AMA 01/15/2019

Opens the AMA live from a conference in Switzerland after traveling from London through Zurich, joking about being tired late at night.

Clarifies that the Berlin gathering was an internal formal methods workshop, not a public event, addressing Twitter criticism about attendance.

Explains what a ledger rule is and why writing a formal specification up front beats inferring rules by reading a reference implementation.

Details the ledger specification covering multi-asset accounting, extended UTXO for Plutus and Marlowe, and preservation of value proofs.

The red team, black team review model used at the workshop, with IOHK, Haskell, Rust and Emurgo engineers challenging every design assumption.

Traces the delegation specification's history from a naive 2016 certificate-style design to the richer proof of stake delegation model finalized in 2018.

Argues that specs written close to the Haskell implementation give a reference version for A/B testing and richer security audits, comparing the rigor to aerospace engineering.

Calls the workshop a major milestone as the first company-wide formal review across multiple teams, noting only minor changes were needed afterward.

Announces that release 1.5 retires Byron and starts the Shelley bridge by replacing the old consensus implementation with Ouroboros BFT.

Release 1.6 will ship the decoupled Cardano wallet backend as a standalone product with Icarus style addresses.

Plans for the following update to swap in the new Rust and Haskell clients, replacing the deprecated Byron based Cardano SL code entirely.

The Shelley address model splitting a spending key from a staking key so a hardware or paper wallet can delegate without ever exposing the spending key.

Introduces enterprise addresses that let exchanges hold ADA without gaining control over staking rights.

Upcoming address and metadata design, including stake pool metadata like ticker names, descriptions and URLs, plus signed transaction intent metadata.

Admits the modular Shelley design trades some complexity creep for flexibility, which is part of why the address and scripting architecture took so long to settle.

Lays out two parallel tracks toward staking: core infrastructure updates and a separate effort helping early stake pool operators get set up ahead of a February testnet.

Details the gradual decentralization plan where Ouroboros BFT's share of slots shrinks epoch by epoch until stake pools alone produce every block around release 1.8.

Stresses that Shelley is entirely new software and new protocols, not a fork, backed by more than 40 original research papers with over 20 already peer reviewed.

Points to the company's track record in its first year: over 100,000 users, listings on more than 25 exchanges, no major security incident, and over 50,000 support tickets resolved.

Jared is updating the ledger and UTXO specs to the extended UTXO model that Plutus and Marlowe require, validated against an internal mock chain.

The four layer smart contract stack of Marlowe, Plutus and Haskell, compiled via GHCJS and WebAssembly so contract logic can run client side in a browser or node.

Introduces chimeric ledgers for multi-asset accounting, launching first for trustless stake pool reward payouts before extending to user issued native assets.

Outlines the next roadmap phase: hardware wallet support, multi-asset APIs, and a final set of staking specific REST API extensions.

Plans to migrate the wallet backend and other IOG product lines to GraphQL for richer, more granular queries from exchanges and the command line terminal.

Plans for multi-sig accounts, including multisig control over staking rights with configurable majority or supermajority thresholds.

Sidechain support and threshold multi-signature schemes that let a light client verify a UTXO snapshot against a block header and start spending within minutes, without downloading the full chain.

Projects Cardano's stake pool model could be 50 to 100 times more decentralized than Bitcoin mining pools or EOS's 21 delegates, citing a forthcoming paper on proof of stake fairness.

Introduces the Daedalus DApp platform built on Electron, aiming for a permissionless app store with no central curator.

Signed application manifests get hashed and stubbed on chain so Daedalus can list apps, auto detect updates, and support community reputation and curation.

Envisions stake pools becoming service providers offering randomness, data feeds and other decentralized services to DApp developers for a fee.

Reflects on the shift from old technical debt heavy code to a fully specification driven codebase, with about half the system already replaced.

Argues most cryptocurrency projects fail because they half-ass individual components and ignore how design flaws in one part cascade through the whole system.

Answers whether satellite relays could decentralize Cardano's network layer, covering satellite ownership tradeoffs, micro satellite swarms, mesh net protocols and Project Loon style balloons.

Addresses reports of a Daedalus password corruption bug from the 1.3 to 1.4 encryption upgrade, recommending affected users delete and restore their wallet database or file a helpdesk ticket.

Answers a CryptoKitties scalability question by setting a 50 to 250 TPS target for Shelley and explaining the throughput versus settlement latency tradeoff from the parallel chains paper.

Sets Hydra's longer term throughput target at 5,000 to 10,000 TPS by converging the parallel chains and sidechains research lines.

Plans to adapt Lightning style layer two protocols onto Cardano's stake pools, which already have reputation and uptime, to handle micropayments and bundled transactions off chain.

Summarizes the three pronged performance roadmap: Shelley first, then Hydra, then zero knowledge proofs and layer two networks for the heaviest workloads.

Confirms none of the company's work is patented, released instead under MIT or Apache licenses with documentation under a Creative Commons Attribution license.

Shares a personal aside about doing judo as a kid and college classmates at Metro State Denver who went into Brazilian jiu jitsu.

Tokenizing intellectual property as a non-fungible master token with a vending machine style license marketplace, enforceable through trusted hardware DRM enclaves.

Reselling DRM enforced licenses with automatic royalty payments back to the creator, plus zero knowledge proofs letting art collectors verify portfolio value without revealing individual pieces.

Daedalus hardware wallet support waits on Icarus style addresses landing in 1.6, so Yoroi will likely add Ledger support first around February with Daedalus following in March.

Hopes to ship Docker images around February and get community help pushing Homebrew, Snapcraft and Chocolatey packages for easier installs.

Jokes about wanting to embed goofy images like David Hasselhoff on a jet ski somewhere in the product.

Congratulates graduate student Dionysis Zindros for getting his first paper accepted at the Oakland security and privacy symposium.

Gives an update on the Cardano Ethiopia education program: an all female class of 23 to 24 students from Ethiopia and Uganda learning Haskell and Plutus.

A training philosophy of favoring a thousand excellent developers over a million mediocre ones, citing prior IOHK Africa cohorts in Athens and Barbados and plans for free online courses.

Investing in Ethereum interoperability through an EVM sidechain so existing Solidity DApps can run on Cardano faster, cheaper and safer.

Wraps up the AMA, thanking viewers for their time before signing off for the night.