[This is a crosspost from LessWrong, reworded slightly to fit the EA discussion better.]

In 2017 Eliezer Yudkowsky wrote:

Coordination isn’t as simple as everyone jumping simultaneously every time one person shouts “Jump!” For coordinated action to be successful, you need to trust the institution that says what the action should be, and a majority of people have to trust that institution, and they have to know that other people trust the institution, so that everyone expects the coordinated action to occur at the critical time, so that it makes sense for them to act too.

That’s why we have policy prediction markets and… there doesn’t seem to be a word in your language for the timed-collective-action-threshold-conditional-commitment…hold on, this cultural translator isn’t making any sense. “Kickstarter”? You have the key concept, but you use it mainly for making video games?

Gabcoh was asking for a similar website in his post titled "Coordinating Commitments Through an Online Service". So does this website exists? Ehhh, sorta? I'm missing certain features I would want from such a website, so let's review what already exists and see how we can improve it.

CollAction

The closest thing I could find is CollAction. These people have gone ahead and build a functional website with actual projects on it. The makers of this site call these kinds of websites 'Crowdacting websites'. Since 'inadequate equilibria' didn't name this idea I shall be calling it that from here on out.

The site allows you to upload your own projects (with their approval) and allows you to join other projects. The visually clean lay-out shows how many people have already joined a project and what the desired threshold is. The projects have clear deadlines and goals and once they are reached they are closed forever. I encourage you to check it out for yourself since it's a good way to tickle your imagination about what such a website may look like in the future.

While the site is very pretty, it's not made by this community and as far as I can tell also not used by this community. Not very surprising since it's very barebones. Let's run through a list of features I am missing that I expect to see on a Crowdacting website.

Expected Features of a Crowdaction website

Broad range of projects

The site currently only has projects about crowdacting ecological problems running. Your project also needs to be manually approved by them and it's not entirely clear what they will/won't accept. I'm guessing some level of review is going to be necessary, but there is no reason a crowdaction site should be so narrow-scale.

A Milestone system

Yoav Ravid suggested we use a milestone system similar to kickstarter. This wouldn't work for all types of projects but could certainly be included. He also suggested badges people can have next to their name if they e.g help push over a milestone or make good on their promise.

Talking to humans

Their site has no real defenses against bots. When you promise to protest something if a thousand other people will show up too, you don't want a thousand bots to sign up and make you the only person to show up to the protest. Something like it's me should be able to keep out bots and make sure the collaborations is always between actual humans.

Gated communities

When you're trying to organise a strike in Amsterdam you don't want people from New York to interfere with your collaboration project. Those projects should be closed off to the wider world with only community members being able to join in. Some way to signal to an algorithm what kinds of projects you would be interested in joining is also nice and communities help with that. MakoYass proposed a community system based on SetTrie. I would couple this with community chatrooms so the members can better discuss things amongst themselves.

A registry

For some projects you want to contribute anonymously, for others it's important that people know you support it. So I would add the option to commit anonymously, but the site still vouches for you being an actual person. On the project page should be a registry of people who are committed to this project with the option to upvote and downvote these names. Next to the names people should be able to write a couple words summarizing their influence over this problem. This way CEO's of relevant businesses (or other relevant organization leaders) will be able to get upvoted so that people see what organizations are also joining in.

A voting system

Sometimes people agree that the current equilibria is bad, but they don't agree where they should move to. On the project page should be a STAR-voting feature, where people can submit their equilibria of preference and vote on the equilibria they want to move to. For example: hospitals have different administration systems and want to use the same one. Most hospitals use administration system 1 so administration system 1 comes ahead in the voting ensuring that the transition inconveniences the least amount of people. This voting mechanism should probably close well before the project itself closes so it's always clear what people sign up for.

Complying with the law

I know this could be a tool to protest stupid laws by e.g joining together to start smoking cannabis in front of the White House. I would urge that the first site does not condone the breaking of laws, since that might kill the reputation of crowdacting and might get those types of sites banned by governments. (this is something CollAction thankfully already does) This doesn't have to be as bad as it's sounds. People will almost certainly find a way to work around this limitation. For example "Let's all smoke in front of the white house to protest strict cannabis laws" can stay on the site because smoking in itself isn't illegal (they could be talking about smoking tobacco).

Letting different people do different things

This kind of goes together with the registry. If certain community members can't participate in certain ways but can in others, let them. Have different counters running on the project e.g: one counter saying 320 out of 500 have signed up to do X and a second counter saying 100 out of 200 have signed up to do Y. Yes this makes it more complex, but in real life you often need different types of people to do different types of things. This could be incorporated into the registry with flairs like: Mr. Robin Handsome will do Action 1, Action 2 and Action 5. This allows for more complex coordination problems to be solved.

In case of failure, follow-up suggestions

Let's say you want to start a project titled "Change hospital administration systems in the U.S". You start the project and it get's some traction but ultimately it falls short of the threshold. Ideally some algorithm could pick out which sub-communities were motivated and automatically suggest the same project on a smaller scale. If the algorithm noticed that people in California were committed to the project, it could suggest a new project titled "Change hospital administration system in California". This can keep the momentum of a project going. People could also do this manually thanks to the registry, but they might not think of it.

Fulfillment Verification

The hardest problem might be to make sure the project actually goes through and everybody does their part. While this might be impossible to check on an individual basis I think I have thought of something that might broadly solve this in practice. Everyone submits 5 dollars to the project. Some time after the end of the project (depends on the project but for most 48 hours will do) you get the question: "Did this project succeed in it's mission statement?". If you say 'yes' while more than 60% say 'no' you only get part of your money back. The exact amount of money you get back depends on how big the gap was between what you said and what everyone else said. So someone who said 'yes' while 69% said 'no' will get way more money back than someone who said 'yes' while 99% said 'no'. The same is true in reverse if you say 'no'.

This means you try to predict what everyone else will say which will usually correspond with what actually happened. (You also get your money back if you say 'yes' but less than 60% say no just so there's a margin of error). Because this means that overtime you will lose money on the platform you can actually win some money by repeatedly being in projects that then end the way you said it ended. This money will come from people who didn't say it ended like the rest said it ended.

Getting this project off the ground

I'm honestly kinda sad that this site hasn't been build yet. I would if I could, but I simply don't have the expertise in webdesign to pull this off. But if anyone starts working on it I will be more than happy to help in any way I can (graphic design, translation, marketing...) I do however have a suggestion as to how we can funding for this website. Use kickstarter! Use a crowdfunding website to get a crowdaction website off the ground. I will personally pitch in of course, but I think a lot of people inside and outside this community will be interested too.

I feel like this idea has a lot of potential, but I would like to start working on it sooner rather than later. If anyone has any more suggestions about what a crowdaction platform should have, please put them in the comments and I might add some to the list.