The design space of Ethereum rollups is huge. I am not going to deny that assessing @0xFacet was challenging for the whole team as its design is not what you would expect from a typical rollup. Nevertheless we love being challenged and as a result some new dashboards related to canonical and externally-bridged assets are in the works. The most challenging aspect of @0xFacet is undoubtedly its non-standard gas token. It's use challenges the core definition of a "canonical" asset as normally we assume that gas token must be bridged canonically but what if you don't really "bridge" gas token at all ? What is a canonical asset and canonical bridge then ? Read the full review and @0xFacet announcement for more technical details.

I have a very different take * many projects managed to brainwash everyone that posting data to AWS or 1/1 DAC is "good enough". End users don't care but we (pro community) should make them care by strongly discouraging such architectures. Instead they are actually encouraged by some rollup and RaaS teams * The price discovery mechanism in Ethereum and Celestia is off-putting - projects want Web2 price reliability so that they can maintain a low and predictable price for their users. Sudden price peaks are not acceptable when you operate at scale * DA projects should focus on promoting legitimate use cases and not centralised projects that won't even bother to publish their code so that the community can check their claims. If all we build are Hyperliquid clones then we will all go to zero. Nobody in the long term wants centralised, custodial exchange that pretends to be decentralised * Data withholding attacks (literally the main reason to use DA) are hard to explain to users - we could do with more education and in general the community awareness of a problem. The vast majority of pro users don't see the difference between data publishing and data retrieval. Non-pro users don't see an issue at all as for them data "appears" to be always available so it's hard to explain to them what problem we are trying to solve here

Why you should check @cartesiproject * Having a honeypot to incentivise hackers to break the system is much better than having EOA with upgrade powers for beta launch * If working as designed, you have an SDK at your disposal to quickly and cheaply build custom AppChains that can be Stage1 or Stage2 at launch easily. Cpp code for honeypot is minimal, check it out * Because it's optimistic, there is no need for you to run complicated prover infra so costs of running the chain are minimal and you can make it fully permissionless from day 1 Now we need more funds in a honeypot and researchers / hackers willing to look at their new fraud proof system

Study L2 trust assumptions. Always assume the worst. Act accordingly. Security and decentralisation doesn't matter until it does.

Are alt-DA L2s a failed experiment at scaling Ethereum ? Few months ago @l2beat announced that having an abysmal DA setup where one or two addresses can rug all users' funds, a practice that became prevalent in the space, will result in change of the classification of these systems. They will no longer be considered Validiums/Optimiums. I strongly believe that we, as the community, want systems that give users some minimum security and decentralisation assurances, and that means having a proof system and - if data is not posted to Ethereum - have some basic guarantees about data availability. Turns out project don't care. They don't care because their only goal was to grab some VC money, and get minimum user's attention just to launch a token in a way that's as cheap as possible. In practice this means using one of the RaaS providers and certainly not paying Ethereum for data. If they used @celestia, @AvailProject or @eigen_da they did not bother "informing" Ethereum bridge about it making all funds in the bridge from Ethereum easily ruggable. These projects will soon fade away - most are already ghost chains w/out any activity. But maybe some projects are legitimate but users simply don't care. In this case the decision not to secure Ethereum bridge is a conscious design choice - however there are much better architecture designs than launching a completely insecure "L2" on RaaS. Two years ago all signs suggested we will have thousands of rollups and @l2beat won't be able to keep up. Now it seems more likely that we might be seeing an opposite trend - just a couple of dozens of really good ones and a huge graveyard of failed projects that nobody wanted in the first place.

For some time now @l2beat is tracking DA usage for various DA layers and projects. If you are interested how the recent Pectra upgrade affected Ethereum blob space, check out

What is your gut feeling re: security of the bridge that claims to use MPC ? Would you trust it (bro, multi-party computation, moon math, that has to be secure, no ?) or you're skeptical (who the f* are the signers ?). Check the 🧵 before answering

AnyTrust chains from @arbitrum were introduced back in 2022. They have argued back then that having, say, 20 committee members with the assumption of just two being honest is a good security model. And it's hard to argue

It's been exactly a year now that @Immutable (powered by @0xpolygon) announced their zkEVM chain. You can check the docs to learn about their zk tech

A (not so) short explanation why we have introduced a moniker "AppChain" to the Stages system