The implications of 3074 and 4337

With the recent approval of 3074 in the next Ethereum update we think it's now a good time to sit and think what are the short and long term implications for the "account abstraction".

• ERC 3074: This proposal introduces a method to allow externally owned accounts (EOAs) to delegate their signing authority to smart contracts without transferring ETH or assets. It does this through two new opcodes, AUTH and AUTHCALL, that enable a contract to perform actions on behalf of a user, based on a signature provided by that user. This mechanism essentially allows one transaction to perform multiple steps, reducing transaction fees and improving user experience.

• ERC 4337: This proposal suggests a more structured account abstraction that uses a new type of transaction called "UserOperation." This system involves several components, including "Bundlers" who package and execute these UserOp. It aims to create a more flexible framework where users can use different types of wallets (smart contract wallets) more seamlessly, potentially supporting multiple types of signatures and mitigating some of the limitations of current EOA-centric models.

ERC 3074 – Deeper Technical Insight

Technical Mechanism:

ERC 3074 introduces two new opcodes, AUTH and AUTHCALL, which allow contracts to perform actions on behalf of a user who provides a signature.

• AUTH opcode takes a user's signature and recovers the address, temporarily authorizing the caller contract to act on behalf of the user.
• AUTHCALL then allows the contract to make any call with the user's authority, including transferring tokens, interacting with other contracts, or performing any action the user could do themselves.

Use Case Scenarios:

• Transaction Batching: Users can sign a single message that authorizes a sequence of transactions. This is especially useful in DeFi applications where multiple token interactions might occur in a single session.
• Meta-Transactions: Facilitates meta-transactions where users can interact with Ethereum without holding ETH for gas, as transaction fees could be paid by third parties.

Security Considerations:

• A significant risk involves the potential for contracts (invokers) to misuse the broad powers granted by a user's signature. Ensuring that only trusted contracts have AUTH capabilities is critical. Any good wallet on the market will disallow signing AUTH from a dApp, problem solved.

Comparing 3074 with 4337

Complementarity vs. Competition

• EOA users can now batch transactions and have their txs sponsored. People that don’t want to move to smart accounts at all or until they can keep their same address will continue to be EOA users and that’s fine.
• EOA users can use proper smart accounts (eg existing 4337 accounts) and ppl will start to experiment with smart accounts more easily. If we do our job well, they’ll then move on to the final category.
• Previous EOA users can now transform their EOA into a smart account. To do this, they can store their seed phrase somewhere safe and remove it from all wallets, meaning they can rely only on the auth of the smart account.