Latest / AI for Founders with Ryan Estes / From China to AI: Ryan Shuken on startup lessons and innovation

Transcript

From China to AI: Ryan Shuken on startup lessons and innovation · AI for Founders with Ryan Estes

0:00
Hey friends. Welcome back to AI for Founders.
0:02
I am Ryan Estes. Today I might be AI Today we're
0:06
talking to one of the most plugged in startup mentors out
0:09
there, Ryan Shukin. Ryan's helped founders raise
0:12
over 100 million, worked with VCs managing 10 billion in
0:15
assets, and spent over a decade in China building startup
0:19
ecosystems. He's been at the frontline of
0:22
tech acceleration and now he's knee deep in vibe coding
0:25
revolution, helping non-technical founders ship AI
0:28
powered products fast. In this episode, we dive into
0:32
the shift from $1,000,000 dev budgets to AI native builds, the
0:36
future of venture capital, how to protect your code from
0:39
hallucinations, and what it means to build with a beginner's
0:42
mind. Also, he brought his puppy.
0:45
Let's go, but for some business. This episode is brought to you
0:48
by Kitcaster, the podcasting booking service that gets people
0:52
like you on top podcasts. Whether you're a founder, a
0:55
thought leader, or just ready to build your personal brand,
0:59
Kitcaster connects you with podcast hosts who are looking
1:02
for your story. We handle all the outreach,
1:04
scheduling, prep, everything. All you do is show up and share
1:07
your message. If you've ever thought I should
1:10
be on more podcasts, go to kitcaster.com and apply today.
1:13
Start showing up where your customers are already listening.
1:16
That's Kitcaster by Mober. Hey, it's great to be here.
1:37
It's great to be on the show. Thanks you.
1:38
Bet man, I want to talk about vibe coding.
1:40
I'm hoping we talk about startups in China, but before we
1:46
do, I want to lead in with kind of some news and just get your
1:48
take. There's a million ways to go
1:50
with this. So Open AI was just given a $200
1:53
million by the department or $200 million endowment for by
1:57
the Department of Defense to do both.
2:00
It sounds like a lot of administrative type AI
2:04
programming, but also defense. So what jumps out at you from
2:10
that story? Well, you know, we have a large
2:17
endowment. I've heard you talk about this
2:19
before about the $500 billion that the government is basically
2:24
promising to fund AI research, technology and projects.
2:27
And that's chunk of it, which is fine.
2:31
You know, I know that other countries are doing this and
2:36
funding with larger endowments even, and it's important you
2:40
just stay relative to stay up to date and to kind of refine those
2:44
systems. But I just think that we go
2:46
really fast. And there's this organization
2:50
out of San Francisco. It's the machine learning, the
2:53
human Machine Learning Foundation and it's been a
2:58
leader in just the rules and and how AI should function to
3:05
protect humanity. The basic rules that would keep
3:08
it from like some people say, taking over and things like
3:11
that. I don't think we're going to get
3:12
there right now, but I just think that it's just one of many
3:18
investment that the government's going to be making into AI for
3:21
their own internal systems and to stay competitive.
3:24
How do you think Sam Altman like got his way into it now?
3:28
GPT is a beautiful product. I use it everyday, it's amazing.
3:32
But I I'm a little bit worried about Sam's intentions for a
3:36
couple of reasons. One, obviously starting out as a
3:38
non profit and then trying to do the old switcheroo, you know,
3:42
starting up a new product for hardware.
3:45
I had Andrew on the show and he was describing GP TS like big
3:49
problem is data. Like they don't have the data.
3:53
And so they're looking for hardware so they can start
3:55
incorporating data, whether that's a wearable or a puck or
3:58
something like that. So a part of my skepticism is
4:02
like, OK, you let Sam Altman and open AI into these government
4:07
databases. Well, I mean, they're starving
4:10
to death for data, you know, is there a potential breach or a
4:14
conflict of interest? You know, and, and secondly,
4:18
like, why isn't Google a better partner?
4:21
Like what do you think It's, it's a loyalty to, to Trump,
4:26
essentially kind of a political thing.
4:28
How Sam Altman is getting like a significant chunk of this.
4:31
I mean, 500 billion to 200 million, It's not significant,
4:35
but something. Yeah, I, I, I mean, I think that
4:41
the, there's a few questions in there just about like what the
4:46
intentions and should we be careful.
4:47
But I think it's something that we learn as we go, no matter
4:51
what kind of company is doing it, no matter who they are.
4:54
It's important that we just kind of have our best policies and
4:58
learn as we move forward, have oversight, have an understanding
5:02
of what we're doing and its impact and how to track and
5:04
trace that data, make people accountable and have a
5:07
transparency for that accountability.
5:09
I mean, that goes just with any kind of project you're doing,
5:12
especially when you have sensitive data like government
5:15
data or HIPAA. And that's, those are the
5:17
questions that I'm most curious about and I'd like to see
5:20
answered. It's just the transparency and
5:21
accountability when those things are touched.
5:23
And I'm sure they have people thinking about that.
5:27
Now, I, I don't know about the political complications of
5:29
whether it was this or that. And, you know, I think that, you
5:34
know, I just don't know. Yeah, it's not, it's not my,
5:37
it's not my court to play in. But I know that there is a need
5:43
for it no matter who it's from and that Google in particular is
5:49
very picky about what they do and how they do it.
5:52
Probably more picky than a data hungry company because they have
5:57
all the data. They've had all of our data, all
5:59
of our searches. Every thought that you've had
6:02
has probably been sent to Google at some point if you've ever
6:06
typed it out and searched for it.
6:07
So you know, they're not, they're not hurting for data.
6:11
They're not hurting for a lot of things.
6:12
And they take a more measured approach.
6:14
But Google has been moving faster than I could ever imagine
6:18
with productization of AI than anything else.
6:21
I mean, they they are snails. The past Google is a snail
6:25
comparison to who they are right now, and it's been really
6:27
impressive. Totally.
6:28
I agree. What an incredible company,
6:30
Google, you know, and open AI disrespect.
6:33
But like, I don't know, Sam, early on, I don't know how to
6:36
just keep my eye on that guy. So Ron, I became aware of your
6:41
work through Denver's Vibe Coding meet up that Danny Newman
6:45
puts together, which is super fun.
6:48
And it's been really encouraging because, you know, for me, I'm a
6:51
non-technical founder and I always had like a little bit of
6:55
like, you know, just nerves around that or embarrassment, I
7:01
suppose. And so vibe coding is like
7:04
giving guys like me the opportunity to build tools very
7:08
quickly. You know, I had kind of a bucket
7:10
list of tools that I wanted to work on and I've just spit them
7:14
out 1 by 1 by 1 by 1. It's really fun.
7:17
And so particularly what the what Howard became aware of you
7:21
is you did this fantastic presentation on kind of vibe
7:25
coding one O 1. Like here are the steps you
7:28
should consider step by step when you're beginning kind of
7:33
doing vibe coding. And also I love the name Vibe
7:37
coding. They couldn't have given it a
7:38
better moniker. Yeah, yeah, yeah, it is great.
7:42
I love it. So maybe we'll we'll just lead
7:45
right here with maybe to summarize kind of that
7:48
presentation. And you've been generous that
7:50
we're going to put the the whole presentation in the show notes
7:52
of this show so people can check that out.
7:55
But to summarize, if someone's getting started with vibe coding
7:59
or actually let's go a little bit more advanced, you know,
8:02
people are are are vibe coding out there, but they're having a
8:05
hard time finishing products. What is something they can do in
8:10
the beginning to help ensure when they're at the finish line
8:13
that they're going to be able to actually ship something cool?
8:17
Yeah, yeah. This is a really good question.
8:20
So I think a lot of people, they run into vibe coding issues like
8:24
if you look online, you'll see post about people saying I have
8:29
a, you know, 1530 thousand line of code and it's all gone
8:34
instantly because I didn't do proper backups or I didn't do
8:37
this or I didn't do that. So you can get really far in
8:39
vibe coding and then all the sudden hallucinations can take
8:42
over and parts of your code are are just gone because there's a
8:47
lot of fundamentals that aren't solid in learning how it works.
8:52
I mean, imagine having multiple people editing a Word document
8:57
and having no way to save it. And, and just like, oh, I wrote
9:01
that really good paragraph for this really good chapter and
9:04
then someone overwrites it with something else and there's no
9:06
save button. That's kind of what coding can
9:08
be right now if you're not using the right setup.
9:11
And, and this is what developers are really good at because they
9:14
have, you know, systems on top of systems, but vibe coding, a
9:18
lot of people jumping into it, they don't understand what those
9:20
systems are, why they're there. So like here, a quick breakdown
9:23
is you first want to understand the terminology.
9:27
The terminology is really important.
9:29
And there's like two or three just big things that if you
9:33
understand as an early adopter of vibe coding that will help
9:36
you get all the way to shipping your product without fail.
9:41
And I'd say the first thing is, is your PRD project
9:46
requirements. Your project requirement
9:49
document is your PRD. And this is something that
9:52
people have encoding whenever you're starting a new feature or
9:55
a new app or some kind of new thing, you're building
9:57
something. These define the goals, the
10:00
requirements and the user stories in which you know
10:03
someone's going to use your app like a user can you know, search
10:07
for the news, get it summarized, you know, listen to that news or
10:10
however it is their user stories that are told using normal human
10:14
language about how users use it. That in addition to how it
10:18
should be designed and what its functions are, are all in the
10:21
project requirement document. And that is like the the
10:25
guideline, the rules of what happens for your your app or
10:30
your technology or or whatever you go got some weather going on
10:36
in the background for me, I. Can't hear it.
10:37
I can't hear it but. Yeah, yeah, You can't hear it.
10:40
The mics. Good.
10:41
I was just like, what was that? All right, you can cut that or
10:44
keep it in back to it. Is that PRD you can build with
10:50
AI. So you first thing you do is
10:52
let's say you let's say you're going to spend, I'm going to
10:54
spend the week building an app. I'm going to spend the weekend
10:57
building something that weekends whatever you're going to commit
11:00
36 hours to it. 12 of those hours or 11:50 should be
11:05
planning period. Not doing anything but planning.
11:09
Getting examples. The more that you can feed AI
11:13
like Grok or Gemini which is Google's 2.5 Flash Pro.
11:18
These models are really good at collaborating with you to create
11:21
this project planning document. This project plan document is
11:26
the like North star of your project.
11:29
And through every single type of feature development that you do,
11:34
you have the AI, whichever one you're using, reference it so it
11:37
doesn't get all whirly out and hallucinate and go somewhere
11:41
else. And that document becomes a
11:44
markdown file that gets saved in the structure of your code.
11:49
And one of the instructions, because you can actually set
11:51
like local instructions and global instructions for your
11:55
coding, is to always reference and remember the project
12:00
requirement document. And so when you say I want to
12:03
fix something, it doesn't redo something else, It doesn't
12:07
rebuild the new idea. It doesn't add scope or things
12:10
when you're building stuff. AI loves to add scope, which is
12:13
new features, new technology, or just kind of searches the
12:16
Internet and just adds stuff. So that will save you from
12:20
losing a lot of night sleep just from having that project
12:25
requirement. That that recommendation was
12:29
amazing and completely transformed my projects.
12:32
When I read that, I was like, oh, of course I now the, the
12:36
instinct is to just sit down and start prompting because it's fun
12:39
and there's a impulsivity of it. But it reminds me of that
12:44
Abraham Lincoln quote where he said, you know, give me 6 hours
12:47
to chop down a tree and I'll spend the 1st 4 sharpening the
12:51
blade. You know, so it's like you have
12:55
the inspiration. Take it to grok, take it to
12:58
Gemini, take it to Claude and start to work out this PRD.
13:02
So the, the first prompt is like, Hey, I have an idea.
13:05
You know, I'm want to work with you on a PRD.
13:08
Here's the, the, the outline of the idea.
13:10
Help walk me through what it's missing and give me ideas.
13:13
Boom. So now you're, you're not
13:15
actually building before, you're not putting the court cart cart
13:18
before the horse. Yep, Yep, absolutely.
13:22
And you know, sometimes if you just wanna like see what happens
13:25
and mess around, it's called one shot coding.
13:28
So there are some apps out there that are one shot apps like same
13:32
dot Dev and lovable can do it as well and a few other ones.
13:38
And you just kinda say, I wanna build a store that sells, you
13:41
know, socks online. And it will just boom, one shot
13:45
put it out. And one shot coding is a little
13:46
bit different. It's like the extreme version of
13:49
vibe coding is one shot vibe coding.
13:52
You're trying to, you're trying to, you're just basically
13:54
throwing darts. You're trying to hit the bull's
13:56
eye, see who can hit it the most.
13:58
Yeah. And that's fun.
14:01
But when you really take an, you want to take an idea to
14:03
production and you know, have it with test and, and actually get
14:07
it out there and make money from it.
14:09
You want to have this PRD, this project requirement document.
14:12
You take that and, and this presentation that's linked here
14:15
has all of the prompts that you should do in order to create
14:20
that and all the suggestions and gives examples and everything.
14:24
The next thing that's really important is saving your your
14:28
code. Now, a lot of people just assume
14:30
that like, it's kind of saved on my computer, right?
14:33
Like it's saved on the cloud or wherever I'm working at, it's
14:36
just saved. But that isn't actually what's
14:39
happening. It's rewriting that file every
14:42
time it you prompt it again and again and again.
14:44
And you can lose what you've worked on.
14:47
You can lose some functionality that you didn't even know you
14:50
lost. Yeah.
14:51
And so the, the core thing that you need to learn is this thing
14:55
called get, which is version tracking Git or GitHub is free
15:00
resource. You can login, you can sign up
15:02
for it and you can prompt your AI, you can ask your AI, any
15:07
kind of vibe code can do this, You know, teach me git, show me
15:11
what I need to know about it. And in simple form, it's just
15:14
saving different versions of your code.
15:17
And you should think of it as like, this is actually hard
15:19
saving when AI is just doing it and randomly saving, you know,
15:24
it's not really doing that. It's just putting more building
15:26
blocks on top of the same file. What you want is you want to
15:30
create a copy and save that and hold that somewhere.
15:32
And when that copy is good and you're happy with it, you merge
15:35
it with the old and you keep on building.
15:38
It's like trees that kind of branch out.
15:41
And that's the way you want to do your version control.
15:44
It's so important, basically version control and saving on on
15:49
encoding is like voting. You want to do it early and
15:52
often and it's it's so important.
15:56
It will save your project. It will absolutely save your
15:58
project and understanding what that is because you think it
16:01
saves, it does not. You have to learn how to do git,
16:04
and you can prompt your AI to save it for you, to manage all
16:08
of that for you. But it doesn't come
16:09
out-of-the-box that way. And so you want to be able to
16:13
prompt it. You want to give it instructions
16:15
on how to do that. That's fantastic.
16:17
Yeah. Because, you know, to your
16:19
point, it's rewriting at every prompt.
16:21
And if it changes something that you're not working on, you don't
16:24
know which version to roll back to, then you're just like, Oh my
16:28
gosh, how deep was this? You know, and you're and then
16:32
your enthusiasm and momentum is destroyed for the day.
16:37
Yep, these are the two biggest thing.
16:38
If you can just it's yeah, everything's destroyed.
16:41
You're like, and you might have spent, I don't know, 400, five,
16:44
$100 on coding this something. And something changed in a part
16:49
of your code that you don't even know.
16:51
And, and, and by the time you are like, great, I'm done.
16:55
Your code's all different, your functionality is all different.
16:57
This is really what happens and I think, you know, places, the
17:02
Ides, which are the coding software, we call them Ides,
17:07
they are getting better at managing this for you, like
17:10
windsurf or cursor, but at its core, like you are really
17:14
responsible for your own save states.
17:17
You know, you're really responsible for going through
17:20
and, and managing that and understanding how to save it.
17:23
It's easy, just every single time you do something, save,
17:26
update, save, you want to do a new feature, you create a new
17:29
save point. You know, it's kind of like
17:31
playing a video game where there's multiple choices and you
17:35
want to experience all the choices.
17:36
You're like, I want to go down this path, so I'm going to save
17:39
before and then after I can check out this path and that
17:42
path. And they're just like save spots
17:44
and video games, but for your code.
17:46
And that's the really important thing like project planning,
17:49
your PRD, understanding how to save like that will save you so
17:53
much money and. Time amazing the saving is what
17:55
I'm doing next. That's that's my next yeah
17:59
that's my next goal. Especially autosave.
18:01
That's fantastic with the get very cool.
18:05
Now when I was doing research on you you've had an incredible
18:08
career You we have a kinship because I work with founders.
18:13
I love founders and you have kind of a similar appreciation.
18:18
I think that I do as I'm reading quotes about you and just the
18:21
the trajectory of your career. You know, I, I, this one stood
18:25
out for me and said I have one simple mission to help founders
18:27
build their company, which is great.
18:30
And, and you work largely in technology as well.
18:33
So when we're talking about vibe coding and we're talking about
18:37
basically seeing the cost of software production run to zero,
18:42
I mean, that'll be the trend eventually.
18:45
How does the, what does that look like economically?
18:48
Obviously tech has been such a big sector, you know, kind of
18:51
the the tip of the need or tip of the spear for, for American
18:55
economy. What does that mean to the SAS
18:58
model? What does that mean to software
19:00
designers and engineers? Where do you see this thing
19:03
going? Yeah, I, I'm so excited.
19:07
I love this question. This is this.
19:08
So this one gets me riled up because it means so much for all
19:13
that. I mean, we are living in a time
19:15
where so much is getting revolutionized about what we
19:18
used to do and how we used to help founders.
19:21
And now what we can do now, what you can do to build your
19:24
company, to make your idea real. You know, the entire way in
19:27
which we help founders is now fundamentally changing and, and
19:31
we have to stay up on top of it. And so I'd say when I was, so
19:36
my, my past, my history, I was, I did a lot of startups venture
19:42
backed and I found my way to China and I spent a long time in
19:46
China over a decade working with startups and helping build the
19:50
startup ecosystem and investing companies in Shanghai.
19:53
And I did that with a big VC that was one of the big early
19:56
stage ones. And we ran accelerators and I
20:01
was one of the accelerator directors.
20:03
And so I have hands on experience with helping startups
20:06
build. Tech cross-border into China,
20:10
out of China, all over the world, all that kind of stuff,
20:13
running startup ecosystems and trying to build these classes,
20:18
these accelerators, these modules, this track for startups
20:22
to be successful and everything has changed.
20:25
And when I say everything, I would average that when I would
20:28
give, let's say we did $1,000,000 in investment just to
20:31
keep the numbers simple in a startup to grow, I would expect
20:35
that they would spend around $800,000 in development.
20:40
That's front end, back end, middleware, security, just
20:44
figuring out project planning, paying the developers, all of
20:48
that stuff. That's 80% of what we would
20:51
give. We'd expect to go to just
20:53
development today, right now with AI, we don't have to.
20:58
That's not where it goes. 80% doesn't go to development
21:01
anymore. It doesn't you know you it's a
21:04
fraction of that now 10,000 or 100,000 at Max is going to
21:09
development and all of the rest of that money is going to actual
21:13
building your team, doing marketing, which is then also
21:16
probably run by AI and then you know your sales and everything
21:20
else and leveraging it. I mean, it allows us to help
21:24
companies and help them scale in ways that we just didn't know we
21:30
could do before. So as someone who helps startups
21:33
and kind of guides founders on what the best practices are,
21:36
it's all changed. The price tag of what the entry,
21:40
you know, the ticket price to enter is so much lower now than
21:43
ever to get something out there. And I genuinely believe that
21:47
we're going to see a billion dollar valuation company, a
21:50
Unicorn come out that is entirely vibe coded by people
21:54
who don't know how to code very soon.
21:56
And that means the whole game has changed.
21:59
I mean, the SAS models change because we're not going to have
22:03
to build the middle Ware for the SAS model.
22:06
We're going to have the database layer.
22:07
That database layer is going to immediately access an MCP
22:11
server, which is the model context protocol, which is then
22:13
going to interact with an AI, which is going to give it access
22:16
to interact with the customer. We don't need even a front end
22:20
anymore. I mean, you can have the AI just
22:23
give a phone call, then update somebody about how you know
22:26
their account is and what their status is and everything like it
22:29
allows us to rethink almost everything and how we help and
22:34
work with startups and what kind of advice we give.
22:37
And, and learning what that is right now is kind of my mission.
22:40
That's why I built that Vibe coding tutorial.
22:44
That's why I'm continuing to do it, is because I think
22:47
everything that we know about how to start up and how to help
22:51
these founders has changed and we have to start AI startups
22:55
ourselves to figure it out. We need to jump in the pool with
22:58
them. We can't.
22:59
We can't sit up here anymore and say, hey, move this chess piece
23:03
there and do that. We have to figure it out with
23:05
them. So does that imply then get for
23:08
venture capital that they're not going to be writing $1,000,000
23:11
checks, rather they'll be writing $200,000 checks?
23:15
Or are they just like encouraging, like, OK, hey, you
23:18
know, you got we're flipping it on its head.
23:20
You got 80% to spend on marketing now and sales, how are
23:23
they kind of approaching the, the, the climate?
23:26
Yeah. Yeah, Yeah.
23:28
So that's a very interesting question because with my
23:32
background in VCs and building accelerators, you know VCs,
23:36
large ones, they're dependent on their LP's, they're limited
23:39
partners. These are large families.
23:41
These are billion dollar funds that expect a certain amount of
23:46
investment and an amount of return within a search period of
23:50
time. Your internal rate of returns,
23:52
really important. All these things.
23:54
What they can't do, what a new VC could do or a new VC model
23:59
could do is that they have a APPM, which is a document that
24:04
says like this is how we are going to invest, this is our
24:07
plan for success, this is how we're going to invest, this is
24:10
how we're going to do it over this fund, next fund, next fund.
24:12
It's kind of hard to change that plan as you're going it and get
24:16
buy in from everyone and not all the firms are going to be fast
24:19
enough and be able to do that. So you, I don't think you're
24:21
going to see a lot of companies. I think you're going to
24:24
basically see people fitting in $1,000,000 or whatever they're
24:29
if they're used to starting $3,000,000 checks or 1 million
24:32
or 2 million, they're going to make enough stuff to fit in the
24:36
$3,000,000 box to get that investment.
24:39
But that's not needed. Early stage VC for a long time
24:43
has been broken. We followed a model that's been
24:46
grown out of Silicon Valley in the 1980s and we haven't really
24:49
varied far from it. You know, we have safe notes
24:52
which are great for startups, but we've stayed on those for a
24:55
really long time. And the way in which we think
24:58
about the money in our LP's is pretty old fashioned.
25:02
And this is the first opportunity, I think that people
25:04
have an opportunity of building VCs in a new way because smaller
25:09
tracks are going to do just as much as the big ones.
25:12
Before you're like, I can't do much without a $20 million fund,
25:15
$60 million fund. Now we can do a lot with a few
25:20
$100,000 fund. We can help a lot of people.
25:23
We can do a lot of different stuff.
25:25
And it's going to change the way that people who are starting VCs
25:27
or starting new funds within VCs think about what they're doing.
25:31
More opportunity for people to just try something and follow
25:35
their dreams. I I'm so excited about it.
25:37
Yeah, I mean it once you to your point, once you've got the two
25:41
person team that vibe coded their whole billion dollar
25:44
company, then, you know, it becomes like like a headhunter,
25:50
like now you're just looking for talent.
25:51
And of course VCs always look for talent, talent.
25:54
But it's it's no longer the, you know, Harvard grad coming out
25:58
with like you perfect SAT scores.
26:01
You know, you're looking for like dreamers or in people with
26:05
the grit to like really push something through.
26:07
It's totally fascinating. It's and this promise to of like
26:12
what you were talking about, of reimagining the interface with
26:15
technology. I'm really interested in this.
26:18
And it was kind of highlighted because I moved recently and you
26:23
know, they're like, Oh, welcome to your smart house.
26:26
I'm like, no, no, no, no, no, no, I want a dumb house.
26:28
Like they're like, no, look at everything has its own app and
26:32
it all has updates and it all text message.
26:34
You, you I'm like, dude, no way. Give me a key.
26:37
I can hit the garage door opener.
26:39
I don't want a smart house. And I, I love that term
26:42
inshittification because we've really kind of gotten into the
26:47
insidification of a lot of different technologies,
26:49
especially these kind of like half baked apps that people are
26:53
putting into their IoT devices and like everything's kind of
26:57
connected kind of, but then it breaks.
26:59
So you don't have the password, then you 2 factor authenticate
27:02
the download. It is, it is a just a minefield
27:06
of microaggressions constantly. So when you're thinking about
27:10
novel ways, you know that that AI can bring new frictionless
27:16
or, or reduce the friction of some of these tensions with,
27:20
with the banalities of technology.
27:23
What kind of comes to mind? Are are we talking about like
27:27
glasses? Are we talking about hardware?
27:30
Are we talking about neural implants?
27:32
Or we talk what? What does it look like to you?
27:38
When we're talking about kind of the future and what that looks
27:41
like and what's most useful, give me, give me one second.
27:45
I think we can. I need to check my dog.
27:48
This is my puppy. Yeah, what kind of dog is he?
27:51
My. Puppy and he needs a medic.
27:53
He's a little Boston, little Boston Terrier and he's 9 weeks
27:59
and I just need to give him a little detention when he woken
28:02
up right now, yeah. He's like, make sure he's you're
28:04
talking and talking and you're not talking to me.
28:07
How dare you? I know you OK, Brody?
28:10
You gonna hang out? Do you know any film?
28:12
OK. I think he's OK.
28:15
That's the only inner Richter. I promise.
28:17
I might need to hold him. OK, so back to the topic.
28:22
So we're talking about like, what AI means for human
28:26
interfaces, what it means for your life and where technology
28:28
can kind of go. But previously, before I was in
28:32
DC, one of the start-ups that I was with, we were doing
28:35
augmented reality vision. This is before it was sexy.
28:40
This was before it was like in everything.
28:43
And we basically created a product on Google Glass that
28:46
would allow you to translate any sign anywhere in the world, and
28:50
it would just kind of transform it.
28:52
And we gave quite a few talks about like how we think getting
28:58
technology closer to the human can transform and help us.
29:01
And this is kind of like the same question.
29:03
How is AI going to do this? How is AI going to get closer to
29:06
us and help us? And the way that I think that I
29:09
would like personally, a little bit of my own idea in this is a
29:14
eyes and technology's goal is not to directly become us or
29:18
interface with us entirely. It's to almost become invisible
29:22
in our day-to-day life and what we do.
29:24
It's kind of like the good leader aspect or the great
29:28
worker, you know, like if the great, the great leader, you
29:31
know, almost when, when they do something, it's like they almost
29:34
did nothing at all. And, and that's kind of how I'd
29:37
like to make it feel. And AI for the first time with
29:40
how smart it is, legitimately smart.
29:43
Like a lot like like VR and augmented reality until now was
29:48
pretty much dumb. You know, we had an Apple Watch
29:50
kind of buzz us to remind us something.
29:52
We have things giving us notifications and messaging us
29:57
and causing a stress and anxiety.
29:59
For the first time ever, we have something that can reach us on
30:03
our level. What if I want like a daily
30:06
meditation where after that I get like a rundown of what I'm
30:09
doing and what I need to do for the day and kind of like
30:12
researches everything. And I can do that.
30:15
Yeah. And I can do it in a soothing
30:16
voice in a smart way. Like for the first time, we can
30:19
kind of have a softer touch of technology because we don't we
30:22
can actually communicate directly with what we want and
30:25
how we want before that community, like technology's way
30:27
to communicate was, hey, look at me read this thing, alarm bell,
30:32
look at me alert notification. And that was very, I think bad
30:37
in design. There's a reason why Steve Jobs
30:39
hated that in terms of design, in terms of like notifications
30:43
and stuff like that, because it does stop the human experience.
30:47
And I think there are more ways now in using AI for us to
30:53
integrate, I don't know, assistance planners, just things
30:56
in our life that we need, that we want, that will feel so
30:59
seamless, like we're just talking to someone and like
31:02
nothing's a big deal. Yeah, I'm, I'm.
31:04
That's kind of why. Me too.
31:08
This episode is brought to you by MO Burst, the AI first growth
31:12
engine trusted by iconic startups like Calm, Robin Hood,
31:16
Reddit, Uber and Google. Moburst is a digital, full
31:20
service, mobile first marketing agency that helps companies to
31:23
scale and become category leaders.
31:26
Whether you're finding product market fit or ready to
31:28
accelerate growth, Moburst delivers creative, data-driven
31:32
solutions that help with scale and efficiency and impact.
31:36
At Moburst, we're not just guessing what works.
31:39
We're predicting it. As an AI first agency, Moburst
31:42
uses cutting edge tools to analyze user behavior, adapt
31:46
creative in real time, and of our hidden growth opportunities.
31:50
The result? Marketing that's faster,
31:52
smarter, and built to convert from early stage startups to
31:56
global brands. We've helped hundreds of
31:58
companies overcome complex challenges, unlock new
32:01
audiences, and optimize every metric that matters.
32:05
At Moeburst, we believe that creative thinking, advanced
32:08
technology, and data are the core drivers of success at
32:11
scale. If you're serious about scaling
32:13
your business with precision and purpose, it's time to talk to
32:17
Moeburst. Visit moeburst.com and let's
32:20
build something category defining together.
32:24
I'm totally fascinated with this and if anybody is it is working
32:29
in this field, reach out to me because this is what I'm
32:32
creating a frictionless interface with technology.
32:36
And I think that there might be ways that you can create kind
32:39
of, Oh my gosh, he's adorable. I think there's ways that.
32:44
He's going to be. In the show, all right, there's
32:47
ways that you can do it that might create new sensory
32:51
experiences for humans. There are so many different ways
32:55
that humans can move or modulate that could be tied into some
33:01
kind of AI trigger. You know, a very rudimentary one
33:05
would be my stupid smart house with you.
33:08
You can go to the app and dim the lights, you know, and it's
33:10
like did I would rather walk over to the the big button and
33:15
dim it then pull my phone out, you know?
33:19
Yeah, it's that it's that extra ass that's constantly being
33:22
asked of you for technology forever that finally can be
33:26
bridged in a truly like just get it done methodology, you know,
33:31
and I can just do. It and you know, so, So what I'm
33:34
thinking though, is like what I want to do is like have like the
33:37
full Jedi. I just want to extend my hand
33:39
and move my my fingers and that dims the light.
33:42
But then can I shorten the gap to just make it my hands and I
33:45
can move my hands a certain way that dims the light?
33:47
Or could I just make it for like my eyelid and I slowly, you
33:52
know, all these days? Just your attention, just your
33:55
attention, just like where your attention is.
33:57
If you're like looking over at the kitchen, the AI could be
34:00
like, oh, he's looking over the kitchen, Just dim the lights up
34:02
a little bit over there. Maybe you need to see something.
34:04
OK, now, now he's looking over there.
34:06
Like it can interface with what our attention and what we need
34:09
in a way that we've just had to kind of like ping and alert you
34:14
and Ding and all this. Like, you know, I, I, I just
34:18
don't like them. They're, they're these alerts
34:19
that create anxiety and stress. And for the first time, we can
34:22
build in a different way. We can actually get rid of the
34:25
human interface. I mean, that's what we, you
34:27
know, how impossible would it be for me to ask you to download
34:30
another app? Yeah, you know, like it's a big
34:33
ass to download another app on your phone.
34:35
It's a big ass. You know, it is like, OK, then.
34:39
Yeah. Then do this.
34:40
And that's how that's how people feel.
34:42
Those are big ass. And so now we can start thinking
34:45
about, well, what if we just have my AI, which I'll download
34:50
that, I'll have that on my phone, I'll have that in my
34:52
house and that's connected to these things called MCP servers,
34:56
model context protocol. These are like the servers that
35:00
give us tools. So if you are a founder and
35:04
you're building something and you're building technology and
35:06
you want to integrate it with someone's life, you can make
35:08
your technology kind of like a micro service in these MCP
35:11
servers that any AI can access and start doing things with and
35:15
make calls to and start thinking about.
35:17
So it's like, you know, that we're we're starting to build
35:20
the process and the foundations of the future of what that looks
35:23
like. And if you, you know, want to
35:26
build right now is the best time ever.
35:28
Even right now, still the bolt hackathon is going on.
35:31
You can get bolt for free. You can vibe code for free.
35:34
You can get all the stuff. I mean, this is this is the,
35:37
this is the moment. It's so exciting.
35:39
It's so. Fun.
35:41
It truly is. I want to go back just a little
35:44
bit to your time in China and especially working in software,
35:48
especially being there for 10 years, especially to where we
35:51
are now with China. Now, one thing China's maybe a
35:54
little bit notorious for is not respecting IP.
35:58
What, what was it like when you were there and working with
36:01
software teams and, and just kind of the, the, the energy and
36:07
feeling around copyrights and intellectual property and
36:10
copying and ripping off like what's, what were the kind of
36:13
guidelines there? So that's a really good
36:17
question. And I think the answer is a
36:21
little sometimes shocks people because if we think about like,
36:26
oh, they're not respecting IP internationally, well, they're
36:30
also not, you know, they are also running really fast and
36:33
trying to copy whoever is successful as fast as possible.
36:38
And it's not about not respecting IP, it's about
36:41
servicing the market as fast as you possibly can.
36:45
So internally within China, the competition is so intense.
36:48
I mean, the America has no idea what actual competition looks
36:53
like compared to Chinese competition.
36:55
Because I've seen it for so many of the iterations of revolutions
36:59
of tech and things happening that it, it is incredible.
37:02
And the, the thing that wins all the time over protecting your IP
37:07
is innovation. And China had learned that over
37:11
and over and over again. And they went from copycat China
37:14
to innovate in China very quickly.
37:17
And they're already at innovative.
37:18
If you're still in the mindset that China is, you know, copycat
37:21
this and that, then you're probably I'd say 8 to 10 years
37:26
like back and you're thinking they've invented and kind of
37:29
innovative so much on stuff that it's that it's really something
37:34
that we started bringing to America.
37:37
Some of the innovations that they've made in restaurants and
37:39
in software and how they manage that and delivery and E bikes
37:43
and everything, you name it. A lot of the things that we have
37:45
are innovations that we've copied.
37:47
But another thing is that China, in terms of IP, if you go there
37:52
and you want to build something like a schematic or technology
37:54
or you're doing some engineering, it's not that
37:56
they're not respecting your IP. It is that they are very clearly
37:59
stating and having you sign a waiver that says you recognize
38:03
that your IP is not protected by our loss.
38:07
They are letting you know and it is very clear that you do
38:10
business here. This is what the cost is.
38:13
You want cheap labor, you want this, you want good factories,
38:16
all that kind of stuff. It's very transparent.
38:19
And so when people are like, I signed this thing that said I
38:21
don't have, I signed away my rights of IP and then they
38:26
complain about it. You're like, that's what it that
38:28
that's their rules and it's their homecourt advantage.
38:31
And it's something that's helped them grow a lot.
38:34
And so, you know, as long as it's a transparent and it's, you
38:37
know, right up front, you know, it's one of the risk of doing
38:41
business there and it's what's helped them go.
38:43
But now we're learning from China, Now we're learning new
38:46
ways of building technology. I mean, it's gonna be, it's
38:50
better for global competition anyways to have more creatives
38:53
and more things coming out. So, you know, I don't think, I
38:57
think we're past kind of that era of fear.
39:00
They have their own IP laws internally.
39:03
But again, you really need to be working in China for a long time
39:07
to to, to navigate that stuff, yeah.
39:09
They're very impressive. You know, the robots that build
39:12
the bridges over there? How?
39:14
Fast you. Know their cars, their electric
39:18
cars are totally insane the the drone swarms.
39:22
Yeah, they're great. I mean.
39:24
Yeah, I remember the the drone swarm shows they would have over
39:29
their cities. Yeah, it's really impressive.
39:31
Very impressive, very impressive.
39:32
It's it's, you know, as the, the, the drum of war is always
39:39
being beat. It's usually off in the distance
39:42
especially, but you can kind of hear how they're starting to
39:45
manipulate narratives and things like that to, to create more of
39:49
an adversarial relationship with China.
39:51
But, you know, I, I think that to your point, what I love about
39:55
technology is it democratizes, you know, talent, you know, so
39:59
there is so much we can learn from them.
40:03
I look forward to doing it. All right, I, I know that puppy
40:06
is, is itching for dad. Are you ready for the Lightning
40:09
Round? Yeah, Yeah, yeah, I'm ready.
40:12
All right. I'm in coach.
40:14
If you're entering college today as a freshman, what would you
40:17
study considering the the future that's coming for us?
40:23
I would study not actually AI, like I wouldn't study building
40:29
AI systems and building the architecture of them.
40:31
Instead, I would study business and how to AI prompt.
40:37
So how to build businesses and structure.
40:39
I'd understand finance, I'd understand like, you know, how
40:44
to market and how to grow things because your opportunity to
40:46
build a business and do things right now and scale, that is
40:51
like no other time in history and to even get that information
40:55
and learn how to do it. So, you know, I'd be like
40:58
gearing up on how to build, what to build, how to sell, and then
41:02
learn how to prompt studying, how to prompt studying these
41:05
systems and how to use them, not necessarily how to do the
41:09
advanced statistical math needed to build your own system because
41:13
they're kind of already out-of-the-box.
41:15
Is there a prompt engineering course or degree now there it
41:20
is. There are, there are courses,
41:23
they're online, I'm not sure if any.
41:24
I mean probably a university is going to send something out
41:26
there, but there are online courses.
41:28
I have a few of the certificates for prompt engineering and there
41:32
are lots of things you can look on.
41:34
A lot of them are free. Some of the best ones are just
41:37
people talking on YouTube for like 3 hours and showing you how
41:41
to do things and what to do. But you know, we're at a point
41:45
where you can spend, you can spend a month doing nothing but
41:52
studying AI and prompting and come out of.
41:55
It a. Like hirable prompt engineer,
42:00
there are very few things in this world you can spend a month
42:03
and come out competent about anything.
42:06
You just need to like, you know, be able to speak and and convey
42:12
your ideas into words pretty well.
42:15
But there are so many guides and practices, it's pretty
42:17
incredible. Another soft spot I need to work
42:21
on Prompt studying. Amazing.
42:24
Is there ACEO or entrepreneur you really admire?
42:30
CEO of Entrepreneur. I really admire.
42:33
It's a good question. I think, you know, I'm not I'm
42:40
not a fan of I don't I don't try to root for anyone big out
42:45
there. I'm kind of for the underdogs,
42:47
my favorite CE OS and founders because I'm a startup guy and
42:52
I've just been there are people who are literally just starting
42:55
out. I know you're a big zero to 1
42:58
fan. I'm a huge zero to 1.
43:00
You know, if you're a college student, like my favorite person
43:03
to work with is, is the is the founder is like the college
43:06
student or the the gal who just quit her job to start this idea.
43:11
And those are my favorite people.
43:13
And I will say it every time. That's the people that I want to
43:15
help. That's why I do it.
43:17
I want to help them not run into the same problems that, you
43:20
know, we've all ran into 1000 times and try to follow their
43:23
dream. And so it's just like kind of, I
43:25
know it's a lame response saying it's all of this group, but it's
43:28
the zero to 1. It's the founders just trying to
43:29
make it for a sale, trying to figure out how to build their
43:32
first thing. That's those are my favorite,
43:34
favorite people to work with. Favorite founder is the passion,
43:37
the drive you know they're showing up they have no idea
43:40
what they're getting into there's a little bit of fear
43:42
there's a lot of excitement like that's my favorite and and just
43:45
seeing and working with them and meeting them that's you know it
43:48
just. I love it.
43:48
I, I totally agree. You know, as you get a little
43:51
bit older, for me, the, the, the goal is to not become bitter and
43:57
not become resentful. And I think people grow bitter
44:00
and resentful for good reason, you know, But so like beginner
44:05
mind that that founder energy, I always want to like keep a
44:09
wellspring of that. It's like Zen talks about that
44:12
beginner's mind is where you always want to be in return to.
44:14
Yeah, Beginner's mind, yeah. Cool, if you could live in any
44:19
other era besides now or the future, what would it be?
44:25
If I could live in any era besides now or the future, what
44:28
would it be? Probably probably be when they
44:33
were exploring and mapping out the world, kind of like when the
44:36
world was fully undiscovered, when they were sending ships to
44:40
like map out new trading routes, when they thought it was flat.
44:44
You know, I'd definitely be on a on a ship trying to map out the.
44:47
World totally I I agree man. I, I, I'm obsessed with like
44:53
Mesoamerica, so like the Aztecs and the Mayans and the Olmecs,
44:57
and like those cultures, which we're just learning now how
45:01
sophisticated they really were, way more so than than Europe at
45:05
the time. And to be some of the first
45:07
people to discover those cultures would be so cool.
45:13
Yeah, absolutely. Plus.
45:14
There's kind of a pirate energy, you know, you're on a ship with
45:17
your mates, you're drinking a gallon of ale a day.
45:20
You probably got some songs. Yeah, I know.
45:25
I I grew up pirating software so I I have it in me.
45:28
You got a you got a peg leg and an eye patch in the closet, I
45:31
bet. Yeah.
45:33
Nice. Ryan, this has been awesome.
45:36
Where can people connect with you online?
45:37
What's the best place? You can connect with me.
45:41
Probably one of the best things is I respond to LinkedIn.
45:44
Just look for Ryan Shuken, RYANSHUKEN on LinkedIn.
45:49
You'll find me. Just connect, say you heard me
45:52
from the show or whatever and I'll connect with you.
45:55
You can also find me at ryanshuken.com and you can
45:58
e-mail me there if you want, but probably just do LinkedIn.
46:02
And yeah, if you're a new founder, if you are thinking
46:04
about doing something, if you want to learn more, if you want
46:06
some of this material, I appreciate you and you know it.
46:09
Takes a lot of bravery not only to launch multiple startups, but
46:13
also to invest in startups. And really, you know, for the
46:16
the founder community at large, we need, we need good guys out
46:19
there like you that are, you know, driving value 1st and then
46:23
has a helping hand to offer people coming up.
46:26
So thanks for that. Thank you.
46:30
I appreciate that. Same to you.
46:31
And you know, we got to get, we got to share the the lion's
46:34
share of information here. We got to get it out somehow.
46:36
And every time that we do that, it's, it's a, it's an act of,
46:40
you know, generosity. So thanks, thanks for putting
46:43
all the stuff out here and continuing with it represent.
46:45
Thanks buddy. Cheers.
46:50
Hey founders, let me ask you this.
46:53
Are you building with AI or getting left behind by it?
46:57
If you're serious about scaling smarter, faster and bolder, you
47:01
need to check out AI for founders.co.
47:05
It's your go to resource for cutting through the noise and
47:08
actually using AI to grow your start up.
47:11
Each week, AI for Founders drops a sharp, no fluff newsletter,
47:16
dives deep with expert guests on the podcast, and hosts monthly
47:20
workshops where you can get your hands dirty with the real tools
47:23
and strategies. Whether you're precede or Series
47:27
B and beyond, launching something new or just staying
47:30
sharp, AI for Founders gives you the edge you need to win in 2025
47:35
and beyond. Hit AI for founders.co to
47:38
subscribe to the newsletter, tune into the podcast, and grab
47:42
your spot in the next workshop, and it's all free.
47:45
Build smarter, move faster, think bolder with AI for
47:49
Founders. Now back to the show.