YKarma is an experimental project to model reputation as a spendable currency.
The basic concept: every person in a community or organization is allotted 100 "karma coins" to distribute each week. These must be given away to other people before they can be used. The recipients can then spend these coins on various rewards (a day off, a conference ticket, a coffee with someone notable, etc.)
-
Quantifying reputation makes it possible to build a real reputation economy. That could benefit many people.
-
It's an experiment which could have a lot of interesting emergent properties.
Giving people public scores or ratings, like China's new "social credit" system or that Black Mirror episode, is indeed a dystopic concept. Fortunately this is quite different. People don't have a persistent YKarma score, they just have coins to spend in the notional reputation economy. If someone doesn't have many, that doesn't indicate unpopularity; it could just mean they recently spent theirs on a reward.
The general hope/notion is for people who excel in fields and communities which don't see much in the way of monetary compensation -- poetry, art, music, open- source software development, indie video games, nonprofits, churches, etc. -- to receive new rewards for their work. Since every participant receives coins which must be given away, transactions are no longer zero-sum. People can be recognized and rewarded without it costing the giving individuals anything ... as long as rewards are available for people who accumulate enough karma.
OK, but why would anyone make any valuable reward available for purchase with a reputation currency like this?
The most obvious and most boring use is within a company, wherein offering employee rewards -- a day off, a course, a conference, a social event at company expense, etc. -- would a) be good for morale, b) identify those people most highly thought of by their fellow workers. In that case the company would be the "vendor" of rewards for their employees.
It's also easy to envision why external vendors would want to offer rewards to people who have established themselves as high-reputation within a particular community. The immediate value for conferences etc. would be simply having them present. The longer-term value would be establishing relationships with people who are, definitionally, influential.
It could also make sense to offer "perishable inventory," e.g. empty hotel rooms or airplane seats, as doing so within a reputation economy wouldn't devalue that inventory's perceived monetary worth.
There may also be a concept of brand reputation, wherein brands make goods or services available in exchange for sizable amounts of YKarma, just as they comp celebrities today. The amount of YKarma spent on competing brands could serve as a proof and/or quantitative measure of their relative values.
It's also possible, crazy as it may sound, that some people might like to give rewards away to strangers who have been shown to be deserving.
But how can a single measure of reputation be meaningful across these different fields and communities?
It probably can't -- though, again, this is all very experimental! -- so when coins are minted, they are "tagged" with the set of interests/fields for the community in question, ranging from fairly specific ("Ethereum", "Consensys") to somewhat generic ("blockchain", "open source") to very generic ("tech"). Then, when vendors make rewards available, they can require payment in karma coins with a given tag -- "tech" coins to acquire tickets to a generic tech conference, or "open source" coins for something more specific.
Sort of, but not really, and to the extent it's true, it's a feature not a bug.
When you spend karma, the money is destroyed (although see below) because reputation is not really transferable, so money is also taken out of the economy. And while it's true there'll likely be considerable inflation over time if/as more people join, this is deliberate: YKarma is inflationary by design. Like reputation itself, the real value of YKarma savings decay steadily, making it difficult to hoard capital. That's one of the things which makes a reputation economy very different from a monetary one.
(At one point I was going to build significant demurrage into the system, but on reflection decided that wasn't needed, or at least not yet.)
It's a web site backed by tokens on an Ethereum blockchain.
Because the ultimate idea is to build a decentralized, permissionless currency which federates an arbitrary number of communities and users, and that is, it seems, a mission which calls for some kind of blockchain.
That said: this is an experiment, and if it goes anywhere at all it will presumably require a lot of tweaking, which, at least initially, can't really be managed efficiently via consensus or voting. So there'll be ongoing tension between initial control of the experimental parameters and the end goal of permissionlessness. It's on a blockchain, though, because that is the end goal.
OK, but even granting the desire for an ultimately permissionless and hence blockchain solution, is Ethereum really the right platform?
It's the right platform right now. Maybe in some distant future it will be migrated to a different one. But in the here and now, Ethereum is running, battle-tested code with a maturing software toolset for its smart-contract system, which is flexible and powerful enough for this experiment.
But how can this possibly work on the Ethereum blockchain, with its high gas costs and low transaction bandwidth?
It can't. So this won't run on the public Ethereum mainnet, until / unless that scales (eg via Plasma and/or sharding) massively beyond its current limits. Instead it will run on a "consortium" / "proof of authority" Ethereum blockchain, just like the Rinkeby testnet, with an eye towards one day federating together multiple blockchains -- each supporting one or more community -- as Ethermint zones.
For now, the code is set up so there's a main smart contract, YKarma, which handles most of the economy, and a separate administrative smart contract, the YKOracle, which can manage communities, tags, and vendors.
Users must have at least one URL which identifies them -- currently either an email address or a Twitter handle, but obviously many more options, ranging from phone numbers to Keybase URLs to IPFS addresses. Coins are sent to these URLs, meaning you can send coins to people who are not yet part of the system and/or any of its communities. However, communities can restrict the URLs their members' coins are sent to, so they can e.g. ensure that coins only go to people with @company.com email addresses.
Users who never want to deal with the blockchain won't have to -- that's the whole point of all the web code in this repo -- but users who want to become "free agents" may do so at the cost of additional complexity.
Ultimately each individual community will validate its own users, vendors, and which other communities they include in their reputation economy. In the very long term I can envision networks of networks of federated communities.
Sure! Let's follow some coins through from genesis to expenditure.
Alice is a YKarma user; every Monday, she receives 100 YK (aka coins aka karma) to give to others as she sees fit. One fine Monday, Alice decides she likes Bob's latest essay on Medium, and sends him 10 YK from that pile. Note that Alice can't spend it on rewards herself; YK has to be given away before it can be spent.
Bob receives an email saying "Alice has sent you some karma!" (Obviously we want to avoid spamming people, so that email is only sent once unless Bob opts in to further notifications.) The email includes a link to the web site. If he already has an account in Alice's community -- say, if someone already sent his Twitter handle some karma -- then he can link his email to his Twitter handle and add those 10 YK to his existing balance. Otherwise, he now has a YKarma account with 10 YK in it.
Ten YK isn't very much. But next week, when Bob's essay goes viral, lots of people in this community send him YK, and he winds up with a balance of 1000. Meanwhile, Charlie, a popular community co-founder, has decided to become a YKarma "vendor" and offer a "Charlie Buys You Dinner" reward to people in the community, so as to keep up with influential people in it.
Bob logs back into YKarma, goes to the Rewards page, and uses his YK to purchase the "Charlie Buys You Dinner" reward, partly for the free meal, partly because he's always admired Charlie's work. The irrevocable ownership of that reward is transferred from Charlie's vendor account to Bob's user account, and then it's up to them to then arrange the dinner.
A note to decentralization devotees: this is all described as happening through the YKarma site, because that is the easiest solution -- but behind the scenes, this is all blockchain-based. Anyone could create an address on the blockchain, receive "YKEther" from a tap, and directly create / interact with communities, accounts, and vendors. This will obviously only be appealing to technically sophisticated users / communities, but the possibility exists.
It's deployed at https://www.ykarma.com/ -- but you won't be able to do much there until/unless you're invited to join a community. Email info@ykarma.com if you want to join the alpha-test community. (Note that, currently, if you do join, then your email address will be visible to other community members. I might change that if people seem dismayed by it.)
Currently they are gone forever, because we're modeling reputation, which isn't really transferable, and because money has to leave the system somehow or we'll have runaway inflation.
However one possible experimental notion is for a fraction of the spent karma to be retained by the vendor, because there is such a thing as proxy reputation. This may be at the buyer's discretion, independent of the transaction completing successfully -- i.e. you can always buy the reward, whether or not you want to transfer some of the spent karma to the vendor. (Yes, this is yet another significant divergence from traditional money.)
Furthermore, we track and display the total number of rewards a vendor has ever sold, and the total karma they have received, as another form of reputation. This is admittedly a "rating", but one still very different from e.g. "social credit", and rating vendors as opposed to users seems a reasonable thing to do.
This is known in the infosec biz as a "sybil attack." Fortunately, there are a lot of notional ways for communities to prevent sybils: two-factor authentication to ensure each account has a phone associated with it, making sybils both inconvenient and expensive; only allowing accounts with emails at a particular domain; requiring X established accounts to vouch for any new one before they can join; simply having trusted administrator accounts approve each new account; etcetera.
Lots of stuff, including:
-
Rewards should be auctionable as well as fixed-price. That's obviously much more complex to code, though, so it's not yet implemented.
-
Karma used for governance / voting; people who want to bring forth a proposal to a community must spend X karma to do so, where X is dissuasive but not preventative, and then the entire community can spend karma as votes, with Z net votes in favor required to pass. Interesting, but again, complex.
-
I'd also like to implement a "vouching" system wherein communities can agree on new members, eg "5 existing members much vouch for you before you can become a member and start giving out karma yourself." One could also limit by domain, e.g. "only email addresses which end in @company.com can be added." For the alpha test, anyone who is sent karma becomes part of the "alpha test" community.
-
In a better world the web site would be less janky (the interactions between React, Firebase, and the blockchain's relatively slow read times make the site behavior and performance ... suboptimal) and would look-and-flow much less like it was designed by an engineer.
You're kidding, right?
You want the project HOWTO.
You want the project MANIFESTO.
That's slightly unclear. Right now I just want to get the code into a stable and reliable state, and I think it's ready for alpha testing. Which means it probably isn't, but that's the chance that alpha testers take. So, most immediately, I'm looking for volunteers to take part in that testing.
I'm also interested in talking to people who find the concept intriguing, whether they be potential collaborators, parallel experimenters, angry critics, or anything else. If you're any of those, please feel free to reach out to me at info@ykarma.com to converse privately, or @rezendi on Twitter to discuss publicly.
Mostly because the domain seemed vaguely descriptive and was available.