MegaETH can censor, front run & steal all user funds without delay! It is so centralized that it literally runs from a single server Flexing capacity & speed in this case is not a flex at all Less than 0.2% of fees make it back to ETH, which is also exceptionally parasitic: 🧵 If we truly care about decentralization & its utility. Then we should stay clear of centralized L2s & advocate for high capacity L1's such as SOL, SUI, & NEAR instead Admin Key Risk: It is the MegaETH smart contract that lives on ETH that exposes users to the risk of theft & loss of funds. As the smart contract can be "upgraded" through a 4 out of 8 multisignature, also known as an "admin key" Such a change could include new rules that could send all tokens present in the L2 to a new address. Effectively draining all user funds deposited in the L2. This is currently the case for all major L2s! We might not have seen any major exploits of this nature in major L2s in the past, but we have in other smart contracts with the same security setup. That means it is only a matter of time before this happens with one of the big L2s, hurting countless innocent people in the process Ofcourse, like all other L2s, MegaETH promises to remove the admin keys one day, but if the history of "L2 scaling" proves anything, it is that the incentives rarely align for that to actually ever happen. As promises should carry very little weight in this industry, based on our shared history Centralized Sequencer Risk: There is currently a single permissioned sequencer that can therefore censor & frontrun. As it alone determines transaction ordering. Allowing it to ignore some TXs (censorship) & prioritize others (MeV). Throwing all pretense of decentralization out the window This is literally a single centralized server sitting in a data centre with insane hardware requirements. As it costs over $100k to operate the hardware alone, that is for a single sequencer per year. Making it at least 20x more demanding than a SOL validator, for example In MegaETH's defense, there are two additional backup sequencers that can be swapped in if a failure occurs. This also lays the critical technological groundwork for a larger sequencer set in the future, something that they plan to do. It is yet to be determined if this will be stake-weighted or PoA-based The irony does not escape me that the only way to "decentralize" an L2 is to recreate the same decentralized consensus that enabled blockchains (L1) in the first place, putting us back at square one. That is how decentralizing L2s is a lot like re-inventing the wheel Parasitic Economics: As long as L2s such as MegaETH help to justify & in some cases lobby against ETH's L1 scaling significantly, the relationship remains fundamentally parasitic. As it slowly kills the host, while private L2s profit & siphon off ETH users to new chains What makes MegaETH particularly bad in that regard is that it does not even settle directly on ETH but on EigenDA instead! We can figure out the percentages by looking at the cumulative cost to ETH: Average cost per L2 user operation (UOP, which approximates a transaction or batch element in this context): $0.000006 We can then compare that to how much MegaETH actually charges its users: Using the targeted fee instead of the average fee, since the latter is higher: $0.003 user fee per tx. Percentage ≈ ($0.000006 / $0.003) × 100 = 0.2% This ratio will only get worse as MegaETH does more TXs. This is efficient from MegaETH’s perspective while being wholly parasitic from ETH’s perspective! Far from this being a reciprocal relationship, this strikes me as being more extractive. MegaETH should be called MegaEigenDA or MegaCentralized instead! False Equivalance Comparing the performance of MegaETH to a permissionless & decentralized system such as ETH, SOL, SUI & NEAR is totally bonkers. It is an unfair comparison because they are not the same thing at all. It would make more sense to compare it to traditional centralized server architecture MegaETH is impressive from an engineering perspective. However, achieving the same performance on any decentralized system is far more impressive. Having a single centralized server solves many of the bottlenecks that real cryptocurrencies have to work around The 10ms speed claim is also misleading. The time it takes for the speed of light alone to travel around the Earth is 130ms. So, the speed you experience depends on how far you are from their single centralized server. That is where decentralized systems have the potential to be even faster than centralized systems one day & truly global. Thanks to the type of multi-leader architectures, the likes of SOL & SUI have been innovating recently Credit Where Credit Is Due MegaETH is the most interesting & best L2 from my perspective, as unlike other ETH L2’s, it actually scales! Not sure how much that praise is worth coming from one of the biggest L2 critics, but this needs to be acknowledged What also needs to be acknowledged is that @bread_ is a great guy. He helped review this critique & he did so very fairly & graciously. Helping to correct any inaccuracies, even when he obviously does not agree with my conclusions, but we can agree on the facts. So, despite our strong disagreement on this point, there is a great mutual respect here. Ultimately, we have the same goals in mind; we are just taking radically different paths to achieve them MegaETH is also my favorite L2 for not pretending to be something it is not. It embraces its centralization as a strength & confidently bites the bullet on trade-offs Conclusion Anyone who cares about decentralization & everything that entails. Should stay clear of L2’s, including MegaETH High throughput L1’s cannot censor & cannot steal user funds, MegaETH can The blockchain trilemma has effectively been solved. We do not need to sacrifice decentralization on the altar of utility. We can have our cake & eat it too That is what makes the entire concept of “L2 scaling” an anathema. An unnecessary compromise in what matters most to this entire movement: Decentralization & more importantly, the goals it was meant to accomplish. From censorship resistance, financial sovereignty, credible neutrality & more. MegaETH supports none of these causes & only drives more people in the opposite direction That is how ETH’s misguided quest to preserve decentralization has paradoxically pushed most of its users away & into centralized systems such as MegaETH. I do not blame L2 founders for finding opportunities in the free market. My gripe lies with the ETH’s leadership for not scaling their L1 & consequently creating this market segment in the first place! MegaETH is the best L2, for whatever that is worth. As it remains an incredibly low bar. As all of the major L2s can still censor, frontrun & steal user funds today. However, none of the major L2s are actually capable of significant scale. This is where MegaETH at least gives something in return for that horrible trade-off We are all free to innovate in our own unique ways, but always keep in mind to “never make a deal with the devil unless you’re prepared to lose”