Why Legacy Companies Can't Adopt AI with Luke Thomas

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Luke, this is a great pleasure for me.

(00:03):
I've been waiting for this episode for, I think, years now.
Well, I'm definitely excited to be here.
And to all your listeners, I'm sorry.
There'll be lots of rants, lots of unhinged, maybe unhinged thoughts.
But I'm really excited to be here.

(00:23):
Well, you've had an open invite to the show for, I think, at least two years now.
And you said, you've been saying for two years when the time's right, and I feel like I have something to say, I will let you know.
And that happened a couple of weeks ago.
And here we are, thinking of how to start off this episode.
I think it's important for the audience to get context.

(00:44):
Luke has been somebody who's helped me immensely behind the scenes here at TFTC, thinking about how to position the brand and how to increase our SEO.
So not only that, how to lean into this AI revolution.
And I think just to start off the conversation, let's get a little history of your career, what you've built, where you've worked, what you're working on now, and what people may learn if they stick around for the whole episode.

(01:15):
Yeah, sure.
So I'll do my best to keep it short.
Yeah, so a little bit about me.
I grew up in the middle of nowhere in Maine.
And up until when I first went to college, I had always lived on a dirt road.
I grew up in a small town in Central Maine.
It was about 800 people.
I think half of the town was on some form of general assistance.

(01:39):
I was homeschooled growing up.
I attended a mixture of homeschooling, private schooling.
in high school, I ended up getting into web design because I wanted to customize my MySpace profile.
I thought it would be cool and maybe more girls would talk to me.

(01:59):
And so that's what I did. I ended up basically devoting, call it half of day. So I would
homeschool for half of the day. And then for like the other half of the day, I would go to what's
called a vocational technical school. And I was in a class with a bunch of kids that wanted to
create video games. And I thought web design was cool. And I thought maybe there's something there.

(02:20):
Little did I know that it would basically influence the rest of my life to date. And so, yeah, in high
school, I kind of learned how to code. I wasn't very good at it, but I started creating websites
for small businesses. I ended up graduating. I went to school one year in Boston because I thought
sports management would be really fun. Turns out that spending tens of thousands of dollars a

(02:45):
year for phys ed classes. Probably wasn't the best idea. So I ended up going to school for marketing
and paid my way through school, kind of did the consulting thing, ended up graduating college
and wanted to join the startup thing. And so long story short, my wife and I, the second we

(03:05):
graduated, we moved to Boston. I started working in early stage tech companies. I originally joined
as an engineer, quickly realized I wasn't good at writing code. But I always viewed writing code
as a means to an end. For me, it was like, I loved to grow businesses. And so I ended up
kind of falling into this industry called like growth, which is how to think about marketing

(03:31):
through a programmatic lens, like a programmer would. And anyway, did that for a few years,
you know, learned SEO, learned a bunch of, you know, internet marketing techniques and
ended up falling down the product rabbit hole, largely because, you know, when you build,
when you're building like a startup that is, you know, like a digital product, a lot of the growth

(03:53):
work, so to speak, that you do ends up actually being product work. It's like, how can you create
a product where these, these like flywheels that help you grow at a low cost? So, you know, for
those who are familiar with like Dropbox, right, it's like people sign up for Dropbox to get free
space. If you want more free space, you share it with a friend. I became very obsessed with that
and ended up doing that for a while. In 2019, I decided to make the plunge, decided to start a

(04:19):
remote work software company. I saw a bunch of these fully distributed companies building their
own intranet is the best way of describing it. Zapier and Stripe and all these well-known
remote companies had kind of hand-rolled their own system. So I started that and then six months
later, COVID hit. And so we experienced a lot of growth. I did that for about two years,

(04:43):
but I ended up shutting the business down because I didn't see a clear path to scaling
on the revenue side in particular. So I ended up joining Zapier for about three years
and went super deep there on AI and automation. And so, yeah, in many ways, a culmination of
what your listeners will probably hear is maybe a little bit more of a deeper dive on AI,

(05:07):
what it looks like, how it's changing things. I also left about two months ago and I started
a new company called Formable. We basically are making it easy for businesses to respond
to inbound leads way faster. And so for a lot of businesses, you have a contact form on your
website. When someone fills it out, they usually have a high level of intent. You want to get back

(05:27):
to them as fast as you can. Well, we're building simple and easy to use tools that make it easy
for businesses to get back to their potential customers way faster using AI and automation
and a better data collection or form experience. So long-term though, we want to build kind of a
new school marketing automation platform, which is just like making it easy for businesses of all

(05:50):
shapes and sizes to market in a way where you're leveraging AI as much as possible in new and we'll
call it creative ways. We see a big opportunity there and we don't think the people in the space
are moving fast enough. So that's kind of the spiel. I live in the Nashville area and I have
four kids, seven and under. So life is good. But I think this context is probably useful for

(06:16):
whatever else I'm going to say over the next hour or so. So that's my story. And yeah.
Yeah. Well, like I said, like you've been helping me a lot behind the scenes, particularly
as it pertains to how we grow this business, I think over the last year through the conversations
we've had, you've made it clear to me that you're, you strongly believe that the nature

(06:41):
of businesses, how they are started, how they operate, how they scale, and how big their teams
get specifically is about to drastically change. And not only that, on the business side of things,
you believe strongly that Bitcoin is going to play a role in accelerating these businesses as well.
Yeah, yeah. So the way I kind of think about it is, the best way I can describe what I think is

(07:06):
going on is kind of we live in an age where there's kind of this like concept of like unbounded
leverage, right? So before there were certain constraints that existed in a business, it could
be in particular teams in the business. And we've entered a phase where, you know, one person can
accomplish a lot more with AI in particular. And so I think like, there's this rise, at least how I

(07:34):
view it of these businesses that understand this reality, and they work their way through
their own companies, trying to identify these areas of leverage, and then implementing solutions
or tools or processes to kind of take full advantage of this kind of unbounded leverage

(07:55):
idea, right? And so I think actually Bitcoin really fits into this thesis quite a bit, right?
Because if you think about it, you know, in my world, you know, if you're an early stage startup, you especially if you're venture backed, you have to go out and you raise money.
Right. And usually I would say usually it really depends on the strategy.

(08:15):
But I would say in general, a business goes out and tries to raise at least 18 months of runway.
Oftentimes, in some instances, it can be years of runway. Right.
And so you basically end up having a bunch of fiat sitting in your bank account.
And so if you think about this idea of where are the areas where there's leverage, it's like your balance sheet is clearly an area where there needs to be some amount of leverage as well.

(08:42):
If you're an early stage startup burning with a burn rate, you're not making as much as you are spending.
Obviously, you have to be a little bit more, we'll call it risk off.
but for a profitable business in particular I think the the ability to be a bit more risk on
here around your balance sheet in particular becomes much more interesting and compelling

(09:05):
but like if you apply that principle of like where can I kind of like tap into this like
almost like digital energy as sailor kind of puts it right obviously the balance sheet is one area
But I think AI represents its own form of digital energy.
And that can be applied in other parts of the business.

(09:26):
And so I think as I survey the nature of where things are going, the businesses of the future will kind of understand this reality and will look for areas where they can implement this everywhere.
I would say to maybe take it a step further, I actually would argue we're not seeing nearly – like we're not seeing that much yet.

(09:53):
And I would argue the reason why is because existing businesses have a lot of – we'll call it baggage, right?
They have – they made certain hires based on certain assumptions.
They implemented certain tools based on other assumptions.
And so they've created this system that is designed and aims to function in a particular way given certain assumptions.

(10:21):
But with AI, it's literally coming in and it's like really rattling those assumptions.
So if you're a business, it could mean that you need to change your pricing and packaging.
Right. And for a lot of like software companies in particular, it could represent saying, hey, we're going to like literally axe a bunch of revenue to change our business model around to be more future proof.

(10:45):
Most businesses will not do that. It could mean laying off people.
Right. And so I think we're only starting to see the beginning of this shift.
And I would argue that the tooling and the tech is far ahead of what is actually in use and in practice at these businesses. And so the constraint is not the tech. It's the existing structures in place.

(11:14):
And I saw this firsthand, and I see this firsthand where it's like, businesses, we want to do things the way we want to do it. You can sprinkle AI on top of it, go for it, have fun. You can have a CEO mandate, but it doesn't drive the necessary change or the change that will be coming.

(11:35):
And so, yeah, I think like we're kind of entering an age of where there's unbounded leverage.
Startups, when you start from scratch, or if you're building a business from scratch, you can start with a greenfield environment, but the existing business cannot.
And so that's why you see a lot of these startups that are like they're shipping code way faster than others.
They're moving way faster than others. It's because they're tapping into all that like unbounded leverage where the existing business has the structures in place that just are not they're not set up in that way.

(12:08):
I'll give you one more analogy. I was maybe too young to experience this firsthand. I wasn't working in tech, but, you know, there was a season.
And I feel like it was maybe like in the early or middle 2000s where there was this concept of like software moves to the cloud.
Right. And so you had these like, you know, code that was on prem or it was like, you know, in the confines of the four walls of the business.

(12:33):
And then like, you know, Salesforce comes along and is like, hey, it's like an address book in the cloud, whatever the cloud means, you know.
But those old businesses had to make a decision around how serious they were about this new cloud thing.
And I think the same is true now with existing even software businesses.

(12:56):
Are we going to fully embrace this AI thing?
And I just think that you're going to see a lot of, we'll call it a flippening, that takes place over the next few years.
Anyway, I digress, but that's kind of like the thesis in a nutshell.
Well, speaking of digressing, let's regress and do like a retrospective on the last one
or two years, the evolution of AI.

(13:17):
And let's get more specific.
You're talking about this unbounded leverage that it provides.
But what has happened since GBT3 launched?
Like how good are the models get?
What specifically are they good at?
And how are you and other founders leveraging them to outcompete incumbents?
And then on the incumbent side of things, let's get into the specifics of what sort

(13:39):
of confines them and makes them unable to make the necessary changes to compete.
Yeah, sure.
So I'll kind of walk you through my personal firsthand experience using AI.
So it's kind of a funny story.
Back actually at my previous startup, I think it was like, I don't know, it was maybe early
2022, late 2021.
And I chatted with a fellow founder.

(14:00):
He's kind of a bit of a brainiac.
And he's like, you got to go get an invite.
Email this guy.
and I think it was Greg Brockman, I think one of the co-founders of OpenAI,
he's like, email him and ask him for access.
So I emailed somebody at OpenAI, and they gave me access to,
I think it was either it was ChatGPT, or it wasn't ChatGPT,

(14:24):
it was GPT2 or 3.
And I started playing around with it, and it had a developer experience.
And I'll just share my screen for a second.
We'll come back to this later as well because there's some stuff I want to show you guys.
But it looked something like this.
It was a developer-first experience.
And once again, nobody ever hired me, at least for a long period of time, to write code.

(14:50):
But I started playing around with it.
And at the time, I was like, wow, this is kind of like autocomplete.
It's like autocomplete.
At the end of the day, like AI in many ways, like this is maybe a simplistic way of describing it.

(15:10):
But at the end of the day it like if it uses math to determine what the next word or set of characters should be Like that kind of that how it works in a nutshell It like it makes guesses
It's trained on data.
It makes guesses about what the next words or words should be.
And so I was like, oh, this is pretty cool.

(15:32):
And obviously, as you know, someone coming from like the content marketing and the SEO
background, I'm like, I get my head of marketing, who's like an amazing content marketer, right?
He writes blog posts, they rank in Google.
Like, we got traffic, I think we were doing like, you know, 15,000 signups a month with like a $2,000 a month marketing budget, right? And I'm like, you got to start using this to, to basically help scaffold up some blog posts.

(15:59):
um and and so he started using it and we went from like if it was like 200 a post right 200
a post like we had you know people off in yonderland writing posts and they were okay but they
weren't that great right and all of a sudden it's like uh this is like 25 30 dollars and it's like
wow that's pretty cool and so you know i used it ended up shutting you know shutting the business

(16:22):
down, joined Zapier. And then I think it was maybe like six months later, ChatGPT comes out.
One of the founders was going down the rabbit hole on AI. It's going down the rabbit hole.
And all of a sudden it's like, hey, we need to care about this. This is like a big bang moment

(16:43):
in the tech world. Everybody needs to care about this. And so they pulled this, what we're looking
back called a code red, which is kind of just like an internal, we'll call it fire drill. It's like,
we need to care about this. And here's how we're thinking about this. And so immediately, at the
time, I was a product manager there. And I started playing around with it because I knew that it was

(17:06):
powerful before, but it was like, how am I going to use it now? And at the time, I would actually
say I was a little bit of a skeptic, even though I'd experienced the power in one particular
area. I was like, Hannah, what is the scope here? What are we dealing with? How useful is it,
especially with this new launch? And so at the time, I kind of had this fun little hobby,

(17:31):
and I still do, where I would read these old books written in old English.
And I find that there's something cathartic about reading old books outside of the current
moment that we're in. And I was like, I bet you, I bet you I could use this to update the text so

(17:51):
that I can make sense of what the heck I'm reading right now. And so I ended up doing that. I was
like, oh, it's good, but it's not great. Like it would maybe make it a few sentences in a paragraph.
It would update it. And then all of a sudden it just kind of go off the rails. I was like, man,
that's kind of a bummer, but like, cool. Well, three months later they launched, I think it was
like a new model. And so I returned to that problem. And it was like, oh, wow, I'm getting

(18:16):
like full paragraphs now. And so I decided to have a friend build me a little tool where I could
update like, you know, 20 pages of text at a time. And then I decided, oh, I'll just keep going down
the rabbit hole. So I'm learning how to prompt. I'm learning how to use different models. I'm
learning how to kind of adjust things. Like, how can I make it less creative and more,

(18:37):
based on that original text with minor updates.
And so I just kept going and going, going down the rabbit hole.
And then I ended up, wow, I'll just put it up on Amazon.
Well, long story short, I think I've sold about 50,000 books in the past three years.
But I modernized Old English.

(18:58):
And now all of a sudden, it's like my world is AI.
And then at work, we launched a feature that used AI, and it just took right off.
It was this basic capability where you could ask questions.
You could embed this widget on your website and ask questions.

(19:19):
And so if business could put a widget on their website, people would ask questions, and then they wouldn't have to ping you and ask you generic questions over and over again.
So I'm like, I'm way down the rabbit hole now.
right and then it's been one of those things where i just consistently as a result of tinkering
um you know just trying things the models have just continued to get better and better

(19:40):
some models aren't as game-changing as others um but you just continue to see this
improvement being made where before you know i could ask a question of a 10-page pdf and now i
can ask a question of a 300-page PDF and get the right answer, right? Or summarize this short text,
summarize this long text. And so I'm, once again, I'm digressing, but as a result of tinkering and

(20:09):
kind of building in the space, you're able to kind of monitor things and see the nature of the
progression. And so as a founder, gosh, where do I even begin? Maybe I'll walk you through
how
yeah, like where do I even begin?

(20:31):
What's the best way of describing? Maybe I'll talk about agents for a couple minutes.
Okay. So for those who
you've probably seen the pieces in the news and all that,
there's this whole kind of idea of like AI agents. What are AI agents?
It's like, well, a lot of it's hype, but there is a reality underneath it. And it's

(20:51):
basically that agents are, I view them as almost like little digital coworkers, right? Where you
can tell them what to do and they'll do something for you and then they'll give you back a result.
So the way I describe this is, you know, if you've ever used ChatGPT before,
you give it an input, hands on keyboard, you type in something that the agent goes to work

(21:15):
and then it returns text for you to take a look at. Well, the chat experience is just one
implementation of an agent, right? It takes an input, it does a thing, it gives you back an output,
right? And this instance, it does it through a chat experience. Well, one of the big things that I

(21:39):
I guess learned is, well, that's cool. But when it's initiated by a human, there's only so like
you have, you basically have to rely on human behavior change to get the result that you're
looking for with AI, right? It's like, oh, I need to go to chat GPT. I need to remember to use it.
Like that's actually really difficult. But there's no reason why you couldn't feed

(22:03):
these little digital brains information behind the scenes where it does a thing and then it gives you
back a result and then you use it further down in some type of workflow and so zapier is a workflow
automation product um it's been around for 10 years and what we saw is like people would build
these automations right when this happens in this tool i want to do something else in another tool

(22:26):
well what we saw people doing was they were saying yeah i want that to happen but in the
middle, I want to use basically ChatGPT to do a thing for me. So, oh yeah, when I get an email,
I want to summarize the email, and then I want to insert that information into another tool.
And so agents is kind of the big topic these days for startup founders. And once again,

(22:55):
there's a decent amount of hype, but there is a very real reality. And it's that these
these digital employees um kind of do work for you and so in the old world when you're building
software you kind of it's like you assemble tools in a toolbox right and um and then they give you

(23:19):
the toolbox you pay for the tools in the toolbox and then they say have fun we'll see you later
maybe we'll do a little training with you um but like ultimately it's up to you the human to um
pick up the tools and use them. Well, with AI and with agents in particular,
you can basically almost have this intermediary between you and the tool.

(23:41):
So now there's like these little digital employees that can kind of start doing work for you.
And so all of a sudden, the human labor that I would need to do, whatever I was doing before,
I don't have to do as much of it because the agents are doing work. And so all these startups
are looking for ways where agents or these little digital employees can basically be plopped into a

(24:07):
workflow or injected into a business process. And so you're just seeing this explosion of
startups looking for places where they can just stuff a little digital employee.
And so every time, though, that Grok or OpenAI or Anthropic releases a new model, the digital employee gets smarter.

(24:33):
The digital employee couldn't do a thing before, but now it's like it received better training and it's able to do some new things now.
And so every time a new thing is released, it gets smarter and smarter.
So anyway, startups, I would say in general, are looking for ways to use and kind of get these agents to do work, to augment human labor, to take work off people's plate.

(25:00):
And the agent is really, really good at what it does.
It can contain a bunch of information about your business.
like it can literally have libraries of content in its brain or in its memory um and it can
reference that so that it does the job in many instances better than the human because the agent

(25:22):
doesn't forget the agent keeps keeps all this stuff in mind and so anyway it's it's kind of
nuts honestly like i've never in the you know decade plus i've been in tech you know there's
been some shifts and all this. This is easily the biggest shift I've ever seen. And it's like,
and we're still really only in the beginning stages. And so if, you know, as a start, like,

(25:44):
this is part of the reason why I'm working on a new thing is because it's just like, I can't sit
by and watch this happen. Like I want to build stuff, you know? So anyway, I don't know if that's
useful, but that's how startups are using AI and what's happening in the market right now.
Yeah. No, I'm thinking of, because you've heard it, startups are even more mature companies. There's always like one or two employees that know the code base extremely well. And they're like a key man risk. Like if they were to leave, they know how everything works. And now you can sort of automate that task. You know, that risk doesn't isn't as high as it was because you have these agents now that can just.

(26:26):
I'll give you an example. So, you know, we have a code base, the new product that we're building has a code base. I don't go in and I don't really know the code base. I mean, I could like fiddle around to see what's going on and I can roughly read what's going on.
But now I can just get the code on my machine, point Cloud Code at it.

(26:48):
And so yesterday I was playing around, and in a startup, you kind of need to create product documentation and help docs so that if someone needs to find their way around the product, there's like a little repository or a series of web pages that you can go to to say, how do I do this?
How do I do that?
I pointed Cloud Code at this and said, hey, look at my code base and literally write the help docs for me.

(27:12):
And it did it.
And it did like five minutes.
And so it wrote like 20 super comprehensive help docs and like literally walked people through how to access certain features because it has knowledge of the code.
I will tell you, there are dedicated people in companies that literally do this all day.
it's just like that's just that's just one example that's just one example

(27:38):
yeah well and then i guess shifting this towards the companies that already exist have these code
bases have these teams and these processes like let's dig into how they're about to get disrupted
and how if at all they can reposition themselves to not get blown out of the water
Yeah, yeah. So this is a fun one. Okay, so I'll kind of walk you through what I'll call the roadblocks or the constraints are. So the first is we'll maybe start with the people. So I think there's two pieces. The first is there's a lot of people that are just like, I don't care about AI and I don't care to know and I'm just going to keep collecting my paycheck and screw what the CEO says.

(28:25):
you know this mandate i don't care like i i don't care right and so i would say that is the fool
um that is a foolish approach um and so you have you know you could have
i mean depends on the company right but i mean even in a tech company you might have 20 percent

(28:45):
of your employee base that thinks that way right because if you just think about it there's kind of
this idea of these technological adoption curves. There's the innovator, the early majority,
the late majority. Some percentage of your business is the late majority.
They're at the tail end of adoption. They're not at the cutting edge.

(29:07):
And so they are going to not necessarily fight the initiative, but they're definitely not going
to look for ways to try to improve. And in some instances, they will kind of block things and be
a naysayer. So that's part one. Part two is the organization itself is structured in a particular

(29:31):
way. Certain people have certain roles and responsibilities, and they like that, right?
People like having consistency. People don't generally like when their job description gets
torn up and they're like, you're doing something new. So when it comes to some of these new tools,
there are new tools that because you have these agents and AI doing more of the human labor,

(29:57):
they rightfully notice that this is disruptive to my job.
It's kind of like, I see these stories on the internet where you have these people in IT or
whatever and they're like they're training their replacement that's overseas it's like it's kind of
like that you know they they know they're like something's wrong here and so they will fight

(30:20):
right they'll stonewall things they will um say we don't need the software they'll that once again
kind of be a naysayer and i like for what it worth i i uh i have a certain degree of empathy for that situation But I don think the answer here is letting the big tidal wave smack you in the face

(30:45):
Don't go to the beach and let the big tsunami run over you.
I just don't think that's wise.
So that's just the people.
That's just the people.
Now, we'll talk about the tech or the assets that the business has, right?
So in my world in software, it's code, right?

(31:07):
The asset is the code in the code base.
Well, for a lot of businesses, unless they were started like recently, they have a lot of baggage tech debt, right?
And so when someone wrote that code, they built it in a particular way based on certain assumptions, as I mentioned earlier.
Um, well, if you, if, if you want to make a meaningful change and like put AI like deeply into an existing feature, you basically have to rewrite the thing from scratch.

(31:41):
Right.
And you have to build it in a way where like you're throwing out a lot of like, you know, old code and like, you know, you're making massive investments.
And like tech debt is not something that the companies generally want to invest in.
They like revenue generating features, revenue generating code that's written.

(32:02):
Right. And so it's easy to kick the can, kick the can, kick the can.
So that's why if you look at a lot of these software providers, if they've done AI, it's very surface level.
It's like, oh, here's a little copilot.
Right? Here's a little co-pilot on the side. You can load it, and then you could talk to it,
and it will answer some questions, but it's kind of useless.

(32:24):
But the whole rest of the experience is the same thing. So that gets into maybe part three,
design. When you onboard people, and when you have been around for a long time, you have a
user experience that has been built in a particular way, and your customers are used to it. And they
might be a little bit of a, I wouldn't say the naysayer, but let's call it the late majority.

(32:48):
So if you go in and you start ripping, like you just start making huge changes to the user
experience, they are going to just moan and complain because their workflow was disrupted.
Once again, rightfully so. So it's like, wow, well, to do some type of change, like let's just

(33:09):
Sprinkle a little bit of AI on top of what already exists.
And that's from a design perspective.
Moving on, pricing and packaging.
Businesses love to sell features.
They love to sell seats.
Seats is probably my favorite example.
Well, if a business, like in the markets that we play in, if a business makes a ton of money selling seats,

(33:35):
and AI comes along and says, you don't need as many employees.
Well, A, like they literally, it's kind of, it competes with their own business model.
Right. And so are you going to cannibalize existing revenue for the ambiguity of the unknown?

(33:56):
I don't think they will unless they're backed into a corner.
um and so in general you know start businesses like to grow and ai is a big enough shift where
you kind of have to rethink your your um your your pricing and packaging you have to shift away from
um selling seats i think you have to shift towards um this concept of agents doing work

(34:19):
right and how do you build based on the work and especially if the work is done well
right um and so that's a pretty big shift so it's a huge shift so that's on the pricing and
packaging front trying to think of what else i mean it just works its way through basically
every single department in the company and so when you add it all up it's like they're mired

(34:43):
in the old world yeah i think the other thing that we've discussed before too is just like
the bureaucratic administrative layer that's been implanted into a lot of these businesses,
particularly the larger ones. Oh, I mean, I could go on and on about that.
But it's like, you know, this whole managerial class concept where you have these people steering

(35:06):
the work that don't have any idea what they're doing, you know, at least on the practice. So
like you have an engineer, let's say, who goes deep in AI and they're like trying to explain
what the heck this new technology does. And the middle manager that's selling their RSUs and making
plenty of money off the old world is not going to really entertain this unless they're forced to.

(35:33):
And so you just have this, the politics is crazy. It's absolutely insane. And so you put all of this
stuff together and this is why i say that the the roadblock is is around the existing structures it
the ai the tech itself is way further ahead than how businesses have implemented it and then you

(35:55):
also have this whole kind of um consultant class we'll call it where they come in and like ai is
changing the game um pay us lots of money and we will recommend mediocre things like you know maybe
I mean, and I don't want to be too facetious here, I guess.
But we will make it so you can process your PDFs.

(36:17):
They implement, in some businesses, this is actually a game changer.
But they kind of just sprinkle it in.
Oh, AI transformation.
Your entire company needs to be on ChatGPT.
OK, I will take my monthly retainer, and I will come back next month and tell you some other basic, obvious thing that you need to do.
And so that's why I think I was reading recently.
I think it's like an MIT story or something.

(36:39):
95% of AI initiatives fail.
It's like, it's no wonder why.
Ever after the last five minutes of explaining every part of a business where things could go wrong, it just makes sense, right?
Yeah, it makes total sense.
And that's why, number one, I'm very happy that I met you and we've been having these conversations.

(37:03):
Number two, I feel fortunate at TFTC and at 1031 that we have relatively small teams.
tftc there's four of us and some contractors 1031 there's five of us and i feel like i've been able
to adapt i mean i'm not saying that we're on the cutting edge of ai um in terms of implementing it
on the business but we have implemented it to a certain degree like we're making i think we've

(37:26):
got some automation on the back end very simple stuff just to get my head of gross and sort of my
multi-tool guy, Ed,
gives more time on his plate.
And then on the front end,
the AI-generated videos,
obviously opportunity costs,
which you were the inspiration
for pivoting that from a mobile app
to a browser extension.

(37:46):
We were able to build that in a month.
And it's been fun implementing it
in this business,
but we're not providing AI products.
We're leveraging it to make our media business
and our investment platform
more efficient.
in many different ways.
And I guess this gets to the question,
like it seems pretty obvious to me

(38:06):
that there seems to be an arms race in Silicon Valley
and in tech more broadly to figure this out.
And there's been billions, hundreds of billions,
probably over a trillion dollars now at this point
that's been thrown at the space.
And even though you just described
that the advantages for startups are massive,

(38:29):
I find it hard to believe
that a lot of that capital is not being misallocated and people haven't really clued in on
what they should build specifically and how this is actually going to change.
Yeah. I mean, I certainly think there's quite a bit of that. And so I could kind of share some
examples that kind of pop, you know, are top of mind. Actually, maybe I'll save it for a little

(38:53):
bit. So I think the first thing is, in many ways, I actually understand the rationale for why the
money's sloshing around. And I think it's because the people that are pretty invested and are in the
weeds, they kind of understand that things are going to change in a meaningful way. And for all

(39:17):
the reasons I mentioned earlier, there will be a meaningful amount of disruption, in particular to
these existing established businesses. And so I actually, I don't know if I,
I understand the rationale for why investors and venture capitalists are plowing a bunch of money
in space. So, you know, I'll give an example. And this is happening, I would argue, like,

(39:41):
it seems like just everywhere, right? So you see it in the late stage where, you know,
Open AI is valued at some large amount of money.
Anthropic is raising.
So you have that at the foundation model level.
And that one feels honestly maybe a little bit, I don't know if I'd say riskier, but
somebody is going to win and the revenues are scaling in a meaningful way.

(40:04):
Profitability may be a different story, but the revenues are scaling massively.
um and so it kind of makes its way throughout all stages of um the tech sector and so i'll give
maybe my favorite example um so maybe a little bit of background before i do though um
these startups that latch on right because there's so many people out there saying what is this ai

(40:31):
thing um how do i use it like i've never seen any time in technology where you could just
basically walk up to anybody on the street and you could have like a conversation and they're
going to have like questions and thoughts and curiosity around AI. Like I've just, I've never
seen that before. Like it's, it's like, it's very top of mind for a lot of people. Like, you know,

(40:53):
I take my kids to the pool down the street and I'm like chatting with a neighbor about it. You know,
it's like, it's just, it's everywhere. And so because there's so much attention there,
you have these startups that come out of nowhere and like scale like super fast so i think notable
examples are like you know products like lovable or cursor if you're a developer there's kind of

(41:18):
these stories right where it's like hey we went from zero to 100 million in annual revenue in like
eight months. And it's like, no one's ever seen that before, ever. And so some of that's maybe
because of the fiat debasement. But people have never seen that before. And so what that means is

(41:40):
if you're a venture capitalist that is in the business of identifying one out of 10, one out of
20 companies where you think it could be a home run, your spreadsheets, you can make your
spreadsheets make sense if that one business, instead of scaling to $50 million in annual

(42:01):
revenue, scales to $150 million in the same amount of time.
So the spreadsheet math, I think, actually kind of makes sense.
So what they do is you have these bigger funds that have a billion dollars that they have
to deploy, they'll still write checks into the open AIs of the world. And that's like,

(42:22):
stroke $100 million check or whatever. But they're also getting involved in the earlier stages too.
And so it creates, it kind of reminds me of the COVID era. I dare say hysteria,
where early stage businesses that had traction, people were ripping massive checks into these.

(42:43):
And so I heard from one of my existing investors that a startup that graduated, this is not a rare occurrence, but a startup that graduates from the last Y Combinator batch, which is a big, really well-known accelerator that fetches a premium for the startups that go through it.
It's like she was saying, oh, yeah, like, you know, this individual like raised at a forty five million dollar valuation after like three weeks of work.

(43:13):
Like barely had a product that was raising at like a forty five million dollar valuation.
And like I have friends that have raised at 50, 60 and like some of them didn't even have any paid customers.
So like there is a level of froth in the market.
um so what that means is if you're an early stage founder you kind of need to know like

(43:34):
hey the model is nine out of ten of us are gonna fail one out of ten of us is gonna make it work
and that will return the fund for the venture capitalists and that's the way that the game
is played and if you don't like it don't raise vc um but i actually think that the money is

(43:55):
in some instances probably not being allocated well, but I understand the rationale. I do.
I will say the thing that irks me is, and once again, I kind of understand this logic as well,
but I think it's kind of maybe a sleight of hand, is you have these startups. I think Cursor might

(44:18):
be one of them. Not hating on Cursor because I think they built a good product, but they basically,
it's a land grab. So they're competing with, you know, Cloud Code and Cursor and all these other,
all these other like code tools. So what Cursor does is they give away a bunch of credits.
So you sign up and they're like, here's $200 in free AI credits to use

(44:40):
so that you can get up and running. It's a great idea, right? Because people
use those credits when they're writing code and then they learn to like your product and they
learn how it works, and then they're less likely to go try the other product down the street.
Well, what happens is maybe if you're a cursor, and I don't know the numbers, so I'm just
using hypothetical examples, you might be doing $100, $150 million in annual revenue,

(45:06):
but you might be spending $50 of that, $75 of it, on literally your open AI bill.
So in the old world, software margins might have been like 80%, 90%.
In the new world, in land grab mode, you might actually have a negative – you may actually be losing money.

(45:30):
And so in the old world, they would still kind of do this with sales and marketing and other things.
But I've never seen this maneuver that I'm seeing now where you're basically just paying the anthropics and the open AIs of the world this monster bill.
And that's top-line revenue.
But you're paying massive infracosts.

(45:52):
Coming in the front door going right out the back door Yeah So everybody goes oh my gosh they killing it My startup is not nearly like and sure maybe your startup isn actually that successful
Sure.
But what they're doing is they're kind of more like a 50 million AR business, not 100 million AR business when you think about it.

(46:12):
But top line is top line.
So I get it.
Once again, I understand why they're doing it.
It's a competitive space.
But these are some of the shenanigans that are currently happening in the market.
Yeah, on the topic of hysteria, it is crazy how quickly people develop amnesia, like both on the investor side, but on the founder side, too, because like raising at insane valuations puts a lot of pressure on you and sort of handcuffs you in the future if you don't get the traction that was expected of you.

(46:42):
Yeah. And I think a lot of founders will basically, you know, a lot of founders look at it and say, you know, when I take a VC check, this is the path that I am signing up for.
Right. Some decide, you know, maybe part of the way through, you know, if they don't have the revenue and the traction, it's like switch to profitability, you know, make the business profitable.

(47:08):
you know it's not going to grow as fast um it's not going to return the fund like the like the
vcs want but it is um but you still at the end of the day have a business that employs people and
you're still delivering our product into the world you know um so which is you know no shame on that
but but when you get on the treadmill you need to know you're getting on the treadmill yeah i mean

(47:28):
on that topic of bringing a product to the world how what effect is ai going to have on what products
look like moving forward like to my point earlier like not only does it seem like vcs are
in this arms race and pouring a bunch of capital and a lot of that's being misallocated
just because of just nature of spray and pray but then on the execution of the startups like do you

(47:52):
how many startups out there do you think have like the right idea of how to implement
ai and bring ai products with new ux new ui to market does that make sense yeah definitely so
So this is a really good question.
I think maybe the – I think all of us are honestly figuring it out as we go.

(48:18):
There are certain patterns that have been, we'll call it, proven in some ways to work.
So this idea of AI being a chat-like experience, it's like for certain jobs to be done, for certain use cases,
That's a great way of interacting with AI.

(48:40):
I think we're going to see a lot of different approaches in the future.
And I actually think a meaningful, feels like almost 50% of the innovation is actually undesigned.
And so a startup, when they hire a product designer, somebody who helps build out the UX,

(49:00):
I mean, we just went through a hiring process.
We brought on a product designer a few weeks ago.
She's amazing.
But there's a lot of people that have an old world view, and they haven't spent time playing
around with products.
They haven't spent time learning AI for themselves.

(49:21):
So you'd be able to kind of understand, at the end of the day, I'm building a user experience
that sits on top of this core piece of technology that has certain things that it needs to function,
a lot of designers in particular just haven't spent the time.
But the ones that do, I think, will really kind of pave some interesting paths.

(49:46):
And so I think the thing that I, and I'll kind of maybe present a high-level way of thinking about this,
but I mentioned earlier that these agents are kind of like little digital employees.
Well, if that is true, right, or even remotely close to true, the user experience may actually
need to end up feeling a bit more like you are working with like a coworker, right?

(50:13):
Maybe these agents, this sounds so cheesy and I can't believe I'm saying it, but like
maybe these little agents need to have little cheesy names, right? Maybe you need to be able
to like tag that, the name of that agent, like you tag them when you're like, you know, using Slack or
Microsoft Teams, right? And so, you know, and like, how do these agents or little employees interact

(50:37):
with each other? Like, when do they hand off data from one thing to the next? Like, let's make sure
that this agent has enough information in its digital brain. And like, if one agent has different
information and another agent has another set of information and they aren't sharing well then you
run into these issues and then there's also this concept of like oh maybe an age like maybe there's

(51:01):
a boss agent too right that kind of like is scanning right it's like are these things working
like maybe you need like little one-on-ones with your with your little with your little
employees like oh let's check in see how they're doing so anyway the reason why i bring this up is

(51:21):
because that's kind of like those are almost like lego building blocks right but there's like these
user experiences on top of that will that will need to be built i just i'm like it's kind of
look different and so that's why to get back to the original comment that i made um i think we're
all just trying to figure this out what have you seen that's been have you seen like a oh
these guys are doing it right this makes sense ux yeah you know i think some of my favorite tools

(51:52):
that i like playing around with are you know uh replit replit's pretty great i'm a big fan of
replit um i like uh v0 which is another similar tool um i've been really having quite a bit of fun
with cloud code but that's more for like a developer like experience because it kind of
sits in your terminal. Honestly, I feel like the number of products that do it poorly is

(52:18):
significantly higher than the ones that do it well. I'm trying to think of this. There's this
one product that I used for a little bit. It was called Howie, and it was almost like this little
meeting assistant, but it would interact only through email. So it's almost like Calendly,
if you've ever used that before, but it was like, there was no user interface.

(52:38):
Right. So I think we'll see more of that, but honestly, like I, I,
I'm constantly trying new tools just because I, I feel like once again,
we're all trying to figure this out.
Yeah. What, um, how do you approach UX UI formable?
I think we, we definitely lean on, uh, the concept of,

(53:00):
so there's a few things that we, that we talk a lot about.
One is this concept of a primitive, or it's like a building block. In the world of, we'll call it
marketing, if you're trying to collect leads and get back to them, you may have a form,
and that's a primitive. You may need to send an email. That's another primitive.
These Lego building blocks, so to speak, that you need to construct together. I'm of the opinion that

(53:27):
all of those primitives, there's an old world primitive, and then there's a new world primitive.
The new world primitive is like, they're still probably called the same thing, but they are actually an agent.
So like the product that we're working on, the initial product, it literally looks like a multi-step form, like if you've ever used type form before.

(53:51):
And so you fill it out almost like you're filling out a form, but it's actually an agent behind the scenes.
And so because it's an agent, it is literally as you're filling it out, it's like it's kind of reasoning on the fly about what should we do.
Right. So someone enters their email. We're like, hey, we can enrich that information.
Right. So we can figure out a little bit more about you. You fill it out.

(54:13):
Maybe we show you questions and we don't show you other questions.
And then at the end, we can actually have like another agent that goes, is there anything else that we need to ask?
It's like, oh, yeah, there might actually be because you have gave us this job that we need to do.
So at the end, it might ask you a couple final follow up optional questions.
So as a result, if you're a business, you're able to collect this really rich information.

(54:37):
And then you can feed that really rich information into like your follow up email.
So your follow up email is way more personalized because you were able to collect this like amount of information that helps you deliver a super personalized email.
And so we talk a lot about these primitives to get back to my point.
And these primitives are actually kind of like agents. So instead of trying to convince the world that you should be using agents, we're like, no, no, no, there's an existing pattern. There's forms, there's emails, there's these things. And we're infusing them with AI and rebuilding them as an agent first.

(55:10):
So you still think about your world through this lens of these existing Lego building blocks that you're used to. So that kind of constructs things. But then you have this kind of coordination that needs to happen, like this intelligence layer above those primitives, where maybe you have a world where in the future, I can describe the funnel that I'm trying to build.

(55:32):
Hey, I want to collect leads and I want to get back to them really quickly.
And then I want to enroll them in some type of drip sequence where they get an email three days later and five days later.
And an agent should be able to construct those different primitives or building blocks, put them all together, connect them all, and you don't have to do it yourself.

(55:53):
Even better, there could be another agent or another piece of your code, we'll call it, that has knowledge about your business.
Maybe it scraped your website.
And it understands your pricing and packaging.
It understands the tone of voice that you're trying to use.
It understands these things about your business.
And it literally gives that information to every single one of those little agents.

(56:16):
So all of a sudden, your emails are smarter because it knows about your business.
Your forms are smarter because it knows about the job that you're trying to do, but also it knows about your business.
And so there's this intelligence layer on top of these primitives that we believe will be really transformational when it comes to assembling the Lego building blocks.
And then there's almost like another level where you'll have another thing that sits above it, and it will analyze everything.

(56:43):
It'll analyze your entire marketing funnel and go, hey, there's drop-off rates here.
There's drop-off rates here.
You're not converting as well as you could be.
Would you like us to make some changes?
Like, heck yeah, of course.
click, right? It applies to change. I don't need to go in and click 18 buttons over and over again.
And so it's like, I honestly don't know exactly what the user experience for that looks like,

(57:06):
but I feel like I have at least, you know, maybe I'm being foolish, but I feel like I kind of have
an idea about what the world will look like in the future. And it's, instead of humans being like
the first thing that you plug in to solve a problem, it's going to be these little digital
employees. Unless you need that extra, unless the variance in the task is too high. And sometimes

(57:33):
you'll send a person in to learn more about it so that you can then build an agent to put it in
there instead. So it's almost like people are gap-filling the work of the agents instead of
agents gap-filling the work of the humans. But right now, it's like human plus agent or human
plus AI. And I'm like, no, no, no. It's AI plus human. Yeah. No, it's got me thinking.

(57:58):
I was just like thinking about how I could use this at TFTC. Just one, just thinking of
a meeting I had earlier this morning. One of our goals is
newsletter has been word of mouth. I haven't really spent any time like focusing on building
that funnel and pushing people towards that. We've put a little bit more effort into it

(58:19):
in recent months with like a read at the end of a of a podcast or whatever but like we can make the
newsletter experience number one much bigger by actually trying to grow it and then number two
much more hopefully insightful by if we had a forum where people were getting onboarded to sign
up for the newsletter like what interests you most about bitcoin what topics do you

(58:42):
are you most interested if you've been reading like what topics that we cover
make you the happiest. I'm just thinking of examples here.
That's something, if we could use your tool to implement that, I don't have to hire
a marketing expert, per se, to build these funnels.
Formable could do it for us. As I'm thinking through that,

(59:04):
it's like, okay, me as a business owner, that really empowers me and allows me
to do things for cheaper and hopefully increase our audience
and ultimately revenues here at the media business.
and I like to think I have agency
and we're trying to lean into these tools
so it would be good for our business.
The broader question I'm getting at is
is this AI revolution truly a 10,000 flowers will bloom situation

(59:30):
or are you going to have a crazy, not even Pareto,
but crazy distribution of companies that exist and succeed
versus those that simply don't exist?
What's the nature of working going to be like?
I think my short answer is it's almost like chaos is a ladder.

(59:50):
We would sometimes, my co-founder and I would say this a decent amount.
Chaos is a ladder.
It's an opportunity, right?
I think there's the first part, which is chaos.
And I'm like, yeah, that's going to happen.
You see it in every other part of our world, right?
Just like total nonsensical.

(01:00:10):
You read the news and you're just like, I can't believe.
If it was maybe a few years ago, I never would have believed it. But now it's like, this is normal. This is pure chaos. And so I think that in some ways it will apply here as well.
right and so you know but but with that chaos there is opportunity um and

(01:00:38):
i think i am of the opinion and i try to be an optimist um you know i try to be realistic about
the situation at hand but i i tend to you know envision a bright and interesting future i think
I just like to follow. I love to tinker. I love to try things. I love to learn new things. And so

(01:01:00):
I think, yeah, being an optimist is kind of part of that job, right? But also, there is a reality
that things are going to look very different. And I don't necessarily think that changed
will happen all at once, gradually, then suddenly. Or actually, I think it'll be maybe a little bit

(01:01:25):
more of a gradual change. Because those bigger businesses that are sitting on their hands right
now, they won't lose all of their revenue overnight, but it'll just kind of go,
right? It'll be this gradual decline, unless you're Chegg. And then if you do quizzes,
online quizzes, you're totally cooked, right? So, yeah, where am I going with this?

(01:01:55):
So I think like, here's what I would encourage you, like forever, whoever who's still listening
to my diatribes, the way that I would encourage you all to think about it is maybe keep two things
in mind. One is that these AI companies in particular are honestly in the business of
narrative. They're in the narrative business. A year and a half ago, it was like, AGI is here.

(01:02:25):
That's wonderful if you're going out and trying to raise a mountain of cash.
right it's like that's a great that's a great idea say agi is near
um turns out yeah two weeks until agi right um and so on one hand they will hype this

(01:02:48):
right they will hype it and then there are moments where you actually i feel like you
see their true colors where they will like in other areas undersell it so it's almost like they're
kind of uh uh like they they it's almost like i can't parse these companies because one minute
they're saying agi is near and then the next minute they're like kind of almost like walking

(01:03:11):
that back and it's like why is that right and i think that there's like two things that are true
right one is like obviously i was changing the game um i think that agi is not near unless you're
fundraising. But on the other end, there is meaningful disruption that will take place
that they are not going to tell you about. And so that organization that maybe had, I mean,

(01:03:35):
this is happening in customer support right now, for sure. You have these chatbot products that
have come out that are way better at answering questions than people. Because once again,
they have knowledge about the business and they can hold it in their little digital brains way
better than a human can. And so as a result, it's like customer support is one of the first things

(01:03:57):
that's just going to be roasted by AI. And we're seeing it like there, but you don't see a lot of
it in the news. You will see more of it. And that's like a huge amount of, you know, there's
a lot of people that are employed in customer support. Like there are a lot of people that are
out in, you know, rural America that work at a remote customer support job. Well, you know,

(01:04:22):
maybe there's fear because of outsourcing, but, um, well, like, you know, and, and they're,
you know, they're working this job and they're for most part, you know, happy with the situation
that they're in. It's like AI comes in, bang. It's almost like, uh, uh, growing up in Maine,
we had these, uh, mill towns, right? So you had these paper mills, um, and, uh, you had these

(01:04:43):
paper mills and they started outsourcing everything. And my grandfather, he would haul lumber and he
would haul wood chips to these paper mills. And I remember being five, six years old and him
complaining about outsourcing happening. I would go on trips with him in his tractor trailer truck.

(01:05:04):
I would be in the back, take naps, and then he would take lumber from Maine, drive it across
into Canada, they would process and he'd bring a load back. And it infuriated
him. And he would constantly talk
about it with my dad. And so I think
there's going to be some element of that that happens, but it's going to happen

(01:05:27):
in white-collar work, which is kind of
wild to think about.
It's wild to think about because
I mean, there are, I mean, I think an overwhelming majority of people in these white collar jobs are asleep at the wheel when it comes to this stuff.

(01:05:47):
I think they're, I don't know, but it's definitely a form of cognitive dissonance where they don't want to believe.
And especially for white collar workers around our age, millennials who've struggled enough to get to where they are.
It's pop culture or popular culture right now.
It's like the millennial trajectory.
You go to school, you get a degree, you get a good high school, good college, right?

(01:06:10):
Major, get a degree.
And then that doesn't panned out well.
A lot of the people who hasn't panned out as well as they expected are in these white collar jobs already not very happy with their position in life.
And about to get a double whammy with this.
I'll give you another example.
um so um there's this guy that i follow on the internet and he basically built um i don't know

(01:06:35):
exactly how he built it but imagine like if ai was in excel right so imagine all these uh financial
analysts that are really good at writing i don't know i hate excel but like macros i guess yeah
yeah v lookups macros whatever right it's like he built this tool and it's like

(01:06:57):
the people that are using it are freaking out because it's like the analyst jobs.
And so it's not only happening to like, you know, the low end, it's like, you know,
somebody who works at like, I don't know, Goldman Sachs or whatever, you know,
some financial business that I'm unaware of, right? It's like, analyst, bye.

(01:07:18):
um and so you know and right now we don't see the i don't you know like because the models are
not perfect not everything can be solved by ai obviously and i think that even in the future
like not everything will be solved with ai but once again if you think about it ai first then

(01:07:40):
human it just like the the amount of labor when you go from human first then ai to ai then human
it's like there's going to be a meaningful delta and so i really think you know i'm an optimist
and i am optimistic on the future and i'm a technologist for a reason um but that doesn't

(01:08:01):
mean that there isn't some type of short-term chaos yeah but moving on from the cast to the
the optimism like for teams out there that are building like it's incredibly powerful it's
exciting it's fun we're having a ton of fun at tftc implementing the stuff and working on
particularly the videos it's um it's been something that we've really been somewhat

(01:08:25):
addicted to over the last three months and it's just getting better and easier for us as we
get the process down and the reception is still pretty pretty remarkable once we get it out there
I was watching one earlier today and I'm going to follow up with you after and be like, tell me your secrets because I haven't spent much time here, but I'm just like absolutely loving it.

(01:08:48):
I think you guys had one video to get like 20,000 retweets or something crazy like that.
Yeah, I think the 19 – going off the gold standard one, that was the most popular so far.
Amazing.
And it was such a great video.
Obviously, it's like, oh, it's a little AI generated in there.
But it was like the storytelling was amazing.

(01:09:09):
Yeah.
I think that's what we phoned in on is just using it to tell stories and really storyboarding.
That's what most of the time is generating.
You have to iterate through the generations.
And thank God I don't have to do that.
That's Trevor Elijah's job to do that.
And I think they definitely have moments where they're beating their head on the table because the generation isn't what they need.

(01:09:33):
But once the end product's there, it's like, holy crap.
A lot of our time spent on storyboarding and deciding what story to tell, which is fun.
I like to think I'm somewhat of a creative person, but I've never been able to manifest that outside of conversations and writing.
I love movies and I like art, but I can't draw.

(01:09:56):
I can't produce a movie, but now we can sort of start that process with AI.
I mean one of my favorite things to do and this is such an easy example right is I have a bunch of kids right I have four kids seven and under And you know one of the favorite activities is to do coloring pages And it like I will

(01:10:16):
generate coloring pages for the kids, right? And so I'll ask them, what do you want? And like,
I want a princess riding a pony, jumping over this thing. And it's like, it gets really wild
really quickly. It's like, it generates it. I print it off. I give it to them. And they're like
coloring they're having a great time yeah and we have um we've been uh in the car ride we've been

(01:10:38):
using chat gbt voice we named our gbt daryl and uh if the kids the boys ever have a question in
the back of the car we just pop daryl open give them the context hey daryl yes we've got a five
and a three-year-old here they're very curious and they're looking to learn and um my my oldest
son two weeks ago we were playing with daryl he's like how long does it take to count to 100

(01:11:00):
Trillion Daddy. And I was like, I don't know, let's ask Daryl.
And the answer blew my mind. I was like, ah.
But then it went from there. His cousin
was in the car and similar to the
coloring
pages, we asked
Daryl to craft a story for
my oldest boy and his cousin
who's a girl. Like a heroic story of them
traveling through a forest and
in 10 seconds had a five minute story.

(01:11:22):
And his parents in the car was like,
oh, we get five minutes of relative silence.
And it was fun.
And they were amazed. Like the fact that they were the
main characters in the story it's like it's so much fun as you know i my son loves uh insects
and so you'll turn on the little video mode and you'll like pull up an insect he comes in he's

(01:11:44):
like got this thing on his you know on his arm or whatever what is it like i don't know let's find
out right yeah and it's like untapped and that's actually one thing the uh the school my boys go to
And now they actually I'm happy to see that they've put feelers out there to the parents.
Like we're looking to before we're thinking on AI, we're putting together like an AI board or committee to help us implement it in the school.

(01:12:09):
I'm like, thank God you're thinking about this because that's like the thing positively, optimistically, like the potential for this to really accelerate human knowledge is unbounded.
Really, it really is.
It really is.
you almost have to like teach yourself when, um, like historically, you know, there would be

(01:12:34):
these questions that I would have, or it'd be like, how do I do X? Right. Or is it possible to do X?
Right. And before, if I were to Google it, right. It's like, I'm probably not going to get
something interesting because it's such a like particular thing, unless it's, you know,
how do I change my door or something, you know, like a lot of people search for, but usually it's

(01:12:58):
like this kind of longer tail thing. Right. And, uh, and now it's just like, you almost have to
teach yourself. I, I should just ask that question. Right. And, and then it will kind of prime my
thinking with some ideas, which I can then use to dig deeper. And I see it with kids. I see it with

(01:13:19):
myself, you know, it's like, is it possible to even build this into my product? Like, oh yeah,
you could probably do it this way or that way or the other way. It's like, that's actually a great
idea. I'm going to use that. Right. So you almost kind of have to have like a, there's a level of,
you know, people would always say in, in for software engineers in particular, like you should
go Google. Like if you ever have a question about code, go Google it. And then if you can't figure

(01:13:44):
it out, come back and ask me. It's almost like we've kind of entered this phase where it's not
just for software engineers anymore. It's like, you have a question, ask, you know, ask one of
these AIs first to prime it. And then if you still can't figure it out, you know, go another level
deeper. You know, someone who's homeschooled, the world has just will massively change in education

(01:14:10):
as well. And we homeschool our children. And a lot of it was we were kind of teetering on like,
Should we do maybe like some private schooling or something versus homeschooling?
And at the time, I was playing around with AI and I was like, time out.
No, we're homeschooling.
Because when you have these superpowers in your hands, obviously, I don't want my kids

(01:14:34):
to be playing around on screens all day.
But if you selectively use it, it is a massive accelerant.
And the homeschooler, at least the homeschooled parent, is always kind of trained or
learns that the most important thing about homeschooling is the personalized instruction.
And so if you apply that with AI, it's like it just makes it even more.

(01:14:57):
You know, it's like personalization on steroids.
Yeah, especially with the memory, with the agentic brain that exists.
Yeah.
And it is scary, too.
It's like, how much information do we give these models?
Like, can open source models and products like Maple get the parodies so you feel more comfortable doing all this?
but that's why i think the big scary thing is uh is really around like the training

(01:15:21):
you know it's like i if there's one area that i feel like some amount of maybe the free the free
market maybe the free market figures it out um but the thing that i really think um should exist
and should be transparent is you know they basically have these i guess i was listening to

(01:15:44):
I think it was Sam Altman chatting about it, but they basically have these big documents that have these very maybe convoluted rules around like if this happens, don't be right.
It's like it's kind of the teaching document for the foundation model.
I want to know what that is.
And I think we have – I think that those companies have a responsibility to tell us.

(01:16:06):
because if anybody was concerned about social media censorship,
this has the potential to be that on steroids.
So publish, get that out in the open.

(01:16:27):
I think that that makes sense.
I think you can keep some of these other trade secrets, sure.
But like that, we should know how the algorithm works.
Yeah.
I have a feeling that's going to be hard to get unless you have the free market come and somebody position themselves as we're going to be the company that does this and you should use us because of this.

(01:16:47):
But tying this and the bow on the overarching conversation, like a lot of how we view a 1031 and me personally as a Bitcoin or businesses transforming is on the capital side, on the reserve asset side with Bitcoin.
Like, let's just expand on how you view Bitcoin's role in not only business formation, but like how you run a business into the future, particularly if you're leveraging this unbounded leverage with AI.

(01:17:18):
There's a lot of places that I could go.
Maybe I'll start with the tech, maybe.
um uh the first is like the idea of agents um exchanging some amount of currency between each
other very quickly um i can see that happening i was never like super bullish on this idea of like

(01:17:41):
i get i pay per amount of minute i consume like i was never a big fan of that but like with agents
because like i think humans kind of are always like oh i'm paying more and i'm paying more and
And there's a psychological roadblock, I think, to that model.
Nick Zabo is the mental cost of microtransactions.
Yeah, yeah.
But when agents are doing it, they don't really care, right?

(01:18:04):
Because they're not like – I think maybe they will care.
I don't know.
So I think that from a technological level, I think we can see more of that in the future or we will see more of that in the future.
It feels like that's coming.
and Stripe and all these other ones are doing it with different stablecoin rails and other things
like that. I'm a big fan of these Bitcoin treasury companies. I think that they're

(01:18:33):
Bitcoin treasury maxis. And so they recognize that there's an opportunity in the market. And so
in some ways, the actual business is the sideshow to the main show. And the main show is
like intelligent leverage, et cetera.
MicroStrategy has done an amazing job there,
and I a huge fan of what they done But I think most businesses will still maintain their normal business operations They won go public but any profitable business will have an opportunity to store their energy

(01:19:09):
or store their capital in a way where it's not debased.
And I think that that will be a wonderful thing for the world.
I think that that will present new opportunities for startups amid their growth journey.
So like one of my friends who I will not mention his name started a VC backed startup, grew it.

(01:19:30):
I mean, the thing's profitable. And over the past few years, like he's accumulated a war chest.
Like he's accumulated a lot of Bitcoin. And and it's like absolutely thrilling.
And it's changed the way that he thinks about how he grows his business.
So like in the past, you kind of you basically only had like maybe one option and it was let's throw money at trying to reignite that growth in a more meaningful way.

(01:19:59):
But when you're able to store your capital in a better way, Bitcoin's a hurdle, right? Are you going to hire that super senior leader that wants a ton of money and may not work out? Or you could just shovel that into Bitcoin, shovel that extra profit into Bitcoin, and play the long game.

(01:20:25):
I think that it's going to be a wonderful thing for businesses.
I think the big question mark is profitability.
Because I think if you're in my shoes, we hold a very small sliver in Bitcoin.
And I'm happy about that.
And I'm proud of it.

(01:20:47):
But the reality is, we actually want to beat the hurdle rate.
We want to beat the Bitcoin appreciation rate.
If we don't, we're cooked.
Right?
And so in some ways, you kind of have to think about, like, there's a lot of areas that I feel like I can deploy the capital and beat the Bitcoin hurdle rate.

(01:21:10):
That may not always be true, but it's true now.
Um, but like for certain capital that I wouldn't deploy, I should think about storing it, um,
in a way that is, uh, is wise and thoughtful, you know, like I, uh, I was in the, I was
in the process of shutting my business down and I had a bank account at SVB, Silicon Valley bank,

(01:21:36):
and I had some money in it. And then that whole SVB thing blew up and it was like, huh, okay.
You know, like, this was never something I thought about before. Now you have to think about it a little bit. Right? And like every entrepreneur is just like, I put my money in SVB, I maybe put it in a savings account that yields, at the time, I think it was like, I don't know, 3% or 2% or something.

(01:22:04):
but you know that's what they did and then all of a sudden you know the financial world kind of
blew up so anyway I digress but I think for for businesses especially profitable ones
you know Bitcoin is going to be an amazing recalibration moment

(01:22:29):
it will recalibrate how they think about businesses, how they think about capital allocation.
And I think that that will be a wonderful thing for the world because there's nothing
that you just, when, when fiat is melting, the, I just think you end up making stupid capital

(01:22:50):
allocation decisions because you, you intuitively know that the ice cube is melting.
um and so yeah i'm really excited to see how things change and i hope to you know do a little
bit of writing and kind of talk a lot about it a little bit more um for even like an early stage
startup because i think that while we have to be we have to think about this different um it kind

(01:23:12):
of feels weird that startups that are encouraged to embrace technology in every part of the business
Like, well, that part's off limits. That doesn't make sense. You know, it's like, no, this digital transformation exists everywhere. And we and we should be thoughtful around how we do that and where and that applies to your balance sheet as well.

(01:23:36):
Yeah, completely agree. I mean, we've seen it within the portfolio 1031, like the companies that have taken a portion of their raises, use it to buy Bitcoin, like the amount of companies that you were describing it earlier, you raise typically 18 months of runway, in some cases, 24 months, in extreme cases, 36 months, but there's a number of companies that raised well beyond that sort of threshold.

(01:24:06):
that typical threshold that's thrown out there and have not had to go back to market because
the Bitcoin treasury has worked in their favor and they're able to use that as working capital
down the line. And not only that, we're investing in Bitcoin companies run by Bitcoiners who
understand the hurdle rate. And so I think we have an anomalous portfolio in the sense that
we have companies focused on profitability and growth alongside each other to make sure that

(01:24:30):
they can stack as much Bitcoin as possible. Yeah. I really think the number one benefit
is not price appreciation over time. I actually think it's that it rewires how you think about
building a business. And that will make you a better founder. It will make you a better operator.

(01:24:51):
And then maybe the second is like, there's a little bit of a hedge. Like you can create a
little bit of a hedge against debasement in particular. And then thirdly, think about price
appreciation last. I like that. I like that framing. And I know you said that like a lot of
particularly other VCs, that's the thing that blows my mind. If we're like in a deal with a VC

(01:25:19):
that's not focused on Bitcoin or quote unquote crypto and the founder has to go to them and be
like, yeah, we're going to take a portion of this raise and put it in Bitcoin.
I get a lot of questions.
And it blows my mind that, to your point,
we're in this digital transformation age
and you're having a company focus on all areas of the business
that are going to be disrupted by this digital transformation

(01:25:41):
except for the balance sheet.
It just doesn't make sense.
But on that note, obviously you're well-connected,
know a lot of founders.
Do you get the feeling that people are waking up to this?
I think it's still super early.
Yeah.
I think it's super early.

(01:26:02):
So if you look in the public markets, I think the public markets are kind of indicative of the private markets, right?
So the most notable example as of late is Figma, right?
Figma went public.
They had, I think it was like 5% or less in Bitcoin.
And I think Dylan Field, I think that's his name, the CEO, was super involved in NFTs and all this stuff. And obviously, apparently, he's a Bitcoin fan. I think that that's generally – that feels like what will become a new de facto standard is like 5% or less for a business in this fast-growing venture-backed world.

(01:26:49):
I don't think it – yeah, because you literally – the game, the way that you grow is like you're okay being unprofitable.
You're trying to beat the hurdle rate meaningfully.
So you just have to think about it differently.
But like I said, if you're profitable and you, you know, I met, I was at a Bitcoin treasury conference last week and I stopped by and I met an individual and his father that run a, like a self-storage facility.

(01:27:23):
And it's like, they're buying, like, I think he said, I mean, this is public. So I think I saw it on Twitter. It's like, you know, 10 grand a week or 10 grand a month.
and they just you know they run a you know a self-storage facility and somewhere in in the
states and but it's like that's awesome that's absolutely fantastic and there's going to be a
lot more of those yeah and those who get on earlier are gonna be very happy that they did yeah that what i and that what i think you an incredible advocate for it because everybody going to be a bitcoin treasury company just like everybody is

(01:27:58):
an internet company to a certain extent at some point and we're obviously in the early stages
that very early stages of that transition um and obviously our core focus my core focus here at
this business at 1031 is on bitcoin specific industry but that's something at 1031 particularly
we talk about a lot is like, when do we sort of begin to cross that chasm? And I think advocates

(01:28:20):
like you or Bitcoin are building a business that really doesn't have anything to do with Bitcoin
and is on the cutting edge of his AI revolution. Like the more Luke Thomas's that exist in the
world that are really putting this message out there and sort of de-risking it for others
is extremely important. Yeah, it's going to be, it's going to be a fun time, but,

(01:28:44):
We're still obviously early innings. And like I said before, I really think the number one unlock is that it rewires how you think about capital allocation. I've seen it with my friends. I have personally experienced this for myself.
Right. So it's like a lot of Bitcoiners experience this personally. And then they're like, would this actually work in a business context? It's like, absolutely. Absolutely. And so I just think that that is the nugget is it's like it's almost like if you were to, you know, there's this concept of, you know, like CEO coaching, founder coaching.

(01:29:22):
right it's like holding a little bit of bitcoin on your balance sheet even a small amount
is like hiring a coach it teaches you to be more thoughtful it's like what is that worth
priceless it's priceless it really is i love you describing it that way

(01:29:44):
it's a founder coach yeah doesn't mean i'm not gonna spend money but it does mean that um that
i'm gonna view it through through a through a different lens i'm gonna think twice yeah weigh
the opportunity cost if you will the opportunity cost what um any final thoughts final pieces of

(01:30:08):
advice as we wrap up here so many things i could probably say but i think if you've made it this far
thank you um if you have comments uh shoot me an email luke at formable.ai um or leave a comment
on the youtube video or whatever and i will uh read it and weep um but uh i think if you if you're

(01:30:30):
skeptical about ai i understand um and i think that um the reason why you're skeptical is probably
because you see a lot of hype.
And that is true.
But what is also is true
is that there is substance.
And if you want to learn
and the analogy I always use is,
I feel like sometimes I'm in a room
and there's like $100 bills

(01:30:51):
to use the fiat world as an example.
It's like AI opens your eyes
and you kind of realize
that there's like a bunch of money
lying on the floor.
And you're like, well,
like, I guess I could pick it up.
Like, I guess I could pick it up.
Because if I don't like,
maybe someone else will. And so AI kind of presents that opportunity, whether you're a small business

(01:31:12):
or a bigger business. I think small businesses, per your comment earlier around how you're using it,
actually have a way bigger advantage. There's just so much more that you can do. And I feel like the
only way to jump in is to just tinker. Start playing around with ChatGPT for a little bit,

(01:31:33):
or Grok or whatever your tool choice is, then upgrade and start using AI in like workflow tools,
whether it's Zapier, there's a product called N8N, Relay, Gumloop, those ones where basically you can
start building out business workflows that kind of consistently do things for you. That's like the

(01:31:55):
next leg. And while you're playing with those tools, they will kind of introduce you to some
of the more advanced concepts as you're tuning the things.
And then you'll be able to kind of build this worldview
or mental model around how to use agents
and how they can work for you.

(01:32:16):
And then I think all of a sudden,
you'll start to notice more areas
where things could be transformed.
I will say the other last thing is that you can actually,
this sounds so bad and weird,
um ask chat gpt how to use it if it has recommendations explain your business ask it
if it has recommendations throw in images of like little whiteboard diagrams of your business process

(01:32:40):
and stuff like that ask it it'll probably come back with some ideas that will help you um come
up with some creative inspiration yeah i will say i mean i know we're wrapping up but i think
it's important just to give an example of how we're leveraging it on the back end business
process side but we use n8n for a couple things right now um for the newsletter we just a scaffold

(01:33:01):
to get a scaffold of the newsletter every day we're we've used it we're currently like iterating
on it should go live again next week but ed when i write the newsletter um ed usually i'm like all
right i'm ready to write a newsletter and he goes in and manually like creates a new post make sure
the ads are there like and then we have the section um which is basically a short summary of

(01:33:26):
a section of a podcast that we recorded within the last week with a quote and a link back to the
podcast and it's been a manual process of we'll record this on riverside we'll get the transcript
we'll put that in dropbox then we'll take the transcript from dropbox and put in claude code
that we've built this prompt that we've been iterating on for for months now at this point

(01:33:47):
And then we'll get a couple of podcast sections that we can choose from to put into the newsletter.
And that's through the manual process.
And AI is sprinkled in, obviously, transcription and the prompt for clock code.
But now with N8N, we can automate that.
Like Logan can put the transcript in the Dropbox, and that triggers the flow where we just go in.

(01:34:08):
And it's like, what section do you want?
Click.
And then you can even connect it to Ghost, and it'll create the post for you.
And that will save ad like 15 minutes a day.
Yep. Perfect example.
Yeah. And it's, it works.
and it's one that works like the power is that you you have this kind of digital brain thinking

(01:34:33):
on your behalf but you still have guardrails on it right you're like no no no like after this
happens i want you to put it over here right so you still feel like you have control um but you
have a little bit of the magic sprinkled in parts of the process where it's most relevant and that i
think is the big nugget is like, that's why I always recommend like play around with chat GPT,

(01:34:55):
sure. Then upgrade these workflow tools because like, you'll still have some control,
but you'll be able to infuse the magic, so to speak, at different parts of the workflow,
but you'll still have the feeling of control. And that's really important because that's where
people go off the rails is they think like, oh, if I have this agent talking to this other agent,

(01:35:15):
and like, all of a sudden, like if it one mistake, like it makes one mistake, the whole thing blows up.
right so with these you have some level of control and that control like businesses want
predictability and so that's why infusing some element of control with the magic is kind of like
the key that picks the lock at least for step two and whatever step three is we'll figure out

(01:35:39):
as ai gets better yeah well hopefully as ai gets better we'll continue these conversations i know
you're a busy man and took you a couple of years to let me know that you're ready to come on. So
hopefully it doesn't take two more years, but I would be happy to come back on anytime.
This is all, you know, I'm, I'm in the thick of it right now with the new business,

(01:36:01):
um, learning every day and, uh, always happy to share, um, what I'm learning. Uh, hopefully it's
right. But always happy to share my learnings. Well, I want to sincerely thank you because you've
been extremely generous with sharing your learnings with me over the years. And it's
helped me really think about how to approach this transformation as it's happening. I think it's

(01:36:27):
helped us improve our business. So thank you. Appreciate it. All right. We'll do it again.
Let's go enjoy our Friday nights with our families. I can hear my boys playing downstairs
and we're going to win.
It's going to be okay.
Absolutely.
Never doom.
Peace and love, freaks.

All Episodes

Episode Transcript

Popular Podcasts

CrimeLess: Hillbilly Heist

Crime Junkie

Stuff You Should Know

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Why Legacy Companies Can't Adopt AI with Luke Thomas

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}CrimeLess: Hillbilly Heist

Crime Junkie

Stuff You Should Know

All Episodes

Why Legacy Companies Can't Adopt AI with Luke Thomas

CrimeLess: Hillbilly Heist