 I was hoping it's late in the day, we could all take a moment to brainstorm together, actually. So the start of my talk was just encouraging everyone to bring to the table a commons that you would like to see, a public good you'd like to bring into the world, and bear in mind that basically anything can be a public good if we make enough of it. And at the same time, anything that we rely on as a commons or as something that has always been publicly shared can be enclosed and can disappear. So just run through the thought experiment in your mind and think of things that you aren't necessarily trying to build yourself, but that you would love to see. And then if there's some prominent idea that come up, we can workshop some of them later, we can take time in one of the other rooms. We've been talking a lot today about how to fund the commons and how like incentive design, ways to encourage collaboration or to support collaborative work. I mainly wanna talk about how you scale a commons and what the right scale is. Most things that we think about like public space and fresh water and fresh air, we don't actually have the infinite, the ultimate version of that commons in the world. We have a version that sort of has fit into our infrastructure that we've maybe built, spent some centuries building civilizational structure around so that there's a particular stream of that commons that's inexpensive and that we can maintain as a society. My computer's going on an adventure. So let's think about how to scale commons and the things that we can do that reliably expand or contract the distribution of a public good, the scope and the strength of a commons. There's a little bit of tension also between developing built things, creating new things and developing commons that especially commons that persist over a long period of time. We are here inside an amazing commons, this museum, it's been around for a couple of hundred years, I think at the oldest museum in the city. And it was here when Central Park was built. Central Park wasn't just there. There were other things in Manhattan. Oh, thank you. Hi. So there were things in Manhattan and then there was this comprehensive change where people said, what if we had a massive Central Park? 20,000 people worked on this project. So now we're inside one commons, this building. We're on the border of another commons inside one of the most built places in the world. And that tension is what I want everyone to think about. I'm S.J. Klein from the Knowledge Futures Group. And I have broken a mirror at some point in the recent past. No worries. And I want to start by talking about society, right? We live in society, we live in a commons. There's a planetary commons within which there are commons of different sizes. And the process of maintaining a commons, right? Commoning itself is an equilibrium. It's affected by social norms. It's affected by what we think is reasonable for a society to do. So much of the uncommons is something that we invented, right, to enclose, to financialize. So we're here talking about funding a commons, but a lot of the funds come from enclosure. So here's a little survey. Just take the commons that you want, whatever you're just thinking about. Go to this URL, enter that code, describe in a couple sentences what you would love to see common to being the expanded, be preserved. You can be creative. It doesn't have to be things that we traditionally think of as public goods. Commons develop in cycles of expansion and enclosure as new inventions or discoveries come about. The things that we can do at scale, the cost of maintaining public and private goods changes. And the enclosure itself is part of a dynamic system. If we ever stopped living in Manhattan, the city will be wild. In the meantime, we're here building commons inside this built structure in a global commons. So how do we design for the right equilibrium? This is a system question, equilibrium design question. And you think we've been doing this for a few thousand years at least, we should be pretty good at it. So how are we doing? Well, we just had probably the most kumbaya event in recent generations. We had a global pandemic. Everybody came together to try and address it. And we probably spent tens of trillions of dollars, years of global product for a number of countries. Did we get any public goods out of it? Vaccines? No, here's the list of all of the vaccines that have been accepted in any country. Corbavax is the only one that might be patented and encumbered, still not really accepted even in India. And they only managed to raise $7 million to do their work. Dozens of other vaccines, all of which were patented. Testing? No, limited labs, little pooling. Masking? Did we seize the tools of mass production? I don't think so. Masks are still some multiple of the marginal cost of production which hasn't changed since before the pandemic. So social norms have an outside influence. We accepted that the way to all come together and solve this was to do things other than building public health goods. Here IP is really the thing that held up that idea of building a commons. We'll come back to that in a minute. So let's talk about scaling. The nature of our public good changes with scale. If you've been brainstorming in this survey, think about the scale you have in mind when you say you want a good to come into existence as a public good. We should consider current advances, past approaches and failures, the ecosystem of creation and maintenance, and forecasts of what its full potential could be if everyone adopted or implemented the good. Do we have fully automated luxury commonism? Not yet. But let's think about some examples. Epidemiology, we were talking about COVID. We totally failed to produce public goods for COVID, but there are some amazing commons. There are some amazing public resources, public compute equivalents that people have been using to sequence the genome of every instance of the virus through next strain. Has anyone used next strain? Awesome, beautiful service. Again, totally undervalued by society. I mean, I'm sure they have most of the funding that they need, but this is a great example of something which could be implemented in a hundred times in many cases. We could just be sequencing every 10th test. We don't, but now the cost has gone down enough that we could reasonably do it alongside, we're all paying some multiple of the costs of the tests. Same with the human genome browser developed by a number of people after the human genome project was more closed. You can browse the entirety of an aggregate human genome in extreme detail. Public data, does anyone use our world and data? They're probably the most popular in our community, but there was Google Public Data Explorer. Inside Google, they failed to take off. Now there's the Google Data Commons. These things have great power, but people haven't figured out how to convert them into a lasting and use public good. Finally, large language models. We have OpenAI, we have GPT-3. Initially OpenAI was thinking about potentially being more open, now it's become less and less with time, and so these more decentralized organizations like Aluther and Big Science have picked up the slack and are producing the new language models. But it's still not clear what kind of public good the resulting model will be. Let's also talk about commons failures. We need to know what the failure modes are so that we can avoid them. Public goods that we rely on can disappear. They can be captured and closed. It's actually easy to build in anti-commons if all you're doing is studying incentive design and you build too many cool incentives. If they conflict and the incentives include veto over the use of something, now there's good luck. You can end up with a system that nobody uses. Once again, IP is eating the commons in this regard. And in parasitic uses, if people aren't maintaining it and other people come in and make use of it for something no one likes, people might decide maybe we didn't want this anyway. It wasn't really such a good commons. And eventual loss of interest. In general, exploits are common if people don't treat a commons as a maintainable public resource like a park because it's easy to see it as a game and say, oh, there's a commons there. If I can figure out how to use the commons as an input to my business, my business becomes much more effective and you can sort of drain the commons of what made it useful to society. A great idea if you are thinking of building a commons yourself is just start by cataloging all the failures. People have probably built the same thing before. Really diligently find out why they failed. Dan Whaley, who founded Hypothesis, did a lot of work and open annotation. When he began Hypothesis, before he decided to dive in, he made this amazing spreadsheet still online of the 120 failed annotation startups. Everyone should do this. So a few examples. Mail. Has anyone lost a phone number or a snail mail address because you were forced to move? And you weren't able to get phone or mail forwarding? Has anyone lost an email address and not been able to get forwarding? Happens to me. That's ridiculous. The cost of mail distribution went down by a factor of a million, maybe more. And what we got was we lost the post office. There's nothing like the post office today. In fact, the rise of domaining as something that sat on top of DNS really ruined many aspects of addressing and now we're forced to start thinking about self-sovereign identity. Not the only reason. Homelessness. You might not think of homelessness as, you might not think of housing as a commons. It's a very rival good. But homelessness is absolutely an anti-commons. In almost every part of the world, the cost of just housing, people who are homeless, is a fraction of the welfare and social costs already being spent every year to support the homeless. But first, what's the most expensive housing market in the US? Anyone know? Close, yeah? Close, I'll give you a hint. 4K for a studio with shared bath and a roommate who you don't know. You're very close. If you lived here, you'd be in jail. The prison system charges something like three to 5K a month for people who are in pretrial or regular detention. However, prisons are often turned into luxury hotels because they're like big, settled buildings. So I guess the hotel industry is actually the highest rent market. Scholarly publishing, I'm preaching to the choir. It's completely broken. It's basically a canonical example of centralization that seemed like a good idea and went wrong. And it felt like a commons creation at the time because every scholarly network was doing their own publishing and they wanted a shared public source of this work. So they outsourced it and they didn't have a model for how to make sure that remained a commons and didn't then become a rentier who when everything became almost free thanks to the web claimed 95% of the benefits of being on the web. Okay, so a quick hit case study. We have learned some things. We do know how to make commons that are joyful. Encyclopedias for millennia, well at least half a millennium, have been a beloved commons with the understanding that there's sort of a group of people who need to come together and spend long, long time building the encyclopedia. Like the first version of CERN wasn't particle physics. It was people summarizing knowledge. And the thing is over the distribution of the results and access to the results were a little bit dicey. The young encyclopedia was the largest ever produced until Wikipedia was seven or eight years old. It was produced by an emperor who did it telling his people, this is going to elevate us. This is going to make us a great scholarly society. And then was criticized by all of the researchers across the country because only one copy was made. There were some fights. It was hard to make another copy. Basically the emperor said, okay, I have one copy. We ran out of funding. We spent 10 years compiling the encyclopedia. We just can't fund a second copy. So there was one copy for a long time. Only 4% of that encyclopedia is still around. Someone made another copy 150 years later and basically they've both been burned, stolen. So bring this to Wikipedia in 2000. The initial goal was really simple. And this is the other thing I want you to take away when you're imagining or scaling your comments. Think of describing the value of a simple, very attainable version. You also want to imagine what will happen when it has scaled and adopted by everyone. But for Wikipedia people said, look, we're just going to make us thousand good articles. And that will be amazing because, you know, FSF had also said something similar. It's just really hard to get access to information about the essential topics in most languages in most parts of the world. Gotta spend some time attending the information fields. There's an understood gestation period. And then Wikis came around the second year too. The initial new pedia project was producing 40 articles a year. And people were like, okay, in a generation we'll have our thousand articles. That's pretty good. But here's the new thing. There's clearly this sea change of people just generating all kinds of material online. What if we try to Wiki? It's gonna be bad. We had, there was a nice seven step editing process. We're gonna lose that process. Let's see what happens. Within a year that was up to 10,000 articles a year. Now we have 50 or 60 million articles. But now things are slowing down again. So these are just examples of walking through some of those patterns that people thought about when they were considering what Wikipedia was going to be. And committing to it being decentralized, forkable, and visibly generative. Not committed to the expertise of a particular editor. So we're in the middle of a revolution. Machine learning, web three. Distributed energy across time and space that doesn't require, it requires new kinds of coordination, but not all kinds. Never waste a revolution like that. In the middle of this kind of revolution is the time to figure out the equilibrium that we want to move towards. And then use some of the energy while everyone is in disequilibrium. Use some of that energy to move towards something that we want. In particular, still talking about knowledge commons. Knowledge is increasingly executable. There's this growth of knowledge generated by models of the world, trained on models of the world that is just going, it's already dominating knowledge production. And the next encyclopedia is gonna be full of that. Wikipedia in its current form is not going to be that. So we need to either make it that or make sure that we have healthy alternatives. And we've been using this example from a knowledge commons because knowledge is sort of a meta commons, right? We need knowledge about the world to understand what's possible, to understand what other people are doing. So the spirit of encyclopedism of cataloging what's known and figuring out ways to learn new things is relevant to a lot of other arenas. And machine learning is something like an abstraction of that spirit. What we do now is gonna decide what happens with machine learning outputs and the near future. So let's articulate the future we want in order to make sure that we fund the thing we think we're funding. So coming back to the initial question, provocation, inception, revisit the question. Now, what kind of commons do you wanna see in the world? What public goods would you like to see? For the public goods that you work on, how would you like their scale or shape to change? In some cases maybe there's a little bit too much of a good thing. Enclosure is not always bad. Sometimes enclosure in order to, like maintenance involves a kind of enclosure. You just have to be very careful to be aware of the failure modes and the advantages. As it's not necessarily a lot of work, but it takes a constant vigilance and reflection of how things are working. What should we build? What should we extend or preserve? You'll hear from Zargum in a minute about algorithms as a public good or at least thinking about algorithms in a new way. To what extent should algorithms in general be a public good? We have this long history of copyright. Going back to the secondary theme that IP is eating the commons, which it is. Copyright was initially just for books and then it got extended to all kinds of other things. It got extended to databases. It got extended to code. Easy to imagine extending it to algorithms even though right now the definition of an algorithm is generally not protective. This is the moment when we need to take in hand the value that we see in a potential future and use that to make sure that we don't diverge. I would love to workshop some specific ideas into implementable or expandable proposals for building commonses. So if you're interested, find me after the break. Find me tomorrow and we'll make space here. Thank you.