Latest / AI for Founders with Ryan Estes / Speed is killing AI startups

Transcript

Speed is killing AI startups · AI for Founders with Ryan Estes

0:00
Speed is killing more AI startups than the competition
0:02
ever will. Everyone's sprinting to ship
0:05
faster demos, faster features, faster funding.
0:08
But speed without structure is just expensive.
0:10
Guessing. You think velocity is your Moat,
0:12
but it's not. Stability is, and most founders
0:15
are one Claude code prompt away from losing the breakthrough
0:18
that they celebrated last week. That's not momentum.
0:21
That's fragility disguised as progress.
0:23
And here's the tension that nobody emits.
0:26
You're tweeting wins, closing pilots watching usage spike.
0:29
But under the. The hood.
0:30
No clean experiment trail, no controlled variables, no
0:34
reliable way to rewind the tape. If your open claw engineers
0:38
fired themselves tomorrow, could you recreate your core insight
0:41
from scratch in AI? If your progress depends on
0:44
memory instead of method, you're building on sand.
0:47
Now imagine this instead. Every test deliberate, every run
0:50
logged, every improvement traceable.
0:53
Not just moving fast, but moving with proof.
0:56
Not just iteration, but institutional memory.
0:59
The real founder advantage isn't who ships first, it's who can
1:02
compound correctly. So are you building something
1:04
that gets smarter overtime or something that just gets lucky?
1:08
Welcome to AI for founders. I'm Ryan Estes, and this is for
1:11
builders who want leverage, not luck.
1:13
Today we're unpacking discipline in AI systems with James
1:16
Everingham and why reproducibility might be the
1:19
real competitive edge in the generative era.
1:22
If you're serious about building AI that scales without breaking,
1:25
go to aiforfounders.co. Join the newsletter with 27,000
1:29
founders thinking sharper about systems. 1 e-mail once a week
1:33
compounding edge. And if this hits, leave a review
1:35
that future you would be proud to read.
1:38
If you're building something big and want to get on the world's
1:40
best podcast, go to kickcaster.com.
1:43
We book founders on top shows that drive authority, revenue
1:46
and investor attention. You build the company, we build
1:49
the conversation. That's kick caster.com.
1:52
Kick Caster by MO Burst. James, welcome to the show.
1:59
Thank you so much for joining me.
2:01
Thanks so much for having me on. Yeah, looking forward to this
2:04
conversation, largely because you're cosigned by some absolute
2:07
Titans. So let's dive right into it.
2:10
We're going to be talking about Guild AI today, so give us a
2:13
kind of an overview of what Guild is now and where it's
2:16
going. Yeah, we've been sort of
2:18
operating a little bit in stealth, but we're starting to,
2:21
you know, pull back the curtain a little bit about what we're
2:23
building. And the story of it is, is
2:27
recently at my last job working at Meta, where I was supporting
2:31
all the developer infrastructure, we we leaned a
2:34
little bit more into agentic agentic platforms to help do
2:39
things within the infrastructure.
2:41
A lot of people at the time were leaning on coding agents, but we
2:44
were using things to do more advanced operations like in the
2:49
orchestration of our infrastructure.
2:51
And we sort of saw at scale what you need to build.
2:55
You need governance, you kind of need access, you kind of need to
2:59
understand what these agents are doing.
3:01
You need to give them, you know, you need to control them like
3:04
almost like you would humans. And there wasn't a platform out
3:07
there to do it. Speaking with some of the
3:09
companies that I'm close with, they were all building their own
3:13
version of this. And we're like, well, you kind
3:15
of need a control plane out there in order to help build
3:19
this and, and productionize this in an enterprise.
3:21
So that's basically at the roots of what we're doing.
3:24
There's a few more layers on top of this, including the
3:29
centralization of these agents. If you're going to manage them,
3:32
you need to sort of centralized them.
3:34
You can think about this a lot like when the iPhone first came
3:37
out and there were no enterprise controls and people were just
3:41
installing apps on their phones. Nobody knew what was happening
3:44
with their enterprise. It wasn't trackable.
3:47
So we think that's sort of happening with agents right now
3:50
is you need something similar to a centralized system where
3:54
you're where you can go manage these things, you can control
3:57
what they have access to, very much like an internal App Store.
4:02
Amazing. So it's basically is it a
4:04
development platform to create and manage agents?
4:07
Is that safe to say? Yes.
4:09
And actually a little bit more is to integrate them into your
4:12
workflow. So it's not just to manage them.
4:14
You think about what an agent needs to be effective in your
4:16
infrastructure. It needs integration with your
4:18
tools. It need, you need to be able to
4:21
specify what agents can and cannot access.
4:24
You need to be able to track what they're doing and see what
4:27
they did. If you're going to be debugging
4:28
these things, you sort of need to understand everything they
4:32
touched and and have a log, an auditable log of what they what
4:36
they did. You need this not just for
4:38
debugging, you also need it for compliance.
4:40
So if you're in a regulated company and you need, you need
4:43
these governance layers. And that's sort of the direction
4:47
that we're going with this. Incredible.
4:49
And it's a timely conversation because the last couple of weeks
4:53
it's just been all open claw. That's, that's all everyone can
4:57
talk about it is open claw and to honestly, it kind of Spooks
5:00
me a little bit the the capacity and capabilities of open claws.
5:03
So maybe put Guild in that kind of perspective.
5:07
What's the What's the differentiation?
5:09
Yeah. So yes, I totally agree.
5:12
It's just been fascinating to watch the whole open claw thing
5:14
and it's, it's validating in a lot of ways is that, you know,
5:19
open claw basically is a personal assistant, right?
5:22
You, you put it on your computer and you give it access to
5:25
everything that you want and it sort of goes on.
5:27
It's sort of a single player thing, you know, and, but if you
5:30
think about the way these agents work in an infrastructure,
5:33
they're going to be running on enterprise servers, not on a
5:35
workstation. So first you need to be able to
5:39
have an environment that hosts those.
5:41
You don't want them just running on an instance of someone's
5:43
laptop or or desktop machine. You need a cloud environment to
5:48
host those in. And then as we've seen, like
5:51
with the open slot movement, there have been some security
5:54
problems, little ones, you know, and yeah.
5:57
Little. So you need like a layer to
6:01
provide that security and control what they have access to
6:04
and that. So like that's sort of the
6:06
difference. I, you know, I think that what
6:09
enterprises want is they want to have a variety.
6:12
They don't want to like be hooked to 1 vendor.
6:14
They don't want to be set in one model.
6:16
You want to be able to, you need a runtime basically that will
6:20
handle all of these different things.
6:22
Open Clot, you can think of as another runtime.
6:25
Clot code can be a runtime, but you need a centralized place to
6:29
run and manage those things. So you know, we we think that
6:33
that that is very complementary to that type of that type of a
6:37
system. Yeah.
6:38
So they're basically owned, operated, managed internal for
6:42
internal use only kind of things.
6:45
Yeah, Good. Then that kind of illustrates
6:47
kind of a kind of a lot of blowback that's coming now.
6:50
So if open claw people really excited about the fact that
6:53
like, hey, they can be detached, we can I can give it all my
6:57
personal data. It's not going to ping any of
6:59
the big LLMS because I want to maintain my privacy in some
7:03
capacity. And enterprise is kind of
7:05
learning the same things. They're like, well, you know, at
7:08
some point, you know, Open AI, Google, Gemini, Anthropic, all
7:14
these people might have a vested interest in being a competitor
7:17
of my company. Do I want to be sharing all the
7:20
information? So this is a way to kind of
7:22
shore up that exposure to say like, oh, we're going to keep
7:24
all this in house. Is that a good way to look at
7:26
it? Yeah, I think.
7:27
That's a that's a good way to look at it.
7:29
And you know, not only keep it in house, but like, you know,
7:31
you, you also need to control access and like have an
7:35
auditable and loggable, you know, because just like humans,
7:39
if you think of agents as the way you think of employees, you
7:42
know, different employees have different access, you know, to
7:46
different parts of the system based on training or based on
7:50
other criteria. So like you also need to be able
7:52
to have that level of control. Yeah, it's really interesting
7:56
the personification that I do this all the time too.
7:59
I wonder if you do, as I'm starting to think about agents,
8:02
I'm giving them names. I'm creating like different kind
8:05
of to your point like responsibilities and access
8:09
credentials and things like that.
8:11
Is that important to the the work that you're doing as you're
8:14
building Guild? I, I think it's an interesting
8:16
way to look at it, but my worry of thinking about it as a
8:20
persona is that we end up thinking about like, this is a
8:23
way you can replace people if you start thinking of them as,
8:27
as a person. I, I tend to think of them more
8:29
as a strength that a person would have, you know, back at
8:32
Meta, I used to say, hey, we're building bionics here, not
8:35
robotics. So like the, or robots because
8:39
you know, they, they, they have like LLMS have a strengths and
8:43
weaknesses just like we do. But you have to, if you're
8:47
trying to figure out what to apply these things to and like
8:50
what you want to do with them, you, you got to look at what
8:52
they're good at. They're good at taking complex
8:54
information and making like and, and, and, and sort of making
9:00
insights based on that. It's all, it's all statistical,
9:03
you know, a model if you think about it.
9:06
So it ends up that they're really good at doing complex
9:08
simple tasks. But when you get into, you know,
9:11
at least not yet. The LLM capability hasn't got to
9:14
the point where you can do really deep, you know,
9:17
algorithmic problem solving. For example, at Meta, you know,
9:21
in dev infra we had the languages team, we had HHVM and
9:25
LLVM and we had this thing called WWW which is the this in
9:29
mind bendingly complex system and stack that runs most of Meta
9:35
software. It's so hyper tuned for
9:38
efficiency that like you go in there and try to vibe code in
9:42
the kernel like it's it's probably going to not end well.
9:46
You're going to, it's going to create a problem, but you know,
9:50
going in and like doing the boilerplate stuff, the coding,
9:53
like going through and you know, I think that we had this one
9:58
example where an engineer used an agent to go fix a bunch of
10:01
accessibility issues. You know, like 10s of thousands
10:05
of these things all at once. It's some, it's like the largest
10:08
source code repository, one of them that exists.
10:11
So going through and doing these types of things are awesome.
10:14
You know, like and that removes that's a lot of what engineers
10:18
are sort of like slowed down by is by this monotonous the ditch
10:23
digging of computer programming. So it frees them up to do more
10:28
feature development, inventing and thinking and and problem
10:31
solving. Yeah, which is amazing.
10:34
It's it's absolutely incredible. I mean, it's just in the, the
10:37
speed at which it's developing right now is just unbelievable
10:42
and really cool. Now, one thing about your career
10:44
and you have a storied career with Meta that the people might
10:48
not understand, like how influential, you know, Meta is,
10:51
particularly with, you know, AI development over the last 10
10:54
years. But when it was called machine
10:56
learning and other things, everybody kind of everything's
10:59
AI now, but people might not understand like really kind of
11:02
how innovative Meta and Facebook products were with AI.
11:07
What what did you learn from that experience and what did you
11:10
bring from that to Guild? This episode is brought to you
11:14
by a dangerous idea. It's called reading.
11:16
Specifically the AI for Founders newsletter. 47,000 Founders
11:21
strong now. And it's the good stuff, no?
11:23
Top ten AI. Tools that will change your life
11:26
written by someone who doesn't ship product.
11:28
It's just real breakdowns of what's actually working in AI,
11:31
who's raised, who failed, who pivoted, and which founders
11:35
quietly built $1,000,000 in ARR while everyone else was arguing
11:39
on. X.
11:40
It's written for people building with AI.
11:42
Every week. You get tactical frameworks,
11:44
unfair advantages, breakdowns of real companies, and the
11:47
occasional existential question about whether your startup is
11:50
intelligent or just well branded.
11:52
It's free. It's sharp.
11:53
It might mildly offend your current workflow.
11:55
Go to aiforfounders.co and subscribe before your competitor
11:59
does because the only thing worse than being replaced by AI
12:02
is being replaced by a founder who reads.
12:05
That's the newsletter at AI for founders.co.
12:07
Now back to the show. Yeah, that's a, that's a good
12:10
question. Like, yeah, you're right.
12:11
Meadow was very was pioneering some of this.
12:14
Like the feed was probably the biggest innovation that Facebook
12:17
put out was the news feed. There were a few things, you
12:21
know, we we build a lot of machine learning teams back at
12:24
Instagram. You know, we were doing things
12:26
that for feed ranking and like ranking content.
12:29
I mean, some of the things that we learned was that these are
12:32
super powerful tools, but you have to use them responsibly.
12:36
Like, you know, there's a difference between, for example,
12:39
optimizing for time spent versus time well spent, you know, So,
12:44
you know, there was always the struggle of, you know, being
12:48
careful about what metrics that you want to move up and to the
12:51
right because you want that you, you really want time well spent,
12:54
not just you don't want to waste people's time.
12:57
I remember I also learned that, you know, it's new, new
13:02
technology is always, people don't like change, right?
13:06
I remember when we first put feed ranking into Instagram, you
13:10
know, people were complaining. We went from reverse
13:12
chronological. And The funny thing was we just
13:15
basically took the same posts, like we didn't add new ones or
13:20
put ones in. But we just, we, we, we, we
13:23
ordered them on something called coefficient, which was basically
13:26
how, how close you are to that contact so that you had a spouse
13:31
or a partner, you would see their content first before
13:34
seeing people that you didn't see much.
13:36
And that was it. And of course, you know, it was
13:41
very loud. It was a very loud community.
13:43
They, they, they, they, they were, they were upset by that,
13:46
you know, and this just change is hard.
13:48
So you have to be careful and you have to be really conscious
13:52
about how you integrate this stuff.
13:55
Yeah, yeah, Public opinion, particularly around really kind
13:58
of tip of the spear technology often gets kind of shot down
14:02
right at first. You know, the, the Google Glass
14:04
is a great example of that. You know, or you could look at,
14:08
you know, the the Apple AR goggles.
14:12
It's there's almost like a, a certain level that people are
14:17
willing to accept in one chunk. I suppose when you take meteoric
14:22
leaps, they get a little bit skeptical, they get a little bit
14:24
afraid, or maybe they just get a little bit nostalgic about the
14:27
past. You know, when we're thinking
14:29
about this in terms of where we are with agentic support, you
14:32
know, it seems like we're maybe on kind of a a third wave now,
14:36
you know, maybe a year ago, you know, the the idea and promise
14:40
of the agents put everybody on their heels.
14:41
You know, they're going to take our jobs, they're going to do
14:43
this and that. You know, everyone's diving in
14:45
to see what they could automate, what they can create with N8 N,
14:48
you know, with even Zapier edits like, OK, how much of this can
14:52
we can we automate? And you know, if a lot of
14:56
founders already talked to maybe are similar to me, 70% of the
15:00
stuff they they made broke and they didn't really fix it.
15:02
And so they kind of let it go, you know, and then you have kind
15:05
of this open claw thing that's been bubbling maybe for four or
15:08
five weeks where, you know, people are really excited about
15:11
like, OK, what can we make? And then at what cost?
15:15
Because all that stuff is like, I don't want to say all that
15:17
stuff. A lot of that stuff is malware.
15:19
And then you just see this new kind of, you know, attack vector
15:23
that has just opened up people's entire, you know, computers in
15:28
tech stack entirely. So we're moving into like maybe
15:31
1/3 realm now of, you know, I think owners, CE OS founders are
15:37
starting to think of like, OK, how can I 10X the output of my
15:40
most talented people with agentic support?
15:43
And how can we standardize what that looks like?
15:46
So I'd be curious from your perspective, if you think about
15:49
enterprise companies, you know, what are kind of like the roles,
15:53
the agents, the job titles that will just become like synonymous
15:57
for automation and agentic support?
16:01
Yeah, we think about that here for a second.
16:04
So, you know, I'm, I'm, I'm old enough to remember when the
16:14
first spreadsheet came out and like all the accountants were
16:17
freaking out and they were like, this is the end of accounting.
16:20
And it turns out that it just made them a lot more efficient.
16:24
And there's 10 times more accountants than there were.
16:28
They're just all using spreadsheets.
16:30
So I think that this is not exactly the same, but I do think
16:35
that there's a lot of similarities is that we're going
16:39
to make our teams a lot more effective.
16:41
So there was a lot in that question and where you started
16:47
is that CE OS want to make their employees 10X more productive.
16:51
I think that that that question alone sort of puts a challenge
16:56
out there that needs a lot of a lot of thought itself.
17:00
It's like, what exactly is productivity and how do you
17:04
measure it? Is it a lot of a lot of CEOs, a
17:07
lot of leaders are out there saying like write more code, but
17:11
does that actually mean that you're writing more?
17:14
Is that you're being more productive or are you doing a
17:17
lot of more, a lot more action without traction?
17:20
So understanding sort of 1 like you have to really deeply
17:23
understand what productivity means in your company 1st and it
17:29
might be feature velocity, but it might not.
17:31
It might be removing code might be like what it is.
17:34
You know what the best engineers, the most productive
17:36
engineers delete more code than they write.
17:38
So like you have to get really specific on what what velocity
17:43
means. And once you back out of that,
17:47
you start highlighting work that you can use agents for.
17:52
Now, the hero use case right now is coding.
17:55
So you're seeing a lot of agents out there writing code, but it's
17:59
quickly going to move into other areas, right?
18:01
Like coding's great because it's the most complex workflow that
18:06
exists out there. And you can imagine that sort of
18:10
graduating into other types of work, you know, product design
18:13
all the way up to HR legal. We're already seeing these types
18:15
of companies left and right pop up that are making amazing like
18:20
impact in their areas. But I think that each company is
18:26
going to require like a different analysis of that.
18:30
I I don't know if I answered your question or confused you
18:33
and myself in that answer, but. No, no, absolutely.
18:37
Yeah, it's I think that's probably a two that a lot of
18:41
this is really confusing because if you can, you get the sense
18:45
you're like, OK, there I have all these tools and coding is
18:48
the perfect example because I can just mumble into the chatbot
18:53
and the the analogy would be like, it creates a Shakespearean
18:58
play and you're like, that's exactly what I wanted.
19:01
Oh my gosh. So it's like, where do you get
19:03
that kind of similar leaps and bounds?
19:06
You know, and, and maybe that's just kind of like the fun thing
19:10
about AI right now in the moment that we're in is we don't really
19:12
understand the boundaries of it. So, you know, things pop up, a
19:16
node pops up and everything expands around that where people
19:19
are kind of feeling for the wall of what's possible, you know, so
19:23
if it's, if it's like, oh, we, well, we can automate that
19:26
social media report and search trends and all those things,
19:28
it's like, OK, that seems pretty cool.
19:30
Or, you know, we can build custom whisper flows for
19:34
everybody and, and build all of that, you know, where software
19:38
becomes starting to be hyper customizable.
19:40
And there's, there's some good things about that as well, you
19:44
know, but I don't, I don't know if anybody really has a handle
19:49
on what we're going to look like in about 6 months.
19:52
You know, and Speaking of six months, you guys, like you said,
19:55
are still kind of like just about to come out of stealth.
19:59
So I would love to talk what, what that kind of launch looks
20:02
like, what kind of like go to market looks like and kind of
20:05
start there like what you're planning for the next 6 months?
20:09
Yeah, sure. Like, First off, the way that
20:11
we're looking at this is like last year was like the age of
20:14
coding with AI. Like we saw so much progress,
20:17
but now that's sort of switching.
20:19
And we think 2026 is going to be more agents.
20:23
You're going to see the agents becoming more into focus.
20:26
You're already starting to see that.
20:28
And our belief is that corporate infrastructure is going to
20:33
evolve quite a bit. It's going to look a lot less
20:36
like a set of internal web applications to a lot of agents
20:41
that are helping with orchestrate work flows in your
20:44
company. So our belief is that companies
20:48
will have hundreds and thousands of agents.
20:51
You know, like, I think we're at that phase right now.
20:55
That reminds me of back at Netscape when we're sort of in
20:59
the browser wars and everyone was focusing on building
21:01
browsers, but a lot of the value that created or came from that
21:05
was in the services built on top the web, the web, the Amazons to
21:09
Googles and all of that. And I think we're in the like AI
21:12
equivalent of that right now where there's a lot of focus on
21:16
model capability. But you know, look, I believe
21:20
that a few years from now that's probably going to be smaller and
21:24
smaller deltas. It'll probably become somewhat
21:26
of a commodity. And then you people will be
21:29
using this and you're going to need rich services to be able to
21:31
integrate this into your corporation.
21:34
So that one is our belief. We think that you need a
21:38
platform to be able to productionize these agents to
21:41
control them to and everything. And so that's what we're
21:46
building. 1 The next part of it is this centralization of agents
21:51
in the company. This is what we saw at Meta with
21:54
what we were building for Dev Infra.
21:56
The team that I was part of is we build an internal agentic
21:59
platform where engineers could go and see what agents were
22:03
there, what they were doing. They could, they could fork
22:06
them, branch them, they could reuse them and extend them.
22:09
And it went viral. So, so, so making these things
22:14
turn them from single player to multiplayer is a big change.
22:18
So we, that's something else that we plan on doing and
22:22
ultimately pushing that out to a larger community so that you can
22:25
you, you can go and find public agents.
22:28
And this is not just something that we think will be good for
22:31
building agents. It's a, it's a key part of a
22:33
security strategy, right? Open sourcing agents, because
22:38
this is a new area and we saw with open claw some of the
22:41
security stuff, but like putting it out open source can be a
22:45
really good effective way to also surface security problems.
22:49
Now it doesn't work unless there's a lot of eyes on it, but
22:52
you know, it is it's you know, the thesis is that you can
22:56
invite researchers and people can look at it.
23:00
By not open sourcing it, people are under the assumption that
23:04
people can't see their code, but that's actually not true.
23:07
Like good hackers will reverse engineer and figure out the
23:10
code. So you're basically just hiding
23:13
it from the people that would help make it more secure you.
23:17
Know it's an interesting point. There's actually a lot of
23:19
parallels, aren't there? From the browser wars, the, the
23:22
pre, the, the pre cloud infrastructure days that you
23:26
know, what we're dealing with now that maybe is holding back
23:30
agents to really come pouring in as like speed, you know, just
23:35
small improvements in speed really matter right now.
23:39
It's how can I assign agents to be able to pay for things
23:45
online, you know, payment portals.
23:47
That's kind of like, it reminds me of back then credentials.
23:51
How, how much of my passwords am I going to give my agents and
23:55
can I trust it? So it's like, it's kind of like
23:59
repeating in, in a certain kind of way, like a lot of the, the
24:02
technology that was built for, you know, over the last 20 years
24:06
is like kind of coming back for like a, a Round 2.
24:09
I think that you're right. I mean, these patterns, it's not
24:12
the first, right? Like we've seen these patterns
24:14
historically, like all the way back to the invention of
24:16
electricity to, you know, the industrial revolution and sort
24:22
of making automate, like, you know, Ford with basically the
24:26
factory and, you know, doing all this stuff.
24:29
But the problem is, or I'd say the difference is that these
24:32
cycles are happening much quicker, right?
24:34
So the Internet, that happened pretty fast, that changed things
24:39
really quick. So each one of these seems to
24:41
get more and more compressed and this one seems to be going even
24:44
like way faster than than than than.
24:48
You know when when we started working on the first browsers.
24:52
It's incredible. What, what else is holding?
24:54
Kind of the the, what are they? They're swarms of agents, I
24:57
believe. Is the, the plural what?
25:00
What's holding the swarms back? Well, you look, you know,
25:07
holding them back from like what?
25:09
What? Well, I think that hold on here,
25:13
let me think about this here for a second.
25:14
So one of the things in order to get agents to work well is yes,
25:18
orchestration is, is something really important.
25:21
And so there's a lot of companies out working on
25:24
different orchestration layers. There's some protocols that
25:28
people are standardizing on so that they can talk to each other
25:31
and tools. But I think that these things
25:33
need to evolve a little bit. They're not quite right.
25:36
They're getting there quickly, but I think that the two things
25:39
in order to make swarms of agents work effectively are one,
25:44
you know, getting the protocols right so that they can handle
25:48
all of the cases that you need. MCPS already falling a little
25:52
bit short on some of the things that people need to do.
25:54
Hopefully it'll evolve, providing an orchestration layer
25:59
that works for specific work cases.
26:02
And you know, also like you need human intervention in these
26:06
things. They're not, they're not right
26:08
all the time. So that's another thing.
26:11
Now, I do think we'll end up with swarms of agents and not a
26:14
single agent. Like this to me is pretty,
26:19
pretty obvious, but I could be wrong.
26:22
It reminds me of when mainframes came out and you we thought like
26:25
there'd be a bunch of giant computers people were talking
26:30
about like one agent that will like do everything.
26:33
I think that in order to, I think where we'll end up is a
26:36
bunch of specialized agents, right?
26:38
One, they're easier to debug. Even if you can build 1 super
26:42
agent that does it all, it's a black box.
26:44
So when things go wrong, you it's almost impossible to go
26:47
troubleshoot. And for the next foreseeable
26:51
future, you're going to need to be able to troubleshoot these
26:53
things, especially as people generate, you know, they rely on
26:57
more output without understanding what it what it's
27:00
generating, it becomes even more tricky and more fragile.
27:05
Yeah. Does that lend itself to then
27:07
there is a marketplace, like there's an App Store, there's a
27:10
marketplace for agents where this particular agent persona is
27:14
very good at this one thing and you can license that and you can
27:17
have IP built around kind of the the creation of that agent.
27:21
I think that that's that's going to be a big opportunity, I
27:24
believe, and we'll see what happens, but I believe that will
27:27
happen. You know, back at Meta, we were
27:30
experimenting with code review agents and we sort of came to
27:34
the conclusion that one code review agent, like one company
27:38
can't really solve it. You need probably 500 different
27:41
code review agents. You need the one that's a
27:43
specialist in SoC 2 or GDPR or graphic specialists or hardware
27:48
specialists. So you, you know, you may need
27:51
to, you're going to need like a variety of these specialized
27:54
tools. And I think that applying that
27:58
across all industries, that's probably millions of different
28:02
specialized agents, millions. Incredible.
28:08
I don't want to be IBM that you know, what was it, there's a
28:11
global marketplace for maybe 5 computers back in the 60s.
28:15
So I'm wondering if we if we if we are, are there a little bit
28:19
with some of the some of our thinking today?
28:22
I love the the, the cyclical patterns, you know, I was really
28:25
stretching my brain to be like, how could I bring this back to
28:28
ancient Rome in some way? If there's some senator that's
28:32
like, wow, we need to get there faster, We need to do
28:35
transaction. I need 10 of you, you know, but
28:37
I don't know if there's a a great analogy there, but.
28:40
I'm sure there is anywhere where there's like innovation, like
28:43
the patterns of innovation are the chain are, are the same,
28:46
right? Like one way to think about it
28:48
is like when a core, like a core input, like let's take AI right
28:53
now, what we're seeing is sort of the cost of intelligence
28:56
collapsing, right? It's like the workforce was
29:00
intelligence and that's what we were paying for.
29:03
Every time we've seen a core input sort of go down in value,
29:06
like society reorganizes around it, right?
29:09
So, you know, the Internet, look at the way stores and commerce
29:13
and business work, we reorganized around it.
29:16
This has happened all the way back since.
29:19
You know, you can think probably, you can probably apply
29:21
that pattern to everything. Now going back to the browser
29:25
wars, you know, we're back there and I was at Netscape working on
29:29
the browsers. I was, I was convinced that if
29:31
we build a better browser, we were going to win.
29:33
And that turned out not to be the case.
29:36
Microsoft is like, watch this. It's actually a distribution
29:39
war. And it sort of makes you wonder
29:42
like you know what, what is the contest today?
29:45
Is it a technology contest? Is it a data context?
29:49
Is it a contest, is it CapEx, is it distribution?
29:52
Might be all of those, but. Is it hardware you know?
29:56
It could be hardware, yeah, yeah, it could be all of that.
30:00
And you know, and it's, it's that unknown right now that it's
30:03
just brought all the fun back to tech for me.
30:05
You know, I've been grinding away for for a long time.
30:08
But it's like, boy, there's a, there's a real kind of Wild West
30:12
feeling right now in the air, which I think is really
30:14
exciting. But to your point, it also makes
30:17
people very uncomfortable and very scared.
30:19
You know, there was, I watched a TikTok and somebody was like,
30:24
oh, you're, you're worried about immigration.
30:27
If you're worried about immigration, you should really
30:28
be worried about agents, AI agents, because these are
30:32
millions of of immigrants coming that will work 24 hours a day
30:36
for less than minimum wage. And they're, they're all at
30:40
like, you know, PhD level intelligence, you know, so is
30:44
there a bear case for agents in some case where it's like, you
30:48
know what, here's a line we probably shouldn't cross as far
30:51
as exposure and risk. If we're talking about like
30:54
creating agents, it was in particularly here, you know,
30:58
inside like enterprise companies.
31:01
Yeah, this, this is. I think that there is, but I I
31:05
would probably say it isn't capability, it's more speed.
31:09
Like, you know, how fast we're adopting these things without
31:12
fully understanding them is probably something that I would
31:17
say is more problematic in my mind or can cause more problems
31:20
than the capability. I mean, again, we've seen these
31:24
patterns over history where, yeah, you know, certain works
31:28
going to get automated away and then there's going to be a full
31:31
other class of work that's going to pop up.
31:33
You know, the question is, are we incrementally making more,
31:37
more, more, more jobs that are more meaningful?
31:41
I'm, I'm an optimist and I'm trying to, you know, I can see a
31:45
world where that's the case, where I try not to think of it
31:50
as how can I, how can I remove engineers?
31:54
It's like more, how can I increase the ROI of engineers?
31:58
Because if you can increase the ROI of them, then don't you want
32:01
a lot more of those, you know, and it what, what was it?
32:05
Is it Jeffrey's paradox or something that I've heard quoted
32:08
over and over, which is like basically the more efficient you
32:13
make a resource, the more of those resources you ultimately
32:16
end up needing. I, I don't think this is an
32:19
exception. You know, innovation is, is, is
32:23
a good thing. New, new technology is a good
32:27
thing in a lot of ways. And, and the only way I can
32:31
think of that stopping is if innovation is finite.
32:34
Like, are we gonna hit an end and then there's nothing else to
32:37
invent or build? I struggle to believe that's the
32:40
case. So I think that we'll create
32:42
more jobs. I think, you know, it's hard to
32:44
imagine what those are. I don't think we can even
32:46
imagine what they are yet, but I'm pretty sure it will.
32:51
I'm with you. And there's been a couple of
32:53
conversations I've had with my family where people are pretty
32:55
freaked out and I'm like, no, no, but come here.
32:57
Just look at what you can do. Like just show them a couple
33:00
things and they're like, that's pretty cool.
33:01
I'm like, I know it is, you know, so you know.
33:06
And like, look, you still need humans involved in this stuff.
33:09
It's not, you know, a lot of people I've heard use the
33:11
analogy of a calculator, right? And it's like, remember when the
33:13
calculator and you couldn't use a calculator in school and then
33:16
it was OK, Well, this isn't that, you know, you can pretty
33:20
much rely that if you type the formula in right, the calculator
33:23
is going to give you the same, you know, answer That's correct.
33:27
AI is going to give you a different answer every time.
33:29
So, you know, you need human intelligence to be able to, to
33:33
to be able to check and make sure that you're getting the
33:35
right thing. So you know, humans in the loop
33:38
are still very important with AI 100.
33:41
Percent. Let's talk about the kind of the
33:44
beginnings of Guild AAI. How did you finance the company?
33:48
How did you build kind of the initial team?
33:51
Yeah, so once again, I was at Meta for a while.
33:56
You know, I did 2 tours of duty. The last one was running the
34:00
developer infrastructure teams. It was, it was pretty awesome.
34:03
My background's developer tools. So it's, it's a passion of mine
34:08
when we got in there, like you were saying, that just brought
34:12
back a lot of the magic in the industry for you and I, I felt
34:15
the same. I was like, you know, I learned
34:18
on these Borland developer tools back in the day.
34:21
I also worked there. But like the first time I used a
34:24
visual debugger that they that, that they had, I thought it was
34:26
magic. And I'm like, AI can bring magic
34:29
back to a lot of these tools and we get to go write a science
34:32
fiction story about how, the how the future looks with
34:35
development and go try to make that happen.
34:39
So I went into meta and that's what that's what sort of really
34:44
lured me back in was just that opportunity work with great
34:47
talent in a enclosed like loop. You know, we, we build
34:52
everything, even our own source. So like you can ship pretty
34:55
fast. It's a meaningful community.
34:57
There's 40 some thousand plus developers in there.
35:00
So you're, you're building tools for, for, for, for a big team.
35:05
And you know, I didn't expect, I didn't think that I would go off
35:09
and want to start another company.
35:11
And this is my fifth one I've started in my career.
35:15
But when, when we were in there, we were experimenting and the
35:18
way that we did this I think is interesting.
35:20
It's like a lot of, you'll see a lot of leaders, a lot of CEOs
35:24
out there like sort of mandating AI tool use.
35:26
You hear this a lot like everyone better be using AI
35:30
tools or you're you're going to get kicked out.
35:33
Like I feel that was wrong. We were trying to earn the usage
35:36
of the tools. I think good developer tools are
35:39
their usage is earned. Like you have to, you know, good
35:41
developers will want to use things that they see value in.
35:45
And so we went in and we put a bunch of business challenges out
35:49
to the team and we're like, and when you put crazy science
35:55
fictiony business challenges out to your team, it it forces a
35:59
different type of thinking. So some of those were like, hey,
36:03
could we build self healing fabric?
36:05
Could we have when the site goes down, could we have it identify,
36:09
fix and deploy fixes? Could we eliminate code freeze
36:13
at holiday during holidays? Could we just make it so that
36:17
the risk of the system going down slow?
36:18
So we put like these 10 challenges out and then people
36:22
had no choice but to pick up these new tools and then they
36:25
found a lot of value in them. And one of the things that came
36:29
out of that was this platform where we had this centralized
36:33
system. Now, when this worked and it was
36:37
working so well for so many developers that you know, when
36:40
you saw the dynamics that like were happening, like the
36:43
community, the centralization, you know, the ease of use, the
36:48
ability to have things run separate from your laptop so
36:52
they could go for longer and you don't have to worry about them.
36:56
You know, we sort of had Naha moment.
36:57
It was like, I kind of have to go do that.
37:00
Like this is sort of, you know, one of those rare moments like
37:03
entrepreneurs or there's there's been one other time in my career
37:06
where I'm like, I I have to go do that.
37:09
I love my job at met. I absolutely loved it, but this
37:12
was something I kind of had to go do and the team itself.
37:19
Look, some of these, we've worked together for a long time.
37:22
I have some of our team members have been all the way back to
37:26
Netscape and this is our 5th, 6th company working together.
37:29
So we have those and you know, and we've sort of picked up new
37:33
members along the way. But generally we start started
37:37
with a group that's sort of tight, kind of like a band.
37:40
You already know how to play together, you know what you're
37:42
doing and, and then you got to bring in new, new, new
37:46
perspectives as you scale. And that's what we're doing now.
37:51
Cool. Yeah, I, I, somebody was
37:53
describing it as like you're, you're recruiting for a pirate
37:56
ship, you know, so you got to have a cook, you got to have the
37:59
guy that can play guitar. You got to have the guy's got
38:01
dogs. You got to have the drunken
38:03
fighter you. Yeah, I always thought it was
38:05
like a like, also like cooking like a chef.
38:08
You know, you, you got to have different ingredients, right?
38:11
You can't have the same ingredient.
38:12
If you want to have a really good recipe, you got to know
38:16
what the complimentary ingredients are.
38:18
And it's sort of that as well. Totally.
38:20
Are you comfortable talking about the finance aspect of how
38:23
much you've raised from who, those kind of numbers?
38:26
I can tell you like I can tell you the I we haven't publicly
38:31
announced the numbers yet. I don't know if Christina is
38:35
here, but anyway, you can edit that out.
38:38
But like, we haven't talked about the numbers yet, but we
38:42
were pretty oversubscribed and we, we ended up going with
38:48
Google Ventures as our lead. We brought in some other firms
38:52
that honestly, I'm very close with, including Google Ventures.
38:56
My, my, my last company was Google Ventures, very first
39:00
company they funded. So when they announced Google
39:01
Ventures, they announced my company.
39:04
So I've known them for a long time and I'm, they're just
39:06
incredibly helpful, like unbelievably helpful firm.
39:09
And I, I just jumped it to chance to have them involved
39:12
again. You know, we had, we had NFX,
39:15
which is also, you know, not team members here, but like
39:20
their team members that I worked with at Meta are there.
39:24
And I've known them for a while and they were very close to
39:28
lobby Capital, which I was part of as a venture partner for, for
39:32
a while too. We brought in a crew.
39:34
They have a great portfolio and they're a fantastic firm.
39:38
We brought in a few seed funds like Scribble Ventures, which is
39:42
Kevin and Elizabeth Wheel. Once again, Kevin was my product
39:46
partner at both Instagram and Libra and you know, people that
39:51
I trust and I'm tight with. Incredible, you know, and your
39:55
your your resume is incredible. You're very successful guy.
39:58
I'm just curious about that voice.
40:00
You know, why start at 01 more time?
40:03
Are you just in love with the game or what is it about?
40:06
Maybe this time right now. Yeah, look, it's just it's fun.
40:11
I mean, I like to look, I like to build things that have value.
40:14
I like to, you know, one once you said being in the game is,
40:19
is is fun, It's intoxicating. But like most importantly, I
40:25
think all of us here at the company are very impact driven.
40:28
Like we want to do something and make a difference and build
40:31
something that's good. We like, we like to work
40:34
together. You know, we, we all get up and
40:37
look forward to going to work to see each other and we really
40:41
believe deeply in what we're building.
40:43
And yeah, you know, I, I find out that like, I feel burned out
40:49
sometimes in my career, but it's usually I'm just not inspired.
40:52
So I would always confuse burnout with lack of
40:55
inspiration. And I wasn't feeling burnout,
40:58
but boy, you know, you know it when you see it.
41:01
Like, you know, entrepreneurs, if I had one bit of advice for
41:05
entrepreneurs, it's like, don't start a company to start a
41:08
company. Go wait for that observation
41:10
that you can't not go to. Like that's the one that's like
41:14
going to you're, you know, you're going to at least have
41:16
the best time. You're going to, you're going to
41:18
enjoy going, going to do the most and you're probably be the
41:22
most successful with. Incredible.
41:25
I love it. And with that, that'll bring us
41:27
into the Quantum round for the next 10 minutes.
41:30
Are you ready? I'm ready.
41:34
You know, I, I love speaking with people that have had so
41:37
much success and you've been in, you know, the valley for a long,
41:40
long time. So, you know, there's a whole
41:42
wave of new entrepreneurs that are looking for the their next
41:46
project over the years. What are the attributes of the
41:51
winners that you've been associated with that maybe our
41:54
our listeners could take with them and and maybe relate to?
41:58
I'm sorry, but the winners as far as companies or like or the.
42:02
Yeah, even whether it's founders or individual contributors or
42:05
managers, like is there kind of a thread that you've noticed
42:08
kind of runs through these people?
42:11
You know, look, I, I think that if I were to go back and redo
42:14
things like lessons I learned that I wish I would have learned
42:18
a little, little faster. Was that like when, when you're
42:24
starting out, more important than the company or the idea is
42:27
sort of the network. So like I, I try to go and find
42:31
like 1, where are all the smart people going?
42:34
And how do I get, how do I get in there?
42:36
And, and, and any way I can. I think that also like you have
42:41
to really deeply care about what you're building.
42:44
I think that success comes as a byproduct of pursuing your
42:47
passion instead of instead of like, I want to go here because
42:52
I think it's the next successful thing.
42:54
I don't think that's the right way to look at it.
42:57
Where there's change, there's opportunity.
42:59
Right now is an amazing time to build because there's so much
43:02
change. Yeah.
43:06
But I don't know is that, is that sort of what you're trying
43:09
to get at or? Yeah, absolutely.
43:12
You know, maybe for just a personal question, you know, for
43:16
the folks that are maybe before I, I'm OK selfishly, I'm asking
43:20
for my daughter as a freshman in college.
43:22
Bring it on. Knowing what you know now, what
43:25
would you study if you are entering university as a
43:27
freshman? You know, that's a good
43:30
question. You know, I'll tell you I
43:31
wouldn't study AII think once again, that's going to be a
43:35
commodity. It's like if you went in to
43:36
study electricity or or browser development right now, I think
43:40
that will become a commodity. But going back to what we
43:44
mentioned earlier is like as that input gets cheaper, the
43:48
intelligence and society rebuilds around it.
43:51
Like, you know, economics becomes interesting, like first
43:56
principles, things like math and physics still very valuable.
44:01
So I think, I think studying the things like even human
44:06
psychology, like what are all the changes that are going to
44:10
compound on top of this is like the things that I would try to
44:13
focus on. But that might be the answer my
44:16
father would have given me because he wanted me to make
44:20
sure I was going to be able to pay for food and, and get the
44:24
hell out of the house. And, you know, and all of that
44:27
stuff. There's part of me and, you
44:29
know, look, my background is like, I, I, I didn't really go
44:33
to college. I worked for the university, but
44:35
I didn't go there, you know, pursue what you're passionate
44:39
about, like you know, what really grabs you and just go
44:43
after that build stuff. Just go play in the areas that
44:48
you are most excited in and regardless of what you think the
44:53
financial future of that's going to be and work on building your
44:57
network. The network like is like is
45:00
going to carry you far. Like I lucked into going to
45:04
Netscape. Really it was friends that that
45:07
that sort of recruited me in there and and the network was
45:11
amazing. They're some of my closest
45:14
colleagues and friends now in the future.
45:16
So get her get around a great group of people, even in
45:21
college. Go find the clubs, the the you
45:25
know the the things that are like driving your area of
45:29
interest and get close to those. Yeah, it's really impressive.
45:33
It sounds like, you know, your network has been a really a
45:36
driving force and a great support for you over the years.
45:40
You know, now people start thinking about distribution more
45:43
than actually network, you know, which is to say breadth as
45:46
opposed to depth. You know, I'm going to put 8 Tik
45:49
Toks out every single day until one of them pops and then I
45:52
have, you know, a million followers or something like like
45:55
that. What what's been like a a great
46:01
way that you've nurtured and and continue those relationships?
46:06
Yeah, I'm probably a little old school here.
46:08
I, I mean, I still call people and talk to them on the phone
46:11
and you know, and you know, you have to, you have to purposely
46:17
put energy into it, you know, and one I try to, I try to, I
46:22
love working with entrepreneurs. So I try to reach out and, and
46:27
work with entrepreneurs as much as I can.
46:29
I think that here in the valley, we have the luxury of having
46:33
like a lot of venture firms that are always looking for people
46:37
technical to help them do tech diligence.
46:39
So if you can reach out and say I'm available, I'd love to help
46:43
vet any technology companies or anything like that, Then you
46:47
know, you can get pulled in. And those are amazing tools to
46:50
help build your network. Local clubs, you know, there's,
46:54
there's events that go on like go attend those.
46:57
Go, go, go meet people in person.
47:00
It's very different than online. You know, it's, it's, it's,
47:04
it's, it's, it's the only way I think to build long enduring,
47:08
deep relationships with people. And decades down the line,
47:12
you'll be glad you did it because those, those engineers
47:17
that are sitting beside you and, and, and at a desk in college or
47:21
in, in a company, they're probably going to be, you know,
47:24
the leaders 20 years from now, the company owners.
47:29
And they're going to open so many doors for you, for you, if
47:31
you can, if you can network with them now.
47:33
Yeah, you said go to VC firms and volunteer to do tech
47:38
diligence. Is that literally just like,
47:40
what are you looking at? I'll give you my take.
47:42
Yeah, I think so. Like, yeah, if you, you know,
47:44
one, if you can, if you can get even a second, get an intro in
47:47
and saying you're looking to help do tech diligence or, you
47:50
know, vet companies like as experience and networking like
47:54
that. That's pretty valuable to a
47:56
venture firm. You know, it's kind of like
47:58
looking for a mentor. You know, my one, my close
48:02
friend says, you know, the way to get a mentor isn't to go say,
48:05
well, you mentor me. It's go and offer to do
48:08
something for them and, and and then they become a mentor.
48:13
And I think like that's the way you can build your network.
48:16
Go to like people that have big networks, venture firms, things
48:19
like that and offer to do something for them.
48:22
Fascinating. That's so cool.
48:23
That's a real gem right there. You know, over the last couple
48:27
of weeks with the newsletter, you know, I had somebody who was
48:29
approaching me with a couple of decks.
48:31
They're looking for investors and I was like, you know what,
48:33
I'm not writing checks for that right now, you know, but I'm,
48:36
I'm happy to throw that on the newsletter and they're like, oh,
48:38
would you do that? I'm like, sure.
48:40
So I posted a couple of them and then I, and then, I don't know,
48:44
dozens of people were like, Hey, will you post mine?
48:46
I was like, absolutely. And it, it is just kind of a
48:48
cool way to like help the community and build and see what
48:52
people are working on. I I'm going to, I'm going to
48:54
take that to heart, James. That's a really good idea.
48:57
To remind, I'm just a collection of stolen ideas from everyone
49:00
else. Hey, Miles Davis said that, you
49:03
know, there's no good musician. There's only good thieves.
49:06
Well, I'm I'm I'm right at the top of that list.
49:11
Amazing. Do you have a book that you you
49:15
want people to read so much you'll buy it for them?
49:18
Well, yeah, I do. I would say I have, you know, in
49:23
different dimensions, there's my personal friends and then
49:26
there's work friends. Now, like if the one that I
49:29
would say that I, that stands both that I'm a big fan of this
49:32
is like Ray Dalio's principles, you know, I'm very principle
49:37
driven. I, I like to think about not
49:39
just how to make decisions, but like, what is the reasoning
49:42
behind making that decision? And I think that's a great read
49:47
and it's really made a difference in my work life and
49:50
also my personal life. So I, I love that book at work,
49:55
though, there's a bunch that has inspired me.
49:58
I love stretch. It's basically working under
50:03
resource constraints and like how more innovation comes from
50:06
that than under abundance. I like extreme ownership, you
50:12
know, that's also a good one. That's like, you know, what can
50:14
I do to get a different outcome? Switching your locus of control
50:18
to internal from external. So those those are a few.
50:24
Fantastic. Kind of wrapping up here a
50:27
little bit, do you have a a life motto or a quote that drives you
50:33
or do you have a horrific failure that we can all relish
50:37
in your pain? I I do have lots of horrific
50:40
failures. Like, I mean, we could, this
50:42
could be a whole series of future podcasts for you.
50:46
Yeah. So look, you know, I think that
50:50
I've learned a lot from a lot of mistakes.
50:52
Like, you know, I, I used to say, like the perp there was
50:54
this joke poster I saw that said the purpose of your life could
50:57
be to serve as a warning to other.
50:59
And like, I sometimes think that that might be my life motto.
51:03
But I'd say some, some mistakes that I made that looked like a
51:09
win at the beginning or interesting ones.
51:11
Like my last company, I raised too much money.
51:15
Like I raised so much money without having product market
51:18
fit and without actually I wanted to go start a company I
51:22
didn't have that I have to go start this and I beat my head
51:25
against the wall for 6 1/2 seven years and and you know, I can't
51:30
get that time back. So that I would say is 1.
51:35
But then look, you know, all failures lead you somewhere
51:39
like, you know, the timing of that worked out so that I ended
51:43
up and Instagram, which was an amazing experience and an
51:47
amazing career. And if I wouldn't have had that
51:49
failure, you know, who knows where you'd be?
51:53
So every everywhere leads you somewhere.
51:55
So you have to kind of not regret your failures.
52:01
Yeah, there seems to be the the further you get away from them,
52:04
there seems to be wisdom in all this.
52:06
You know, I, I, I, you know, appreciate failure stories
52:10
because I can really relate to it.
52:11
And failure is obviously very scary.
52:13
But now I realize that I should also be afraid of my wins.
52:17
And I'll attribute that to you, James.
52:21
Sometimes you're win. Oh my gosh, there's a whole new
52:25
category of of terror I can I can experience now.
52:28
I'm here for you. Amazing.
52:31
OK, ladies and gentlemen, go check out Guild dot AI there you
52:37
can join the wait list coming out it.
52:39
It sounds like it's going to be available very, very shortly, so
52:43
definitely hop on the website Guild dot AI.
52:46
James, what's the best place for people to connect with you
52:48
personally online? Probably on LinkedIn, you know,
52:52
I'm there. That's that's that's my main,
52:54
that's my main social network is LinkedIn.
52:58
Amazing and we will link that in the the show notes below.
53:02
So if you're listening to this just go down below and you can
53:05
find Gil dot AI. You can also find James's
53:08
LinkedIn and if you're interested in your raising got a
53:11
pitch deck, you know for the foreseeable future, I'm going to
53:14
keep publishing it to the newsletter and I'm happy to to
53:16
help folks out with that. If you want to shoot it over to
53:19
me, you can reach out to me through AI for founders.co.
53:23
Other than that, James, wonderful to talk to you.
53:25
Thanks for the time and congratulations on all your
53:28
great success. Thanks, Ryan.
53:30
It's been a fun conversation. I appreciate your time.
53:34
This episode is brought to you by MO Burst, the AI first growth
53:38
engine trusted by iconic startups like Calm, Robin Hood,
53:42
Reddit, Uber and Google. Moburst is a digital, full
53:46
service, mobile first marketing agency that helps companies to
53:49
scale and become category leaders.
53:52
Whether you're finding product market fit or ready to
53:55
accelerate growth, Moburst delivers creative, data-driven
53:58
solutions that help with scale and efficiency and impact.
54:03
At Moburst, we're not just guessing what works.
54:05
We're predicting it. As an AI first agency, Moburst
54:09
uses cutting edge tools to analyze user behavior, adapt
54:12
creative in real time, and cover hidden growth opportunities.
54:16
The result? Marketing that's faster,
54:19
smarter, and built to convert from early stage startups to
54:22
global brands. We've helped hundreds of
54:24
companies overcome complex challenges, unlock new
54:28
audiences, and optimize every metric that matters.
54:31
At Moberst, we believe that creative thinking, advanced
54:34
technology, and data are the core drivers of success at
54:38
scale. If you're serious about scaling
54:40
your business with precision and purpose, it's time to talk to
54:43
Moberst. Visit moberst.com and let's
54:46
build something category defining together.