EthMagicians Trust & Reputation Ring DISCUSSION POINTS: --- UBI & Voting - Use KYC process and tweak to make it less silo’d - Do this through a “phone tree” type process. Algorhythms for reputations may need to vary by context: Examples of Contexts for Trust: Accountability Latency Friendliness Humanity Hygiene Considerations for whether people change - if you believe people change then you’ll weight more recent determinations more heavily. If you do not, you’ll value the entire history equally. Can we humanize the decentralized world? Ethics, Morals, Objective and Subjectivity - A Trustless Marketplace has existed.. Never. What is good? How do people know what is bad? Thresholds are advisable for qualifying reputation. ----- Bounties Network: Difficulties faced in a viable reputation system? Avoiding Evil reputation - challenges lie in validating that signals are good ie. over-reliance on reviews (reviews are not a ‘good’ reputation system) .. how do we show signals of good reputation - through action-based metrics, how many have been accepted, how long taken compared to value.. What identity solution for profiles: not ERC725 ..solely around Ethereum Address. Offchain storage of their contact info so they don’t have to pay with gas to make changes. IN this world, lose your address, lose your reputation history. Contract-based wallets for recovery purposes is a thought for the future.. ---- Some questions to consider for the Ring: How do we define TRUST? What is wrong with Web2 that can be better with Web3 to address Trust & Reputation? ---- Legal Enforcement systems are policing systems and are the “back up” to trust when you deal with someone at a counter, etc. Ebay, etc: dispute resolution processes are the backend of trust. RECOURSE.. alternatives? Public shaming in community for failure (this introduces spite action possibility).. --- Do I trust you enough to have a conversation with you? Do I need to trust that you are good to your family at home, elders and children.. This is NOT USEFUL in most contexts, and highly useful in SOME.. EXPECTATIONS are useful. What do we agree to expect from one another based on an exchange? What is the best we can do given our constraints? Trust Predictive and deterministic algorhythms.. Competitive algorhythms in the marketplace that over time prove out reliability and accuracy Browsers are a clear example of this.. Which browser has the most trustworthy algorhythms to return what you are looking for? Trustchains - we trust things that our peer group trusts, or someone we delegate our trust to because of their reputation. A trustworthy entity enables trust through transparency, accountability, feedback process, customizability.. Basically access and control of data and options. AAA or gems - arbitration or mediation firms - expensive but fast. The cost for abitrator comes from hiring trusted people. People trust in: Meeting a specification - is spec vague, did fulfiller not meet.. Clarifying that it is a CLEAN SPEC before actioning on it. Build a website - what kind of website - how detailed. Crappy Bounties - good work on bounties require good specs! DEFINE WHAT YOU WANT, OR YOU WON’T GET IT! How do we get to Win-Win-Win for both parties AND the platform.. If we don’t have a clear agreement everyone loses. Define your BIAS!! Are you a ‘customer is always right’ or ‘our business model dictates decision in X direction’ CROSS Platform reputation.. REPUTATION AGGREGATION.. Different dapps have ‘action-based’ reputations .. building trust in one place can help you gain entry into places where your trust must be pre-established. Multiple aggregators provides that algorhythm variety and complexity for leaders to emerge. Ridesharing - equal pay across.. Reputation requires selfie photos.. Links to the people in the photo to show step connections.. Selfie becomes the action defining reputation. Contact becomes the verification. CULTURE based reputations.. Give the users the option to choose their metrics and plug into the algorhythms. Where can you acknowledge your failures? Say ‘I regret this’ .. to demonstrate change or request forgiveness to have an action ‘forgotten’ by the metrics. Resumes show off what we are proud of, where can we show our lessons, what we are not so proud of and what we learned from it? How to aggregate our pluses and minuses. Someone who has made big mistakes may be more trustworthy for having learned from them! Universal reputation is not realistic. Contextual reputations and custom-defined reputation forums are appealing so far. Reward good actors, incentivize good actions. . AND actively exclude bad actors. It is currently illegal in many countries to say something that hurts another person's reputation, even if it’s TRUE. --- ETH REGISTRY: Running Code Exists! Centralized identity, total hearsay and gossip.. Self attestation is second level.. Then 3rd Level is the trusted parties within the ecosystem such as wallets, entities whose success is based on their trustworthiness. Trust model interoperability .. are our datasets compatible? Layers.. Data - Facts - is the bottom layer, we have to agree to add our data from different places to the aggregator. Social Justice - Judgement - is the second layer, how to use the algorithms. A third level is how you choose to use this. NEXT STEPS, SubRing Makers: Category - Trust & Reputation Conversation threads - the following subjects: Dispute Resolution - Dean & Mark Types of Reputation - Daniel & Blake Forgiveness - Chaals Reputation Aggregation - Michael & Morgan Ethereum Ecosystem / Contextual Trust -- Chelsea
Last changed by  
Guest
295
Published on HackMD