Transcript of the Ampleforth office hours on 09.02.2022

7 min readFeb 10, 2022


The meter passport bridge contract had a bug so the operators have paused the bridge until it’s resolved. It was isolated to some assets but did not affect AMPL (as we just used the bridge as a generic data transfer mechanism). We are working with them to get it up soon. Will give an update once it’s live again

Question 1:
When Balancer boosted pools?

ATM we are waiting on them to review and merge these 2 PRs. Once that’s done we’ll then have to work with their governance for their system use the new relayer.

Question 2:
Will there be any more bridges added so that there isn’t a single point of failure?

Bridges were always designed to be an open platform, meaning that anyone can bridge AMPL in a permissionless way. The best way to think of AMPL on the end of bridge X is that it is “AMPL secured by X”.

So the bridge and the L1 platform determine the security model of that bridged token.

There can be multiple flavors on AMPL on, say, Avalance, just like there are multiple flavors of BTC on Ethereum (WBTC, RenBTC, etc.).

L2s are special in the sense that the main value proposition is that it inherits the security of the underlying L1. So AMPL on L2s make the most sense if it uses that L2’s enter/exit methods. It’s likely that you’ll only ever need one flavor of AMPL on any particular ETH L2.

Yes I think we should pursue more L2 integrations which don’t have bridge risk.

Side note: We have an existing implementation for Arbitrum bridging, which is functionally complete but we’ve put a hold on it as we’ve been rethinking how we bridge AMPL over to different chains / l2s. I’m currently leaning toward using wAMPL as the bridging asset, as it can be supported by all native ERC-20 bridges without any changes or additional integration. Users can then “un-wrap” into AMPL on the other side (which gets the new supply over the bridge).

This unlocks wAMPL liquidity in every chain or L2 through any ERC-20 bridge. Feels like a better strategy since a lot more CEXs are directly integrating with L2s (so we can just benefit from all of that).

Question 3:
The newly proposed AIP-5 parameters (forum post) seem to make the rebase a lot more linear compared to the old settings (on the AIP website) in important price range. Could you elaborate on this change?

The idea is first introducing the Sigmoid function with parameters that makes it close to the current linear function. And in later discussions and proposals configure parameters to achieve better performance. As opposed to bundling the changes in one go.

Follow-up question:
The new settings will be a lot slower in adjusting the supply starting from 4% off target. Couldn’t this be more dangerous while having a small market cap than the overjustment in the smaller range, i.e. the old AIP 5 settings?

It may take more days but it still significant enough to catch up with demand. While giving time to distinguish between short bursts of high demand vs sustainable increases in demand.

Follow-up question 2:
So basically if rebase is capped, the thinking is that people will continue paying above peg prices for it if the demand is legitimate. This will make the positive rebases last longer if indeed legitimate. Do we thinking this will make the negative rebase periods even longer than or shorter?

In this mild version it may not have significant effect to make it shorter but at least shouldn’t make it longer.

Question 4:
How will the X-chain bridge situation affect geyser rewards, if at all? In the future, if the bridge is turned off — how would it affect borrowing/lending?

If a bridge goes down, meaning there are no longer any messages going across, the AMPL on the satellite chains continue to operate as normal tokens. They just can’t be sent back to the base chain, and they don’t rebase.

This is a fairly graceful degradation mode. When the next rebase does go across the bridge it immediately gets up to date with the base chain.

Follow-up question:
So instead of gradually rebasing, it’s just gonna slam back to target supply?

Yep exactly.

Question 5:
In light of the serious issue with Meter, are the devs looking into Chainlink’s CCIP as a potential future avenue for building an AMPL bridge across multiple blockchains?

I’d say the bridge landscape is still very nascent. We’re looking at them closely for sure.

Question 6:
I’m assuming that you’ve read Manny’s latest essay. How would you say Ampleforth fits into it?

That’s a good question, loved that essay! Manny classifies crypto into 1) deflationary assets and 2) DeFi. He compares Bitcoin to deflationary assets in the 70’s and DeFi to equities. He notes that equities bounced back much more quickly.

My sense is that 1) Powell isn’t quite Volker and 2) It’s not obvious that Bitcoin will trade like a deflationary asset.

Staring down markets and hiking interest rates aggressively isn’t for the feint of heart, especially not in the face of rapid information proliferation and political pressures. We saw interest rates hike up to like ~20% by the end of Volker’s war and massive market contraction along the way. I think the Fed will have to hike far more gradually in this case and would be surprised if they felt comfortable getting anywhere near 1979–83 levels.

Lastly, I think Bitcoin and other cryptocurrencies could end up following equity trading patterns. Many treat Bitcoin as a growth / risk-asset vs a hedge-against-inflation / deflationary asset etc (at least today).

With regards to AMPL, I think of it as more similar to a commodity-money (though not a deflationary one). That being the case, we are far more integrated with DeFi platforms and would imagine that, if anything, AMPL would tend to follow DeFi patterns.

Question 7:
It was mentioned here before that BenQi demands a certain liquidity condition to be met for AMPL on Avalanche before they will list AMPL can you give a rough description what this condition entails or how it can be met?

BenQi has not communicated any liquidity requirements to us so far 🤷‍♂️. More liquidity is obviously better though.

Question 8:
Does bridge going down mean that AMPL on Avalanche and Binance Smart Chain are not rebasing right now?

Yes they are 1 supply change behind Ethereum at the moment.

Question 9:
One interesting thing about the paused rebases on side chains at the moment is that AMPL price on those chains has diverged from on Ethereum — eg. it’s about 1.04 on Ethereum but 1.20 on Avalanche.

What are the team’s thoughts about how oracles should report on AMPL price under these circumstances?

If there’s a deviation between supplies on the chains, it’s natural to expect AMPL to trade at different prices on each platform.

This could be because changes in demand are unable to be arbed by movements across the bridge. Or in the case where demand is steady, the different supply values could result in a different equilibrium price.

Our oracle does not report in price data from satellite chains, and I think that’s the best structure since it minimizes exposure of the core protocol to particular bridges (Naguib is confirming this right now).

Follow.up questions:
1. Do Chainlink and Tellor oracles follow that rule at the moment? I’d expect them to be sourcing price data from all (or at least, many) markets, multichain DEX and CEX?
2. In a future where AMPL exists across many chains, will Ethereum have sufficient market activity to accurately report a price that reflects the AMPL economy?

Naguib is confirming #1

For #2, hard to say but I’m not so worried about that because it’s a solvable problem if we get there. The multichain/crosschain space is very difficult to predict… perhaps arb makes it a non issue.

If we had to, we could have confidence values attached to particular bridge/chains, but we’d prefer to not have to do that.

Question 10:
(Regarding AIP-5) If we are headed into the direction of constant parameter adjustment designs for the endless future I’d rather just stick with linear.

I don’t see anyone advocating for constant parameter tuning.

For AIP-5, it makes the most sense to decouple the question of “Is sigmoid a good structure we can move forward with?” from “What rebase response speed do we want?”

I see it more like an expected two-phased rollout more than anything else.

Smaller adjustments are likely to be safer and less contentious than one big change and hence starting with the minimal change.

Follow-up thoughts:
A linear curve can accommodate unlimited swings in demand proportionally. Especially in the early years of Ampleforth’s life changes in demand will be stronger than later on when the market cap is large and inert. Then a Sigmoid function may be more appropriate when rebases of a few percent are to be expected, not when by natural growth in demand rebase may outpace 10%

An unbounded curve with high increase rate as cubic is too dangerous. No bound or high increase rate can be OK but a rebase function with both would cause undesirable volatility.

Link to the start of the office hours on Discord