I am keen to find out if there is a timeline in which Flashbots believes it will open source it’s builder? ( this is a followup to Discord )
I think it would both provide much greater clarity to the users of Flashbots Protect as to what the protection effectively provides, as well as make the builder market more contestable by lowering the entry costs.
A AGPL license like in the relayer would be extremely appropriate to ameliorate the risk of helping efforts to monopolize the builder market behind the network effects of exclusive orderflow.
For me personally it would make experiments around user welfare maximizing builder mechanisms much more feasable than having to start from the builder example and having to rebuild the bundle and inclusion logic.
While I can’t answer your question on open source builder either way for now, I wanted to touch on the topics you bring up:
What do you think could be more clear on protection within Flashbots Protect? It offers full privacy, and the transactions are not being used in any way.
What kinds of improvements to user welfare are you thinking of? Any kind of welfare that impacts the privacy of the transactions is a very difficult subject. We are all ears if you want to share how we could improve Protect.
Hi @mateusz , thank you for the interesting points.
Without knowing what the algorithm that is being used to build the block, the assertion that the transactions are not being used in any way is nice to hear, but it be nicer to be able to verify that the code in fact carries it out. Trust, but verify.
A minimal improvement to user welfare would be for the builder to do the backruns at the end of block across exchanges and refund the users who had the suboptimal routing which created the back-run opportunities. There are many interesting AMM mehcanisms that require changing the builder too.
I am delighted to have an open conversation around improvements to the builder to maximize user welfare with anyone who is working on an open source builder. As things stand right now, helping improve the user welfare of the closed-sourced market-dominant builder, runs the risk of further entrenching the dominance. Two sided markets have strong network effects, so the better the closed builder becomes, the harder it is to switch the market away from that instance, even if later its underlying code is public.
I did read the thread by @bert, who I consider a friend. I would be keen to hear arguments for how keeping the builder closed instead of releasing it under the AGPL (like the relayer) serves the purposes and principles the thread expresses.