atiselsts.eth

@tallyxyz 🫡

Most yield products optimize for APY and bolt on risk as an afterthought. All @chaoslabs vaults are powered by the same Chaos AI risk models that have secured trillions in vol and billions in asset deposits. The appetite for Low-Risk DeFi is real. Higher.

Their excellent writeup: x.com/CoWSwap/status…

Correction: multiple @CoWSwap solvers submitted quotes, but just one could be executed. And if you're wondering what's the point of a "minSolvers" flag since the user can set max slippage: info asymmetry. Solvers know the fair price better than the user. x.com/atiselsts_eth/…

In addition to everything else, @CoWSwap orders should be able to specify: "only execute if ≥ N solvers participated." Here just one solver submitted a solution. There was no competitive auction. x.com/Ehsan1579/stat…

Gauntlet decided to cancel the proposal, it's now gone from Snapshot. Shows that the Uni DAO still has some power left! Broadly speaking, this is what DAOs should do: check and call out BS, instead of trying to execute on their own. x.com/atiselsts_eth/…

1/ stETH CAPO Misconfiguration Today, a misconfiguration on Aave's CAPO oracle caused wstETH E-Mode liquidations, resulting in a loss of 345 ETH. No bad debt was incurred, and all affected users will be fully reimbursed. More below.

Will be voting against Gauntlet's proposal to extend incentives for @Uniswap

> can't use GPL code can through core (BUSL) and periphery (GPL) separation, in all except the core code > hurts users v2 clones hurt users far more than v3 clones > real protection proliferation of v2 clones vs. most teams waiting for April 2023 to clone v3 > victory is unrelated for shorter term dynamics it's absolutely related, e.g. Quickswap was completely dominant on Polygon in 2021, for longer terms far too many other factors also -- no team will reimplement v3 swap code from interfaces, from scratch in 2 weeks. LLM is going to simply copy existing implementation

No. Argument zero is just a licensing fact. Argument one hasn't applied. Argument two is also a licensing fact. Argument three is strengthened by Uniswap's history; v2 is Free software and Sushi lost because Uniswap was simply the better team. Uniswap is absolutely killing it right now. If Uniswap flipped v4 to a Free software license this instant, they would still be killing it. A lot of the recent work, like the CCA contracts, aren't even BUSL either.

TLDR: I joined @ethereumfndn as DeFi Coordinator 1] I got introduced to DeFi back in 2019 and stuck to it ever since. As narratives appeared and faded away, my general belief in DeFi stayed. I think today, more than ever, Ethereum is the right place to grow DeFi further. I'd like to help make this vision a reality 🙏

@_Enoch The history of Uniswap literally has a counterexample for each of your arguments

Hi @ivangbi_, thanks for replying. Here is my overdue response. (A note to the reader: Free software means Free as in Freedom #fs-definition" target="_blank" rel="nofollow noopener">gnu.org/philosophy/fre…). In short: the use of BUSL and similar unFree software licenses is not Defipunk, not cypherpunk, and unEthereum. It is worse for users and ultimately I argue that it is not even necessary for projects to include this restriction. I hope these arguments can reform you for Free software sins once committed, and ignite in you a Crusader's passion. 😁 > Second, indeed the EF agenda is to support protocols that are open-source (not just source-available) because they need to pass the walkaway test alongside maintaining a few other decentralization traits. I love everything about this agenda. I love everything about this mission. The Foundation must maintain, with militant fervor, the stance that software must be Free in DeFi. The Foundation is among the few actors who can afford, by currency of their virtue, to promote this. I am especially in favor of the Foundation taking a hardline stance that the carrot of their patronage shall never touch unFree software, and that they will never promote the use of unFree software. This is truly for the good of all DeFi users and ultimately neutral-if-not-good for DeFi builders. Without mentioning at all any of the _benefits_ of having a strong copyleft license with a friendly forking culture, I'll limit myself to the arguments about why BUSL doesn't meaningfully help DeFi users or builders. Argument zero: being BUSL means you can't use any (A)GPL-licensed (or other copyleft, viral licensed) code. This can be an annoying limitation when developing things, given the proper prevalance of that license in the ecosystem. Argument one: it hurts your users in adversarial conditions. DeFi must be trustless, and that means it must pass the walkaway test, and that means it must be forkable. The Foundation is correct to care about this. The recent AAVE disaster proves this. Stani ruined his DAO but, because of the licensing, ultimately users are the ones who suffer. ACI and BGD Labs can't instantly spin up a fork. They need to do something else instead. Which brings us to how easy that is... Argument two: the license ultimately does not give you any real protection. Bad actors will simply fork it and ignore the license. Good actors can simply fork it too. Interfaces, tests, and "simple mechanisms" cannot be licensed. This means that clean room reimplementations are trivial, especially in the age of LLMs. A talented engineer could take BUSL software, create a copy with no specific implementation code (only your test cases, audits, and interfaces), and then have an LLM assist them in filling the gaps until all tests pass. Et voila, it would take less than a week and they can relicense the new reimplementation however they'd like. Using unFree licenses therefore is really just a very flimsy paper shield which turns away certain discerning users. Argument three: victory is entirely unrelated to the license. I have had discussions with a few different DeFi builders about this. The license is never the moat. Liquidity is heavily biased by network effects and the sense of security that users feel when you have a good reputation from being a consistently-honest actor. If any BUSL projects decided to relicense to AGPL today instead of waiting until it automatically rotates later, they would not suffer: they've got established users and liquidity and reputations. Every successful DeFi protocol which launched with a Free software license immediately, from day one, and succeeded anyways is proof of this. They're not suffering from forks because a fork does not have their team, network, reputation and ability to onboard people. The smart contract is never the full product; it's the frontend, the socials, the community, the token holders, etc. Ultimately, if DeFi builders launch under AGPL instead of BUSL, it would 1) be courageous and 2) have fuck all impact on whether they succeed or fail. People use products because they trust the brand and the team whose reputation inspires confidence. With those arguments out of the way, I will address your questions. > I am just a fan of the argument that at the start (unless it's a public good infra) it's natural to want that moat. SO if one agrees with this argument, what's the license that would permit this? A license restriction is a user-hostile position; there is no permissible license. It doesn't matter anyways because the license gives no moat. > Could "a path to full decentralization" be a viable option? No. Like you said, money corrupts. Training wheels are a lie, progressive decentralization is a lie, it never happens. The failure of major L2s to achieve proper trustlessness is billions of dollars of proof. > (1) a moat isn't allowed in this sense Moats are natural and good teams will have good moats without using restrictive licenses. They'll just have better products or better adoption. > (2) a moat must be at a different layer of the stack meaning the codebase can simply be open-source (4)... Speaking of number 2, that can be achieved via a not open-source UI at the start, but then that's still scary as a no-moat in bull markets and isn't a good thing by itself either (?) A closed source interface is not ideal, but it is nowhere near as bad as an unFree contract. The same argument as earlier applies: there is no moat from the contract. If you put a gun to my head and told me that closed software needed to exist, and a project was compelled to have an unFree component, the interface is the layer where I think it is "least bad". > (3) BUSL with a max legally allowed 2 year timeline In the grand scale of evils, BUSL with short timelines is better than something that is closed forever. It's still ultimately something that is entirely unnecessary and more a reflection of fear than the realities of any actual protection. I didn't mention any of the specific benefits of Free software licenses here, but there are benefits. The one thing I will mention: builders confuse cuck licenses like the permissive MIT license with viral, copyleft AGPL-style licenses. There is this misconception that if you license something as Free software this means someone else can take it and run with it. This is only true of cuck licenses. Choose copyleft, choose the AGPL, choose the VPL, choose the Seppuku license: you forever bind the forker to make freely available their own implementation. May the best execution win. Ultimately, I don't need to convince anyone too afraid to choose a proper Free software license. I simply know they will never make it into network heaven, and continue to espouse the correct cypherpunk principles at all times. My life for the network.

@MattFiebach Not sure if there have even been contentious votes on asset listings, but my point is more that the DAO had to select and fund SPs like ACI & Chaos otherwise the protocol would not exist today.

@atiselsts_eth If ACI+Chaos support a new asset listing did the DAO ever really need to vote on it?

Uniswap UNIfication and Aave Will Win are the same proposal. Labs becomes the funded operator with consolidated control over the direction of the protocol. UNI passed with consensus. Aave Will Win barely cleared temp check with Labs wallets tipping the vote. The difference is Uniswap spent 2 years quieting its DAO first so by the time the vote happened it was a non-event.

@koeppelmann not that surprising, we are more likely to be born at a time when there are many humans around

Currently, it seems quite likely that it is exactly during our lifetime, in the ~300,000-year history of humans, that we live in the period with the most humans on Earth.

@DefiIgnas For the record, Uniswap Unleashed asked for $165M, not $70M (op budget + grants + liquidity incentives).

I also support Aave will win proposal in principle: - 100% rev share and IP rights to foundation is a win. A sexy proposal. The difference between Aave and Uniswap is the DAO disagreement: There is so much feedback that Aave Labs fails to address. I wrote a post that spending is not necessarily an issue: Lido's 2026 budget is ~$60M. Uniswap asked for ~$70M. MakerDAO/Sky was burning ~$100M in 2025 including sub-DAOs. Labs asking for $51M in that context is not inherently unreasonable. The question is what the DAO gets in return. x.com/DefiIgnas/stat…

The "Aave Will Win" Snapshot vote is live. Aave Labs made one change: softened V3 deprecation from "should" to "could" and removed the 8-12 month migration timeline. Everything else went to vote unchanged: - Still bundled (many delegates asked to unbundle) - No accountability report - No AAVE holdings disclosure - No KPIs per product - Revenue deductions still at Labs' sole discretion Same $51M ask. NAY is leading. But ACI traced 663K voting power to wallets connected to Aave founding infrastructure. That's enough to win the vote alone. They are yet to vote.

I've had a thesis for 10+ years why AI will not take over the economy: Many economic problems are hard, so even superintelligent actors cannot solve them optimally. Examples: resource allocation; optimization, search & planning with constraints: all NP hard in many instances.

💛💙

From risk perspective the safest option is probably to put your money in a smaller and slightly customized pool! AMM attacks typically require lot of liquidity manipulation, so the largest pools get attacked first. Any custom logic disincentivizes too. x.com/fiddyresearch/…

Using "--" instead of dashes is the new way to signal "this text is human written"? Feels natural for Latex users.