desmos-labs / desmos

Improving the well-being of users on social networks through the blockchain technology.
https://desmos.network
Apache License 2.0
102 stars 44 forks source link

Vision and Mission of Desmos #119

Open kwunyeung opened 4 years ago

kwunyeung commented 4 years ago

I would like to keep a pinned issue for us to keep the Visions and Mission of Desmos so that we can keep discussing and enhance the specifications of different parts of the network in an open manner. We can keep adding references and why things have to be done like that in this issue. It is also a thread for us to bring examples from other protocols how they achieve or ruin some ideas.

Background

End users become the products of the centralized social media

Social media has been a part of our daily life. We cannot get rid of it if we have to interact with different people or acquire information in a modern way. Tech giants like Facebook, Google, Twitter have been providing "free services" to end users while hijacking their personal information. I believe these centralized social media follow all those privacy policies and not selling a particular user personal information to other companies. However, what they are doing is even worse. They sell the results of the calculated user behaviours with the concept of omni marketing. Whoever you have contacted online, wherever you go, whatever terms you search on search engine, whatever piece of information you read, they take these to make calculations and prediction and users are forced to read ads and being drive to spend. All these platforms become ad channels and the aim is to make end user spend money.

Centralized social media are full of censorship and loosen people's awareness of finding out the truth

The current social media platforms are full of content censorship which endanger the freedom of speech. We see posts, videos, messages and user accounts being blocked or deactivated in different occasions. Some messages with sensitive terms on WeChat cannot even be sent out to friends. People are afraid of expressing their real thoughts. This is because we all have exposed our true identity. In the early days of internet, when we use ICQ, MSN or writing blogs, we preferred using monikers to hide our true identity. ICQ was a p2p chat protocol and we can chat freely with anyone we like. Facebook first introduced users using their own true identity online. This is in favour to their advertisment strategy. From time to time, people tend to believe and follow what is being said from a checked social identity rather than finding out the truth.

Vision and Mission

Desmos aims to be a permissionless decentralized social network. It will be the backbone of different social apps. User interacts with each other in the cryptography way. There is no central authority to censor content. Users choose their own way to express themselves in their profile. They can appear as anonymous or choose to disclose their own identity. This is their own choice. Messages and data broadcasted are all belongs to the users. The protocol itself encourages mutual respect. It is a protocol to let people communicate freely, take good are of your own content, see your own data as assets and avoid being overspend.

DEStroy MOdern Slavery - Sodizon

Target Audience

The target audience of Desmos are very important as they are the people who eventually form the network itself. Every action they have done on the network is a contribution and should be rewarded and take care of.

The 3 groups of target audience are supporting while relying on each other.

As we are building Desmos based on Cosmos SDK, we rely on the functionalities the modules provide. However, there are some problems we have seen happening on Cosmos Hub with the currently working BPoS. In order to achieve the mission of Desmos, we have to understand those problems and make changes on those modules and dynamics.

Block Proposer Selection

Problem

With the existing reward distribution on Cosmos Hub, it only suggests Rich Get Richer. With more stake, a validator has more voting power, and the validator has more chance to create a block, turns out it has more opportunity to earn the block proposer bonus. This is not fair to validators with less voting power as all of them do the same operations for processing and validating the transactions. Basically have the same work done but the bonus always go to the validators with higher voting power. This make the validators with lower voting power difficult to survive and they do not want to make the network have higher thoughput as the bonus will all go to the validators with high voting power. We agree that having more delegations from delegators result in higher voting power. The commission from the delegators' block rewards already pay their effort in marketing the protocol and securing the network.

Solution

Validators should be rewarded based on their output in the right proportion. If the validator is there and have help generating the blocks, they should be rewarded. The existing incentive of proposerBonus and precommitBonus have no problem if we can change how the block proposers are being selected. One possible way is to decouple the relationship between voting power and proposer priority in Tendermint. Voting power is defined by the bonded stake which has no problem as all. The staking module has solve the Proof-of-Stake issue pretty well. Consensus will still be made with enough signatures from 2/3+ voting power. However, if the proposer selection can be selected randomly (randomly not related to voting power), the validators will have similar chance to get the proposer and precommit bonus in a longer term. If this reward can be high enough, it helps validators to sustain and they will encourage more development on the network in order to gain more activities which incur more income to the whole community.

Governance

Problem

On-chain governance is experimental in the space. On Cosmos Hub, parameters can be changed, community fund can be transferred with on-chain governance. If a stake holder can control governance, it has the abilities to change all the rules set in the protocol. In the existing governance module, the votes are casted based on voting power, again. Which means with more stake, with more power to define the result of governance. This introduce the idea of wealthier people have more power than average people. This is totally wrong. Would the vote from Donald Trump has more power than a normal office lady? Would you say they are different rights in the same community because one is richer than another?

Most of the time, the validators are the actors who vote. Delegators usually not able or eager to vote as not many of them have the knowledge of what the protocol really do. Validators have the obligations to explain what the governance proposals are doing to their delegators. Validators also represent the delegators to vote. In this sense, on-chain governance also open a window for bad actors to hijack their delegators.

Exchanges and VC are another issue on governance. In some pervious proposals on Cosmos Hub, we see that some validators running from exchanges or VC never response to the details of the proposals. They never voice out in the community and give reasons on why they voted to the proposals. The dynamics of the community will be altered by those actors without community contributions.

Solution

Governance is controlling many aspects on the network.

Validators are those who should understand the proposals extremely well as they are running the network. Validators will earn commission from the delegators. The involvement in governance of validators become an important criteria to consider if it's a good or bad validator and delegators should take this into account. Telescope will take the involvement on governance of a validator as one of the scores of a validator. To even take it further, this can be something would happen on-chain directly. We may utilize the posts modules to let validators send messages to their delegators to make sure they have communicate with the delegators and wait if they have done their job.

Defining the vote from voting power is wrong but validators still have to act on behalf of their delegators. One way to compensate this would be to count number votes based on number of delegations, but not voting power. For example, a validator with low voting power only means its delegators do not have so much money to acquire so many tokens. If the validator does community work well, many delegators will trust it and delegate to it. In the meantime, exchanges never promote the network protocol as they only want their customers to stick to their centralized platform. They cannot let their customers to vote as their customers don't have their own address holding the assets. They will only have a few delegations and the main delegator will be the exchange itself. In this case, the exchange validator will only have few number of votes. This truly reflect a validator is representing a large number of real users or just one actor controlling all the funds.

Another way to have variance in voting power would be in consideration of voluntary locking being used in Polkadot https://wiki.polkadot.network/docs/en/learn-governance#voluntary-locking. DOT holders can consider increasing their voting power by locking a longer period voluntarily.

Reputation (Proof of Engagement)

Problem

One typical incentive for content creator is ad profit sharing. However, there are too many censorship and biased on content categories which make many content creators cannot get what they should get even they have help the platform created a lot of user engagements and public concern. A typical content censorship is the "yellow labelling" on YouTube. https://support.google.com/youtube/answer/7561938?hl=en YouTube's AI (without letting the public know how the machine was trained) would label the content of creators as yellow which show limited number of ads. For those content, even if the video is being viewed with large audience and draw a lot of traffic to the platform, YouTube will not share the ad profits the content creators because of less ads being shown. Many content creators have shown concern on being yellow labelled even they are only expressing their own views on the recent Covid-19 pandemic or politics. As a result, content creators will reduce creating content in those topics. Platforms like YouTube is in a way of censoring those contents avoiding content creators to publish anything their AI doesn't like.

The existing up/down vote system in many platforms also encourage spamming and dummy supporters for some artificial influencers. There is no cost to create fake IDs and follow or upvote the artificial influencers.

Solution

The "legitimacy" of the content and content creators should be vetted by network users but not a private trained AI. Each reaction (like, share, comment, flag) of the user should be paid with transaction fee to avoid spam. The reactions act as vote by the users. The value of the votes will be calculated with their stakes and the corresponding stakes will be accounted for slashing to avoid sybil. Users will share the rewards earned by the content creators while being punished if they have supported bad content. A reputation index is being calculated to reflect the contribution done to the network.

kwunyeung commented 1 year ago

I strongly believe that such a document does not belong into an issue tracker.

Thanks for interacting. Desmos started with this github repo and it was the easiest way to pin some information with the repo.