October 7, 2022: Benny
In our occasional interview series, Benny is unique in so many regards.
For starters, you won’t find any of “Based” Benny’s wit or wisdom on Twitter. The best way to invoke Benny is to ask a tough question within a DeFi Discord or Telegram groups, the kind of question that stumps all the other big brains. Benny will appear almost instantly with the correct answer and a good joke. You’re too busy laughing to feel intimidated by Benny’s brilliance.
Yet with no known Twitter handle, Benny’s presence is a privilege reserved for insiders.
We’re fond of saying that the best alfa is the alfa you make yourself. Benny is a great encapsulation of this principle. Benny doesn’t waste time with empty Twitter feuds. Benny is smart enough to comprehend the most complex concepts in cryptocurrency, and proactive enough to use this knowledge to directly influence the direction of DeFi. Be the change you want to see in this world.
Our interview shows off Benny’s talent for identifying opportunities to add value and improving protocols with these ideas. If any protocol is actually decentralized enough to permit such outsider contributions, these community-led ideas will rapidly ascend to reality. Yet even though these doors are open to everybody, few are willing to walk through them.
This interview played out over several months. Admittedly, part of me didn’t want the patter to end. For every question, Benny delivered a thorough and thoughtful essay we could study for hours. Just as you’d hope for infinite questions when consulting with the Oracle at Delphi, we couldn’t bear to close the chapter on Benny.
So many questions answered: topics like Curve, Llama Airforce, DeFi, Vyper, academia, airdrops, alfa… yet so many more questions remain. Sadly, we had to force the dialogue to a conclusion, otherwise the interview might stretch on forever. Also, we wanted to have it released roughly in time to celebrate the Llama Airforce 1 year anniversary.
Fortunately this need not be the end. Should we ever need to profit from Benny’s intellect in the future, we know we need only go to the #dev channel and ask a stupid question…. Benny will be quick to offer some sharp advice and a dank pepe emoji.
As always, interviews are paywalled for 72 hours, to give subscribers first crack at combing through the text and unpacking any alfa. If you prefer the free tier, we’ve got an archive of unpaywalled interviews to keep you entertained through the weekend:
CMC
Benny, tell us a bit about yourself!
benny
Ha, yeah sure. My first encounter with crypto was in 2016 when my kid was born and I got him some Bitcoins instead of the traditional savings account. The year after that I wanted to learn golang and a friend suggested looking at the codebase of crypto projects b/c many used the language. That's how I found out about ChainLink and that kind of opened my eyes to all the cool stuff you could potentially build on-chain with reliable oracles. when covid hit in 2020, I got a lot more time to spend on crypto, so I started doing more to contribute to DAOs I was interested in. Mostly dev stuff and mostly DeFi. Still going at it two years later. It's been a great ride with lots of learning and fantastic people all around.
CMC
So it's safe to say you wouldn't consider yourself a maximalist of any coin or token?
benny
I see DeFi as a collaborative endeavor, so no. There's more to gain from better integrations and communication between protocols than from infighting. That being said, I do spend most of my time on Curve and things related to Curve.
Monomaniac but not maximalist.
CMC
Is this something you're willing to elaborate on, your role with Curve? Are you hobbyist? Fanboy? Teammate? From my perspective you appear to be a vital part of Curve's foundation
benny
LOL yeah fanboy sounds about right.
As for what I do, I don't know if it's vital but I hope it's useful.
A couple months ago I started working on a metaregistry contract that would make it easier to find pools and pool related info. The goal was to use that as a stepping stone for an improved router and then also maybe make the job of integrators like 1inch and Paraswap easier. Fiddy's spent a lot of time improving it since and hopefully it should be out soon™️ and then we can start working on other cool things built on top of it.
Other than that I've been mostly doing analytics. Some of them are on Llama Airforce, others were ad hoc notebooks to answer specific questions. To get the data for that work, I wrote a couple of subgraphs and Curve's now using them for some of the stats on the UI. So there's a bit of pressure now when a bug or a calc error shows up.
I see analytics as a key to the protocol's short and long term success, so I'll definitely keep building in that area in the future.
CMC
What is your bread and butter? Royalties from Llama Airforce? Leveraged trading? Day job?
benny
Nice try IRS.
I still have a dayjob, yes.
CMC
Well, what could I do. They were hiring like crazy and let me carry a gun for work...
No need to dox too much, but is the day job Web2? Are you considering making a jump to Web3 anytime?
benny
No, current job doesn't have anything to do with web2.
Web3's already very much a 2nd full time job
I like my dayjob and I like to think I'm good at it, so I don't think I'd quit
But I'd definitely look to scaling back, maybe do it part time, if I ever find it hard to manage the web3 commitments timewise.
So far that hasn't been the case.
CMC
You could probably work wherever you like in Web3 given your talents -- what sort of teams or technologies would you be interested in working on?
benny
I already work in web3. I don't think traditional wage employment is the sole way to "work" in an industry. I do things that I find meaningful, spend quite a lot on time of them and sometimes receive [REDACTED] when these efforts prove useful to others.
I like to work with teams that care more about decentralization, privacy and censorship resistance than things like sanctions compliance.
As for technologies... well I don't know if it's technology per se, more at the nexus of tech, economics and social sciences, but I'd say i'm interested in how market mechanisms can be used to make the governance and management of large organizations more efficient. Being built on a tech that allows for the free and easy creation of programmable markets, DAOs are a great experimental ground for this. (and the Curve ecosystem one of the most interesting case studies). My intuition is that the current doomer's view on DAOs, that they're inefficient, that people don't care to vote or are too retarded to make the right choices mostly comes from trying to impose a political model (direct democracy) on an economic problem.
CMC
What are your thoughts on the Curve DAO relative to other DAOs in the space?
benny
Well one thing that I like is that I think it's gone further than many other DAOs in decentralizing the decision process.
Of course it doesn't mean that there isn't still a lot of work to be done.
But a lot of projects have adopted a model where they retain a traditional corporate organigram with a CEO, CTO, PMs, devs, etc...
and then you have a so-called DAO on the side with some governance theatre for token holders who get to vote on proposals from the team.
By contrast Curve has done a much better job at fostering bottom-up governance.
To give you an example, just a couple months ago ahab kept complaining that Curve was losing business to Uni because of the high fees. I wrote a quick notebook that showed that aggregators indeed almost always routed through Uni for the most popular stable pairs like DAI-USDC, but that if the fees were lowered, in particular on the 3pool, Curve could capture significant amount of volume from Uniswap, which in turn could offset the losses. I talked about it with fiddy who further discussed the issue with Charlie and Mich. Eventually nagaking ran a multipool simulation that showed that lowering the 3pool fees would also increase overall revenue by making metapools more efficient. He posted a very detailed proposal that passed almost unanimously.
I thought the whole thing was pretty cool, going from random people talking on Discord to a rigorously tested proposal on something critical to protocol revenue with hardly any intervention from the core team except for providing intelligent feedback along the way.
And there are other examples of course with the risk team or veFunder which were all community initiatives.
You may say that this stuff happens all the time in the open source world and that's true. But in blockchain, outside of the classic L1s like Bitcoin or Ethereum, you don't see it often. It's certainly rare at the application layer despite the larger number of projects.
The other and perhaps even more exciting thing about Curve's DAO is the way in which protocols like Curve and Convex have brought a market logic into the governance protocol.
But i've already written a wall of text so maybe i'll stop here
CMC
How decentralized do you think Curve is in practice? How many tactical bus strikes would be needed before you were in charge of the protocol, and do you think you have everything you need to keep it in operation?
benny
LOL, i don't know about bus striking your way to a protocol take over, there's enough decentralization in the original design that there's little chance of that happening. But the part about keeping everything running is not trivial for sure.
For instance, the collection of trading fees can technically be done by anyone, but the complexity of the process and the lack of incentives to cover the gas costs means that in practice the Curve team handles it. So if that were to stop, the DAO would have to scramble to find a sustainable replacement mechanism. You also have contracts like the address provider and pool registries that act as guarantees that factory contracts and non-factory pools are legit, but they're only managed by a couple of people.
I'd say that knowledge of the protocol is also still very much a centralization point. A lot of the knowledge about the protocol, its implementation details, its history etc... is in the team's heads. And currently the only way to access that knowledge is to hit them up on Discord or TG. Now Mich will even field the noobest of questions on TG and the team is very approachable so that may not seem like a problem. But in the mid to long term, it's going to be essential to ensure the perenniality and transmissibility of that knowledge base. It'll be necessary so that important operational governance decisions don't always come from the team (or a team) and for the DAO to onboard new maintainers/contributors.
So nothing outrageous, most of these can be addressed with a good incentive/market designs and better analytics/documentation
CMC
Llama Airforce is certainly stepping up to educate users about several aspects of the broader flywheel. What is your vision for where Llama Airforce is going?
benny
I don't know about educating, the risk team and llama academy do a better job than we do on that front.
But we do aim to have more quality analytics available that people can use for their research on Curve and its ecosystem.
To that end Alu and I have been working on separating the union from the Airforce website
It makes more sense to have the union be standalone, especially if we expand to other protocols or chains and it gives us more freedom to further develop the analytics part of the site, so we'll be open sourcing the front-end repo as well as the ones for the backend (some services are in C# others will be in Python).
Hopefully this helps onboard new contributor and makes it easier to interface with other people in the DAO.
CMC
You have deep exposure to academia through your non web3 life, if I'm not mistaken. How does the intellectual rigor of trad-school compare with crypto? Where are the bigger brains? Will we see traditional schools disrupted in any way by cryptocurrency?
benny
Both spaces are very diverse, so any comparison is going to be in very broad brushstrokes.
I mean in DeFi you have Curve and then you have people who think a stable swap with prices hardcoded to 1 is a revolutionary idea. Likewise in academia you've got people doing fantastic work sharing departments with complete cooks.
But one thing that's been interesting to observe is that the industry has reached a stage where there's a lot of work that crosses boundaries between academia and industry. I guess it's already been the case for a few years on L1 research, but now we're seeing it with DeFi too which is very exciting. I'm thinking for instance of the work of people like Guillermo Angeris, Tarun Chitra, Guillaume Lambert, or even the Curve Whitepapers.
I don't really see crypto disrupting education for the foreseeable future, maybe if AAVE starts doing student loans
But the state of crypto education is symptomatic of some larger issues/question with academia.
While lots of people in crypto have some sort of secondary education, the space is still run by people who are largely self-taught, in the sense that they haven't had a formal education when it comes to crypto, simply because that doesn't exist yet.
Some universities will have a general blockchain or cryptocurrencies course, but it's still really basic and you could blame universities for being out of touch or too slow to adapt, but ultimately we still have a thriving, innovative space in crypto even without that formal training path.
So on one hand the large number of autodidacts in crypto does point to the decreasing relevance of a college education. there are many other ways to learn. On the other hand, there are still a good number of people with BS or post-graduate degrees in the space who put what they studied (whether that be maths, physics, law, communication...) to use in an industry that they've never learned about in school.
So I think that also shows the limits of the argument that education should be practical or more tightly following what's done in industry. If you focus on fundamentals (again not just in CS), you still give your grads the ability to adapt to entirely different fields.
If you switch your focus to purely practical training, it's going to be hard to keep up with the pace of industry because academia is naturally slower paced and longer term oriented. so in that case not only do universities end up being subsidized employee training (with tax money or student loans, i.e. people paying to have a job), but those subsidies will favor established industries over innovative, nascent ones.
As for where the bigger brains are...
There's been a general pattern of brain drain from academia to industry for years now. across all disciplines, not just CS or STEM.
It's unsurprising since universities are increasingly ran like large corporations but with lower financial compensation and decreasing social prestige.
And that's a consequence of focusing on education (where the money is with rising tuitions) over research (where the money is harder and harder to come by).
It's unfortunate since I think universities should focus on the latter rather than the former.
So I guess for crypto specifically, if you're interested in doing research, there's probably better opportunities in the industry as opposed to say doing a PhD.
CMC
What's your most unpopular opinion about cryptocurrency?
benny
Hmm maybe that cryptocurrencies should be inflationary / inflation is good.
CMC
Blasphemer! That harms the ever-important P/E ratio! What's your case for this bold claim?
benny
Well even if L1 tokens are a different game and should probably be called something else than crypto"currency", I'd point to what happened to Ethereum post-merge. It was good that there is still an inflationary mechanism in place in these days of low gas to keep the network secure. You could make a similar argument for Curve and its ability to retain liquidity in this bear by issuing CRV.
I guess you could see inflation as kind of an operational budget for a DAO. Stakeholders don't pay directly but they agree to be collectively diluted by X% so that the money can go to critical operations like network security or liquidity incentives (or with veFunder, key elements of the Curve ecosystem like Vyper). They do this because they believe that this will allow the protocol to create a higher amount of value (reflected in price appreciation or yield) in the mid- to long-term. So it's like a decentralized way of funding an ecosystem's public goods.
So I'm not advocating for 4 digit APR, but a moderate amount of inflation that is put to productive use.
CMC
The DeFi space has been a bit quiet as TVLs have mostly stagnated in the bear market. Are there interesting trends or concepts you're seeing emerging at the moment?
benny
I wouldn't say that the space has been quiet, I can still hear the screams of all the people who got rekt this year.
But that has been interesting in its own right, the unraveling of protocols that went from the scammy to legit but more experimental. It's giving us a better sense of what ideas are worth pursuing and which are not, which protocols and designs are robust and the ones that are not.
The recent regulatory attacks, especially after the Tornado sanctions have been interesting as well, because they've pushed protocols to take strong stances and it makes it easier to distinguish the ones truly committed to building censorship-resistant decentralized solutions and the ones with more questionable ethics.
In terms of interesting trends, in DeFi specifically it's been interesting seeing all these options-based protocol show up. even if they don't succeed, they've been developing libraries to do on chain stats, complex number operations, etc... that will be useful for the space in the long term.
Related to that, my favorite non-DeFi trend's probably been the development of Vyper. Charles' doing amazing work on the language and we're seeing so much awesome tooling being released, leading to more adoption.
CMC
The options have been quite interesting -- I steer clear for now because I've seen them explode too often in real life. A lot of the protocols rely on Black-Scholes, which seems like a rough but adequate estimate, until it doesn't... But agreed on the point that the technology and learnings make the next generation really work.
To your point of Tornado Cash -- a lot of people presume a future split between DeFi and KYC-fi. I'm not so sure though, a split implies both halves are roughly equally weighted. I'm not sure KYC-fi, CBDCs, et al ever get enough adoption to get critical mass. Do you perceive a future for those with, as you put it, "more questionable ethics"?
benny
Well I may be overly pessimistic but it doesn't seem impossible to me that KYC-fi / CeDeFi will turn out to be the most popular platforms. A few have imploded this year, but CEXes are still around and integrating DeFi products on top of their other offerings. And as long as DeFi is around, there'll be startups trying to add a thin layer of friendly UI and custody over DeFi products.
I'm more concerned with the long term survival of actually decentralized DeFi, and if not giving up to surveillance and censorship means that it has to remain niche, so be it.
So yeah hopefully people learn their lesson from the recent CeDeFi shitshows and realize that transparency, self-custody and censorship resistance are essential.
Or maybe they think new regulations will solve old ills and they sign up for reversible ERC-20s with a copy of their passport and enjoy giving custody of their money to some obscure third party because they can't be bothered to do otherwise anyways.
CMC
Any alfa you can gib? If not alfa you're making, alfa in terms of where people should be looking in these times? Anything we should have asked that you'd like to mention?
benny
There's no more alfa, we're all in it for the tech now…
…but I'm very much looking forward to $crvUSD.
CMC
What did Alunara mean by this??!?
benny
He's just off his meds
Again
Disclaimers! Author has no position in $LAF