opt out of them and become permanent named sponsors

of that orc and then as a final step you would have an oracle of retroactive funders that reward that reward the owners of the open sign certificates and we could organize this for example through a dow with quadratic funding and as a second example we could actually also combine the classic funding instrument of grants with impact certificates so this example here would be a lot more difficult to implement it's much more complex but it also has a lot of advantages it basically takes a lot of the boxes that that funders and scientists have in mind it would create a lot of opportunities for

basically closing the liquidity gap and for giving the the impact funders the opportunity to also claim full causal credit for the work that they have funded even though they've been using a retractive funding mechanism so how would this work so first thing is a funding pool is open so this could be general purpose or it could be dedicated to a particular research topic or field and then you open the the pool and you make crowdfunding possible and then a part of the funding pool would get reserved for retroactive funding and the other part would actually be given out in the form of grants to scientists and to teams that are applying for this and here at this third point the the funders can actually decide if

they want to decide which which proposals they want to sponsor or they can completely outsource that entire decision making process to an arc that would review the proposals and allocate the funds and at this point we would have a lot of opportunities to experiment with more efficient ways to to allocate grant grants right so one thing you could do here is rather than what we're currently having in our current system where you get typically give out relatively large grants to a small number of people so like grand funding rates are usually like around 10 of the applications which is really really low so instead what you could do is you could give out a lot of microgrants to a lot to basically allow a lot of different

groups to start experimenting with their ideas and then additional funders can come in and they can invest into the projects by buying shares in these impact certificates on the marketplace and the most promising projects at that point they would get additional funds and initial investments of the impact funders would be leveraged and then finally an oracle would reward later stage buyers of the impact certificate share owners so the good investors they would make a profit and the impact investors they could again they could opt out of the financial rewards and become permanent name sponsors of of that research so these are some starting points some some ideas that we're currently turning around with this is all work in progress we'd love

you guys to join us this is our qr code here for the discord so hop on the discord join our discussions and help us build this thanks [Applause] is there any questions so thanks for the talk i must say that this site is interesting as a whole and in particular what interests me is the new metric you're introducing to this world coming from [Music] some years in the academia i see that the main metric for value of a paper or research is probably

the age index of the author or either number of citations and right now you're introducing this ic or oscs that have some value in the d5 market and can you try to explain how does this not compromise the true value that was defined before d psi right so i mean the big question is what is the true value right so how do you determine that and i think this is really the big question that we have to resolve uh so the current metrics that that people are using such as you know in which journal have you published so what's the impact factor of that journal how many citations does your paper get these are all like rough proxies for how important that idea was relative to

other ideas in the same fields but these metrics they're problematic because for example they do not take into account whether the work that was published is actually reproducible and replicable right so one thing we could actually think about is if we want to stick with the traditional metrics and if we would take them as you know potential indicators of true value we could combine them with markers of replicability and and uh and reproducibility which we can actually also generate in these nodes and have them uh basically verified through a peer review in our system so that is one approach and the other the other thing is also like whether so what the ultimate value of research

is is probably not an objective criterion to begin with right so i mean if you're if you're suffering from cancer then any research that is related to curing your disease is much more important to you than any other research right so there's also something to be said here that uh you know it should be ultimately the funders to decide what uh what is actually impactful and meaningful to them so i would see that as you know as complementary approaches and you know i'm arguing for you know various for for plurality of different approaches to measure impact yeah thanks a lot philip um so martin

kodar is my name i'm working on a project called fleming protocol we we're um involved at the kind of patient side of this where we want to pull biomedical innovation from the the academic domain and they're working we see the patient groups as possible funders for these types of projects but how do you see this type of a situation integrating with the existing intellectual property uh situation so the traditional patent situation and and the kind of trade secret side of things as well yeah so so i think there is there is an obvious conflict between doing science totally openly

and doing it under an ip regime where you know one particular entity actually owns the property and can license it out and basically has a temporary monopoly right so in a way um so this this entire ip mechanism is sort of like it's it's a broken crutch to solve the public goods you know problem that we're having because instead of you know actually solving the problem it introduces a new one which is that yes you you basically give an incentive for people to to develop these things but you're doing it at the cost of creating a temporary monopoly that is going to be awfully exploited right so i think there's obviously a conflict

between these two things however so i completely understand that you know ips will remain to be an important mechanism so especially in pharma and so on and so forth so i don't think that it's going to go away but what could happen is that you sort of have agents or companies that are starting to use these impact certificates to sort of incentivize basic research that they think is necessary to be conducted before they can actually do research at the ip stage so they can use that for for basically outsourcing part of their research that they don't think is necessarily applicable to ip exactly

any other questions no thank you thank you you