womenbuildweb3 / grants

Help build web3!
0 stars 1 forks source link

[BUIDL '22] Sharetape #4

Open theekrystallee opened 1 year ago

theekrystallee commented 1 year ago

Sharetape

Name of Project: Sharetape

Proposer: theekrystallee

Do you agree to Encode Club's Terms and Conditions?: Yes

Do you agree to the grant process outlined by WBW3?: Yes

Project Description

From black box algorithms to discretionary content moderation enforcement to ambiguous monetization policies, social media giants like Instagram are notorious for enacting arbitrary changes that negatively impact their end users. These advertising-based platforms are incentivized to lock users in their data moat and monetize that data, often at the expense of user privacy.

In contrast, blockchain enables an open, permissionless internet where users fully own their digital identity, data, and assets. Decentralized social media empowers users to take back control of how their content is distributed and monetized.

Leveraging user-owned infrastructure, Sharetape is a decentralized video streaming platform centered on collective ownership, privacy, and community governance. Built for creators by creators, Sharetape supports networked collaboration through joint channels where multiple creators can co-own and co-curate content on shared feeds. Users can easily monetize their content through subscriptions and gate their videos for their most loyal subscribers. Users can also tailor private feeds and share them with their closest friends.

With a reputation-based recommendation engine, governance-based content moderation, and privacy-first features, Sharetape tackles issues that plague current social media incumbents like biased and subjective censorship, dissemination of fake news, and data privacy violations.

Sharetape is where content you want to watch thrives.

Tech Stack

Frontend Framework: Next.js Smart Contract Language: Solidity Wallet Connection: RainbowKit, wagmi.sh Ethereum Library: Ethers.js File Storage: IPFS, Web3.Storage Data Querying/indexing: The Graph/GraphQL Social Graph/Network: Lens Protocol CSS Framework: TailwindCSS Ethereum Development Environment: Hardhat Layer 2 Blockchain: Polygon Backend Server: Node.js Backend Database: SQL Encryption: Lit Protocol

Next.js will be used to build out each of the web pages as well as being the main UI where users will login/create accounts and upload and watch videos. Solidity will be used to set up the entire smart contract of the website where transactions can take place in terms of video upload and wallet connections. Ethers.js will be the main ethereum library so we can perform transactions. IPFS will be used for file storage. The Graph will be used to query our videos for the main feed and other feeds. TailwindCSS will be used for the design of the website. Hardhat is used to deploy your contracts, run tests and debug Solidity code without dealing with live environments. Polygon is the Layer 2 blockchain that we will be transacting in, with the Matic coin. The backend server/DB will be used for other user information.

Development Roadmap

Milestone 1

Number Deliverable Specification
1. Account creation Users can auth with their wallet and create an account based off their Lens handle.
2. Channel creation Users can create and customize channels. Ownership of channels are represented as NFTs.
3. Video upload Users can upload videos and mint them as NFTs. Video data will be encrypted using Lit and stored on IPFS.
4. Video playback Users can play video content, which is streamed using Livepeer. Play behavior such as total plays and time spent watching will be recorded.
5. Discovery and search Users can find content and other users. A custom subgraph will be created and deployed to support search functionality.
6. Social features Users can subscribe to channels and like and comment on videos. Subscriptions, likes, and comments will be NFTs built on top of Lens Protocols and Unlock Protocol.

Milestone 2

Number Deliverable Specification
1. Reputation system We want to minimize the number of attack points for potential bad actors with our reputation system. To derive the reputation ratings the system uses probability density functions to combine feedback. It includes comprehensive mathematical implementations consisting of feedback weights with a discounting agent, and scalable feedback collection. We also want a forward biased algorithm as the algorithm understands that old feedback may not be relevant for the actual reputation rating as user behavior changes over time. The key parameters that are necessary for this reputation system include watch time, likes, comments, video volume, video interaction, shares and favorites. The discounting factors will include dislikes as well as deleted after receiving enough spam or rude or abusive flags.
2. Recommendation algorithm Our recommendation algorithm has four main pillars: platform value, creator value, and user value. With respect to a given user, videos get scored by our machine learning algorithm and user behavior using four data points – watch time, comments, likes, and indication of plays. The data points are used by our predictive model to estimate whether a given user will like a video, leave a comment, watch multiple times or even engage at all. After each subsequent iteration of the predictive model, the accuracy of our recommendation algorithm increases. Here’s a simplified version of the formula used to rank videos: predicted(likes) value(likes) + predicted(comments) value(comments) + estimated(watch time) value(watch time) + predicted (plays) value(plays). The highest ranking videos are then recommended to users based on the output of this equation.

Milestone 3

Number Deliverable Specification
1. Collaboration features Users can co-own shared channels and curate content together. Reputation system and recommendation algo will be expanded to incentivize co-creation.
2. Membership features Creators can paywall their content, and their subscribers can upgrade their membership NFTs to access token-gated content.

Total Budget Requested

$14,650.00

Project Links

https://github.com/theekrystallee/sharetape-dev https://sharetape-dev.vercel.app/

Team Members

Paras: GitHub, Twitter Krystal: GitHub, Twitter Sarah: GitHub, Twitter