Vitalik Buterin on Decentralized Science, Aging, AI and Scientific Progress

31:02 What’s holding back scientific progress?

36:37 The digital world’s impact on longevity

41:49 Potential Curveballs and unknowns

45:09 Decentralized science and prediction markets.

‍

Snippet:

Vitalik Buterin  0:00  

What if we fix aging? Right? And you know, what if all those numbers are off by a factor of two, you know, what if there is a radical cultural change that causes people to have like much more, much fewer children, there's just so many different curveballs that could happen. What if, you know, we do discover evidence of alien life, like even if it is bacteria, right? And that just creates a big, big philosophical awakening in people; there's just so many of these big unknowns.

‍

Vincent Weisser  0:36  

Yeah, it's great to have you, Vitalik, for this conversation about the intersection of science and crypto called "decentralized science," hosted at DeSci London at the Francis Crick Institute there; maybe we get directly into it. So on decentralized science, I would be curious how you define decentralized science and which aspects of it are you most excited about.

‍

Vitalik Buterin  0:59  

Hmm, I guess I view a decentralized science as a collection of ideas that basically have to do with trying to answer the question of what are some ways in which emerging open and decentralized technologies can make us, science better and can improve the institutions of science, in a broad sense, can improve mental health, people who are just excited to collaborate with each other and get a get things out to the broader public? So I think some like, examples of this would include one is just funding right. So I think the interesting thing about cryptocurrency is, first of all, that it's just a very effective fundraising tool by itself, right? Because I've used cryptocurrency to donate money to various kinds of charities, like some of them giving money directly to the poor, some of them, you know, doing existential risk research, some of them also doing longevity research, or doing other kinds of science.

I think my most recent, really big and significant one was donating some amount of money to Impetus Grants, which is doing longevity. So just crypto as payments, I think was one big part of it, and like just in addition to providing more funding for science, I think it's fascinating how it provides for more diverse funding, right, like, it's not just traditional institutions, it's also a yet, you know, pretty wide array of, of individuals that can now provide funding in a lot of cases. And then going a bit beyond that there's, you know, things like DAOs, and things like quadratic funding, and all of these different tools that tried to create a more crypto native and more structured approach to funding different kinds of science projects. So quadratic funding, in general, it's a set of the decentralized and democratic ways of choosing where the funds for a particular funding will go. So one way to think about it is that it's like a matching system. So if people donate to one project, then the matching pool also contributes its money. But the unique thing about quadratic funding is that the match is greater the larger the number of distinct people that contribute to a project, right? So if project A, let's say, gets $10,000, each from two people, but project B gets $1, each from 20,000 people, then project B is going to get a much, much higher match than project A does. There's some really interesting mathematical reasoning for why quadratic funding is economically optimal in a certain sense. And if people want to read up on, like, my explanation of that, I'm on my blog. vitalik.ca. I Yeah, have uh, it's just like a quick roll up for quadratic, you can find my post on this, the quadratic mechanisms in general. And then I think we also want to kind of go a bit beyond quadratic funding by itself.

There's the concept of DAOs in general, and there have been more and more DAOs. There is VitaDAO, there are other kinds of DAOs now that are trying to organize the process of deciding who to fund in all kinds of structured ways that even go beyond just the funding question. There's also potentially the concept of you know, using like Blockchain traceability or NFTs or soul bound tokens or anything like that, to try to do what better job of just maintaining the graph of like, who contributed how much to what, right because one of the big issues in academia is the whole peer review system. And the way that that by itself has basically turned into its own game that is unfair in a lot of ways and gets exploited in a lot of ways and exclusive in a lot of ways and creates a very misaligned behavior as people try to chase citations in a lot of ways. And the question of, can we both try to make better versions of what the peer review and or what the citation graph tries to do and what the peer review process tries to do, and also use blockchain and these toolkits to really kind of weave that into the process of doing things and the process of funding things and, you know, really make rewarding people for any contributions to the things that we set, you know important results downstream, be something that happens and much more seamlessly and much more automatically. You know, retroactive funding by optimism is one good example. Right? So I think just zooming out of edits, so there's a large collection of things that people are doing with blockchains, with DAOs, with cryptocurrencies, you know, with soulbound tokens, and you know, with ZK, and all of these different tools that try to reform or replace or improve on all kinds of processes in, like, how we do science as a civilization.

‍

Vincent Weisser  6:37

Yeah, makes a lot of sense. I think, maybe we want to double-click on public goods. So I think it's really interesting to see the first experiments with quadratic funding also showing some of the ways that it could potentially be improved. So we'd be curious, on public goods experiments, like for quadratic funding and retroactive funding, which experiments, would you like to see in the context of science that you haven't seen yet? Also, are there potential improvements to quadratic funding or retroactive funding? Yeah, that we should attempt.

‍

Vitalik Buterin  7:13  

Yeah, I think science is actually a great place to try to do experiments with, like, more sophisticated forms of retroactive funding, because science is the one place where it's, like, more obvious than anywhere else. I think it's true everywhere. But it's particularly obvious in science, like just how complicated the contribution graph leading to any particular result is, right? Like, if, let's say, someone invents some new zero-knowledge proof protocol, usually, that protocol itself is a version of some other idea. And it uses, you know, some particular polynomial commitment scheme that was invented by someone 2 to 10 years ago. And then it builds on top of this other results, and that polynomial commitment scheme also builds on top of other results, and it's basically, you know, people building on stuff, that's built on stuff, that's built on stuff that, you know, goes all the way back to RSA and Shannon, and other stuff in the 1940s, the 1970s, and, ultimately, stuff even long before then, right. And there's more of a tradition within science of like being explicit about that graph, which I think is fascinating. And what that means is that it's an opportunity to try to solve what I think might be one of the critiques of retro funding in practice, which is basically the last step in a particular chain that leads to some outcome getting funded but some of the less glamorous intermediate steps don't. So everyone has a big incentive to try to play up how big their contributions are and, like, be really loud about what they're doing. And that's like trying to reward not just the people directly responsible but also, you know, the people who are responsible for the parents and the grandparents of a particular result. I can do that in a way that's incentive-compatible. I think it's a really fascinating thing to think more about. And I think in the, it's a good idea at every space, but in the decentralized side space, in particular, be interesting to really try to start

‍

Vincent Weisser  9:32  

Which ideas from decentralized society, also called "DeSoc," that you coined with Glen Weyl and Puja and broadly in social technologies are you most excited about in the context of science?

‍

Vitalik Buterin  9:47  

I think there is definitely a big overlap right because DeSoc is in a lot of ways all about exploring the nonfinancial ways to try to encode different kinds of relationships. So we have things like POAPs that encode, you know, whether or not someone attended a particular event, we have things like proof of humanity, we yet, you know, have a bunch of various attestation protocols that people are working on. And I think all of those have a big voice in DeSci. Because, like, as I was mentioning before, one of the big things that I think that DeSci space could do was really try to, like do better things with having a higher quality, I mean of contribution graph. And that's something that we could get more of automatically just by, you know, having more of that kind of information committed to the chain everywhere, right? So the more of, you know, there are these public commitments of like, what contributed to what, who participated in what and you know, who was a part of a particular community, the more it becomes possible to do more of those kinds of experiments. So I view DeSoc as being a space that's going to create and really yet build out a base layer that will make available information that a lot of these decentralized protocols are going to end up making use of.

‍

Vincent Weisser  11:29  

Yeah on decentralized society. I wonder, do you see the first use case being mainly around identity and soulbound tokens and then enabling a bunch of other use cases? Like, what could some MVPs of this look like? If a builder is listening, who wants to build something at the intersection of DeSoc and DeSci, like, for example, the Gitcoin passport, which I think is interesting, but there might be other identity solutions or other DeSoc ideas that one could build?

‍

Vitalik Buterin  11:57  

I think the other really natural one is basically trying to, think through it and figure out what it actually would mean in practice to have some kind of POAP. That's proof of attendance token that represents the contribution to some particular piece of scientific research, right? Now it could mean that like, you get one by being an author of some paper. Or some posts, it could mean, just being involved in a particular community or at a particular event that's doing research in a particular area. I don't know yet. But that's one of the challenges that's worth figuring out.

‍

Vincent Weisser 12:51  

You mentioned earlier some of the amazing initiatives that you supported; maybe you could paint a picture of the areas of science that you're most excited about and why you're excited about them.

‍

Vitalik Buterin  13:04  

Sure. So one that I've obviously been excited about for a long time is longevity research. Yeah, I read Aubrey's book, Ending Aging, when I was a teenager, and I was immediately convinced and definitely really inspired by the vision. And since then, that's always something—a space that I have dreams that I could somehow be helpful too. And so I started donating to well, longevity pretty much as soon as I started having significant amounts of money, I try to continue being more supportive of different groups in the space in various ways.

‍

In addition to longevity, one of the others that I've started going into recently is COVID research; this was actually one that kind of happened accidentally, in that, you know, I got these Dogecoins back in 2021. And I needed to send the Dogecoins somewhere. And right at that exact same time, India was getting hit really hard by the COVID situation, right? This was like late April, early May 2021. And there was this group called Crypto COVID Relief India that was just starting to get help in India, trying to make its way through the crisis. And I sent some Dogecoins to crypto relief, like these are the Shiba Inu tokens that I got that just somehow ballooned to a really huge market cap. And at the time, I thought that it was not going to be a huge amount of money. And so we didn't really think much of it, but they ended up being able to get far more money out of it than I had imagined it would be possible like I imagined they'd be able to get maybe $10 or $20 million, and then Shiba would crash to zero, but they ended up cashing out like a bit over 400 million USDC then They started funding a lot of projects in India. And then I also started basically working on this effort called Balvi with some other people who are into infectious disease science.

‍

And we really understood the COVID situation and I knew that this would be what also simultaneously use some of that money for some more kind of potentially speculative and high-risk, but also potentially really valuable projects. Right. So this is basically like full spectrum COVID support. So it included projects that are working on masks, projects that are working on vaccines, UV radiation, air filtering, you know, wastewater surveillance, just like the entire spectrum of figuring out a lot of ways to detect COVID, Testing is the other one, you know, both betas, like antigen tests and lab tests and PCRs, and everything in between. And then finally, try to deal with COVID, treat it, and make it less of a problem in the event that someone does get it. And what was interesting was that Balvi did end up funding a lot of different work and even ended up funding a lot of open-source work. So one of those things was RaDVaC, which was an open-source project that was working on vaccines for COVID. And they were one of the early ones that started working on nasal vaccines, right, because they know, there's this fact about COVID, that the place where it's like easiest and most practical to stop, it actually is like basically in the middle upper respiratory system. And if you don't get that, then basically, you're not actually going to create a vaccine that completely stops COVID. Because like, you could easily get it and pass it on to someone, even if the rest of your body outside the respiratory system is immune. So if you want to make what's called a sterilizing vaccine, then, you have to get pretty close to the source. And also like nasal vaccines, you know, they get don't require an injection, are potentially easier to administer, potentially, you could just give the belt to people at a less scary and all that. And they were one of the earliest projects that we're doing that, there are a bunch of other projects as well other vaccine projects, other projects in a whole bunch of other areas.

So what was fascinating there, I think was that, like one just cryptocurrency ended up being are an extremely convenient funding method, especially for all of these different international things, right, it's just so much easier than sending a bank wire, even with the controls that we had, like we had a multi SIG wallet and multiple people had to sign off to prevent the money from getting stolen, right, like, which definitely adds like a bit of bureaucracy and checking, but like, even with that, it's just so much easier than, you know, sending bank wires for everything and dealing with all kinds of banking systems and it arrives instantly, then the other one, I think, is just like the value of, you know, a model of philanthropy that's more willing to trust people, right, like a lot of donors to this project say, yeah, be like, try really hard to inject to their own opinions into everything and basically say, you know, hey, you know, you have to put all your money into this and put all your money into that and that there's no way I'm gonna let you spend anything on overhead because everyone agrees that overhead is, you know, totally unacceptable and bad. And you know, these things add up actually, hamstringing a lot of people right, and with Balvi, we're trying to be a win trying to intentionally not do that and I think that led to a lot of good outcomes. And then I think it also counts as science you know, even though it's not biology but the whole “zero knowledge” space.

I feel like I've been trying to support the zero-knowledge space in multiple ways so one of them is obviously the Ethereum Foundation funding a lot of zero-knowledge research. I did like actually contributed a bit myself to a couple of zero-knowledge projects. A tiny feature accredited to myself ended up making its way into one of the original plug protocols, which I thought was kind of fun. The other really big thing was just making these exploiters, write these blog posts to try to make the math behind zero-knowledge protocols and a whole bunch of other cryptography really accessible to people. What I basically felt like I was doing was trying to fill what I thought was this missing middle niche where, on the one hand, you have academic papers, which are just like horrible and unreadable, and, like, part of the reason why they're horrible and unreadable is not because, you know, academics are like bad people who are elitist, No, there definitely is some elitist.

So it's just because, you know, the incentive structure encourages writing to a particular audience, and I think changing the incentive structure is part of what DeSci is supposed to be about. So, either you have PDFs that are basically unreadable, unless you're almost part of the same lab. And then on the other hand, you have Pop-sci writing, which is extremely accessible, but it just, like, doesn't explain, you know, pretty much anything right, it's like trying to explain physics to people. And then, oh, so there's this thing called a light cone. And you can imagine a light cone by imagining something like one of the pyramids of Egypt, except it's circular. And like, you know, it gives an image in people's head, but it doesn't actually help people, like, understand the science in the sense of like, being able to, like getting insights about how the world works, that would like help them actually understand or predict anything. And I wanted to fill the big missing middle of like writing that's both accessible to some number of people. And wouldn't, you know, yes, you will have to know a bit of undergraduate math and you also have to work hard, but at the same time, he could still understand it, but still, like actually being accurate and like actually describing the whole protocol.

And I think, part of my vision, there was definitely just inspired by my own experience as a self-taught person in a lot of areas and just being frustrated by you know, all the unreadable PDF papers and the, you know, extremely superficial Pop-sci material and trying to create something better for people who are not taking the conventional University path. And that's good, you know, ended up getting a lot of good feedback on a lot of those pieces. And it feels like that's turned into more of a trend, with a lot more people writing that kind of stuff now, which I think is good, right? I actually think educating people in a decentralized education about science really should be viewed as being part of decentralized science, right? Popularizing knowledge and even popularizing knowledge across disciplines is so valuable, right? Because so many of the important contributions, I think, come from intersections of different disciplines. So those are some of the things that I've done, and I definitely look forward to doing more things in the future.

‍

Vincent Weisser  22:39  

Amazing! What strikes me as unique with your kind of efforts is just like the variety of different approaches like even from the Ethereum foundation to Balvi to some of those academic efforts to much more on non-traditional ones like what have been some surprising learnings between for example,  how something like Arc institute even or Impetus and Balvi setup and being kind of in the trenches with something like Balvi be super curious like which learnings you incorporated from remember Ethereum foundation and other things into setting up Balvi and how you structure it, which might be different from like a traditional foundation or brand given body?

‍

Vitalik Buterin  23:21  

I think being lean is one of the warnings, right? So basically, another one is setting up an organization that is explicitly not trying to have an infinite lifespan. And I think to some extent, those two are connected, right, in the sense that, there was this, I think it was Venkatesh Rao for one of the new things called internet philosophers, who has a lot of interesting ideas, but one of them is, I think he called it, Neotenism, it's like the opposite of the Lindy effect, where the Lindy effect is like saying, I trust things more if they're old. Because they stood the test of time that he is saying, like, I trust things more if they're new because they haven't been corrupted by these yet inevitable pressures of social entropy yet. And like part of the reason why you know structures gain entropy over time is because they go through, you know, disasters, and those disasters will lead to calls for defending against things and creating layers of accountability and defending against more problems, over time that just leads to structures that have like, good defenses against though leading to problems but also unfortunately, you know good defenses against being able to move quickly into new things as well. And like the thing that the Ethereum foundation and Balvi have in common is that they're explicitly not entities that are meant to survive forever. The thing with the Ethereum foundation, it's, like, more open-ended, but there's definitely a big strand to the original vision that says that, well, maybe the thing is going to disappear eventually.

Or maybe the thing is going to turn into a DAO eventually. And Balvi is explicitly intended to start with a pool of money, spend it and then say that we're done, right? So, like, explicitly say that the goal is to disappear so as to not have the pressures that we had to create some of these permanent bureaucracies and, you know, avoid the pressures of negative people more and more people trying to exploit it over time. So that kind of approach I think, has been fascinating to see it. Another one is just appealing internationally. Right? So one of the big differences between crypto and a lot of other frontier tech industries is that like, if you look at AI, for example, AI is completely based in the Bay Area, like OpenAI Bay area, Stable diffusion Bay Area, you know, Deep Mind and like the, you know, the entire kind of Google alphabet bubble, you know, largely Bay Area, right, Meta, you know, what's in there? It's, like, huge Bay Area dominance. Crypto is not like that at all, right? Like the Ethereum Foundation, which is legally based in Switzerland with a major satellite in Singapore, it has a large number of developers in Berlin.

It has, you know, people in you know, the US and even within the US, actually, Ethereum's primary American hub is Denver, which is, it's like, you know, neither SF nor New York, nor one of the kinds of mainstream contrarian hubs like Austin or Miami; it's like, hey, Denver, and it's like, you know, right in the middle, and then it's, it's there, and it's cool. The Z cash is also there, by the way, well, technically Ethereum foundation is more in Boulder, and Z cash is more in Denver, but like, you know, to me as an airplane person, if it shares an airport, it's part of the same city. But then, you know, Canada, we have developers in Taiwan, we have developers in all kinds of European countries, and Justin has been creating more and more of a research bubble in the UK. And there are places like Cambridge, you know, strong community in Taiwan all the way through more and more of a Chinese community growing, right, more and more of a Latin American community growing. A lot of Ethereum people in places like Argentina. So crypto has been, I think, practically more successful than most tech industries at international diversity in that way. And I think, what Balvi in particular, also had that always as part of its vision, right, I mean Balvi had crypto relief. And that crypto relief was about supporting, you know, India, and this was happening at a time when this was, as I remember, one of my bigger disappointments. The whole India crypto situation was just bubbling up on Twitter. And I remember, but even in the US, like, almost nobody was talking about it except for basically a couple of middle Silicon Valley people.

And this was actually one of the things that I still respect Silicon Valley for, I thought the Silicon Valley elites were more on the ball and recognizing that there was a problem happening in India and trying to think through like what, you know, what, what they could do to help than like pretty much anyone inside the Biden administration. And it was only when the Twitter noise became really, really just like unbearably wild doesn't, you know, they came out and made one kind of a statement. Right? So that was, I thought kind of testaments to, like more non-political organizations' greater ability to care about things that time you know, go beyond the kinds of countries that you and me and probably even most people listening to this are going to be based in. But then Balvi itself, you know, went all the way beyond the idea and supported, you know, stuff in Thailand, stuff in the US, stuff in the UK, a lot of stuff in Latin America, stuff in China, stuff in all kinds of countries. And I thought, you know, it was good that we did that. And it was only possible, I think, because of the international networks that we leveraged and international networks that were actually only available to me because I had made that decision on you know, nine and a half years ago to start roaming around and be a crypto Nomad instead of being based somewhere. So yeah, being international, I think those are really key.

Just being willing to trust people and process grants quickly. That was something that the EF definitely has done at various times and sometimes it takes longer, but it's I think, it's obviously not always the right approach for the long term, especially because if you do it too much, then eventually the Grifters kind of catch up And that's probably one of the main reasons to become more careful. But like, when your goal is scaling up quickly, and when your goal is helping a lot of people quickly, and especially when you start off being kind of not an established donor in a space it can be, it could be a good approach. So, yeah, those are some examples. And, you know, I'm hoping that those aren't just the models that we're going to be stuck with, you know, I hope that things like DeSci and, you know, reputation systems, and POAPs and all of these things are going to do an even better job at helping to solve the problem of, for example, who is trustworthy and what kind of people are worthy of more of a fast track process, right? And trying to, like, really get closer to the frontier between, like, accountability where it's necessary and, you know, giving people the freedom to do their job in places where that's obviously the right approach.

‍

Vincent Weisser  32:01

Yeah, totally agree. Like it's kind of crazy that how much of the science and everything else like a centering around for example, SF or Boston, and like, how many cities and places kind of leaving out kind of shifting gears towards scientific progress? I would be curious, on your take on like, what is broadly holding back progress, maybe like in a scientific and technological sense, globally, but also kind of relates to add, like which scientific breakthroughs you're most excited about to see in your lifetime? And hopefully, your lifetime is fairly long, but that might be a breakthrough. That might extend the amount of braces you see in your lifetime. But we'd be super curious, especially also under the lens of like, Tyler Cowen and Peter Thiel and others and Patrick Collison saying we are in stagnation and that we'd like need to get back onto the track of growth and scientific progress. Be curious on your take.

‍

Vitalik Buterin  32:00  

Yeah, I think there are a bunch of factors, probably the biggest is always just the fact that public goods are chronically underfunded because it's hard to create a business model for them. Right? Like, if you're just a small actor, and at a global scale, everyone is a small actor that you only get, at most, a few percent of the benefit, maybe even a fraction or percentage points of the benefit of anything that you fund, right? And that is just always the biggest problem. That's a problem that humanity has been stuck with, and that’s a problem that is going to be hard to get past. But you know, if we can get from, only funding things that have a 10,000:1 benefit-cost ratio, and then all the way down to, you know, funding everything that has more than, let's say, 100:1 benefit-cost ratio, that's already a huge amount of progress. Other things, I think, definitely, like exclusive in essence silos. The fact that there is a lot of knowledge that's sort of generated by particular academic communities, and it's stuck in particular academic communities and very smart people who would be able to contribute just don't have that knowledge be accessible to them, because we've invented a path that basically says that you only get the required background knowledge by you know age 24, if and only if you're wealthy and privileged enough to like go through a particular university path. And the reality is if you enable people, there are lots of people who would make huge contributions when they're 15. Right. So trying to create paths that are accessible to more people and, you know, across wealth levels, across age levels across, you know, kind of regional and national levels, I think, its something that's going to be really valuable. ideology is also, I think, another one, right, like, there is the whole line.

And I think Tyler Cowen is one of the people who has kind of, you know, reblogged things about this, that basically sometime after, you know, World War II, and then kind of going further over time, a lot of people became much more conservative in certain ways about the physical world, and there was this mentality that, you know, actually kind of keeping the world pristine is good, and like large-scale, intentional human change is, by default, a bad thing. And, you know, the physical world is something that we should not touch. And that, you know, like, we see that in a lot of, in a lot of different contexts. So like, they're like that probably, you know, massively reduced the the amount of research into nuclear power, for example, similar things are definitely, reducing the level of interest into longevity research, similar things are reducing the level of interest into just you know, all kinds of areas that were improving as human lives ultimately involve some, you know, changing either ourselves or the physical world. But the big irony is that the digital world ended up being completely exempt from that mentality. And so he basically had a couple of decades where a lot of the smartest people basically went into, you know, creating social networks, creating all kinds of these, like Internet tools and creating all kinds of just ways for people to play, in some cases, increasingly negative of some internet games with each other. And lately, we've been seeing a backlash against that, which is interesting, right?

Like, we've been seeing a lot more interest in solar power, a lot more interested in nuclear power, a lot more interest in nuclear fusion, a lot more of just kind of like, human positivity and even, you know, positivity available large scale human intentional action, right. This is one of the things one of the ideas that I feel Palladium does a good job of promoting, for example, also the rise of the developing world, right, because the developing world, I think, never really went through the same phase of like getting a little bit disillusioned with everything that these you know, the rich country West did. And, you know, we've been seeing a lot of Indians, for example, getting into science with some very different perspectives. And also like the crypto space by itself, one way that I think you can critique the crypto space is you can say that, oh, actually, it's just part of the exact same phenomenon of people in the 2000s and in the 2010s. Like just forgetting about the physical world and thinking that they can do everything they want by just, you know, creating new tools for people to interact with each other digitally.

I think one of the interesting things about like network states, for example, is that its an instance of people in a lot of cases definitely kind of like crypto ideological ill lineage, but like rediscovering the fact that the physical world is important. And aside from network states, probably another example is the radical exchange movements, which have definitely different ideological beds, but I think it's also about physical spaces to some extent, in the same way. And there's a lot of crypto adoption going on, like crypto-inspired people going in that direction, too. Right. So, I think, you know, just having an ideology that's, you know, positive about the collective human ability to improve our understanding about the world and, you know, even have an impact on the world, and on physical space, and, like, an understanding of what it means to like this, you know, do the right thing and not do the wrong thing in those in those areas without just like falling back on saying, like, you know, let's set you know, let's not go here entirely, is one of the things that I think is like changing and slowly coming back and they just going also bring more people into science in the future.

‍

Vincent Weisser  38:33

Being curious on this century broadly, people like Holden Karnofsky called it potentially the most important century for many reasons, including, of course, existential risks, AI progress, and broader technological and scientific progress. What's your view on how you could imagine the century to play out, especially in regards to, for example, AI progress, the singularity, or specific existential risks from AI to others? Kind of like, top of your mind for you for this century?

‍

Vitalik Buterin  39:08

It's a good question.

‍

Vincent Weisser  39:10 

It's kind of like a vision or forecast. 

‍

Vitalik Buterin  39:12

The interesting thing is that there's a lot of curveballs this century, right? So, like, I mean, 2020 to 2022 have been this, like, crazy, long, three-year period of, you know, curveballs and some, a lot of negative curveballs and some curveballs that are, you know, either positive or negative, depending on your views on AI risk. But definitely be lots of curveballs, either way. And over the next century, there's just all these different potential curveballs that we just don't know yet. right? So one of them is what the AI future is going to look like.

Like, basically, are we in more of a kind of film world where it's just inevitable that some broad AI project, once it gets powerful enough, is just going to self-improve until it can basically control the world? And it's up to us to make sure that it controls the world in a way that's agreeable to us, despite us being very, very far away from being good at being able to put concrete goals of any kind into AIs at this point, or are we in more of a Robin Hanson world where there's no single AI that's might have particularly far ahead of all the other AIs. And so we're going to see a politically multipolar world, going into the foreseeable future. And then, like, there isn't really an existential risk, but there is just like this ongoing battle of offense versus defense technology, and our job is to try to make sure that no offense doesn't overtake defense by too much. Or, you know, also are we in a world where, you know, the current LLM approach is basically you know, that plus scale literally is all we need or are we in a world where, look, we're already starting with what GPT4 is going to be close to the last major piece of progress that we get from LLMs. Otherwise, LLMs are exhausted, and we need some really significant, more innovations. Another curveball is geopolitics. Right. So, you know, obviously, last year has seen a really unfortunate, you know, negative curveball in relations between Europe and Russia and the next two decades, or we could easily see a very negative curveball and relations between the US and China, we could see a very negative curveball in relations between India and China. Or nuclear proliferation. That's another one. Like, you know, this is something no one's expecting, but you know, what, if, like, just any one of these countries that has a big grudge against the United States just like decides to go into, you know, sell nukes with full plans for how to build more of them for, you know, 10 million bucks a pop to a bunch of randos. And we're up to having 30 nuclear states within two decades. And, you know, some of them start declaring a nuclear war against each other, like, Bio is another curveball right?

So we get, oh, we have one big COVID, which, I guess right now, the smart people are still like confused and like, what it's still 50/50, about whether it was like a wet market or a yellow lab leak. And but you know, blood leaks do happen all the time. And it's, you know, we could see more of them as we see more labs. And you know, we could see more intentional ones at some point, right? There's just that. But then on the other hand, you know, what if, like, it turns out that longevity does get solved, like, literally 20 years from now? And so, you know, what's that going to do? Like, what's that going to do to Peter Zeihan's, you know, demographic projections, right? It's like, everyone has these numbers that say, like, oh, in 2100, India is going to have a population of 1.4 billion China is going to be 0.7 Billion US is going to be 0.4 5 billion Africa is going to be like, somewhere around four. But look, what if we fix aging? Right? And you know, what, if all those numbers are off by a factor of two, you know, what if there is a radical cultural change there causes people to have like, much more much fewer children? You know, what if there is that the combination of you know, having a richer society and having like one tier or more of automation suddenly makes it more practical for every family to have five kids each? There are just so many different curveballs that could happen.

What if, you know, we do discover evidence of alien life, even if it is bacteria, right? And that just creates a big philosophical awakening and people, there are just so many of these big unknowns. And you know, a lot of the ones that I've mentioned are negative, but no, I think we could easily see a lot of positive unknowns as well, I'm, yeah, it's definitely going to be a really big century, and it's definitely going to be a really important one. I think I mean obviously the 21st century is more important than any century that came before. But the question is, like, is this 22nd century going to be even more important? Like, is that you know, is this the final boss of humanity? Or are we really at step n minus 2? Or do we really have the model wrong, and we're gonna see it, you know, extremely radical change that like, looks very radical, even from the inside keeping you going forward for millennia? I don't know yet.

‍

Vincent Weisser  44:36 

Maybe, to close off to kind of like also the audience of builders in decentralized science. You mentioned some of the ideas you'd be excited about. What are maybe some other ideas you haven't mentioned yet that, like the DeSci London hackathon, you would be excited for people to see built?

‍

Vitalik Buterin  44:54

Things they’ll be excited to see built. Yeah, I think one would definitely be some kind of like, you know, whether it's attestation-based or soulbound token-based or POAP-based or something, some kind of solution for trying to like assign credit and assign credit recursively for things, possibly an implementation of retroactive public goods funding, tries to incorporate those kinds of ideas, possibly, what else simply build the education side would be interesting, right? So like trying to whether proactively or retroactively public funds, education, in different kinds of, you know, scientific disciplines, and the different ways that I've tried to evaluate like, Which ones of those are actually effective? prediction markets, actually, this is one we didn't talk about at all right? Well, this is—I know, Robin Hanson is really into this. And this is one of his big primary ideas, but there's just such a natural tie-in between decentralized science and prediction markets, right? And basically using them as a way of, like, trying to predict what paths in scientific research are more likely to be valuable retroactive public goods funding to some extent, it'd be like saying that kind of applies to prediction markets, right? Because the whole idea is that you know, you're gonna have investors that try to get participate in these retroactive rewards and that's how this stuff gets funded but like prediction markets more generally and their applications to science is another interesting area and with layer twos and Poly markets and prediction markets are getting more and more viable and more and more adoption so expanding number definitely be really interesting. So yeah, those are a few examples; otherwise, I'm sure people can come up with a lot of their own too.

‍

Vincent Weisser  46:48

Amazing yeah, I'll collect all the older show notes and things like the DeSoc paper and other things we've discussed. Maybe it's like the very last but is there a specific book, blog post or podcast you would recommend to people specifically in the context of science you enjoyed reasonably

‍

Vitalik Buterin  47:08

That's a good question. I mean, you know, 80,000 hours is always good. I feel like there are lots of very interesting and very smart people talking about all of the different areas that we brought up today. I don't know; I feel like we need to. It would be interesting to just have a podcast focusing on decentralized science, particularly at this point, and like, does something like that exist yet? Maybe you should start one?

‍

Vincent Weisser  47:37 

This will also go up on the DeSci podcast. So we also have a long conversation with a bunch of different people. I'll share the link in the show notes. But we have a bunch of other awesome conversations coming up, and all the talks and workshops from the DeSci conferences are already up. Find it in the comments. And on the view of the talks, I really enjoyed the intersection of DeSci and DeSoc, so I think that's a good related one, but thank you so much, Vitalik, for taking the time, and yeah, I wish you a great day.

‍

Vitalik Buterin  48:09

Have a good day too.