Imagine if you could build a Tumblr that runs itself and pays the people that write blogs on it? The bigger the site becomes, the more worth your contributions become. It also means no ads need to be run on the site.
This is the concept of a content-producing decentralized autonomous organisation (CPDAO). The Cypherfunks was an early attempt at that (which we are still busy with), but I believe the parts can be improved to more succinctly introduce feedback loops. I’ve spoken to various people (The Cypherfunks, guys from Ethereum and Wendell from Humint), and have put together some thoughts on how this could work. There is still a part or 2 missing, but I feel if I write this up, and get more minds onto this, we move towards making CPDAOs. Organic, emergent networks that don’t require a central entity to run and pays out the whole ecosystem, instead of one company that captures all the value.
The basic features of a CPDAO is that it should host the content, distribute it and incentivise people to contribute to it. CPDAO’s content is dependent on how the tech is set up & how the network expects to interact with the content. In other words, different CPDAO’s could eventually be built for various content. A group of people making music together (The Cypherfunks) or people working together on a blogging network (ala decentralized Tumblr). A blockchain (not necessarily, it could work with something like Ethereum hypothetically) along with a token forms the core of the system. The token serves several purposes: as quantification of the network (like Bitcoin), payment to keep it running and is required to add content to the CPDAO. Additionally, a “voting” system seems like it is required to incentivise content that isn’t bullshit.
So. I’ve thought 2 ways this could possibly work. I hit a technical roadbloack, so I’m not even sure if it is possible. I need to do more research, but hopefully if someone more technical reads it they can add their ideas.
Version 1 (New coins go to hosters of popular content).
The important part of a CPDAO is: who gets the new coins? And how can you do it cryptographically unbreakable? In version 1, the new coins go proportionally to the hosters of popular content. You become a hoster, by voting for the content you like. Voting for content is a tip to the content-producer. By tipping, the software takes the content and you help store & serve it from your computer (as a usual p2p network).
This means. If you tip the right content (content you think is good and popular for the network), you host & store it. If you serve the most content, you get the biggest reward (from the issuance). This means: Content-producers gets rewarded through tips as the “hosters” want to make money as well. The hosters thus play the role of curators as well: hoping to find good content to host, as serving the most popular content, nets the biggest rewards. Becoming a larger hoster (hoping to get more new coins), involves finding new content, tipping said content-producers, and growing the content you host.
The technical problem here, and I need to read up on how things like Ethereum’s Swarm, MaidSafe, Filecoin, Storj, etc reward storage hosters. If it is simply by size, then it becomes a bit more of a difficult feedback loop. As you can see, this technical part is still hazy. But if you can cryptographically issue new coins proportionally to people that host the most popular content, and they can only get new content to host by paying content-producers, it feels like it is getting closer.
- BuzzFeed, race to the bottom, Upworthy-style. What’s considered “good” for the CPDAO, might not be the most “popular”. A meme CPDAO might work with this version.
- If you have a lot of money, you can tip all content, and overpower the network by hosting most of the content.
So, to divorce the form of content (size, amount, etc) from the equation, I thought up a second way.
Version 2 (New coins go to voters).
This version works by hopefully employing a modified form of proof of stake. Some parts work similarly to above.
If you like content, you tip it (giving content-producers money). That tx then gets a special tag (it’s a special form of transaction). The tag is based on a specific address you own. You still then duplicate the content to your local drive to host and distribute. Instead of somehow being paid proportionally to how much content you serve, instead that special transaction becomes a record of a tip. Over time, each tip has a chance to solve a block (and like proof-of-stake, it accumulates over time). If those special tx solve it, then at that point, you still have to prove you host the file. If you don’t host the file anymore, that special tx you’ve done won’t be able to solve a block anymore. That’s the premise.
However, again, I’m not sure if that is possible. I’m thinking hash challenges could work, but hash challenges won’t be verifiable by the rest of the network. Or maybe it is possible?
So based on the above 2 versions, the basic formation of a CPDAO is as follows:
- New coins are issued to content hosters & curators.
- New users discover content of the CPDAO and want to to be a part of it.
- Content-producers buy coins to post new content to the CPDAO (hosting it themselves first). Content grows.
- Content hosters & curators can get new coins if they tip content-producers. Because when they tip, they agree to host the content.
- Go back to the top.
So, as you can see. The following incentives need to be created:
- Content relevant to the CPDAO, because hosters/curators that tip content that don’t fit the paradigm, won’t get more coins (this works better with v1 than v2). But not perfect yet. It still seems possible to create a scenario where people “attack” the network by tipping blog posts containing “aaaaaaaaaaaaaa”. To mitigate around this, a tip could be turned into a “follow”, so if you do try to game it by tipping purposefully poor quality content, your stream will be full of shit. It won’t stop it, but I don’t think the point will be to stop garbage, rather like in Bitcoin, make it so that there’s spam control. This where the purpose of requiring the coin to add content comes from. But guessing what that fee should be is difficult. You want it to be high enough so that people don’t spam blogs such as “aaaaaaaaaa”.
- Content is hosted &distributed through the p2p network. This incentivised by new coin issuance in the blockchain.
Other possible problems: If content-producers get rewarded through votes, there is possibly manipulation possible. Stealing content and passing off as your own becomes an incentive. Don’t think reputation has to be dragged into a CPDAO, but perhaps it must, to make sure people don’t steal content?
So, theoretically, the feedback loops seem to work okay (can need some refinement). It’s just solving the problem of proving that hosting/curation worked of content that fits the paradigm of the CPDAO.
Perhaps it won’t be so farfetched to think that one day in the future we will all live alongside “CPDAO”s and working with them will be our “keep” (perhaps “keep” is an outdated word in this context). Bitcoin is already a DAO and absorbing loads of talent & mindshare from all over the world.
So, I recently found that Filecoin does proof-of-storage through hash challenges. However, more exciting is proof-of-custody that uses both private key + storage elements. So technically, it *does* seem like it is possible. If you are interested in working on this, find me on Twitter (@simondlr)