In recent months, DeFi on alt L1s has dried up. The community believes the reason for this is a lack of tokens that people want to trade, and the rallying cry has been encouraging developers to launch more tokens, like protocol governance tokens or meme coins.
My hottest take is that L1 tokens are the only remotely investable tokens, and that DeFi protocols should almost never launch them. Protocols with governance are the worst protocols, governance-minimized protocols are better ~100% of the time. You may initially disagree for smart reasons, like concerns about contract immutability, changing risk parameters, or governing the oracle. You may also disagree for dumb reasons, such as believing protocol developers need the asymmetric upside of a token launch as an incentive to build, or that they absolutely need to raise funds from venture with token deals. Regardless, I hope to alleviate all of your concerns in this post.
Without lots of new tokens, what saves DeFi?
The best thing for a chain’s DeFi ecosystem is a strong L1 token. It’s been argued that deflationary economics is bad because it “encourages saving and discourages usage”. This is wrong, because on blockchains and in DeFi, saving IS usage. Holders of long positions on L1 tokens are some of the top users of Ethereum’s DeFi ecosystem.
Users lock their long positions on-chain, often using DeFi to get liquidity by collateralizing decentralized stablecoins, synthetics, and derivatives positions. DeFi allows them to do this without taking on centralization risk in the form of depositing their long position on an exchange or on a lending platform like Celsius, or bridge risk from bridging elsewhere to some other specialized chain. A major problem is solved, in that they can self-custody their long position while doing lots of useful and interesting things with it… a great option for holders who are long on the L1.
One could argue that DeFi is useful, even without a strong native token. It’s true you can use DeFi with wrapped and bridged assets, but without self-custody, it’s less useful than DeFi with native assets. You might be thinking “you can still self-custody bridged and wrapped assets” NO YOU FOOL! There is no such thing as a trustless cross-chain bridge- remember when SBF stole the bitcoins from Ren and Sollet? Plus, even if you could, if your goal is to be the best place to conduct finance with non-native assets, you have to compete with both CeFi, and an endless slog of new L1s and L2s, possibly with venture-subsidized incentives.
Your country’s government could ruin your value prop by getting its shit together and passing a permissive regulatory framework for financial innovation. While that day may never come, and continued bureaucratic incompetence is generally a smart bet, this possibility should make us think harder about lasting use-cases for financial disintermediation.
Some argue that even without self-custody, there are big benefits for decentralizing other parts of financial supply chain, such as execution and settlement, and that we benefit from the transparency and accountability properties, even if we can’t actually force CeFi intermediaries to play by the rules. This is definitely true, but public blockchains with decentralized validator sets and native crypto tokens are an over-engineered and unnecessarily complex means to achieve that outcome.
What makes a strong native token?
In contrast to the “ultrasound money” ETH memes celebrating high fees, aeyakovenko has brilliantly argued that a blockchain will be undercut by a competitor unless its fees are kept as low as possible. Low-fee blockchains get revenue by selling priority access to state hot spots for a premium during periods of high contention, and near-zero fees attract developer and user activity, leading to creation of state hot spots, and more revenue for the chain.
This is extremely clever, but in practice some of the most popular transaction types don’t create any state contention. The chain doesn’t make as much money as it could from those transactions, and the base fee could be considerably raised before anyone flees to a subsidized chain in pursuit of cheaper fees. The base fee should be as high as users are willing to pay, but obviously not too high so that people are priced out.
It’s smart to anticipate that many users will always flock to the cheapest option, but a blockchain without a strong native token is simply much less useful. A decentralized stablecoin with 1¢ fees for transfers is more-useful as a cryptocurrency than a centralized stablecoin with zero fee transfers, since the latter fails to unplug users from centralized financial intermediaries, the purpose it serves is unproven and speculative.
The base fee could be dynamically priced, such that the chain collects as much revenue from fees as possible, per block. Control theory should be applied to design such a pricing algorithm. A naive example (designed with ignorance of control theory) might look like this:
// if we lost money last block by pricing too many people out, decrease the fee
if lastBlockRevenueChange <= 0 && lastBlockBaseFeeChange > 0 {
currentBlockFeeChange -= 0.01
}
// if we increased profits last block by raising fees, raise ‘em again
else if lastBlockRevenueChange > 0 && lastBlockBaseFeeChange > 0 {
currentBlockFeeChange += 0.01
}
// if we decreased fees too much last block, raise ‘em
else if lastBlockRevenueChange <= 0 && lastBlockBaseFeeChange < 0 {
currentBlockFeeChange += 0.01
}
// if decreasing the fees attracted more users, run it back
else if lastBlockRevenueChange > 0 && lastBlockBaseFeeChange < 0 {
currentBlockFeeChange -= 0.01
}Secondly, if all the value accrues to stakers, not to the unstaked token, it’s harder for all that precious natively-created value to be used in DeFi. Staked tokens can be re-used in DeFi with LSTs, but LSTs have problems in their current form. Some of the most popular LSTs are closed-source, and controlled by multisigs or plutocratic “D” “A” Os. Proposals have been written to decentralize LSTs, but it’s a hard problem to solve because tokens staked with different validators have different risk profiles, and are valued accordingly.
Fun fact: if the chain is a sovereign rollup, it doesn’t need validators or staking, and all of its revenue can accrue to the unstaked token. Wow!
Don’t protocols need tokens, and create value by launching them?
On-chain governance is inherently plutocratic. Cryptocurrencies are meant to be autonomous and free-standing, and not require humans in the loop ruling over them. Governance-minimized protocols are the best. MakerDAO voted to infect itself with centralized USDC, making DAI just like a CeFi platform, while immutable LUSD, free of governance, beautifully fuses the convenience of USD with the powers of a decentralized cryptocurrency.
Some protocols like Aave have a DAO that controls important risk parameters. I don’t have an answer for this, there may be no way to avoid it which would simply be unfortunate, but automating those parameters might turn out to be possible somehow.
An often overlooked reason to oppose protocols launching tokens, is that it leads to misalignment with the L1, and creates some perverse incentives for developers. Create protocol, launch token with incentives, let insiders dump, anger the users who got dumped on, create resentment, move on to the next protocol with a different set of insiders is a tired loop and it’s destroying this space. How many lending protocols are there now? It seems like a new lending and DEX protocol launches every month. Imagine if every update to Microsoft Word was launched as a brand new word processor. It’s getting ridiculous. The true motivation for launching a new protocol is rarely to solve problems with previous protocols, rather they’re created because some devs who didn’t get to be insiders on the last dump-n-rug want to be the insiders for the next dump-n-rug.
It would be much better if all DeFi devs on an L1 could rally around one standardized DEX, lending, synthetic stablecoin, etc, and keep it fully aligned with the L1, no value accrual to unnecessary tokens which creates an incentive to fragment the ecosystem and pile on complexity and ugliness.
What about bugs? How can open source devs upgrade and improve the software it if it’s a locked smart contract?
Once the protocols are truly fully L1-aligned, they can be enshrined into the blockchain. While this may seem extreme, this solves a number of major problems and should be explored as a solution for some of DeFi’s biggest problems.
Firstly, it allows the protocols to be governed and upgraded without governance tokens. EIP-style off-chain governance is more meritocratic, more opt-in and voluntary, in contrast to token governance which is unforgiving and plutocratic.
Secondly, it allows the protocols to be resilient against code bugs in smart contracts because they get security from client diversity. Securing a smart contract against bugs is hard. Formal verification is an incredibly slow and difficult process, and normal security audits are no silver bullet. Hacks are devastating, and remain a glaring unsolved problem in crypto that attracts negative scrutiny to our space. Though it’s an extreme solution, it may be time to explore enshrined DeFi as a solution to this ongoing scourge.
How do devs make money by developing fully L1-aligned protocols?
it’s simple: they make money by going long on the L1. They could get L1 token grants from whales, who are incentivized to give grants like this to drive value to their investments. They could also raise cash from institutional L1 whales, or those institutional whales could hire devs to develop enshrined protocols, and profit by creating utility for their L1.
You’re too idealistic. Devs want asymmetric upside.
Asking “why would I build a protocol if I can’t get rich by launching a token” makes no sense. It would be nice if I had exposure to asymmetric upside from visiting my grandma, I’d probably visit her more often, but that’s not how the world works. Just because you want protocol development to be lucrative, does not mean that it is, or ever will be. A few lucky protocol devs have gotten rich from token launches in the past during irrational bubbles, occasionally by doing dishonest things, and always creating fragmentation and misalignment in the L1 ecosystem.
It’s conceivable that governance rights over risk parameters is so valuable that a protocol’s governance tokens become highly valued for valid reasons, but in general most DeFi tokens are meme coins pumped with incoherent cope narratives.
Besides governance rights, DeFi protocols don’t have scarce resources that necessitate tokenization (DePIN could be an exception). This is in contrast to L1s, whose tokens have very straightforward utility and purpose. Rally around the L1 token, share the love, get rich together.
“You’re describing a circular ponzi”
As I’ve described it so far, yes.
But don’t forget that cryptocurrencies existed and were valuable for several years, even before DeFi was conceived. They are convenient digital payments solutions, often have better privacy properties than centralized alternatives, are censorship resistant, borderless, and we love them.
Thanks for reading! Feel free to roast me on twitter!
