DAPPCON 2025: Alan Scott - Product-market-fit in onchain Privacy

Transcript

Uh, I guess before I get started, just a quick show of hands. Who's like heard of rail gun? Okay, cool. Because it was going to be really ironic if I was talking about product market fit and no one no one knew what rail gun was. Um, yeah, so I'm Alan. Uh, I'm one of the core contributors to the rail gun privacy project. Uh, I've been doing this for about five years. My background is actually in finance. Uh, very specifically offshore finance. So you guys have heard about rich people sort of hiding money offshore and Cayman entities and the Turks and K coast and stuff like that. I did that for about seven years before I got into the onchain privacy stuff.

Um I guess like before I get started, you know, I want to talk a little bit about like why rail gun sort of made like product market fit for me and caused me to quit my job that I've been working for uh about seven years. Um I started out in crypto pretty similarly to a lot of people um you know speculating. Um it was really after a um a doxing event in a telegram group. Uh a guy showed me my own wallet on Ether Scan. Um who here knows like about like token gated communities on Telegram like collab bot and stuff like that. Yeah. So I was in a small one of those had the token talked about a bunch of different positions that I had held and sold and things like this and he walked me through how you can do onchain analytics through something as simple as Ether

Scan. And I sort of opened up my mind to um it's not private. Pseudonymity isn't enough. Um and so I started looking at a lot of different things. Uh naturally tornado cache was the thing back in the day. Uh but it sort of lacked uh a lot of the uh various things that I felt like I wanted in um uh a privacy solution. And then there was a white paper that had come out called uh uh basically rail gun. And uh the the supposition of the whole paper was that you don't have to do a layer 2 or a side chain or try to lure people away from the watering hole that they're in already to to get private DeFi. You could do it all within the rules of consensus on Ethereum. Uh and I thought

that was really really cool. Um and so I'll try to like very anecdotally talk about product market fit. U none of this is gospel. Uh I'm a bit of an idiot. So, uh, but hopefully this will this will, uh, resonate with some people. Um, my biggest pet peeve when people talk about product market fit is they they talk about it as an instate. Uh, you know, how do you plan to reach product market fit? Especially when you're talking to people who are like VCs and things like this. Um, and I think that this is sort of a a systemic issue of solutionism in the world of tech where people will build something and then try to find the the the nails for the hammer that they have made. Uh but I I would argue and and present to you guys today that it's

actually a mirror, right? Uh it's a reflection of what you're doing at that given time. Uh and that can sort of um you know that can that can change over time. Um and I'll try to talk about rail gun. Uh maybe look in the mirror a little bit. Um so a lot of projects use the same metrics uh in crypto, right? TVL creasing, uh, Discord, Twitter, uh, people liking and retweeting, following a project or something like this, uh, a token price mooning, uh, maybe some partnerships and integrations being announced. Um, and I think that these are really, really cool things, but I think one of the uh, uh, one of the things that you have to be wary of is that bullish metrics can actually be a um, uh, a net negative and sort of a

false signal. Um, you know, Rail Gun TVL has gone up and to the right, uh, which is really cool. Uh, the Discord and Telegram, uh, have people in it talking about stuff and, uh, 20, however many followers on a a community, Rail Gun, Twitter, it's pretty cool. Um, but, uh, the reality is is a lot of this stuff, it takes a long ass time. It's it's a long long time. In fact, rail gun when it was deployed uh, 300, let's see, 3 years and 45 days ago. I took that screenshot yesterday. um it was not great. In fact, after about six months of being deployed, the TVL was a whopping $10,000.

Um and so, um I think that, you know, sometimes we can sort of try to stimulate these uh these growth metrics in the wrong way. Uh and this is a bit of a bad capital, uh allocation. Um and so I want to talk a little bit about, you know, you know, exactly what product market fit is, at least to me. Um, we'll talk about some of the clap traps. Um, so, uh, crypto's full of miragages, right? We have, uh, these really, um, dastardly inflated incentive programs. I think Kerala Luna was a great example of this.

Uh, you know, they were offering yields of 20% in a market that was bearing five. Um, which people didn't see as a red flag, but we can let somebody else talk about that at a different presentation. Uh, now we've we've changed that to points. Uh, people are points farming instead of airdrop farming. Uh there's all these different KO deals. Uh and uh I would even argue that sometimes like uh ecosystem funds are uh common clap traps. We'll see these uh venture capital backed L2s come in with a lot of money, lure people to build. And um just a quick show of hands, who's developers in here?

Sick. Hell yeah. My crowd. Um if somebody offered you a couple hundred,000 to build some dumb on an L2, like you'd probably do it, right? Yeah. And so like this is actually a really this is like a a false narrative in my opinion, right? Does the world actually need your your L2, right? Is that a product market fit? Maybe, maybe not right? Um and so you really have to ask yourself, are people farming your project or like is it actual user growth, right? Um is it solving a real world problem that if you turned all that stuff off, people would use it?

Like would they still be using it? Um, and so let's talk about identifying these people, right? So, um, Rail Gun is a suite of smart contracts, right? It's a it's an SDK, uh, and it's a bunch of smart contracts that allow you to do cool DeFi stuff, right? Uh, so primarily the the the product is for builders, right? So, I'll talk a little bit about that uh market at least from my hot take, right? So, as we can see here, uh, it's not measurable in the same way, right? Right? You want to look at integrations, uh developer attention, what are the repos look like? Uh do you guys have issues on the GitHub? Are people doing pull requests? Uh what are the integrations looking like? Uh and then one of my favorite things that we've done over the years is go to

hackathons and see how hard rail gun is to build with. Um and I think that that's one of the biggest criticisms I have of the project. It's super hard to build with. Uh rail gun SDK is over half a million lines of code. It's all intertangled. uh and the docs are pretty Um and so this is something that we sort of take and you know look in the mirror and say where where where are the problems with rail gun, right? Um and so I I guess you know sort of pointing back to the the the points I was making earlier about uh spending money on points farming and uh paying developers, you really have to be careful uh that you don't misallocate too much capital to that kind of stuff, right? Um, and a great way to see if you have product market fit is to actually

turn all that off uh, or even better yet, make people pay for it. Uh, Rail Gun charges uh, 25 basis points uh, on transactions. Uh, and in spite of that, it still does uh, a quarter of a billion dollars in transaction volume every month. Uh, so it people really want to use it at least on the end user side. Um, I think another thing that I've seen at least in the the crypto development world is that uh we have a hard time reconciling that people have to pay for stuff. Um, we have a lot of developers in here. If anybody wants to raise their hand uh that they would work for free on stuff, right? Sick. Hell yeah. I'll talk to you guys later. We got some cool stuff that I want you to do. Um, so the reality is is most people uh need to get paid,

right? they have uh families uh worldly needs uh they have to eat these sort of things right uh going to hackathons uh conference participation me coming here all the way from Japan uh was not free it cost me money to get here uh one of the things that actually cost a shitload of money that surprised me when I got into this field was audits uh and security um rail gun has gone through I think 18 different audits uh and they range from $10,000 to a quarter of a million dollars for an audit uh and those things can be really really really expensive. Bug bounties uh as well. Uh if you want to actually lure in white hat hackers and not black hats, you have to have a pretty stop bug bounty program for people to come and report bugs to

you. Otherwise, they'll just yink the cash instead. Uh legal, right? Lawyers cost tons of money. Um if the whole crypto thing doesn't work out, I'm probably going to become a lawyer. Um hosting, tooling, marketing, you guys know. Um and the reality is is we have to ask ourselves who the hell is going to pay for it, right? uh you have the the cool things that people do initially which are you know venture capital grants ecosystem funds public donations um you know some of the things that you could do during ongoing development or fees um like rail gun subscription models uh I believe NIM uh has a subscription model on their VPN uh there's uh various premium services that exist out there in the world uh one of the things that I really want to drive home and I hope people appreciate is

that we can't really rely on the benevolence of um you know early ETH adopters forever. We have to sort of like, you know, grasp, you know, that reality. And so, Rail Gun sort of did the the fee model, right? Uh, and what it does is it takes 25 basis points of every transaction uh that goes uh in the um the vault logic. Um, and so I guess one of the interesting things and one of the challenges for that is that it takes uh a myriad of tokens. And this is really cool because it creates treasury stability, but also creates a lot of volatility. As you can see, ETH is a very popular token for rail gun and uh ETH did really great until it didn't do very great. Uh and then it didn't it it

did really bad, right? Uh and then it did really good. And so you have like a lot of this volatility. And so if you're going to be doing fees uh in this way, I guess a a a two cents from someone who's sort of participated in an ecosystem like that is sort of be aware of the volatility. Uh the runway you think you have isn't actually there always. Um, and so one of the things, uh, I guess I'll I'll leave with, right, is that probably, you know, PMF isn't static. Uh, it changes over time, and I think that's one of the biggest challenges in crypto because people sort of want to do things in an immutable way. Um, and that actually isn't um true of the market around you. It's not immutable. In fact, it's uh very nebulous and changes over time. Uh and so making sure that you

architect uh for immutability and long-term development is something that you really have to balance. Uh and this is something that we work on uh as contributors to rail gun. Uh you know, we did a beta contract. Uh we're on a v2 now. Um and the architecture the way it exists now uh isn't actually very um uh fit for the way the market exists today. It's made to do DeFi interactions, right? So you can do swaps. It's done hundreds of million dollars in swaps. you can provide LP to unis swap and do all sorts of cool stuff, right? But um whenever this stuff got scoped in 2021, um DeFi wasn't what it is today, right?

There wasn't uh limit orders. Cal swap wasn't a thing in 21. Um and so how do we deal with that kind of stuff uh has really been the crux of research for V3 uh for us. And so one of the things that we really got to be aware of is not only how are we going to pay for it, but also sort of plan for the uh I guess the expectation that ideals in the market um change and what people want and need in the market changes pretty drastically as well. Um and so yeah, that's really it. Um I really wanted to kind of end it and hope for people having questions if you had any.

One check check Mike. Yeah, one question for me from my side is like how uh rail gun will make money in let's say five years. So um you know it's really kind of interesting. I think that like you know it'll probably continue down the the the fee model, right? Um I hope that maybe eventually you know longterm the fees won't be as high, you know? So, we're researching some things with Nosis Guild on like uh sub accounts that can be held like from the private uh wallet uh outside of the vault. Um and that should alleviate a lot of the uh the cycles and round trips that you make to do DeFi interactions. Uh but I think that that'll be something that'll probably continue into the future. Uh I think

it's really underappreciated if you're doing something. Rail gun's ultimately sort of a a financial product if you will, right? It's made for wallets that people hold money in. Um and fees have worked for a very very long time. Um, you know, if the Earth, Moon, Sun, and Stars all align and rail gun sticks to landing on perfect eventually, maybe the Dow will vote to kill the fees. But, um, I think that that's probably TBD. Check. Check. More questions to Alan. Yeah please. Hi, thanks so much for your talk. Um, product market fit is something like it's um a little bit of a mystery. You can't plan it. you can't build product market fit.

Can you share the moment when you realized we have product market fit now? So, how did you go that's a question? Um, I think like the first like two years was like hm maybe this isn't great, right? Because like uh the beta contracts came out, right? And gas in 21 was really really high. high. I mean, I think a rail gun transaction was like 500 bucks. Uh, which no one's going to pay. Uh, it was kind of ridiculous. Uh, and so that sort of like spurred it to go out into uh B&B chain and then Polygon. Um, and then the the V1 contracts came out and I think we really sort of got like our first taste of it

whenever we had to do a second ceremony for a rail gun. So it started out originally with eight circuits uh for the contracts and then we realized that that really wasn't enough uh for the balances that people were maintaining within the contracts and so we expanded that out to 54 circuits and now uh 57 if we include the private proof of innocent stuff. Um, but I think when it really sort of became abundantly clear is when uh like the news started talking about like people like Vitalic using Rail Gun and we could start to see um people making deposits from places like Coinbase Prime uh and then you know unshielding to Coinbase Prime uh and seeing a lot of I guess like interesting users come to

the the system. Um, but yeah, I mean having people like Vitalic say like Real Gun's cool is like Defi Jesus turning water into your brand of wine. So that was probably uh a good one for us. This kind of like a followup to your question, but is it do you find it possible to do like direct user research with your users given the specifics of what you're building where they probably are kind of sensitive to deanmization, but you can observe their behavior to understand like what you could do better. But do you get to talk to them directly? So yeah, we definitely don't do any sort of like tracking. That's sort of antithetical to the project. Um, and so it really relies on a lot of people coming and like raising their

hands, which as you've alluded to is pretty few and far between. Most people don't say, "Hi, I'm going to try to use this privacy project to do this very specific thing, right?" And dox themselves. Uh, and so we go to hackathons a lot for that, right? Our our major product is builders, right? So people building things like wallets and integrations with with rail gun is what we're after. And so the acid test for us has really been um you know go to a hackathon and see how hard rail gun is to build with. Um and sort of reflecting on that it was it's a huge challenge. Like rail gun's an absolute bear. Uh the SDK is huge. Uh it's very very complex.

It's all intertwined. Um and um over the first three years of building it um the community sort of geared it towards wallets which sort of put arms length uh against like other people who are trying to build other things. So like for example if you're going to do um I don't know an accounting software for your private address on rail gun uh you probably don't really need a lot of the gas estimation stuff. You don't need the um uh the zk stuff. you don't need account um generation. You don't need to know how to make a zero ZK address, for example. Uh the only thing you probably need is balance scanning. And that's uh a couple of thousand lines of code, not you know the half a million that the SDK

is that you have to set up. And so um the feedback for us has really been your SDK super hard, your docs suck. And so that's sort of been the last six months of brain damage for people uh is you know peeling that apart and uh putting it. If you guys want to check it out, by the way, all the devs here uh it's in GitHub. It's called rail gun reloaded. So rail gun- reloaded. You can see the the work that's going on there. So that's sort of where we get the feedback, but end user feedback is definitely few and far between for privacy stuff. I would say one last here. Alan Alan here, here here. And I would like to have a certain connection between privacy projects that reached PMF and the ones who are on the way. And Antonio is here from uh fluid key.

Hi Alan. Uh thanks for the nice talk. uh around product microsoft fit we are in the phase of we are trying to find our product market fit we're not profitable yet no and uh what uh for you has been like a moment or or something that uh help you guiding because you know sometimes you feel lost in this that you have a lot of options and you have to figure out what to choose for and what are the best so how can you suggest what's the best way you know to find one uh I mean I guess you kind of assume that there's some people out there like yourself right I mean you can use your own indiv individual needs as like a bit of a a compass if you will. Um, you know, so whenever I started in crypto, you know, I was very dumb many years ago

and thought it was private and then realized it wasn't. And then the stuff that existed out there for DeFi privacy was really quite limited. Uh, you know, at the time, Tornado Cash was fixed deposits of ETH and only ETH. Um, and then even after that, you know, the development was sort of unalvered bridging over to XDI. Um, I don't want to get in like philosophical stuff, but suffice it to say, for me that was sort of like a security assumption that I didn't want to make, which was I have to rely on a bridge. Um, and so I felt like, you know, you just feel like people will want the same things that you want, right? And you sort of have to go out there and build it and see if people do. Um, and so, yeah, there isn't like a a magical way to do it, but you

know, more often than not, you can you can go out and talk to people and see what they think. Uh, and you're probably in a good space to do that given the amount of hands that got raised for developers in the crypto space. So, probably do it here. Thank you, Alan. Hello. Applause.