Latest / AI for Founders with Ryan Estes / Your API Keys Are Killing Your Productivity: Mitchell Jones of Lava.so

Transcript

Your API Keys Are Killing Your Productivity: Mitchell Jones of Lava.so · AI for Founders with Ryan Estes

0:00
Your AI tools aren't making you tired because you're working too
0:03
hard, they're making you tired because you just got promoted
0:05
and nobody told you. You went from an individual
0:08
contributor to manager overnight.
0:10
You're not writing code anymore, you're managing instances.
0:13
You're hovering over clawed code with four terminal windows open,
0:16
thinking which one do I run first?
0:18
And all that mental load, that constant contact switching in
0:21
your own brain. This is the new work.
0:24
The exhaustion is just the job now, and most founders are
0:27
burning out trying to plug in API keys, case subscriptions,
0:30
and copy paste secret keys into tools that are already behind a
0:33
pay wall. By the time Claude even finishes
0:36
its sentence, you're spending more energy on the plumbing than
0:39
on the building, and the plumbing is leaking.
0:42
The real unlock isn't working harder inside your AI stack,
0:45
it's removing every single point of friction between your idea
0:49
and execution, so the only thing standing between you and the
0:52
output is the quality of your thinking.
0:54
So if all the friction disappeared tomorrow and your
0:56
tools just worked, would your ideas be good enough to fill
0:59
that silence? Welcome to AI for founders.
1:02
I'm Ryan Estes. My guest today is Mitchell
1:04
Jones, founder of Lava, a Yale grad, Forbes 30 under 30
1:08
honoree, and the guy building an invisible payment layer that
1:11
powers the AI web. This one's for founders who are
1:14
shipping AI products and want to stop leaving money on the table
1:17
or stop spending money they shouldn't have to.
1:20
The AI for foundersnewsletter@aiforfounders.co
1:23
is where founders go when they're tired of being the last
1:25
to know. Every week I'm dropping tools,
1:28
frameworks, and conversations like this one.
1:30
So you walk in already warmed up.
1:32
Go get it at AI for founders.co before your competitor does.
1:36
If this episode made you think, please leave us a review,
1:39
because five stars won't write themselves and neither will your
1:42
margins. And finally, if you're a founder
1:44
or an expert who wants to be a guest on shows like this one,
1:47
you need to talk to Kit Caster. They are the premier podcast
1:51
booking agency connecting brilliant minds with the right
1:53
audiences. No cold pitching, no guessing,
1:56
just targeted placements that build your brand and grow your
1:59
business. Check them out at
2:00
kickcaster.com. That's Kick Caster by Moburst.
2:05
All right, Mitchell, welcome to the show.
2:06
Thank you so much for joining me.
2:08
Thank you so much for having me. Yeah, I'm excited to talk to
2:11
you. You, your product is really on
2:13
the cutting edge of what people are looking for right now and
2:16
very germane to the conversations we have.
2:19
Pricing is always a topic of consideration, you know, the
2:23
different models, whether it's token base, if it's outcome
2:25
basis, is it subscription, is it a classic SAS play.
2:29
So I think this is going to be very relevant and probably very
2:32
useful for a lot of folks. So give us a breakdown real
2:35
quick of lava. Yeah, well, hello everyone.
2:39
My name is Mitchell. I'm the founder of Lava.
2:41
And what we do at a very, very high simple level is make it
2:44
easy for two groups of people to work very well together.
2:48
We make it easy for end users of AI apps to be able to call any
2:52
service or provider, whether that's a model provider or an
2:56
API, be able to use it on a usage basis.
3:00
So imagine you want to use Notion or Slack or GPT or
3:05
whatever tool you want. You can add our AI gateway to
3:09
your Cloud Code instance, your GPT instance, and no connecting
3:13
MCPS, no secret keys, no things like that.
3:16
You can use it immediately. On the flip side, for merchants
3:19
or for folks who are selling services and providing, we make
3:22
it very, very easy for you to get paid by agents and other AI
3:27
services and other users of your services.
3:31
No need for wallets, no need for simple things.
3:35
As long as you have AP is we help you monetize those.
3:38
AP is you can set any pricing you'd like, whether that's token
3:42
markups, whether it's usage based, outcomes based, and it
3:45
works existing with your existing stack, no new tools.
3:49
Our goal is to just make it so that agents that are already
3:52
trying to access your service can access your service, but
3:55
also so that you can get paid for it.
3:57
And so we're just trying to help you convert new users and new
3:59
traffic. Amazing.
4:01
I love it. OK, so 2 use cases.
4:03
Let's go to the first one first, which is basically I want usage
4:08
base payments. Lava takes care of that.
4:12
Is that to say then? But like my Claude Pro
4:14
subscription, I don't need that anymore.
4:16
Does it kind of bypass the the monthly service cost that
4:20
they're getting you to subscribe to?
4:22
The real thing is that's actually they're complimentary.
4:25
So the way most people use AI these days is Claude Code Codex
4:31
Cursor. It's actually becoming your
4:33
command center. So that is your base agent, and
4:36
from that agent you can do a lot of things.
4:39
You can write code, you can read emails, you can send Slack
4:43
requests, you can look at financial data, you can access
4:46
information. But if you want to be able to
4:48
use it to do all of those things, most of those things are
4:51
behind a paywall. What we do is you can add our
4:55
Lava MCP and you can have access to all of those paid services.
5:00
You don't have to have all of these like complex things that
5:03
usually engineers are familiar with, like secret keys and
5:06
signing up for new subscriptions and all those things.
5:09
You just download our MCP, put money into your Lava wallet, and
5:12
you access all of those services.
5:14
So for that use case, you can almost think about us like an AI
5:16
service provider. Why that's valuable for the
5:20
other side is be a a legacy existing company.
5:24
You have this valuable data and this service.
5:27
You don't really know a thing or two about AI or AI payments or
5:30
how do you make it so that agents can pay you and all this
5:32
stuff. But you need to be able to make
5:34
sure you're getting rightfully paid for the products that
5:38
you've built that people can access.
5:40
We make it easy for them to access your product through us,
5:44
and then we just remit you exactly what you're owed.
5:47
So we just help everybody make this stuff simple.
5:50
No new requirements, no new understanding.
5:53
If you have AP is you can work with Lava.
5:56
Love it. So, OK, so on the one side, if
5:58
I'm building and this is me everyday, I'm trying out all
6:01
kinds of AI products and I'll push them to the absolute limit
6:04
until they're like, oh, you know, you're going to need, this
6:07
requires a pro package. So in this case, I could, I
6:10
could actually go in through lava where I could just kind of
6:12
Max it out for like two or three days, see what's going on
6:15
without actually having to commit to the subscription and
6:17
then canceling it. Correct.
6:19
Exactly. And because you get lava credits
6:22
to just access services, we're effectively subsidizing your
6:25
first few. So if you sign up on Lava's
6:26
website, you'll get $10.00 of credits to just access models,
6:30
API services and all those things.
6:31
And so we're footing the bill on those.
6:33
And then I think that's where the magic is.
6:35
You didn't see like, wait, Claude code is great, but if you
6:38
have an agentic gateway with it or an agentic service that gives
6:41
you access to all these tools, you can do so much more, right?
6:45
So you know, one of the things that we, for example, one of the
6:48
things that people like to use for is like for finance, right?
6:51
You can look at different filings, like Edgar filings and
6:53
you can just immediately call them from our gateway.
6:56
No context needed, no understanding needed.
6:58
You just go in the cloud and say use the Lava MCP to find this
7:02
information. Or for folks who are on the
7:04
blockchain, you can just immediately query it by just
7:07
saying use Lava to find what is the price of Solana and we will
7:12
do it for you using whatever tools are available.
7:15
Most of those tools are oftentimes paid.
7:18
And usually what ends up happening is then Claude spits
7:20
back to something to you like, you need to go sign up for this
7:22
tool. And then once you sign up, you
7:24
need to choose a plan. And then once you choose a plan,
7:26
you need to copy the secret key. And then you need to add this
7:28
key back to me. But make sure you do it securely
7:31
and you don't waste a bunch of time and don't expose it.
7:34
And then it's you're somebody who doesn't even know what
7:36
exposing it means. You're like, what is all this
7:38
stuff? So we've just gotten rid of all
7:40
of that and said, hey, you want to use this to do your best
7:44
work. And it is very powerful because
7:46
it can connect all your sources in one, understand how to use
7:49
them and get the job done. But you can't do all of the
7:53
plumbing. So we do the plumbing for you,
7:55
but we also make sure the person who is offering the service gets
7:59
what they're rightfully earned. We're helping them convert free
8:01
traffic into paid traffic. I love it.
8:05
You know it's it's, but you obviously could see this coming
8:09
before me because it was probably January where I was
8:11
like everything's AP is and I need a folder where I'm storing
8:15
all my freaking keys because it's like it's getting insane.
8:18
So in this case, like all the the API keys, it's actually it's
8:23
pulling from lava. So you don't have to worry about
8:24
it. Rather you're just like kind of
8:26
in straight workflow, you're just like, hey lava, go check
8:29
out XYZ 'cause I'm putting together this thing and it's
8:31
like, correct. And it's only pulling in kind of
8:34
what usage you need on a in on a unique per basis.
8:38
Correct. And the, the thing that's really
8:40
powerful about it is there's versions where let's just say
8:43
you have your own instance or your own keys that you want to
8:47
use. You can securely pass them to us
8:50
and then we store them securely. So then you never have to fumble
8:53
around with them ever again. So then you can go to any single
8:55
service, call that service and connect it and you just call it
8:59
through lava and we'll do it all for you.
9:01
No passing keys. No, no, no, anything like that.
9:04
And so that's actually one of the things that a lot of people
9:06
who like open claw actually like about our product is they can
9:09
give our gateway to open claw and an open claw never needs to
9:12
manage keys anymore. It just it just uses lava to
9:15
call things and manage that manage the keys.
9:18
And it also kind of, it limits your exposure to like dropping a
9:21
mint through open claw because that's kind of the, the scary
9:24
thing, you know, that like lava is kind of your budget.
9:26
You're like, once you hit that $100 threshold, like I, I can't
9:30
do it anymore. Yep, exactly.
9:32
And that's like one of the big things that also folks used us
9:34
for was like, oh, wait, not only do you guys enable my open
9:38
collage to call every single service in the world, it's it's
9:42
a literal natural budget. It's like what's in the wallet,
9:45
right? And that is that was always the
9:47
primitive that we thought was really important was like, you
9:50
know, a lot of services right now are trying to make you learn
9:52
new things and new ideas, right? Like they want humans to now
9:56
learn what secret keys are. That makes sense for everybody
9:59
in the engineering world. But Claude is spanning beyond
10:01
that. Open claw is spanning beyond
10:02
that. These are things that like
10:04
humans don't naturally enjoy learning new things.
10:07
So the more we can make it where it's like, no, I just, I got
10:10
into Claude. I asked it to set something up.
10:12
I only had to do one thing and set up one thing and it it had
10:16
access to all the main tools I use.
10:18
That's the powerful thing that we're trying to get to and make
10:20
sure it's easy for folks to to be able to do.
10:23
It just removes all the friction because I mean, what's important
10:26
now, it's not engineering thinking.
10:28
So anything that's like engineering and structure, it's
10:31
it's I, I want to be thinking like a scientist.
10:33
I want to be able to iterate and do experiments rapidly and very
10:37
quickly. And I don't want to be limited
10:39
by like, man, I got to sign up and I get a password and like go
10:43
find API key. It's it's awful.
10:46
Like truly like it slows you down.
10:48
And one thing I've been talking, talking about a lot, and I
10:50
wonder if you can relate to this, is that a part of like,
10:53
you know, AI augmented engineering, There's kind of a
10:56
depleting feeling to it. Like when I, when I would spend
11:00
my days designing, it was like kind of like stimulating and
11:03
like building and fulfilling. But like after a day of like
11:07
sitting in Claude code, I'm like drained.
11:10
And I, I, I've really, and, and it's not because of lack of
11:14
enthusiasm. I mean, it's incredible.
11:16
You know, it's like, it's so fun, but there's something about
11:19
it. And so I've been trying to kind
11:20
of figure out, and I think this might be a part of it that that,
11:23
you know, Claude code's moving so fast that when you have to
11:27
stop and like, like break flow and then go do this thing.
11:30
And now it's a, it's a whole nother authorization and an
11:33
e-mail blah, blah, blah, blah, blah.
11:35
Maybe that's contributing to it. Yeah, and like, I'll, I'll,
11:38
there's two parts of that and I'll break them both down. 1 is
11:40
like you're talking about something that's actually bigger
11:42
than just managing keys, which is why am I technically doing
11:46
less work but feel more exhausted?
11:49
And then there's another thing here, which is like, why am I
11:52
doing so much work managing these things?
11:54
And like, let's go back to the scientific example you talked
11:56
about and like using it like a researcher, right?
11:58
Well, you know, oftentimes you want to research something where
12:01
you need to figure out like what are the sources Claude can
12:03
actually access, right? And so, you know, with a gateway
12:07
like Lava, you can already access things like Pubchem,
12:10
right? And that's all of the chemical
12:11
compounded database in the NIH. You can just call it directly
12:15
from our gateway. No new things to add, no keys to
12:18
add. You can just call it right?
12:20
Or like open library, right? Like a library catalog of every
12:23
single scientifically researched work is right there.
12:28
And so our goal is to take away that one fatigue for you where
12:32
it's like we're just trying to give your agent super your
12:35
powers, right? How do we give your agent access
12:37
to all the things it needs? And then on the other side, make
12:40
sure people are getting rightfully paid for that.
12:43
The second piece is, why do we feel like we're still using up
12:48
so much energy? And it's because you went from a
12:52
really like an individual contributor to a manager.
12:57
And that is actually a different type of stress.
13:01
You know, if you think about it, one of the things that Claude
13:03
empowers you to do is you can use actually you can.
13:06
You can have multiple instances running and you can go do all of
13:11
those instances in one sitting. You just have multiple terminal
13:13
windows for it. But the problem with that is now
13:17
you're not the the energy isn't doing the work.
13:20
The energy is now understanding the context yourself.
13:23
It's the mental energy of trying to make sure it's doing the
13:25
right thing. It's in thinking through those
13:28
problems like a manager as opposed to now like an IC.
13:31
And that is the kind of step up of what worked.
13:34
Looks like you're now a manager of instances as opposed to an IC
13:39
working on one specific narrow project at a time.
13:42
And I think that's a lot of why we kind of feel that exhaustion
13:45
sometimes at the end of a day of working with these tools.
13:47
Technically, they did so much work for us.
13:50
And you know, we're still like, Oh my God, why am I so tired
13:53
after hitting inter all day or, or, or or those things?
13:56
And I think you're, I think. You're completely right.
13:59
Dude one and two, which one do I do?
14:01
And then for some people, they already have dangerously skip
14:03
on, so they're not even doing that part.
14:07
This episode is brought to you by a dangerous idea.
14:10
It's called reading, specifically the AI for Founders
14:12
newsletter. 47,000 Founders strong now.
14:16
And it's the good stuff. No top ten AI tools that will
14:19
change your life written by someone who doesn't ship
14:21
product. It's just real breakdowns of
14:23
what's actually working in AI, who's raised, who failed, who
14:27
pivoted, and which founders quietly built $1,000,000 in ARR
14:31
while everyone else was arguing on X.
14:34
It's written for people building with AI every week.
14:37
You get tactical frameworks, unfair advantages, breakdowns of
14:40
real companies, and the occasional existential question
14:43
about whether your startup is intelligent or just well
14:45
branded. It's free, it's sharp, it might
14:48
mildly offend your current workflow.
14:50
Go to aiforfounders.co and subscribe before your competitor
14:53
does, because the only thing worse than being replaced by AI
14:57
is being replaced by a founder who reads.
14:59
That's the newsletter at AI for founders.co.
15:01
Now back to the show. Oh, and a, a part of it too is
15:04
just like, I, I, I totally agree with you.
15:08
You know, it's like, OK, I can do so much.
15:10
So a lot more needs to go into like, what am I going to do?
15:13
Because I'm going to get deep and then be like, ah, there's
15:16
some elements. So there's kind of like this
15:18
like feeling of like, Are you sure all the time kind of
15:21
holding it with you, you know, similarly, it's kind of like
15:24
using whisper flow. So like if I'm composing an
15:26
e-mail, it's so fun, you know, but it's also, there's kind of
15:30
like a, a ramp up. So I'm used to like kind of
15:33
hunting and pecking, which means that I can kind of formulate my
15:36
thought and my sentences as I'm moving, as opposed to like when
15:39
you just hit the function key and I'm like, OK, now I'm
15:42
talking. And it's like, you know, so it,
15:46
it, it's a little bit different to like be like, OK, I'm, this
15:49
is a little bit more spontaneous and I can, I just need to start
15:51
talking. So I think there's like
15:54
something similar that when I'm working on a project, I'm like,
15:56
OK, this is going to take me 6 weeks.
15:58
So I know I can kind of move the, the ship slowly and kind of
16:02
massage it as we're going as opposed to six weeks now it's 6
16:06
minutes. And it's like, OK, we got to
16:08
have this point in the right direction right now.
16:10
So a part of that might also be kind of like reverse spotlight
16:14
bias, you know, where I also have a feeling of like I'm the
16:19
last to know everything. So I'm like, Oh my gosh, lava.
16:23
This is so cool. But everybody already knows
16:25
about lava except me. Yeah, well, that's actually an
16:28
interesting thing. Like you're there's this FOMO
16:31
like that people feel right now and like to anybody listening,
16:35
like if you're just trying to figure this stuff out for the
16:37
first time, you're not behind. Like truly, if you're listening
16:40
to this podcast, you're not behind.
16:42
And you got to remember, social media is excellent at putting
16:47
the thing in front of you that feels like it's like everyone
16:50
knows that's just because it's gotten 100 likes.
16:53
But like, if we zoom out here and there's hundreds of millions
16:57
of people in the United States and billions of people in the
16:59
world, you're not behind. And so, you know, I think
17:03
there's a lot of like, Ryan, to your point, like a lot of panic
17:06
around, like, am I behind? Do I not know these things?
17:08
And it's like, you know, you can, you can get up and running
17:13
so quickly. And like people forget that most
17:16
institutions still have not implemented some form of actual
17:21
like when people say their organization is using AI, they
17:25
literally just mean right now that they just they've asked
17:27
questions to ChatGPT or Claude. Like they don't mean a lot of
17:31
these deeper things that we've been talking about on the
17:32
podcast about how do I like actually stop running my
17:36
day-to-day from like all these different app points like Slack
17:39
and Notion and Gmail and just do it all from Claude code.
17:42
No one's talking about like that's not most people.
17:46
And so, you know, I think you're completely right, which is like,
17:48
there's this like impending feeling of like, if you don't
17:51
know the latest cutting edge thing, you're behind.
17:55
And the answer is you're not. And the, but the big thing is,
17:58
is how do you make sure you approach this stuff in a way
18:01
that doesn't feel so daunting? And our that's that's what we're
18:06
obsessed with is like there's just two very obvious big
18:10
problems at the same time, which is that existing solutions need
18:15
to figure out how to make their products very easy to use from
18:19
what is a now a new interface to do it, which is like clawed
18:22
code, clawed Co work, clawed desktop, all of those things.
18:25
And secondly, how do they make it so that they still are
18:28
getting the value out of it? But from the reverse side, how
18:32
do you make it approachable, right?
18:33
It's so powerful, but how do you slim it down and make it
18:36
approachable for you? And that's really what we're on
18:39
a mission to do. And a lot of that is just like
18:42
even knowing how to ask the first question, like I'm, you
18:45
know, I'm curious for you, how did you get into all like, how
18:47
did you actually build your work flows?
18:49
Like how did you get into it? Well, through curiosity,
18:56
experimentation, and through this paralyzing feeling that I'm
18:59
behind, you know, it, it's actually really encouraging to
19:03
hear you say that. You know, I think the, the tone
19:06
of the country as far as, you know, our corner of the nerd
19:10
universe is I feel like I'm behind.
19:13
And you know, I, I feel very fortunate that I get to talk to
19:17
Mitchell and guys like you every single day and have great
19:20
conversations and figure out what's going on.
19:22
And I still need to be reminded that like, it's not too late.
19:25
You're not behind it. You're like, just just keep
19:27
moving. And I, you know, I have 4
19:29
terminal instances open and I'm just ripping all day long, you
19:32
know, now, no, so that's for me. Now I also because I'm putting
19:36
out tons and tons of content, like a, a lot of friends and
19:39
people reach out to me and they're like, hey man, I like
19:42
what you're doing. This is what I'm doing.
19:44
I feel behind. That is the tone.
19:47
So for you, I know that like you on the cutting edge of this
19:50
stuff and I know for a fact you feel like you're behind.
19:53
So what do you do to kind of remedy that feeling inside
19:56
yourself? Do you take deep breaths?
19:58
Are you doing push ups in between phone calls?
20:00
What's going on? Yeah, man.
20:02
Well, I, I do. I try to work out every day at
20:04
at night, you know, me and a couple teammates, we just go to
20:08
the gym at 8:00 PM every day. We get a sweat in and then we
20:11
come back to the office. So that's one, your health is
20:14
important. Two, sometimes just just hop off
20:17
the keys. Like take a day off of social
20:20
media is usually one of the really good ones.
20:23
And then you recognize, wait a minute, like life is good.
20:26
And then the third one is, you know, it's, it's really one of
20:30
the systems that we've built up here as an organization is it is
20:35
a, an, an expectation for us to share what we are like what has
20:41
helped us use tools better. So I think the best teams and
20:45
organizations, they don't really set goals, they set systems.
20:49
And the biggest system right now is actually knowledge share.
20:53
We're in the early stage and tools change so much.
20:57
They do change often, but like, you don't want to just
21:01
necessarily always use the thing that is the newest piece because
21:04
per everything in the world, it might have bugs, it might be
21:08
broken. And also when you're building
21:10
something, especially something that we'd call production ready,
21:14
you, you can't really afford for it to break for your users.
21:17
You can't expose their data in an unsecure way.
21:20
There's things you actually have to do properly, but what we do
21:24
oftentimes is as teammates are experimenting and are encouraged
21:27
to experiment and do new things, they share their winnings back
21:31
and they share their learnings back.
21:32
And then what we do is effectively, you know, once a
21:35
week we are like, well, what have we learned?
21:37
What have we changed? How do we add that to core files
21:40
and memory pieces that allow our whole organization to by default
21:44
use something. So it's not AO you can choose to
21:48
use something or not. It's like you, we have now
21:51
implemented this into the defaults.
21:52
So whether that's the MD files, whether that's like, hey, we
21:56
actually don't use, you know, for example, like here at Lava,
22:00
everyone works from command line.
22:01
Like if you're someone who's like, no, no, no, like I want to
22:04
write my messages in Slack and I want to type my Notion docs by
22:06
hand. And like, I want to like do my
22:08
emails by hand and I want to do all these things not in like the
22:12
central terminal. We're like, hey, like that's not
22:14
actually how we we do things here and we know that's going to
22:17
be a learning curve up front, but it's also because we want
22:19
you to use our own product. And so that's like a, a very
22:23
important thing is you got to set a system and that system
22:26
makes it less daunting because you didn't get this like hive
22:28
mind of thing. For folks who are hobbyists or
22:30
working on their own, though, the biggest thing I would say is
22:33
like find a couple folks who you actually think are very
22:37
interesting. Follow them on X or LinkedIn or
22:39
whatever, whatever they're on, and just like follow those
22:42
people. Like get a knowledgeable like
22:44
reference that you think is pretty good at this stuff.
22:47
And then just try to try out the things that they they suggest is
22:50
probably the simplest way to do it.
22:52
You're hearing that. Sign up for the newsletter at AI
22:54
for founders.co because you'll find products like Lava every
22:58
single week. Exactly.
23:00
Coming to you hot off the presses.
23:04
Amazing. OK, I love this.
23:06
Now a part of, you know, feeling like you're the last to know it
23:10
is largely the, the, the terrible thing about that is
23:13
that there is so much to know and that when you're trying to
23:16
think about products and solutions and you're, you're
23:20
tackling different things. If you don't know what you don't
23:22
know, then you can't even expand your consciousness to, to have
23:26
that kind of information. So let's just not say
23:29
consciousness, but let's just say data points.
23:31
Now you mentioned some really cool APIs that you can query.
23:34
So what are some other ones? Like what are what are some data
23:36
sources that maybe I've never even thought about that with
23:39
lava are now available to? Yeah.
23:42
I mean like there's so many, like we think about it almost by
23:44
use cases internally and you know many, many different ones.
23:49
Like if you're doing go to market work, you can enrich data
23:52
with Apollo. If you are needing to do neural
23:56
web searches, we have exasearch that's native.
23:59
If you need to obviously access LMS and models, every single LM
24:04
and model is now accessible to your agents.
24:07
So like, let's say you want to use use Claude code as your base
24:10
coding agent, but you want to quickly consult with GT 5.4, but
24:15
you don't have any setup. You can now say, hey, Claude,
24:19
actually ask this question to GPT 5.4 and you guys can
24:23
deliberate on it. You can do all of that directly
24:24
from our product. So we really think about it by
24:28
use cases, right? Go to market is one of the
24:30
really, really big ones that a lot of people like to use our
24:32
product for. So how do I use this product to
24:35
send emails, draft like follow-ups, enriched contacts,
24:41
do pipelines and store them? A lot of people do that through
24:44
our gateway Finance is another pretty popular one right now.
24:48
We have a lot of different financial tools and I think
24:52
that's one that we've we've seen kind of continue to grow and
24:55
it's because we have a lot of different like data sources.
24:57
So I think I mentioned like Edgar before, we have like the
25:02
BLSBLS data, a lot of the things that like you're often times
25:06
going to be using as a as an investor, we can have prediction
25:09
markets. Pardon me.
25:10
Yeah, exactly, exactly. And so it's like those are the
25:14
things that we really like to do.
25:16
And even even tools like Frankfurter where you can get
25:18
currency rates in real time and exchange rates in real time.
25:20
Like we have a lot of different things like that.
25:23
And one of the things that's pretty clean is you can
25:26
literally just download the Lava MCP and literally just ask the
25:30
question, what are all the tools you have for finance?
25:33
And it will literally spit them out, right?
25:34
Oh, we have Alchemy and Dune for crypto slash blockchain.
25:38
We have polymarket, we have Kaushi, we have, you know, all
25:41
of these different places where you can get data.
25:43
Yahoo Finance like we try to, we have all of those pieces of
25:46
information. So now instead of you having to
25:48
fetch and do things, you can just call them directly from
25:51
your cloud code agent. And you know, the value on the
25:54
other side is we're just trying to help you.
25:56
And as we see more and more of that traffic, what we do is, you
26:00
know, we just work with the provider to say, hey, like these
26:03
are your APIs and your services. If you want me to help you
26:06
monetize on top of those, right? We're, you know, we can, we can,
26:10
we can do that on your behalf, right?
26:11
Like we can do all of that through our gateway.
26:14
And so that is the the big thing we do.
26:17
And like, that's what we enable you to do.
26:18
It's because it's already priced in and because all of these
26:21
folks who are using our product already have wallets, it's
26:24
actually very, very easy to do that for our our customers, both
26:28
end users and the merchants and service providers that are
26:31
sharing their services. Amazing.
26:33
OK, so something like this. I got the terminal open.
26:37
I hit whisper flow. I'm like, OK, what I want you to
26:40
do is pull Mitchell's headshot from Crunchbase.
26:44
I want you to do iterations of of a sketch design.
26:49
Give me one from I'll ChatGPT. Give me one.
26:52
Give me two from Gemini. I want you to go into the
26:54
script, add 20% compression to his voice.
26:57
I want you to check pitchbook for his his leading investor.
27:01
Give a little summary of that when you're doing the show notes
27:03
and incorporate that from the transcription from the script.
27:06
Exactly. And So what, and what really
27:09
actually happens there on the back end is like, you know, we
27:12
take your query, we look at that query and we say, OK, so he
27:15
mentioned Crunchbase, he said voice notes that's probably
27:18
going to need to use retail or 11 labs.
27:20
We got that covered. And then it breaks down all of
27:23
the pieces of that, then actually then sequences it for
27:26
you and says, well, the order we need to do this in is AB and
27:29
C&D. And then it says, OK, all of
27:31
these services we have, let's just say we already have in
27:35
memory that you prefer to use 11 labs for, for voice services.
27:40
We'll do that and we'll break it down or you use retail for, for
27:44
translation, we'll break that down.
27:45
And then it does that automatically for you.
27:47
So instead of you having to say, Oh, well, I don't have a, an 11
27:51
labs account or a retail account or, or all of those pieces, it
27:54
actually just goes into end. You were able to one shot that
27:56
thing that you weren't able to do before because of the fact
28:00
that, oh, I didn't have all the keys.
28:03
I need to go sign up for the service.
28:04
Claude got blocked because it's a paid thing, right?
28:07
Well, you already have a wallet through lava.
28:09
You've already have your credits.
28:11
You can always recharge and add more.
28:13
And then we also, you know, we're handling the API payments
28:16
for you, right? Because for the service
28:18
provider, we're already paying, they're, they're already
28:21
charging us for what you're doing.
28:22
So it's clean because it essentially allows both parties
28:27
to both list their services and other people to use their
28:30
services. It's that simple.
28:33
So cool, man, it's so cool. It's and it's basically, you
28:36
know, and even if you have accounts, you don't have to log
28:37
in and toggle tabs and go back and forth and copy paste and
28:41
blah, blah, blah. One shot the whole thing to your
28:43
whole tech stack and then it delivers it back in one place,
28:47
you know, and. And like, and if you think about
28:49
it like what, what are we doing? Like, what are we?
28:50
What is the point of like, you know, we're trying to move up a
28:54
level of abstraction, right? Like natural language is the
28:57
best form of coding. Natural language is the best
29:00
form of working, right? Like when I feel connected to
29:03
you, it's probably because we have some ability to connect
29:06
through language. Oh, I learned a little bit about
29:09
what you care about. I learned about what you
29:10
understand. And the more we move towards
29:14
that as the just the purest thing and not having to workshop
29:17
and move tools around and do all those things, that is where we
29:20
need to get. And our job is to make sure that
29:22
for the average person who kind of like we were talking about
29:25
feels behind, it's like, no, no, no, no, you're not behind.
29:28
And actually, we actually now have the version.
29:30
It's probably the cleanest for you all to use these services.
29:34
And it's like, so actually you, you, you get a second mover
29:37
advantage here. Other everyone like, you know,
29:40
Ryan and and Mitchell ran in were like had to do all this
29:42
annoying, annoying stuff. And then, you know, now there's
29:46
all these services that make this easier for you.
29:49
Like we have curious actually for you, I was wanting to say
29:52
what's one of the most annoying things you've ever had to set up
29:54
with with AI? Boy, the most annoying I, I
30:00
think it's really probably just the daily toggle back and forth.
30:04
I'm always looking for tools. That's like predictively telling
30:07
me it's like, dude, I know you're going back to the
30:08
spreadsheet. Do you just want me to automate
30:10
that? Like predictive automations and
30:12
all the toggle swapping. So like I'm already thinking
30:15
about lava just just simply for for podcast production because,
30:19
you know, all the tools make it very fast.
30:20
But if I could just have my prompt that is just like, oh,
30:24
it's a skills to be like I every time I stop hit recording, it
30:28
just automatically produces everything.
30:30
And then, you know, in an hour, I come back, I'm like, that's
30:33
beautiful commit, you know, I mean.
30:37
What? So I need to ask you then like
30:38
what are all the tools you're using so we can make sure
30:40
they're in a gateway so you're ready to go?
30:42
Yeah. You know the big one for the
30:44
podcast production is descript. Do you guys use have descript?
30:48
We don't yet, but that's actually a huge one, like, and
30:51
then the other thing too is I think the, the thing that's
30:53
pretty valuable is so long as there's an API, we can do it,
30:56
right? I think that's the, I think
30:58
that's like the powerful thing. I think right now we're in a
31:00
very interesting world where a lot of people are trying to
31:02
figure out how to give agents money and cards and all of these
31:06
things in order to enable them to pay for things.
31:08
And we actually have a different approach.
31:10
We're like, you know, usually people don't want to do all that
31:12
stuff. If you just make it really easy
31:14
where it's like I handle the payment and like, yes, I'm
31:17
getting billed by descriptor, whoever that is.
31:20
Everybody else gets to just keep doing what they're doing.
31:23
We're the ones who help you use your APIs and make them agent
31:26
friendly. And then we can do the rest.
31:28
And then Ryan gets to then just say I created a skill and the
31:31
skill already understands that it uses lava to call Descript.
31:35
And then it's it's it's it's that pretty, it's that easy.
31:37
So, so the answer is, yeah, we'll be on it.
31:40
We'll have it by the end of the day.
31:44
It's such a good point. You know, I mean, a couple of
31:46
things come to mind. One, this is also very
31:48
encouraging for people that feel like they're behind because to
31:50
that point, you know, I had to bunt to get on first base.
31:53
I had to steal second base and somehow it threw an air.
31:56
I'm on 3rd. You can enter the game on third
31:58
base right now, you know, So that's really exciting.
32:03
That's a great way to put it. It's like, yeah, we we had to
32:06
fight and claw. We still haven't even gotten a
32:07
run in, but we had to fight and claw to get to 3rd and like now
32:10
you just hit a triple and it was it was easy, right?
32:13
It's. No problem.
32:14
So, so that's that's cool. And I really like your point
32:17
too. And I've spoke to folks that are
32:19
building kind of like, you know, blockchain enabled, like
32:22
wallets. So, and the use case is always
32:24
like, oh, if you want to book your travel thing, it can pay
32:27
for you or if you want to order DoorDash, it could pay for you.
32:31
But for me, I'm like, well, one, I cook and two, I'm the only
32:34
travel I do is hunting and fishing and you're not buying me
32:36
tickets to that. It's like, that's not going to
32:38
really help me. But if it removes barriers in my
32:41
workflow and makes everything much easier, you know, so like
32:45
literally I stopped recording and all the assets are ready to
32:48
go. I mean, that is incredible.
32:51
Like truly like that is that's the, that's the world I'm
32:54
waiting to exist with AI, you know, it's like, it just, it
32:58
understands me and then I can subtly move and change things
33:02
with different skills and it's butter.
33:04
Yeah, and like you're, you're, you're bringing out a good point
33:07
too, is like, in some ways it, although people forget this,
33:10
we're still in the toy face. Like people are still trying to
33:14
like, build stuff for problems that don't really exist yet.
33:18
Like, I still really don't need an agent to buy me a shirt.
33:22
Like I still need to look at the shirt and see if I like the
33:25
shirt. And then by the time I see that
33:27
I like the shirt, I'm on Shopify's web page or I'm on the
33:31
merchant's web page. And guess what?
33:32
They probably already have my card saved on file.
33:35
Like, so a lot of these things we're still, but people are
33:39
still trying to like add new things that we don't necessarily
33:43
need yet. And again, like, I think, you
33:46
know, if, if people take away anything from this podcast, it's
33:49
number one, just just dive in like it's you're, you're and
33:53
you're not as behind as you think.
33:54
And then #2 is just you're at the end of the day, you know,
33:58
Ryan, I think you maybe made the most golden point.
34:01
Screw having me on the podcast. I think you made the most golden
34:03
point, which is that the beauty of AI is that companies are
34:07
trying to make it easier and more accessible to use.
34:11
So long as you're just staying abreast of like what's making it
34:13
easier to use, that's probably the most important thing you can
34:17
do, right? And and that's, that's really
34:20
it. It's like, how are AI teams
34:23
trying to make this actually fit back on the way people work
34:27
today is going to probably be the the real unlock.
34:30
And you know, I think that's AI think it's a very astute point.
34:33
I think that's that's probably how this is going to end up
34:35
looking. You know, let's, let's take this
34:38
to the, the other side of this marketplace too, because where I
34:42
see everything going is people are creating agents and there's
34:45
going to be a, a marketplace for agents.
34:47
Now, the use case I'm really excited about right now is
34:50
podcast production, right? So it's like this, I feel like
34:53
Lava could really, really help and speed me up.
34:55
Now the, I have a, an advantage if I'm developing a agent for
35:00
podcast production because I made 500 podcast episodes.
35:04
So it's like I know what I want and what what Spotify wants and
35:08
what YouTube wants and what Apple wants there.
35:11
All these things go into that, such as to say I have domain
35:14
expertise. So when I create this agent,
35:16
it's going to be sick. And so once I have it kind of
35:19
polished into place, it's like, OK, this is working great for
35:21
me. Why not bring this to the
35:23
market? So let's say, OK, now my podcast
35:26
production agent I want to bring and give to other people like
35:29
here's here's the stack, blah, blah, blah.
35:30
I license it through lava. So now with Lava I can basically
35:35
if it's API enabled then I can sell it to other people for a
35:39
per use basis. The way that I was using all the
35:42
lava tools as well. Is that kind of how it works?
35:44
Correct. And you can also, if you want
35:46
to. So think about it like this,
35:48
let's just say you have a model where somebody can call your
35:52
like. Let's actually really break this
35:54
down to what agents really are. They're just chained calls on
35:58
top of a APIs, right? And then like, you know, we then
36:02
expose them to an end user in the form of a CLI or an MCP.
36:08
But really all it is is there's APIs.
36:10
We build layers to help people understand those AP is and then
36:15
we sell them that that is. And then agents then use that
36:18
understanding layer to call AP is that is really what is
36:23
headless SAS. That's really what all of that
36:25
is effectively doing. And so Ryan, you're, you're
36:28
making this, this agent that is literally doing that.
36:31
Then you're selling it off as this package of your your API
36:34
calls that someone else can use in their day-to-day from Claude
36:37
code. And let's just say you want to
36:41
enable them to do 10 actions before they then have to pay you
36:46
a subscription, right? And so those 10 actions are
36:49
just, at the end of the day, API calls that are getting counted
36:53
on a counter. And then after a certain amount
36:56
of them, then you want to actually have a what is
36:59
effectively a paywall of some sort, whether that's an e-mail,
37:02
whether that's a hard block, like whatever you want to do.
37:05
What Lava does to make that very easy for you is we've already
37:09
tracked all of that usage per person.
37:13
And then when they hit that paywall, we make it easy for you
37:16
to monetize on top of that. And we can do that by just
37:21
blocking the traffic. Now, one of the things that make
37:24
that's very powerful is there's a trillion ways to do this
37:27
already in the more traditional human billing sense.
37:30
You can go use Stripe and set up a plan and count things and do
37:34
all that for an agent that is actually your new pie, which you
37:38
would love to monetize. Somebody finds you through an A
37:41
clawed code agent and they start doing this.
37:44
Well, there's a whole freemium problem, right?
37:47
How are you now? You're gonna, they have to now
37:49
go to your website, sign up, start integrating from there.
37:52
Or were they able to just the agent be able to just
37:55
autonomously do what it needed to do without interrupting a
37:58
human? Because it notices.
37:59
Wait, you already have access to this service in Lava?
38:03
And then once you hit the limit, we just ask you, well, if you
38:07
want to keep using Ryan's service, you have to actually
38:10
pay $20.00 a month for it. And then because you already
38:13
have a wallet, you add $20 into your wallet and we just remit
38:16
that back to you Ryan. That's it.
38:18
And so now you didn't have to learn how to make agents pay on
38:21
the blockchain. You didn't have to learn any new
38:25
primitives. You just built an you built AP
38:28
is you didn't expose those AP is to agents or actually in matter
38:32
of fact, you didn't even need to do it.
38:33
We did it for you and now people are using your product.
38:36
That is the magic of how we we build lava and the real pinch
38:40
there isn't just for the AI forward companies like yours.
38:44
It's it's for the legacy company.
38:46
If you're a legacy company and you're trying to figure out
38:48
right now how the hell do I monetize my service and how do I
38:52
start converting all these agents who are using all my
38:54
data? Well, the way you do that is you
38:57
put your data on a service like ours where we can securely
39:01
manage the flow of that traffic and also convert those those
39:05
usage into payments. And that's what we do.
39:10
So cool. And there's like an arbitrage
39:12
play here too. I could see is your user and
39:15
like a seller. If you're on both sides of the
39:17
table, then basically you've got your, your lava balance and
39:20
like, OK, I'm spending 10 bucks a month on lava, but I'm earning
39:24
30 bucks from lava on the backside.
39:27
You know, So as you're spinning up tools, then you're not kind
39:30
of like reconciling like, Dang, I'm spending 500 bucks a month
39:33
on all these AI tools. Like what's the return on that?
39:35
It's almost. It's right there in the lava.
39:37
Right there and you see exactly how much it's costing you.
39:40
So the whole thing is the whole point of this is just like we're
39:43
trying to turn what was your cost center into your revenue
39:46
center. And we do that by on both sides.
39:49
If you're just a user of tools, you now don't, it doesn't cost
39:53
you to get up and running. If you're building a product,
39:56
you can track everything that's happening in your product.
39:59
You can label it so you can understand the cost of it and
40:02
then you can serve it out. You can then charge whatever
40:04
price you want to on top of it. And we do all of that for you in
40:07
one fell swoop. And you can do all of that
40:10
directly from Claude code. You can also do it if you want
40:13
to. The old fashioned way you can go
40:14
to. Www.lava.sosignup and like do
40:18
all that lava dot. So I shouldn't say lava dot so
40:21
and you can do all of that, but our goal is just to make this
40:24
very simple for you. And like, we understand how
40:26
people are building things these days, and we just want to make
40:28
sure we're aligned with that. Yeah.
40:31
Well, it sounds like, you know my assumption that there's going
40:34
to be an agent's marketplace. I was behind and you, you were
40:39
on top of it. But like, yeah, this is, this is
40:41
exactly what is the necessary step as people are getting, you
40:45
know, more and more nuanced with kind of their agent builds and
40:49
they, you know, you, you can find very specific use cases.
40:52
And I think that, you know, a year ago people were kind of
40:54
talking about that, that like, hey, man, if you've got siloed
40:57
data and you have like domain expertise around a very tiny
41:00
thing like this is where it's going to be the, the solutions
41:03
are going to arise. Because you, you know, even if
41:06
you have 1000 customers, like the, the value of what you're
41:08
providing is going to be huge. And if you just tee it up for
41:11
them, you're like, boom, there you go.
41:12
And again, you're, you already have the in, in that aspect, you
41:16
already have the vault, you already have the walled vault.
41:18
What you just need to do is you're just trying to figure out
41:20
how do I expose this to an agent and get paid for it.
41:24
And you know, the answer to that is like, well, there's, there's
41:27
two parts of that. And we, we, we solve both of
41:29
those for you as a, as a, as a, as a company, it's like 1, you
41:33
need to make sure it's metered and measured so that they only
41:35
get a certain amount. And two, they need to pay for
41:38
every amount that they get, whether that's me paying on
41:41
their behalf for you or that's them paying me.
41:44
And I'm remitting it back to you.
41:45
However you want that to happen. You need to get paid for what
41:48
you're what's rightfully going through.
41:49
And you need to be able to see how much usage is happening and
41:52
how much is happening in real time.
41:54
And that's what we really saw for the, for the service
41:56
provider or the OR the OR the startup or the company, whoever
41:58
that is that wants to list services.
42:02
But, you know, we just see, we see that there's two things.
42:04
And there's a is a very clean way to make this work for both
42:07
parties without having to learn a bunch of new stuff.
42:10
Cool. How's the remittance paid out?
42:12
Is that AAPI call to my bank? We do old fashioned.
42:16
We just we, we, we, we, we get a connection to your bank and then
42:19
we, we have a Ledger for you because everyone's just got a
42:22
wallet and then you can just cash out your wallet like you
42:25
would with any other service. So very, very simple, I look
42:29
like you're, you, you have a Venmo balance or a cash app
42:31
balance. And that's somebody, imagine it
42:33
like this, someone's Venmoing you.
42:35
Over time, that balance is getting bigger and bigger and
42:37
bigger. We have a saying, it's called
42:39
being Venmo rich. And now you're Venmo rich.
42:41
And then you say, you know what, let me cash this out to my
42:43
actual bank. It's the same thing.
42:44
And then of course for like, you know, for merchants, we have a
42:48
pretty really, really, really ironclad record of payment,
42:52
which makes it actually easy for you to do just like any types of
42:57
financial operation to understand and do all your
43:00
accounting and all those pieces as well.
43:01
So like that's the, that's the other big thing there is that
43:04
like a lot of tools weren't really built for like how to
43:07
like do Rev rec on like usage based pricing and like more of
43:11
this incremental based pricing. And you know, that's, that's
43:13
another thing that we we try to nail for our customers.
43:16
Yeah, man. I mean, it's worth the the price
43:18
of payment if I just never have to touch Stripe again, you know
43:21
what I mean? Yeah, Yeah.
43:22
Tell me about it. And like this is, I think that's
43:24
the thing that we love about this is it doesn't actually
43:25
compete with Stripe. Like if you want to keep billing
43:28
humans, great. Like that's the way to do it.
43:31
I think the big thing we're saying is, hey, there's an agent
43:34
that's trying to use your site right now.
43:36
You have no ability to convert that agent who's currently right
43:39
now trying to use your crap traffic into revenue.
43:42
That's what our goal is. So that's actually just an
43:44
additional revenue stream for you.
43:46
It's not necessarily competing with your existing customer
43:48
base. It doesn't have to like if you
43:50
want to use us, of course you can use us for that as well.
43:52
We'd love to work with you, but we also can help you just with
43:55
converting more users and who doesn't want more money on their
43:57
product. And that's really what we try to
44:00
focus on on for that as well. It's like our goal is to try to
44:03
turn what it's been like your biggest problem and to your
44:05
biggest advantage. Right now, a lot of companies,
44:08
especially legacy companies, are like, my biggest problem is that
44:11
people are scraping my stuff and then they're like trying to
44:13
resell it. And I have no ability to fight
44:16
back. And what we try to do is say,
44:18
hey, like, you know, well, a lot of that is because you're trying
44:21
to bolt this off and like, you're doing that because you
44:24
have no way to make the agent pay for the thing.
44:27
And so if you actually put it in a secure store, right, agents
44:31
are going to access information a lot.
44:33
You didn't have actually a very secure pipeline for them to be
44:35
able to go and pay you for it. And then we make sure you're
44:39
getting paid every single time they're taking a piece of
44:41
information. And what are the pricing rules
44:43
around that? They're the same as what they
44:45
are for your AP is. And that's really what we've
44:47
been able to build and we make it very easy for you to to be
44:49
able to adopt. Sounds like you might be talking
44:52
about Apollo. I wonder, I wonder if if you
44:56
have the capability to say like, OK, I don't want the
44:58
subscription and all the things from Apollo.
45:00
What I want to be able to do is filter Apollo's database and
45:03
just get the top 12 examples on that query from that thing and
45:07
bring that in and only pay for that.
45:08
Is that a possibility? So the big thing with tools like
45:12
Apollo is the one thing you have to make sure you want to do is
45:15
we, there's a lot of like, you know, a lot of people right now
45:18
who are very, we'll just scrape all these other tools and then
45:22
we'll resell all their information.
45:24
I think that that like can get you a lot of usage right now.
45:27
I think that that in the long run, it's not very on the side
45:30
of making a marketplace that's fair for both sides.
45:32
So what we do is it just it's like per provider.
45:36
So what we do is we will always make sure we're kind of doing
45:40
the things that we know are like already use its base and pay as
45:43
you go. And we'll make sure those are
45:45
available in our existing product.
45:46
And then like if you if like a service provider like Apollo
45:50
then wants us to also be able to give people access or convert
45:54
them to their like paid tiers and things like that, We'll just
45:56
try to, we just try to make sure we work with them and make that
45:58
very clean and easy. And so then it's like the other
46:01
thing too is like we actually have a like, if you go to our,
46:05
our website, you can actually go and request us to do your AP is
46:08
for you. So it's like, let's say you just
46:10
want to like make your AP is agent agentic.
46:12
You can just go to lava dot. So log into your dashboard, go
46:17
to storefront and you can just request us to do your AP is for
46:19
you. It's that simple.
46:20
And then we'll go and do them for you.
46:21
We'll let them know when they're ready and then you claim your
46:24
records and then you can see all of the data that's flowing
46:26
through and you can start getting paid on it.
46:28
So. Man, you said Venmo rich.
46:32
You know I always keep 100 bucks in my Venmo wallet.
46:34
You know why the? Venmo Rich.
46:38
Well, it it's also to rub it into my son because we have $5
46:42
bets that go back and forth all the time.
46:43
And honestly, I probably owe. We never really pay out.
46:46
It's just I bet you 5 bucks. I'm usually losing that bet
46:48
these days. But he bet me 5 bucks that I
46:52
couldn't run a 12 minute mile. And I was like, I was like,
46:56
dude, 12 minutes. I was like, I was like, bro, I
46:59
do yoga like everyday. He's like this not cardio.
47:01
I'm like, dude, OK, OK, that, that that's insulting.
47:04
I'm not, I'm not doing it for a bet.
47:06
He's like, I'll bet you 100 bucks.
47:08
And I was like, let's go to the track right now.
47:10
Went to the track, I I did a 7 1/2 minute mile and almost died,
47:16
but I made him pay me. Now, typically I wouldn't make
47:18
him like pay on the bet, but I was like, no man, you need to
47:21
learn a lesson that you could never count your dad out.
47:24
OK, yeah. So now whenever like he's like
47:26
questioning me on some stuff, I'll pull up that Venmo balance
47:28
and be like, oh, need a stable. Wait, question for you.
47:31
So did you like, rub salt in the wound by taking him out and get
47:34
him like a burger at like, I don't know, like Shake Shack or
47:37
something afterwards? Yeah, man, I still have money
47:39
leftover. From I think for me, I think we
47:42
got like a smoothie or something.
47:44
But yeah, you know, Mrs. Estes was there too, so I want to make
47:47
sure she witnessed this. And it was funny because he was
47:49
like, I'll run with you. So he was on the inside of the
47:51
track as I'm running. And he's like, you're not going
47:53
to make it. You're not going to make it
47:54
after like the first lap. He's like looked at the pace.
47:57
He's like I'm. Let me, let me get that.
48:01
Let me get, let me fire up the keys on that Venmo request.
48:04
You know, it's funny, my dad always, he's his, his I never
48:07
bet my dad. And the reason why is I remember
48:10
when I was young, I bet him once and I absolutely lost.
48:14
And he's like, I, I don't, I only bet on things I know I'm
48:16
going to win and what, and that was a lesson I learned from him.
48:21
He never. So then that was it.
48:23
That was it for me and betting my pops.
48:24
It was just like, I can't. And it's just it's not it's
48:27
often not a a winning battle. I think it's it's actually so
48:30
funny too. It's like, you know, it like not
48:33
like outside of just all this AI stuff, it's like it's just so
48:36
cool and interesting to see so many spaces take off, you know,
48:40
like I think one of the things it's been interesting to see and
48:41
it's going to be interesting to see how it plays out.
48:43
You know, there's a lot of regulation things are going to
48:45
still come down the pipe with AI, but you know, it's going to
48:47
be interesting to see what happens also in this like, you
48:50
know, prediction market space as well.
48:51
When we're, when you're talking about betting, it's one of the,
48:54
the, it's one of the things that people call from our gateway, a
48:56
lot like they call Kaushi and Polymarket and pieces like that.
49:00
And like, you know, it's, it's, it's going to get into an
49:04
interesting point of like spirit of the law versus letter of the
49:07
law and like, what are these things allowed to do?
49:11
But I, it's going to be an interesting to see how that
49:12
shapes out in the next couple of years when it comes to things
49:15
like, you know, bets and, or predictions and, and, and things
49:18
of that nature. Like the the the prediction on
49:22
polymarket, you know that they took off, I think it was last
49:25
week, which was the probability of nuclear war.
49:30
Let's take that one off. Well man, you're bumming me out.
49:34
Yeah, I know, But I guess isn't that what the Doomsday Clock is?
49:37
Yeah, that's true. That's always, that's always
49:40
ticking. So why?
49:41
Why not, you know, you know, be like you just can.
49:44
You can you can go out in the mushroom cloud with a big grin
49:47
on your face. Oh, I made 100 bucks.
49:49
Yeah, I've been made 100 bucks, but you know, let's, you know,
49:53
we're on the side. We're we're on the side.
49:55
I'll take the I'll take the under.
49:57
It's not going to happen. We're going to we're going to be
49:59
able to learn a lot of new stuff.
50:00
Abundance, abundance, abundance. And you know, we humans, humans
50:06
figure stuff out all the time. So we'll, I'll, I'll, I'll root
50:09
for that side of humanity on that one.
50:11
Me too Mitchell, this has been such a fun conversation.
50:14
I really appreciate it. Congratulations on your awesome
50:16
business. Thank you so much and Ryan,
50:19
thank you so much for having me on the pod and everyone
50:21
listening. Remember you're not that behind.
50:24
If you do want to try to learn how to catch up a little bit on
50:26
the end user side, just go to Lava dot SO and get signed up
50:30
and you can install our MCP and and get running and give your
50:34
claw code codex instant superpowers.
50:37
And if you're a merchant or a business service provider
50:40
wanting to make sure you can actually capitalize on agentic
50:44
traffic and upsell that into the existing APIs and services you
50:47
already have, you can also get signed up on Lava dot SO or you
50:51
can always reach out to me on LinkedIn.
50:52
All right, thank you guys so much.
50:54
That's enough of my my pitch there.
50:56
The big take away. Take a breather, get a workout
51:00
in, get back on claw code, learn some new stuff.
51:04
You're not that behind. Words of wisdom that I needed to
51:08
hear today, Mitchell. I appreciate it.
51:10
Before we get out of here, I want to give a quick shout out
51:12
to Saint Jude's Children's Hospital, which I know is near
51:14
and dear to your heart. I'll link to them in the show
51:17
notes. I'll have the link to your
51:18
LinkedIn and Lava, of course, in the show notes as well.
51:21
Thank you so much, Sir. Awesome.
51:23
Take care. You are about to spend anywhere
51:26
from 100,000 to $300,000 on your kids college education, and
51:31
that's not a small bet. And right now there's a very
51:33
real chance that they graduate and can't find a job.
51:36
Not because they weren't smart enough, not because they didn't
51:39
work hard enough, but because the job market changed and
51:41
nobody told them. Employers are hiring AI native
51:45
candidates, people who don't just know their field, but know
51:47
how to work with AI at a professional level.
51:50
And most colleges are not teaching this.
51:52
Not even close. AI native student.com is the
51:56
program that fills that gap. It takes high school and college
51:59
students and teaches them how to actually think and work with AI
52:02
the way top companies expect on day one.
52:04
So your investment in their degree actually pays off.
52:08
You're already paying for tuition, housing, textbooks.
52:11
This is the one thing that actually makes it all worth it.
52:13
Go to ainativestudent.com because a diploma is a starting
52:17
point. Being AI native is what gets
52:19
them hired. That's AI native student.com.