Latest / AI for Founders with Ryan Estes / Most AI prototypes die before launch

Transcript

Most AI prototypes die before launch · AI for Founders with Ryan Estes

0:00
If you have an app idea and it's still living in notes, you're
0:02
not early. You're stuck.
0:04
Because the new founder trap is not a lack of code, it's fake
0:08
momentum. Vibe coding gets you a pretty
0:10
good planet side fast, and then the last 30% hits like a brick
0:14
wall. Architecture, requirements,
0:16
integration, security, the stuff that makes you want to delete
0:19
the repo and move to the cabin. Isan Bill Codalio to be the
0:23
between a CTO and a technical lead layer.
0:26
You prompt in plain English. Codalio interviews you back,
0:29
pulls the media out of your head and turns it into real
0:31
requirements and technical scope.
0:33
Then it generates a full app front end, back end around 80 to
0:37
90% of the build. And this is a use case founders
0:40
are sleeping on. Use it to build a real internal
0:43
tool in a weekend. Think OPS dashboard, customer
0:46
intake, workflow, distribution centre, digitization, or a data
0:50
ingestion pipeline from drones or.
0:52
Robots for real. Not a demo, a viable product
0:56
that can actually scale. So think about.
0:58
It are you delaying your product because the idea needs more time
1:01
or because you're afraid of finding out it's real?
1:04
Welcome to AI for Founders. I'm Ryan Estes.
1:06
This interview is for founders who want speed without building
1:09
a fragile spaghetti monster. You will learn how to avoid the
1:13
70% wall, how to scope like a real CTO, and how to ship
1:17
something that does not collapse the moment users show up.
1:20
And if you want an unfair founder advantage, grab the
1:23
newsletter at AI for founders.co.
1:26
It's. Bite size playbooks, spicy
1:28
tools, and the kind of clarity that makes shipping feel
1:30
inevitable. Join and your future.
1:32
You will stop hate scrolling and start building.
1:35
That's the newsletter at AI for founders.co and if this has
1:38
happened to you, please leave a review on the podcast like
1:41
you're tipping your future self. Founders, you do not have a lead
1:45
problem. You have a trust problem.
1:47
Ads are expensive. Cold outreach is a knife fight.
1:50
Social posts disappear in 6 minutes.
1:52
Kick Caster fixes that by booking you as the guest on the
1:55
exact podcast your customers already trust you Show up, tell
1:59
the story, teach the lesson, drop the proof.
2:02
Then the audience goes, oh, this person actually knows what
2:04
they're doing. We have helped founders drive 7
2:06
figures in revenue from podcast appearances.
2:09
We have seen founders raise serious capital after the right
2:12
interview hit the right investor at the right time because.
2:15
Podcasts are the last. Place on the Internet where
2:17
people still listen like humans. So here's how it works.
2:20
We dial in your positioning, your talking points, and your
2:23
founder narrative. We pitch you to shows that match
2:25
your ICP. We handle booking logistics and
2:28
prep. So you walk in sounding like the
2:30
main character, not a random LinkedIn post with legs.
2:33
If you're building something real and you're ready to be
2:35
known for it, Kick Caster is the shortcut.
2:38
That's kick caster.com by MO burst.
2:55
Hassan, welcome to the show. Thank you so much for joining
2:58
me. Hello, thanks for having me.
3:02
Absolutely. My pleasure.
3:03
I'm really excited to talk about your tool Codalio.
3:06
So give us the the 30,000 foot view of your, your tool and the
3:11
company. Absolutely.
3:13
So Codalio is an AI generative, AI augmented platform allows you
3:21
to do both product and software development a lot faster, more
3:25
quicker and it simply allows you to build enterprise and the
3:30
scalable full applications 80% faster and more cost effective.
3:38
Incredible. Would it be safe to say it's a
3:41
vibe coding platform? Well, it is not.
3:46
I think there's a big differentiation between US and
3:49
what vibconing in general means. I think from our perspective,
3:54
Vibconing allows you to prototype things a lot more
3:57
quicker and but it doesn't give you the sort of, I guess the
4:06
whole experience of building full scalable enterprise grade
4:09
or you know, let's say good quality application.
4:12
Vibe coding is great in just you know to start with a simple
4:16
prompt and give you a good UI very quickly, but it's not
4:21
designed to be the tool to be used to build scalable.
4:26
Applications 100%. You know, vibe coding is
4:31
probably the the favorite conversation of this of this
4:35
podcast, largely because you have all these new people that
4:38
have enthusiasm and ideas and they're bringing their ideas to
4:41
bolt, to replict to lovable. You know, if they're a little
4:44
bit more advanced, they're getting into clawed, they're
4:46
doing cursor, you know, so I kind of how should we think
4:49
about Cudalio in this kind of ecosystem?
4:53
Is is this kind of a graduating platform?
4:55
Like you kind of tinkered around with a couple things, but you're
4:58
like, you know what, it's actually time to bring this to
5:01
kind of a retail ready type. Platform is.
5:04
Is that a good place to kind of describe where you kind of fit
5:07
in the market? That is that is in a good
5:11
description of where we land in this market in this spectrum of
5:14
different services. But if I could just give you
5:17
maybe our one version of how we categorize the whole industry
5:20
and and the other tools in the AI encoding space, maybe that
5:23
gives you better perspective into into what Caudalib is.
5:28
So we have categorized these AI coding or AI encoding tools into
5:34
three main big buckets. One of them we call them the
5:39
generalist code assistant that the copilot cursor and the other
5:43
tools that they plug into your IDE, the developers can prompt
5:47
them and then did they generate some code for you?
5:51
They're adding more layers into it, they're kind of adding more
5:54
planning features into it and so on and so on.
5:57
But at the end of the day, those tools are designed to help the
6:01
developers to build code or write code a lot faster.
6:07
There are journalists that are agnostic to the tech sack that
6:10
he used. And so they're designed to be,
6:13
you mean as part of the toolbox for every type of developer out
6:19
there. Then it comes a second category,
6:21
the prototyping tools, the lovable V-0 and Bolton and the
6:24
other ones that they are great in kind of get you from prompt
6:30
to a good UI that you can start imagining and kind of bringing
6:37
some of your ideas to life. You can click through it, you
6:40
can kind of get a sense of the workflow.
6:43
And again, for the most part, they do a good job building
6:48
front end for for the applications and it is designed
6:54
to be prompted by almost anybody that and you can just simply
6:59
prompt them in plain English. So you whether you are just a
7:03
hobbyist or a developer or a product manager that you want to
7:08
quickly prototype something that actually do a fantastic job.
7:11
And the third category are the ones that we're actually calling
7:15
them in our own words, saying that they are life cycle agents
7:20
or life cycle agentic flaws like Devon Optimal or the card code
7:25
that again, they're essentially designed to augment engineering
7:29
teams. And they basically accelerate
7:33
the core existing workflow of writing code, committing to the,
7:36
you know, get repose, you know, getting the code reviews, maybe
7:40
debug and so on and so on. And so they kind of have a have
7:43
an overarching focus on the entire software development life
7:47
cycle, again designed to be used by software developers and
7:52
engineers. So where we come is that we are
7:57
kind of looking at it as an evolutionary step into all of
8:01
this that we do have a combination of all.
8:04
Again, we're not targeting all of them.
8:05
Obviously, that's the right thing to do for any startup to
8:08
just be able to just go after everybody in that sense.
8:12
But we're essentially saying that we wanted to have a closed
8:15
loop, tightly integrated loop to build fully functional
8:20
applications. And the 1st and the primary ICP
8:26
for us are the founders. Anybody that wants to build a
8:31
new application and that they can simply just prompt in
8:36
English, get all of the details sorted out from the scope
8:41
requirements, even understanding the ICP and the type of users
8:45
that are they're going to bring on board.
8:47
And then it kind of generates from there all the technical
8:50
details, the technical skull. And then we go through a process
8:54
to generate both the front and the back end, but obviously not
8:58
entirety. And there's always a last smile
9:00
that we're fully aware of that we want to make sure that the
9:03
developers can actually bet on. But we actually get you to a
9:07
good chunk of the code on the front, on the back end, so about
9:11
80 to 90%. So it is this incense replacing
9:18
CTO for example. I think that is a very good way
9:22
of you know characterizing it. We think it's somewhere between
9:26
CT OS and the tech leads because not every CTO is is hands on,
9:31
you know, dealing with the day-to-day development by cycle.
9:35
You bet, you bet. But at least to to spin up
9:38
ideas, get them all the way there and then using.
9:42
So then what do you recommend for kind of completing the last
9:47
mile, really tying a bow on it and getting it ready to ship to
9:50
market at? The end of the day, for any
9:55
developer to be able to really execute on, on the plan that
10:00
that is aligned with the product, they need to have all
10:03
this specific tasks and requirements defined and, and
10:07
provided to them. So once you generate all of this
10:11
scope and requirements early on, and then we generate again, some
10:15
good amount of front and on the back.
10:17
And then if we put on the gate repo and then we we do have
10:20
these, let's call them Halpers that are fed to the LLMS,
10:26
vanilla island Lambs or any coding, you know, agent.
10:31
Then from that point on, the developers not only have the
10:34
clearview of an understanding of what exactly they should be
10:37
working on, but they also have the proper guidance and
10:41
assistance with the specialized agents that they can simply not
10:46
only understand the existing code, but also just start
10:51
building that last mile with the proper guidance and assistance.
10:55
And and you know, the way that we do it is that we properly
10:59
manage the context in that, in that sense, which means that
11:03
they will have all of the, everything that they need from,
11:10
you know, in from the traditional sense from the tech
11:13
leads or CT OS. See, I love it.
11:16
So the last mile is going to be a mix of agentic support and
11:21
then there's kind of some human checks on the on the backside as
11:23
well. That is correct.
11:25
Cool. And then what about
11:27
collaborating? So like if I, I'm going to go to
11:29
Codalio, I've got an idea. I just lob it out there.
11:33
I'm going to have a, a, an exchange and, and basically
11:36
Codalio is going to interview me, kind of refine the idea,
11:39
make sure it knows exactly what to expect, maybe probably
11:43
predict some of the things that that I don't know.
11:47
My biggest problem right now coming from a design standpoint
11:50
and a Rev op standpoint and going into some of these coding
11:53
apps is I don't know what I don't know.
11:56
So I really do rely on Codalio, for example, to like be like,
12:00
Hey, we know what he wants to do.
12:01
He's going to need all these things as well.
12:03
So through that exchange, it's going to basically put together
12:07
all the requirements for the application and then it's going
12:10
to build. Now if I'm in this process and I
12:13
got a buddy and he's in Toronto with you, you know, and I'm here
12:17
in Colorado, I'm like, hey, I want to bring in my buddy.
12:20
How do we collaborate with the the kind of design of these
12:23
MVPS? What's up everybody?
12:26
Let me hit you with some math that will blow your mind.
12:28
A wealth manager using Warm Start dot AI grew their assets
12:32
under management by 8.5 million in three months.
12:36
Their monthly fee? 500 bucks.
12:38
That's a 1700% return on investment in three months.
12:42
And I get it. Another AI tool, another monthly
12:45
subscription. But here's what makes Warm Start
12:47
different. It's not trying to find you new
12:49
people to bug. It's helping you reconnect with
12:52
people who already know you. As a founder, your biggest asset
12:55
isn't your product or your website.
12:57
It's your relationships. But most of us are terrible at
13:00
nurturing them systematically. Warm Start connects your e-mail,
13:04
identifies warm contacts, and helps you craft personalized
13:08
outreach. That doesn't sound like a robot
13:09
wrote it. The best part?
13:11
You review every message before it goes out.
13:13
No automation without permission.
13:16
If you're ready to stop leaving money in your inbox, try Warm
13:19
Start dot AI. Use code AI FF for 50 bucks off
13:24
your first month because the ROI on your existing relationships
13:28
is probably better than any investment you'll make this
13:31
year. That's warmstar dot AI and use
13:33
code. AIFF this episode is brought to
13:40
you by MO Burst, the AI first growth engine trusted by iconic
13:44
startups like Calm, Robin Hood, Reddit, Uber and Google.
13:49
Moburst is a digital full service, mobile first marketing
13:52
agency that helps companies to scale and become category
13:55
leaders. Whether you're finding product
13:57
market fit or ready to accelerate growth, Moburst
14:01
delivers creative, data-driven solutions that help with scale
14:05
and efficiency and impact. At Moburst, we're not just
14:08
guessing what works. We're predicting it.
14:10
As an AI first agency, Moburst uses cutting edge tools to
14:14
analyze user behavior, adapt creative in real time, and cover
14:19
hidden growth opportunities. The result?
14:21
Marketing that's faster, smarter, and built to convert
14:25
from early stage startups to global brands.
14:27
We've helped hundreds of companies overcome complex
14:30
challenges, unlock new audiences, and optimize every
14:34
metric that matters. At MO Burst, we believe that
14:37
creative thinking, advanced technology, and data are the
14:40
core drivers of success at scale.
14:43
If you're serious about scaling your business with precision and
14:45
purpose, it's time to talk to MO Burst.
14:48
Visit moburst.com and Let's build something category
14:51
defining together. So I have to say, you know, you
14:58
had a very great way of, you know, providing an overview.
15:04
Appreciate that. That makes my job a lot easier
15:06
to explain. But you know, to your point,
15:09
when you prompt in English and you're guided throughout the
15:11
process and you figure out what you know the, the other things
15:14
that you don't know, that you don't know.
15:17
And once all of those, let's call them artifacts or the
15:22
detailed requirements are, are generated, you at that point are
15:27
able to invite, you know, anybody from your team or the
15:31
external teams to the flow. And they're going to be able to
15:34
see all of those details in the flow.
15:38
And then obviously they can continue editing or prompting to
15:41
make sure that they get it, they get it to the finish line from
15:45
that sense. We also have the versioning
15:47
feature that allows the collaborators to essentially
15:50
first maybe freeze the scope on a version and then start, maybe
15:55
continue prompting to, to iterate and, and define the next
16:00
iteration and the version on that.
16:01
So that makes it a lot easier to for the collaborators to to not
16:05
only collaborate with each other and edit an update, but also
16:10
through the versioning, create a clear understanding of how they
16:14
want to iterate again on the future iterations, again in a
16:20
more collaborative way. What's a man?
16:23
What a cool tool. Now one of the my favorite
16:26
things is getting inspiration. So here in Denver we have a
16:29
great, we call it AI clubhouse. It's ID 345 here in Denver and
16:35
basically on a bi weekly basis we get together and demo demo
16:38
products and I love it because there it's almost like the it's
16:43
a with AI, it's a blank palette. You know, sometimes you need
16:48
somebody to give you kind of paint by number.
16:50
So what are some great products that you've seen developed with
16:55
Codalio and what are some tools you'd like to see developed with
16:59
Codalio of? Course.
17:03
Well, again, one of the things that we've always focused on is
17:07
to make sure that we enable our customers to build things that
17:13
are even from the day first are scalable because the term MVP
17:19
and again we call ourselves as part of it 4th category MVP
17:23
builders, you're essentially doing a minimal Bible product.
17:29
It's not a prototype that you want to still experiment the
17:32
ideas with. We want to make sure what we
17:34
generate is a Bible product. So in our examples, we've seen
17:39
companies that wanted to not only ingest some massive amount
17:45
of data from let's say robotics or a drone, pull them into the
17:50
application, pull them into the software properly, put them into
17:53
the proper data pipelines, pre process and post process using
17:58
AI or whatever may be the case. And then although their
18:02
requirements started to go almost in this case that
18:07
required those requirements started to grow very quickly
18:10
after the first iteration, they were able to still just continue
18:15
on the same trajectory or the same pathways to basically add
18:19
more bells and whistles, update the architecture and then add
18:23
more capabilities to it. So the best example we've seen
18:29
distribution centers wanted to digitize the whole process and
18:32
wanted to experiments maybe for just some time and then start
18:37
scanning from that point on understanding where they can
18:40
start building more efficiency into the process.
18:43
Robotic companies, drone companies that wanted to ingest
18:47
massive amount of data, react on it properly and so on and so on.
18:53
What I would like to see hopefully in the future is that
18:56
we can hopefully build ways that we can automatically use MCPS
19:04
other than just using AP is to do the manual integration.
19:09
And how we do this effectively, again, a lot of it is yet to be
19:12
seen, yet to be figured out. But I would like to see that
19:16
people come in from, you know, their own ideas in ways that we
19:20
can simply connect to many different types of solutions
19:24
that makes them better without having to really think about
19:27
building them and without having to thinking to think about how
19:31
to integrate with those. Because that's the way of the
19:34
future. The way I see it, you don't want
19:36
it. You want to move a lot faster in
19:38
this time of day. And you just want to connect to
19:43
a slew of different possibilities that other people
19:47
are working on. And you can bring something into
19:50
existence that, although it may sound simple, but does obviously
19:57
provide some impactful outcome. So cool.
20:02
Now these tools as a category seem to be about 12 months old.
20:06
So talk about the inception of Codalio.
20:10
Did you finance around or did you raise a round of financing
20:14
to to basically to launch? But the original idea and the
20:19
original code that we started with actually came about five
20:23
years ago. So myself and my ex Co founders
20:28
at a company called New Binary, which was a fractional CTO
20:32
service company where we were providing some very senior CTOS
20:38
to many earliest stage. And later such companies that
20:40
were dealing with, you know, the whole exercise of building
20:44
complicated solutions. And we figured out a way that
20:49
very much early odd that the, the biggest challenge for us to
20:52
get started was to simply bring Dwarf's ideas or extract those
20:56
ideas from the minds of the founders.
20:59
And so in that process we started, you know, coming with
21:03
our own ways to interview our customers better to understand
21:08
the business objective first. Again, the job of the CTO back
21:11
in the day for us was to make sure that we not only understand
21:14
the business objects, but also we can properly tie them to the
21:17
technical requirements. So we started with that mindset.
21:21
We build our own processes internally.
21:23
We we had an interesting run. We grew so fast and so quickly.
21:28
That was the need that we didn't think that that existed to that,
21:32
to that amount to that significance.
21:36
And so right there, we obviously realized that we need to enable
21:40
our customers to move a lot faster.
21:43
And one of the things that we paid a lot of it, you know, put
21:46
a lot of emphasis on and paid attention to was the fact that
21:50
most of the new software had good chunk of them.
21:56
Or let's say 70% of every new software was simply the repeat
22:02
of the same features that existed before.
22:04
So for example, if you've been in your login system, you're
22:06
doing absolutely no innovation. A lot of the deaf shops,
22:12
obviously we don't consider, we didn't consider ourselves as a,
22:14
as a software development company.
22:16
But a lot of the software development companies, deaf
22:18
shops, outsourcing companies where a lot of the senior
22:20
engineers, they created their own boiler plates.
22:24
That basically meant that they had all of these modules or
22:28
standard modules that they, they were weaving together to produce
22:32
a new product. We decided early on, 5-6 years
22:35
ago to create an abstraction layer on top of, on top of the
22:39
Ruby on Rails, you know, textile that simply generated all of
22:45
those commodity components automatically instead of trying
22:49
to bring in those components together and try to stitch them
22:54
together. So we build that.
22:56
We started thinking about how we can simply, you know, elicit or
23:01
extract those requirements and then push it through our own,
23:06
you know, algorithms, you know, processes, code and then down to
23:10
our framework. And so at some point around 2
23:14
years ago, we realized we're on to something very interesting
23:18
because we could actually reduce the time and the cost of the
23:21
development significantly for our customers.
23:24
So we then basically said let's just spin this off.
23:28
We spawned off as a separate company and about just a few
23:33
months ago we raised the pre seat capital and then we started
23:37
wrapping up and we are in in our second alpha and soon we are
23:43
going to be going out with our major beta release in about
23:46
couple months and hopefully you know expect good amount of
23:51
adoption by the market. Yeah, incredible and
23:54
congratulations. How much is you Have you raised
23:57
so far? In Canadian dollars, that's
24:01
about $900,000, to be exact, $860,000.
24:07
And from angels, from, you know, some local government entities,
24:12
they actually do equity investment.
24:14
And we do have some more that we expect to land very soon, and
24:21
hopefully that gives us enough, you know, horsepower to push
24:25
this faster, quicker and better. I love it.
24:28
Wait till this audience gets a hold of this.
24:30
You know, there's nothing more fun than experimenting with
24:33
tools like this, you know, especially if at the level that
24:36
you're designing for, it sounds like a lot of applications
24:39
internally that will be like very useful for a lot of
24:42
different kind of folks. But then there's the tinkerers,
24:45
the dabblers, you know, I love that this is also an offshoot of
24:50
your your other company because that's like kind of exactly what
24:54
you could use this for, right? Like you have, you have a, your,
24:57
your primary company that you're working on, but you're looking
25:00
for new projects. Cudalia is a place where you can
25:02
go experiment with it. Now I want to talk a little bit
25:05
further differentiation because I think you like spelled out
25:08
kind of your position in the market very clearly.
25:12
There's going to be folks like me that just stumble in talk
25:14
about vibe coding this and that, and that's not exactly it.
25:17
What are some common misconceptions that people might
25:20
have and maybe how do you clear those up?
25:25
Well, you know, it's very much evident today that a lot of
25:29
people had this misconception that by coding basically means
25:34
that you get a deployable ready to go ready to in the production
25:38
applications. And it's not the case.
25:40
Yeah. And so yes, by coding allows you
25:45
to really bring your vision into something that you can play
25:49
around with very quickly. But then if it's not, if you're
25:55
not thinking about your requirements early on and you're
25:58
not really figuring out all the details that come right after
26:04
the first prototype is produced, you are starting to to basically
26:11
seeing some sort of a wall that you cannot bypass that easy.
26:16
There are so many articles coming out that talk about the
26:19
70% problem, which basically means that you get to the 70% of
26:25
your, you know, conceptualized code.
26:30
But then the last 30% is so difficult to build that you have
26:37
to rethink about the entire UI, the workflow, the architecture,
26:42
the business logic, the integrations and so on and so on
26:45
and so on. So I think we have to give the
26:48
byte coding code, byte coding tools, the credit to really just
26:55
allow people to envision things and bring them into something
26:58
that they can click and play around with.
27:01
That actually solves one of the biggest, I guess, aspects of the
27:06
scoping project. But then you fall short right
27:11
after very quickly after just having the first word to type
27:14
ready for you. Yeah, that it's so true.
27:17
And you know, I was talking to my buddy Danny in a previous
27:21
episode this week and we were talking about like if you're
27:24
using vibe coding broadly, like the importance of momentum, you
27:27
know, and I, I think particularly there's a low
27:30
barrier of entry for a lot of people with ideas to start to
27:32
materialize those ideas. And you're exactly right.
27:36
You know, if it's 70% or 80% or maybe you can get 99% of the the
27:41
idea out in 10 minutes, but that last 1% is going to make you
27:45
wish you were dead, you know. So that's really where you're
27:50
going to need a a different scale of product to kind of help
27:54
you through as far as like providing the extra human in the
27:58
loop is there. How involved is your team
28:01
actually like in deploying projects, getting it past that
28:05
last mile, and then also kind of talk about your team, how many
28:08
of the folks are on the team? We are about 6 people full time,
28:14
but I do have, you know, some great advisors and in total with
28:18
our interns we're about 13 people today.
28:21
Amazing. That's fantastic.
28:24
You're HQ in Toronto. Sorry to interrupt.
28:27
In, in Toronto in I mean mostly in, in Northern Ontario, in the
28:32
city called Sodbury where we received some good amount of
28:36
support from the economy development and that and the
28:39
local government there. And but yes, we do have, we have
28:47
to be closer to some extent to, to, to what happens in the
28:51
Toronto core because of all the great things that are happening
28:55
within the start of ecosystem in Cuba as the accelerators and the
28:59
whole scene of the earliest stage, you know BCS and the
29:03
founder. So we have to maintain some
29:06
proximity to that. But I wanted to go back, if you
29:12
don't mind, to the whole notion of the OR the the whole concept
29:17
of the misconception about the other tools.
29:20
One of the things that relevant to the to to that conversation
29:25
is that obviously a lot of people do the prototype.
29:28
They unplug from the lover one the both of the platform, then
29:31
they go into their own preferred coding to like cursor or the
29:36
other ones and they seem to continue the prompting and
29:39
hoping to get you know some, you know, strong or high quality
29:45
code out of those platforms for the on top of what they already
29:48
produced in those in those tools.
29:51
The challenge is if you are prompting those tools, you need
29:57
to have a good understanding of the context of the architecture
30:01
of the code, the requirements, the business requirements and so
30:04
on and so on. So yes, it is true that you can
30:09
indeed as a founder, as a non-technical or a non developer
30:14
founder to prompt those other tools with plain English and
30:17
produce some code. But if you do not have a good
30:21
understanding of the software architecture about the type of,
30:25
you know, or the best practices in coding or even the existing
30:32
even thin layer of the code that you generated in your
30:34
prototyping tool. If you do not have a good
30:37
understanding of that, prompting those tools is not going to get
30:40
you that much, you know, far ahead.
30:45
So that's another thing that, you know, another misconception
30:48
in the space that we see out there.
30:52
Just wanted to put it out there. Yeah, I appreciate that.
30:54
You know, I'm interested about Toronto too.
30:57
You know, I, I, I know about Denver's kind of AI community
31:01
and startup communities, Very strong Austin, of course, San
31:06
Francisco, New York's got a great community learning a lot
31:09
about Zurich and their kind of AI community.
31:12
Talk to me a little bit about the the startups and companies
31:14
coming out of Toronto right now. Well, first of all, I I think by
31:22
numbers Toronto and Saturn on 2 ecosystems.
31:26
So Toronto Waterloo Corridor is considered to be the third
31:31
largest startup ecosystem in the world.
31:34
Again, I haven't checked the numbers, but apparently at San
31:38
Francisco, obviously still #1 then you have the Boston
31:42
ecosystem. And then based on the numbers
31:45
that I've been tracking for the past few years, Toronto Waterloo
31:51
Corridor has become the number three.
31:55
It is said that at any given time you're looking at about 60
32:00
to 80,000 active start-ups, tech start-ups and they are mostly
32:06
incubated at the university linked incubators.
32:11
So the ones that are at the campus or at around the campus
32:15
that do have their close proximity and collaboration with
32:20
engineering schools with let's call it today with any type of,
32:26
you know, with any of the departments at the university
32:28
settings. And then there are some really
32:32
big success stories that led to creation of some industry
32:37
specific ecosystems in the in the in the city that are
32:41
notable. One of them is Fintech because
32:45
of the Ethereum and the fact that we in Canada have some of
32:48
the largest banks in the world. The second one is related to the
32:54
medical device and life science, life sciences.
32:57
And that's mostly attributed to the fact that we do have some
33:00
great universities, University of Toronto in the city that that
33:04
that actually produces a lot of great IP and intellectual
33:07
property. And I brought those.
33:10
And then AI has always been big. Toronto is University of Toronto
33:18
is considered to be the birth place of AI in general with Jeff
33:24
Hinton just starting his research for the most part in
33:29
Toronto. And then he was a contributor to
33:33
the tensor flow at Google. That is the underlying
33:38
technology to the Google Translate that later became I
33:42
guess the the foundation to a lot of the LMS.
33:47
You know one thing, and I mentioned that because I, I read
33:49
on LinkedIn that you'd been a little bit, I would say vocal
33:55
about Canada's innovation commercialization problem.
33:59
What did you mean by? That so Canada as a country, as
34:08
a nation is injecting over $53 billion into R&D at the
34:14
university settings, you know, colleges, universities,
34:17
hospitals and also funding some business, you know, innovation
34:24
through many different programs. I think 40 if I'm not mistaken,
34:27
about $40 billion of that goes into the research institutions
34:30
directly. So university colleges and
34:33
universities. And we do have a very small
34:36
share of commercialization of those generated, you know, IP
34:40
coming out of those programs. The number is very low and when
34:49
you compare it to some of the best universities in United
34:52
States, MIT, Stanford, I think the collective of, you know,
34:58
what happens in Canada today, maybe close to one of the
35:03
universities, one of the top universities in United States.
35:06
So that is a big blow to to to the investments of of our tax
35:12
dollars as as Canadian citizens. And so I think if if in this new
35:18
world that you need to shorten the cycle of innovation, it is
35:23
important to not only investing in deeper tech or high tech type
35:29
of place. And because then obviously
35:35
accelerate that process, not commercialization process by
35:38
many by any means that you can. And the reason that maybe the
35:43
biggest reason of it is that just like the white is
35:51
democratizing the whole software development, there isn't much of
35:56
an intellectual property in many of the solutions, traditional
36:01
type of solutions out there. So if you could, if you want, if
36:05
you decided to build a new, let's say CRM, you could
36:08
actually do something very decent in about a day or maybe
36:12
maximum in about a week. So when that happens, your
36:18
barrier, you are, you know, your ads, your competitive edge
36:26
cannot be on the building a software tool as it used to be.
36:31
It should be around, you know, some deeper intellectual
36:35
property play. And that becomes extremely
36:38
important for a nation like Canada that that that brings in
36:42
good amount of talents from around the world, produces great
36:46
talents and great quality talents locally and and you
36:52
know, invest massively into R&D, but they're not properly.
36:56
Capitalizing on that sounds fair enough.
37:01
All right, we're we're ready to go into the quantum round.
37:05
Are you ready for this? All right, You think the Raptors
37:09
are going to beat the Pistons this year?
37:13
I honestly don't know. You a basketball guy?
37:18
I am more of a soccer guy because I was not born here and
37:23
but you know, I grew up, I learned to play soccer and I'm
37:28
more, more of a soccer guy. Where were you born?
37:31
I. Was born in Iran.
37:35
In where? In Iran?
37:38
Iran or Iran? Yeah.
37:41
Yeah, they they've got some pretty good soccer teams down
37:43
there. Yes.
37:45
Yeah. It's what World Cup is next
37:48
year, right? Yes.
37:50
Cool. Well, and I don't blame you.
37:52
You know, you get you end up in Toronto, they give you the
37:54
Raptors. You can't it's not the right
37:56
team. Yeah.
37:56
It's like you want to be a basketball fan, you come to
37:58
Denver, I'll I'll take you to Nuggets game.
38:01
We'll do it properly. Amazing.
38:06
OK, let's go back to I business ideas again, because it sounds
38:12
like you're a business idea guy and you, you're kind of at the
38:15
fulcrum of other people's business ideas.
38:18
The audience of this program loves business ideas.
38:21
Now it the, the, I think they're the kind of the benchmark for a
38:26
successful idea is a quick ramp up to that $1,000,000 run rate.
38:31
So if you had an idea for another business or maybe you
38:34
have an idea that could quickly ramp to $1,000,000 annual run
38:38
rate, what's that business idea you put?
38:44
Me on the spot. Think of something.
38:48
Fast. Well, I think one of the
38:57
smartest things I've seen in the retail space is the way that
39:01
Costco operates. If you pay close attention, I
39:07
think they have developed maybe one of the some of the smartest
39:11
algorithms to pick what people need at the right time and they
39:15
price it very accurately. And so they have created this
39:21
massive customer confidence that when you go to Costco and you
39:26
buy something, you actually get the best deal.
39:31
In this time of day. We do have access to a lot of
39:35
public data on, you know, custom reviews.
39:40
We do have access to a lot of the specifications of, let's
39:45
say, laptops. Even as an engineer, even as a
39:49
computer scientist, as an engineer, it's not easy for me
39:54
to do a quick comparative analysis on laptop one versus
39:58
laptop 2 to say, you know how much of A performance boost I
40:02
get from this CPU versus the other CPU.
40:08
So if anybody could be inspired by what Costco has done with all
40:14
of the publicly available data, they could actually run an
40:18
engine. That you could potentially run
40:22
on your browser as an extension that when you're looking at the
40:26
product you you you gain good amount of confidence that you
40:32
can actually you are actually getting a good deal or not.
40:38
And so that's a good deal. I love that.
40:44
Yeah. I mean, you know, Costco and,
40:47
and a big part of AI, the future of AI where I'm most interested
40:51
is in its predictive capabilities.
40:53
And this kind of ties right into that, you know, based on your
40:56
patterns, based on, you know, your spending behavior or even
41:02
also just based on kind of the cultural milieu.
41:05
Because I mean, Costco, I, I went in there like 6 weeks ago
41:10
with my wife because I'm like, you know what, I need 20 fours
41:12
frozen burritos, you know, so of course I'm going to Costco.
41:16
But on the way to get the frozen burritos, I was like, oh man,
41:19
they got a, they got a dry sauna for like 2000 bucks.
41:23
And so I'm like, sweetheart, I'm just going to, I'm pulling the
41:24
trigger. We could put this in the, in the
41:26
garage. She's like, no, no, you can't,
41:28
you can't buy a sauna right now. But I was like, I, I mean, I was
41:31
one good argument away from buying a sauna when I went in
41:34
for burritos. So like they, they definitely
41:36
understand their customers. And of course I am the consumer,
41:41
you know, American man, you know, they know exactly what I
41:44
want. I'm like, oh, what, what's this
41:47
electric griddle? Oh, I do want electric griddle.
41:50
You know, so I think that is a brilliant product.
41:52
I I wonder how you would find it would be data based on like the
41:56
user's browser habits kind of? Is that what you're thinking?
42:00
That and also the publicly available data because like we
42:05
are now, consumers are just bombarded with the amount of
42:08
information and the ads that we get.
42:11
But when you, for example, click on an ad and you land, let's say
42:15
on Best Buy website, how do you know what you're actually
42:19
getting is the is the absolute best deal you can find in the
42:23
market? So that's one thing.
42:25
But to your point, to simply pull you back into the, into a
42:30
process, yes, you look at the historical patterns you can
42:36
through, you know, over time, understand the personality and
42:39
the, and the needs and the wants of the user.
42:41
And then you can do wonders. And then you know, you know,
42:45
understanding the timing, understanding, you know, when
42:49
you're, you know, your, your son is getting closer to a
42:53
graduation party and so on. So that actually gives you a lot
42:57
of personalization experience. But I think that the, the, the
43:00
biggest, the bigger challenge that this type of a solution can
43:03
solve is the fact that, you know, as a consumer today, you
43:08
are not sure if you can, if you're actually getting the best
43:11
deal. You can't be buying things on
43:14
sale. But is that really the best
43:18
thing that you can find in the market or not that that's a
43:21
that's a big question for for a lot of the consumers today.
43:25
Love it. Cool, So you just recently
43:28
closed seed round. Congratulations, that's amazing.
43:31
You'll you'll be going to for larger rounds soon.
43:34
I'm an investor myself and you know, you make a very strong
43:38
case, a very bullish case for for these kind of tools.
43:42
But as an investor, what should I be cautious about when I'm
43:45
thinking about AI infrastructure tools?
43:52
Well, I think we can always be excited about the amazing things
44:00
AI can do, but understanding the next wave of innovation and how
44:08
it translates back into solving real problems is extremely
44:12
important. Because excitement is one thing,
44:15
but at the end of the day, technology needs to solve humans
44:20
problems. And the challenge today is that
44:25
there are so many exciting flashy tools out there that get
44:28
you excited, but a lot of them can actually become, you know,
44:33
very quickly become irrelevant. Because one, the market is
44:39
moving so fast. A lot of the bigger players do
44:42
have great plans that they can they can think of and they can
44:45
always look at their usage powders and be even more
44:48
creative than you are. And second thing is that a lot
44:53
of the excitement that we get may not actually solve the real
44:56
problem in the real business settings.
44:59
Like for example, you can actually upload a document, the
45:02
legal document and get some high level understanding, better
45:06
understanding of that legal document from the AI.
45:10
But is it really something that can overtime be compliance and
45:16
you get some solid feedback from it or not?
45:18
Those are the questions. I do understand that.
45:20
What I'm telling you is a simple example of it.
45:22
But you can extrapolate from that and really understand that
45:26
no matter how effective the AI tools can be, but they may
45:31
never. A lot of the use cases that
45:32
we're thinking about it, they may never or they may not that
45:37
easily land into into environments where people where
45:41
the adoption can actually happen effectively.
45:44
So those are the things, I think it's the second big trend that
45:49
we should be cautious about is that how the big L&M models are
45:53
evolving and how easy it is for them to turn on a feature that
46:00
provides, you know, some value to the client.
46:03
So for example, the challenge to a lot of the prototyping tools
46:08
is that a lot of what they do can be done by the big L&M apps,
46:16
the open AI, the ChatGPT and Gemini from Google and so on and
46:20
so on. And so for example, I think just
46:23
recently open AI launch these different versions or this, you
46:28
know, slimmed down versions of a Photoshop that once you actually
46:34
generate your prototype where you are, you can you can use a
46:38
drag and drop features of the Photoshop that is just plugged
46:43
right into your workflow. So if you're not continuing on
46:49
the same trajectory or go deeper into the user workflow, you can
46:56
essentially become irrelevant very quickly by the big element
47:04
providers. Yeah.
47:05
Basically, it's make sure you got a customer.
47:08
I mean, I had a great conversation with a buddy of
47:10
mine's father. He's, you know, 76 years old.
47:13
He's worked, worked 3 generations, 70 years in a
47:16
flooring store. And we're just throwing around
47:19
AI stuff. He was curious about it.
47:20
So we're talking about it. And he's like, I don't know
47:22
about that, but I know about cash flow.
47:25
I think, you know, it's like, I was like, hey man, actually that
47:27
sensibility goes a long ways in building any kind of tools, you
47:31
know, so like making sure that you've got a customer, making
47:34
sure you can cash flow a business as an investor.
47:36
That might be like the brass tacks, hey, this technology
47:39
seems amazing, but what's the protection?
47:42
What's the Moat you're building around that Gemini can't just
47:44
scoop up in one second with 1 update?
47:47
Exactly. Yeah, I love that.
47:53
Is there ACEO or entrepreneur you follow closely?
48:00
Well, I usually learned a lot from Bill Gates.
48:04
Warren Buffett usually comes up with some very interesting
48:08
perspective, although I don't necessarily agree with him on
48:11
the AI hype. I don't think it's a pure hype.
48:16
There are a lot of great things that are happening in AI and,
48:23
you know, obviously, just like everybody else, I'm following
48:26
Elon Musk as well. I like the fact that he's so
48:31
brave, doing many things at the same time and just pushing the
48:35
boundaries so much. But I don't know, it's very hard
48:39
to be inspired by him because you kind of feel that he's just
48:43
like a superhuman. Maybe a villain yet.
48:47
Maybe both. Maybe both.
48:51
Yeah, 100%. Cool.
48:54
Just a couple more questions here.
48:55
This has been so fun. So the inception for Codalia was
48:59
like 5 years ago. What are maybe some some some
49:06
conceptions about start-ups that have changed since then or that
49:09
you like to update that you have updated?
49:13
Worse, I think one of the biggest patterns that I'm seeing
49:18
and I'm and it's just sometimes mind boggling is the fact that
49:26
the whole PPC and growing by spending on ads type of a
49:32
situation is seems to be gone. You can't win your way out to
49:38
growth that easily by just paying to Google and the meta.
49:42
They just eat up your entire margin and it's just very hard
49:46
to, you know, cut through the noise, get the attention of your
49:50
eye. So a lot of the new companies
49:54
are learning or coming up with creative ways to use social
49:58
media in many different ways and shape to just simply encourage
50:04
user adoption. Some of them are learning that
50:07
we're coming up with creative ways to float social media in a
50:11
very short span of time to just get the 1st 100,000 clients or
50:16
prompters very quickly. I think that's very interesting.
50:21
The other thing is that, you know, the whole BC model that,
50:28
you know, building many hyper scalable companies seems to be
50:32
impacted by this. And so raising capital has
50:35
become extremely difficult for a lot of entrepreneurs with
50:40
traditional way of thinking. And I have to say, people do
50:48
realize the fact that they need to move to market a lot faster.
50:53
So that traditional or you know, maybe previously, maybe up until
50:57
about a year or two years ago, people were spending maybe about
51:03
a year, up to about two years to really get to their product from
51:06
the, from the, from the inception to the product market
51:10
fits. So it would take them that long
51:12
to get to the market. In today's world you cannot
51:15
afford that. Yes, we talked about the fact
51:19
that we started the whole vision about five years ago.
51:22
But the moment we said we want to get into the market, we knew
51:24
we didn't have much of A time. So I am actually getting into
51:27
the market within about the year since we started to in publicly
51:31
talk about it. And we do realize the fact that
51:36
this is a, this is going to be an interesting time for us that,
51:39
you know, if you don't learn from those marvels, you may just
51:45
miss the boat. Didn't agree with you more.
51:50
Social media seems to be very strong, but I think you already
51:54
know and you're a little bit crafty.
51:56
That 2026 podcast, guest interviews and newsletters,
52:00
founder LED marketing. That's the best way to go to
52:03
market now. So I think I see that as part of
52:06
a social media, you know, distribution strategy.
52:11
You need to get to the communities, you need to the,
52:14
you need to get to environments where you can have a, you know,
52:20
more, more intimate conversation that you can get to the details
52:24
of what you're trying to do. Because you cannot be selling
52:28
concepts like what we're trying to do in Codalio by just, you
52:33
know, pushing some, you know, short videos.
52:37
The underlying technology of these types of solutions are not
52:40
that that as easy or as simple to be explained that quickly.
52:45
And the founder story has become extremely important to you.
52:49
So that's why the podcast and the way that that you folks are
52:53
actually running your interest in podcasts today is in my
52:58
opinion, is extremely important and can be very much beneficial
53:01
to a lot of other founders. I'll tell you what, interviewing
53:05
you is extremely beneficial to me.
53:07
This has been awesome. Hey son, I love it.
53:09
One thing I want to do is give a shout out to the Red Cross
53:12
because I know they're special to you.
53:14
Any particular Red Cross stories?
53:18
Well, I think yeah, we live in a time and era that we see a lot
53:21
of unfortunately crisis around the world.
53:29
I am just, I'm just happy that there is an institution like Red
53:36
Cross that deals with one with some of the most difficult
53:40
situations where people have to go through war, hardship and
53:46
crisis. And even if you can put a dance
53:52
in one day of one day of their life and make it just a little
53:56
bit better, I think that's very important.
54:01
You know, you always become you. You often become numb when you
54:07
look at all of the crazy things that are happening around the
54:09
world and when you look at them and when you hear them, see them
54:13
every day. But I think it's important to
54:15
just, you know, take a pause at some point and say maybe we can
54:19
contribute, maybe we can help out.
54:20
I think Red Cross has been amazing, doing some amazing work
54:23
around the world. For decades, you know, and I, I
54:26
love mentioning charitable organizations just from the pain
54:29
point of like not really knowing, you know, sometimes
54:32
like they're this things will come under your, your nose, you
54:36
know, charitable organization and you just can't tell who the
54:39
charlatans are. But Red Cross, the the work that
54:42
they do is so instrumental for humans across the planet and
54:45
just their reputation is absolutely bulletproof.
54:49
So I'm with you, man. I I'm really good that there's
54:51
are really glad that there's, you know, just such a
54:54
unbelievably robust humanitarian organization like the Red Cross.
54:59
So definitely shout out to those guys.
55:03
ISON the website is Codalio CODAL io.com.
55:10
What's the best place for people to connect with you online?
55:13
LinkedIn. LinkedIn.
55:16
Amazing. I'm going to I'm sending you a,
55:17
A friend ad right now. Sir, thank you so much for
55:22
joining us. I really appreciate you telling
55:23
your story. Best of luck with the next round
55:26
of revisions and in the next round of raising.
55:29
Thank you so much Ryan. Thanks for having me.
55:31
You made it a lot like a lot easier for me.
55:34
Thank you so much. It was casual the same time,
55:38
friendly and and technical some I mean in some areas and thank
55:44
you for having me. Absolute pleasure, thank you.
55:47
Hey founders, let me ask you this, are you building with AI
55:51
or getting left behind by it? If you're serious about scaling
55:55
smarter, faster and bolder, you need to check out AI for
56:00
founders.co. It's your go to resource for
56:03
cutting through the noise and actually using.
56:06
AI to grow your startup. Each week, AI for Founders drops
56:10
a sharp, no fluff newsletter, dives deep with expert guests on
56:14
the podcast, and hosts monthly workshops where you can get your
56:18
hands dirty with the real tools and strategies.
56:22
Whether you're precede or series.
56:24
B and beyond Launching something new or just staying sharp AI for
56:28
founders gives you the edge you need to win in 2025 and beyond.
56:33
Hit. AI for founders.co to subscribe
56:36
to the newsletter, tune into the podcast and grab your spot in
56:40
the next workshop. And it's all free.
56:42
Build smarter, move faster, think bolder with AI for
56:46
Founders. Now back to the show.