A Bonded Curation Community
Ever since I came up with the original meme markets designs, I’ve been delving continuously into countless permutations of token design systems. I’ve received amazing feedback and insight into many token models after my previous post:
Thank you for all that gave feedback and emailed me with ideas. The more I dig, the more I realise just how incredibly deep this rabbit hole goes. The design scope is massively broad. Just have to keep iterating and poking holes. So, this is another token design. Please poke holes.
Recently, in the context of Ujo, we’ve been exploring variations of these models to enable collaboration around the development of an open music industry and more specifically attempts to add skin-in-the-game signals to the validity of shared metadata.
The issue with some of the previous designs is that it seemed unreasonable to expect that holders of the token community would actually want to bond/stake/dispense their tokens against information that could amount to millions of records. Locking or spending tokens in this long-tail have low incentives.
Thus: This is a new variation of a continuous token model (tokens are algorithmically minted as desired) that fosters any curation in any community/meme.
A Bonded Curation Community
The gist is:
Tokens are algorithmically minted and then used to deposit it against participants in the community that curate and foster the shared goals of the community. These participants can then use their standing (proportional backing) in the community to curate any information in this community. Anyone, at any point in time can become a curator. It is dynamic. The deposits can be shifted around to other curators as desired.
In detail, it works as follows:
- Mint the community’s token according to a hardcoded algorithm with ETH. Token price is correlated to active supply (eg, more tokens in circulation == higher price to get in). (the curve mostly linear + slow exponential).
- All ETH that is used to mint the tokens are kept as a communal deposit.
- One can leave at any point by taking a proportional percentage of the deposit pool with you. Eg: 1/100 (where 100 == current total supply) of a pool of 200 ETH == taking 2 ETH with when leaving.
- The only explicit usage of the token is used to deposit it against a set of curators.
- This results in having a list of curators that get a percentage standing in their community.
- Eg 1) Ujo 80% (160/200), 2) BigchainDB 10% (20/200), 3) Mediachain 10% (20/200). (Where 200 == active supply).
- Curators can back their percentage stake in the community to any information/meme.
- Eg: Ujo can back “Re-design standards” with their 80%.
- Eg: Mediachain back “Re-design standards” with their 10%.
- Thus, 90% of the community regards “Re-design standards” as important information and signals it as such.
- They can use their backing simultaneously with any information. Equally Ujo can use their 80% to also back “Buy SoundCloud” without forfeiting any other information they back.
- Any backed information can be revoked, simply signalling that it’s not regarded as important anymore.
This system thus rewards early adopters for actively curating information to the benefit of that community. As the supply grows, the pool grows, and although percentage-wise if they choose to leave, they will get less of the overall pool, the value of the floor will continue to rise.
Newcomers to an existing community needs to weigh up the higher cost of joining and the possibility that by doing so will cause an earlier adopter to leave. The tokens provide value in it being used to signal commitment to a shared goal. It is “proof of work” to get inside to coordinate. It is eustress in social systems design.
Rationally, if this system provides value in terms of coordinating information with skin-in-the-game economic signals, then a holder will keep the token as long as they feel that it can guide attention accordingly. Thus, at some point the value of the token will increase to a point where the holder would want to exit as they believe the perceived coordination benefits is less than the current value. They can then sell the token on a secondary market or leave.
The algorithm is designed in such a way that the ceiling (cost to enter) and floor (reward to leave) grows wider as the community grows. Thus there is a middle-ground where the token will trade on secondary markets until it hits either the ceiling or the floor.
Example value for different users:
I’m a curator: I will buy a token and deposit it to my own reputation. I’m a good curator because I do x,y,z. If you think I curate and provide novelty well, please also back me.
I’m not a curator but want to contribute: I use/consume the artefacts of what this community produces. I will buy a token and bond it to participants in the community whom I regard that fosters the goals of the project.
I’m a speculator: I think this community will be successful and the shared goal will attract attention. I will buy a token now and bond it to a whomever will produce more value/attention.
I’m a developer: I will buy a token that backs the curator that fosters the developers of this community.
Permission-less, Bonded Curation
One of the design goals of these systems have always been that one can create them without permission. In other words, one can bootstrap such a community around any meme and attract the meme’s attention and network effect.
To be more specific, let’s take the example of creating a curation community around Truffle [an open source dev framework for Ethereum].
I can create the contracts, deploy them, start buying the tokens and start using it to curate what *I* regard to be important issues to work on. Through this I can show to the Truffle team what issues are important to fix and what features to develop by adding skin-in-the-game signals to the repo.
The question now is: should the Truffle developers pay any attention to this curation community? Nope. They don’t have to. However, if the curation community succeeds in attracting the attention of the Truffle developers, the value of the curation community goes up, as it has succeeded in putting it’s skin-in-the-game signals to use. As it grows, and grows, it becomes hard *not* to pay attention to all the signals.
In order to make the Truffle developers pay attention, the curation community can even decide to bond their tokens to the developers themselves.
There is also potentially a good likelihood that many derivative communities might spawn. Many might simultaneously compete for the Truffle developers’ attention. It might not even be directly Truffle related. It could attached to other meme’s, such as Ethereum development in general or Solidity development.
The meme curation communities will compete for attention. With the original meme market designs, there was the addition of a global namespace in order to reduce competition and increase coordination around hashtags. eg #truffle, #ethereum, #vitalikbuterin, etc.
For the sake of simplicity, I’ve decided to avoid this namespace design for now, until there is proof the underlying coordination mechanism works. However, one could potentially build a meta curator community that essentially pays curators to take a namespace and point it towards a specific curation community. eg: “The grand meta curation community regards curation community at address 0x42 with 80% certainty to be *the* #truffle community.”
From Simplicity to Complexity: Namespace Messages, Filters, Rankings, Futures.
Increasingly with these token designs, I attempt to design a simple core that could potentially lead to emergent complex behaviours, instead of pigeon-holing the token. This design allows some immediate interesting emergent behaviours.
Namespace Messages
Eventually, the community can craft unique sets of information to signal for various things that’s specific to their community. Let’s say it’s again an open source dev community. The namespaces can signal different actions to be taken: eg
“merge:#34” == backing whether to merge issue 34.
“upvote:#44” == support a specific feature.
In Ujo and an open music commons, we can use this to help disambiguate metadata. eg:
“work: Taylor Swift — Shake it Off, metadata_hash: Qm77xx33” == The correct metadata for Taylor Swift’s Shake it Off is contained in the CBOR object Qm77xx33.
Filters
By backing your reputation to a community, you will have 2 signals -> the current proportional standing, and the proportional standing at the time of first backing it. You can design separate filters that incorporate time as well (like a normal Reddit-based algorithm) that takes these signals and modifies them.
Rankings
As with previous designs, if many of these curation communities flourish and use the same underlying algorithms, you can start ranking different memes against each other, much like how you can compare market caps of companies.
If a curation community wants your attention with a total communal pool of 1000 ETH and 80% backing, vs a curation community with a communal pool of 100 ETH and 100% backing, you can discern the relative value. There’s more skin-in-the-game in one vs the other.
Futures
Currently, one can speculate by buying a token early, hoping it will be useful to coordinate later. There are other ways to hedge as well. The active supply is a direct metric that can be measured in prediction markets. Thus, one can attach a prediction market to essentially create “curation futures” that predict the growth/size of the community some time into the future.
Drawbacks & Potential Modifications
Unfortunately, unlike previous designs, none of the ETH are flowing to actors/beneficiaries behind the meme/community. Whilst it can succeed in attracting attention, the Truffle developers don’t get anything from it. For them to make money, they have to take part in their own attention community (buying tokens early).
Thus, whilst a fully decentralized, permission-less curation community is useful, one could potentially design slightly modified versions that also foster and keeps the beneficiaries behind the system “alive”.
For example, a 50/50 split could be done that pays it to the Truffle team directly, with the rest going into the communal pool. The curation community can then use a new namespace to signal how Truffle should spend its funds.
“fundmanagent: Pay Tim Coulter”
“fundmanagement: Hire SEO/Marketing”
One then just need to trust that the Truffle team won’t use the funds to attack the curation community and buy up reputation itself. eg:
“Get paid. Take money to buy up rep. Get paid again. Take money to buy up rep, etc, etc”
Outstanding Questions
This system adds skin-in-the-game signals that attempt to produce more novelty. It adds transaction costs in order to do so. It seems useful, but the value of additional coordination benefits need to outweigh the additional transaction costs (vs something like just using a sub-reddit to coordinate). Where this level is, is still uncertain. It might not be useful for the long-tail and only useful when noise starts becoming an issue.
Next Steps:
If anyone has feedback, would love to hear it. Barring glaring issues, I feel fairly certain that in its current form, it’s worth giving a spin. I have started with preliminary contract writing.
Thanks:
As always, thanks to Maciej Olpinksi who has been pushing on this thinking in his own way. This design is heavily inspired and very similar to Maciej Olpinski’s Userfeeds → the act of using ETH alone to back information. Except, the difference here is the ability to create these interest networks without a centralized intermediary.
For this particular design, also big thanks team Ujo who always listens to these crazy ideas. Additionally, Rabbyte who shared a similar idea after my continuous token models post. Thanks.
