Ethereum's latest account abstraction proposal is its best yet. EIP-8141 [1] adds a new transaction type called a frame transaction.
Existing transaction types operate in the context of a single identity, whoever signs the transaction. Adding the ability to bundle operations from multiple entities into a single transaction was done at the application layer, creating confusion for users and using a lot of gas.
The frame transaction includes a list of frames (essentially calls) that can originate from different identities. This allows for gas sponsorship that is much simpler than a paymaster [2]. Nothing is worse than user churn because they don't have a gas paying asset. It also allows for native multicall. No more sending two transactions (approve/transferFrom [3]) to deposit a token into a defi app. This is huge for apps, every additional click is churn.
Perhaps most users will send frames to the network instead of transactions in the future. Bundles [4] can be implemented natively and maybe there is a way to prevent unbundling at the app layer.
It reminds me of the Solana instruction [5] model, where frames are similar to instructions.
Probably too late in the dev cycle for this to be considered in the Hegota upgrade (late 2026) but generally think the only EIP that is more impactful than this one is 6 second slots [6]
[1] eips.ethereum.org/EIPS/eip-8141
[2] alchemy.com/overviews/what…
[3] ethereum.stackexchange.com/a/145358
[4] docs.flashbots.net/flashbots-auct…
[5] solana.com/docs/core/inst…
[6] eips.ethereum.org/EIPS/eip-7782
Existing transaction types operate in the context of a single identity, whoever signs the transaction. Adding the ability to bundle operations from multiple entities into a single transaction was done at the application layer, creating confusion for users and using a lot of gas.
The frame transaction includes a list of frames (essentially calls) that can originate from different identities. This allows for gas sponsorship that is much simpler than a paymaster [2]. Nothing is worse than user churn because they don't have a gas paying asset. It also allows for native multicall. No more sending two transactions (approve/transferFrom [3]) to deposit a token into a defi app. This is huge for apps, every additional click is churn.
Perhaps most users will send frames to the network instead of transactions in the future. Bundles [4] can be implemented natively and maybe there is a way to prevent unbundling at the app layer.
It reminds me of the Solana instruction [5] model, where frames are similar to instructions.
Probably too late in the dev cycle for this to be considered in the Hegota upgrade (late 2026) but generally think the only EIP that is more impactful than this one is 6 second slots [6]
[1] eips.ethereum.org/EIPS/eip-8141
[2] alchemy.com/overviews/what…
[3] ethereum.stackexchange.com/a/145358
[4] docs.flashbots.net/flashbots-auct…
[5] solana.com/docs/core/inst…
[6] eips.ethereum.org/EIPS/eip-7782
• • •
Missing some Tweet in this thread? You can try to
force a refresh
