DeFi Weekly #56

Set launches Social Trading, Aave's Flash Loans, Wave (Uniswap but only for Stablecoins)

Latest This Week

  • Set Labs (formerly Set Protocol) just launched their new social trading and managed to secure over $700k on launch! If there's one team that I think most people don't pay enough attention to it's easily this one. From the conception of being "5 tokens in 1", to auto-rebalancing tokens to social trading it's clear to see the reasoning and thinking around each pivot and how it gets them close to product market fit. Despite still having some additional work to find rocket-ship traction I think the potential of Set is extremely high, mainly due to the mechanics of a Set works. In its essence, a Set is an abstraction of the underlying assets contained inside it. While those assets are locked the opportunity scope of what's possible with the underlying assets is quite high. For example, lending out the assets is one possibility, however what could more mind-blowing is implementing a delay for redeeming sets and then fine tuning parameters to allow a fractional reserve banking/superfluid collateral scheme where the same collateral locked up inside MakerDAO can be used to create a CDP. The reason why it works a lot better in Set's case is because people actually care about the abstraction, have an incentive to trade with it as well but then also may accept a delay to redeem a set. Take for example the following scenario:

    1. User locks up 2 ETH ($100) and 200 DAI ($200) to mint one XYZ Set

    2. The XYZ Set lends out 200 DAI to Compound (already possible)

    3. The remaining 2 ETH is then used to open a CDP and generate additional DAI

    4. 1 XYZ Set can now be traded as if it has the value of exactly 2 ETH + 200 DAI but the user also has additional DAI and has a leveraged long position in ETH

    A few readers are probably thinking "but can't you just use a leveraged long token" or "isn't this creating more systematic risk". My answer the first point is yes, however liquidity is further fragmented depending on the type of long token which reduces the liquidity of the set itself. My answer to the second question, not sure but it sounds pretty cool as an experiment. Another scenario I thought of while writing this was that if Set was to ever get enough traction where sets have an aggregated market cap of $100m+, then they could create their own stable coin which has the ability to mint more out of thin air. Without going into crazy detailed mechanics you could have a situation where they could run a fractional reserve banking scheme and if everyone tries to withdraw there's simply a time lock to redeem the collateral. Provided people have a strong incentive to only deal with the abstraction of the assets and not the underlying actual assets it has legs to run for a while. I'd probably describe it as Tether but there'd be algorithm to determine how much should be kept in reserves versus how much can be created out of thin air. Furthermore, if people are interested in depositing ETH to generate DAI, you could replace DAI with this new stable coin (via forking MakerDAO) and bootstrap a new lending protocol at the same time since you have real borrow demand (attracting lenders is the easy bit). I'm probably missing plenty of details but financial engineering like this is super neat to think about. Stepping back into current time, Set can break out of relying on DeFi whales to lock up more capital as retail has a solid way to grow their ETH stack beyond the nominal 1%-4% in money markets. Their best performing set’s are up over 60% YTD!

  • While flicking through CT, I came across the following: "Larger blocks are a feature, not a bug" (in relation to Aave's flash loans). For those of you who aren't aware about what they are, to put it simply you can borrow money you don't actually have as long as you return it by the end of the transaction. What use case is this good for? Well the main one is closing out a MakerDAO CDP where you have enough collateral inside your CDP but need some extra DAI (which you don't have) to close out the CDP. With Aave flash loans you can borrow ETH with no collateral, use it to buy DAI, close out your CDP, then pay back the ETH borrowed from the collateral in your CDP all in one transaction. If you can't pay back the ETH, the entire transaction reverts meaning it's impossible to steal from the pool itself. Referencing back to the start of this point, a larger block would allow you to execute more operations inside a single block which could allow you to do more for the duration of your "flash loan". Looking at it more fundamentally, a flash loan is relying on an atomic guarantee that both:

    • You are agreeing to take out a loan with no money in advanced

    • You are guaranteeing to pay back the loan at the end of the transaction

    If any of the either conditions fail, the entire transaction reverts. It'd be pretty neat imagining what if we could guarantee atomicity across multiple blocks? There might be a path with VDFs and STARKs but I'd need to think about it more. Speaking of Aave, they raised via their LEND token via an ICO ~2 years ago. The current market cap of their token is $36m, up from $12m a couple of months ago. I tried looking into what the tokenomics of Aave actually is but was surprised to realise that I actually couldn't find it. Maybe they still have enough money in the bank, but they're wasting the use of a potentially powerful shilling tool (see: SNX spartans). Apart from tokenomics, I definitely think they could improve their branding to attract more liquidity. The dominant use of the blue/purple gradient is reminiscence of low quality ICOs and doesn't deliver confidence to users that their financial assets are safe. It's a minor point but commonly overlooked by engineering heavy teams. Since most tech in crypto is experimental, investing in making sure your brand has a good outlook is just as important as the tech itself. VCs if you're reading this - please help your portfolio companies a top designer. It'll pay multiple times over in perception.

  • New project alert: Wave! TLDR: It's Uniswap but just for stable coins. Why? Help build on-chain liquidity for transferring and supporting multiple stable coin pairings. It's a great public infrastructure tool but I think precisely just that. For starters, it allows more types of stable coins to be used in DeFi without DAI or USDC having an exclusive monopoly at the moment. From a value capture perspective, a 0.3% cut for liquidity providers for a paring which is theoretically meant to have the same price is absolutely tiny. Wave was started by the current CTO of NuCypher which I never understood the detailed value proposition for apart from "encryption as a service". It seems that there's definitely a rise in 2017 operators and founders in the space realising the path to their creations being used is much further out than even they had anticipated and as a result working on more side projects. Just this week we saw Jae Kwon from Cosmos step down as CEO to work on another crypto side project. For many token based projects the incentive to continue working goes down as time goes on. Longer vesting periods are only part of the solution but don't fully help since some tokens doesn't have a chance of regaining investors confidence again especially after they've been pumped up like crazy before. The DigiX crew hit this point a few weeks ago as well after voting to dissolve the "DAO" and take home millions of dollars in ETH. If we don't see a bull run by the end of this year many more companies will fold as only the strongest will manage to survive through crypto winter. Those who do have capital to hire great talent should especially be on the lookout as the same talent will be priced much higher during the next bull run.



Ganesh Swami wrote a pretty interesting piece about DeFi data availability and MakerDAO, check it out here: https://www.covalenthq.com/blog/defi-data-availability-makerdao-csv-export/