I’ll give a few notes, but bearing in mind that some arguments are speculative. Generally the post tries to augment the status quo, giving more optionality rather than less, so any configuration we have today may certainly be realised under rainbow staking too. I’ll then argue for why I believe outcomes where protocol rewards are generally better distributed, are likelier under this framework rather than the status quo.
I would expect the dynamics of the heavy layer to be somewhat similar to what the status quo of staking dynamics currently is. The post suggests making more staking gadgets available to the protocol, so that the playing field may be more level between different LSPs, and allowing features such as fast-redelegation. This is nothing particularly novel, borrowing from a few different places such as existing LSP constructions on Ethereum or gadgets deployed in different ecosystems such as Cosmos. Professional operators may still receive most of the stake via delegation, as they do today, but hopefully the environment can be as a whole more competitive and feature greater diversity and resilience.
The post then suggests to create a new type of service, light services, which have fundamentally different economics. Taking censorship-resistance gadgets as an example, and the construction of some inclusion list as mechanism archetype, the mechanism can be designed without “economic security” in mind, its goal really is to surface a signal to the protocol. Liveness of the mechanism matters (we want it to be useful after all) so we want to incentivise people to do a good job at it still, but we need not require them to have much or anything at stake. This means that we can create something akin to token-curated provider selection mechanisms, where the token (ETH) can be trustlessly re-issued as (provider-specific) “light LSTs” to still align incentives between delegators who share in the rewards, and their chosen operators.
Could professional operators still get most of the stake via delegation? Certainly, and we would expect them to be active on the light layer too. But note that without separating light and heavy layers, a.k.a., the status quo, whatever distribution of stake we currently have between professional and solo operators will mechanically also be the distribution of stake prevailing for censorship-resistance mechanisms such as inclusion lists, since staking signs you up to all services at once including FFG + LMD-GHOST (Gasper), sync committees, (execution) block production and potentially in the future inclusion lists too. By unbundling, my thesis is that the competitive advantage of solo operators as providers of light services (the preference entropy argument), as well as the social dynamics at play allowing anyone holding ETH to back risk-free any provider that they choose, will allow solo operators to command a more significant amount of (light) stake than the amount of stake they currently command (which is about 6.5% according to rated.network). By directing a moderate amount of issuance to light services, we may help foster an ecosystem of small node operators who are a strong counterweight to the more professionalised heavy layer.
The outcome where solo operators on the light layer represent less than 6.5% of the delegated light stake is theoretically possible, but imo is unlikely given that the light layer has in some sense weaker restrictions than the heavier one. The next step will be to provide a more complete specification of each layer of services, making some of the claims and constructions more precise, to evaluate whether this is workable.