| This whole post reads like someone who is frustrated with a new technology they don't understand. >We should stop building key-management plugins and start thinking about a standardizable web API. We must stop training our users to install shitty browser plugins! "standardizable web API" is a total pipe dream. The space moves too quickly for standards. Why is installing a browser plugin bad? Is metamask shitty? The millions of MAU must not have noticed. >We need to make light clients work as soon as possible and become independent from third-party services like thegraph and Infura. What does this mean? You can use metmask and web3/ethers.js with your own node. When $$$ is on the line (as with Uniswap, et. al.), it makes sense to farm out your infra to someone else. Should we also be moving away from AWS and its reliability / productivity gains to appease the author? >We need to improve our client libraries (ethers.js and web3.js) by dramatically simplifying them and making them bug-free (god damn it!)! Yes, first good point of the article. People are working on this. Could be said about ANY technology. Web3 is still VERY early. >We need to take advantage of some of the blockchain's fundamental properties. Most data is immutable so let's start caching things. Not sure what this means either. What are we caching? Where? Why? Are you aware we already (kinda) do this with the strata of different ETH nodes via say syncmode in geth? Or are you talking about front ends, many of which are already distributed via IPFS? |