Latest / TFTC: A Bitcoin Podcast / #711: How to Turn Any App Into a Bitcoin On-Ramp with Harsha Goli

Transcript

#711: How to Turn Any App Into a Bitcoin On-Ramp with Harsha Goli · TFTC: A Bitcoin Podcast

• 0:00
  get that warm, get that warm cup on your hands. I am. I am using it. Uh, this, this is my current
• 0:06
  source of warmth. All right. It's working well. Oh my gosh. We got the deep freeze. It's good for,
• 0:13
  uh, focused work. I've found that over the last three days, I've had a lot of focused work because
• 0:17
  I've been forced to stay in the house. It's very punishing if you go outside. Yes.
• 0:23
  It's like the roads are icy. If I go visit a friend, I'm risking life and limb. I might as
• 0:27
  get work done you know yeah yeah why why risk it you can just sit at the computer you can play with
• 0:32
  all these ai tools you can push towards a goal you may have but uh i think the idea of focus work uh
• 0:41
  is apropos for this discussion because i know that you've had a lot of focus work building out magnolia
• 0:50
  relatively under the radar behind the scenes for the last couple of years.
• 0:57
  And as we were discussing before we hit record, I'm sitting down with Harsha Goley.
• 1:02
  And for those of you who are not familiar with Harsha, he's been in the industry for quite a while
• 1:08
  and worked on many impactful projects and companies.
• 1:12
  And so I think instead of me giving your background, it probably makes sense for you to do that, Harsha.
• 1:18
  to sort of explain your journey within the Bitcoin industry
• 1:23
  over the last better part of a decade,
• 1:25
  over a decade now, I think.
• 1:26
  Dude, thank you for the incredibly warm intro.
• 1:29
  Yeah, we're over a decade now.
• 1:30
  I started working in the space back in 2015.
• 1:33
  I was just like a sophomore in college
• 1:36
  and I like back then was really into like open source,
• 1:40
  you know, like anything, right?
• 1:41
  So I just like me and my friends,
• 1:42
  we would just go to hackathons and just work on,
• 1:44
  you know, like whatever get out projects we could find,
• 1:46
  make them better or like create something completely new and like ship stuff.
• 1:50
  And one of them, you know, someone was telling me about Bitcoin.
• 1:53
  And like to me, the best part of it about Bitcoin was that it was a financial thing that was open source.
• 1:58
  And I'm like, OK, neat. I can mess with this.
• 2:00
  Of course, I didn't know C, you know, I only knew Python and JavaScript.
• 2:03
  So I would I would like try to like find whatever clobbered together libraries other people put together.
• 2:08
  If anyone else is listening, who was a dev around the time that the state of Bitcoin tooling back then was abysmal,
• 2:14
  which, you know, was pretty awesome for me to work in.
• 2:16
  Um, and yeah, like from there, I decided like, okay, this stuff is pretty cool.
• 2:22
  I'd like to work on this stuff a little bit more.
• 2:24
  So at the time I had like, you know, a junior, uh, you know, like job at a company called
• 2:29
  GE, uh, they like GE power specifically that division.
• 2:33
  And I was like, okay, well, you know, I'd like to work on Bitcoin stuff that they work
• 2:36
  on turbines.
• 2:37
  So that's not really the same thing.
• 2:39
  Uh, so I got a job at BitPay, which was in Atlanta and yeah, they just like had me working
• 2:44
  on effectively this is very this was very controversial we're way past it but they had
• 2:48
  me working on like a bunch of segwit 2x tooling right so i was like the guy who like spun up the
• 2:53
  initial segwit 2x test net i remember like getting a bunch of like usb miners all right and like
• 2:58
  sticking into a bunch of raspberry pi's we had like 50 of them you know in a room uh just buzzing
• 3:02
  away and i wasn't there for the actual activation but i can imagine what what day that was because
• 3:08
  the whole thing just fell over immediately off by one off by once and so that was that was pretty
• 3:15
  cool but it was cool like setting up a test net you know like from scratch and yeah like after that
• 3:21
  um my friends and i we like you know realized like okay there's a lot of work in this space and
• 3:25
  there's a lot of like tools we can work on so we started our own startup basically like consulting
• 3:29
  firm that like had some internal products and that did really well until uh eventually i met
• 3:35
  some folks from Fold who were looking for
• 3:38
  a back-end engineer, and they
• 3:40
  wanted to work on some really, really cool stuff.
• 3:42
  So I was like, okay, well, you know,
• 3:44
  I know I have this startup, but it's doing pretty well.
• 3:46
  It can probably chug along without me for a while.
• 3:48
  Let me just hop on this
• 3:50
  team and see what we can do here.
• 3:51
  I still remember the job that I ended up
• 3:54
  applying for was
• 3:54
  a Clojure dev.
• 3:58
  I don't know if any of you guys know about Clojure, but
• 3:59
  it's a crazy language.
• 4:01
  I didn't know any Clojure. Still got
• 4:04
  the job, got hired.
• 4:05
  ended up working on their like back end uh like wallet infrastructure for a minute um which is how
• 4:12
  i started getting familiar with like how enterprises wallets really really worked
• 4:15
  and yeah like from there we shipped lightning and you know bitcoin uh for fold and eventually like
• 4:22
  you know launched the actual uh like full debit card right which actually is what got traction
• 4:27
  fold did not have too much traction on the gift card stuff the debit card really took off um even
• 4:31
  no gift cards never really stop being part of the bottom line. And then, yeah, like, so from Fold,
• 4:35
  and, you know, I've worked in a bunch of different places. I think Fold was a place where I kind of
• 4:39
  decided, like, okay, this stuff is better than, like, college. You know, this stuff is better than,
• 4:46
  like, working on anything else. Just the energy of, like, you know, of thesis and, like, Will and
• 4:52
  Tom and all the guys there was just awesome. So after I left there, I went to work at a couple
• 4:57
  different places including lightning labs just to work more on lighting stuff uh that was really
• 5:01
  really cool uh met some of the smartest people in my life there and then worked at voltage as well
• 5:06
  got my like chops and like product management um and then worked at swan another like a company i'm
• 5:12
  sure all you guys know uh again doing product management stuff specifically uh for their data
• 5:16
  team and their lightning teams and all this was kind of like going in a very specific direction
• 5:21
  because i wanted to learn how to run teams effectively on a myriad of different problems
• 5:25
  because there's been like this problem that I've been kind of seeing in the space,
• 5:30
  which is like, you know, the tech,
• 5:32
  like even though I've been working on Lightning for a long time
• 5:34
  and trying to get it implemented with more people,
• 5:36
  I noticed that most people don't use the new and shiny in Bitcoin.
• 5:42
  That's never really been a thing about Bitcoin.
• 5:45
  Tech enthusiasts like myself do, sure, but the masses do not.
• 5:50
  And the more I looked at it, the more I realized,
• 5:52
  like the problem is it isn't like you know bitcoin necessarily needs a new you know functionality in
• 5:58
  order to like get mainstream it already was the problem always was that in order to actually get
• 6:03
  bitcoin you could only go to reliably a couple of the big exchanges and if you were a small company
• 6:10
  smaller than you know coinbase uh in order to get that same functionality integrated into your
• 6:16
  application the exact same way Coinbase would, you had to do so much, right? And you had to rely
• 6:23
  on sometimes really sketchy partners. And it was a huge burden. It was one of those things where
• 6:30
  the payoff was immense, but the costs were somehow even more immense, right? I remember at Swan,
• 6:37
  we had like five different migrations between like all these different providers. It was nonsense,
• 6:42
  You know, especially when like 80% of your, um, 80% of your like engineering budget goes
• 6:49
  toward this one specific problem.
• 6:50
  That, that's not a good thing to me.
• 6:51
  But anyway, that's like my background in a nutshell.
• 6:53
  I've, I've been working in Bitcoin for, uh, I think, yeah, over 10 years now.
• 6:57
  It's, it's the thing I, I love.
• 6:59
  I've, I, uh, yeah, I, I think I'll be working in Bitcoin for the rest of my life, but you
• 7:03
  know, why not at this point?
• 7:06
  Yeah.
• 7:06
  And so this brings us to Magnolia, which is a company that you found and have been working
• 7:10
  on for a couple years now. And again, highlighting the end of your introduction there, which is this
• 7:20
  problem that exists for many companies in the space who basically just want to interact with
• 7:26
  Bitcoin and make it easy for their customers, most importantly, to interact with Bitcoin.
• 7:31
  The third party dependencies that have existed in the industry throughout the years have stood on,
• 7:39
  for lack of a better term, shaky foundation. And through my conversations with you and what you're
• 7:45
  doing at Magnolia, that's exactly what you're trying to solve is providing a sound foundation
• 7:51
  for companies trying to build the space. Yeah, it's quite surprising, actually. You know, like,
• 7:57
  I got to empathize with all these builders, you know, I feel like if you're a builder in Bitcoin,
• 8:02
  you've got a pretty like tough task ahead of you. And, you know, everyone seems to like solve it,
• 8:07
  but you spend all this time and energy cultivating an audience online,
• 8:13
  getting them to use the application that you're trying to get them to use,
• 8:17
  whether it's a wallet or a Bitcoin lending app or even a white glove service.
• 8:22
  That takes a lot of effort, a lot of energy.
• 8:25
  And it kind of sucks because after you've done all this effort, all this work,
• 8:30
  you try to make it as easy as possible to onboard your customers.
• 8:33
  And you realize getting these kind of integrations in your application
• 8:37
  isn't like, you know, signing up for a credit card, right?
• 8:40
  Where it's like one form or going to a bank and, you know,
• 8:42
  like signing a couple of papers and then, you know,
• 8:44
  getting access to the financial system.
• 8:47
  When you try and do this as a business,
• 8:48
  specifically like, you know, in Bitcoin and, you know,
• 8:51
  stable coins or anything else, it's so tedious.
• 8:55
  It's so expensive.
• 8:57
  And it's really just out of reach, right?
• 8:59
  You basically have two options.
• 9:00
  One is you just get like state licenses.
• 9:03
  You get money transmitter, like, you know, like access, right?
• 9:07
  And it's state by state.
• 9:08
  It's not like if I go to a bank and I sign up for a bank account, I can use that bank account anywhere in the U.S.
• 9:13
  That's not how it works if you're a business, you know.
• 9:16
  And then alternatively, you set up an account with someone else who runs the infrastructure for you.
• 9:22
  And that process is incredibly tedious.
• 9:25
  I think like for us, it can go over like three months when we were doing this ourselves.
• 9:30
  And then on top of it, they charge incredibly high minimums in the neighborhood of like $20,000 to $30,000 a month.
• 9:38
  It's like this is for a massive business like Polymarket, not for someone who's got a pretty effective retail audience, but not like I've dominated the world effective.
• 9:51
  And that's just kind of wild to me.
• 9:54
  Um, but, but yeah, so like what we're trying to do, what, what the, I've, I've really kind
• 9:59
  of identified as a huge problem in Bitcoin is I'm trying to turn every single one of
• 10:04
  these applications, like big ones for sure, but also smaller ones as well into bridges
• 10:08
  where you can just go with, you know, your credit card, your bank details, whatever,
• 10:13
  and be able to convert, uh, yourself from a precoiner to like a Bitcoiner.
• 10:18
  That that's the, that's the idea here.
• 10:19
  All these applications should be able to take someone's money and give them Bitcoin, right?
• 10:23
  Instead of shoving them off to Coinbase or something.
• 10:28
  And so what was the process of building this out like?
• 10:30
  Let's get a landscape of the technical architecture that exists.
• 10:34
  And then I guess the compliance architecture necessary to facilitate this as well.
• 10:40
  Yeah.
• 10:40
  So, you know, from a technical point of view, it's actually like, I guess for me, it's not that hard because it's something I built like three or four times in my career.
• 10:47
  right but it's like having an enterprise wallet system having uh what's called a like a double
• 10:53
  entry accounting ledger so it's just a it's just a well-built like accounting system um and then on
• 10:59
  top of that having a robust api that is not overwhelming that part's you know honestly not
• 11:04
  the most complicated part um i've done that multiple times most people have some version
• 11:08
  of that in in the systems you know um and then of course like a connection to your banking partner
• 11:15
  And that's where things get hard. The hard part is interacting with the banking providers. The hard part is being able to talk, really, step one. That was actually kind of wild when I realized just how hard that was going to be at the time. I think I had really long hair. I had just done a huge hippie dive. It was awesome. I lived in the van for a couple of years.
• 11:35
  and uh i realized like okay i'm going to these conferences you know i've got my hair tied up all
• 11:39
  nice and stuff uh in order to meet like a lot of these bakers but um i i've got to like you know
• 11:45
  make some changes to my appearance here you know i've got to like pick up a different kind of uh
• 11:49
  like repertoire and i've got to like cut my hair i've got to stop referring to ourselves as crypto
• 11:54
  to try and explain myself to them just be like no we do bitcoin you know the bitcoin is the thing we
• 11:58
  do you know we don't get involved with anything else and after like honestly a large number of uh
• 12:04
  of conferences and conversations, we managed to find a lot of really healthy inroads into the
• 12:09
  banking world with really good compliance folks on our backs too. There's a whole separate world
• 12:15
  of Bitcoin dedicated just to compliance. And it's kind of fascinating, right? Haley Lennon is
• 12:20
  a great name in this space too. She's the general counsel of Fold, right? And she was our lawyer at
• 12:27
  one point. And she was able to intro us to all these people who are on the Bitcoin compliance
• 12:32
  side of things. And they know all the right people in the compliance world to get you past
• 12:38
  all the big roadblocks that exist for everyone else. That was honestly pretty wild for us to
• 12:42
  realize, okay, compliance is not that effective. And the way how compliance tends to work is it's
• 12:50
  all about making sure that who you know is someone with a lot of moral accountability,
• 12:57
  essentially. And that does apparently go quite a long way in the compliance world. And I do kind
• 13:02
  see it and understand it now. Yeah. And so let's take a step back, a high level view.
• 13:11
  Who is your ideal customer and how do they interact with you going from zero to one?
• 13:16
  So right now, like a huge focus for us is we realized that a huge pain point is OK. So
• 13:21
  like Bitcoin, right? Like the biggest value of Bitcoin is the value it has. And I think like
• 13:28
  2026 is turning out into this year where a lot of people are moving their Bitcoin into some kind of non-custodial setup, whatever it may be, where they can also take a loan out against it.
• 13:38
  Right. And be able to utilize the value they've accrued in some kind of way.
• 13:43
  This has been like kind of a silent explosion over the last, like, you know, I think six months to a year.
• 13:48
  And I think it's going to really, really, really get big in 2026.
• 13:52
  I have a crazy high ARR now, which is wild.
• 13:55
  L-E-D-N.
• 13:57
  They're like, I think, a leader in this space.
• 13:58
  I think it's like $100 million or something.
• 14:00
  And the reason why I bring this up is because all these different Bitcoin-backed lending folks, they all issue the loans in a stablecoin.
• 14:09
  And something I've learned from talking to these folks is the whole narrative for stablecoins for trading is kind of like dead, right?
• 14:17
  But people are still taking out these loans.
• 14:19
  And that's because they're taking out these loans at decent rates specifically to use them in the real world, you know, to pay for their like daughter's college or, you know, to upgrade a property they have or whatever.
• 14:32
  And so what we're plugging into for a lot of these folks is we're plugging into their systems to convert their stable coins to fiat directly for their end user.
• 14:41
  That way their end user doesn't have to take a couple more points at Coinbase or something like that.
• 14:45
  So that's someone that we're like really kind of charging headfirst in.
• 14:48
  But basically, our ideal customer is any kind of Bitcoin business that's got a high retail audience that they're looking to monetize effectively.
• 14:59
  We've interviewed a couple other wallets as well that are doing just that.
• 15:03
  Oh, and also anyone who's looking to access the US and EU financial markets, essentially.
• 15:10
  So South American companies and Asian companies, too, who would like to onboard folks from the EU.
• 15:18
  Super, super valuable for us right now.
• 15:21
  Let's dive into the stablecoins conversation.
• 15:23
  What are you seeing?
• 15:24
  You mentioned the trading narrative, which was the stablecoin, the predominant stablecoin narrative for many years.
• 15:31
  I think it makes sense.
• 15:33
  Yeah.
• 15:33
  is birthed out of bitfinex which is this big exchange and there was obviously the ability to
• 15:39
  send these stable coins between exchanges to um our prices or just get better margin
• 15:45
  rates on on different exchanges at any given point in time but it seems like it's breaking
• 15:49
  out of that the confines of trading only and breaking into um true for lack of a better term
• 15:57
  utility for everyday retail consumers who aren't traders um do you see it coming to america
• 16:04
  yeah i think that's a big question i have in my mind right now like do stable coins make sense
• 16:09
  in the united states and then even if they do or don't what um is the activity like outside of the
• 16:17
  united states from your perspective yeah i mean to to your point i think everyone everyone knows
• 16:23
  that there's a lot of activity for stable coins in developing nations, specifically in South
• 16:27
  America and Central America. But the US, now what's happening in the US is actually quite,
• 16:32
  quite fascinating. Right now, because of the Genius Bill, the Genius Bill
• 16:36
  creates a lot of pressure to implement stable coins somewhere in the US, right? Like in the US,
• 16:43
  like banking system in particular, that's fundamentally what the Genius Bill does.
• 16:46
  And if y aren aware of like how the Genius Bill really works it basically creates a ton of incentive to buy up T and then issue stablecoins against these T right
• 16:59
  And so now you have a lot of stablecoins
• 17:00
  that are rolling around, like, okay,
• 17:02
  we need to find a use for these stablecoins
• 17:04
  in order to justify our business, right?
• 17:07
  Because we're giving most of our return
• 17:10
  from the T-bills that are generated
• 17:12
  back to the end user,
• 17:13
  which is a whole conversation in and of itself.
• 17:16
  But we have these stablecoins,
• 17:17
  and we need to fit them somewhere.
• 17:19
  And so what we're kind of seeing
• 17:20
  is we're seeing this like really kind of
• 17:22
  backwards thing happen
• 17:23
  where normally if people implement products and stuff
• 17:27
  after there's some level of demand,
• 17:29
  but what we're seeing with stable coins
• 17:30
  is people are implementing stable coins everywhere
• 17:33
  and seeing where they fit,
• 17:36
  you know, what sticks and what doesn't stick.
• 17:38
  And one of the things that's sticking quite well
• 17:40
  is this concept of, okay,
• 17:42
  everyone's got some amount of like, you know,
• 17:45
  like Bitcoin or maybe even like some other assets
• 17:47
  as well that they want to take out loans against. Taking out the loans typically ends up being part
• 17:53
  of this, you know, like protocol, either it's a smart contract, you know, on Aave or something
• 17:57
  like that, or, you know, more like I think gaining more popularity, it's some kind of like Bitcoin
• 18:01
  backed loan, right? But either way, it doesn't really matter. You're going to get a stable coin
• 18:06
  as a result of this loan. And getting a stable coin is cool, but a lot of these people are not
• 18:12
  using these loans to like gamble more. That's just not what a lot of these folks are doing.
• 18:17
  They're using it to like actually improve their lives in some way.
• 18:21
  These loans are not that high anymore.
• 18:23
  I think before it made a lot of sense to like, you know, kind of gamble on another project or whatever when they were between like, you know, 15 and 12.
• 18:34
  But now you see them going kind of low in some platforms.
• 18:36
  It's as low as like 6 percent.
• 18:38
  I think a pretty good price you can get in Bitcoin is 10 percent.
• 18:41
  And those are like low enough to actually justify like building, you know, real value in the real world.
• 18:47
  you know, and that's kind of fascinating. So seeing that take off is really kind of interesting.
• 18:53
  Where else will stable coins fit? I don't know. You know, I'm a skeptic in general. I like seeing
• 19:00
  things before I believe it. You know, I, you know, call me crazy, but I don't know where else it fits
• 19:05
  right now. But I see that one really kind of like actually getting use cases right now. I'd be very
• 19:11
  curious to see if anyone uses them on the day-to-day elsewhere yeah it's obviously a big topic of
• 19:18
  discussion right now with the genius act and again it seems like they're trying to manufacture
• 19:24
  demand for these treasuries and that's the one thing a question in my mind is um is it a be
• 19:32
  careful what you wish for scenario with the genius act where um you have to recognize that stable
• 19:37
  coins, the success of stable coins to date has been driven by this regulatory gray area
• 19:42
  that's existed and allowed individuals outside the borders of the United States to get access
• 19:48
  to dollar instruments rather trippily.
• 19:52
  Yeah.
• 19:53
  I mean, like to their point, the reason why they're doing it, why they're so interested
• 19:58
  in doing it is because if it's not the U.S. dollar, it's the Chinese yen, right?
• 20:05
  It's the Russian ruble.
• 20:07
  And so everyone wants to like take advantage of these like, you know, deteriorating economies and just plant our flag in the first thing.
• 20:14
  You know, we've done this with real dollars like in the past, you know, especially like in Africa to stabilize economies.
• 20:20
  And, you know, in Zimbabwe, like, you know, after the currency like went to hyperinflation, people started using the U.S. dollar, you know, that that is like that is what happened.
• 20:29
  But I do agree that it is quite literally artificial demand.
• 20:33
  You know, that's what the genius bill is.
• 20:35
  and I do think it's interesting that we're just trying to like fit it in anywhere we can.
• 20:41
  I look forward to the day when my Walmart app offers stablecoin as a rewards option.
• 20:48
  You know, like I'm sure that's coming.
• 20:50
  You know, that's that's going to be wild.
• 20:54
  Switching gears a little bit and building on conversations we've had over the years.
• 20:59
  I think Magnolia, the impetus for you starting it again, as you alluded to,
• 21:05
  is that there are companies that try to do something,
• 21:08
  have tried to do something similar in the past,
• 21:10
  I think most notably Prime Trust and have failed epically.
• 21:17
  And so I guess what are the lessons learned from their failures
• 21:20
  and how you're trying to prove on that and make sure that Magnolia is sound?
• 21:26
  Yeah, dude. Yeah. I remember Prime Trust,
• 21:28
  the first version of Fold built against Prime Trust, you know?
• 21:32
  And we, you know, frankly loved it.
• 21:34
  It was great from a tech point of view.
• 21:36
  It was everything we ever wanted.
• 21:38
  It's just that it turns out the way how their operating side like worked was not, you know, everything we wanted.
• 21:45
  There's this like really kind of incredible story with Prime Trust and I think also like Fortress Trust where they ended up just losing a ton of money because they had some like really complicated key ceremony.
• 21:58
  right and the result of this really complicated key ceremony was they ended up losing the actual
• 22:07
  private keys that backed the deposit addresses right at prime trust and so people were depositing
• 22:13
  money at prime trust uh and for i think a couple weeks no one realized that they were they were
• 22:19
  putting money into a black hole you know they really were and when it came time to withdraw
• 22:23
  the money, you know, for typical operation purposes. Only then did they realize like,
• 22:28
  we don't have this cash, you know, Prime Trust wasn't hacked, you know, or they may have been
• 22:32
  hacked at some point there in their history, but this specific instance, they weren't hacked.
• 22:37
  And that is, that is just baffling to me. So like really my career is kind of built on
• 22:43
  the failures of these entities. And it's not just Prime Trust, you know, there, there's a couple of
• 22:47
  them out there. Uh, because folks like me and myself, you know, specifically, we had to deal
• 22:52
  with these ramifications, right?
• 22:54
  So we learned a lot.
• 22:55
  Some of the big things that we really learned
• 22:57
  is like the compliance folks,
• 22:59
  the operation folks,
• 23:00
  and the engineering folks,
• 23:01
  they have to be in the same room.
• 23:02
  They cannot operate independently, you know?
• 23:05
  They need to constantly talk.
• 23:06
  They need to understand like,
• 23:07
  okay, here's what we're implementing.
• 23:09
  This is how we're doing it.
• 23:11
  This is what our own like policies say.
• 23:13
  And here's what needs to get modified.
• 23:15
  This has been a really, really valuable exercise for us
• 23:19
  every time it happens,
• 23:20
  specifically because really often we realize something can get shaved down.
• 23:27
  Something doesn't have to get complicated, right?
• 23:29
  Or even, okay, we're implementing something that is actually very dangerous for us.
• 23:34
  And that is kind of a huge boon.
• 23:37
  From the security side, we made the decision that my computer is just completely segregated away from all of our systems.
• 23:43
  and from ensuring that our security is on top,
• 23:47
  our CTO, he's 100% in charge of that.
• 23:51
  And that's his focus, right?
• 23:52
  Alongside building features and stuff.
• 23:54
  But this separation of concerns
• 23:56
  makes it so that I'm able to work on my stuff
• 24:00
  and he's able to work on his stuff.
• 24:01
  And my job is just ensuring the machine
• 24:03
  gets to keep going forward.
• 24:04
  And his job is making sure
• 24:06
  that the machine is well-oiled to go forward.
• 24:08
  The separation of concerns works out,
• 24:10
  I think, a lot better.
• 24:11
  I know that the Prime Trust
• 24:13
  method of doing this stuff was a little bit more democratic where, you know, keys were
• 24:20
  separated between a bunch of different people.
• 24:22
  We have separated keys out.
• 24:23
  It's not one individual, but our keys aren't separated between individuals.
• 24:26
  They're separated in like, like actual different key servers, you know, that only one individual
• 24:32
  who's, you know, like got a shotgun at home and takes security very seriously, like is
• 24:37
  in charge of.
• 24:37
  And I think that's probably like the right way for how we handle things right now.
• 24:43
  So what is this key setup?
• 24:48
  I mean, you mentioned multiple different servers.
• 24:50
  One guy, how do you ensure that there's no key man risk with the setup?
• 24:56
  Yeah, so we use multiple different key servers from different other custodians specifically.
• 25:02
  So they're all watching out.
• 25:03
  I think a lot of it is automated, but for deposit addresses, those go into cold storage directly,
• 25:09
  but it's not cold storage that's run by us.
• 25:11
  It's cold storage run by someone else.
• 25:13
  That way, you know, like there isn't really a situation where we mess up.
• 25:16
  Our money that we send out is completely segregated from money that we receive.
• 25:21
  So there's just an operations focus to make sure that, you know, we always have enough money to withdraw.
• 25:27
  Yeah, like we've had to go through kind of a lot of iterations of this design for our compliance program specifically.
• 25:35
  And so we have this all really well documented in our policies.
• 25:38
  And whenever regulators or banking partners or even customers ask us for these, we're able to provide specifications on hand, which is nice.
• 25:48
  But it's really, in my opinion, nothing that isn't industry standard.
• 25:52
  This kind of separation of concerns is what everyone should be doing.
• 25:55
  It's something super crazy or magical.
• 25:58
  Are the regulators getting smarter on this, on how this technical architecture should be built?
• 26:05
  They're not really on the technical architecture side of things.
• 26:08
  They're really concerned about like over leveraging rehypothecation and making sure that certain individuals are touching it or not touching it.
• 26:19
  Right. Like that's that's really what they care about.
• 26:21
  And they they've gotten smarter in the sense that they're not going to allow like fancy diagrams to get in the way of their understanding.
• 26:30
  Right. So if we send them diagrams and they don't understand it, they'll tell us it's not understandable.
• 26:34
  at the back you know um make it better fair enough you know like so um which we've done a
• 26:41
  couple times we have like a whole uh like project for uh diagrams specifically for regulators like
• 26:47
  for this topic and they're huge they're like powerpoint presentations that are like over 200
• 26:51
  pages long right with 200 slides but it works you know they get it they get the idea yeah
• 26:57
  and so now i mean you've got a big announcement coming up we're recording this on tuesday the
• 27:04
  27th but i think over the weekend you guys are going to be um coming out with some news and this
• 27:11
  will be published after so by the time this reaches the audience the news will be announced
• 27:16
  um what uh what do we have in store so uh like later this past year right i think in december
• 27:25
  we launched a price oracle right we want this yeah it would do um we found some people really
• 27:31
  liked it and you know onboarded it and this price oracle works specifically with uh discrete log
• 27:36
  contracts right so it's like a a much more uh like trustless way of uh handling these bitcoin
• 27:43
  back loans and we we were also we also like announced that okay we're supporting um and so
• 27:49
  instead of stable coins you know you can do disbursements for these loans via fiat so you
• 27:53
  get, you know, cold hard cash in your bank. And people are really excited about that.
• 27:58
  The big announcement that will be out by time that by the time you guys hear this
• 28:03
  is that we are releasing our full fledged like banking products for everyone to use right in the
• 28:12
  U.S. and soon the EU. And so what I mean in the U.S., I mean, we can we can launch this in New York,
• 28:18
  florida california texas and all the states in between and what this product is it's it's a full
• 28:26
  bank right over an api so you can spin up individual accounts right like you know routing
• 28:32
  account number for each of your end users through this system if that's what you want to do put fiat
• 28:37
  into it or have it do an ach poll which is like auto pay convert it into bitcoin and then send it
• 28:43
  whoever it is you need to send it to this is to my knowledge the the the quickest way to convert
• 28:50
  someone like convert someone's fiat to bitcoin or lightning um and the most uh programmatic and
• 28:58
  like extents or expendable way or whatever um oh my gosh i'm forgetting the term but yeah
• 29:03
  this is the most like programmatic way you can do it uh so that's kind of a huge huge win so
• 29:09
  there's no iframes involved in this it's not moon pay or anything like that uh which means that
• 29:15
  people can create their own custom flows as as they'd like right and that also does include
• 29:22
  with ai so as long as you're onboarded with uh with with us and past kyb which we've gotten down
• 29:28
  from six weeks to less than a week now so it's really easy to to get turned around on that you
• 29:33
  can just like implement our api into existing infrastructure and turn any application into a
• 29:39
  full-fledged bank that's huge that's what i wish i had five years ago how's this work on the licensing
• 29:46
  side of things similar prime charts in that way where you guys have licenses and charter
• 29:52
  so we have some licenses uh but really like the big chains that we made in 2045 is we realized like
• 29:58
  what we should do instead is we should really focus on getting like everyone is just saying
• 30:04
  this is great come back like we we'd love to talk when you have more than you know a couple states
• 30:08
  right we realized people want to go to all these states so we figured out a way with our banking
• 30:14
  partners we have three different banking partners to use their licenses in different states so that
• 30:19
  we can actually just go to all these different states all at once right it turns out some of
• 30:23
  our banking partners also have uh locations in eu as well so we can do native european transfers
• 30:29
  so not swift in the eu um we're working on implementing that right now that has to come
• 30:34
  in a month or two after we do the US launch.
• 30:36
  But this brings our services kind of everywhere right now.
• 30:40
  In the future, I think it's going to make sense for us
• 30:43
  to finish up these applications.
• 30:46
  Unfortunately, these applications normally take a year or two.
• 30:50
  Right now, because of all the stuff that's going on,
• 30:52
  everyone is backlogged.
• 30:54
  So it's actually going to take even longer than that.
• 30:56
  So we realized we're going to have to come back to the applications
• 30:59
  whenever the regulators get back to us
• 31:01
  and just kind of handle it from there.
• 31:03
  But until then, we can go in all these states and bring our services to just about anyone we can.
• 31:10
  And that's a pretty big win for us, right?
• 31:12
  It's gotten people really excited.
• 31:15
  You mentioned AI.
• 31:17
  How do you see AI playing into all this?
• 31:19
  Because I've been leaning into it as heavily as possible here at TFTC, doing research at 1031, things like that.
• 31:29
  And it is becoming more and more powerful by the day.
• 31:32
  I was actually playing around with, I had my agent playing around with Money Dev Kit last night, and it was pretty cool to see what it can do.
• 31:40
  So how do you see this intersection of AI and not only what you're doing at Magnolia, but Bitcoin more broadly?
• 31:49
  Dude, the financial system should not be locked out to people who are building.
• 31:53
  It just should not be, right?
• 31:54
  We want to make it as successful as possible.
• 31:56
  So if people want to utilize banking in their VibeCore products, we can make sure that they do it safely.
• 32:04
  But AI can actually do the implementation, create incredibly delightful flows for people.
• 32:09
  And yeah, I agree.
• 32:11
  We've been leaning into it pretty hard too.
• 32:13
  It's helped us out kind of tremendously.
• 32:16
  The different teams at Magnolia handle it very differently.
• 32:20
  right um so our engineering team has a very like complex and precise method of handling ai that's
• 32:27
  very different from how our compliance team handles it and our compliance team was able to
• 32:31
  reduce its you know like our kyb process our onboarding process from like six weeks down to
• 32:37
  less than one with the help of ai just like you know optimizing what it can and reducing flows
• 32:42
  and trying to get as many uh like you know multiple birds hit with one stone so to speak
• 32:46
  and I use it kind of all the time,
• 32:49
  specifically with the Ralph Wiggum loop,
• 32:51
  if you're familiar with that.
• 32:52
  I use that all the time for our docs
• 32:54
  and our like main webpage as well.
• 32:57
  That way I don't have to bother our devs about it
• 32:59
  and I can make sure our docs are as accurate as possible.
• 33:02
  I'm actually watching it run right now
• 33:04
  because I'm going to El Salvador tomorrow
• 33:06
  and I need this new website like out.
• 33:08
  So I'm just like watching it just iterate.
• 33:11
  But we use it kind of a lot
• 33:13
  and I do kind of have this belief
• 33:14
  that I've always been against
• 33:15
  hiring as a growth metric, right?
• 33:19
  I do believe AI is going to really be able to like supercharge
• 33:22
  really, really powerful operators at a company
• 33:25
  and let people go even farther with, you know,
• 33:28
  their seed or pre-seed stage, right?
• 33:29
  We've come pretty far on pre-seed.
• 33:33
  And, you know, at some point we're going to like raise our next round,
• 33:35
  but we don't want to just, you know,
• 33:37
  like burn capital here for the sake of burning capital.
• 33:39
  We want everyone to go as far as they can.
• 33:42
  And AI really does supercharge folks It insane And to your point I think if you a founder out there not leveraging these tools obviously you got to be careful to an extent
• 33:55
  It's easy to foot gun yourself
• 33:56
  if you're not using it appropriately.
• 33:59
  But to your point,
• 34:00
  if you're a sort of high level operator
• 34:03
  that understands the risk
• 34:04
  and understands how to use it,
• 34:05
  if you're not utilizing it,
• 34:07
  you're going to get,
• 34:08
  I don't want to say maybe left behind,
• 34:10
  But I think from a sort of balance sheet management perspective, there's no reason not to lean into it so you can keep headcount as low as possible, get the revenue, profitability, and begin stacking sats as quickly as possible.
• 34:26
  Oh, yeah.
• 34:27
  Oh, yeah.
• 34:27
  Plus, with all the Twitter stuff, it's all everyone gets to see.
• 34:31
  So, you know, you have no excuse.
• 34:33
  You got no excuse.
• 34:34
  All the tricks are right there.
• 34:35
  BodBod is big.
• 34:36
  Or I think now it's renamed.
• 34:38
  But, you know, whatever the next trick will be.
• 34:39
  I'm sure it'll be very useful.
• 34:41
  Moltapot, Moltapot, no.
• 34:43
  Do you think that, like, I wonder if we're in a bubble on Twitter.
• 34:47
  And yes, it's obviously a massive topic of discussion in our Twitter bubble.
• 34:55
  But, like, how many people outside of that bubble are actually utilizing these tools or even recognize that they exist and are as powerful as they are?
• 35:02
  i i think like from a retail point of view most people are already on like chat gbt specifically
• 35:10
  because it has a free tier but using it in this kind of like complex way um is definitely in the
• 35:17
  realm of tech enthusiasts right it absolutely is there's still like a lot of setup a lot of uh like
• 35:22
  stuff you have to do you have to have like a prerequisite amount of knowledge to understand
• 35:25
  that code clog code exists right for something like um bod bot you know maybe like folks buy
• 35:32
  like the Apple Mini or whatever,
• 35:34
  where they figure out they can just run it locally.
• 35:37
  But as time goes on,
• 35:39
  you're going to see a lot of products
• 35:41
  that just spin up the whole thing for you.
• 35:43
  So one that's already existed before PlotBot
• 35:45
  is Interaction.
• 35:48
  If you're familiar with Interaction,
• 35:50
  it's a Telegram bot, essentially.
• 35:52
  Pretty delightful startup flow.
• 35:55
  But as things like this get bigger,
• 35:59
  you're going to see more full-fledged agents
• 36:00
  that I think are a little bit more appealing
• 36:02
  to the mass audience, right?
• 36:04
  So like a good check for me that I always have
• 36:06
  is my little sister.
• 36:07
  She's a chemical engineer, right?
• 36:09
  That's what she does.
• 36:10
  She just wrapped up her education,
• 36:11
  but she is not in tech in the slightest,
• 36:14
  like the way like, you know, you and I are.
• 36:17
  And so I asked her like, you know,
• 36:17
  what do people think about like AI and stuff like that?
• 36:19
  And her point of view is, listen,
• 36:21
  it's cool for like cheating on tests,
• 36:22
  but as far as like actually doing
• 36:24
  like chemical engineering work,
• 36:26
  it's fucking useless, you know?
• 36:28
  And so like, there's still a lot of room
• 36:30
  for this stuff to get in the hands of everyday people,
• 36:33
  people in other sectors.
• 36:35
  It's really, I think, focused on us right now,
• 36:39
  people in the tech enthusiast world.
• 36:41
  Yeah, that's fun.
• 36:42
  I was saying a lot of focus work with the snowstorm
• 36:47
  that hit the Northeast over the weekend.
• 36:50
  And I had some downtime on Sunday
• 36:52
  and spent an hour and a half spinning up ClaudeBot
• 36:55
  and making sure that it was secure behind Tailscale,
• 36:58
  behind a firewall and all of that stuff.
• 37:02
  And then I've been slowly but surely conversing with it
• 37:06
  and iterating on sort of how I'm interacting with it
• 37:10
  over the last two days.
• 37:11
  And it is insane what this thing can do.
• 37:16
  So what's the difference between Claudebot
• 37:17
  and like a Ralph Wiggum loop?
• 37:21
  Like would you say Claudebot is like significantly like better
• 37:24
  than like just doing fundamentally the same thing
• 37:27
  but in a more like narrow context.
• 37:31
  I'll preface this with I'm not an expert on the intricate differences between Ralph Wagon,
• 37:36
  which is essentially you give a, what is it, a product development roadmap
• 37:42
  with very specific outcomes that you want to complete in sort of a modular fashion.
• 37:50
  You give a task and it just basically runs and checks if it's actually completing those tasks
• 37:55
  as laid out in the PDR.
• 37:59
  And once it's done, it's done.
• 38:01
  With Claudebot.
• 38:02
  And I tried Ralph Wiggum just to play with it for one thing
• 38:07
  and didn't really, I wasn't really able to get it working.
• 38:14
  So my personal experience, I found a good secure setup
• 38:18
  for Claudebot on Sunday.
• 38:20
  I was successfully able to set it up.
• 38:22
  I think the one common thing they have is like
• 38:24
  with ralph wiggum you're storing context in the task files so like it has context that it can
• 38:31
  constantly pull from after it clears context and that's what i think claude bot does as well just
• 38:36
  in a different way and then from what i understand clawbot the open source projects this has a bunch
• 38:42
  of sort of overlaid integrations that's been that have been built out over the last month it's still
• 38:47
  a really uh nation project but um i don't know just intuitively for me if somebody is not
• 38:53
  Highly technical, technical enough to be dangerous and get in the terminal and do some stuff with some direction.
• 39:01
  It's been a much smoother experience for me personally.
• 39:05
  You guys should check it out.
• 39:07
  I've definitely I've I've heard really interesting things about it.
• 39:10
  I think the biggest interesting thing is like, unlike any other like system, it's it's not user driven, you know, which is very, very valuable.
• 39:19
  What do you mean by that?
• 39:20
  My understanding of autonomous.
• 39:21
  yeah well it's it lets its loop run like regardless of what you're doing with it so
• 39:26
  it'll always be checking your dms it'll if you ask it to like keep an eye on this or that you
• 39:31
  know it'll constantly just keep an eye on this or that and let you know when things like when you
• 39:34
  need it like get surfaced but alternatively from everything else you know for ralph wiggum i have
• 39:39
  to set the like checklist up and then it executes and the thing the task i was running here just
• 39:46
  like execute like finished iteration 38 you know and now it's done that's time to check the work and
• 39:51
  give feedback and go on. It's just
• 39:53
  a bigger loop, but it's still a
• 39:55
  user-driven loop. But the difference
• 39:57
  between CloudBot is it's not user-driven.
• 40:01
  Yeah,
• 40:01
  this is clearly the direction that development is going
• 40:03
  in, right? I think people
• 40:05
  are going to rely on this stuff to a crazy
• 40:07
  degree. I spent like five years
• 40:09
  from now, some narrative loop like this.
• 40:12
  Yeah.
• 40:13
  Like I said, no enough to be
• 40:15
  dangerous, which means no enough
• 40:17
  to know that I shouldn't just give CloudBot
• 40:19
  everything. So I've been very
• 40:20
  restrained on
• 40:22
  the access that I can
• 40:24
  access to cloud code and I can
• 40:27
  interact with it via
• 40:28
  Telegram.
• 40:31
  It's actually funny. We had a long
• 40:33
  conversation in the Magnolia
• 40:34
  all hands yesterday
• 40:36
  about the security concerns of a lot of
• 40:39
  these AI tools.
• 40:41
  We handle PII.
• 40:43
  We definitely do. We can't be turning
• 40:45
  on, I think it was, the bypass
• 40:46
  permissions on
• 40:48
  like, you know, we can't be piping the stuff in.
• 40:51
  And that's just, that's for the folks
• 40:54
  who don't have production access, you know?
• 40:56
  And so we have to add to our compliance documents,
• 40:58
  like, okay, you know, we're going to use AI,
• 41:00
  but this is how we use it.
• 41:01
  We're going to use it in these like, you know,
• 41:02
  specific environments where, you know,
• 41:04
  it doesn't have access to all this other stuff.
• 41:06
  It's actually kind of like, kind of interesting
• 41:08
  because we're on the side, I think,
• 41:11
  of we care about security a lot.
• 41:13
  You know, we care about making sure
• 41:14
  that this data isn't going to go in the wrong direction.
• 41:16
  and I can only imagine how the companies or other places like, you know, that lack of, you know, security.
• 41:22
  And I'm just really looking forward to like, is 2026 going to be the year of just a lot more hacks, you know,
• 41:27
  like across the whole industry, because most people are just leaving like their credentials and just, you know,
• 41:32
  like white text, right? Plain text. And it's going to be a great year for hackers, for black hats.
• 41:39
  Yeah. Yeah. Getting familiar with .env files is very important, I think, for a lot of people out there.
• 41:46
  oh yeah well like how do you think this changes um
• 41:56
  a company building right like we mentioned like headcounts and all that like how
• 42:00
  like i guess to phrase this another way like how far do you think you can go
• 42:06
  um with a headcount under 20 at magnolia very far um but
• 42:15
  But when I came to, when I moved to New York last year, I moved into this like hacker house, right, called The Residency in Chinatown.
• 42:23
  It was awesome.
• 42:24
  You know, it was my first time living with a lot of like, you know, kind of deep like Silicon Valley tech folks who are really just, you know, in the grind and genuinely a great time.
• 42:34
  And I remember something that they always pointed out about me that I'd never heard anyone refer to me about.
• 42:38
  They always said that I was a very like agentic person, you know, and I asked them, like, what does that mean?
• 42:44
  Is this like, you know, an AI thing?
• 42:45
  they're like no no it just means that when you see a problem you do something about it you know
• 42:50
  and i i've noticed that in all my founder friends that's very much who they are when they see a
• 42:56
  problem they do something about it uh and i think fundamentally what this is going to allow uh in the
• 43:03
  long run is it's going to allow that barrier entry so if you just want to do something you can do it
• 43:08
  you know i you see these tweets online where you know being an idea guy is like now the alpha you
• 43:13
  know? Um, and I, I do kind of agree with that. I think things have changed now to, to the point
• 43:18
  where it's really just a battle of willpower, you know, like how much willpower do you have?
• 43:22
  It's always been a question of willpower, but sometimes, you know, like people are able to go
• 43:28
  really, really far in the wrong direction, you know, and that's detrimental to them. That's part
• 43:31
  of the risk, you know, are you going, uh, are you waiting until the right moment for your thing,
• 43:36
  or are you being boneheaded in whatever direction you're going in? Um, and I think with, with like
• 43:42
  the advent of these tools, it's going to specifically make it a lot easier for people
• 43:46
  to iterate and get to the stage that they want to get to, right? Like being able to launch a product
• 43:52
  and then being able to test the thesis, you know, out, uh, while also being able to maybe get more
• 43:56
  information on what it is they're trying to do. You know, like it's, it's never been so easy
• 44:00
  to get feedback on like the greater, uh, like context of your space than asking, you know,
• 44:06
  like a research bot, Hey, spend like three hours researching anything that has to do with this
• 44:11
  space here and give me like a competitive analysis of these different companies and how they're doing
• 44:16
  and you know what flaws might be um it's never been so easy uh and i think i think that does
• 44:22
  change like the landscape of business quite a lot in the sense that it just we're reducing
• 44:28
  information asymmetry you know and that lets people make more uh well-informed decisions
• 44:35
  while also allowing them to like get more done.
• 44:40
  You know, before I think the things that limited founders
• 44:43
  were very much in the resource like division,
• 44:47
  you know, like how much money you have, obviously,
• 44:50
  but also like what your talent pool is,
• 44:53
  what connections you have.
• 44:55
  But with modern AI, if you're simply creative enough,
• 45:00
  you can overcome a lot of these problems.
• 45:01
  Like I've been, you know, for this launch here, right,
• 45:04
  we're fundamentally a B2B product. I have to talk to people in order to like, you know, get them to
• 45:08
  like to be aware of our product. I've just had Claude, you know, just running against like Twitter
• 45:14
  and Chrome and say like, hey, you know, look at everyone who follows me, you know, are they working
• 45:19
  on something that makes sense for like integrating with us, you know, and create like a leads list
• 45:24
  here and then rank it in terms of like how well I probably know them, you know, and that is
• 45:30
  something that wasn't possible, you know, before that's, that's reducing information asymmetry here.
• 45:36
  There's a lot of things you can do with that. Um, but really that's not like, it's, it's not a
• 45:41
  question of like how smart is Harsha. It's just a question of how bad does Harsha want it? You know,
• 45:47
  and I think now more than ever, uh, that, that barrier is lowered, you know, and I think that's,
• 45:53
  that's fantastic, frankly, for innovation as a whole.
• 45:56
  I completely agree. I think you put that beautifully. People with agency are going to thrive in this world. And I'm not so sure that I'm convinced that it's going to be the job killer apocalypse that it's being marketed as. Certainly it's going to have an effect on the jobs market and will displace a lot of jobs.
• 46:16
  But I think on the other side of things, I think it was Aaron Levy, Aaron Levy, Aaron Levy, the guy from box.com.
• 46:24
  He put out an article basically saying like Jeven's paradox compute is upon us.
• 46:30
  Like you think just because you're becoming more efficient with these resources, in this case compute, you're going to use less of it.
• 46:37
  But what we'll actually find is that we just use more and more.
• 46:39
  And by extension, one could speculate that that means that the potential jobs that exist in the future are way more bountiful than they currently are.
• 46:55
  And I do think, to your point, we do need some sort of social resurgence of individual agency and creativity to emerge, to make that a reality.
• 47:09
  And there will obviously be a period of time where we're sort of reacting and getting comfortable with this change.
• 47:20
  But I think people begin to realize, like, OK, this is what these tools can do.
• 47:25
  If I just start thinking in a different way and thinking more creatively and actually getting off my ass and typing into these LLMs, maybe something will happen.
• 47:36
  And I think that'll have a net positive effect.
• 47:39
  on the job market. Maybe that's naive. Maybe that's too optimistic. Maybe it's a pipe dream, but
• 47:43
  I can see it playing out that way.
• 47:46
  Let me ask you a question. Before the power drill,
• 47:52
  how did people fasten
• 47:55
  two pieces of wood together? Before we were able to drill things in,
• 48:00
  how did we screw drills in?
• 48:03
  Do you know the answer to this?
• 48:04
  uh that spurs even exist was just like hammered i think yeah but there were there were screws it's
• 48:13
  like this like tool it's like a rod that comes out with a crank handle right um and there's two
• 48:19
  there's two parts of it one is you have to drill the hole first right because it was just you know
• 48:23
  it was too high torque to go through the wood and then you had to put the the long screw all the way
• 48:28
  in you know and this is only really feasible for for long screws right for anything smaller you use
• 48:33
  a nail. But it was a pretty tedious job, you know, frankly. And this was like a specialized
• 48:40
  part of construction as a whole. And whenever like my friends like, you know, mentioning,
• 48:46
  you know, they're worried about AI, like taking everyone's jobs, making everyone like homeless,
• 48:49
  right? I bring this up, you know, we're long past that era. And no one even remembers that tool
• 48:55
  anymore. Everyone only thinks about how easy it is to take your drill and screw it in. And I ask
• 48:59
  people like, do people still have homes? You know, do we still build houses? You know, are there,
• 49:04
  is there a lack of construction jobs right now? No, there's actually a huge boom of construction
• 49:10
  jobs because these tools allow people to build more faster. You know, it just increases the
• 49:16
  economic output of them. And yes, when, when electric tools, especially the corded ones and
• 49:21
  the cordless ones came out, you know, people who like specialize in, you know, like doing this,
• 49:26
  they got displaced. Absolutely. And that probably sucked for them, genuinely. My family and the
• 49:33
  things that my family, like my mom and dad have done over the years, they've been displaced
• 49:36
  multiple times. I've been displaced. But economic displacement is a very normal part of any kind of
• 49:41
  healthy economy. All it means is you got to switch gears. You got to do something else.
• 49:47
  It doesn't mean you lose your job forever. People oftentimes learn a lot of skills that you can use
• 49:52
  somewhere else. That's a very normal, healthy part of the economy. But the part I want to
• 49:56
  to forget is like the economic output of an individual does compound think about all the
• 50:03
  all the hand cranks that are saved now every time you screw in a drill you know and i think the exact
• 50:08
  same thing happens with ai right now building all these like products and all these like you know
• 50:13
  simple like sass and limitations are you know hundreds of million dollar like you know enterprises
• 50:18
  that don't necessarily have to be they don't have to charge such massive amounts now people can
• 50:22
  spin up their own competitors to all these massive SaaS products.
• 50:26
  And now that's a good thing for everyone.
• 50:29
  The costs go down for everyone as a result.
• 50:32
  And I do think that not even long short to mid we going to see a lot of economic boom that comes out of this Now that comes with a caveat and everyone waiting for this in the AI world Will enterprise actually pick up
• 50:48
  tooling that reflects a lot of these AI things, right?
• 50:52
  That's what everyone's waiting for
• 50:54
  in the meta of the AI world.
• 50:55
  And if enterprise is not able to pick up
• 50:58
  direct AI tooling to fund the AI bubble,
• 51:01
  then the bubble does pop.
• 51:03
  And we're going to have to wait
• 51:04
  for our electric drills a little bit a little bit longer but you know well dig into that a little
• 51:10
  further what do you mean by um enterprise figuring this out and getting the tooling right yeah so i
• 51:16
  mean okay so right now like the ai world's in a very specific spot we put a lot of money in to a
• 51:21
  lot of these like you know data farms and a lot of these server farms you know ramp prices through
• 51:25
  the roof and you know gpus are pretty much hard to like impossible to find all that stuff um the
• 51:30
  The thing is, you know, Microsoft and, you know, OpenAI and all these folks only have so much cash to burn on this stuff.
• 51:36
  And so they've been utilizing a lot of really interesting economic tools to fund a lot of these projects.
• 51:44
  A lot of these tools, I think the term I've heard from The Economist is they are economic innovations, which is, you know, always a great thing to hear.
• 51:51
  And so everyone's pretty sure that if these server farms don't work out, it's going to be a relatively big pop.
• 52:03
  And we've already seen a big pop in these huge hype-driven investments.
• 52:09
  Like EVs are a huge one.
• 52:10
  Regardless of how you feel about EVs, that project is kind of dead right now, mostly due to the death of government subsidies.
• 52:18
  and some companies, some manufacturers spent billions of dollars building battery plants in
• 52:24
  the US, billions of dollars that they're not going to get a good return on. Those same lenders are
• 52:31
  the ones backing the AI bubble as well, building a lot of these data farms. And so everyone's just
• 52:37
  kind of like waiting with bated breath in 2026. Does enterprise actually pick up the bill and
• 52:42
  start paying significant amounts for a lot of these AI tools. You know, people kind of feel like
• 52:48
  retail is tapped out for AI. And people are trying to figure out ways to monetize retail,
• 52:54
  you know, more so. I think like the tech enthusiasts are happy to pay the $200 a month,
• 52:58
  but most retailers are just happy with free ChatTBT. And if not free ChatTBT, free DeepSeek,
• 53:04
  you know, so ChatTBT is going to have to stay free. And so, yeah, right now, everyone's looking
• 53:10
  for what the next thing is. And, you know, the initial reports are, frankly, enterprise isn't
• 53:15
  getting that much value out of AI tooling. You know, if anything, they're actually losing money
• 53:20
  because AI tooling is kind of expensive for them. So we're going to have to see, like, you know,
• 53:24
  what can a lot of the big guys do to actually change this? And if some decisions that they do
• 53:29
  don't make that much sense, like implementing ads in ChatGPT or, you know, making it harder to use
• 53:35
  cloud code outside of cloud code it's due to this like large-scale macroeconomics thing they they do
• 53:42
  need the enterprise angle to pay off um and so that that that's what i mean by that spot you know
• 53:47
  there's a lot going on over there you know i i hope the bubble doesn't pop but it could
• 53:51
  no i mean as somebody's been using claude um pretty heavily over the last six months i look
• 53:58
  at my usage look at my bill and i'm like just like knowing how hard i've hammered it in some
• 54:03
  days, I'm like, I'm definitely not paying enough for this as being subsidized to some extent.
• 54:07
  And it's like, how long can that subsidization persist before the reality sets in and you don't
• 54:15
  have a deeper sort of investor pockets to dig into? Which is interesting. It's like, are we
• 54:24
  going to see a dot-com bubble like pop and sort of valley of despair for a few years and the market
• 54:33
  eventually figures it out between now and 2030.
• 54:35
  Like, okay, here's how to do it.
• 54:36
  And not only do it, but do it in a way where you can make money.
• 54:39
  It actually makes economic sense and consumers are paying for what they're consuming.
• 54:44
  And the producers of that product are actually making money because that's all it comes down
• 54:49
  to.
• 54:49
  You have to make money at the end of the day.
• 54:51
  But then you also have this existential geopolitical angle to it where it has been deemed existentially
• 55:01
  critical to the future of freedom in the digital age, America versus China. So you do have daddy,
• 55:09
  daddy, uncle Sam, who seems pretty motivated to step in and back it up as well.
• 55:15
  Oh yeah. Oh yeah. There's, I, it's, the only thing I know is that technology advancements
• 55:22
  always come in waves, right? There's, there's always a big one and the crashes, that's the
• 55:28
  Same thing happened with EVs.
• 55:30
  We're on, I think, the third or fourth waves of EVs.
• 55:32
  People forget there was a wave in the 80s during the oil crisis.
• 55:38
  There was a wave in the mid-2010s, and there was a recent wave that just crashed there.
• 55:44
  Now, the Ford Lightning is going to go and throw in a gas generator.
• 55:48
  That's about how that's going.
• 55:49
  I might buy it in complete transparency.
• 55:51
  I think that's awesome.
• 55:53
  Will AI have a crash?
• 55:55
  it would follow the pattern of every other technological advancement if it did yeah just
• 56:01
  using pure heuristics it it would make sense that it should at some point um bringing this back to
• 56:08
  bitcoin what um what are your thoughts on the current state of bitcoin both from a economic
• 56:15
  technical adoption standpoint and then uh the industry the state of the bitcoin industry um
• 56:22
  where it is right now where it could be 12 months from now particularly if uh magnolia takes off
• 56:29
  or a changing i don't know if you saw but uh optx hash has uh a bip now right which is interesting
• 56:38
  i think it came out like like an hour or two before i hopped on um so that's interesting
• 56:42
  um are like basically at the high level of bitcoin right you zoom out as much as possible
• 56:48
  So there's where is Bitcoin going at an enterprise level?
• 56:53
  Where is Bitcoin going at a legal level?
• 56:56
  And then where is Bitcoin going at a tech level, right?
• 57:00
  And when it comes to the enterprise plays, there have been some interesting plays over 2025.
• 57:05
  But by and large, the value of Bitcoin has always been the actual buy-sell price of Bitcoin.
• 57:17
  People are trying to take advantage of that in really interesting ways.
• 57:21
  Like MSTR, you know, has always been like, how do we acquire more of this on the bet that it's going to go up?
• 57:27
  And, you know, we're going to use all these different like devices to get that, you know, to get that going.
• 57:32
  Right. And I think Bitcoin treasury companies are always some version of that.
• 57:35
  Right. You know, I'm not trying to say if it's good or bad, but that's they're playing on the price of Bitcoin.
• 57:40
  And that seems to be like more of the enterprise play when it comes to Bitcoin.
• 57:44
  When it comes to the legal aspect of Bitcoin, it's pretty much just riding the coattails of the genius building the Clarity Act.
• 57:51
  I think all those conversations are about stablecoins, but Bitcoin just kind of follows along.
• 57:59
  It's just hanging out in the same conversation, but not the part that everyone's fighting this way or that way.
• 58:07
  People are fighting over the stablecoins and the yield from the stablecoins and what to do with that.
• 58:14
  And then when it comes to the tech side of Bitcoin, the thing is, Bitcoin has had a lot of tech improvements over the years.
• 58:23
  Almost none of them getting serious adoption, right?
• 58:26
  I mean, not just Lightning.
• 58:27
  Lightning, when I mean serious, I mean serious compared to Bitcoin, right?
• 58:32
  Lightning is there and has been awesome for a while.
• 58:35
  But now we've got Spark and ARK and a lot of other layer twos that are coming up.
• 58:41
  And, you know, those are all fine and do like interesting things.
• 58:46
  But then there's all these like other upgrades to Bitcoin coming along, like, you know, whether or not covenants come to Bitcoin, all these different questions.
• 58:53
  But the thing is, compared to the other two things I described, the enterprise and the legal changes that are in the Bitcoin space, all these tech innovations, they do seem smaller, right, significantly.
• 59:06
  And as a developer, right, that's my entire background, I would like to see those innovations be bigger and make more of a splash.
• 59:14
  But that's just not the part of Bitcoin people are really interested in, you know.
• 59:17
  So for Magnolia, I want to follow the things that like have much bigger splashes, you know, which to us seems to be tooling and services that have to do regarding with the price of Bitcoin.
• 59:30
  You know, like pragmatically, that's just the direction that everything is in.
• 59:34
  And so if there's a way we can use the price of Bitcoin to make other people money or to, you know, like convert precoiners in, that's where I think Magnolia wants to be.
• 59:43
  For these other projects, I want to support them the best I can because there might be like some way to, you know, have a really cool like lending protocol related to some of these things or a really cool like way to bridge gaps here with Lightning.
• 59:57
  We support Lightning at Magnolia and that lets us plug in things like ARK and maybe even Spark.
• 1:00:02
  And that's fantastic.
• 1:00:04
  But I think the future for 2026 is going to be a lot like really focused on these bigger folks right now.
• 1:00:12
  I think the thing that people haven't really talked about is that retail isn't pumping in Bitcoin and not even AI.
• 1:00:20
  You know, so I think people are feeling squeezed right now.
• 1:00:23
  But enterprises are out here and they're doing things.
• 1:00:26
  So that's where McDonald's got to be.
• 1:00:29
  Yeah, there's an order of operations to Bitcoin's long term success is something I've long said.
• 1:00:34
  I mean, expressed in other ways, like store value, medium exchange, unit of account, whatever it may be.
• 1:00:38
  It seems like we're still squarely in the early phase of adoption where it's monetizing.
• 1:00:45
  And like it or not, that is where a lot of the money is going to be made at this particular point in time and where a lot of the activity is.
• 1:00:53
  But to your point about the second layers to that, I'd be interested to get your thoughts on that.
• 1:00:56
  Like how much do you think people in the industry should even market them?
• 1:01:00
  So you think of Spark, Arc, Chami, Mintz, Liquid, these second layers.
• 1:01:07
  I think 2025 was the year, at least in my perspective, that people in the industry begin
• 1:01:13
  to realize, oh, all these things are interoperable.
• 1:01:15
  And if we can combine them in creative ways, we can eke out efficiencies and really improve
• 1:01:24
  our processes as a Bitcoin service provider.
• 1:01:28
  One thing that comes to mind is being able to automatically rebalance channels by combining
• 1:01:32
  ARK and Lightning.
• 1:01:36
  Obviously caveat, at scale you'll probably need covenants, but we're not at that point
• 1:01:40
  yet.
• 1:01:43
  Point being, the point that I'm trying to make is how much of these second layer technologies
• 1:01:48
  should even be marketed to the public and should they just be viewed as a backend infrastructure
• 1:01:54
  that app developers and product developers integrate and don't even
• 1:01:58
  don't even really market to the end user that that's what's happening in the back end.
• 1:02:05
  I think that's the second part.
• 1:02:07
  Like, well, like all this stuff is really complicated for even people in this space,
• 1:02:10
  you know, all these different like different protocol schemes.
• 1:02:13
  But to your point, it's it's becoming pretty interoperable, you know,
• 1:02:18
  and I think we're going to have like it's going to have its moment when there is a killer application
• 1:02:23
  somewhere in in this like whole you know ecosphere but all these different ecosystems work together
• 1:02:30
  quite effectively so let's say you know we partner with like arc for example and some of those
• 1:02:34
  partners with spark and both of them are trying to like interact with some application on like
• 1:02:39
  some layer two that is neither of those right but everyone has bridges we can all interact with the
• 1:02:44
  application uh just like quite fine right and i think that is a world like worth kind of like going
• 1:02:51
  for you know like if that explosion occurs everyone wins especially when it's interoperable
• 1:02:56
  um how soon will that well how soon will that it will take to get there i have no idea i think we're
• 1:03:03
  firmly past the era where there's like like a lot of you know frankly just like scam projects out
• 1:03:08
  there i think that the guys that are left are like probably pretty good at this point um i've
• 1:03:13
  been paying attention to i think like a couple days ago i checked in on a couple projects that
• 1:03:17
  I was always a little skeptical of just to see what they were up to. I hadn't heard about them.
• 1:03:23
  And I don't think they're doing super well, which is good to see that they're not going to get
• 1:03:28
  involved in any of our conversations. But that's a good thing because now I can kind of generally
• 1:03:32
  start to trust like, okay, we've weeded out a lot of folks from 2025. You guys are left. They're
• 1:03:37
  still building, they're launching, and they're doing things. Let's see what they come up with.
• 1:03:41
  And if they come up with something good, hopefully there's a way we can all bridge over there with
• 1:03:45
  minimal cost, you know, then everyone really wins out on, on one person's game.
• 1:03:50
  You know, I think that's like kind of the strategy other projects have really taken,
• 1:03:54
  uh, when they had like, you know, a lot of, uh, you know, different projects that needed
• 1:03:59
  to interact with each other for whatever reason.
• 1:04:00
  I think that was a winning strategy, you know, as long as it's all based on Bitcoin, I'm happy.
• 1:04:05
  Yeah.
• 1:04:06
  And considering where the price is too, I got actually, I get a, I get a bit of guilty
• 1:04:12
  pleasure out of the price consolidating.
• 1:04:13
  like it has some people freaking out and i think it's been a trope for at least uh as long as i've
• 1:04:19
  been in bitcoin like building happens in the bear markets not to say that this is a bear market like
• 1:04:24
  we've experienced in the past but it feels like that to many particularly many who are relatively
• 1:04:28
  new to bitcoin and if you just zoom out and look at all the progress that's been made whether that's
• 1:04:37
  uh in the business sector the industry being built on top of the protocol or the protocol and
• 1:04:43
  And the second layer protocol is being built on top of it.
• 1:04:46
  It's been pretty massive from my perspective.
• 1:04:49
  It's the signal in the sea of price noise that exists in my mind.
• 1:04:54
  Yeah, I think if you're someone who thinks that this is a bear market,
• 1:04:59
  then, you know, just don't worry.
• 1:05:00
  Like, you know, you'll lose that idea over time.
• 1:05:03
  Every now and then I like see a tweet like the price is down.
• 1:05:06
  And it's like actually higher the last time I knew the price.
• 1:05:10
  I've been in Bitcoin long enough to
• 1:05:15
  what even is the price right now? I have no idea.
• 1:05:18
  It's an 88. That's funny.
• 1:05:22
  I think the last time I saw it was an 86. I'm not too
• 1:05:26
  upset. 88. Lucky.
• 1:05:29
  Lucky number 88. This has been
• 1:05:33
  awesome. What parting thoughts, messages,
• 1:05:37
  ideas around Magnolia that we haven't touched on
• 1:05:41
  do you have for the freaks out there?
• 1:05:44
  Yeah, I mean, like, really, what we're trying to do is just make it easy
• 1:05:47
  to get banked, right? Like, to get you banked, to get your customers
• 1:05:51
  banked, so that you can just implement whatever interface you want, right?
• 1:05:55
  Like, you can go as far as creating a neobank if that's what you want to do
• 1:05:58
  with Magnolia. If you want just auto-pay, or even just auto-DCA,
• 1:06:03
  you can do that with Magnolia. Like, what we're trying to do here is just
• 1:06:06
  create as much financial tooling as possible for the builders. And if there's something that,
• 1:06:11
  you know, that you would like to see that you don't think we have, reach out, you know,
• 1:06:16
  maybe there's a way you can do it. We're, I'm a very customer first person, right?
• 1:06:21
  If people tell me they want to go to EU, we'll go to EU. We figured out a way to do it. People
• 1:06:24
  told us that they wanted to go to all 50 states, you know, and that's what they wanted to focus on.
• 1:06:29
  They don't want to focus on geofencing people from like Alaska, right? Or Indiana. So we figured
• 1:06:34
  got a way to do that you know um we're very like customer focused and i think that's that's really
• 1:06:39
  worked out for us and our launch is kind of based on what people have really wanted so you know if
• 1:06:45
  you're building something and you've got a customer base that you want to to like you know connect to
• 1:06:50
  tradrails for whatever reason reach out there's probably some way we can help you like genuinely
• 1:06:55
  i think all bitcoin apps can can utilize something like this um yeah and even even if uh you just
• 1:07:01
  like say hi also we love to just talk with people i love chatting with people bitcoin
• 1:07:05
  y'all are my people you know so i just love to talk oh yeah well thank you for your time today
• 1:07:11
  congrats on getting this over the line i know it's been a lot of a lot of hard work on your end in the
• 1:07:17
  in the uh end of your team there at magnolia so looking forward to seeing you guys really get out
• 1:07:23
  there and uh we'll have to do this at some point again in the future dude till next time man thank
• 1:07:28
  Thank you so much for having me on and I'll catch you when I see you.
• 1:07:32
  Peace.
• 1:07:33
  Peace and love, freaks.