I’ve read the paper. I’ve talked with quite a few people interested in implementing the idea, from the social financing side, as well as from the public donation side, as well as from the extracting promise out of mayors promising to not do something within their term side and so on.
Just for the sake of benefit of the readers and viewers of this recorded video, would you like to outline some of the main ideas and topics that you are personally now working on or focused? Then we can...
Around quadratic voting, the thing that we are most interested in is first of all, the idea of having a democratic system that allows minorities to protect themselves rather than to have bureaucrats or judges or something like that be in charge of protecting minorities...
By giving every citizen an equal budget of what we call voice credits that they can allocate to support or oppose issues and candidates that they most strongly favor or oppose, but not allowing people to just be extremists and dominate an issue.
This can be applied to voting situations. It can be applied by politicians to poll to figure out positions that might form a legitimate basis for legitimate public decision-making. It can also be used for funding.
If you want to fund local public goods, Vitalik Buterin and I worked on a variant of this idea where rather than it just being a way of voting, you would actually have public matching funds that could be given to different local projects or even, say, to media.
Individuals could make contributions. The amount that would be received by, say, the charitable cause or the candidate or whatever would be the sum of the square...Come. No, I’m just checking that she’s not here. Sorry.
Would be the sum of the square root, all squared. What that would mean is that smaller contributions would receive more matching funds from the public. Causes that had received contributions from more people would also receive more matching funds.
It’s a way of overcoming the usual free-rider problem, where when you have public projects, people don’t want to individually contribute to them because they would only do it if other people would go along with them. I’ve been working on all of those.
We’ve been thinking about applications for everything from funding news media, because it’s not usually well-funded, just through really private means. On the other hand, you don’t really want the government funding it because it could control the media and undermine democracy. Everything from that to making decisions in local councils.
At the same time, we’ve also been working on these identity solutions that would be necessary to support a system like that. It requires a notion of different voters. If you don’t want that to all be done by some central government authority approving people to participate, we’ve been working on identity solutions as well.
Say I’m a person interested in participating in crowdfunding you just mentioned. I’m a regular funding person on Patreon. Now Kickstarter has introduced the new Drip. I’m sure that you’re aware of many other such platforms.
Of course, Kickstarter is a B Corp, supposedly. They drive their social purpose and so on, but nowadays we’re also seeing, because of technologies out there now, quite a few what we call platform cooperatives. With any other name, people are basically putting up their own crowdfunding, distributing schemes up on open collectives and other open co-op movements.
There’s a donation platform on Ethereum called WeTrust that actually put $100,000 or 500 ETH behind matching funds for liberal radicalism, for donations to charity. There’s also a lot of different mostly Ethereum-based platforms that have been using quadratic voting for various governance things.
Everything from regulating the process of electing people to do block-making within a permissioned system that is used for doing import-export regulatory compliance to commercial real estate developments that are tokenizing real estate and that are governing some of the choices about how to invest the community resources using this mechanism.
I was just looking at this. Fang, just for context, I’m looking through this crowdfunding website that is currently applying Glen’s idea of what we called QV or quadratic voting. Can you see my screen?
In short, it basically says if you have a lot of money or if you can mobilize a lot of people to donate a small amount of money each, it’s going to be roughly the same by taking the square roots of each donation and matching them accordingly.
There’s the usual suspects -- MIRI, SENS, the Ubuntu Foundation -- joining this crowdfunding experiment. There’s also African Advocacy Network, as well Surgeons of Hope, the more traditional charities, and, of course, some people in between like Code for America, which I’m not surprised at all as being listed here.
One of the underlying assumption in the QV idea is that the projects themselves compete somewhat for resources, so that the, I wouldn’t say winner-takes-all, but the most well-known charities, or most well-known causes, or most well-known participatory budget items or whatever, dominates the resource in a network effect, increasing-returns fashion.
QV is designed to mitigate that. Taking this very concrete example of quite a few people funding the Lupus Foundation, and at the moment, not much at all at the African Advocacy Network, how does it help?
One property that QV absolutely does have is that, in this particular formula, the more people that are contributing to something, the more the effect of a marginal dollar you contribute on that particular one.
Instead, the notion is, you could give a little bit of funding to some things, more funding to others, etc. The notion is that it should allow for an optimal balance between you not wanting things to bee too de-fragmented because people feel they can free-ride on the things that already have momentum.
I think you do need to have probably some view of what the current funding levels are, and you actually saw that on that site. They make it pretty transparent what the current funding levels are. That’s helpful for the users.
I haven’t come up with a really compelling visualization of it, but in some of the articles online, I believe there’s one where they show this cool diagram which shows, I don’t know if you saw this one, but they show little blocks stacked on top of each other by their height, and then their width...
That’s a great visualization. Just think that you have any number of square votes, really. [laughs] You can buy areas, but the areas is going to count toward their height. I think that’s a beautiful thing.
Yes, of course. It ties very well with the idea of the experiment actually that’s going on because it’s a spring. You can spring that goes into a funnel and it’s all very water, common spaced liquid solutions.
All right. Fang, would you like to quickly introduce yourself to our viewers? [laughs] We can chat more freely I’m sure afterwards, but it’s just I have maybe only 40 minutes after this, so we can maybe switch from topic to topic.
Sure. [laughs] My name is Fang, and I’m a Service Design/Consultant at PDIS. One of Audrey’s agenda is open government, so I help facilitate the mechanism which is called Participation Officers Network.
It’s a network of 70 civil servants across 34 ministries. We hope to use that mechanism to cut across the governmental silos and help people to work towards different issues more openly and creatively, not just within the circle but also onto the wider stakeholders.
At the moment, we’re still using old school approval voting to pick, every month, which topic or priority to work on, and because they come from e-petition, you can also think of it as a kind of approval voting. At any time, anyone can decide to countersign a petition to raise their priority about which that we’d take an interest in looking at.
At the moment in Taiwan, there’s 23 million people. E-petition network online is being used by 5 million people, so one-quarter of the population, which is not too bad. People just countersign each other.
We have some machine recommendation algorithm, like Netflix or Amazon, that recommends similar interested petitions. It’s also of course for budgeting visualization, regulation pre-announcement, participatory budgeting on a city level, so it’s an all-in-one participation platform.
The way it works is that whenever there’s anything that receives 5,000 signatures over a two months period gets a VU from all the all participation officers. They can explain and defend whether it needs a cross-ministerial collaboration. We do a approval voting anonymously on it and then we select two cases every month to collaborate on.
Each and every one that reach the 5,000 people threshold automatically gets a binding power to basically be interviewed, be in talk with by the stakeholders, and publish the full transcript of the conversation, and get a point-by-point response within two months from the respective ministry.
That’s one of the more successful direct democracy-ish experiments that we do. It’s been working pretty well, because people, essentially, when they’re petitioning, have unlimited number of votes. They are somewhat authenticated through SMS. It’s difficult to get 5,000 SMS numbers, as you know.
That’s the system that we’re currently working on. Just as you were talking about this crowdfunding idea, I was just wondering how QV or a similar design can help. It seems like when It’s just agenda setting or priority setting, and it’s not allocation of resources, it’s kind of OK to use approval voting, no?
I think approval voting is better than some systems, but I would prefer a QV-based system. What I’d like people to be able to express is the same thing they, in other cases, express through a protest. When you get out and protest, it’s a more costly action than just signing a petition, but it shows that something’s very important to you.
If every citizen had a budget of credits, and they could say, "This issue is incredibly important to me," maybe, only you would need 300 of things like that. You would need 10,000, or 15,000, if people just say, "Well, I’m interested, but I don’t really care."
At the moment, for example, at any given moment, the Join platform may have 100 petitions going on. Truth is that maybe, after two months, only five of them will get the 5,000 people threshold. That’s the reality we’re now facing.
What you’re proposing, essentially, is that if you can get any number, like 500 people, feeling that this is really important -- for some definition of really important -- so much so that they’re willing to forsake their capability of petitioning for that particular model, any particular issue, any other issue.
They dedicate their petition resource, so to speak, on this issue, then it only takes a square root of our current threshold to basically pin it into our "Must Respond" board. The number is going to be very low. It’s going to be 70 per people, basically.
Yeah, and conversely, like you said, you have these recommendation things. There might be some people who just are having fun, and on the website, they click on one, and then they follow the recommendation. They click the other, and they don’t even think anything about it, right?
Those things you might want to require 10,000 or 15,000 such signatures in order to respond, because they’re not really driven by passion, they’re driven -- depth of importance -- they’re just driven by entertainment. You know what I mean? You want to have some way of measuring that. The idea is that quadratic voting could help you do that.
From an interface or experience design perspective, because we have a professional experienced designer here, how would that even work? Medium says if you are just passing by, you just click on claps, and it’s one each.
I quite enjoy that Medium interface. I think it’s pretty good. I think it would be better to have some sort of a token, if you can build the infrastructure that necessary for that. Obviously, that requires a more persistent identity than just an SMS code. That’s the disadvantage of it.
You just raised a very important point. How can we distinguish if the vote is really valuable? Am I passionate about that, or am I just doing this for entertaining? The question is, how can we tell? What are criteria that we can set up to evaluate that?
I don’t have the address saved on right here. Sorry. The advantage of Brave is that it doesn’t remember everything that you’ve ever done, but the disadvantage of Brave is that it doesn’t remember everything that you’ve ever done.
You have 100 credits left, and here are various referenda, which you could vote in favor of or against. An immediate tax cut for wealthy individuals and corporations.-- let’s say we’re opposed to that. We put one credit on that.
Background requirements for all gun purchases. Let’s say we’re in favor of that, but we’re actually strongly in favor, so we want to put more than one vote on t. You see, my votes are going down faster and faster, as I put more and more votes on it? You see what I mean? Whereas if I just put one vote on that it just goes very quickly.
This measures how much you care about it, by making it increasingly expensive to have more votes, so that you go buy votes just up to the point where you care enough. Then that will be proportional of the number of votes that you’ve already bought.
In this case, this was a poll that we did for a political candidate in the United States, but in general, it doesn’t have to be that. It could be actually citizens proposing these things. Then once they cross the 5,000-vote threshold, they could be allowed.
When you want to propose something, it depends on you’re allocation of attention. Your allocation of attention, based on the information you receiving and also, the people you interact with. What if your echo chamber is limited? It prevent you from seeing the people’s view from other sides, even within the same topic.
What I was going to say is the voting mechanism itself can help shape the incentives, people out to get information, under Quadratic Voting, having very extreme opinions is very expensive to do. Having more moderate opinions is cheaper.
Unlike in much standard voting, if you take somebody who you really disagree with, and you cause them to disagree a little bit less, even if you don’t completely change their mind, that still makes a difference in political outcomes.
The cost of the votes goes up the more votes that you get, in terms of the units of the credits that I was just showing. I don’t think that the voting mechanism itself can solve all these problems. Of course, education is hugely important and so forth.
I do think that the voting mechanism can help create an environment where the incentives are aligned with that. It can actually be pretty powerful, because if you think about it, the founders of the American Republic, they didn’t want a two-party system.
They created a set of incentives that created a two-party system, in spite of themselves. Once you have plurality vote, one first-past-the-post, it creates a two-party system. I think that some of these incentives can filter back into the way that the politics is organized.
That’s also true. That’s a very good point, which is that actually, one thing that we found when we used this survey with people, is that because they have a constraint, and they have to make these trade-offs, we often get comments from people that they learned a lot about their own preferences.
Just to follow up on that, very quickly, because when you talk about trade-offs, the interface you just showed has upvotes and downvotes that cancel each other out. The crowdfunding experiments this spring is entirely up-vote only.
It also can be a little bit more complicated for people. Without downvotes, if you have things that are genuinely harmful, for example, a petition that might be hate speech, or directed, targeted against some group in the population -- it’s actually quite important that you allow downvotes on that as well, in order to try to limit the possibility of having potentially hateful perspectives.
In some politics, for example -- this doesn’t happen, I don’t think, in the Taiwanese system -- in the United States, there’s often something that happens where a politician that’s very not popular will do well, not because they’re popular, but just because people are afraid of the other alternative. If you could vote negatively on the other alternative, that issue wouldn’t show up.
I see. That’s a powerful argument right there. I’d like to show you the real interface of petitioning that we are talking about, because I think that will help massively. That’s something that Fang-Jui can carry on in the conversation afterwards.
This is machine translated, but very quickly, just to give you an idea, for example, there is someone who is Mary X, who we know the SMS number, but we don’t reveal it. They don’t have to be under a real name. They can be a pseudonym.
It’s quite like Ethereum voting in this sense. They can choose a nickname, basically, but that’s consistent over time as an identity. It’s not mapped into a real world identity, only that we know there’s an SMS number behind it.
Then they proposed to amend the provision of our public service leave rules. At the moment of petition, it was at least half a day for each vacation. They wanted to change into by hour, which by the way, takes effect this week. I took an hour off yesterday.
Yes, right. It’s a successful petition. As you can see, there’s a timer that says two months. Within the timer, there has be 5,000 signatures, this check passed, and the response is done. There’s agency response in each and every step.
You can see the supporting argument of each person participating. Of course, nobody has the time to read through the 5,000 people’s commentaries. That used to be Fang-Jui’s largest headache, reading through those 5,000 people’s commentary.
That’s right. Instead of text analysis, we just did crowdsourcing. This is the actual interface now that saves Fang-Jui a lot of time. Basically, we have two columns underneath every petition that people can post their supporting arguments on the left-hand column, and the not exactly counter, but other arguments, on the right-hand column.
What we are now doing is something that is out of the playbook out of Better Reykjavik. I don’t know whether you know the Icelandic experiment from the Better Party. The Better Reykjavik has the same design.
We took a page by not actually showing the bars in a proportional to their number of comments, because that only encourage spam, and nothing good happens. You can see there’s 46 supporting arguments and 11 counterarguments.
Then each one can receive any number of upvotes and downvotes. Of course, there’s a flag button for truly hate speech stuff. Otherwise, we just sort. We discovered very early on, of course, exactly as you mentioned.
Although it takes the troll away, because they cannot really reply to anything. There’s no reply button, so there’s no incentive to attack people. Still, people just casually use downvotes to censor the good arguments.
That’s interesting. I think this is another area where something like quadratic voting could be interesting, because it could make it costly to just censor. On the other hand, it could allow you to flag borderline hate speech. You know what I mean? That you don’t have time to actually investigate and so forth.
I think that’s another potentially interesting thing. In principle, you could even unify the two systems together, and say that both the signatures of the petitions and the up and down votes for the arguments could be part of a unified system.
Or you can cost some credit to add a supporting argument. It’s the credit pool everybody receive anew every month. Then it’s used basically just to sort the signals, not the substantial deliberation which will happen afterward, according to Fang’s road map.
Exactly. That’s why we call them voice credits, is we think voting is a voice argument, or a sort of voice. All these things are sorts of voice. We’d like to have a token that can represent a currency for that voice, rather than just a currency for buying things.
No. In fact, I wouldn’t even call it credits. I would just call it voice. You can just think of it about a unit. In fact, in the book, we used to denote the currency marker, rather than a dollar sign, we use a voice bubble.
I think the V here, which is the shape of a vote, really actually carries the idea of quadratic voting also. It’s basically a stamp of approval on what your voice matters like. We can make the dot here proportional, exactly as you visualize.
Another very interesting resonance that I wanted to mention is that many of these ideas in the book came from someone named William Vickrey. William Vickrey was a famous economist who won the Nobel Prize.
Most of his ideas were based on the work of another economist named Henry George. Henry George has a very deep relationship to Taiwan. I bet most people in Taiwan don’t know about it. George inspired the ideas of Sun Yat-sen almost as much as Karl Marx inspired the ideas of Lenin.
Exactly. The book is very connected to these ideas of Henry George. Taiwan, Scandinavia, and Singapore are the countries that have been most influenced by those ideas. There’s a natural affinity, I think.
Ah, OK, right. I think all of those are very relevant nowadays, because we’re essentially building a code-based normativity around the same ideas. That is, legal by design, instead of by interpretation.
In George’s time, it would have to have a lot of post-fact interpretations, negotiations, and whatever to make the system that he designed actually work as intended, instead of as people just randomly interpret it to be.
Nowadays, we get to code those algorithms into code. The communication effort, of course, is the most important. That we can find the intuitive interface that makes people get it, so that it becomes the social norm. Then we compile that into code so that...
That’s the most important thing, is that there’s a notion of social legitimacy around the ideas. That’s the reason why what we’re trying to do with this ideas is not just to go to government bureaucrats, but very much like the approach you’re taking, of trying to be open, trying to communicate with the public, and engage them.
We believe that ultimately, these ideas will be successful if and only if they are able to become part of people’s widespread notion of legitimacy. If they don’t do that, then they’re imposed by a state, and they’ll be rejected.
If they do, do that, then anything that the state does will have to follow that. Otherwise, people will be upset. That’s why, rather than taking the usual economist approach of, "We just talk to the central bank. We just talk to the IMF," instead, what we’re trying to do is actually build a social movement.
We have dozens of clubs all around the world that are forming around these ideas. We’re working with entrepreneurs to experiment with them. We’re talking to folks like you, who can try experimenting with them in participatory democracy.
Instead of fighting the system, fighting the existing reality, you’re building a new model that eventually makes the existing model obsolete... which is a Buckminster Fuller quote that Fang-Jui always uses in her slides.
That’s awesome. March is parliamentary inquiry period, but I have a way of appearing through telepresence robots, double robotics, holograms, and the sort. I’m happy to virtually be there. There’s also a very quick prototyping system that we work with the g0v movement called vTaiwan.
Basically, every Wednesday, anyone can come up with an idea of saying, "Oh, how about let’s do an experiment this way?" Everybody just tags along. Literally, I think this week’s experiment is the social physics tags from Sandy, from Alex Pentland.
It also is a kind of voice credit, voice token, so to speak, because it only measures the volume of your voice. Although voice is not -- strictly speaking -- quadratic, but if you count the distance, it’s just the proximity.
How loud people are speaking, and how much attention they’re monopolizing, so to speak. We can distribute it more fairly in a physical space. Even if the physical space has its own attenuation parameters, we can change those parameters. It’s almost like speculative...
It’s very interesting what you say, because actually, one inspiration for us calling it voice credits is the physical voice. In ancient Sparta, the way that they used to do the vote, was to try to incorporate intensity of preference, they allowed people to shout in favor or shout against.
One of the things that we did in virtual reality is attenuation design. Right now, with Skype, of course, everybody here, it’s just two of us. We hear each other equally. One of the experiments we did in virtual reality is to change the position by having people physically walk toward the position they take, and then change the attenuation factors.
Because it’s virtual reality, you see, you can normalize people’s input voice level to the same level. Then your position determines the sound dissipation. That’s one of the interesting experiments. I can go on and on.
What I mean is that it’s very easy to prototype new ideas, including QV, through the vTaiwan meet-ups. Even the national petition mechanism, if we can find there’s a beta version, there is a beta website, we can also test this dynamic out on the beta website, anyway.
We tried on the beta site for a year for the visualization of the budget of the entire national budget of more than 1,300, actually, governmental projects. Because the ministries in charge were afraid that if they let everybody literally see the relative budget allocation, how it’s being executed, and all the 1,300 cases, that they will be swamped with comments.
We only tried an initial pilot with 65 national priority projects before we let people see that actually, responding publicly has a lot better properties. For things like social housing, which everybody cares about how well we’re doing, and how exactly, which procurement and spendings went on, people won’t waste each other’s time.
They will actually ask quality questions. Once you respond to them fully and in public, everybody just found them through search engines. The respective authorities don’t have to pick up phones, each one not knowing 550 people have asked this particular question before.
It saves everybody over time, amortized, but in the beginning, we have to put it on the beta stage to show to the competent authorities in all the different 34 ministries that this is going to be a time saver and not a time-waster for them.
I think that’s a great...Yeah. I would like to do something like that with QV. If you guys have relevant developers, and you can do it on your own, that would be great. We’re happy to consult. Also, there’s now a whole movement.
We have hundreds of people who are organized around these ideas. Probably thousands, in the Ethereum community and different communities. We’d be happy to find people to collaborate with you, if you need support in trying to build a prototype.
The last thing I want to show is that this is the official participation platform, including municipalities, the corrective and auditing agency, and of course, the administration of Taiwan is in join.gov.tw.
If you change the O to a 0, as is customary with the g0v movement, you get into the shadow government, which is join.g0v.tw. Anyone can just leave their email address there, and join the Slack channel on the g0v movement.
At the moment, I think it’s 4,000 people or so. I’m happy to donate a GQV domain and see what you guys can come up with, and basically prototype the join.gov.tw system with a re-imagination of the QV system. We’ll see how far we can take from that.
I really appreciate it. Thanks so much. If anyone is watching, and is interested in being involved, we’re working on this in Taiwan. The movement’s called Radical Exchange, and you can find me @glenweyl on Twitter. We can be in contact.