I once made a large part of a reddit substitute with a couple of algorithms for doing queries over intersections and unions of user categories (your "communities"). The data structure is called a SetTrie. We would do well to remember its name.
I can easily imagine wanting to target queries at any of (bus drivers ∧ new Yorkers) (EG: organizing union activity), (bus drivers) (EG: Organizing the funding of some assistive piece of software that all bus drivers everywhere can use), and (new Yorkers) alone, so I don't think the heirarchical structure is always adequate. Imagining Earth and Mars sharing software.
Thinking about it... society still really needs a new reddit. The current one has some huge problems that make it inadequate for the functions its been assigned. I should probably write up a concrete proposal at some point...
I am not sure i understand your middle sentence, but if i did then the system i proposed allows it. can you explain again?
It's probably not important.
I'm concerned about
it reminded me of set theory, but thinking about it more, it ended up merely resembling it
When did it diverge?
Can your system express sets that have multiple parent sets? Can israelifarmers be inside of both earth/middleeast/israel/farmers/ and in work/primaryindustry/agriculture/horticulture?
I think in the design of systems like these there's often a tension between tag heirarchy and tag intersection as a way of talking about increasingly specific categories, and intersection should be used more often than it currently is. Under intersection, as long as the "israeli" and "farmer" categories exists, "israeli∧farmer" category exists implicitly as a subset of israeli and farmer, and there is no ambiguity as to where it should "go".
Can your system express sets that have multiple parent sets?
Yup, that was the idea. a strict hierarchy wouldn't be an accurate map of the communities landscape. I think you would probably get some really weird nesting in some places (especially if we're talking about individual users), but as long as it works and is intuitive to the user, it's fine.
Consider first what worked and what didn't work historically for coordination as a means to move between equilibria. Revolutions/uprisings/unionizing/boycotting sometimes worked and sometimes didn't. What determines the odds of success?
I don't think i have an answer to your question, but something related that i thought of, was to have a guide with game theory and behavioral economics to coordinating actions on the site, and maybe even a team that helps people optimize their CAs.
did you have specific thoughts when you asked that question? I'd be glad to hear them
"that's a lot communities..", you say? "like, a ton of communities", yeah, that's true. but that's how it should be.
I'm interested in hearing more about your thoughts on this point.
My intuition is the opposite - that in general there should be very few communities, and a person can realistically only belong to a few of them. I feel this way because I am strongly convinced that in order for a community to have any meaning people need to invest energy and attention in them over time: a community is a unit of action.
If there are dozens of communities, how will a person become familiar with other members, with whom they are coordinating to change the equilibrium? How will they establish that they really share values? How will they build trust? How will they determine who does which task, and who can be counted on to do the task for which they volunteered? I expect there to be minimum amount of wrench time to get these things figured out per community.
I can see the sense in what you're saying, here's where I'm coming from -
I'm aiming to open up as much possibilities for coordination, so i also want communities to be very versatile (When a CA is made for a community you're a part of you get notified).
I expect most communities to not being very active (most don't try to initiate massive changes all the time). you probably won't have to devote attention only to some and not be able to afford attention to the rest.
now, think of all the communities a person is actually a part of, not just the ones he identifies with - through your geography, workplace, hobbies, lifestyle, areas of interest, social circle, and maybe more - it becomes quite a bit (but In second thought i might have went up too much with saying "upwards of hundreds", it's a little hard to quantify before actually doing it). and still, all of these can be important for you (and you for the communities), at certain times and for certain actions. think of it as trying to have a community map that matches the community territory. I think that more resemblance will open more possibilities, thus making it a more powerful tool.
Another thing i thought of is to let users rank communities by their activeness or how important it is to them. and that will both change what they're notified of in the community, and shown to other members in that community, so it's easier to know who's in the community's core.
So to answer your last paragraph - I don't think everyone who's going to commit to a CA has to be familiar with other members of the community (It depends on the situation), eg Vegans may join a CA with other vegans even if they're not into activism. I think the second question is interesting but a bit more complicated. for the last two questions - a lot of the trust should come from the structure of the website (like in KickStarter), else you don't only rule out many communities, but also large communities.
What you made me think of, is how much other tools should this website give to communities, since they already exist on it
So trying to restate your viewpoint, would it be fair to say that you see communities as essentially the recruitment pools from which people to coordinate with are drawn? So the maximize-communities strategy is maximizing the contact surface for coordination opportunities?
resources Hunter Glenn recommended to me on the topic:
I'm not sure I approve of the social contract in an algorithmic tutu. That's what I'm getting from this anyways; An algorithm to recommend who should partner up for group projects. I'd love to have a magic bullet too, but an implementation without clear justification sounds similar to brute force search to me. Force may be necessary anyways, I just want to clarify my first point.
It seems your website is assumed that for a new project to succeed it's possible to define a few clear actions and if only people can effectively to commit to those actions, the project will succeed.
My model of the world is that for most projects a like of what's necessary to be done won't be known exactly ahead of schedule. They are better solve by people coming together and building a shared organization where they don't know beforehand what they will have to do.
than it seems i haven't explained myself good enough.
I didn't mean that you create a project or some long term goal, and ask people to commit to the whole plan thinking that this will be the solution.
Rather that you open a new coordinated action for each action that you want to be taken simultaneously (or a small bundle of action that need to be obligated to together). no need to try to foresee the future and pre-plan everything ahead.
Why do you believe that the problem of getting people to take an action simultaneously is a central problem that has to be solved?
It's my impression that most projects need coordination where people talk together about what needs to be done and then see that there are people with the necessary expertise, connections and resources to do the task.
In many cases it's also not possible to commit to getting certain results.
If I hold a LW meetup then a person might decide to ask their workplace whether we can hold a LW meetup in their rooms.
If a government bureaucrat is a member of my community that bureaucrat might work towards us a government permit to do X, but that's not something the bureaucrat can pledge to do.
Well, i wrote this article with that premise without justifying it as it's generally accepted on lesswrong. as i said, it's based on problems from game theory that we know also happen many times in real life, and on yudkowsky's book inadequate equilibria.
the structure of the site doesn't require you to commit to results, but to actions. if we're taking your example, then the bureaucrat commits to do certain actions (probably ones that are believed to be helpful in getting a permit), not on actually getting it.
but it's not actually a good example for what this site would be about. it's not about small groups or complex projects, but about simple actions taken by a lot of people. humans can coordinate effectively in groups of up to roughly 150 people, above that it doesn't work as great. If you want to see better examples, look at this post.
But here a tailored example for you -
Say you have a organization with the goal of making the scientific process better, for most of your work (research, discussions, negotiating, surveys, etc) you don't use this site. but when you want to implement the solutions you found, you find that it's quite hard, there are many gears in the mechanism (researchers, grant-makers, journals...), and one won't change with out the other changing too, at the same time. so you open a coordinated action on this website, something simple and not output-based, and people commit. if enough people committed, they all take the action at the same time. much less effort and strain, instead, you can smoother transitions.
Of course i can't promise that it works that way, and i can't quantify beforehand how beneficial it'll be, but to me it seems promising.
There's nothing in inadequate equilibria that claims that the cooperation that's needed doesn't require complex coordination.
If you take the example of nutrition for babies, it require coordination between different institutions and isn't just a matter of a lot of people deciding to do the same thing.
humans can coordinate effectively in groups of up to roughly 150 people, above that it doesn't work as great.
That's not true. While there are companies that don't effectively coordinate, a company like Amazon is able to coordinate a lot more then 150 people in an efficient way.
Outside of companies, there are also unions that can effectively coordinate for the benefit of their members with more then 150 members.
Of course i can't promise that it works that way, and i can't quantify beforehand how beneficial it'll be, but to me it seems promising.
It seems to me fairly straightforward to do this with the old technology of Google Forms and email without a new crowdcoordination website. It will require a bit more work for the person who organizes the coordinated action, but if you believe in the concept it would be the minimum viable product that's worthwhile to do before building the website.
A number of websites (such as Facebook) are valuable because of the people there. As it is (usually) useful to a community/group/etc. to communicate using the same channel, channel switching is an activity that makes more sense to do all at once. (If everyone except you left Facebook, would you still use Facebook?)
I don't think that's an effective way to think about Facebook.
Whether or not I use Facebook has little to do with whether I use Google+. There's no need to leave Facebook to use a new social network.
We switched our local LessWrong community to using Slack by gathering with 5 of the people who do the most in our community for a weekend where we decided to move the community to Slack and afterwards various people who attended took actions they believed would be helpful for establishing the Slack channel.
When I announce events then I crosspost them over multiple channels.
I agree that the type of action needed to address the parenteral nutrition example would be very unlikely to arise from a hypothetical online crowd action platform.
I think much of the purpose here is to brainstorm what the most effective version of such a platform could look like, and thereby also get a picture about what class of problems such a platform might be able to address or help with.
Is your position that that class of problems is small, and/or those problems are better addressed by existing solutions anyway? Maybe that's right. We are largely brainstorming here to gauge whether a marginal improvement is possible (and probably hoping that a big improvement is).
Is your position that that class of problems is small, and/or those problems are better addressed by existing solutions anyway?
The philosophy of the lean startup suggests that one of the best way to learn is to start with a minimum viable product. In this case the minimum viable product isn't building a platform but to run a campaign and run it with existing tools in a less automated way.
I think much of the purpose here is to brainstorm what the most effective version of such a platform could look like, and thereby also get a picture about what class of problems such a platform might be able to address or help with.
Depending on the problem you want to solve with the platform, the website might look quite different.
There are problems where you would want to be able to organize civil disobedience and need to be resistant against attempts of the government to shut your website down.
If you want to organize 150 people to come late to school to push back the starting time of the school, it might be important to keep the identity of the people secret until the project is finished. In the case of founding a new academic journal on the other hand you likely want the identity of the people who already committed to the new journal to be public.
I haven't said IE claimed that, but that "getting people to take an action simultaneously is a central problem that has to be solved", and he does hint at a kickstarter-like site for coordination, so it's not far fetched from the book.
pay attention that i did give examples of more complicated 'contracts' than "everybody takes action X". the more the site evolves the more options it will have. but that still doesn't reach project management level. it's simple, but not too simple to not be useful.
of course there are ways for humans to cooperate at larger numbers than 150, or else we wouldn't have modern society. or any society for that matter. I meant that it's hard for humans to do it with some kinds of social structures. but it's not so important, we can leave it aside (the important bit isn't the number, but the notion that the more people, the more coordination gets difficult).
I think the minimum viable project is much more than google forms (no need to even bother if it doesn't get better than Facebook and Twitter). and actually there is a similar site, but for me it also doesn't reach minimum viable product. most of what i described in the post is what i think should part of the minimum viable product.
I haven't said IE claimed that, but that "getting people to take an action simultaneously is a central problem that has to be solved", and he does hint at a kickstarter-like site for coordination, so it's not far fetched from the book.
I don't think the difficult thing about coordination is to get people to take an action simultaneously. I think it's a far bigger problem to get people to stick with a certain project for a longer timeframe.
Getting people with different views to a consensus about how a problem is to be solved is another general coordination problem.
yes, I agree that both of these are also large problems of coordination that aren't solved by collective action. we will need to find ways to do all of these better.
I think the minimum viable project is much more than google forms (no need to even bother if it doesn't get better than Facebook and Twitter).
This sounds to me like you ignore the actual problem. The problem is organizing people. The problem isn't the tool to organize.
Do you believe that it's currently impossible to run such a campaign with Google Forms and email and at the same if you put everything on the website, it would suddenly work?
Why do you think people can't effectively commit by filling out Google Forms?
Could you fund a video game with google forms? (assuming it has a PayPal widget, if needed)
There are three main problems i see with google forms.
First, i think it'll be harder to get people on board, much harder to get many people on board, and much harder still to do it consistently.
Second, there's probably a high value to a system that ensures (or at least makes it more probable) that someone who obligated to take an action actually took it. if you don't have such a system, then whatever you did, you're arguably almost in the starting same position. everyone might still be suspicious that no one else will take action, and so nobody will take it.
the third connects to the second, but may be different enough - A platform could gain trust. not only would i not fund a video game through google forms, i may not even want to fund it through certain kickstarter-like sites, cause i don't trust them as i trust kickstarter.
That's not getting into functionality that to me seems important.
Could you fund a video game with google forms? (assuming it has a PayPal widget, if needed)
I see no reason why you shouldn't be able to do that. But for a MVP you might start with a project that's even smaller.
The sign-up for our LessWrong Community Weekend works well through Google form and we are confident to get 150 people to attend.
Second, there's probably a high value to a system that ensures (or at least makes it more probable) that someone who obligated to take an action actually took it.
I think it's likely that it's easier to solve the system of ensuring that as many people as possible who committed to do something actually do something for one particular campaign then to solve it in a fully general way.
the third connects to the second, but may be different enough - A platform could gain trust.
That's a bad reaction to proposing to do a MVP. Whatever you do, a platform won't have trust at the beginning.
At the starty, you likely have to get trust by getting influential people or organizations to endorse your campaign.
1. I'd love to see that. surely, it's not probable that a funding campaign which uses google forms instead of KickStarter would be better off, right? i think we have concluded this line of thought - i'm skeptical, and you don't see a reason why not.
Though, again, 150 people attending a LW meeting is far from the type of coordination this site will be for. i'm not blind to the fact we have better coordination tools than going door-to-door and talking to people - i say that the coordination tools we have today aren't sufficient for more complicated coordination - where there are a lot of people involved, several types of actors, many incentives, etc...
2. maybe you're right. still, it's worth a try and a little effort :)
3. that wasn't my reaction for doing an MVP, but reasoning for a platform being good, and saying that google forms isn't a platform (more specifically, it isn't a platform for coordinating action, but a more general platform, so it can't be trusted as one)
surely, it's not probable that a funding campaign which uses google forms instead of KickStarter would be better off, right? My point isn't that Google Forms is more effective then having
That's moving the goal post. Finished well-engineered products are usually more effective then generalized tools with you can use for an MVP.
That doesn't mean that the way to proceed isn't building a MVP before desgining the bigger platform.
Running the MVC gives you a much better idea of what the actual challenges happen to be.
I really like the idea of milestones, I think seeing the result of each milestones will help create trust in the group, confidence that the end action will succeed and a realization of the real impact the group has. Each CA should probably start with small milestones (posting something on social medias) and ramp things up until the end goal is reached. Seeing actual impact early will definitely keep people engaged and might make the group more cohesive and ambitious.
Regarding communities: I don't think it's necessary for such a website to have a particularly strong conception of communities, outside of each particular collective action. For each collective action, there should be robust tools to make it clear who is / isn't qualified to take part in that action, but this should be handled on an action-by-action basis, and there's not much benefit to having a community structure outside of individual actions. Such a structure would simply add needless effort to the development and maintenance of such a website, which will make it less likely to achieve its mission in the long run.
Hi! Someone sent me this post as it closely resembles what we are building at https://mattermore.io - except that we believe coordinated action to tackle the grand challenges of our time cannot be facilitated within the app. Still, many of the premises of this post + the comment section are very valid.
I'd love to have a chat with you, if you're interested in exchanging ideas!
Here's a short little manifesto we wrote about our work: https://www.notion.so/mattermore/Solving-real-world-problems-A-manifesto-d54473855d014a6d8bfe3676932dcaa5
Hey, just seen this.
Sure, i would love to have a chat if you want. I'll DM you my Email and my Discord, feel free to contact me :)
inspired by Inadequate equilibria, and following the 'KickStarter for Coordinated Action' sequence.
This is an idea-dump post for a website i thought of after reading Inadequate equilibria.
Today, tools like Facebook and twitter help us coordinate better and faster. but still, that is not enough to solve problems of "high-inadequacy" - where we're stuck in bad Nash equilibrium, and moving away from it demands many things to happen together.
The idea here is to take coordination much further, so we can solve as much of the game theory problems bound in moving to other Nash equilibria.
The goal is to allow 'Exoduses' from bad Nash equilibria, to better a Nash equilibria.
Note - none of this (at least as of writing this) is being worked on, nor are there currently plans to do so. Though, of course this doesn't mean there shouldn't be, or else i wouldn't have wrote the post :)
basic concept
big problems can't be solved by a single person. but sometimes even a large group of people who agree on the problem and the solution, and are even motivated to bring the change - can't do it. That is because some problems are more complicated than that.
one such complication is coordination problems - where "everyone is taking some action A, and we’d rather all be taking action B, but it’s bad if we don’t all move to B at the same time." or each individual wouldn't want to take action A unless he knows the rest do too.
solutions to coordination problems include common knowledge (Where all know the intentions of all actors) and pre-commitment. that's the site's goal.
Coordinated actions
Any user can become a coordinator/initiator by creating a coordinated action (CA), in order to solve some problem. Each CA gets its own page (similar to kickStarter). Any user (unless otherwise specified) can obligate to the action - and here's the catch - the commitment is to be realized if and only if a certain number of other users obligated the same. In KickStarter, we commit to pay, here we commit to take action.
Versatile coordination:
Since reality is very dynamic, a rigid structure will work for few situations and be less useful. So to allow a large array of projects, options for CAs should be versatile. I see it as an ongoing project that will develop next to the community's needs. think of how tesla is implementing features in dialog with their community's wishes. here are some examples of "contracts":
Basic: all of us obligate to some CA if X others do too.
milestones: many actions, listed on the same initiative, which are taken at different amounts of obligations. It is either required to obligate for all actions, or possible to obligate to specific actions only.
Obligation for obligation: a group or individual obligate for something, if a different group or individual obligate for the same or a different action (only one side can be an individual).
Please comment with more types that you can think of, it's very interesting.
further support
And sometimes just committing isn't enough for you, or you support the initiative but done that action (in cases of one time decisions, like going vegan, going zero-waste, getting rid of your car, etc...). So we want to give users a way to support further than just obligating. One example is an option for users would to donate money to the campaign, which will be used to further spread it, through some kind of advertising.
Communities
This is where it gets quite complex, but it has to. people don't want a thousand random humans around the world to something with them. they want people from their country, from their city, from their profession/hobby/Interest-area, from their social circle or organization.
That's where communities come in, the ideal is that for every real-world community you would be able to create a community on the site to resemble it, and that if people from that community are already on the site, they will find out that a new community they belong to has been opened, and will join. the reason is, so it's possible to coordinate action with and within certain communities.
When i first thought about how communities will work in practice it reminded me of set theory, but thinking about it more, it ended up merely resembling it (and probably breaks some of its laws). still, hopefully mentioning it helps to visualize.
the communities structure
There are many communities, it's easy and accessible to create new ones, or join existing ones (Unless said community has some requirements). any community is a sub and/or parent community of other communities. The goal is that the communities on the site will be able to reflect the communities in the real world.
some examples of communities: The earth (The parent of all communities until we colonies mars) and the user base[1], all regions/continents, and all countries, are some that can be added per-launch. examples for user added communities: cities, EA, farmers is Israel, bus drivers in new York, vegans is the US, LessWrong, etc...
"that's a lot communities..", you say? "like, a ton of communities", yeah, that's true. but that's how it should be.
it might sound a bit like FB, but except the community grouping aspect, there are two more important differences. here we don't want two groups which are basically the same (Cause that's ineffective, if the real world community is divided between them on the site), and the user doesn't have to be aware of all the communities they're part of[2], which may feel weird in a online platform, but that's how it is in real life.
Whenever a user joins a community he is suggested sub-communities he might fit in, and is automatically added to all parent communities, which he can manually exclude themselves from (I bet the set theoreticians winched). part of the account creation process would be spotting the user's communities (easy ones, for example, are countries and cities)
organizations
not only individual humans are "agents" in this world, but also some communities, like, corporations, non-profits, and any other goal oriented group. organizations will be able to create a community around themselves, but also act as a user on the site. this is important, cause some coordinated actions seek not only the cooperation of single humans, but of the groups they make. if vegans think of taking a coordinated action together, they want to know that the businesses and organizations they support will go with them too. some CAs will be relevant only for organizations, where it doesn't matter how many individuals commit to take some action, they have to have cooperation from their own group.
Motivation and verification
At least until all of our society is nice, educated and rational - We need two mechanisms, one that verifies who really cooperated and who defected, and a mechanism to discourage defecting, and encourage cooperation.
fulfillment verification
Verification is the much harder of the two. how do you verify that someone tried veganism for 28 days? how do you verify that someone has/hasn't posted certain types of posts to FB? Can you verify whether someone really voted third party? whether they went to work that day? Some stuff are easier to verify, but if we stuck only to actions that are easy to verify, this tool won't be very useful. It's a hard-shell to crack, but it needs cracking - ideas?
user verification
In this system it's important that we know user = person, even better if we know user = which person. It can also help with fulfillment verification. There are many ways to do it - Email, phone, face, PayPal(?), ID. it just needs to be. maybe not all users have to be verified, yet still possible for communities and CAs to require verification.
Motivation
given that we solved verification - motivation is simpler. a few options:
cooperator/defector score: users have a publicly displayed score that shows how they acted on their obligations.
Achievement badges: I envision something similar to khan academy's badges, but harder to get so they're more meaningful. an example of one "good to have: on average, you referred to each coordinated action you obligated to, at least 10 fulfilled obligations." You can display these badges on your profile to signal how awesome you are :)
putting your cash where your mouth is: for each CA, either there's a set amount or the initiator sets it, each user deposits cash against his cooperation - if he defects, he looses that money, if he cooperates, if he wins some extra money. This is a pretty much bound-to-work motivator (unless bill gates starts using this too), But i'd rather incorporate money as a last resort, If we find that we really need this extra motivator, since it makes everything more complicated. It also makes the verification task harder, since if people can use the site as a money-pump, they'll be more likely to look for a backdoor to exploit.
I believe this concept, if it was successfully realized, could bring great benefits to the world.
yeah, both are all users, but i see reasons to differentiate, sorting wise - communities might have a counter feature of how many people in the real world community are registered to this online community, in the users it's a 100%, in the world.... you'd target different CAs to the user base and "the whole world". and, Sub community sorting is based on the real world, not the website, so the farmers of Israel are only part of the earth community and not the website's User Base. possible sub-communities are active users, contributors, etc...
Hard to estimate exactly how much, but it's at least a few dozens and maybe more than a hundred