Francesco

@fradamt

31 Following

153 Followers

Francesco pfp

I know this might be more confusing than helpful, but trying to condense into cast-sized bites is a bit too much for me in the morning. Feel free to DM if you want more details @ellied

2 replies

0 recast

2 reactions

Francesco pfp

Two-slot PBS kinda requires us to get rid of committees (SSF, essentially) for complicated LMD-GHOST security reasons. This also gives us reorg resilience, i.e., any honest proposer (execution or beacon) can ensure the success of their proposal (in fact, reorg resilience is what fixes two-slot PBS)

1 reply

0 recast

2 reactions

Francesco pfp

Another complication is that now you can't base your bid on whether the tip of the chain is stable. To help with both, we'd want the execution proposer to have its own slot and be able to ensure inclusion of its payload, i.e., two-slot PBS with reorg resilience

2 replies

0 recast

1 reaction

Francesco pfp

Have to think about it more, but the rough idea is that, if an execution proposer is not chosen by the previous beacon proposer, it is even more important that they be able to deliver their block regardless of what that beacon proposer does.

1 reply

0 recast

1 reaction

Francesco pfp

0 reply

1 recast

2 reactions

Francesco pfp

Down horrendously

1 reply

0 recast

6 reactions

caspar pfp

wrote down some thoughts with @ansgar.eth, exploring the status quo of staking economics, its drawbacks as we see them, and what the endgame of staking economics could look like + what we should do in the short term until endgame.

4 replies

7 recasts

31 reactions

Francesco pfp

If it's any consolation, we suffer from the same ilness

0 reply

0 recast

1 reaction

Francesco pfp

Can you expand on "beacon block throughout holds back decentralization"?

0 reply

0 recast

0 reaction

Francesco pfp

It would be an even more radical redesign, where stakers are really just attesters. There may be ways to do it which preserve all properties we care about, not sure. Execution tickets is already pretty radical and may or may not be the way to go, I am somewhat skeptical about something even more extreme :)

1 reply

0 recast

2 reactions

Francesco pfp

Everything in the beacon chain works through beacon blocks. For example finality is defined by the state transition processing attestations in a particular chain, up until the state of that chain finalizes something new. Also the whole incentive system is built around on-chain inclusion of messages.

1 reply

0 recast

1 reaction

Francesco pfp

mmm I think it doesn't quite get you that, because each client would still be convinced of its own assessment of the payload. E.g. if geth thinks a payload is valid but no one else does, all validators running geth will just continue undisturbed, while others will see the beacon block as canonical but not the payload

0 reply

0 recast

0 reaction

Francesco pfp

I think we can have this feature by itself, without the whole epbs machinery. I think we can even do it in a way which is compatible with future more radical redesigns that separate consensus and execution slots (like execution tickets or regular epbs itself)

2 replies

0 recast

6 reactions

Francesco pfp

How desirable is the separation of beacon blocks from execution payloads w/o the other guarantees that epbs has been associated with? (unconditional payment, builder protection w/o relay) How much value do we get just by ensuring that an execution failure still leaves a valid beacon block?

2 replies

0 recast

4 reactions

Barnabé Monnot pfp

Barnabé Monnot

Here is a new RIG Open Problem on the specification of a Multiplicity gadget in the context of Ethereum’s Gasper consensus. Good problem for a consensus researcher interested in censorship-resistance! https://efdn.notion.site/ROP-9-Multiplicity-gadgets-for-censorship-resistance-7def9d354f8a4ed5a0722f4eb04ca73b

3 replies

10 recasts

19 reactions

Francesco pfp

Research team AMA happening tomorrow, still time to leave any questions you might have :) https://www.reddit.com/r/ethereum/comments/191kke6/ama_we_are_ef_research_pt_11_10_january_2024/

0 reply

2 recasts

3 reactions

Francesco pfp

A classic already. Rarely informative panel

1 reply

0 recast

0 reaction

Francesco pfp

Definitely no decision nor timeline but MaxEB can be a first step to many of these approaches to SSF. It would be nice for it to be in the next fork, so some consolidation can start earlier rather than later

0 reply

1 recast

1 reaction