(noun40 voting via agora)

I personally don't think the new attack vector raised (keep accumulating nouns with "no risk" post 20% threshold while blocking spending) is viable attack vector. my reasoning below.

the threshold to capture the dao (theoretically 51% but in practice likely closer to 40-50% since there's likely votes you can never galvanize in any circumstance) is unchanged.

let's assume the arber has accumulated the needed 20% to fork but decides to continue. let's look at two scenarios of what they could now do.

scenario 1: they try to block spending as much as they can with their voting weight as argued.

in this case they face immense risk. they have to actively use their tokens to vote to block props so their actions become visible and identifiable. if they are unsuccessful in blocking spending and important props are still getting through (which was our case this time around; as much as sentiment is struggling, the dao still passed a ton of props this year including the last two months) then the dao might just let it slide. but if it truly becomes successful at meaningfully interrupting substantially all funding activity, then the flood gates of defense mechanisms open up.

we could vote to continuously send excess eth to a time locked smart contract to erase the arb profit as charlie suggested in his blog post. we could attempt mutual destruction buy purchasing a high priced nft or even choosing to burn the eth. we could even do something more akin to "slashing" by identifying the portion of the dao that has substantially voted against all recent props and passing a prop to send the entire treasury eth to a fork dao and assigning ownership of all the identified attacker tokens IDs in the fork dao to the treasury (deleting attacker ownership). that would mean not only the arb profit gone but the entire principal involved in the attack to be zeroed and actually just become a donation to the dao. the world would also not be very confused in figuring out what the legitimate dao is in that case so we'll just continue onward with our provenance.

the moral of the story is that you're absolutely not safe until you actually completely capture the dao. even if you're 1% less than that target the super frustrated dao could totally retaliate and you don't have a way of stopping spending without frustrating the dao.

scenario 2: you silently continue to accumulate without any attempt to block or reduce spending.

in this case you're arb profits are still absolutely at risk. any prop that spends eth is reducing the book value of all nouns thus reducing your per noun arb profit (percentage return on capital invested in the arb).

one could argue that well, while they lose on percentage returns, they could still be interested in increasing their absolute eth profit and continue for that reason. well note that as your position grows, the portion of dao treasury expenses that you're financing (in reduced total arb profit) also grows linearly.

let's actually just do some rough back of napkin style math. this year our monthly eth spend has been roughly the same as the monthly auction revenue (treasury eth amount has been roughly flat; source: https://dune.com/hildobby/nouns). for easy round numbers let's say nouns are sold every day at 30 ETH and the book value is 35 ETH. our monthly spend in this case is 900 ETH (assuming monthly spend = monthly auction revenue). if you hold 20% of the supply then you're paying for 20% (180 ETH) of that 900 ETH expense. in other words every day that you continue the arb you have a chance at making 5 ETH more (gap between book value and auction price) but you're on average losing 6 ETH per day (180 ETH / 30 days).

thus it's not economically rationale to continue in both percentage of return terms AND absolute total profit terms unless you can meaningfully reduce spending (but note that spending this year didn't really stop even with the arbers trying and we analyzed in scenario #1 the added risks arbers would be facing such as slashing by further frustrating the dao).

for these reasons I don't think the new attack vector raised is really viable and doesn't change my view on v3 being the right path forward.

P.S. just one last note on the dao capture situation. this attack vector is already alive and well without the fork and I would argue our options are worse without the fork. if someone has 40%+ voting power then while they can't unilaterally pass any prop b/c of veto, they can absolutely reject every prop. we have no recourse here. they can ask for their "share" of the treasury. they could ask for more. we have no other recourse than to accept their terms or else the dao is permanently bricked. imo it's naive to assume to that in the current mode only economically irrational actors would attack the dao. it can be absolutely rational to capture the dao even without the fork if you can get a low enough cost basis since you can take the entire treasury hostage and negotiate with the dao for a profit. the only real defense here is something like the fork that limits the damages to their "share" of the treasury and ultimately not having auction prices dip too far below the book value (which the fork mechanism contributes to as well). in a way there's a viewpoint one can have that creating market competition amongst arbers through something like the fork mechanism ensures that the cost of such a capture attack remains reasonably high at all times.

this proposal passed already, the cancellation based only on the last-minute opinion of one individual despite the 214 votes in favor was very unfair.

anyway, the "crypto-economic issue" is something so remote and theorical and that needs so many unpredictable factors to be perfectly aligned to make it happen (like a steep discount to bv for a very long period of time and the complete absence of other arbers etc.), that it's not an issue at all.

what are the 2 most important assets for a Dao? the treasury and the community. And BOTH THESE ARE NOT AT RISK in the case of the theorized attack, thanks to the vanilla rq feature of the forked dao.

it's false that if someone gets 20% of the supply can block the dao from spending and it's even more false that going from 20% to ~40% is "risk free", such an attack would be extremely expensive and a crazy bad r/r to even try it.

it's also insane to think that an attacker would join the forked dao just to grief it, simply because THERE ARE NO ECONOMIC INCENTIVES TO DO SO, thanks again to the vanilla rq feature.

so throwing away $200k that we paid for the audit, $900k/year for the dev team, and months of intensive discussions, just for this is very stupid especially because this can always be fixed at a later time in many ways e.g. set a reserve price for the auctions.

additional views on the topic that I suggest reading: https://twitter.com/noun40__/status/1691252232808017920, https://twitter.com/noun12LFG/status/1691180505520123904

Copying my reason from the previous prop:

Thank you Noun 40 for all the context that you've authored and shared! As a newer participant in the game of Nouns, it helped me inherit a nuanced and optimistic perspective on this proposal. I'm sure it was super helpful to those who have been embedded in the community as well. 💖🙏

I support the V3 upgrade that includes a fork, provided that future forks aren't guaranteed. If an arbitrage opportunity exists, it will be pursued. Nouns should be resilient to arbitrage, instead of being vulnerable to it or treating it adversarially. While I’m personally not a fan of arbitrage, it will always exist, and should be addressed in the DAO’s design. The best way to address this is in two ways: increasing spending and increasing uncertainty.

noun40 makes many thoughtful points here, especially on spending. As noun40 points out, spending makes arbitrage less profitable. Ensuring that Nouns does not hold a significant treasury, but instead spends its revenue relatively quickly, would ensure that arbitrage opportunities are minimal. There is no reason that Nouns should sit on a substantial treasury. It's arguably better to inject ETH into the market now, given the scarcity of funding, than to hoard it and spend reluctantly. Nouns has the opportunity to keep builders and artists in the ecosystem and to make it viable for them to participate for the long-term.

Nouns should spend more, it should spend quickly, and it should take into account that funding deployed now is more valuable to the ecosystem than funding deployed later. Given the time value of money, spending now yields the greatest impact. While it is of course important to spend well, and ensure that funds go to the right places, there are many good actors ready to absorb the $50M of liquidity that Nouns holds. If a portion of that is misused, it is still less expensive than the costs of enforcement. Enforcement around spending, or not spending at all, is very likely more costly in terms of impact than spending misuse. If $49M goes to good actors and $1M is wasted, that is better than barely spending at all. Cherry picked examples of waste will always be touted on Twitter, but the alternatives are worse.

In addition to spending, Nouns has another mechanism that it should leverage: uncertainty. Right now, it’s possible to estimate how many days Nouns has until arbitragers have a majority of votes. (There will be ambiguity here from groups that don’t vote, but it’s relatively easy to estimate a range here.) If Nouns are below “book value”, then arbitrage is nearly guaranteed. Nouns can raise these costs, however, by adding uncertainty. As I write this, the odds of a fork are very close to guaranteed. If someone is buying Nouns for an arbitrage opportunity and there is a 90+% chance of a fork, arbitrage looks very appealing. By reducing the certainty of a fork, we can raise the costs of arbitrage.

Nouns has an excellent mechanism for injecting uncertainty: the Nouns Foundation veto. The veto is relatively open-ended, and specifically mentions the possibility of a veto for using the Nouns treasury for personal gain. Announcing that the veto will be considered for future forks, without committing to a particular path, will reduce the expected value of any arbitrage. This can be dialed in based on messaging — stronger statements will hold more sway, as well statements with greater specificity. We can inject uncertainty without resorting to more desperate measures, like burning or slashing. Simply announcing the possibility of a veto and letting the arbitragers incorporate that uncertainty will reduce the expected value of the arbitrage.

Nouns should spend more, but it should also ensure that future forks are seen as far from guaranteed. This is a path to a DAO focused on public goods that does not reduce the flexibility that Nouns has or commit to a particular path. I am in favor of the V3 upgrade + the fork, but I encourage Nouns (in particular, the Nouns Foundation) to become more aggressive about arbitrage and reduce its expected value. That will ensure that arbitrage is not guaranteed and make it a less appealing target.

P.S. A brief note on procedural objections: I understand the procedural concerns of bundling multiple proposals together, but with the many cancelations of prior V3 upgrade proposals, there was ample opportunity to split these out into separate votes before now. I also understand the concerns about the preamble, which I am not a fan of, but the transaction data is what matters. The preamble is less relevant here, where all of the work to date on V3 speaks for itself.