Joseph Poon

@mteamisloading @VitalikButerin @tyneslol I am confident the real reason is laziness. I think if it was included as default in the software no one would change it. There is no reason for L2s to include reverted transaction in its L2 blocks, let alone in L1 blobs. Every L2 should not include reverted TXes as the default.

@VitalikButerin @tyneslol On L2s I think someone did the math and realized they would make a lot more money from gas fees if they included reverting transactions With that being said it does seem like there is a lot of room to differentiate between arbitrage spam and other users

Revert protection is a scalability and UX improvement. Less gas spent on noop transactions, less downside risk for racing to land a transaction on chain. EIP-7640 is underrated. It adds revert protection to the protocol itself github.com/ethereum/EIPs/…

Hi Mark, thanks for talking about this. I believe this is the most important change needed for keeping Ethereum decentralized and acting in the users' interest as the first priority. I believe a big part of Base and the OP stack taking market share from Solana is fair processing of transactions, and expect this growth in market share to further (and dramatically) expand this year.

psst hai @jcp

This proposal was made because I am confident that the principal problem with Ethereum centralization boils down to revert costs. It seems counterintuitive, but the risk of paying for reverted TXes means people inevitably gravitate towards centralized providers. This is especially magnified in AA, as bundling moves deeper into the mempool. It is in the interest of keeping the spec as minimal as possible to not specify any solution for a DoS resistant p2p mempool, as solutions do not require a consensus protocol change whatsoever (only the revert protection part requires it). An example way to achieve this on a public p2p network is an opt-in network which requires the submitter (e.g. AA providers) to have a minimal amount of ETH staked, which affords a certain amount of throughput (change it into a stake-weighted flow problem). These can be innovated on by many, and I would presume AA developers would find it an interesting area to explore. If you want decentralized AA, where anyone can submit/bundle transactions on a publicly gossiped network, revert protection in consensus is the ONLY known way to do so. There are significant centralization pressures with current directions (AA, bundlers, sequencing, etc) on-chain, and revert protection is the fundamental prerequisite towards developing further solutions, without there is no way to even begin tackling the centralization pressures because decentralized gossiped systems are at a fundamental disadvantage to centralized providers which offer this as a service because you "TRUST" them.

@tyneslol This seems incompatible with DoS-resistant public mempools? And on L2s, surely the sequencer can just not include those transactions I feel like the more interesting approach is to expand EIP-7701-style AA, to allow transactions to have more complex preconditions.

👀

Adventure Time "Come Along With Me" finale song :))) youtube.com/watch?v=mIkS8e…

@DeaterBob Lightning on Bitcoin matters a lot wrt decentralized scaling! It is about multiple *potentially complimentary* strategies, and exploring different technical tradeoffs. Work on Plasma can end up being an alternative path to fedpeg sidechains.

While many of us care a lot about decentralization, that isn't how we should pitch it. For most, decentralization is the means, the ends is mass coordination and agreement. Scalability is how we are going to get there and for that reason I am also deeply focused on Plasma on ETH

@homakov Of course, it is experimental now, but can be used to demo real non-monetary usefulness of the blockchain today

@homakov It works, as in there's a demo C library (chjj is a beast!) which does light client name resolution, ssh key exchange works now. Light client via merkelized ommitment to global name state. Mitigates a lot of squatting (but not all) via submitting ownership proofs directly onchain

I'm contributing to a new Decentralized CA and naming project with a bunch of friends. The goal is to replace the Certificate Authority system entirely and to not have that annoying SSH fingerprint question. handshake.org

Excited to be backing Handshake.org- not only because it is a big idea but they are focused on community alignment, giving tokens to who matters. 1. Raise money/sell small stake of tokens 💰 2. Give said money away to FOSS 💸 3. Give rest of tokens away to FOSS Devs 💻

@bosstiat Different. Let's Encrypt does awesome work, but uses the existing CA infrastructure and it's trust models. This is an experimental decentralized naming chain and builds upon the fantastic work by Namecoin, ENS, and Blockstack.

Certificate Authorities and DNS are the foundation in which all of the internet services stand upon. The green lock icon in your browser may not be as secure as you think, there have been many known security failures (see DigiNotar in 2011).

The ssh client can verify that the name record matches the server's pubkey record, it's working on the testnet blockchain now (it's experimental though, but has SPV too!).

Don't mean to be too preachy about it; there were lots of discussions after the event with how to positively handle these types of situations in crypto.

Both @Excellion summarizing his views on Bitcoin Core's position and @rogerkver emphatically presenting his concerns on LN&BTC should be seen as an example of (mostly) respectful discourse. It's important to recognize that honest, respectful disagreements are preferred.

The confrontation at the conference should not be seen as a normal response to disagreements in the crypto ecosystem. It was an extremely strong response to extreme, intentional deception. True, respectful disagreements of philosophy should be celebrated.

@PeterRizun People said I should crash the party, I heard there was a big focus and discussions on LN, but got trapped by the Japanese snacks.

Small blockers: is it time to boycott the Lightning Network? Its co-inventor showed up for the Satoshi's Vision Conference in Tokyo! (Or maybe LN isn't the panacea Blockstream/Core would lead you to believe.)

Announcing lnd v0.4 beta! blog.lightning.engineering/announcement/2… 🔥⚡️⚡️🔥

Just posted a quick Plasma MVP Overview: youtu.be/jTc_2tyT_lY 🐹🦔🐹🐸🐹🦔🐹