for the end user:
- they see far fewer reverts
- and get much closer to theoretically optimal price (we'll share some data backing this up in the coming days)
present state:
1. trader comes to an aggregator and requests a quote at block n
2. aggregator shows a quote based on state of liquidity at the END of block n
3. trader signs approval and locks in the route, and blocks have moved to block n+1
4. txn is now sent to mempool
5. txn gets included in block n+2 if we're lucky, liquidity has already moved somewhere, that route the user locked is now far from best price
6. txn is executed and user got a much worse price because the route was stale
post fynd inblock:
1 - 4 remains the same.
5. builder sees special fynd inblock txn, they run fynd (super fast router) DURING block building on state of block n+2 right before inclusion postion (ie. current block)
6. they replace the route if/when there is better route computed live
7. txn is now executed, and user gets price closer per realtime state
possible only because fynd can run trustlessly within builders and no one needs to worry about leaking state.
user swapping a billion dollars gets 2bps more. tradfi only theorised this but could never ship.