Francesco
@fradamt
Followers
2,267
Following
981
Media
18
Statuses
1,023
Math student fallen in the Ethereum rabbit hole. Research at the EF, currently focused on consensus
Joined January 2021
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
#ขวัญฤทัยตอนจบ
• 574087 Tweets
オーロラ
• 477961 Tweets
Joost
• 327556 Tweets
jeonghan
• 127459 Tweets
Fulham
• 79631 Tweets
#SixTONESANN
• 76439 Tweets
Man City
• 74328 Tweets
#الهلال_الحزم
• 61542 Tweets
自分これ
• 61279 Tweets
HeavenlyVoice WithMrC
• 54936 Tweets
Gvardiol
• 51763 Tweets
Ali Koç
• 44051 Tweets
カクレンジャー
• 38465 Tweets
太陽フレアのせい
• 38155 Tweets
キンスパ
• 37413 Tweets
Dremo
• 36451 Tweets
マリノス
• 29498 Tweets
Burnley
• 27714 Tweets
SUPER READY FOR SUPERNOVA
• 25151 Tweets
SEE YOU NEXT MONTH JIN
• 24910 Tweets
Luton
• 19603 Tweets
Haley
• 18797 Tweets
Dursun Özbek
• 17255 Tweets
#AdayOlAzizYıldırım
• 16137 Tweets
やまとなでしこ
• 13543 Tweets
Check out this ethresearch post if you're curious about DA scaling after 4844!
Tldr: we can gradually scale to 32/64 blobs, 10x 4844, and eventually beyond, up to full Danksharding, mostly with tools we are already familiar with. More blobs are coming!
11
38
162
New paper with
@luca_zanolini
, trying to expand the family of dynamically available consensus protocols with more practical ones (especially for the Ethereum context), to get better candidates for the "available chain" of a future SSF construction
4
20
94
I recently published this post, , discussing something I have wanted to bring to the protocol for a long time. It’s very long and I was *politely encouraged* to make it more digestible, so this is for all of you that can't read two paragraphs in a row🧵
7
23
87
I find it surprising that people seem to be ok with a future where eth is replaced by LSTs (possibly one, possibly custodial). What good is decentralization of staking if there is no trustless asset?
Imho the staking economics debate should focus on this point first
13
11
77
Has anyone written an overview of current research on L2 sequencing mechanisms? Would love to know what people have explored
7
3
46
Blobs are great, L2s are surely getting cheaper and/or reaching more users, and it is something very much worth celebrating. But let’s also wait a second before drawing conclusions about fees while blobs are free 😄
0
4
39
💯. Every MEV discussion is derailed by this misunderstanding. Lack of user harm does not mean it’s not MEV, because it can still have a centralizing effect, consensus-destabilizing effect etc.. If it affects the delicate balance of incentives of a cryptoeconomic system, it’s MEV
2
7
38
Thanks for organizing this and for giving me a chance to talk about some of what I have been working on!
Feels a bit surreal to be on the other side of this, after learning about MEV through roasts just a few months ago
Hope today can be someone else's beginning in the MEV space
Wanna learn more about why Ethereum may Splurge on PBS in addressing MEV?
Join us tomorrow for our live discussion:
MEV Roast - PBS on Ethereum Roadmap
w/
@VitalikButerin
@drakefjustin
@fradamt
@phildaian
@thegostep
& Roast Master
@hasufl
Agenda:
10
46
181
2
7
37
As promised, here's an easier to digest ethresearch post about two recent papers with
@luca_zanolini
(, ): . At the end of it, we specify a simple SSF protocol which we think fits the requirements of Ethereum.
1
6
33
A paper about timing games came out earlier this year, and a ROP (RIG open problems) about it existed since already last year. Problems are talked about openly and researched, not hidden or obscured by a system that only 10 people understand.
Still, what can we do better?
4
6
30
If you want to burn MEV, how about doing it in the 1000s of ways which do not involve fossil fuels? Let's not use PoW as an auction mechanism?
1/
@OffchainLabs
is developing a PoW mechanism to disincentivize sybil attacks by MEV searchers on the Arbitrum sequencer relay. In this post, I describe why the proposed design is economically wasteful and suggest more efficient designs
6
50
283
1
0
32
@TusharJain_
The whole point of rollups is that they can stay secure and decentralized while achieving a scale that an independent L1 cannot. If a very successful rollup were to try to become an L1, it would be unable to do so without compromises, regardless of how valuable its token is
3
2
32
Explainer on the recent MEV exploit. It is NOT as simple as "the slashing amount is lower than the reward for attacking", as has been suggested. There was a bug in the relay validation logic, without which a rogue proposer that is willing to get slashed would still fail
Block 16964664: A user managed to drain five MEV bots by exploiting a bug in mev-boost-relay.
Here's the block:
Here's the user:
Here's the patch:
Here's the longer explanation:
79
509
2K
2
5
32
Wrote something about this and other potential alternative ways to harden PBS without full enshrinement
Imho you get most of the value of epbs by having a gossip topic for bids and signed headers, which gives observability to the process and allows nodes to locally determine with good confidence whether a relay is being malicious and they should disconnect from it
2
5
21
1
4
29
@lightclients
Its main advantage is log(n) load with n validators, so extremely scalable. In practice, this comes at the cost of only √n validators being needed to stop liveness, preventing n from being high anyway. Also other downsides like lack of accountable safety
5
3
28
This is not how it works.
Yes, there's more than "just a bug" to this attack, as it shows that even a low MEV block can be exploitable for millions by unbundling.
No, a proposer can't just do this if reward > penalty. Without the bug/missing relay logic, it requires much more
Regardless, even in the best/worst case, the initial slashing penalty is 1 ETH (1800 USD).
25,000,000 profit for 1,800 penalty? Sure, i'll take that trade.
The economic incentives are broken here and it was only working due to a gentlemen's agreement to not do bad.
14
13
237
5
2
26
Basically works like this: proposal is made, head votes are cast, validators receive those head votes and (attempt to) confirm the current proposal, and if so they can use it as a target of their FFG vote. Finally view-merge is used for synchronization with honest proposers
The implemented consensus protocol of Ethereum, Gasper, has an hybrid design: it combines a protocol that allows dynamic participation among validators, called LMD-GHOST, and a finality gadget, called Casper.
1
14
40
2
5
20
Imho you get most of the value of epbs by having a gossip topic for bids and signed headers, which gives observability to the process and allows nodes to locally determine with good confidence whether a relay is being malicious and they should disconnect from it
mev-boost is fundamentally flawed. Some builder(s) are producing invalid blocks and relayers are recklessly providing them to validators, leading to a missed/invalid block proposal.
There are no in-protocol consequences for this behavior, so let's talk about ePBS in electra.
9
11
91
2
5
21
@koeppelmann
@casparschwa
@gakonst
@VitalikButerin
This only happened because about 51-53% of validators had not updated to a release which uses proposer boost, while the rest did. You can try to do this by an attack now (maybe for a few more weeks?), but as soon as everyone is using boost you won't be able to
2
4
16
(Not so) Little Christmas gift from one of the new members of the research team! Finally there’s a single resource to get introduced to the consensus protocol of Ethereum, its past, its present and the many changes in between
122
135
366
1
2
16
@cyber_hokie
@dabit3
@hellzbellza
Right now, yes. After 4844 (or something like that), not even. The market for “data blobs” can be completely separate from that for computation (including all things which are priced in gas). An NFT drop clogging L1 would then have almost no impact on prices on L2
0
1
16
The explainer a lot of people on CT should read, possibly *before* tweeting
This is regarding the recent
@RollkitDev
release of "Sovereign rollups" on bitcoin. I am not going to get into the terminology, but want to address the security properties, and whether it borrows security from bitcoin or not.
16
84
336
2
2
15
@bramcohen
Or that has been the plan for a while and you just don't know what you are talking about?
People really feel the need to express opinions about anything
2
0
10
@IgMosqueira
@sreeramkannan
A larger overhaul should happen with the SSF upgrade, for which there already exist candidate protocols (roughly derived from Goldfish) with optimal security. Check out and/or the two papers referenced at the top of the post
2
3
13
@seunlanlege
That’s not what single slot finality means when used in Ethereum research. The so called available chain would still exist, but finality would be faster in good conditions
1
1
12
@krugermacro
Quite the opposite. The rollup-centric roadmap is precisely not trying to do everything by itself
0
0
12
@epolynya
@dj_d_sol
@armaniferrante
building widespread awareness and acceptance of the modular paradigm is definitely building something
0
0
12
This is in some sense a continuation of the work started with
@jneu_net
and
@ErtemTas
in Goldfish, expanding our understanding of LMD-GHOST-like dynamically available protocols, and proposing one which we think might be viable in practice as the DA component of an SSF protocol
A major requirement for the consensus protocol of
#Ethereum
is that it should be dynamically available, i.e., in a context of dynamic participation among validators, safety and liveness of such protocol should be ensured.
6
9
35
2
2
12
@_danielmarzec
@MaxResnick1
I wouldn't worry too much about formal analysis in the consensus model, I can do that part 😀 What's helpful is something in between the vague explanation of a paper and the exact explanation of a spec: a hackmd that describes a specific way to do this in the context of Ethereum
2
0
12
@AlgodTrading
Crazy to think that a bunch of millionaires would try to scam people, in the middle of a bull market where just they are already multiplying their wealth every month...
0
0
11
@ralexstokes
make 50 anon accounts and flood everyone's TL with threads about you made 500k per month for the last 3 months through obscure data availability sampling opportunities (?)
They'll figure it out
1
0
11
@zkcat_eth
Some speccing/prototyping is ongoing, and clients will be more involved after 4844. Timelines are hard, but what I can say is that there are no research breakthroughs needed or anything.
1
0
11
It doesn't seem like there's any good out-of-protocol CR solution (am I missing something?). Then, should we just do inclusion lists asap? All we need is a fork-choice change (the (block, slot) fork-choice, enforcing the ILs) and a p2p message. No hard-fork needed
3
0
11
@scott_lew_is
This is quite the misrepresentation tbh. Ask Uniswap, and in particular the people which have been advocating for this EIP, if they feel like they are getting any preferential treatment. They’ll likely tell you the exact opposite.
1
0
10
@ittaia
@lightclients
Not that I know. The Avalanche whitepaper only claims O(log n) rounds termination when f <= O(sqrt(n)) though, and for good reason: convergence when views are 50/50 split relies on random fluctuations moving away from the equilibrium, but an O(sqrt(n)) attacker can avoid that
1
0
10
@SimonORN
@orion_protocol
Solana, because of its role in the Alameda ecosystem, and because it is specifically built for high speed trading. I think Serum has a chance to become a fairly relevant trading platform in the future because of that
0
0
10
A rogue relay would still be able to cause a handful of missed slots before everyone classifies it as malicious and disconnects, but I don’t see how preventing that justifies the complexity of epbs and doubling the slot time. What do we get for all that?
2
1
10
@koeppelmann
@casparschwa
@gakonst
@VitalikButerin
And it won't happen by accident as soon as > 50% of validators are using boost, which is probably imminent.
1
0
9
Would need to see a quantification of the social cost. Napkin math suggests we have better things to worry about imho (there just aren't that many solo stakers). Meanwhile censorship resistance falls mostly on them.
Over the last month a solo staker has made 49k usd in eth price appreciation on their 32 ETH staked. But god forbid we penalize them 1000$ for the social cost that staking on a toaster imposes on the network.
10
1
23
0
0
10
Length aside, it's actually a quite accessible presentation of a topic that should be understood and discussed well beyond the research realm. Anyone in the Ethereum community that cares about staking economics is encouraged to get informed and take part in the conversation!
1. In my last thread, minimum viable issuance was introduced as an important guideline for staking economics. This thread will take a closer look at how issuance level affects Ethereum’s equilibrium staking conditions and guide us toward a utility-maximizing reward curve.
6
18
70
0
1
8
I am a simple man, I see “reduce transactions costs on rollups”, I retweet and I go read/watch what’s new
@ethereum
@parithosh_j
@vdWijden
@lightclients
@peter_szilagyi
@alexberegszaszi
@EthMagicians
@greg_colvin
@nethermindeth
After
@lightclients
,
@VitalikButerin
had two proposals he shared about how we could reduce transaction costs on rollups. Both introduce a new transaction type which contains a blob of data, a bit like a "mini chunk of shard data".
1
4
20
0
0
9
@sassal0x
@0xPolygon
Paying for checkpoints is much cheaper than what rollups would be paying though, right? I don’t know what exactly goes in checkpoints, but I am assuming it’s a merkle root and maybe something else, but nothing like batches of transaction data
Regardless, nice thread!
3
0
9
@jon_charb
@nero_eth
@koeppelmann
@weboftrees
@ballsyalchemist
Even if they did all follow some randomized strategy, there's so many validators that it would be very easy to figure out exactly how late you can get away with publishing, with high probability
3
0
8
How does distributed building deal with merging bundles without access lists or other information which leaks what state you're touching?
2
0
9
@SamuelShadrach4
I’d recommend listening to this if you are interested in the security implications of different design choices, and what does it mean for a chain to be secure. They go off topic a lot but the good thing about that is it ends up exploring lots of aspects, especially for ETH2
The perennial full-node debate gets a refresher!
@aeyakovenko
(Solana) vs
@dankrad
(ETH2)
Motion: Security is about maximizing the minimum set of colluding miners.
(as in, not so much about long-tail of independent watchers)
Recording next week. Get your questions in!
7
9
74
0
3
8
@Fiskantes
If you build a “sovereign rollup” on Celestia (essentially an L1, including the ability to hard fork), it won’t have trustless bridges to other such Celestia-based rollups.
An Ethereum-based rollup won’t be sovereign (the settlement is governed by Ethereum) have trustless bridges
3
0
8
@safetyth1rd
It was never meant to fix gas prices. The only way to fix gas prices is increasing the gas limit, or reducing demand by moving it on rollups/other chains
0
0
7
@ChainLinkGod
Imagine you could protect users from frontrunning and sandwiching and also auction off extraction of "ethically neutral" permissionless MEV of the kind that will either way extracted by actors with the most resources.
3
0
8
where do I go to figure out what's the deal with the Cosmos-ecosystem infighting?
3
0
8
@koeppelmann
@peter_szilagyi
L2s leveraging L1 blockspace is only a temporary state of things. In a few months, if we want to increase L2s’ throughput we should increase the blob count, not the L1 gas limit. And as Ansgar pointed out, the two are actually competitive to some degree
2
1
8
🤡
Why would Putin kill Navalny? He already had him under lock and key. Navalny had little support in Russia. No threat whatsoever.
Of course the same was true about Gonzalo Lira, and the Zelensky government murdered him anyway. So who knows.
2K
1K
8K
0
0
9
@ittaia
@lightclients
The reason is that the deviation from the equilibrium is given by the variance of a Binomial(n, 0.5) distribution. If you have enough nodes to control a few standard deviations, you can prevent convergence for very long. Very similar to the balancing attack on Gasper
1
0
7
That said, we can, and imho should, do better than relying on the relayed block to win a network race (though starting with some advantage, given that the relay can release it first) in order to guarantee theft-protection against equivocating validators
1
0
7
Being able to efficiently process votes from the entire validator set unlocks so much! Feedback much appreciated both from researchers and client teams. Let's figure out if this is the best way to go about it or what we could do instead, and let's bring SSF to Ethereum soon™
1
2
7
@semenov_roman_
It is exactly another L1, but one which does not need its own validator set or even token, because it does not need consensus. It just needs a state transition rule. The difference between this and an Ethereum rollup is that Bitcoin itself can’t verify state transitions.
1
0
7
@toghrulmaharram
@0xMert_
@fede_intern
No offense to anyone, but Gasper and Gasper derivatives are way better understood than Solana's consensus. I have tried to understand how it works through the docs and it's really hard, someone should really give grants for formalizing the protocol and making it understandable
4
0
7
The initial goal for me was to figure out if we can get a simple, practical protocol which *does not need view-merge*, as RLMD-GHOST/Goldfish do, so that we don't have to have the proposer send a message summarizing its view. Basically trying to remove view-merge as an obstacle
2
0
7
@jon_charb
The insistence on focusing on confirmation rules is one of the few instances of consensus people being pedantic for a good reason
1
2
6
Is there any other chain that has concrete PBS plans? I’d be particularly interested in chains with fast finality but no guarantees for proposers if finalization fails (like in Tendermint), because it doesn’t seem so straightforward how to go about it there
0
0
5
@kevinsekniqi
@bertcmiller
Your MEV thesis was that Avalanche has none because it's leaderless, then you went leader-based and said it "kills pretty much all forms of MEV", which it clearly doesn't (and no, it's not just because annoying MEV people are being pedantic). But you want to lecture people on it?
0
0
6
@Piotr37889350
@CryptoPlunderer
@haydenzadams
@Uniswap
Uniswap can’t do anything about the fees except trying to accelerate L2 solutions, which they are already doing
1
0
5
@CollinCusce
@cryptoMaxi420
@vitmantug
@avaxdao
Data shards are fungible. A rollup can use post data to all of them without breaking composability. So no, they are not like subnets or sidechains
1
0
6
@Keremcrypto
@Uniswap
@PancakeSwap
The garbage blog post was actually incredibly informative and explained a complex and interesting product in a very clear way, but you're obviously not interested in anything of substance
0
0
6
@0xdoug
@uriklarman
Is it so unrealistic to think that staking pools should run multiple ELs and avoid attesting if they disagree on validity?
3
0
6
@weboftrees
@casparschwa
@jneu_net
@barnabemonnot
@ethconomics
@EFDevconnect
@skylar_eth
@TimBeiko
I just sat where Caspar put me so I could have a t-shirt
0
0
6
@dystopiabreaker
@sassal0x
@PhABCD
Fast finality is not the same as safety under asynchrony. The finalized prefix of the beacon chain is safe under asynchrony. The available suffix isn’t
2
0
6
@cklee_
@epolynya
@SwagtimusP
It's not unfortunately. The proposer still needs to sign a header, and if they are attacked via ddos the slot will be missed because no header is published
1
0
6
@nic__carter
The basefee could go to miners or eth holders, from transactors' perspective it is irrelevant. What you are saying makes no sense
1
0
6
@vdWijden
@wazadak
@peter_szilagyi
Get on twitch and stream yourself doing stuff, for educational value and fun
1
0
6
@wsfoxley
@VitalikButerin
@drakefjustin
@phildaian
@ObadiaAlex
*hereinafter just called MEV to purposefully stir controversy"
0
0
6
Thanks also to
@sproulM_
for reviewing the post and giving very useful feedback on it, and for raising the right practical questions a year ago, and to
@casparschwa
and
@CarlBeek
for working on the original idea back then and helping with this one now as well.
0
0
6
@fede_intern
Verkle allows for stateless EL clients. It doesn’t do anything to make it easier to follow the Beacon Chain (except for the EL part inside it)
1
0
5
@FranklinLynam
@FreyaHolmer
Writing this stuff in pseudo code is unnecessarily complicated... And this is high-school math?
0
0
4
@MaxResnick1
What's a clearly superior mechanism which can actually implemented in Ethereum, either off protocol like MEV-boost or in protocol with reasonably small changes?
0
0
5
Correction: "a gossip topic for signed headers" only, not bids. This is all you need to observe: if a signed header is gossiped on time and the slot is missed, it gives you a good signal about whose fault it is. As in epbs, but you use this signal locally and not for voting
1
0
5
Imho, we need more people thinking about this stuff in L2 teams. Looking forward to more protocol research happening there!
As a Scroll protocol researcher, you'll have the chance to explore a wide range of research topics, from rollup economics to state growth, EVM, MEV, PBS, AA, Layer 3, etc.
Your work will shape the future of a more secure and scalable infrastructure for the blockchain ecosystem.
3
1
21
0
0
5