Dankrad Feist
@dankrad
Followers
27K
Following
3K
Media
62
Statuses
5K
Some points on this:. 1. The current process of upgrading Ethereum IS BROKEN. Builders need long-term commitments to capacity, otherwise they cannot build on Ethereum. They need to know what the plan is in 5 years. So I hope that we can finally agree to timelines in the core dev.
So finally we are talking about how the current plan is just way too unambitious. If Ethereum wants to win, we need to be ambitious.
147
304
2K
I have decided to resign from my Eigenlayer advisorship. While I believe that the role was negociated in good faith and with the aim of making sure that Eigenlayer is well aligned with Ethereum, I understand that the perception of this relationship has been different and that for.
92
98
1K
So finally we are talking about how the current plan is just way too unambitious. If Ethereum wants to win, we need to be ambitious.
@notnotstorm IMO this isn't even close to aggressive enough. We need to figure out how to pull up the 2027+ timeline to early 2026. cc @VitalikButerin.
62
106
1K
A lot of people making jokes about WETH. Please be aware it may not be obvious to outsiders that it's completely different from bridged assets and there is literally almost zero risk. I think it would be better to mark these more clearly as jokes.
188
209
882
I'm thinking about a new sharding design, where instead of having independent proposers for each shard, all shard blocks in one slot are proposed together with the beacon block. This leads to a major simplification of the sharding design 1/n.
14
152
768
This is great! Celo is showing the way forward. We need to make sure that Ethereum will provide a cheap and secure data availability layer, so chains like Celo can use native data availability asap.
After many months of research, I'm excited to share cLabs' proposal for turning Celo into an Ethereum L2! . This is the start of Celo's return home to Ethereum!.
246
340
549
Max was very helpful in finally opening the Overton window to speak about stuff that isn't perfect in Ethereum land. "but he fudded my bags" yeah you know I don't care about how your bags fare over the next two weeks. In the long term, we need to build a good product, can't do.
39
62
740
Decentralization does not work like this. No economic incentives will automatically guarantee our values such as immutability or censorship resistance -- the values of the Ethereum community do this. Our technology, as well as the incentives built into our systems, are a tool for.
@kunalgoel @dankrad @TimBeiko @dannyryan Hoping on social agreement holding in a trustless system is not smart. The entire premise of crypto is that it's economically guaranteed. Hand shakes are not enough to contain staking and restaking.
90
227
484
OK guys, good time to get back to buidling! We need. * Layer 2 ready. * EIP1559. * Eth2 running.Then we can have a real party :).
33
70
574
Toly trying to find gotchas in my blog posts. Bullish Ethereum.
How far the mighty have fallen. Solana is built for raspberry pi version N+10, where N is the Ethereum version.
11
24
553
Calldata cost reduction will make rollups 5x cheaper: I am for doing this ASAP. As in before Christmas. $100+ fees for using Ethereum smart contracts should be seen as an emergency, and this is the (temporary) solution (before data shards). Let's do it!.
15
99
534
Ethereum: Actually realizing the crypto vision. Yes, it looks messy, but that's because Ethereum is building for the real world. But it's solving ALL the necessary problems and not a tiny subset.
bitcoin: digital gold. solana: global composability. cosmos: chain sovereignty. celestia: modular stack. monad: parallelizable evm. berachain: native staking. ethereum:
41
68
535
I think it is time we increased the gas limit. If you run a validator, you can contribute to this -- every block acts as a vote on whether to increase it. If more than 50% of the stake agrees to an increase, the gas limit will increase. I just set my validators to vote for up.
32
86
530
ZK rollups are awesome. But universal smart contract execution proving, at a scale similar to what the EVM is doing now, is several years away.
33
52
484
@aeyakovenko Isn't Solana running at 10GBps now?.Remember you telling me a few years ago that one day you could run a validator at home thanks to Moore's law, how is that going?.
18
23
498
Crypto markets value hype over execution. Ethereum is currently in execution mode.
45
32
488
Together with Justin Drake, I have recently decided to become an advisor to Eigenlayer, on the same conditions -- I am taking this position personally, not representing the Ethereum Foundation, and with a focus on risks and decentralization. I am therefore fully expected to take.
75
42
452
Look it's simple:.Being aggro doesn't grow your ecosystem. It looks bad from the outside. Most people looking for a tribe don't go "oh look, this one is the most hostile to outsiders, so I'll join them". If you don't grow your ecosystem, coin don't go up.
45
42
446
Let's say you build a lot of roads. But there is nobody lives there to use them. Is it valuable? Not yet. Then you build a lot of houses. But nobody wants to live there, because there are no jobs and no other people. Is it valuable? Not yet. Then you build a lot of factories.
39
65
435
Check: It was years, but definitely fewer years than I expected. Even though I have to eat my words I'm very happy about it! It looks like this is going to be the year of the rollup.
ZK rollups are awesome. But universal smart contract execution proving, at a scale similar to what the EVM is doing now, is several years away.
15
68
361
If you are staking on Ethereum and currently running Prysm, this is for you: Run the majority client at your own peril!.In this post I detail the enormous risks that staking with the majority client incurs.
180
178
362
Anyone working on a rate limited version of Tornado? Say 1000$/week based on Proof of Humanity. Good privacy for all casual users and useless for money laundering.
40
34
401
Quick writeup to explain what's happening about the gas limit raise:.Different from almost all other protocol matters, the gas limit is directly controlled by the validators. With every block, a validator votes whether to raise or lower the gas limit. So, very unlike the EIP.
to understand the dangerous path ethereum is taking right now, we have to understand *why* the gas limit isnβt being raised. itβs not just a question of the gas limit not being raised, which would lower tx fees significantly. itβs not just a question of why, when this is.
15
55
405
This is a very bad idea. ETH and liquid staking derivatives are different assets with different risk profiles (and rewards), and users should have a choice about which one they want to use. The same is true for different types of stablecoins.
Introducing Blast: The only Ethereum L2 with native yield for ETH and stablecoins. Weβve raised $20m from @Paradigm and @StandardCrypto to build the L2 that helps you earn more. Details on how to get early access at the end of the threadπ
25
51
361
EIP-4488, which would have lowered calldata cost to 3 per byte to make rollups cheaper, was proposed just over two years ago. We really should have implemented it.
33
58
335
Statelessness/Verkle does so much more than decrease the cost of running a node by the cost of one SSD. First and foremost, the most important question should be: Do you want to increase mainnet gas 10x? Verkle will enable this.
what % of people with 32+ unstaked ETH aren't staking because of full node operations being too cumbersome?. what's the "TAM" of Verkle trees?.
20
70
356
PSA: If it doesn't use Ethereum for data availability, it's not an (Ethereum) rollup and therefore not an Ethereum L2. [Exceptions: Plasmas/State channels or other constructions that *do not* require data availability for security].
51
42
341
This is such a lazy take if you can't also tell me what your supposed solutions are that let local block producers extract MEV. For everyone who hates so much on Mevboost now. In 2022, before the merge, we talked to Flashbots what their plans where. They said that they would.
MEV became an issue, so instead of trying to find solutions against it, we've glorified it and are now catering the protocol and infra around it to proprietary MEV builders instead of local block producers. It's futile to fight against MEV os might as well lean hard on it, right?.
15
31
328
Actually true. Let's fix this asap.
12
31
340
If you bully people for "doing too little", you create a climate where people would rather not donate anything at all.
27
39
304
Teaching GPT-3 about KZG commitments. I crashed it at the end (as probably most human interlocutors would do long before), but the ability to take in new instructions/information on the spot is absolutely astonishing!. The age of AI starts now. No more winters I would predict.
17
42
314
$172k in new bounties for breaking ZK-friendly hash functions! Rescue Prime, Feistel-MiMC, Poseidon and Reinforced concrete: Happy #Cryptanalysis!.
282
125
274
I am for increasing the gas limit, but we need to be more careful. I think after 4844, the best way to do this would actually be: . - Increase the execution capacity by increasing the gas limit BUT ALSO increasing CALLDATA cost (ETH L1 is for execution, not data). - Increase the.
Thatβs a framing of the upcoming Dencun update that I havenβt seen yet - given that the blob data will be in addition to the existing block limits, it is equivalent to a raise of the gas limit by ~11M, just with that gas reserved for L2 data.
13
40
282
Future stakers, remember that you will need a synced Eth1 node! I recommend setting that up now. Don't leave it to the last minute [it takes a long time to sync].
29
148
277
I'm generally a gas bull and think it is time for a gas limit increase. I think the best would be a package of measures:. - Increase execution gas limit (to 40m or even more). - Increase blob count (3->8). - Implement EIP-7623 to limit the max block size
Today, @nanexcool and I are launching an effort to help raise the Ethereum gas limit from 30mn to 40mn. This can result in a 15-33% reduction in L1 tx fees. We are calling on solo stakers, client teams, pools and community members to help. #pumpthegas .
19
48
273
This is the correct take. Lido is a great addition to Ethereum between 20-30%, maybe the best of all staking pools. But its ambition to go far beyond this threshold, and failing to see a problem with that, is what makes it problematic. I think it is not correct to call Lido.
25
43
263
You have been lied to about ERC20s. They aren't secured by Ethereum. It's just social consensus, any ERC20 community can just fork away.
44
29
257
@smsunarto @0x_cooki @zengjiajun_eth Blockchains are a tool for coordinating social consensus. Your values seem to be pure individual profit maximization which are not mine. My values (and most of the Ethereum community) include many others, including decentralization, censorship resistance and building an immutable.
16
39
241
In order to increase transparency in this space, I am further creating a disclosure page to disclose all potential conflicts of interests, including advisory positions I've held prior to Eigenfoundation. The reason I have not advertised these positions is because I wanted to.
17
18
249
Actually, crypto will stop being a joke when crypto twitter stops talking about the price literally 90% of the time.
6
13
231
My reasoning on why I'm for EIP-7514. It is currently unclear if (especially liquid) staking will keep growing indefinitely. In the case that the withdrawal queue does not empty over the next few months, the lower churn limit will give the Ethereum community the time needed to.
16
70
174
Many misunderstandings out there why Terra failed. It wasn't paying 20% on Anchor, though that was clearly one of the reasons it became so huge. But the design was fundamentally flawed in another way: The circular dependency between UST and Luna.
13
31
222
@norswap @TrustlessState @eigen_da The definitions are important. A rollup has trust guarantees very close to those of the settlement layer, because it has it's data on a layer that's coupled. In particular, it doesn't need any honest majority assumptions. Any chain that has it's data on a different data layer.
12
25
204
There's another reason to be upset about this: The native yield thing can already be done on any L2. You can already hold those yield bearing stablecoins and liquid staking derivatives, and chose exactly the asset you want and do DeFi with it. You can also pay your gas fees.
Introducing Blast: The only Ethereum L2 with native yield for ETH and stablecoins. Weβve raised $20m from @Paradigm and @StandardCrypto to build the L2 that helps you earn more. Details on how to get early access at the end of the threadπ
15
38
189
@lex_node @mteamisloading L1 does not need to be as fast as Solana to be competitive. It just shouldn't be so slow as to push users away. TPS are not the main metric on which chains compete. To me it looks like >10x is achievable without sacrificing home staking, we just have to quantify what the minimum.
15
31
215
When someone from your tennis club contacts you for advice on which Eth2 client to run on their NUC for diversity. That's how you know that at home staking in Ethereum is very real.
6
15
210
Counterpoint: I think people overestimate how much execution scaling is a bottleneck, even for the EVM. The current, naive EVM runs at over 100 mgas/s on consumer-grade hardware (no parallelization). That's around 5000 tps on transfers. So when the EVM is parallelized, your.
Eventually your rollup faces the same βlocal state hotspotβ issues as Solana and may want to price/queue different parts of the state access hotspots differently. Parallelization doesnβt solve the problem if you have two buyers each wanting to buy 100m of $BALD at the same time.
13
25
176
EIP-1559 has multiple aims, and I support all of them. More predictable inclusion of transactions, gas price oracles and burning of fees to decrease ETH supply are all good for Ethereum #supportEIP1559
3
41
195
@sreeramkannan has already made a strong argument why choosing the base of your L2 is not about "which L1 is the most secure", but about "which L1 do I want to settle on". Assets which settle on the same L1 that you have data availability on can get safety that goes far beyond.
By today's market prices, Ethereum has $110b staked while Solana stands at $70b staked. Solana's economic security is getting really close to that of Ethereum. Isn't economic security one of the major reasons L2s choose Ethereum as its home?.
14
40
211
@TrustlessState @eigen_da Rollups will not because if they use a non native DA layer, they aren't rollups.
10
15
206
It's always dangerous to talk about current price action. Next week everything might be differentπ
. BUT: I think Jon has a point. The Ethereum community has built a lot of cool stuff. Is ETH the asset Ethereum's product?. Or is it blockspace to power a global decentralized.
Ethereum does not have a marketing problem, it has a product problem. A marketing agency canβt help a company that doesnβt have internal agreement about what their core product even is. We need to be straight with the shortcoming if weβre going to improve it.
25
12
210
Ethereum research AMA in ~2 hours!.
the EF research team will host a Reddit AMA on Jul 7 at 1pm UTC.
92
108
140
Given the recent debates, one thing I wanted to be clear on: I haven't heard of any team writing a fork to remove Lido and it's not something I would support. Another thing to maybe clarify: My opinions are not representative of the Ethereum Foundation, and there are quite a.
Hearing rumors about clients being released to activate a soft fork that removes the Lido attackers. π.
6
12
187
Shared sequencing + shared proving give you synchronous composability. They also effectively turn all of the rollups into one rollup. There are reasons to do this or not, but we should be honest about it. .
Can we stop saying that shared sequencing enables synchronous composability by default?. The only current way for one to "synchronously compose" between the L2s sequenced together is to operate the full nodes for all of those L2s.
19
19
175
Reminder that data shards do not have any logical restrictions on which shards to use for a transaction. Any rollup can run on any shard, so censoring on one shard does nothing. Anatoly hasn't caught up yet, I guess it's because he doesn't understand rollups ;).
@FigoETH @nik_hayes @aliatiia_ @dankrad Eth2 design doesnβt minimize the cost to censor a message because shard capacity is ridiculously low. To an outside observer, which network looks compromised:. * expensive nodes, low fees extracted .* cheap nodes, fees extracted are 1000x the cost of running nodes.
7
27
173
Loved this take by Udi. Definitely worth a read.
@jon_charb hereβs my take that no one asked for:. every smart bitcoiner i talk to agrees that socially, ethereum in 2024 is EXACTLY where bitcoin was in 2020 (thatβs not a good place). basically the influencers and podcasters dictate the tone and itβs cringe. the important thing to.
16
12
176
It shouldn't really be called Proof of Work, but Proof of Waste. Proof of Waste is just an indirect and extremely wasteful form of Proof of Stake. You prove your "stake" which is in the form of mining hardware by running it and finding high difficulty hashes.
@Foobazzler Proof of Work is a simple proof of capital allocation and to certain extent, your influence and geography also. With lots of money, you are at a massive advantage. Not many can spare $100,000-$500,000 to start a profitable mining operation. But they can purchase bits of ETHs.
15
21
171
I do not know who claimed that royalties are enforced by smart contracts, but you should probably unfollow anyone who did and never take them seriously again. It's more than just a lie. A smart contract cannot enforce to take X% of the sales price, as it cannot know the price.
do eth nerds understand that royalties are still the canonical example for normies (including many artists) of what makes smart contracts useful and different? at a high level, how do you go from characterizing them as iron-clad to "eh, we'll ignore what we don't like".
23
15
171
What are proofs of custody -- and how are they different from data availability checks? Do we actually need both?. New explainer here:
68
101
161
Funny because single threaded EVM execution can reach Solana's current speed at 2000 tps. Is this all SVM can do?.
@ufukaltinok @0xMert_ @drakefjustin Weβll see. Currently EVM execution and sequencers are single threaded. Iβd expect any of those queues to force the fee market to blow up eventually.
19
16
170
Proof of reserves: Possible on a blockchain. Proof of liabilities: Entity commits to all their liabilities in a Merkle root. Gives each creditor proof that their liability is contained. Then proof of sum of all liabilities.
5
11
167
I am in favour of this. My feeling is that the name "Data Availability" has caused confusion and "Data Publishing" is clearer. Still open to better names but so far DP seems most descriptive.
Data Availability is by far the most confusing term we ever came up with. Data Publishing + Data Storage are better terms that are more intuitive. DA = Data Publishing, not Data Storage. Here are few facts that you may be unaware of: π§΅π.
10
14
164
L1s are massively overpaying for security.
Yesterday Base made $673k. Yesterday Base paid $2,500. This is what sovereignty gets you. Let a million rollups bloom.
11
16
166
I want to highlight this: Core devs don't have the last say on what happens on Ethereum, it's the Ethereum community. I have laid out why I believe the churn rate limit is the right thing to do now (although I was originally tending against). Unfortunately this came very late,.
@sassal0x @dapplion @FortuneMagazine @dannyryan I agree, a week for community input is **very** short. But mainnet hasn't upgraded to Dencun. (maybe December). There is still time to shout if the community really disagrees. We are only at testing on devnets. The case for EIP7514:
20
22
151
This take is wrong, but I understand where the many people thinking this are coming from. Ethereum is not a DPoS system. DPoS systems are designed so that staking is a highly professionalized operation that most people cannot take part in other than by delegating.
@epolynya @mewn21 @statelayer i'm sorry but what exactly do you think is happening on ethereum today. it is a delegated PoS system, just a more obscured, harder to participate, and worse one.
16
20
159
Many people I talk to seem genuinely surprised that using a TEE means just trusting the hardware manufacturer.
23
18
162
It's time to learn about Kate commitments! They are amazing at enabling ZKPs and are now being considered to replace Merkle trees in Ethereum to reduce witness sizes. Here's an introduction: #eth2 #ethereum #statelessethereum.
35
93
145
One of the problem with crypto culture is that we call projects scams and ponzis all the time. But a bad or poorly executed project is not a scam. Warnings about projects that deserve to be called a scam and ponzi, like Terra, are drowned out in the noise.
7
11
153
So we can't delay withdrawals by couple of months for 4844 (which most of the community wants desparately) but then instead will most likely delay both of them for shipping EOF. Crazy ACD decision.
32
14
155
Smart contract security is absolutely essential for Ethereum's L2 roadmap. It will be interesting how this plays out, but in practice this seems to be true: Teams that want to build L2s also prioritize security and do everything they can to avoid bugs. In contrast, those who.
"Bridge hacks are one of the largest problems holding us back but L2 will solve that!". If you think that: think again! Because the majority of bridge hacks are not "bridge operators" getting compromised but bugs in the smart contract systems. Those can (an will!) happen with L2!
11
29
141
As Eth2 sharding is coming closer to implementation, analyzing the Legendre PRF primitive is more critical than ever. We are therefore *doubling* the bounties on you can now claim $20,000 if you find a sub-exponential key recovery algorithm!.
14
90
145
Hey I've got an idea. Why don't we make a token that you can only buy and never sell. Guaranteed up only. No investors dumping ever.
51
3
158
What everyone gets wrong about 51% attacks. Correcting the myth that if you control more than 50% of the hashpower in Bitcoin, Ethereum, or another blockchain, then you can do whatever you want with the network.
4
37
147
My arguments on why we should go all in on statelessness, and not aim for an easier intermediate goal. Partial statelessness or state rent is an unhappy half way house that will make UX (and Devex) worse, with few benefits.
41
76
133
What we want is a value based DeFi sector. The difference between finance (positive sum) and a casino (zero/negative sum) is whether there is underlying value in the traded assets.
not very cypherpunk to want only a RWA based defi sector. crypto native businesses will exist as more and more usecases and industries come onchain or be created onchain. if you think defi is a closed loop because it uses only crypto native assets you simply do not consider.
20
19
146
Channelling my inner @aeyakovenko here: I am more tolerant to downtime since Ethereum is currently an unfinished project. It is just unusable for the vast majority of people and therefore a few hours of downtime matter much less than delivering sharding a few months earlier/later.
@das_grasshopper @skylar_eth Yeah except that for 99% of people on this planet, Ethereum is constantly down in practice because they simply can never afford to use it.
19
11
133
The news of FTX being "hacked" just makes me really sad. I can't say why, since much worse already happened in the last week. Maybe the pure evilness of SBF likely draining the last funds out hoping to retire somewhere.
4
9
143
@lex_node Surprisingly, my plan would medium term decrease requirements for running nodes. The tradeoffs are around:. - adding dependencies for more resourceful parties (DAS provider nodes, zk provers), but with an honest minority assumption. - giving up on home block building (which.
9
8
161
Is Solana's Proof of History necessary?. Comments welcome :).
15
17
137
This is wrong. L2s are indeed more secure because to be an L2 you have to provide the same security as the base layer. A PoS sidechain has the security of the weaker chain, usually the sidechain.
3/8 Second (and much more important/concerning), implying that L2s are *by default* more secure than PoS sidechains (commit chain in our case) is absolutely wrong and potentially dangerous. To explain why, let's quickly compare our PoS chain and StarkWare's zkRollup or Validium.
11
11
137
The year is 2030. Ethereum is processing 100,000 tps at 1 cent/tx. This is a neat $1000/s in revenue. In @koeppelmann's fantasy, fees spike 7000x, increasing that to $7m/s. That's over $25 billion per hour. Never gonna happen (as much @drakefjustin would love to burn those fees).
If demand for block-space is inelastic (insensitive to price) even a small (25%) increase in demand can cause 7000x in price. Longterm we either need to hope for elastic demand or develop other strategies to protect apps and user from fee explosions.
9
18
136
Even now, he just can't stop lying. This makes zero sense. If this were the case, there is NO WAY IN HELL YOU WOULD HAVE ASKED CZ FOR A BAILOUT!. And also, YOU ARE SUPPOSED TO HAVE THE SAME ASSETS THAT CLIENTS DEPOSITED. NO MATTER HOW MARKETS FLUCTUATE ALL DEPOSITS SHOULD BE OK!.
4) FTX International currently has a total market value of assets/collateral higher than client deposits (moves with prices!). But that's different from liquidity for delivery--as you can tell from the state of withdrawals. The liquidity varies widely, from very to very little.
6
11
132
@CryptoTruthsee1 @DomSchiener @VitalikButerin @lunfardo314 Good to see that more people want to work with verkle tries. If Iota community is looking for collaboration on this I'm happy to work together.
2
16
134
@aeyakovenko Nah that was 2019. We can update things for 2024. The blog post doesn't even claim that it needs to be a Raspberry Pi.
2
3
135
@Crypt0Panda @aeyakovenko Where can I get that 10gb connection at home and what's the monthly cost, including unlimited traffic?.
6
0
132
I've seen a lot of fear recently that the KZG commitments introduced with 4844/full sharding will be hard to use in ZKRollups. This is not true! You can make quite efficient use of them in proofs. More details here:
7
24
123
The nice thing about this curve is that it's dual purpose: You can use it to make Inner Product Arguments (like the ones needed for verkle tries) directly and it's fast for that, and you can *also* efficiently reason about these arguments inside BLS12_381 SNARKs. Great work!.
Introducing Bandersnatch: a fast elliptic curve built over the BLS12-381 scalar field: joint work with @SimonMasson2.
31
70
108
And so it's true we run a real risk of "cosmosification": The L1 is retreating more and more to its base function, and other projects fill in the gaps. And as they are building the product, they will ultimately accrue the value. This is in short, Ethereum's value accrual.
8
7
129
"issuance is not an expense". Technically correct, but you can model a blockchain as an equivalent business: The business sells blockspace and pays validators to maintain the chain. In that model, the issuance to validators is an expense to the business, and fees that are burnt.
Q1 β24 income statements for Base (L2) and Solana (L1).
25
19
122
Blockchain researcher discovers that value is social π€―, blames rollups.
7
5
118
@peter_szilagyi Complexity is like debt. Ideally you want to have none, but then you usually also can't create anything. So you of course want to minimize the complexity while still creating a system that is useful. Creating complexity in isolation is definitely bad, but are we doing that?.
1
2
117
Now on the one hand, I think it is really cool that we have this way to raise the gas limit even outside of a hard fork because hard forks take a really long time to coordinate. On the other hand, I am kind of sad by what has now become a very small "consensus" increase to only.
12
5
125
@peter_szilagyi So I think your criticism sounds like you think we are designing the protocol to satisfy monetary interests rather than to build a fair system. I can assure you that this is absolutely not how any researcher thinks. Many of us go heavily against the grain and take a lot of flak.
7
5
119