LEAD: Researchers propose computation workers that keep no state. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187]

TEST EDITION | 28 October 2026

A research post proposes separating computation from custody of application state. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187] Its “L1.5” pattern places settlement directly on Ethereum. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187]

The proposed arrangement divides responsibility among wallets, Ethereum and replaceable computation workers. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187] Wallets retain private user data, while Ethereum stores shared state and verifies transitions. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187] Workers calculate results and produce proofs without retaining anything required for the next job. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187]

The authors specify four requirements for private settlement: replaceable workers, Ethereum verification, payment without identity linkage and submission without a persistent operator account. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187] Their distinction concerns who holds state, rather than who sequences transactions or executes them. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187] They describe an architectural pattern rather than necessarily another blockchain layer. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187]

Native account abstraction and proof aggregation supply the proposed machinery for submission and verification. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187] In the authors’ order-book example, a matcher proves trades and another matcher can replace it. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187]

The authors acknowledge earlier designs that already separate some computation from state ownership. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187] Their application sketches also depend partly on Ethereum upgrades that remain proposals. [https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187]

ALSO TODAY:

[NETWORKING] A researcher identifies a separate quantum migration problem for networking. [https://ethresear.ch/t/inventory-of-pq-p2p-migrations-in-the-consensus-layer/26183] Nyx Foundation’s adust09 argues that replacing validator signatures leaves discovery identities and connection handshakes unchanged. [https://ethresear.ch/t/inventory-of-pq-p2p-migrations-in-the-consensus-layer/26183] The 27 October 2026 inventory selects neither an algorithm nor an activation schedule. [https://ethresear.ch/t/inventory-of-pq-p2p-migrations-in-the-consensus-layer/26183]

[VOTING] Proof of One seeks scrutiny before claiming independent review. [https://ethresear.ch/t/proof-of-one-one-person-one-identity-polling-with-encrypted-ballots-design-and-reference-code-requesting-critical-review/26184] Its maintainer’s 27 October 2026 post describes passport-based enrolment and encrypted ballots. [https://ethresear.ch/t/proof-of-one-one-person-one-identity-polling-with-encrypted-ballots-design-and-reference-code-requesting-critical-review/26184] The maintainer says AI agents drafted most design and code, with no independent human review completed. [https://ethresear.ch/t/proof-of-one-one-person-one-identity-polling-with-encrypted-ballots-design-and-reference-code-requesting-critical-review/26184]

[VERIFICATION] Researchers report proofs for Ethereum’s withdrawal-request contract. [https://ethresear.ch/t/lean-verification-of-eip-7002-system-call-success-and-fifo-requests/26180] The 26 October 2026 post reports Lean proofs of system-call success and request ordering against the Jaune execution model. [https://ethresear.ch/t/lean-verification-of-eip-7002-system-call-success-and-fifo-requests/26180] Its stated system-call ceiling is 210,000 gas against a 30 million gas allowance, under the covered forks and canonical-code premise. [https://ethresear.ch/t/lean-verification-of-eip-7002-system-call-success-and-fifo-requests/26180]

READING LIST:

[https://ethresear.ch/t/ephemeral-rollup-the-new-l1-5-of-ethereum/26187] Read the requirements behind the proposed architecture. [https://ethresear.ch/t/inventory-of-pq-p2p-migrations-in-the-consensus-layer/26183] Follow the distinction between encryption and authentication. [https://ethresear.ch/t/proof-of-one-one-person-one-identity-polling-with-encrypted-ballots-design-and-reference-code-requesting-critical-review/26184] Examine the maintainer’s explicit trust limits. [https://ethresear.ch/t/lean-verification-of-eip-7002-system-call-success-and-fifo-requests/26180] Read the assumptions attached to the proof claims.

METHOD: The sweep covered Ethereum Research, the Ethereum Foundation blog and the rehearsal archive. Coverage targeted the 48-hour date span from 26 October 2026 to 28 October 2026. The rehearsal page failed to load directly; its GitHub source and both linked editions were fetched instead. Both editions led with Hoodi’s postponement. [https://raw.githubusercontent.com/openinterest-ai/paper/main/rehearsal.html] No candidate item was cut for failed verification. Technical claims remain attributed to primary authors; no included claim rests on secondary reporting. Proof code was not independently executed.