Remember how Ethereum was supposed to be the 'World Computer'?
Ethereum was supposed to be the 'World Computer', with infinite on-chain scaling allowing people to use it as a real computational resource. This was always an obviously dumb idea. All on-chain validation has to be replicated by all full nodes, which extremely redundant.
This plan has failed. Even dialed up to near the breaking point and with a lot of engineering put into raising that limit Ethereum still only supports roughly double the transaction throughput of Bitcoin. This lackluster even for a blockchain performance is because Solidity and the EVM are a garbage platform and there's little work going into the mundane engineering problem of making something better. But even with a great constant factor performance of 10x the capacity it does now the 'World Computer' idea was always a delusional fantasy.
I bring this up now not to taunt them but because this change of plans has conveniently been forgotten about as if it never happened. Instead Ethereum development has managed to make the whole thing much more centralized and exacerbated their MEV problems.
It appears even plans of sharding (also a bad idea) have quietly been abandoned in favor of support for ZK rollups and proof of publication, which is a reasonable plan but still just a plan, and gets a lot less benefit for a lot more work than the mundane boring problem of getting state channels working.
There's this public narrative that Ethereum developers are a bunch of geniuses doing amazing things. They aren't. Development to date has accomplished very little and continues to focus on fun exciting projects instead of the much more prosaic problems like Solidity being a security nightmare where literal billions of dollars continue to be stolen.