Thanks for the post! I liked it.
I have a couple questions wrt the overall directions.
-
If we assume that - as you suggest - local builders shouldn’t limit the throughput of the network (and even I don’t like it, it’s true and we should face it).
And if we give block building to an oligopoly effectively (MEV builders). Then, we need to lower the hardware requirements such that we can build the most protocols that help validators to fight/retain a power equilibrium right??.
That being said, would that mean that the focus for PR&D should be working on exactly that? (ie. lowering hardware/bandwidth requirements for validators/IL builders/IL attesters and also creating protocols like FOCIL, Rainbow, Distributed Validators etc… that allow having more validators that have more tools to fight for power equilibrium within the network?). -
Would we double down on externalizing heavy parts of the protocol to these parties? Meaning, now we put block building onto them which will entail holding state, generating proofs(at some point), managing A LOT of compute and bandwidth etc…
Is the plan to keep designing and basing some of the “scaling L1” on the fact that these parties with this much compute power exist? Ie. ZKEVM-based proof of chain validity for instant sync. Hold both cold and hot state in case state-expiry happens to reduce the burden on users and dApps/wallets? etc… -
Final question, is more related to the section
How can we ensure we hear the community and at the same time the decision doesn’t drag for years so that we can figure out what the priorities are and work towards them??.
For instance, this recalls me the different stateless flavours (see this post from Julian). If we decide we are scaling by having these super-heavy builders. This makes it easier to decide to go towards weak statelessness rather than other options.
And also helps scoping how much of a priority things like state expiry are.
Thanks, and nice to see discussions like this being triggered. Feels we should decide. As the only clear thing we have now is that we need more blobs. But there’s a lot more to do 
