"Amulets" might be a better word than "tokens" for some of the crypto assets emerging on blockchains today.
Version 2.0 of Uniswap is now live and it is, among other things, an amulet-minting machine, though probably everyone's still going to call them "tokens." Tokens are little more than keys that allow you to access something, such as a game or a laundry machine. Amulets have surprises inside; they can generate something new from within, for anyone with the skill to unlock them.
"Version 1 was almost this like proof-of-concept," Hayden Adams, Uniswap's founder, told CoinDesk. "It was the first implementation of this protocol. It got a lot of things very right. And that's proved by usage and traction."
Uniswap is a system on Ethereum for trading any ERC-20 token for any other, using Ethereum's core cryptocurrency, ETH, as the medium of trade. It currently has $43 million of liquidity (assets locked into the protocol) with $13 million in activity on the platform in the last 24 hours, according to Uniswap.info. It's one of the key so-called "financial primitives" powering decentralized finance (DeFi).
But Adams believes the platform can be pushed a lot further. Uniswap v2 – which he described in a blog post in March – will work the same way as Uniswap v1, but it will take off its key limitation – running all token swaps through ETH. Plus it will add some new features, such as a new oracle system, that many other DeFi projects may find useful.
The headline feature of the new Uniswap will be the ability for anyone to create any token pair they want, so long as it exists on Ethereum.
This is likely to lead to unexpected use cases, but the most obvious use out of the box is a new way to use stablecoins. Popular ERC-20s like MKR, ZRX, WBTC, OMG and others are very likely to be quickly paired up with stablecoins, such as USDT, USDC and dai.
Coinbase has already added liquidity to the ETH/USDC pair on Uniswap v1, after all. "Having stablecoin pairs on Uniswap is a pretty huge improvement. It's probably the most requested thing since Uniswap launched," Adams said.
The top-line advantage of having a direct pair of any two tokens is it should lower transaction fees for trades (v1 always requires two trades between any two ERC-20s). Another advantage: It allows liquidity providers to lower their volatility on one side of a pool. A MKR/ETH pool is volatile on both ends, but a MKR/USDT pool should only be volatile on the MKR side. "A lot of people don't want the ETH exposure," Adams said.
As mentioned above, these token pairs are created by users. If someone thinks WCK (wrapped CryptoKitties) and SOCKS (actual socks) should have a token pair, then they can set it up. Users pay a tiny fee to the liquidity pool for each trade (0.3%) and that ultimately goes back to liquidity providers (each gets it proportional to how much of the pool they have posted).
So returning to our amulet thesis: Any time someone contributes to a Uniswap pool, they get a token whose value grows as people use that pool. It's a token that represents value in two other tokens. It's like owning a deed to a piece of land that's growing beneath your feet.
Adams said he expects these liquidity-pool tokens to show up in other DeFi applications soon.
Uniswap is also adding an oracle service, because the first job of any magician is to learn to see the unseen.
Uniswap isn't going to use these oracles but it's a way of improving the price signals the protocol already yields.
"Uniswap basically is a price discovery mechanism at its core," Adams said.
While the company behind the protocol has never recommended Uniswap as a price oracle, projects do use it that way. Uniswap does a good job tracking the market; however, if a project needed the price at every block to really be on point, it's not safe right now.
"There is this huge demand for oracles, and it's a very valuable thing to have an on-chain price feed, especially a decentralized one," Adams said.
Uniswap is adding a time-weighted oracle service (time-weighting makes shenanigans expensive) that can flexibly deliver average prices over any length of time. Adams expects many oracle-reliant projects will at least incorporate this feed into their data streams, if not relying on it entirely.
"This is a service that Uniswap v2 provides to the world, and I think will indirectly benefit the Uniswap protocol," Adams said.
Uniswap is also adding flash loan functionality. Flash loans have been controversial but they really enable composability. Additionally, they have the benefit of only going through if everything works.
Flash loans permit a user to borrow any amount up to the total liquidity available, so long as the whole sum gets returned in the same transaction. A talented software developer can code up a bot to watch specific places in the market where prices get out of sync and take advantage of arbitrage opportunities, earning quick profit but also helping to restore equilibrium. With flash loans, traders don't even have to front the cost of running those trades.
Flash loans also allow Uniswap users to make fewer transactions, thereby reducing fees. However, by creating a whole new use case with lower upfront costs, it should increase the number of fees that liquidity providers earn.
Again, assets that can leave and return in effectively the same moment but somehow someone out there got richer along the way? Amulets.
Uniswap makes no income off its protocol. All the revenue goes back to liquidity providers. But it's a VC-backed project, with a seed round in April from Paradigm, so one would expect that there must be an intention to exit.
Uniswap v2 leaves an option open to redirect 0.05% of any trade to pay the protocol itself. Adams declined to say more about this than what he wrote in his original blog post: "The best version of Uniswap will be one that autonomously incentivizes contributions to its own growth and development as well as to the broader ecosystem in which it exists."
"I'm a huge fan of automation first and governance absolutely last," he said in an interview. "It's easier to rely on and trust math."
Uniswap has had an impressive run. Roughly 18 months in, it entered basically the same space as Bancor, a token-swapping project that raised $150 million in an initial coin offering (ICO). Uniswap simplified the design by using ETH instead of a new ERC-20. When we last made the comparison in February 2019, Uniswap and Bancor were closely matched. Today, DappRadar shows $213,000 in volume for Bancor over the last seven days and $38 million for Uniswap.
"Uniswap is doing $2 billion in trading, annualized, over the last three months," Adams said, meaning it's running $6 million in fees per year.
The genie is very much out of the bottle, and now it's making moves.