The shasta inbox is built in a way that rewards both aggregating multiple consecutive proofs and finalizing multiple transitions. But to take full advantage of this it requires the off-chain client and prover software to be aware of this, and be smart and coordinate.
This is still very broad, but we should discuss:
- what should be the default strategy for the proposer for when to submit proofs
- proofs should be submitted in a way that don't force the proposer to finalize often, but instead when there is a longer chain to finalize
- how can multiple proposers coordinate to submit proofs together to lower costs(long term some prover market, but short term maybe Taiko or some entity generates, aggregates and submits the proofs?)