So I’m probably responsible for this (I’ll let Vitalik correct me if I’m wrong). Thank you for discussing the issue further.
That won’t be me! I’m afraid your concerns are valid.
If you haven’t read my posts on this I urge you to. Flashbots style full block auctions incentivize both private dark pools and the formation of censorship-as-a-service markets (bribes to exclude other people’s transactions). Both of these come with powerful network effects.
The dominant builder will be the most exploitative by definition, and the endgame is a perfect extortion economy (maximal taxes levied on transactors) that fails to compete with centralized alternatives.
CaaS undermines the whole idea that bribery ordering (and therefore Ethereum) can be censorship resistant because you can be bribed more to exclude a tx than to include it. That is not a small point, it’s fundamental. Getting paid to censor is MEV too (what I call Private MEV).
Flashbots position on MEV is that if it can be extracted, it will be. As such, they should believe more than me that CaaS is inevitable. By their own logic, they must allow censorship and censorship markets or they will be limiting MEV extraction and will therefore create an incentive for extractors to buy validators and centralize.
As significantly, because we have seperated block building from the need to own a validator, there is no requirement for the dominant builder to have an Eth denominated stake at risk. They can exchange the funds needed to win blocks in auction just-in-time. As a result, if the value they can extract from blocks results in damage to Ethereum/Eth price, it may actually be in their interests to do so (quite unlike the assurances of a 51% attack). I call this the Unstaked Hijack.
All of this comes from the long term issue of block content being fully trusted/centralized in Ethereum. Full block Flashbots auctions represent a full and final failure to address this and will have severe consequences.
I have been charged with organizing a discussion group about this with some Flashbots people and other interested parties in a few weeks time. Perhaps you’d like to take part?
Yes, this is a very good point.
It is literally banning altruism.
@fradamt proposed crList after I discussed these issues with him. It’s preferable to nothing, but formally non-existent in that it relies on altruism and is less censorship resistant than what we have under PoW now.
I’m glad you appreciate that is a weak argument. I would say that despite other people trying to build blocks regularly, it does not alter that a dominant builder will likely monopolize block content in the ways I have described above. Arguably, it just means they have to be more exploitative to maintain their position.
You’ll be able to do that with my zeromev project (launching in a few weeks). It tracks exactly which transactions were inserted by miners and which are from the mempool in each block.
But why wait for it to happen? It’s like running Ethereum on a single node and waiting until it double-spends to decentralize it.
I think we need to look very seriously at base layer decentralization of content proposals like this one instead (my only real objection to this being that it is optional).