Steven Pack | RockSolid

2.5K posts

Steven Pack | RockSolid banner
Steven Pack | RockSolid

Steven Pack | RockSolid

@paladin_eth

Co-founder @rocksolidHQ. Vault guy. Product guy. Rocketpooler. Ex @Cloudflare, Ex @Mina. Tweeting on crypto/ethereum/LSTs/vaults and general tech.

The ether เข้าร่วม Ağustos 2009
1K กำลังติดตาม1.2K ผู้ติดตาม
ทวีตที่ปักหมุด
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
Today we’re announcing RockSolid’s $2.8M pre-seed fundraise and launching our platform to the world! RockSolid vaults let anyone access the best of DeFi rewards with a single click. For protocols, asset issuers and DATs, it’s a platform to integrate that single-click experience for your token holders directly into your existing flows, giving them greater access to DeFi and more reasons to hold your token. We’re backed by the best in crypto! @CastleIslandVC led the round with @theBBFund, @Kindredventures, @IDEOVC, @GSR_io, @FinoaConsensus, @PierTwo_com and a host of angels and we’re ecstatic to have them on board. And we’re not launching alone! I care deeply about Ethereum and as such, have been a big supporter of @Rocket_Pool for a number of years. Ethereum is strong because it’s credibly neutral and credible neutrality comes from stake distribution. No one does more on that front than Rocket Pool. BUT, stakers want utility for their tokens and today rETH holders get it with the RockSolid rETH Vault. With a single click, rETH holders can access the best of DeFi rewards via the RockSolid rETH Vault. Better rewards for token holders, more reasons to stake and hold rETH! A stronger rETH DeFi ecosystem means a stronger ethereum. And we’re just getting started. Crypto is reshaping finance and vaults are at the forefront of that. They need to be simple; they need to offer great returns; and they need to be RockSolid. Wherever there’s a crypto balance, expect to see a RockSolid vault.
RockSolid Network@rocksolidHQ

1/ Excited to share that RockSolid has raised $2.8M in pre-seed funding led by @CastleIslandVC w/ backing from @Rocket_Pool, @GSR_io, @kindredventures, @theBBFund, @ideoVC, @StanfordSBA, and others. Today, we're also launching the first official rETH vault with @Rocket_Pool.

English
21
4
47
5K
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
@build_aus Actually it's not the certificate. It was flagged by security software. You should submit it to Cloudflare Radar and other security filters to get it categorized correctly. Looking forward to checking it out!
English
0
0
0
59
Build Australia
Build Australia@build_aus·
The old Australian dream is dead. Today, we launch a new movement for those who think seriously about the future of Australia, and believe in it. Website is now live. Along with our first essay, outlining what we envision as the future of Australia. Join us to #buildaustralia
Build Australia tweet media
English
75
96
647
209.2K
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
@LayerZero_Core Could you guys share if the attackers did anything specific to cloak the op-geth swap from the EDR system? Did it not pick up and flag that? Or was it for too short a time to give notice? Understanding that would help others improve opsec.
English
0
0
0
142
Vladimir S. | Officer's Notes
This is some of the OFT adapters at risk. Urgent for teams! Yes, the @KelpDAO exploit is different. But also relies on this principal of trusting 1 source Because any OFT that inherited LayerZero's default DVN config (single LZ Labs DVN, no custom multi-sig threshold) shares the exact fault domain that let the rsETH forged packet through. If the root cause ends up being a signer-side compromise rather than Kelp-specific misconfig, every OFT on that default path is theoretically mintable the same way until they add independent DVNs or a threshold. Most OFTs never bothered overriding defaults — that's the "at risk" list. Not all of them will actually get hit (attacker needs liquid destination markets too), but the surface is real. drive.google.com/file/d/1azNbR1…
English
4
15
91
13.2K
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
@delitzer I think they'll let it burn on L2, try get some external contributions onto mainnet, socialize the rest where hopefull "the rest" is < Aave's Umbrella coverage.
English
0
0
0
157
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
So ~$200M of rsETH released during a bridge with no corresponding burn, supplied to Aave on Arbitrum and Mainnet and borrowed against. With rsETH missing ~12% of its backing. Aave holds the bag (bad debt), but some combination of Kelp/LayerZero likely the cause. Scenarios? - Aave umbrella? Most of the rsETH on Aave is for looping, so nearly all the borrows are ETH. There is $50m in Umbrella for ETH on mainnet and $1.2B rsETH supplied, so a 12% hair cut on that is > Umbrella ETH coverage. I.e. there will be bad debt if the 12% haircut is recognized on mainnet. (rsETH redemptions currently frozen). No Arbitrum Umbrella AFAIK? - Some combination of parties plugs the hole? rsETH goes back to fully collateralized before redemptions are re-enabled, liquidations and bad debt avoided? - Limbo? rsETH redemptions stay frozen while the parties battle it out. No liquidations until it is. ETH borrow rates stay elevated.
English
3
1
16
2.1K
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
Yessir 🫡. Short term sugar hits are tempting wrt to TVL, but as you'd appreciate @rektdiomedes; it's empty calories. Better to grind and build up rocksolid, sticky customers and TVL.
rektdiomedes@rektdiomedes

It is easy to be disheartened about defi right now... However at the same time... there are fresh, idealistic teams absolutely getting after it... And sowing the seeds of what will be our (imho) pre-ordained long-term subsumattion of the archaic and ossified legacy tradfi financial rails... @rocksolidHQ is a perfect example (see QT'd tweet from co-founder @paladin_eth below)... I joined up with these gents as an advisor late last year and have watched them build and ship in the time since then and they have literally bent over backwards to do everything by the book: - Obsessive, hyper-intentional focus on security (again see tweet below) - No mercenary deals for short-term TVL gains - No weird gimmicks or campaigns - And in addition, they have even taken the bold position of eschewing a points program despite being tokenless, and are attempting build a fully-functioning successful on-chain business without any airdrop-incentivized infusion of users and/or TVL... And they've been pretty darn successful thus far... with one 8-figure $ETH vault and one 7-figure $ETH vault despite $ETH cratering in the time since they launched (right around the market peak in early October)... And imo they are perfectly positioned to thrive when the bull market arrives later this year, as I agree with the suggestion many have put forth that "vaults" will be the defining term of this cycle and the primary manner in which defi reaches the masses 💪 But yeah - @samkamani just published a few great clips from his recent interview with them, so I am going to embed all of those below in the thread! Also I am going to flip the partnership tab thingy on this since I'm an advisor for them (although I've noticed half the time it doesn't show up when I do threads for whatever reason)... And then will also provide a link to their two vaults at the end of the thread too :) And finally, if the above sounds appealing to you, make sure to... - follow the official @rocksolidHQ account - follow co-founder @wardy_ben - follow co-founder @paladin_eth - check out their vaults if you want some good hearty yields on $ETH! And as always please RT to support 💪🙏

English
1
1
8
1.2K
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
A few thoughts on Resolv and other vault blows ups recently 1) Resolv was an OpSec hack. An EOA controlled an admin role and the key was hacked or used by an insider -> Check your vault platform uses either an MPC or a Safe for all admin roles.... If it's not possible to do onchain, an auditor can help. 2) The hack is a specific version of a generic type of hack; oracles. The SERVICE_ROLE wasn't an oracle in the traditional "price oracle" sense, but it was an oracle in the sense it was an external entity verifying some action had taken place (USDC deposits had been made and a swap requested). DeFi in general and especially vaults keep getting hit by this over and over. Even sophisticated designs and sophisticated teams (e.g. the Aave wstETH oracle) make mistakes. Vault platforms have essentially 2 options; trust the code only or add human oversight. -> RockSolid vaults require 2 parties to agree in order to "oracalize" an NAV update. A valuation manager submit their suggested price (can be automated or human) and a secondary MPC that uses secure enclaves and humans in the flow is required to confirm. Hacking one or the other is insufficent. It's not possible for "block-by-block" live NAVs, but it's RockSolid and it's how we've avoided many of the recent issues. 3) Stream/Elixir was just blindly handing deposits to another party with "Trust me bro" agreements. -> RockSolid allocations are visible realtime onchain (see Debank link on our vaults) and we add human commentary as allocations change so our depositors can understand the why, not just the how. -> RockSolid vaults operate under a mandate. A curator can propose a change in strategy, but they can only execute whitelisted strategies that have been reviewed and are then published. Separation of roles make our vaults more RockSolid.
English
5
1
10
5.2K
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
@evanashapiro Yeah I found recently “model this as a state machine” is a good first step to making the model use a reasonable abstraction rather than keep specializing
English
0
0
1
51
Evan Shapiro
Evan Shapiro@evanashapiro·
My best model of ai coding agents right now, is of the most brilliant software engineer, that never thinks to apply any kinds of abstractions. >Working on a game tree? It will be happy to iteratively keep adding javascript cases in huge nested ifs, without understanding failure cases. Tell it, "This seems close to a dynamically generated DAG with end states", and it will happily recognize it and simplify the logic. > Working on a DOM animation? It will happily add random javascript and css to realize something that seems like it could work, without having a sense of if the pieces fit together. Tell it, "Use FLIP and make sure to represent the animation as a before state, after state, and animations between the two", and it will happily reduce complexity to something that is clearly right. What I find interesting, is it can clearly handle far more complexity than at least I can handle, but it still reaches some point where it breaks, without realizing there is some abstraction to collapse the complexity and arrive at something more easily verified as working. It seems like the last thing I've had to do while agentically coding that feels even a little like programming / CS. Once noticing and applying opportunities for abstractions becomes part of how agents work, I suspect they'll be quite good at the whole end to end process of software development.
English
5
2
11
1.1K
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
UX reminder: you can forcibly remove footguns from your app -- and you should.
Steven Pack | RockSolid tweet media
English
2
1
4
1.3K
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
On #1, it's not fertility but net population change that informs the "number of bodies" that need a house, so highly desirable countries with net immigration can stave off population decline for as long as they're willing to accept high numbers of migrants. Ergo, i'm not as bearish on housing.
English
0
0
1
354
Steven Pack | RockSolid
Steven Pack | RockSolid@paladin_eth·
@JoshKale Sweet. Backups coming in hot! "Dude, what happened on that skydive?! Ugh, chute fail. Got bodylossed. Only lost a few hours though, backed up before hand. New shell is sweet"
English
0
0
0
161
Josh Kale
Josh Kale@JoshKale·
Scientists just copied a Fruit Fly's biological brain and trapped it inside of a computer. Not an AI model trained to act like a fly... A total digital copy of a fly !! This is some sick sci-fi stuff: - They scanned and copied the brain, neuron by neuron, synapse by synapse, from electron microscopy data. - Then dropped that brain into a simulated body in a video game like environment. The fly walked. It groomed. It fed. Nobody taught it anything. The behavior was already in the wiring. The entire premise of modern AI is that intelligence is something you train into a system. This is proof it's something you can transfer out of one. Wild times
Dr. Alex Wissner-Gross@alexwg

x.com/i/article/2029…

English
753
2.7K
19.2K
2.7M