Hi @mattstam, thank you for your thoughtful reply!
Indeed, this post was an attempt at laying out a somewhat coherent vision. If the post managed to convince you of the coherence, it performed its function adequately. For the complexity part, I can discuss a little bit about what are some of the challenges I foresee moving forward.
The starting point would be the “three levels” I was describing in my answer to Alon. As proposals such as EIP-7251: Increase the MAXIMUM_EFFECTIVE_BALANCE are discussed, I believe enshrining some measure of operator–delegator separation at the same time may leverage a lot of the work currently done towards specifying EIP-7251. The idea of increasing MaxEB is to allow a single message to carry more stake; enshrining the operator delegator separation would be providing in-protocol a way to functionally distinguish “operator stake” from “delegator stake” on such a message.
The next step would be to offer in-protocol the ability to tokenise such messages. This is essentially the purpose of the LSM, creating shares out of delegated stake. I currently consider this a more “nice-to-have”, and its second-order effects seem to exclude anything worse than the status quo (it provides more flexibility, without removing functionality). Specifying this in the context of Ethereum does seem like a valuable exercise either ways, as it would surface some relevant questions for sure.
Moving to the “next level” (Level 2), again quite a bit of complexity seems embedded into the design and specification of some censorship-resistance mechanism. An extra mechanism is necessary for light delegators (essentially, delegated-staked and un-staked ETH holders) to give weight to the heavy operator of their choosing for the construction of the inclusion set. How this weight feeds into the mechanism itself needs further thought. In the context of inclusion lists as specified in EIP-7547, an idea could be to sample the producer of the list according to the light distribution of stake. With more sophisticated, committee-based mechanisms such as Multiplicity, the sampling of the committee could be again performed according to the light stake distribution.
The real step into the unknown for me is Level 3. Allowing a separate set of participants to become light operators would require something akin to a “second deposit contract”, or at the very least some way of tracking the registration and assignment of weight to light operators. This point may be technical, but thornier questions lie imo in the relationships between the heavy operator set, light operator set, and full nodes more generally. I describe some ideas in the section Mapping the unbundled protocol, essentially we would have to decide on what kind of coupling we wish to go along with, to endow heavy operators with the function of enforcing outputs from the light set such as light certificates. More generally, this also has unknown second-order effects on the political economy of the network, and I agree that much more careful thought is needed.
As you also mention, it would be important to do all this with liquid claims and re-staking in mind. This is something I’ve been thinking more about with recent posts, but still incomplete and more about definitions than properties anyways. If you have some specific questions in mind, I would be happy to discuss them further, feel free to send a DM or reply here 