Opening up listing of assets on Tornado

With reference to this post on meta-protocol for listing.

I think Tornado should consider significantly opening up their listing criteria for the simple reason that if they don’t, other protocols are at liberty to fork Tornado’s codebase and deploy privacy solutions themselves. Tornado deploying them would ensure they remain canonical - even if Tornado is not the one hosting the frontend.

Potential harms to listing multiple assets:

1. Dilution of anonymity sets

(a) If multiple versions of an asset - say yearn ETH, compound ETH and plain old ETH are listed, if everyone does not use the same solution, everyone loses anonymity. We can expect the market to automatically and rapidly converge to those tokens which are most used because everyone wants this privacy.

(b) This is less of an issue when comparing different assets - say YFI versus COMP - because the people who want to hold YFI and the people who want to hold COMP are distinct. Nobody is diluting anybody else’s anonymity set here, because the alternative for say a COMP holder is to just sit without any privacy.

One could argue that a lot of pools with insufficient anonymity doesn’t help much, but it’s still better than the status quo of zero privacy, as long as it doesn’t dilute anon sets as described in (a)

2. Supporting illegitimate projects

In DeFi, being permissionless does not amount to supporting illegitimate projects. For instance, Uniswap is not called a scam or given moral authority to dictate whichh tokens are or aren’t scams. Similarly I think Tornado should adopt a permissionless or near-permissionless approach, and let the market decide which solutions are best.

Tornado can choose to not be the one hosting the frontend, or else show warnings on its frontend, for unvetted assets.

3. Market may not be able to judge quality of anon sets

This is a distinct problem, which is not solved by restricting listing because projects will eventually want to fork our codebase.

Tornado could fund researchers trying to break its anonymity, this opens another can of worms, but is the only real way to objectively judge privacy afforded by various coins. Until then, letting the market judge where they think they get sufficient privacy might be good enough. We as governors of the protocol aren’t necessarily a better judge for this than individual users - both casual and professional - who may do any amount of due diligence and research before using Tornado.

Many users are anyway not looking to escape from state-level actors, they just want to break the link between deposits and withdrawals so that casual users (friends, twitter followers, etc) can’t chainalyse them. Tax evasion is the other common use case for Tornado but Tornado has a clear policy of not wanting to support this.

In short, imperfect privacy also is useful to some users and an improvement over status quo.

4. Imperfect convergence

In the case of competing assets (yearn ETH versus compound ETH), it is possible that the market converges upon the sub-optimal asset, assuming there is an objective way to determine which assets are bad assets. In this case, perhaps Tornado could run liquidity mining on whichever projects it wishes to be opinionated on. Those protocols could also contribute to liquidity mining. Liquidity mining may not significantly increase anonymity, because users who are here only for the profits will leave soon after - but there is a possibility some subset of users stay and continue to use Tornado.

Summary

Significantly open up listing of new assets on Tornado

Seems like the only purpose of listing obscure coins is to next propose AP mining on these coins because ‘if we don’t someone will fork tornado.cash’

I honestly don’t feel like ‘if we don’t do it someone might fork and do it’ is ever a valid argument and proposals should be argued on their actual merits and not speculation of a fork.

All significant DeFi projects have had attempts to fork them or compete. Being proactive is better than being reactive in my opinion.

I have listed actual merits too - happy to discuss on those grounds as well.

Awesome post @ghosts_in_the_code. Thanks for linking to the meta-protocol thread. Obviously, I fully support this initiative

This is exactly the correct attitude

Valid concern. It could be a slippery slope. Although, for the record, I would not support AP mining on too many other pools than what we already have

An alternative option might be that we allow other projects to run their own AP mining program with their own governance token as rewards instead of TORN

For example: Yearn decides they want a YFI tornado pool. The tornado community supports it (or at the very least is agnostic to it), but does not want to waste treasury TORN incentivizing the YFI pool. So instead the Yearn community votes to issue 1,000 YFI for 6 months of their own Yearn Anonymous Points (YAP) mining. Yearn deposits these 1,000 YFI into the Tornado AMM (or clone) and YAP miners bid on the deposited YFI over the next 6 months

Yeah, we’ve only recently made it to the point where the community seems to have the bandwidth to focus on this topic again. In reality, we really only have a finite amount of time before we start to feel the pressure of clones coming online. Best to get ahead of this now and draw some soft consensus around the best path forward

Even for the biggest most popular coins such as USDT and WBTC I think not enough deposits exist currently to warrant anonymity. Diluting the choice of assets will even complicate this more.

Key to this protocol is massive use. If there are only ~500 deposits of 1K USDT (like there are now), that means a 1K withdrawal has to come from one of those 500 wallets. That is a very small sample and will create correlations easily. Only ETH is looking to come close to higher usage able to warrant better anonymity.

Again; diluting the choice of assets will only move true anonymity further into the future because of this.

I agree

An alternative option might be that we allow other projects to run their own AP mining program with their own governance token as rewards instead of TORN

For example: Yearn decides they want a YFI tornado pool. The tornado community supports it (or at the very least is agnostic to it), but does not want to waste treasury TORN incentivizing the YFI pool. So instead the Yearn community votes to issue 1,000 YFI for 6 months of their own Yearn Anonymous Points (YAP) mining. Yearn deposits these 1,000 YFI into the Tornado AMM (or clone) and YAP miners bid on the deposited YFI over the next 6 months

I love this idea

Valid points. My reasoning is basically
a) some privacy is still better than no privacy for people who are holding non-competing coins (YFI versus COMP)
b) market will anyway converge rapidly for competing coins like ETH and yETH. Everybody wants privacy, which means they’re likely to use what everyone else is also using

We could restrict listing competing coins to start…

I created some prototype code for the meta-protocol proposal for the commitment pools that include a deactivatePool() function with an onlyOwner modifier for governance. Meaning, we can open it up to the market to pick new assets and then governance can intervene if we collectively feel an asset is inappropriate for whatever reason (including for technical reasons, like in the case of rebase-able tokens) or competes too closely with existing liquidity.

While this would be great and advocates decentralization I think it’s too early to implement. Anonymity in the base pools beyond ETH are still quite weak, if we were to open up listings of any asset we would be left with countless “ghost pools”.

Whitelisting assets can find the halfway point to moderate any impractical utility, we are still left with weak anonymity. Unless of course people preach to spread the AP mining across all pools, which I might say would be a horrible idea. Rewards should be restricted to WBTC, ETH, DAI, CDAI, at least with it’s current allocation. These assets bare the most utility and usage across the network and therefore should be prioritized.

While opening listing of any asset on tornado would allow individuals to engage in more privacy advocating techniques, it doesn’t particularly mean it would provide privacy nor would their be an incentive to pool ones assets there.

If you haven’t already read it, you might find the proposal on “commitment pools” to be a good compromise. It would allow us to measure the demand of new assets without prematurely ending up with “ghost pools”