DeSci, Independent Labs, & Large Scale Data Science

Transcript

all right hello everyone my name is juan i'm here to talk about desai i want to talk about how we can use seaside to fund organize and open science uh first off let this is what we're going to be talking through i'm going to talk about science in general for a moment then i'm going to talk about what the design movement is then about how we can fund the science commons then how desai is organizing the people the projects and the works around science then i want to speak a little bit about open access and reproducible science and i want to finish with a call to action so this will be pretty this will be pretty quick we'll have all to cover so i'll move up fast

first off i want to start by saying that there's been an enormous amount of progress in the last few centuries almost every humanity has been improving by almost any measure that we can we can think of the human condition has been improving dramatically and a big part of achieving this progress has been the scientific enterprise so by extending what we know by being able to trans uh transform what we know into technologies and solutions to various problems we have been able to lift a huge fraction of the world out of poverty we've been able to feed tons of people give shelter to everybody cure all kinds of diseases and so on so an enormous amount of progress has been achieved thanks to science

science is a massive enterprise with lots of different subfields many different areas of knowledge you can think of any particular field and any area of study and so on science is a big part of it at the end of the day what science is about is the uh the process of finding things out the process of creating new knowledge and coupling new new concepts so think of the scientific method and so on uh there's a famous quote from feynman if it disagrees with experiment it's wrong and that is is the key to science you can think of science as a large scale enterprise that involves humans around the planet there's all kinds of efforts and

systems and so on you think of everything from various universities on the planet various research groups different fields and journals and so on lots of different activity around the process of synthesizing what we know coming up with new ideas transforming those ideas into research projects turning those into um into actual testing and of hypotheses and creating gathering of data to be able to test whether hypothesis is correct and so on all the way through uh writing up those results into some kind of um paper that then gets uh reviewed by a by scientific scientific community gets added to the tree of knowledge and then extends what we know um and maybe the

story stops there or maybe later it figures uh we turns out that um hey actually that wasn't reproducible and we have to like uh unwind that or or hey actually that was uh correct but it opened up the door to tons of other new knowledge so it's a highly dynamic field lots of different activity um and so on now science has a ton of problems there's all kinds of issues with the scientific enterprise even though it has been an enormous engine for progress there's all kinds of things that have been going wrong with it in particular there's lack of funding across a variety of fields at the same time even though there's a lack of funding there's there is a lot of money overall going to science uh there's a feeling that um the the money is not going as far as it

used to that science is not getting as much for its buck anymore there are many um the fields across the board are are way too comp competitive in terms of getting grants then the once the studies are done and produced and so on uh only a fraction of those uh replicates so there's all kinds of science that has been published and accepted and thought to be correct only to find out later that um a huge fraction of it can't actually be reproduced so it's not not even correct so there's a huge reproducibility crisis and there's even scientific uh discoveries whose artifacts are going missing so think of the actual papers or the code or the data associated with

some result going missing from our um our knowledge banks and so on so there's all kinds of issues around around science that needs to be fixed and this is part of what desai is about so this is tackling a range of these problems not wholesale not completely but the design community is trying to tackle a number of these these issues um starting with some some parts and trying to scale solutions great so so what's design design is a movement to improve science using web3 tech and tools so think of being able to use all of the magic of hash linking and blockchains and so on and smart contracts to create systems and structures

that can improve how we do science around across fields around the globe and so on there's a bunch of different focus areas so think of just being able to have open access papers and data and data commons to having better reproducible experiments having um being able to kind of organize labs and organization groups better so think of creating structures like dao's and so on that can enable research groups to form and organize and raise capital and and distribute rewards to participants and so on uh there's new entire funding structures uh things like uh ipnfts and so on uh there's protocols for for peer review peer review with rewards uh historically

peer review has been this kind of kind of predatory situation where lots of academics put in an enormous amount of time and effort to peer review all of the work and journals don't actually uh pay anyone for that labor um and there's all kinds of new incentive structures that are being experimented with and and so on so there's uh it's a this is a fairly new movement uh it has been uh with us for a while uh when i started ipfs um it was kind of a design movement before decide was a thing um i started to profess with a goal of um enabling people to distribute data much better for the purpose of doing science um so a lot of these ideas are are part of what um the core of the project um however the movement has been picking

up a lot of steam over the last year year or two and a lot of new organizations have been have been appearing so um this map has uh i think like doubled or tripled in size in the last year which is really great to see there's all kinds of new organizations coming coming into the ecosystem there's now several groups doing the central uh decentralized biotech funding um groups like vitadao and and molecule and others uh there's many organizations trying to come up with new structures for funding science uh there's several daos that are like um uh scientific organizations themselves and trying to do r d um different kind of kind of research and so on uh there's several foundations and institutions that are um supporting a lot of the design work or that associate

themselves with design one way or another there's many groups exploring different ways of doing polishing many science nfts and so on and so this is a community that is growing a lot has been yeah has been growing a lot over the last few um year or two there's also now um a lot of different meetups and conferences and so on that are gathering uh these communities things like desai day or desi dot you know some place like design berlin shelling point from the github community are funding the comments these conferences are gathering a lot of the conversations around around design so let's talk about funding the commons so i maybe some of you have seen this diagram that i've used in the past around the

innovation chasm where this kind of the science of technology translation so in the science part of the site part is it's mostly focused on the left part just the science part and trying to think of better incentive structures better ways of coordinating groups to produce better scientific output it's worth noting that like the total global r d funding is is both you know from one perspective kind of massive but on the other perspective um not that large and not and hasn't changed that much in the last few decades even though um the the kind of throughput and outcome of the technology that we're building has has grown tremendously um so these these kinds of funding scales are not outside of the reach of blockchains so my my guests uh you know think of um the u.s uh non-non-defense r d is kind of on the

order of like 70 billion a year um which is a lot for sure um but it's not massive uh even just like isolating nsf which is kind of around 10 billion a year it's kind of like the red part in the bottom of the graph that is totally achievable through blockchains so think of um now the crypto space having on the order of like you know trillion to 3 trillion depending on when you look at it um and think of a system that's able to uh so imagine if if blockchains were to devote some fraction of their their supply to r d on a yearly basis imagine taking like one percent of falcon or ethereum or bitcoin or something like that and and pouring it into r d every year um

you start hitting the numbers that are that are starting to get in range of funding signs at a nation state level so if crypto grows by another order magnitude or two hours of magnitude and we have those kinds of structures crypto is going to be able to fund r d and science at the scale of nation states which is pretty crazy to think about so um it'd be great to figure out the structures and figure out good funding pathways um ahead of when when we get there uh so when you start breaking down funding um you know down from those agencies and so on you encounter all kinds of problems uh things like certain fields being under um you know with too little attention or or um way too you know the the programs themselves being um

having kind of uh perverse incentives in certain structures or being like um way too competitive or yielding a condition where um scientists are spending an enormous amount of their time just writing grants there was this effort called the fast grants around coved um and was replicated the same effect uh was replicated in the impetus grants and so on where the these programs just structured a grant program that was very very fast that was able to give out grants on the order of like you know anywhere from 20k to 200k or more um with with a tiny fraction of the amount of time that scientists were putting in instead and just this this is a the results of a of one survey for from the uh fast grounds program

that shows kind of like how much time uh this kind of a survey from scientists that were applying for those grants how much time they normally spend applying for grants think of like you know 25 to 50 percent of the time their time of a scientist time just going to spelling out what they're doing and applying to various different grants like this is kind of insane ideally you would want scientists to spend the vast majority of their time thinking about their work and coming up with new ideas and analyzing the work and so on there's also this other effect that happens which is that grant programs tend to constrain what people um end up exploring exploring many scientists have much more ambitious research than this but they're not able to pursue them because they're not able to get funding for them and so they end up stuck

pursuing all kinds of other work that is not nearly as impactful as perhaps their work might be but they end up following that structure because of the of the grants programs so web 3 public goods to the rescue there's a lot of different groups and of course like this is tiny still the the weathering movement is very small compared to global science or nd funding but if we can get the structures right and we can align the incentives well and we can demonstrate that it works then we can scale it um so then we can scale it by orders of magnitude along with crypto so that's why this stuff can be pretty important so we should explore many different kinds of funding for scientific processes so many different grant programs things like

impact certificates um impact markets um and so on there's there's a whole exploration happening now across many different structures there's the funding the commons uh community has been um sampling a bunch of different different um mechanisms and so on um a lot of this stuff is like now on on youtube you can go and like explore a bunch of the ideas there um and just to give you an example like you know there's groups like vitadao that are like creating a data structure that are giving out grants to groups in exchange for um data and knowledge and ip then they're bundling that ip into ipnfts that carry legal weight and they're using those with then granting ip rights to

to biotechs and then they're going to be funding those biotechs and aiming to return investment through the success of those biotech companies so they're creating this kind of like um there's a structure for this kind of fun um i tend to call it a fundamental development fund uh which is trying to do and fun a bunch of important um work through labs that are not themselves companies with a goal of generating iep and then using that ip to uh to fund the companies then there's groups like molecule who are you know creating marketplaces for that that kind of work to happen either those kinds of grants or that kind of ipnft generation uh certificates of impact are another

fascinating structure that enables you to do these kind of like in retroactive funding where where you enable participants to once they achieve some impact mint a a certificate around that impact and sell it in the market to to anybody that wants to claim that impact that enables a speculative market to emerge around that impact so um you can you're able to then close a loop um across time and be able to kind of retroactively fund uh extremely important work uh this is extremely crucial because many times you you only realize how valuable something is uh long after the work has been done so think of some important research that leads into some really important

discovery and only five or ten years later you realize that that was a fundamentally deep and important contribution and so think of impact certificates as being able to generate cash flows that can throw like move back in time um great so that's kind of you know a whole bunch of experimentation is happening around funding uh still you know very significant already in several fields like there are many groups that are now kind of either entirely funded by by these things or or starting to but this this can be scaled now to much larger um uh levels where many more participants can can um if we can prove this this works we can then scale it to whole industries uh so now some quick thoughts about organizing

people so um in the past like the github has been tremendously successful as just helping organize lots of scientific discovery um there have been entire textbooks and fields developed through github uh there's a great book uh called reinventing discovery uh that kind of goes into into this kind of thing uh many you know particip many groups have just used the the very basic primitives of github around issues and and code collaboration and so on in version control to organize communities of practice and communities of science and so on so that's all uh really good um but what what hasn't um what isn't complete there is is you don't have a way of of creating organizations that do research you don't have any any kind of way of dealing with

with capital you don't have any way of paying contributors and so on so the same kind of structures and problems that open source development um has so some interesting uh experiments so so things like uh there's a lab down group that is creating kind of lab teams where teams can um form around um a group use the dial tooling and blockchain tooling to them form be able to raise funding and then distribute it to participants um you're able to kind of encode the different levels of contribution of different participants to be able to reward them fairly uh there are more ambitious projects of being able to then start thinking about the credit assignment across participants in a larger network being able to think of

various different teams and how they might couple and be able to to propagate reward across them there are groups that are experimenting with peer review so think of many so peer review whole peer review protocols being proposed where you start thinking about the economics and the dynamics of the peer review system and and you arrive at a way to uh in both incentivize that work or rather properly because it's already happening but rather properly rewarding the fact that that work is happening um there's a protocol called ant's answer review that that uh is doing this already and you can find a a demo you can like run in you can use metamask and so on uh which is that's pretty sweet uh worth mentioning as well that bitcoin grants has been pioneered a ton of the

work that can be used here and is already supporting a lot of tooling for participants that want to organize in these ways i mentioned that i mentioned earlier that if i started with this goal because has been used for a lot of the data set building and distribution and and so on one of the really key components here is that you can link content by hash you can freeze the you know some bundle of information get a content address um hashlink it and then be able to reference things so this is think of that as the core primitive that you want in literature anyway so so when you have a citation from one paper to another paper or from a paper to its data or or from a paper to you know some important code that you need in order to

regenerate the data a cid is precisely the thing you want you imagine being able to sort of freeze the entire literature acro with version control and being able to freeze all the important data sets and code required to um uh to be able to run those experiments again uh there's many groups that are that are exploring this kind of thing there's many groups that have proposed uh different ways of doing kind of peer review and science science development through uh through ipfs a number of these projects i think are getting off the ground um and then you can think of bundling that kind of activity and that kind of data generation with um something called the data down these are new so unlike the the dials that i mentioned before which are just already getting started

data dials are very new there aren't that many of them yet but think of it of a group that is able to create collect curate transform uh and compute on data and then govern how that data is used over time govern how it's monetized and figure out how it gets shared some final notes on open access and reproducible science um ibfs has been used already a ton for many kinds of open science work it's already kind of like living the dream of opening up access to to a lot of science so think of you know everything from like the the distributed wikipedia copies to you know there's there's massive um archives of papers uh distributed through through ipfs already um there's there's um all kinds of data sets and facilities and so

on um move this way so it's already enabling an open data commons and there's already people trying to make use it for this reproducibility we're not quite there yet here uh this is where that an area that needs more work but a lot of people have already done the thinking of how to do it so there's really good specs and ideas around how to um take a lot of the the work around standard reproducibility and use ipfs to freeze all of these assets and be able to get to um you know fully reproducible pipeline where you can kind of summon back like specific experiments from the past and be like oh we want to rerun that and make sure that it works and you can bring back totally frozen vms or containers or or functions that you can rerun rerun all the data

pipelines and and verify that the experiments are correct um there's also a whole other angle around doing the data science itself um in this design oriented way where you the actual like notebooks and the data analysis and the artifacts and so on that you're using are using all these web3 powered uh applications there's the first few are starting to do this like they're things like jupiter notebooks and ipython notebooks and wolfram notebooks that already couple uh with tids so you can start referencing all of these these kinds of things um but i think that's going to get supercharged in the future um as now so the falcon network is growing tremendously it's adding an enormous amount of data um and the the falcon network has a lot of storage coupled with compute so you have these

source providers that have tons of gpus right next to the data so those are going to get wired up in the next year with the ability to then issue computational pipelines around that data so think of now being able to generate a platform for scientists to be able to do data science at massive scale leveraging web3 computing platforms so both the addressing of information the storage of information and the computation of that information to have like a full end-to-end pipeline of data science a pretty large scale so finally a quick call to action science is the engine of progress by doing by extending what we know we're able to produce more technology and

improve improve our lives if we can improve the lives of scientists and make their work easier if we can accelerate their development if we can cut down some of their costs if we can enable them to spend more of their time figuring out the problems then we can all as a society advance much faster so this the design movement needs you think of either experimenting with new funding mechanisms or building open access tooling building open science tooling um think of playing around with all these kind of public data or public data sets that are being accumulated think of joining a design team or a dao go back to that to that um diagram and kind of explore those communities and hope to see you in the in the movement thank you very much and

see you around [Applause] you