Video to Elon: Decentralized Twitter Using Doge and Cardano

Opens November 6th 2022 with a stream-of-consciousness whiteboard aimed at Elon Musk and Twitter staff, designing a decentralized Twitter from the crypto industry's best components.

A decentralized social network is hard to define, since a network of millions of people is already decentralized, with high clustering and a diameter like the Six Degrees of Kevin Bacon.

Centralization comes in via curation behind storage, and that crypto is good at peer-to-peer, cryptography, and incentives but not AI, decentralized computation, or storage.

Uses Twitter as a case study where you sign up, create a profile, select preferences, enter the network, and view the tweet space.

Curation and information flow is the algorithm Elon talks about, the source of shadow-banning and bias claims, since what you see is filtered from the social graph.

Preferences are discovered via AI, and interacting with the tweet space means tweeting, replying, liking, quoting, and retweeting, which drives the contagion factor.

The algorithm maximizes engagement and time in the tweet space, which is negatively correlated with quality since deep nuanced information takes time and pulls you out.

Argues you are the product, since your profile lets the network micro-target you for advertisers, so the algorithm disincentivizes good information.

Cites the book Surveillance Capitalism, saying adding the D for decentralized does not remove the engagement-maximizing algorithm that makes billions for Google, Facebook, and Twitter.

Illustrates the storage challenge that summing all crypto storage across every blockchain's 13-year history is less than an hour of Twitter's traffic, so a blockchain cannot handle that data flow.

You need massive infrastructure and a business model where information quality increases over time, unlike social media where it decreases.

Proposes representing users as DIDs, decentralized identifiers, a W3C open standard that Microsoft and IBM worked on for people, smart contracts, IoT, and vehicles.

A DID has a did address ID string and a did document of metadata, so signing up gets you a DID and creating a profile fills the document.

A key decision is the level of verification, noting a researcher at Virginia Commonwealth University created a proof-of-human super-captcha to verify a person is not a bot.

For claims like being a doctor an IDV provider like Accurate and credentialing bodies verify, and that most networks minimize friction, causing false accounts.

DIDs create a blockchain-neutral identity ecosystem via Atala Prism, designed for Cardano but cross-blockchain, with Prism V2 coming in Q1 built on the Ethiopia work.

Proposes a verification-as-a-service API ecosystem for the document, augmented with Algorand's Falcon, a NIST post-quantum lattice-based compact signature scheme.

Putting Falcon in the did document gives quantum-immune cryptography and unforgeable signatures for every tweet, so a hacked account cannot impersonate you and tweets are verified.

Proposes layering zero-knowledge crypto via Prism for Boolean proofs like are-you-over-18, enabling age verification for the OnlyFans-style content Elon discussed.

Turns to the tweet space, noting each piece of content has claims, evidence, and an origin, so a huge claim like Trump died of a heart attack raises what-evidence and who-tweeted-it questions.

Current options like the heart and retweet are limited, since liking is meaningless and retweeting spreads without context, increasing contagion and pushing it into more tweet spaces.

Proposes more nuanced options than a like, such as suspicious, downvote, and tell-me-more, since crypto's payment superpower lets you put bounties, creating a marketplace for curation.

Retweet-with-context, mechanisms to slow contagion by flagging fake versus real news, and fixing that replies lack memory and deteriorate.

Proposes turning each tweet into an NFT with a reputation score and associating each user's DID with a progressive NFT that grows over time, like Crypto Bison.

Treating a tweet as an object lets it be moved, analyzed, and archived with a reputation that flows to sharers, so habitual misinformation-sharers rank lower and good sharers gain weight.

Curation marketplaces become prediction markets for the truth of reported information with money behind it, like betting on the November 8 election outcome.

Introduces a veracity bond where the tweeter stakes money behind a tweet and loses it if false, giving the tweet space nuanced tools and economic agency behind information.

A realm for incompleteness distinguishing unverified from verified, noting curation is where centralized networks hit trouble as fact-checkers de-platform content and ban accounts.

Digital credentials make accounts super-secure, and that instead of following you subscribe like pub-sub, using Poldercast, a 2011 decentralized pub-sub protocol that proposed decentralized Twitter.

The Poldercast author was paid to update the paper, so a follow becomes a meaningful subscription unlike Twitter's algorithm that does not guarantee follows appear.

Proposes breaking the unified view into extended domains like Mastodon, with programmable circles for friends, follows, and trending that you move your eyes into.

Advertisers can restrict ads to a curation regime matching their values, a super-woke-corp box versus an evil-gun-corp box, so brands do not appear in the wrong part of the network.

Users can filter to only see tweets with a veracity bond, so CNN, MSNBC, Fox News, and Bloomberg must stake money on their journalism or lose it.

NFT'd tweets have a tradeable, immutable, timestamped history that cannot be de-platformed, so you can see a user's or corporation's track record and hold the powers-that-be liable.

Cites the Santa Clara Principles on content-moderation transparency endorsed but not followed by Apple, Facebook, Google, Reddit, Twitter, and GitHub, proposing a Santa Clara view versus an Anarchyland view you toggle between.

Proposes advertisers interact via a Basic Attention Token model from the Brave browser, where users own their data and a token creates a direct ad-owner relationship across curation domains.

Suggests combining BAT with Kachina, IOG's private smart contracts paper, for private data mining, so the advertising token gives economic agency through bounties, tips, micropayments, prediction markets, and NFT marketplaces.

Doge cannot provide this infrastructure as a Bitcoin fork, so it should be a side chain and a web 2.5 application rather than a fully decentralized protocol.

A side chain takes Cardano's 3,000 stake pool operators as a route of trust, sorts them, and runs a fast BFT protocol like Narwhal and Tusk by George Danezis at UCL at 600,000 TPS.

The right security model gives immutability, censorship resistance, and auditability at high TPS, augmented by service providers for storage and IDV, suiting IBM Hyperledger Fabric chaincode.

Doge becomes the advertising, veracity, and NFT-issuance token via extended UTXO with Plutus or an EVM, airdropped for users to redeem on the side chain.

Prism for identity and Dionysis Zindros's proof-of-proof with chainless log-space mining, so consensus maintainers hold only a small part of the state and archive nodes act as service providers.

The core innovation is a marketplace for information curation, a marketplace for advertising with values behind it, and a giant identity engine, achievable with a bespoke BFT protocol in 18 to 36 months.

Recaps the design as post-quantum, with reconstructed identity, a verification marketplace, a curation marketplace, a reputation system demanding media honesty, and subscribable left and right views like Spotify playlists.

Mithril lets users verify every tweet client-side without a full blockchain copy for inclusive accountability, and the advertising Doge could be worth hundreds of billions, printable like Tesla's equity offerings to retool the system.

Proposes adding the Signal protocol with post-quantum channels for direct messages, plus MPC protocols for disappearing messages and trusted enclaves, all with verified users and public keys.

The result is tethered to a decentralizing layer one that gets more decentralized without losing speed, an open protocol others like Jack Dorsey can build on, with a tunable information-quality marketplace.

Argues the engagement algorithm is destroying society and making politics toxic, so it has to become more profitable to be nice, which is why he designed Cardano for side chains like this.

Economically the stake pool operators maintaining this ledger earn Doge block rewards, so ADA holders get both ADA and Doge, a partnership where Cardano provides security and Doge is the social-network currency.

Closes hoping it starts a conversation and saying the ecosystem is open for business for anyone building a decentralized social network.