Intentions: Potential Solutions to the Complexity Issues of DeFi

Before the Luna crash, I was managing stablecoin yield strategies for a cryptocurrency novice friend. We operated weekly through video calls, diversifying funds into DeFi protocols across various chains. This process was extremely cumbersome, involving a lot of approvals, transfers, swaps, deposits, and so on, spanning multiple DEXs, yield aggregators, and cross-chain bridges.

For example, converting USDC to FRAX/DAI LP on Polygon and staking it requires 12 transactions:

1. Exchange USDC for DAI on Ethereum (2 times )
2. Cross-chain USDC and DAI to Polygon ( )
3. Merge USDC and DAI on Polygon ( transaction )
4. Deposit LP into the staking pool (2 )

This process is very complex to operate manually. But from a higher perspective, our goal is clear - to convert USDC on Ethereum into staked FRAX/DAI LP on Polygon. This is known as "intention," and the 12 transactions are the specific steps to achieve this intention.

A powerful routing algorithm can greatly simplify this process, reducing it to 1-2 steps. Users only need to express their final goals, and the algorithm can return the best path and process the transactions. This mapping structure is known as "intent" and is an important direction for the future development of middleware in Ethereum.

Although there is currently no unified definition of intention, it is generally believed to be "declarative", meaning that the user states the expected outcome, and the "solver" provides the implementation method. Unlike trading with specific parameters, intention requires third-party mapping and the setting of constraints to narrow down the possible set of paths.

Intent-based architecture already exists within the EVM. When you use a DEX, it automatically seeks the best trading path. For example, the Curve interface finds the optimal execution path among multiple LPs and provides information such as slippage.

In addition to trading aggregators, there are several other types of intentions on Ethereum:

1. Limit Order
2. CowSwap-style auction
3. Gas Sponsorship
4. Entrust
5. Transaction Batch Processing
6. Cross-chain exchange

The intent is essentially similar to a limit order, consisting of two parts: the final state expected by the user and the transaction initiated by the solver.

A core feature of intent is data exposure. When users sign intent messages, they are essentially agreeing to extract MEV for convenience. Intent is not directly broadcast to the Ethereum mempool but is stored in a private off-chain Interpool. These Interpools can be permissioned, permissionless, or hybrid.

Currently, the most popular Interpool is centralized and permissioned, not sharing information with competitors. This creates a monopoly on most intention-based transactions and introduces the risk of additional fees.

Intent can be likened to Robinhood's payment for order flow (PFOF). Users sell order flow in exchange for "free" trading, and market makers profit from the bid-ask spread. Intent is also a form of PFOF arbitrage known as MEV. The arbitrage opportunities created by long-unsettled orders may be more valuable than regular trading.

Unregulated solvers may provide the worst routes as they yield the highest profits. Users need to carefully choose solvers and leverage their negotiating power to force solvers to bid against each other. CoWSwap employs a batch auction design to find the best settlement price for traders through open competition.

Currently, some protocols are developing intent-based hybrid system infrastructure. Although the winner has not yet been determined, this is part of the middleware revolution occurring in the crypto space, aimed at improving user-friendliness. In the future, intent is expected to be applicable to a broader range of data and use cases, opening up new possibilities for projects like Frax.