jgriz.dev

322 posts

jgriz.dev banner
jgriz.dev

jgriz.dev

@jgriz_dev

trading / crypto / dev / follow for bad takes

Denver, CO Katılım Mart 2018
977 Takip Edilen111 Takipçiler
jgriz.dev
jgriz.dev@jgriz_dev·
This will ultimately be the killer feature of an L1- the ability to remediate bugs and hacks. The L1 could have a "canonical fork path" which is decided by some kind of court process. Protocol clients by default follow this fork path, however anyone can configure their own fork. IE you can never force people to follow a fork but most will follow what is considered "legitimate" such that RWA issuers honor on chain state. Building legitimacy is they key, which probably requires some degree of democracy and sybil-resistant identity. Some kind of hierarchy of chains or sub-VMs is needed to isolate state such that only affected part of state is frozen during court proceedings while rest of ecosystem continues. L1's with no fork governance will devolve to chaotic "social forking" (as it is today) which even the threat of will discourage large institutions from having funds on. Most anti-governance never-forkers would change mind if they are personally affected or bug/hack big enough.
English
1
0
1
79
_gabrielShapir0
_gabrielShapir0@lex_node·
this is brilliant, have been brainstorming at @delphi_labs about similar things for a while...I think onchain courts, onchain 'league of nations' mutual security agreements, etc., are inevitabilities
ALEX | ZK@gluk64

Smart contract implementation risks remain the biggest unsolved problem of Defi. L2s are equally affected. Let me pitch an idea: L1 Fork as the Court of Final Appeal. First, why existing solutions don't work: 1) Time-locked upgrades are great for scheduled changes, but obviously don't work for emergency situations. 2) Converting to an alt L1s only adds more problems and solves nothing. Even though an alt L1 can be forked, assets bridged over from Ethereum cannot, and the fork choice resolution is no different than an upgrade problem. See this thread for example: x.com/gluk64/status/… 3) Security councils can mitigate the problem, but not solve it. Any gov multisig empowered with emergency upgrade rights poses regulatory and security risks of a different kind, rendering a Defi system immature (@ChrisBlec won't let us forget about it). 4) Combined gov mechanisms are better: a security council could only freeze the contract temporarily, requiring a token governance approval for an emergency upgrade. But now a malicious majority of undercollaterized stakers could perform an evil take-over upgrade and steal all the assets. L1s don't face the same problem because they are forkable: any user can opt-in into the fork branch which they subjective believe to be correct and canonical. This doesn't work for L2s and Defi protocols, unfortunately, because we can't fork underlying native assets bridged from L1 (such as ETH). Or can we? What if we forked the L1 itself and let the L1 social consensus resolve an issue for a specific protocol? This is tempting and perfectly possible technically, but creates two problems. First, it would only work for large protocols – for the smaller ones, L1 users will likely not bother to take any action. But second, even with the big protocols, we run into a risk of overloading L1 social consensus, which Vitalik has explicitly warned against: vitalik.ca/general/2023/0… Now, hear me out. What if we built a hierarchical system of onchain courts similar to the real-world judiciary? Every protocol will have its own governance with normal and emergency upgrade mechanisms defined. The protocol must also designate a special contract that serves as an instance of appeal. We'll need a standard ERC interface for it. For an emergency upgrade there must be an appeal period, during which anyone can submit a challenge to the higher court. They will have to put a pre-defined bail deposit. This court can then cancel the emergency upgrade (and do nothing else). Different courts will have different members, prices and reputation, in a completely decentralized manner. Each court will also have to specify the a higher court where any decision can be appealed – until at some point we reach the Ethereum Supreme Court. The decision of this smart contract can only be determined by a (technically soft) fork of the L1. Of course, anyone can deploy a contract like this, and there is no guarantee that the users will bother to soft-fork. So the idea is to form a social consensus argound this idea, and deploy a "canonical" instance of the court that be configured to be so expensive to use, that only truly extrodinary cases will be brought before it, and thus will likely be worthy of the attention of the entire Layer-0 (the social consensus) of Ethereum. Think of a bug in @Uniswap, a major L2, a Defi protocol with a systemic risk, etc. The most important function of such a system will be to protect protocols against political inference from the outside. It will serve as a great deterrence mechanism, and will elevate the role of Ethereum as a powerful network state. What do you all think? Is anyone working on something like this already? We at @zksync will be more than happy to fund such a research.

English
7
5
45
16K
jgriz.dev
jgriz.dev@jgriz_dev·
Either put dapps in own arbitratable VM or spoke chains of court hub which specifies canonical fork path. Withdrawal from dapp VM/chain can be delayed to allow quick-reacting security council to freeze VM or specific accounts (so hack does not contaminate state of other chains), then court rules whether bug/hack occurred, and if so how to alter VM state. Lots of difficult technical implementation details to work out, and I suspect democracy (and thus sybil-resistant identity) will be needed to give the court legitimacy.
English
0
0
0
6
ALEX | ZK
ALEX | ZK@gluk64·
Smart contract implementation risks remain the biggest unsolved problem of Defi. L2s are equally affected. Let me pitch an idea: L1 Fork as the Court of Final Appeal. First, why existing solutions don't work: 1) Time-locked upgrades are great for scheduled changes, but obviously don't work for emergency situations. 2) Converting to an alt L1s only adds more problems and solves nothing. Even though an alt L1 can be forked, assets bridged over from Ethereum cannot, and the fork choice resolution is no different than an upgrade problem. See this thread for example: x.com/gluk64/status/… 3) Security councils can mitigate the problem, but not solve it. Any gov multisig empowered with emergency upgrade rights poses regulatory and security risks of a different kind, rendering a Defi system immature (@ChrisBlec won't let us forget about it). 4) Combined gov mechanisms are better: a security council could only freeze the contract temporarily, requiring a token governance approval for an emergency upgrade. But now a malicious majority of undercollaterized stakers could perform an evil take-over upgrade and steal all the assets. L1s don't face the same problem because they are forkable: any user can opt-in into the fork branch which they subjective believe to be correct and canonical. This doesn't work for L2s and Defi protocols, unfortunately, because we can't fork underlying native assets bridged from L1 (such as ETH). Or can we? What if we forked the L1 itself and let the L1 social consensus resolve an issue for a specific protocol? This is tempting and perfectly possible technically, but creates two problems. First, it would only work for large protocols – for the smaller ones, L1 users will likely not bother to take any action. But second, even with the big protocols, we run into a risk of overloading L1 social consensus, which Vitalik has explicitly warned against: vitalik.ca/general/2023/0… Now, hear me out. What if we built a hierarchical system of onchain courts similar to the real-world judiciary? Every protocol will have its own governance with normal and emergency upgrade mechanisms defined. The protocol must also designate a special contract that serves as an instance of appeal. We'll need a standard ERC interface for it. For an emergency upgrade there must be an appeal period, during which anyone can submit a challenge to the higher court. They will have to put a pre-defined bail deposit. This court can then cancel the emergency upgrade (and do nothing else). Different courts will have different members, prices and reputation, in a completely decentralized manner. Each court will also have to specify the a higher court where any decision can be appealed – until at some point we reach the Ethereum Supreme Court. The decision of this smart contract can only be determined by a (technically soft) fork of the L1. Of course, anyone can deploy a contract like this, and there is no guarantee that the users will bother to soft-fork. So the idea is to form a social consensus argound this idea, and deploy a "canonical" instance of the court that be configured to be so expensive to use, that only truly extrodinary cases will be brought before it, and thus will likely be worthy of the attention of the entire Layer-0 (the social consensus) of Ethereum. Think of a bug in @Uniswap, a major L2, a Defi protocol with a systemic risk, etc. The most important function of such a system will be to protect protocols against political inference from the outside. It will serve as a great deterrence mechanism, and will elevate the role of Ethereum as a powerful network state. What do you all think? Is anyone working on something like this already? We at @zksync will be more than happy to fund such a research.
ALEX | ZK tweet media
English
138
170
616
199.1K
jgriz.dev
jgriz.dev@jgriz_dev·
19/19 This thread just scratches the surface of what is enabled by decentralized Sybil-resistant identity. The explosion of innovation that will follow will dwarf everything crypto has seen so far. I predict it will happen 1-3 yrs after @nillion launches, so get ready!
English
0
0
2
108
jgriz.dev
jgriz.dev@jgriz_dev·
18/19 The US gov will be skeptical, but Americans may out-coordinate and out-fundraise the existing political parties with "AmericaDAO" or similar. Note, this will require some KYC providers to attest AmericaDAO members are in fact US citizens.
English
1
0
1
312
jgriz.dev
jgriz.dev@jgriz_dev·
1/19 Why Sybil-Resistant Identity is the next bull market 🧵 For years hackers have used bots and now AI posing as humans to exploit various systems. In crypto, airdrops are gamed with thousands of wallets made to appear like unique users when they are really controlled by 1.
English
1
1
4
840