fede: if we are really breaking contracts, this is wrong
05:15: "never ever break working things for users", and he asks for concrete contracts that would break.
Glamsterdam repricing and "do not break deployed contracts"
lightclient: I asked people to check this a while ago
13:01, and raised it on ACD before that.
Marius van der Wijden: read the 20-line blog post before talking
14:16. The team is investigating internally with auditors and has opened a contact channel, and will decide how to proceed afterwards.
dankrad: "never break an existing contract" is unworkable
14:31. It would mean the gas schedule can never change and that signature schemes can never be deprecated, which will most likely have to happen. Instead we should be clear about which kinds of contracts will not break.
Eli Ben-Sasson: my two cents on the Glamsterdam delay
19:34. He argues that Starknet becoming its own L1 is even more urgent, and that "upgrading blockchains is hard".
kassandra: turn the repricing around?
21:54: instead of raising the gas limit and making a few opcodes more expensive, make most opcodes cheaper.
EIP-8487: separate execution credits and billing credits
EIPs PR #12637: Separate execution credits and billing credits
Opened at 22:58. Both counters start from `tx.gasLimit`: credits use the legacy schedule and govern EVM-visible execution allowances (including CALL and contract creation); billing uses the new schedule and sets the transaction fee. If billing runs out, the whole transaction aborts and the abort cannot be caught, even if credits remain.
ETH as a store of value, not a fee-burning token
Vitalik: why "ETH as store of value" makes more sense to focus on than "ETH as fee-burning token"
19:06. Three reasons: the fee-burning story was never as strong as it looked; optimizing for fee burning would require compromising decentralization and capture resistance (to justify a market cap of about $270 billion, Ethereum would essentially have to act as a world-scale fee platform); and optimizing for store of value is aligned with decentralization and CROPS.
optimizing for fee burning would require compromising decentralization and capture resistance
Vitalik: why "ETH as store of value" makes more sense to focus on than "ETH as fee-burning token", 19:06. Three reasons: the fee-burning story was never as strong as it looked; optimizing for fee burning would require compromising decentralization and capture resistance (to justify a market cap of about $270 billion, Ethereum would essentially have to act as a world-scale fee platform); and optimizing for store of value is aligned with decentralization and CROPS.
optimizing for store of value is aligned with decentralization and CROPS
Vitalik: why "ETH as store of value" makes more sense to focus on than "ETH as fee-burning token", 19:06. Three reasons: the fee-burning story was never as strong as it looked; optimizing for fee burning would require compromising decentralization and capture resistance (to justify a market cap of about $270 billion, Ethereum would essentially have to act as a world-scale fee platform); and optimizing for store of value is aligned with decentralization and CROPS.
Barnabé: security and utility in a programmable store of value
20:20.
Other updates
Ethlabs week 20 update
(X article at 22:45): the Devcon L1 R&D days covered Glamsterdam, Hegotá and I-star; post-quantum was the most important topic (Justin Drake presented the roadmap to be fully quantum-secure by 2028); FCR cuts deposit times from L1 to L2s and exchanges to 12 to 24 seconds, and Gnosis, which adopted it about a month ago, has saved its users more than 12 days of waiting. The update cites Quicknode's 11-04 announcement that the FCR beta is open to all its customers, and other RPC providers already support it. How to use it: enable FCR on a Quicknode mainnet endpoint and keep calling the `safe` tag, with no code changes.
EIPs PR #12610: ML-DSA-65/87 verification vectors and reject cases for EIP-8355
Merged at 04:56, filling out test data for the post-quantum signature precompile.
ethresear.ch: folding vs full-verifier recursion for PQ signature aggregation in leanVM (with negative results)
19:33. On an 8GB M2, hash-based accumulation did not beat leanVM's full-verifier recursion, and folding leanXMSS directly needs at least 7.1 million R1CS constraints per signature, about 450 times leanVM's cost.
Magicians: ERC-8489 Agent Information Flow Paradigm
18:26, matching ERCs PR #2152.
ERCs PR #2152
Magicians: ERC-8489 Agent Information Flow Paradigm, 18:26, matching ERCs PR #2152.
Georgios: almost two years since Ethereum Acceleration
10:35. Paradigm keeps building the code the Ethereum and EVM ecosystem needs (Paradigm: the original Ethereum Acceleration post).
Paradigm: the original Ethereum Acceleration post
Georgios: almost two years since Ethereum Acceleration, 10:35. Paradigm keeps building the code the Ethereum and EVM ecosystem needs (Paradigm: the original Ethereum Acceleration post).
Community
Vitalik disagrees with this week's strong emotional reaction against the anti-data-center movement
17:34.
Barnabé: the roadmap accelerated beyond what anyone imagined this week; "you're not ready for Ethlabs season 2"
23:48.
dapplion asks Coinbase what it would do with enough stake to censor at the attester level
22:29. The answer: "we're gonna fight against policies with which we don't align".