This is a good article. With Optimism, Ethereum actually looks like a complete, usable solution that can scale.<p>That said there is a lot of complexity around bridging which leads to high transaction fees, long waiting times (e.g. some projects have a 7+ days waiting time for bridging tokens back from L2 to Ethereum mainnet).<p>Another downside to Ethereum is that it forces all this legacy on everyone... As many L2s and other ERC20 tokens will be abandoned in 20+ years, much of their data will still be stored on Ethereum mainnet blockchain. Smart contacts are more like accounts, they aren't like old transactions, you can't just prune them after some time without major ramifications.<p>With old transactions in single-token blockchains, you can potentially prune old transactions while keeping account data and balances the same. The trust and centralization implications of doing this are minimal and constrained to that single blockchain only. With Ethereum, Ethereum community can't just decide to prune old unused smart contracts from the mainchain since it would affect the integrity of L2s and it's noth Ethereum's place to decide whether or not they can delete tokens that belong to other communities.<p>Overall, Ethereum ecosystem has a legacy problem due to the fact that they are multi-token. L2s accumulate unnecessary legacy from Ethereum and also from other L2s.<p>IMO, the benefits of decoupling tokens from Blockchain infrastructure aren't worth it as the value of a token is ultimately derived entirely from its infrastructure/integrations.
From a high level engineering perspective (and not being an expert in blockchain systems), the amount of complexity in the Ethereum ecosystem seems... excessive? And it seems likely the amount of complexity will only continue to increase over time. We tend to prefer 'elegant solutions' and simple systems in engineering, and this seems quite the opposite. Is it truly necessary to solve the problems that it's trying to solve?
As far as I’m aware, there’s no way of validating something without observing it. If that’s true, then you will have to trade the ability to validate the entire chain in order to scale on chain, at which point, who cares what’s on chain vs off?
I would like to point out that the alternative method of focusing on maximizing execution bandwidth and speed at the base layer is being done on Solana, with great success.<p>Layer 2s if they’re needed are as easily added on top as it is for Ethereum. But Solana is so fast and cheap already right now that it hasn’t been necessary.<p>I’m of the philosophy that since you only get to build one base layer, you should really spend the time to do it well and have it designed to be as efficient and performant as possible.