Zoom Logo

MEV Roast 13 - The Blockspace Bazaar: Past, Present & Future - Shared screen with speaker view
Alex Obadia
12:36
Wassup everyone
Alex Obadia
12:47
Impressed that you’re here before the Host
Alex Obadia
12:48
kudos
Alex Obadia
13:06
Oh no I’m alone lol
yahsin
18:25
:D
Georgios Konstantopoulos (Paradigm)
19:40
Big gang
thegostep
19:52
👋👋👋
yahsin
20:10
👋👋
phil
20:14
so excited to see everyone here. please feel free to unmute and/or participate in the chat I will try to keep it lively during the roast
Alex Obadia
21:34
Full agenda: https://github.com/flashbots/pm/issues/44 :)
thegostep
21:43
Please drop your questions in the chat and we will call on you to unmute and ask your question :)
Brock Elmore
27:13
please mute him?
Georgios Konstantopoulos (Paradigm)
27:15
Did we get zoombobmed
Brock Elmore
27:18
lol
Alexander Lindgren
27:26
lol
James
27:26
Mute everyone except the speakers
Harith Kamarul
27:28
😂
Alex Obadia
27:37
Ahhh
Alex Obadia
27:40
Ok gimme a sec
Alex Obadia
28:03
Not porn is a big win
Alex Obadia
28:13
Yep indeed
nathan
29:41
Can Phil comment on if he forsees a rise in cross-chain MEV?
phil
29:54
oh i didnt see that nathan sorry
Raul Jordan
30:10
One thing to chat about is regarding reorg incentives in eth2 due to MEV
Lamboshi aka Nolan
30:11
don't forget to disable annotations for participants.
phil
30:16
I see cross chain as being a generalization of cross l2 / cross shard / etc, in some ways they just share a base layer
nathan
30:29
Thanks
Georgios Konstantopoulos (Paradigm)
30:34
nope
Enchanted Broccoli
30:45
nope
phil
30:46
also alex let us turn on cams :) <3
Samuel Shadrach
30:49
no
Enchanted Broccoli
31:24
Enchanted Broccoli
Enchanted Broccoli
31:26
is Leo
Trent Van Epps
31:45
if you make others cohosts they can potentially help with moderating (depends on the settings)
Surya Bakshi
31:47
careful with th video
phil
31:55
yes tina make me co host plz
thegostep
32:13
Tinaaaaaa
x
32:33
1
Raul Jordan
32:34
yes
phil
32:44
i also cant unmute myself lol
Georgios Konstantopoulos (Paradigm)
32:44
Can u also unmute me :d
phil
32:45
tinaaaaa
Trent Van Epps
32:45
when you are sharing screen it's hard to see chat =) so tina might not see this
phil
33:16
ill message her out of band
Trent Van Epps
33:21
I pinged in discord
Georgios Konstantopoulos (Paradigm)
33:45
Any q’s that go to chat I can also answer
Georgios Konstantopoulos (Paradigm)
33:47
So feel free to write
Georgios Konstantopoulos (Paradigm)
33:51
Anything u wanna ask
yahsin
40:07
Alvin E. Roth game theory, experimental economics and market design, and shared the 2012 Nobel Memorial Prize in Economics - https://web.stanford.edu/~alroth/
Trent Van Epps
41:01
what's the source for 80-90 GPUs?
Trent Van Epps
41:17
just asking because it is often a contested #
Trent Van Epps
41:43
sorry, didn't mean to interrupt!
yahsin
44:35
https://www.aniccaresearch.tech/blog/ethereum-blockspace-who-gets-what-and-why
yahsin
45:23
Slides: https://docs.google.com/presentation/d/12v2SRpoyUQuiVwtQfpkNbneW9fAlzo0yl-msTVBPGCs/edit?usp=sharing
nathan
50:17
Long term: what about building new financial primitives that exclude miners from being able to take MEV?
Trent Van Epps
50:50
i believe that would require separating ordering from execution
Trent Van Epps
51:05
or ZK defi primitives
Dev Ojha
51:08
Another alternate for long term is to SNARK the state transition, which makes the block time linear in amount of state accessed, but not computation
Rick Dudley
51:13
I wouldn't phrase it that way.
Rick Dudley
51:24
sorry, that was to trent, I agree with dev more.
Rick Dudley
51:30
but we could just have a sane composability model.
Rick Dudley
51:37
we don't need snarks for that.
Rick Dudley
51:50
but the change is so big, might as well throw some snarks in there too.
Trent Van Epps
52:06
you get a snark, everyone gets a snark
Harith Kamarul
52:20
Would a so-called ‘leaderless’ chain be a long term solution as well?
dave
52:36
there is also work on fair order consensus that can help curtail MEV
dave
52:39
e.g. https://eprint.iacr.org/2020/269
Rick Dudley
53:38
i'm guessing that's the same"separating ordering from execution" but I don't think we want that sort of "double consensus" but maybe we do.
Dev Ojha
53:56
I believe there’s honest threshold assumptions in the fair-order consensus algo’s, and adding fraud provability into that comes at a huge bandwidth cost
phil
54:01
@Rick and Dave I want to debate this issue in the Q/A part and fair ordering in general
phil
54:08
I also want to do a separate roast just on tihs eventually
nathan
54:14
I know it feels undesirable, but there could be horrible consequences if we arrive at a situation where MEV rewards to miners exceed block rewards
nathan
54:39
Miners no longer have an incentive to secure the network as intended, but rather to act against other users
Samuel Shadrach
54:49
MEV already exceeds block rewards on blocks that have hacks
Samuel Shadrach
54:56
DeFi project hacks*
nathan
55:00
Good point
Alex Manuskin
55:31
In the long term, ETH tx price should go down.
Rick Dudley
55:41
I think we should just upstream MEV-geth and help miners adapt, this is seems like the easiest lift engineering wise.
Alex Manuskin
56:46
But if Tx fees out-value block rewards, it might be a problem for chains where reorgs are a problem
dave
56:57
@Dev there is a honest majority assumption. Not sure about fraud proofs. I believe there is also more recent followup work that has similar honesty assumptions to Nakamoto consensus. https://eprint.iacr.org/2021/139
phil
57:43
@Dave the security also relies on honest majority in some way though. also it doesn't handle network attackers
phil
57:57
but I do love that work and am very excited abou tit!!
Surya Bakshi
58:37
woo order fairness
Surya Bakshi
58:47
when's the ico
dave
58:51
@phil yes but I mean Nakamoto consensus also has honest majority assumptions.
Tomasz Stańczak
58:53
What is MPC-EVM?
Surya Bakshi
59:09
executing EVM in an MPC setting without reveling participants inputs
Robert Miller
59:09
multi-party compute EVM, I am guessing
James Hsin-yu Chiang
59:16
I don’t quite follow why separating MEV mempool from non-MEV helps efficiency. It seems both can be translated to comparable bids for block space, no?
dave
59:25
not sure about network adversaries but you might be right especially nation-state level network adversaries
Tomasz Stańczak
59:33
thanks
phil
59:43
@dave some would argue nakamoto is secure in "majority rational" or "supermajority rational" even if not "majority honest"
phil
01:00:02
also there are different degrees of honesty -- eg reordering is considered "honest" in nakamoto security proof but isn't "honest" in the formal definition of "runs the protocol"
phil
01:02:15
georgios how do you see the miner behavior split you mentioned evolving over time
phil
01:02:26
(also i still cant unmute but we are fixing this so things should be smoother soon stay tuned)
Dev Ojha
01:02:28
I don’t think homomorphic execution is a desired property, since I think you can leak almost all the data for contracts you knew about ahead of time
Georgios Konstantopoulos (Paradigm)
01:02:48
Ah I can tlak aobut that
Georgios Konstantopoulos (Paradigm)
01:02:55
Can I unmute
phil
01:03:06
hahahahah someone unmute georgios
Georgios Konstantopoulos (Paradigm)
01:03:17
We’re in mute jail
phil
01:03:20
@tina we cant talk lol
Surya Bakshi
01:03:27
tinaaaaa
Gengmo
01:03:35
The unmute mempool haha
Surya Bakshi
01:03:50
pay tina to get in unmuted first
Samuel Shadrach
01:04:01
via a private channel
thegostep
01:04:03
I stand against this permissioned unmutepool
Surya Bakshi
01:04:04
5 memes is the price
Ansgar Dietrichs
01:04:17
zoom bombs as Attendee Extractable Value…
Georgios Konstantopoulos (Paradigm)
01:04:34
I can talk about all that just unmute meee
Surya Bakshi
01:04:46
tine colludes with zoom bombs to make unmuting more valuable and drive up prices
Mark Tyneway
01:04:52
@georgios about re-using the sequencer bond, there could be arbitrary “batch validity conditions” that people could prove fraud over, one example could be the sequencer promising not to extract a defined type of mev
Rick Dudley
01:04:58
I have some thoughts about the L2 stuff.
Trent Van Epps
01:05:03
connext state channels? or are you asking more philosophically
Tina Zhen
01:05:06
I am going to unmute everyone i recognize
Tina Zhen
01:05:16
or makes a good point in the side chat ;)
phil
01:05:17
if you remute yourself you cannot unmute again
phil
01:05:24
so recommend muting your mic in client side mixer
phil
01:05:31
we are fixing this and will have a new room ready (zoom sucks so we cant fix it on this room)
Surya Bakshi
01:05:34
oh
Surya Bakshi
01:05:36
woops
Surya Bakshi
01:05:47
i got unmuted and decided to mute myself till i needed it
Jun-You Liu
01:05:56
private rpc endpoint to tina is the only way
Daniel Goldman
01:06:46
General “ideological” q: *should* the goal of L2s (that have ordering power, which will be most/all of them) be to minimize MEV (with things like fair ordering consensus discussed above, etc)? Or to try to capture it & use it for stuff?
Rick Dudley
01:07:32
I think people will just yolo it.
Daniel Goldman
01:07:50
lol
Martin Köppelmann
01:07:51
Could someone unmute me?
thegostep
01:07:58
Let’s get Rick and Martin on stage
Georgios Konstantopoulos (Paradigm)
01:08:02
+1
Tina Zhen
01:08:07
just did
Daniel Goldman
01:08:54
Roast: Georgios’s surname is too long #roasted
James Hsin-yu Chiang
01:09:33
@Martin - that comes without atomic guarantees right? You trust the miner-relayer to execute sequence as promised?
Tina Zhen
01:09:33
Daniel are you inviting unmuting yourself to give the correct pronounciation of Georgios' last name?
Daniel Goldman
01:09:55
*leaves call*
Tina Zhen
01:10:14
Jun-you, you are absolutely spot on about the "private rpc endpoint" hehe
Trent Van Epps
01:11:13
boom roasted
Georgios Konstantopoulos (Paradigm)
01:11:25
That seems OK with me
Georgios Konstantopoulos (Paradigm)
01:11:26
@Rick
Georgios Konstantopoulos (Paradigm)
01:11:31
Yes
Georgios Konstantopoulos (Paradigm)
01:11:38
Opt-in though
Rick Dudley
01:12:37
yeah, it would be opt in.
nathan
01:12:41
While perhaps undesirable I would still prefer that to a single-player monopoly taking all MEV
Georgios Konstantopoulos (Paradigm)
01:13:09
Ideally the relayer should be open to anyone
nathan
01:13:18
@ the comment about MEV market emulating traditional markets in composition
Georgios Konstantopoulos (Paradigm)
01:13:21
And you could have trusted-relay bridges for bypassing the trust-minimized relay latency
nathan
01:14:01
I guess an ethical question is if miners deserve MEV at all?
Georgios Konstantopoulos (Paradigm)
01:14:02
But basically any latency sub fraud proof verification (or validity proof generation)
thegostep
01:14:05
Daniel had a q about ethic
Georgios Konstantopoulos (Paradigm)
01:14:10
Requires some market maker or a trusted signature
thegostep
01:14:29
Maybe unmute him so he can ask?
Samuel Shadrach
01:14:30
can we explicitly frame why we want to minimise MEV in first place? I would assume censorship (due to centralisation) and bad UX - are there ways to more directly solve these issues without a general "solve MEV" approach?
_
01:14:30
Can hide trx data using zk?
Dev Ojha
01:14:41
> Can hide trx data using zk?
Dev Ojha
01:14:45
Not really
_
01:14:46
So that its can read the trx data
nathan
01:14:48
After all, they don’t contribute to value creation in MEV at all
Dev Ojha
01:14:51
All changes to state must be public
Dev Ojha
01:15:07
and you can often infer the tx data from the state change
Tina Zhen
01:15:44
Unmute request sent to Daniel...
nathan
01:16:03
Thanks
Martin Köppelmann
01:16:14
Could I comment?
Georgios Konstantopoulos (Paradigm)
01:16:54
Batch auctions !
Daniel Goldman
01:18:05
Martin’s “meta question” is what I was most getting at
Trent Van Epps
01:18:19
i want to hear Daniel talk
Daniel Goldman
01:18:28
Applying to L1 is hard (politically), but on L2s we’re already making these decisions
rogervs
01:18:28
Possibly a noob question (please ignore if it doesn't make sense), could MeV be used to prevent a side chain from syncing with L1 if there was enough financial incentive?
phil
01:18:34
[ <3 batch auctions and fair sequencing, definitely check these things out ]
Surya Bakshi
01:18:46
only once you move to PoS would it be possible at all. My intuition is miners wouldn't adopt changes that force the to order txs in a pre-defined way
Surya Bakshi
01:19:17
giving them the same reward but outsourcing would be better, i think even with mempool privacy you can still extract some MEV by reveaing information and making deals wit miners
Trent Van Epps
01:19:40
someone is sharing screen.. but forgot
Tina Zhen
01:20:04
good observation trent
Martin Köppelmann
01:20:05
For sure
Surya Bakshi
01:20:06
ctrl+t some new memeshaha
Tina Zhen
01:20:19
I was testing it on my other devices
Trent Van Epps
01:20:25
I've done that myself @tina
Tina Zhen
01:20:36
definitely learning my zoom lesson in prod
Georgios Konstantopoulos (Paradigm)
01:20:40
The high level observation is making tx ordering a consensus rule while now it is not
Surya Bakshi
01:21:34
unmute ples
Samuel Shadrach
01:23:20
users care about MEV only due to UX. would this lead to users their txs to a private mempool that makes certain UX guarantees in its MEV extraction? even if this whole relationships is "trusted" - no way to enforce besides real world reputation
Samuel Shadrach
01:23:31
users sending their tx*
Samuel Shadrach
01:23:56
and this private actor does bundling
SiegeRhino
01:32:53
Is this a public chat ?
Trent Van Epps
01:33:03
yes
Tina Zhen
01:33:05
(me test in prod again pls ignore)
Alex Obadia
01:33:09
Yes Siege :)
Trent Van Epps
01:33:10
lol
phil
01:33:19
that can only end well
SiegeRhino
01:33:27
It's not as active as a twitch stream hahaha
Samuel Shadrach
01:33:37
omegalulz
Tina Zhen
01:35:16
we should really explore twitch or discord haha (push to talk) thanks for that suggestion. in the past, we have constantly have a great small audience despite being a public conference. this is the first time we switched to zoom.
Hasu
01:36:01
yeah keep the zoom small, but stream it to YouTube/twitch
Lanre Jonathan
01:36:06
Zoom’s Webinar capabilities may help w/ this also + will prevent Zoom bombing
SiegeRhino
01:36:11
You made me install zoom :D
Trent Van Epps
01:36:27
was also going to recommend webinars
Mo Dong
01:36:34
Colluding bidders in second price auction/generalized VCG also have incentives to collude through bidding cartels. There are some solution to mitigate this through probabilistic winner selection.
Robert Miller
01:36:49
webinars + slido + raising hands could work
phil
01:37:03
I think zoom+cohost is really good too we just can't enable it
Alex Obadia
01:37:57
It used to be a default setting but in new Zoom is buried in advanced account settings
Trent Van Epps
01:38:24
tru
Tina Zhen
01:38:28
Alex can you make me host again, I cannot unmute people again.
SiegeRhino
01:38:43
i'm a mevgeth and bundle user but omg i don't understand a thing of what's being talked about
Shayan Eskandari
01:38:58
you can have co-hosts too, don’t need to hot potato the host permission :)
Tina Zhen
01:39:08
you will soon, in the next talk
Tina Zhen
01:39:23
@SiegeRhino
butta
01:39:44
how do i get my poap
Surya Bakshi
01:39:45
hopefully it made some sense in framing the proble
Gengmo
01:39:48
I believe “credible mechanism design” is a good keyword to search in the literature
Surya Bakshi
01:39:54
yes!
Surya Bakshi
01:40:05
credible auction is what i referred to in the slide
Surya Bakshi
01:40:07
can you unmute again
Vitalik Buterin
01:40:17
http://web.stanford.edu/~mohamwad/CredibleMechanisms.pdf
Samuel Shadrach
01:40:57
we're assuming bidders to be hyper-rational traders. in practice they just care about certain guarantees on UX (user experience). would they prefer a private channel that provides these guarantees in its MEV extraction
Surya Bakshi
01:42:18
shit i muted myself again
Surya Bakshi
01:42:20
i'll stpo doing that
Surya Bakshi
01:42:32
i'll hardware mute my mic next time
Robert Miller
01:42:35
yes we can
Surya Bakshi
01:42:37
instead of through zoom
nathan
01:42:37
Yes
Surya Bakshi
01:42:42
yeuhh
Surya Bakshi
01:43:19
unmute again Alex? sorr
Surya Bakshi
01:43:24
pef
Surya Bakshi
01:43:30
pperf*
Tina Zhen
01:43:45
i will actively unmute everyone who have something to say
Robert Miller
01:44:16
can you unmute
Robert Miller
01:44:18
and give me screen share
Feihu Tang
01:44:33
Tina you need an unmutebot.
Alex Obadia
01:44:37
Steph are you raising your hands
nathan
01:44:46
Would it be helpful to expose more variables in the executing state of the EVM?
thegostep
01:44:55
To let me out of the lock cage
Surya Bakshi
01:45:06
i think unless we can make protcol level changes to the mempool the negative effects of MEV extraction even on users who aren't front run or who aren't trying to extract MEV are sufferig from the costs of higer agas
nathan
01:45:15
For example, if transactions can know the tx position they are in, they could fight against sandwiching
Alex Obadia
01:45:17
Robert can share his screen if you stop sharing Tina
Georgios Konstantopoulos (Paradigm)
01:45:26
Tina you should just stop share screening
Alex Obadia
01:45:28
^
Georgios Konstantopoulos (Paradigm)
01:45:29
And he’ll be able to
Alex Obadia
01:45:40
Yes we can
Alex Obadia
01:45:59
Tina
Alex Obadia
01:46:02
You’re the host!
Surya Bakshi
01:46:10
so private channels are preferred to at least to stop users from facing bad UX
奇迹哥
01:46:20
Yes we can
unvetica.eth
01:46:27
when POAP?
Surya Bakshi
01:46:45
woo sgx!
thegostep
01:46:52
Freedoooom
butta
01:46:57
WHEN POAP
Alex Obadia
01:46:57
Hes alive
Tina Zhen
01:47:21
thegostep is unmuted
Georgios Konstantopoulos (Paradigm)
01:48:00
It’s stephane i think
Surya Bakshi
01:48:03
should i know what POAP is?
Alex Obadia
01:48:09
No Surya you gud
thegostep
01:48:10
Noo not me
Surya Bakshi
01:48:13
haha okay
Alex Obadia
01:48:13
It’s Tina maybe
Alex Obadia
01:48:15
Or Surya
Alex Obadia
01:48:22
I heard the same noise during Surya’s pres
Tina Zhen
01:48:23
nooo not me either
thegostep
01:48:29
Whoever was eating cerials
Tina Zhen
01:48:32
the ghost host
Surya Bakshi
01:48:39
carbs bad for zoom
Dev Ojha
01:48:39
POAP=Proof of attendance patch
Alex Obadia
01:48:44
Revolving host and cereals
Sarah Allen
01:48:46
@Surya yes you should since you got one from IC3 camp
Surya Bakshi
01:48:48
> POAP=Proof of attendance patch+1
Surya Bakshi
01:49:22
> @Surya yes you should since you got one from IC3 campwe should get patches I can sew onto a bomber jacket
Surya Bakshi
01:49:27
my crypto bomber jacket
SiegeRhino
01:49:38
MEV SGX is v2 of mevgeth ?
Tina Zhen
01:50:12
Flashbots is currently in Alpha, MEV SGX is potential V1.
Tina Zhen
01:50:34
We are sharing the MEV-SGX design here to be Roasted
Surya Bakshi
01:51:11
SGX increases CPU temps, roasted
thegostep
01:51:25
😂
Tina Zhen
01:51:51
nice pun ;P
Tina Zhen
01:53:24
there needs to be POAP for ppl who succeeded in getting unjailed from Zoom
Surya Bakshi
01:53:36
is SGX a good way to have intermediaries who can collect bundles andeventually submit whole block sequences to the miner? we want combinatorial work thatthe miner probably won't do so intermediaries could do that for them and bid to the miner
thegostep
01:54:23
Yes @surya
Surya Bakshi
01:54:28
dope
Surya Bakshi
01:54:40
without it intermediaries just extract rent
Alex Obadia
01:54:46
Ahhh who’s making this noise
Surya Bakshi
01:55:01
the participant slist says it's coming from Robert's mic
thegostep
01:55:02
But also provide a credibly neutral way for transactions to be included on chain without being manipulated
Dan Elitzer
01:55:05
What type of problems could occur if someone did a side channel attack on SGX?
David Felsen
01:55:14
is it the noise of a miner unwrapping a bundle?
thegostep
01:55:19
Just mine everyone and unmute for q&a
thegostep
01:55:22
Mute
Surya Bakshi
01:55:33
it's his computer being roasted by sgx at 100% capacity
_
01:55:42
Who the miner will know the bundle have any value over other truss?
_
01:55:54
how**?
Surya Bakshi
01:56:13
so SGX just ensures the miner handles all bundles in the "correct" way
Surya Bakshi
01:56:28
so the SGX code will decide he most profitablebundle and the miner can't do anything to stop it
Surya Bakshi
01:56:43
miner can't be malicious and deviate from the auction
奇迹哥
01:56:43
😂
Samuel Shadrach
01:56:51
Besides the manufacturer, how can anyone know or verify what components are inside a given piece of hardware? Can Intel backdoor the network?
Surya Bakshi
01:57:25
SGX generates an attestation proof that they ran the correct code, but yes intel is trustd here
Samuel Shadrach
01:57:39
Intel can run a POA chain then?
_
01:57:53
Most of the bundle only include 1 Gwen trx then ?
santi
01:58:11
Isn’t The ‘speed limit’ naive in the sense that someone could send 60 bundles per address?
thegostep
01:58:28
The elusive cohost ability has been unlocked
Surya Bakshi
01:59:19
+1
Surya Bakshi
01:59:20
woo
Alex Obadia
01:59:37
Let me just say these animations are top notch
Surya Bakshi
01:59:49
yes! i love these slides
Surya Bakshi
01:59:55
good job Robert
Jonathan Passerat-Palmbach
01:59:57
Robert is clearly elligible for Chief Powerpoint Officer
SiegeRhino
02:00:11
But then u have to trust the relay instead
Alex Obadia
02:00:13
Jonathan speaks the truth
Surya Bakshi
02:00:15
turning tricks in google slides
Surya Bakshi
02:00:36
chief animator
David Felsen
02:00:46
is there a risk that these additional steps add latency ie increase the risk that a block is mined by someone else?
Alex Obadia
02:01:12
I am watching the chat and will unmute peeps for questions
thegostep
02:01:17
@david yes
Surya Bakshi
02:01:32
These bundles are submitted whenver and the miner just chooses the best one when it's forming theblock template so there's some mitigtion to your concern
thegostep
02:01:46
Remains to be seen how much latency will impact the system
Surya Bakshi
02:02:01
in that it doesn't have to run a live auction wen it's forming the block it already has a pool of bundles/bids to choose from
Surya Bakshi
02:02:17
so it's latecy affecting a bidder's ability to get win the auction if they are late in submitting
thegostep
02:03:13
That’s correct @surya
Harith Kamarul
02:03:36
Since the tx are hidden from miner, there could be instances where searcher B’s bundle fails because their target tx was mined earlier in the same block by searcher A’s bundle. Is this correct?
Rick Dudley
02:04:07
I'm pretty sure we can make partial blocks and compose them safely if we provide witnesses with the blocks.
Rick Dudley
02:04:17
with the partial blocks.
Rick Dudley
02:04:22
so we don't need a lot of this stuff.
thegostep
02:04:32
@harith this design looks at complete blocks to avoid that problem, but it would be great to use partial blocks
Dev Ojha
02:04:52
> I'm pretty sure we can make partial blocks and compose them safely if we provide witnesses with the blocks.
Dev Ojha
02:05:03
We still want to hide the txs to avoid sandwiches?
Dev Ojha
02:05:44
You could do a randomised ordering for how you combine the partial blocks together
Rick Dudley
02:05:52
yeah, just use the tx hash.
Robert Miller
02:05:53
don't think the noise was coming from my end btw! I could hear it too
Dev Ojha
02:06:11
But you can make front-run attempts with very asymmetric risk profiles for whether or not you actual front-ran the target
Georgios Konstantopoulos (Paradigm)
02:06:44
My understanding is that we need to do encrypted lookups on the gasPrice
Dev Ojha
02:06:48
so I think in a threshold encryption setting you need to compose partial blocks + tx hiding personally
Dev Ojha
02:06:53
> My understanding is that we need to do encrypted lookups on the gasPrice
Dev Ojha
02:06:55
No
Dev Ojha
02:07:00
You can just do AEAD-style encryptions
Georgios Konstantopoulos (Paradigm)
02:07:04
Can u expand?
Dev Ojha
02:07:16
Yeah, I can explain it for threshold decryption and for timelock
Dev Ojha
02:07:55
So I can threshold encrypt a message M, with an associated public data AD, to get a cipher text C, with the following property.
Dev Ojha
02:08:27
Given (C, AD), any party can check that C is a valid encryption of some message, where the message author also attests to AD
Dev Ojha
02:08:43
So the gas price can just be public in that AD
Dev Ojha
02:09:27
In VDF-style Timelock Crypto constructions, you can make you’re permutation depend on the additional data
Dev Ojha
02:09:42
(e.g. imagine in Poseidon / Rescue, you make you’re ARKs derived from H(additional data)
Tamlyn Rudolph
02:09:43
Thanks for this talk. So, at Vega (DPOS blockchain for trading bridged to Ethereum) we’ve been developing a pre-protocol widget to enforce fairness of ordering at the consensus layer. Would love to hear what people think about it. We have done some modelling on latency and have had thoughts around customised “fairness”. It should be applicable for other POS chains. https://eprint.iacr.org/2020/885
Tina Zhen
02:10:42
https://github.com/flashbots/pm#mev-roast-recordingsOn threshold decryption, timelock crypto as potential alternatives, check out the 1st talk by Sunny Aggarwal on Jan28 MEV Roast: Privacy Solutions for MEV Minimization
phil
02:11:24
im muted again
phil
02:11:25
:(
phil
02:11:26
sorr
phil
02:11:27
sorry lol
Charlie Noyes
02:11:31
@Vitalik what phrase did you just use that is phonetically “day quaseh”?
Tamlyn Rudolph
02:12:32
Admin - could someone let Klaus Kursawe in to this chat?
Alex Obadia
02:13:10
He’s been let in :)
Klaus Kursawe
02:14:33
Thanks
SiegeRhino
02:15:27
Why can't we just allow searchers to chose to not send bundle to a specific miner if they notice that they don't behave ?
thegostep
02:16:29
It might be difficult to identify
Georgios Konstantopoulos (Paradigm)
02:16:31
Gotta hop! Thank you all. Great mtg.
nathan
02:16:42
Thanks George. Great talk!
Tina Zhen
02:16:55
i cannot unmute myself
thegostep
02:17:14
Get a taste of your own medicine
Feihu Tang
02:17:23
😂
SiegeRhino
02:17:37
How do you intend to fix the fact that highest bidders are ordered by implied gas price which does NOT reflect the bundle that end up with the best miner revenue ?
Daniel Goldman
02:17:41
Thanks all, very interesting & useful!
thegostep
02:17:49
I want to talk about account abstraction!
thegostep
02:17:59
Don’t leave
Alex Obadia
02:18:27
https://github.com/flashbots/pm/issues/44 agenda
lightclient
02:18:35
Can I be unmuted to talk 3074 later
john
02:18:58
i don't hear anything
Alex Obadia
02:19:08
John it’s on your side
Surya Bakshi
02:19:12
that and othe intra-block things that can be done like flash loans over bundles @thegostep tweeted about it too
Alex Obadia
02:19:21
Light client - you been unmuted
Surya Bakshi
02:19:22
or he tweeted it period, it's his thing
lightclient
02:19:38
ty
Alex Obadia
02:19:48
Lol helicopter
Alex Obadia
02:19:50
Phil on the run
Surya Bakshi
02:19:53
miners coming for you
Surya Bakshi
02:20:09
miners hate him, learn this one simple trick to beat MEV
Robert Miller
02:20:15
gotta jump - thanks everyone for tuning in today :)
Alex Obadia
02:20:30
Thanks for your great presentation Robert!
Ansgar Dietrichs
02:20:40
would be happy to be unmuted
Alex Obadia
02:20:52
Ansgar - unmuted
SiegeRhino
02:20:55
multi bundle does not fix the fact that highest gas price bundle is not going to be the one that gives the most profit
Alex Obadia
02:20:57
Thegostep - unmuted
phil
02:21:16
oh no i remuted myself
SiegeRhino
02:21:16
if two bundles are concurrent one could give higher miner revenue with lower gas price
phil
02:21:23
alexxxxx
Alex Obadia
02:21:29
unmuted
phil
02:21:41
ty :)
unvetica.eth
02:22:00
Will the POAP be distributed now?
lightclient
02:22:06
Sorry didn’t realize I couldn’t unmute myself once the host unmuted me
Alex Obadia
02:22:21
I just re unmuted you
lightclient
02:22:22
Ansgar can talk mostly
Alex Obadia
02:22:34
Just ask here and I will unmute
lightclient
02:22:35
I’m driving so have background noise :(
Alex Obadia
02:22:40
ahh
lightclient
02:22:42
Okay thanks
Alex Obadia
02:22:45
How are you driving and typing
Alex Obadia
02:22:52
Be careful
Samuel Shadrach
02:23:05
DeFi project hacks = MEV is unbounded
Alex Obadia
02:23:43
Lol at Phil’s muting reflex
Alex Obadia
02:24:55
Thegostep I unmuted you
Alex Obadia
02:25:24
Birds chirpin
David Felsen
02:25:25
i like how we have actual *crickets* when no one is talking
Alex Obadia
02:25:48
Facts
Sam Wilson
02:27:03
Hey. I have a comment.
thegostep
02:27:12
Unmute Scott!
Alex Obadia
02:27:13
Hello Sam
Sam Wilson
02:27:18
Hey!
Alex Obadia
02:27:20
Scott unmute
Alex Obadia
02:27:40
Y’all unmuted
Alex Obadia
02:27:43
Hello Mariano
Sam Wilson
02:27:44
<3
Mariano Conti
02:28:14
Hi! Had to jump to another call, but back :) didn't want to miss account abstraction
Alex Obadia
02:28:27
Thank you for joining initially, and coming back <3
Alex Obadia
02:29:05
Scott - lmk when you wanna chat, I see you muting yourself back
Tina Zhen
02:29:11
im sorry its crickets chirping on my side today, much better than 3 months ago, i had island roosters chiming in on our roast
Scott Bigelow
02:29:18
Oh I thought I would remain in control over my mute!
Alex Obadia
02:29:23
No unfortunately
Alex Obadia
02:29:28
The mute is not permisionless
Tina Zhen
02:29:42
i wish we could go back to our permissionless roast...
Alex Obadia
02:29:46
Tina-amazing-background-noise-Zhen
thegostep
02:30:22
Get Scott on stage !
Alex Obadia
02:30:28
He is being shy
Scott Bigelow
02:30:45
I’m in a windy env, don’t wanna accept unmute until I have something to say :)
Alex Obadia
02:30:55
Fair enuf
Alex Obadia
02:31:07
Just lmk here I am watching the chat
Alex Obadia
02:31:22
That counts for everyone
Sam Wilson
02:31:42
Woops, muted myself
Alex Obadia
02:32:00
I am the mute master
Alex Obadia
02:32:10
Unmute master*
Alex Obadia
02:33:23
Welcome back Santiago
Alex Obadia
02:33:29
It’s nice to have you with us
Alex Obadia
02:33:38
Also nice 1 on that rescue using Flashbots
Alex Obadia
02:35:55
Softly re-opening unmuting and video
Alex Obadia
02:36:11
Let’s nice if we can have nice things on the internet
Alex Obadia
02:36:14
Let’s see*
Alex Obadia
02:36:36
Phil you can put your vid back on
Daniel Goldman
02:38:11
I didn’t leave when I said I would b/c I was pressured to stay for account abstraction talk (and glad I did) but now I really do need to — thanks again all!
Alex Obadia
02:38:32
Thank you for attending and for staying Daniel! :)
Surya Bakshi
02:42:34
yes!
Alex Obadia
02:42:45
Thank you all for attending and the great discussion <3 thanks for bearing with us with the video call issues
nathan
02:42:51
Thanks everyone. Super interesting discussion. Have a good morning/afternoon/night!
Alex Obadia
02:43:22
Great job by Surya, Robert, Tina, Georgios, Leo, Vitalik <3
James
02:43:36
Thank you so much everyone!
Rick Dudley
02:43:42
Bye everyone it was a pleasure.
Mariano Conti
02:43:44
Eye opening work, thank you!
Gengmo
02:43:45
Thanks!
yahsin
02:43:46
thanks everyone!!
tesa
02:43:46
Thanks everyone!
Tamlyn Rudolph
02:43:52
Thanks!
Surya Bakshi
02:43:55
vaccine party at my place!
Josh Groves
02:43:58
thank you!
Diane Dai
02:44:42
Thanks💕
奇迹哥
02:44:50
Thanks💕
Surya Bakshi
02:46:05
ohh i completely misunderstood the question
Surya Bakshi
02:46:09
i migh have misheard it
Tomasz Stańczak
02:48:02
I am on the call.
Klaus Kursawe
02:52:35
@Tina, can I bother you for a sec on a question on research scope ?
Alex Obadia
02:53:49
Light is cutting off for me
Alex Obadia
02:53:53
Not sure if I’m the only one
David Felsen
02:53:59
same
Alex Obadia
02:54:21
Very nice
Alex Obadia
02:54:26
We can hear you
Alex Obadia
02:54:30
Oh no
Josh Groves
02:55:08
thank you!
Alex Obadia
02:55:22
Thanks everyone :)
rogervs
02:55:37
Thanks everyone
Tomasz Stańczak
02:55:40
Sorry I was not able to jump in immediately ^_^!
lightclient
02:55:51
:)
Francois
02:56:10
Thanks everyone :)
Alex Obadia
02:56:25
https://discord.gg/7hvTycdNcK Discord link
Tina Zhen
02:57:35
https://github.com/flashbots/mev-research
dave
02:57:37
@klaus, if you haven’t seen this already, there is followup work on the cornell paper for permissionless PoW https://eprint.iacr.org/2021/139
Alex Obadia
02:59:41
I gotta run, thank you all for joining <3 <3
David Felsen
03:02:30
thanks Tina & everyone else
Mariano Conti
03:02:33
<3