Transcript 0:00 [gentle music] Dara, good to see you. How you doing, my friend? Great seeing you. Cheers. Thank you for having me. Oh, man, thank you for coming. Mm-hmm. 0:17 Do you know, we're at Fintech Node Con, uh, in the Galileo Podcast booth. Uh, shout out to those guys. And when I was putting this event together, uh, you were my ideal client profile. I was... 0:30 Uh, w- I remember having a conversation where you said something of the lines like, "This needs to exist, and I will take my budget from another big conference and give it to you if you make it exist." 0:40 I- I'm glad that we're doing this on audio 'cause otherwise my blushing would be more embarrassing. [laughs] Thank you very much. I'm fantastically impressed with the event that you put together. 0:48 All of the staff have been great. All of the production have been great. This is, um, aspirational for everyone else. Thank you very much. Heck yeah. 0:55 So thank you for being involved and, and y- the reason I said you were my ideal client profile is 'cause you kinda have an interesting story. 1:02 When I was building an event for Fintech Nodes, Dara is, like, the Fintech Node. So tell me your story, like from Stripe to Increase to how you got really hooked on some of the infrastructure. 1:13 I, I run a company called Increase, and Increase is a banking API. Mm-hmm. Um, my journey to Increase, uh, uh, probably most notably came through, uh, being early on at Stripe. 1:25 Um, early on at Stripe was building on top of banking providers and seeing the data and capability that existed one layer down. 1:33 And, um, like any engineer who's building on top of a system over an instruction becomes, "I wonder what would happen if we got to run that layer a little lower," right? Mm-hmm. 1:41 Like, I wonder if there's fun product that we could build there. And that was the motivation for Increase, right? 1:46 Uh, to give a specific one, we, uh, Stripe never did disbursements to merchants, so payment to users by wire. Um, and it's not because wire systems in the US are bad or slow or anything else. 1:57 It was because, uh, the failure mode when a wire was not able to be delivered, uh, Wells Fargo would send you an email that said, "Oops, sorry, wasn't able to deliver it." 2:06 Like, uh, uh, and it's just hard to build good, reliable, uh, trustworthy product on top of, you know, IMAP and an email flow. And it's no fault of Wells Fargo, right? 2:16 Like, they built their systems for a different contemplated use case. And when you bump into the edges, like, uh, um, uh, kinda on you, right? 2:23 Like, so it's, it, it, uh, uh, wasn't, I don't blame them in any way, shape, or form. They built a fantastic institution and fantastic bank. Uh, but it was obvious that we needed something else, right? 2:32 Like, we needed some technology company to drop down a layer more. And so, uh, uh, that's what Increase is. Mm-hmm. 2:38 Increase is we run what's known as our own banking core, our own maintenance of balances and transactions, and we connect directly to the Federal Reserve for the U.S. 2:47 rails of BCH, wire, Check 21, um, uh, and expose that for our users and use cases. 2:53 I, I love this idea that those invisible layers that you only experience when you're higher up the stack are the ones that you immediately wanna go fix. As such an engineer- Yeah, yeah... 3:02 engineers always wanna get, like, a layer deeper. And I wanna get closer to the metal and closer to the metal. Am I allowed to tell a story about this, though? Yeah. I feel like this is one that you might enjoy. 3:10 Yeah, go on then. So we had, uh, uh, Dita was leading international, like, non-US rollout at, at, uh, at, at Stripe. And, uh, Dita's a good mate of mine. 3:19 And the two of us were sitting there lamenting one evening of, you know, we're rolling out country by country. You're going to, initially it was Canada, then the UK, and, and, uh, every country involved, 3:30 uh, you were having a technical conversation and a business conversation with all the financial institutions in different places, and it was just hard, right? Like you're trying to solve two problems at once. Mm-hmm. 3:39 You're trying to understand whether or not, like, uh, they wanna do business with you, and at the same time, like, assess whether or not you could build, like, you know, the pri- the thing that you- The stuff... 3:48 would be proud of- Yeah... underneath. And it was just looking... You can imagine a dejected conversation on a Thursday night after you've gotten out of five of these calls where you're like, "Ah, I'm not... 3:57 Is this possible?" Yeah. [laughs] Yeah. And so the, the, the insight there was can Stripe at a technology layer go a layer deeper, right? 4:04 Like, can Stripe connect to Visa and Mastercard as opposed to National Australia Bank? You know, that unlocked a lot of Stripe's non-US per- was like dropping down a layer. Um, I'm not a very creative person. 4:17 I'm a pretty boring person in my day-to-day life and in my work life. And in the same way, just like applying that same, uh, uh, that same insights in, in Increase, right? Like, can you just drop down there? 4:28 Can you connect directly to the Federal Reserve? Uh- Yeah... can you solve yourself some problems? But it turns out to do that you have to get all kinds of new licenses and stuff. Have you thought about that? 4:37 Especially as, like, an engineer who's like that's, that's something else. There's another team that thinks of that. 4:42 But ha- have you manifested that as an engineering problem, and have you thought about that as a founder? Y- you're correct, yes. 4:49 I take a view that this, philosophically, what I view as engineering is you have your kind of a set of constraints and understanding the natural laws of the world, and you're assembling the natural laws of the world to kind of affect the thing that you're trying to do. 5:03 And like what is a compliance regime but like a set of, like, rules and regulations and tests, that if you look at it from an engineering point of view, and if you bake it in from an engineering point of view, first of all, you don't view it as like a dismissed thing, right? 5:15 But, uh, uh, if you're able to bake in those, both the explicit rules and regulations and also the kind of like culturally received, uh, uh, parts, uh, uh, deeper in, you're able to build, first of all, like, better compliant systems, but also to, uh, allow yourself build better products. 5:32 Uh, to give a specific, it would be, 5:34 uh, there's a rule called the travel rule to do with, uh, strictly it's, uh, kind of like, uh, wires above 3K, but, uh, you can apply it down lower, and there's certain data elements that must be included. Yes. 5:45 And one option is that you go through your audits and you pull the samples of all the wires, and you ensure that it's all present. 5:52 And, like, this is a, uh, time-honored tradition, so, uh, uh, uh, who are we to dismiss it? But an alternative is that you just bake that in as a, like, test within your integration suite, right? 6:02 Like the exact same thing that you try to assure yourself that your product works as expected. If it doesn't have these data elements, it wouldn't have worked. 6:10 Um, and that, that allows me sleep better at night with respect to, uh- Yeah... building good systems, but it also allows, uh, better velocity. 6:20 I think with a, with a charter as well, people forget that if you fall foul of, uh, some of the regs as you are getting closer to the Federal Reserve, that, like, people get pulled over at airports if they get this wrong, and they start having conversations with law enforcement. 6:34 And the, the consequences can be material, so solving it well is very different to, um, many, many other things. 6:41 And solving it well inside of financial institutions is very, very difficult because when they built these systems in the '70s, the computational power to be able to build things the way you're doing it just didn't exist, and nor did none of the software engineering norms of how you do that stuff. 7:00 Memory didn't exist. Networking didn't exist in the same depth. Like, mainframes ran on megabytes, not gigabytes or petabytes. And so the marvel of what they built is, is incredible. Correct. 7:11 The correct reaction is admiration. Right. Like, the correct, uh, uh, reaction isn't to scoff, right? Yeah. Like, it's, the correct reaction is, "You did far better than I ever could," right? 7:20 I hope and aspire to, to, to achieve. 7:22 Do you know my favorite, um, line from my early days in, in engineering in a telco was the reason, um, that startups can build things so quickly, r- as, like the old parable of the Bible, um, the reason God could create the Earth on six days and on the seventh day rest is he didn't have millions of customers and legacy systems. 7:39 And, and most financial institutions have thousands of legacy systems, and the people who put them all together in w- have long since left the company, so nobody has an institutional knowledge of how the whole thing fits together. 7:52 Whereas when you're building it from scratch, you do have an institutional knowledge of how it all fits together. In fact, I'd say a m- decent percentage of it probably fits in your own head. 8:00 Y- y- you know, uh, my head is pretty small. Yeah. So fortunately, we've outgrown that. But, um, I love this attitude, right? Like, this attitude I feel like is the correct one for, uh, um, for engineers to hold, right? 8:11 Like the, uh, rather than the scoffing, dismissive attitude, the, um, "Oh, I'm about to learn something. My curiosity should be piqued by this thing that looks, uh, inconsistent or that I do not understand." Yes. 8:23 Um, uh, I feel this is, it, it's the conversation that I always end up, uh, having with, like, a new financial technology engineer after about a year, right? Yeah, yeah. 8:31 You know, there's usually, like, uh, um- Th- this moment [laughs] Yeah, the, like, hand on the shoulder side of like, "May you hope that you too will write some code someday that lasts 50 years." Yes. 8:40 Like, uh, um, uh- Wha- what an incredible, uh, achievement. 8:44 So talk to me about, um, some of the esoteric, uh, issues you've had to solve, some of the hidden pain that, the friction points that don't get really talked about publicly but make a meaningful difference. 8:54 So I'm a money mover. I just want to, you know, make payouts, or I want to accept funds, or I wanna run payroll, or whatever it is. 9:02 What are some of those friction points underneath the hood that people just sort of accept that is some of your favorites? 9:10 Th- this is a great question, and I feel like my enthusiasm for this is because, you know, when you, you, you, as a consumer, you're interacting with, uh, financial services, it, it at times can seem very easy, right? 9:21 Yeah. Like, it, it, it's great, right? Like, when, uh, your Apple Pay tap just works, it's fantastic. Yeah. 9:26 Um, and then there's the second, uh, version, where someone a couple years into their career is, um, uh, "Ah," like, "My sweet summer child, you don't understand the, uh, complications that exist here." 9:36 And then there's another layer, which is just like, yeah, it, it, it, like, uh, it, it gets back down in simplicity again, which is fantastic, right? 9:44 Like, when, when you're interacting with, uh, uh, with the underlying systems, like at Stripe when you're interacting with the card brand directly or Increase when you're interacting with the Federal Reserve, uh, uh, uh, directly, um, simplicity springs again, right? 9:57 Yeah. Like it, uh, uh, um, oh, of course you can't be acting on that piece of information. You didn't have that piece of information. Like that, uh, um, uh- I, I love that curve. 10:06 You, you start out as like, oh, there's this really, uh, big abstraction. It might be a UI or a mobile app or an API. It's super abstract. It's nice and easy. 10:15 You get a little closer to the metal, maybe to improve your unit economics. Maybe you start working in the industry, and you're like, "This is so hard, and everything's complex." And then that's funny. 10:23 You go the other way, and it's, you get all the way there, and it's actually quite simple again. Uh, um, I'm watching out whether or not there's another one. 10:29 You know, I, uh, um, I don't know, maybe when you start running a central bank yourself, things, uh, uh, get complicated. Yeah. Um, yeah. 10:37 I'll have to ask the mZero guys and, and what's happening in crypto for some of that. You call your infrastructure, like, lovingly boring. I feel pretty strongly about this one, right? Like, which is, um, uh, I 10:50 want reliable, transparent, flexible systems. There's a limit to the excitement that I want in my bank, right? Mm-hmm. Um, uh, I... 11:00 Uh, different people should attain their excitements in different ways, and maybe you're kind of, uh, uh, um, um, your blood is pumped by day trading, et cetera, et cetera. 11:08 But I'm fine with my bank enjoying making things just so, right? Like, ev- everything will be, like, predictable in how it works. Everything will be, uh, uh, arranged as I was expected. The numbers will all sum up. 11:21 I am not going to get my pulse racing, uh, uh, based on, um, based on my interactions with my bank. It happens to align well with my, with my personality. Uh, it makes it, it makes it perhaps unattractive to, to sell. 11:35 Like, to say lovingly boring is not a, a, a- No, ordinarily, I, I used to work for a cards issuer processor, and, uh, what they sold was, um, five nines of uptime, uh, bulletproof mainframes. 11:47 And, uh, we've built it before. We've seen it before. Um, we're gonna give you the best SLAs in the business. 11:54 And that works consistently because, uh, your financial institution, when you have millions of customers, not being available on a Black Friday is existential, and it's very costly, and that is your job gone. 12:06 Like, you are the CEO that gets fired if that's the case. And if it's out for longer, uh, you might even end up in jail. Like, there are examples- Yeah... of that in history. 12:14 So lovingly boring is a sales pitch to a certain type of person, but it, it's not the, it's not the fintech sales pitch. Yeah, it co- correct. 12:23 It's, uh- It's what I find is, is boring and very much filters and self-selects, right? Language, uh, uh, that, you know, um, um, is advantageous. Yeah. My wife is in sports marketing. 12:32 Her world is, uh, i- is, is much more exciting than my world. And as you can imagine, this, uh, introduces dinner conversations of, "Oh, can you not shout about it to the world more?" And like- [laughs]... 12:43 uh, um- Quite the contrasting couple. Yeah. Yin and yang a bit. She is right. She's right. Uh, uh, but, but lovingly boring is appropriate for, uh, uh, for the kind of research here. I, I love it. 12:53 I love the fact that you're doing you still though. Like, there's something, um, to your point about something that might last 50 years, um, and compounding, there's, there's something to be said there. 13:02 And I'm interested in like what are the problems that now need to be solved because I, I... One would argue Stripe solved developer experience- Mm... arguably canonically. 13:14 Um, and you were involved in that in, uh, in your own way. What are... What's the frontier of developer problems, especially in financial services, that need to be solved today? Yeah. 13:22 Uh, I feel my, uh, preamble for this is always, uh, like financial technology is not, uh, uh, all that different from technology generally, right? 13:32 Like 90% of it still remains the same of just like build things people want. Of course, we end up talking about the differential, like, you know, uh, uh, 10%, but, uh, uh, to make sure that we place the right emphasis. 13:44 If you're not building things that people want, like don't worry about it, right? Like, uh, uh, uh, start- What a crazy idea... the, uh... 13:51 But there are things that are differential to financial technology, and, uh, uh, uh, that's oftentimes that there are, uh, a necessity to do a partnership pretty early on in your, your life cycle, a necessity that, uh, you are going to need to access these necessarily gate kept features. 14:06 Yeah. Um, and these are, uh, necessary but insufficient for your product. These can be barriers to existence. Give me an example. 14:14 A very, like, present one all the time in the, uh, uh, at this stage cliched financial te- technology app is the bill splitting app, right? Yeah. Like, uh, um, uh, great. Why hasn't this been done? 14:24 First of all, it hasn't been done because no one wants to pay for it, but let's leave that aside. Yes. Uh, uh, why am I able to iterate on it quickly? 14:29 Well, you'll get your bank partnership up and running in the next nine months, let's say, and then, uh, you may or may not have access to the things that you were hoping to have access to. 14:37 Um, uh, and then you'll be able to like finally interact and like iterate on your product, and that's, uh, that's hard learned- So how, how do you, how do you solve that though? Like how- Yeah. 14:47 Uh, uh, uh, I think that there are parts of compliance that should be done within the four walls of the bank, right? Mm. 14:53 Like, I, I feel this is, uh, uh, the kind of lesson from the last couple years worth of, of, uh, regulatory action of, uh, heads up, this part should be done by folks who are most best placed to do it- Yeah... 15:05 and who must sit in the room anyway and explain it to me, right? Uh, and therefore, uh, very much it... If you're able to have those reside within four walls of the bank, that makes sense. 15:14 Now, obviously there are other companies who, um, are at that stage of maturity where they can handle their own kind of bank secrecy act compliance, uh, uh, uh, program. 15:22 Uh, but if you're able to make it such that, um, early stage companies le- needed to think about that less, uh, uh, that would be a way that you would unlock that. Yeah. Talk stable coins to me. Of course. 15:34 [laughs] I live a, um, a life between Ireland and, uh, and the US. Uh, it's... You can imagine that involves some sort of economic money movement at times between the two. 15:44 So I'm very sensitive to the non-US money movement version of, of stable coins and very energetic about that. And in fact, that is like sufficient in my mind as a, a use case. 15:54 I don't think I, I, I need- You don't need it solved, huh? What's wrong with you? Why- [laughs] Again, boring. Uh, what, what, what are you expecting? Mm. Uh, um, that's, uh, uh... 16:06 That to me is, is a very resonant and real example for me. Yeah. And, and so you're living and dreaming that, but you, you sort of get that. 16:13 But what does somebody need to be able to do that, um, and still offer into TradFi? Like there's, there's an infrastructure that's parallel here and, and i- Are people finding that confusing? 16:25 Are you getting clients asking about it? Are you doing anything in that space? Yeah. Uh, um, of course folks find it confusing, right? Like I view, uh... As a society, I love that we are running these experiments, right? 16:37 Like, uh, uh, that we should be doing multiple things. Um, if you'll forgive the, uh, perhaps a bit of a stretch analogy here. 16:44 Uh, um, uh, like, uh, stable coins and crypto can be the Hyperloop of, of, of, of the world, and people should be running that experiment. And Increase's job is to make the subway run better, right? Mm. 16:56 Uh, and, and I would love for us to have both of these and to run these both as societal experiments and, and, and let us all improve. Do they need to be interoperable in some way? 17:07 Uh- Oh, certainly they need to come back at some stage, right? Yeah. Like, uh, uh, we have a... Presumably we will have footpaths between subway stations and Hyperloop stations at some stage, right? 17:15 Like, uh- But, but it, it strikes me that that's the bottleneck right now of, of like, uh, there's probably one major sponsor bank in the US that supports this. 17:24 Um, and there's a need for like 20 who do it in different ways with different risk appetites, and that do it more programmatically and can make it more efficient. I agree, right? Like it's, uh, uh... 17:33 There will be the different use cases for it. 17:35 If, uh, um, uh, the flavor that you put on the product as you're exposing it to someone else is in fact the fantastically important part, uh, we should have, uh, um, um, more institutions do so safely and soundly. 17:47 Um, uh, it... I don't think there will just be the one provider for that. Talk to me about building thoughtfully from the beginning. There's a, a lot of 17:59 people that just kind of are, are really hooked on blitz scaling at the moment, but what's separating the teams that are, uh, kind of scaling from those that hit infrastructure walls? What, what's the, what's the gap? 18:12 Yeah. I'm pretty fortunate to have come from a place where I got to build a, uh, a system before and having worked with folks, you know... So this was at Stripe. 18:21 Having built their, uh, uh, um, um, worked with the team building out their money movement functionality there. Um, and getting to do that, albeit at a different layer, was very informative, right? 18:31 Like, uh, in particular, there's lessons like keeping engineers in the nontraditional engineering places, right? 18:38 Like being, uh, uh, having engineers work some of the manual allocation tools because that, like, develops the, uh... You'll spot the cases where you can, uh, uh, automate. Uh, uh, that many of those have... 18:49 were good at Stripe, right? Like, one of Stripe's differential advantages was, in fact, keeping financial operations as an engineering function for, uh, painfully long. [laughs] Yeah. Yeah. So you, you must suffer. 18:59 Uh, you must experience the pain, uh, in order to build an engineering solution to it. Uh, I, I reject the characterization of suffer. Uh, uh, but it is- It is. This, this will be your learning opportunity. 19:09 I know, but this is Jensen, uh, once, uh- Yeah... wrote, "May you experience great suffering." 19:14 Um, what he means by that is suffering in the, in the Eastern sense, not the Western sense of, like, may you experience great growth. Yes. May you experience something that makes you better and stronger. 19:23 A- and oftentimes there's, like, uh, interesting details that crop up there, right? 19:27 Like, it, it's the certain personality that finds it interesting when the, um, uh, you know, when that, uh, bit is set on the check, right? Like, an, "Ah, I guess this should update my model of how this works," right? 19:38 Like, and, and, and that curiosity by an engineer when in financial operations is productive. I, I love it, man. And so I, I'm gonna leave you with a fun one. Um, so 19:48 what is the one part of the payment stack that you would rebuild from scratch if you could if it wasn't, like, the layer you're building at now? Um, a- again, I'm not a creative person. I'm boring with this one. 20:01 I'm just doing the parts that I want, right? Yeah. Like, the, uh, uh, the parts that I want, that we need is, um, the US depository rails are actually fantastic. We just need some last mile delivery, right? 20:12 Like, uh, uh, uh, FedAge, FedWire, Check 21. Like, they're actually, like, good, solid, universal, cheap, fast. Uh, we're missing the last mile, uh, work of it. 20:22 Uh, uh, and that's w- the, the, why we're building Increase. That's why folks enjoy using Increase to interact with those. Um, uh, so sorry to shirk the question. 20:31 Uh, uh, uh- Well, I'm gonna- My interests are pretty narrow. All right. So I'm, I'm gonna have to, I'm gonna have to ask you another question now. Oh, my goodness. 20:38 Well, in that case, what's the, what's the one stablecoin misconception you gotta kill? Like, what's the big one? I know what I- I feel that it needs to be, like, vastly different than the existing systems. Yeah. Right? 20:49 Like, that, that, like, uh, um, that we need them to be a million miles different from each other, right? 20:54 Like, connecting the two together is actually far more fascinating to me than, uh, whether or not you've made them- Than whether they compete. Yeah. C- yeah. Yeah. Yeah. Um- It's not a versus thing. I, I, I buy that. 21:04 The one that I'd kill is that they're cheaper. Uh, it's like, no. You... When you avoid compliance costs, of course it's cheaper. 21:12 Um, when you don't cover, uh, gas fees and account abstraction and all the intermediaries and you don't account for FX because it's all dollarized, then it's cheaper. So the... We got one each. Yeah. [laughs] All right. 21:24 Dara, if people wanna find out more about you or Increase, where do they go to do that? We are at Increase, and that's increase.com. Great. Thank you very much for having me. I appreciate you, sir. 21:32 [laid-back music]
