[0x64]

Not to mention, a new groundwork for light clients was laid out in the first upgrade to the beacon chain. Come merge, clients could effectively poll for e.g. balances directly from the network, instead of going through Infura.

More on Ethereum light clients from Chainsafe, the builders of Lodestar, a Typescript implementation of an Ethereum consensus client: https://blog.chainsafe.io/the-road-ahead-for-ethereum-light-...

TL;DR: "Light clients enable more people to participate as first-class citizens, verifying on-chain data without relying on single and centralized JSON RPC endpoints."

[davidgerard]

"Could" is another word for "they don't."

The typical user could use something other than Infura - but they don't.

The typical user could use some future system that doesn't exist yet - but they don't.

The typical user could cash out somewhere other than Coinbase - but they don't.

The typical user could use stablecoins other than USDC - but they don't.

This is covered in the article already. Default behaviour matters, a lot. Almost no Ethereum users are going to telnet 30303.

[randomran01234]

How much of this is a real problem?

Same argument can be made for key management. A typical user could use non custodial wallet, but they don’t. But that option exists, for when they become unsatisfied with their centralized wallet custodian, or when they want to change custodians, or when they want to use a multi sig that allows for a mix of custodial and non custodial.

Importantly: the option exists. The option to use non custodial EOA and smart contract wallet, DAI or RAI, Alchemy or a custom node RPC, Kraken or another CEX or DEX, the option is present and is chosen by plenty of users.

[0x64]

I don't disagree. However, I see no reason why light clients would not become the default option, as they effectively remove the trust placed on an RPC provider.