UX for AI

EP. 81 - No Code: How AI Lets Non-Tech Founders Build MVPs Faster w/ Carl Meran

• Bonanza Studios

Share episode

Send us a text

In this episode of UX For AI, we sit down with Carl Meran, a founder and technologist at the forefront of AI-driven product development. As AI continues to reshape the landscape of design, engineering, and entrepreneurship, Carl shares his firsthand experiences leveraging AI to accelerate product-building, automate workflows, and lower the technical barriers for founders and startups.

Carl discusses how tools like V0.dev, Claude, and Vercel are enabling semi-technical and even non-technical people to build products that would have once required full development teams. He breaks down how AI assists with coding, debugging, and workflow automation, allowing him to work as an engineer without a traditional engineering background. We explore how AI is shifting the startup landscape, making it possible to build a functional MVP (Minimum Viable Product) with a much smaller team—and in some cases, even as a solo founder.

One of the biggest areas of transformation Carl highlights is the Figma-to-code revolution. He believes that one of the next major breakthroughs in AI will be tools that can take Figma designs and instantly generate production-ready code. While some early attempts exist, Carl shares his thoughts on where these tools need to improve and how they could eventually streamline the entire product development lifecycle.

The conversation also covers how AI is affecting UX design, with new capabilities in automated interface generation, content creation, and even real-time translations of complex UI flows. As AI advances, the traditional boundaries between designers, developers, and product managers are becoming more fluid, allowing for a more collaborative and iterative process.

We also discuss the evolving role of founders in this new era. With AI-powered automation reducing the need for large startup teams, Carl argues that what once required ten engineers can now be done with three to five people. This fundamental shift in startup formation has massive implications for entrepreneurship, funding, and product innovation.

Carl also shares his bold predictions for the future of AI-driven development, suggesting that in the near future, founders may be able to simply describe a product idea, and AI will generate the entire application. While this vision is still a work in progress, he believes that we're getting closer to an era where AI acts as a true co-founder, developer, and designer all in one.

Whether you’re a designer, developer, startup founder, or simply curious about the future of AI and UX, this episode is packed with insights into how AI is reshaping the way we build and innovate.

Tune in to hear Carl Meran’s take on the most exciting trends in AI, UX, and product development—and what it all means for the future of tech.

You can find Carl Meran here:
https://www.linkedin.com/in/carl-meran/

Interested in joining the podcast? DM Behrad on LinkedIn:
https://www.linkedin.com/in/behradmirafshar/

This podcast is made by Bonanza Studios, Germany’s Premier Digital Design Studio:
https://www.bonanza-studios.com/

1
00:00:00,625 --> 00:00:04,541
Welcome to UX for AI, where we explore
how artificial

2
00:00:04,541 --> 00:00:07,958
intelligence is transforming
design and development.

3
00:00:08,333 --> 00:00:12,625
Today, we're joined by Carl to discuss
how AI is accelerating

4
00:00:12,625 --> 00:00:16,166
MVP creation and reshaping workflows.

5
00:00:16,375 --> 00:00:19,375
Let's dive in.

6
00:00:20,041 --> 00:00:20,583
Carl.

7
00:00:20,583 --> 00:00:24,208
So we finally turned
one of our conversations into a podcast.

8
00:00:25,166 --> 00:00:26,125
Absolutely.

9
00:00:26,125 --> 00:00:28,250
There have been many before. Yeah.

10
00:00:28,250 --> 00:00:31,791
So the reason I wanted you
to come on the pod,

11
00:00:32,000 --> 00:00:35,875
that I know without going so much in depth

12
00:00:36,750 --> 00:00:40,416
and that you experiment with a lot

13
00:00:40,416 --> 00:00:43,583
with the no code builder tools.

14
00:00:44,250 --> 00:00:48,416
I stuff, so I leave it up to you.

15
00:00:48,416 --> 00:00:52,416
I really want to, you know, spend
the next 30 minutes learning

16
00:00:52,416 --> 00:00:56,666
from your experimentation experiences
as you have with all these tools.

17
00:00:57,000 --> 00:00:57,833
What worked?

18
00:00:57,833 --> 00:01:00,500
What didn't work?

19
00:01:00,500 --> 00:01:03,833
What do you think the trend is in 2025?

20
00:01:03,833 --> 00:01:06,833
Basically.
Let's go. Yeah. Yeah. It's nice.

21
00:01:07,333 --> 00:01:09,250
Nice to be to be here. Yeah.

22
00:01:09,250 --> 00:01:10,666
Where should I start?

23
00:01:10,666 --> 00:01:15,500
Maybe a bit of a background, I would say
about your your sort of journey up to now.

24
00:01:15,500 --> 00:01:18,666
I know there are certain things
that we should not even,

25
00:01:18,750 --> 00:01:21,750
you know, touch about,
and maybe too soon anyways.

26
00:01:21,958 --> 00:01:24,541
But how did you end up here?

27
00:01:24,541 --> 00:01:29,416
And I know that you have a lot of fashion
fascination with all the tech

28
00:01:29,500 --> 00:01:33,291
related stuff, but I think that maybe
like five minutes into would be actually

29
00:01:33,291 --> 00:01:33,958
useful as well.

30
00:01:35,250 --> 00:01:36,208
Absolutely.

31
00:01:36,208 --> 00:01:36,416
Yeah.

32
00:01:36,416 --> 00:01:37,208
So I'm Karl,

33
00:01:37,208 --> 00:01:40,708
I'm originally from Austria, now
based in Berlin since almost three years.

34
00:01:42,041 --> 00:01:44,833
I studied computer science,
although I then went

35
00:01:44,833 --> 00:01:47,833
into, management consulting,

36
00:01:48,625 --> 00:01:52,250
but always kept my fascination
for software, for building, for tech.

37
00:01:52,958 --> 00:01:57,500
And then jumped ship, went to Berlin and.

38
00:01:57,500 --> 00:02:02,375
Yeah, I've been essentially active here
in the startup and tech world ever since.

39
00:02:02,375 --> 00:02:05,041
Since then.

40
00:02:05,041 --> 00:02:08,916
And yeah, I think
and it's probably the main theme,

41
00:02:08,916 --> 00:02:13,750
I think I'm
like one of these semi technical person.

42
00:02:13,750 --> 00:02:15,916
So I think I, I have a fairly good

43
00:02:15,916 --> 00:02:19,625
understanding of tech, of coding
as I've done it at university.

44
00:02:20,000 --> 00:02:23,000
However, I'm not an engineer by any means.

45
00:02:24,833 --> 00:02:27,541
And it's, it's really it's really this
when you ask me.

46
00:02:27,541 --> 00:02:32,000
So what's, what's what's the big trend in
AI that I'm seeing at the moment?

47
00:02:32,000 --> 00:02:35,250
At least it's, it's that it gives

48
00:02:36,166 --> 00:02:40,583
all these AI tools,
both in engineering as well as in,

49
00:02:41,666 --> 00:02:42,833
design.

50
00:02:42,833 --> 00:02:45,833
It gives semi technical people like myself

51
00:02:45,958 --> 00:02:50,250
such a powerful tool in our hands
so that that we can actually expand and go

52
00:02:50,625 --> 00:02:54,833
probably towards these 2 or 3 steps
further than we would have been able

53
00:02:55,166 --> 00:02:58,166
before,
and especially at a much faster pace.

54
00:02:58,208 --> 00:03:01,500
And so I think this is probably the theme
that you'll hear me talking,

55
00:03:02,250 --> 00:03:05,250
talking about today. And,

56
00:03:05,750 --> 00:03:07,791
yeah, happy to go deeper into that.

57
00:03:07,791 --> 00:03:10,583
I can testify also as well, folks.

58
00:03:10,583 --> 00:03:15,458
The stuff that he's doing is, I'm baffled
because for this stuff

59
00:03:15,458 --> 00:03:19,166
that Carl is doing three years ago,
let's say, yeah,

60
00:03:19,791 --> 00:03:24,625
three years ago would take at least
a team of three developers, right.

61
00:03:25,000 --> 00:03:27,166
To move to at least totally.

62
00:03:27,166 --> 00:03:31,125
And so so I think and I think
this is one of my, my main hypotheses,

63
00:03:31,750 --> 00:03:34,708
that we are at least seeing at the moment.

64
00:03:34,708 --> 00:03:38,541
Is that in particular
in the early stages of the company,

65
00:03:38,541 --> 00:03:43,375
in the MVP stages, in the stages where you
where there isn't a huge codebase, where

66
00:03:43,375 --> 00:03:48,083
there isn't, like large complexity and
and also where there aren't things like,

67
00:03:50,291 --> 00:03:52,250
security,

68
00:03:52,250 --> 00:03:56,125
GDPR or at least not
not as much as as before,

69
00:03:56,458 --> 00:04:00,083
but then also, yeah,
all of these like enterprise,

70
00:04:01,166 --> 00:04:04,833
features that that you get into once
once your company gets gets larger

71
00:04:05,125 --> 00:04:06,208
at this stage.

72
00:04:06,208 --> 00:04:09,458
This is really in my
this is really where I see AI,

73
00:04:10,166 --> 00:04:12,666
having a huge impact at the moment.

74
00:04:12,666 --> 00:04:17,291
In engineering
and partly in AI, in design.

75
00:04:17,333 --> 00:04:21,125
I mean, and maybe,
maybe a question to you, but have you had,

76
00:04:21,416 --> 00:04:25,208
access to, Sigma AI already?

77
00:04:25,916 --> 00:04:27,250
It's in beta.

78
00:04:27,250 --> 00:04:30,250
Yeah. Very excited.

79
00:04:30,625 --> 00:04:33,166
I think, you know, we

80
00:04:33,166 --> 00:04:36,750
we long we are past long.

81
00:04:36,750 --> 00:04:39,750
The phase of AI is our enemy.

82
00:04:40,916 --> 00:04:43,166
It's AI we

83
00:04:43,166 --> 00:04:46,291
our approach
here is how we can make friends with AI,

84
00:04:46,291 --> 00:04:49,625
and we can incorporate it
as soon as possible in our processes.

85
00:04:49,833 --> 00:04:51,208
Yeah, absolutely.

86
00:04:51,208 --> 00:04:56,125
And so, so I've, I've, I've had,
I've had the pleasure to see it and it's,

87
00:04:57,375 --> 00:04:58,166
it's on one

88
00:04:58,166 --> 00:05:02,416
hand, it's still, it does,
it does weird thing as we know, like all

89
00:05:02,708 --> 00:05:05,583
and sometimes they just do weird things.

90
00:05:05,583 --> 00:05:09,125
But I think on the, on the whole it's,
it's mind blowing

91
00:05:09,125 --> 00:05:12,125
and it just kind of like points
to the way that we are going.

92
00:05:12,708 --> 00:05:15,416
And I'll give I'll give two
very specific examples.

93
00:05:15,416 --> 00:05:16,583
One thing that we,

94
00:05:16,583 --> 00:05:21,083
that I've already done, been doing with
it is just create live graphics.

95
00:05:21,500 --> 00:05:24,833
So, for example, what you can do is, hey,
I just want to have like

96
00:05:24,833 --> 00:05:27,833
a small icon here or a small graphic

97
00:05:27,833 --> 00:05:30,833
that shows sustainability.

98
00:05:31,041 --> 00:05:33,333
As a, as an example.

99
00:05:33,333 --> 00:05:36,833
You just you just ask it to,
to feed my eye and it produces something.

100
00:05:36,833 --> 00:05:40,000
And, you know, after five iterations,
you're actually at something

101
00:05:40,416 --> 00:05:43,166
that is, at least usable.

102
00:05:43,166 --> 00:05:48,083
And the second one was, is translation
of of entire web pages and web flows.

103
00:05:49,208 --> 00:05:50,500
So I've, I've seen it actually.

104
00:05:50,500 --> 00:05:51,416
And, and screens.

105
00:05:51,416 --> 00:05:54,625
I've seen it do that in in seconds now

106
00:05:54,625 --> 00:05:57,625
which is, which is obviously great.

107
00:05:57,708 --> 00:05:59,875
Kind of like going from in Figma files.

108
00:05:59,875 --> 00:06:02,333
Yeah. Going from,

109
00:06:02,333 --> 00:06:05,333
kind of like an English translation
to a German one,

110
00:06:07,041 --> 00:06:10,500
within, within 35 seconds of, of like

111
00:06:10,500 --> 00:06:13,916
and these where we're talking
maybe 50 plus screens.

112
00:06:14,250 --> 00:06:18,083
So I think I think this is kind of like,
where, where I see at the moment like,

113
00:06:18,958 --> 00:06:19,916
a huge impact.

114
00:06:19,916 --> 00:06:22,916
Not in, like end to end workflows,
but in, in like

115
00:06:23,541 --> 00:06:26,541
point problems
that, that you always encounter.

116
00:06:27,208 --> 00:06:30,208
And that's just
and that's on the, on the design side,

117
00:06:30,791 --> 00:06:34,041
I think on the, on the engineering
side, my, my experience is actually

118
00:06:34,041 --> 00:06:35,208
is actually pretty similar.

119
00:06:36,208 --> 00:06:38,583
As I said before,

120
00:06:38,583 --> 00:06:42,583
I'm not an engineer,
but I am currently doing engineering work.

121
00:06:43,625 --> 00:06:47,125
Simply because AI is enable me,
enabling me to do that.

122
00:06:48,291 --> 00:06:53,541
I, I'm able to communicate
what I want to do, to the AI,

123
00:06:53,541 --> 00:06:57,083
and it gives me
at least the code snippets that I want.

124
00:06:57,083 --> 00:07:00,875
And then as I said this, as a kind of
I would, I would say

125
00:07:00,875 --> 00:07:03,291
semi technical person
and somebody who knows,

126
00:07:03,291 --> 00:07:08,125
like how code works and how to deal
with code, this is, it's a game changer.

127
00:07:10,083 --> 00:07:11,750
Because I would have

128
00:07:11,750 --> 00:07:15,833
I would have it would have taken me
probably days or weeks to produce code

129
00:07:15,833 --> 00:07:19,583
that has like that high quality
and, and works that.

130
00:07:19,583 --> 00:07:23,041
Well, before I went to very concrete.

131
00:07:23,666 --> 00:07:26,291
And and I don't even use,
like, very specific tools.

132
00:07:26,291 --> 00:07:26,833
I use,

133
00:07:28,000 --> 00:07:31,000
a V0 slow and,

134
00:07:31,500 --> 00:07:34,500
and simply clod, for, for coding.

135
00:07:34,708 --> 00:07:36,208
And I just have a conversation with them

136
00:07:36,208 --> 00:07:40,041
and, and work with them
as kind of like an engineering partner.

137
00:07:40,500 --> 00:07:46,666
So v0 and cloud and I wanted to ask you
what's your tool stack right now.

138
00:07:46,666 --> 00:07:49,625
So fantastic that you brought it up.

139
00:07:49,625 --> 00:07:52,583
What do you do on cloud
and what do you do on Vercel.

140
00:07:52,583 --> 00:07:52,791
Yeah.

141
00:07:52,791 --> 00:07:55,791
So so versus I have my entire,

142
00:07:56,666 --> 00:07:57,833
my entire code base.

143
00:07:57,833 --> 00:08:01,708
It's my it's my, code editor, and,

144
00:08:01,708 --> 00:08:06,416
and I, I also work with it
and integrate it, integrate with it.

145
00:08:06,750 --> 00:08:10,875
And then on on cloud, it's
it's, it's not fully consolidated yet.

146
00:08:10,875 --> 00:08:13,875
I could probably also do it in virtual,
but I'm floored is when I have like

147
00:08:14,166 --> 00:08:15,083
a specific,

148
00:08:16,166 --> 00:08:18,083
a specific problem,

149
00:08:18,083 --> 00:08:20,916
that I, that I want to solve
with like a specific function

150
00:08:20,916 --> 00:08:24,833
that doesn't have a lot, to do with my,
with the rest of my code base

151
00:08:25,250 --> 00:08:28,791
that that's where, that's where
I usually go for for stuff like that.

152
00:08:30,000 --> 00:08:31,583
I'll give you I'll give you a problem.

153
00:08:31,583 --> 00:08:34,791
I wanted to do a, an example of that.

154
00:08:34,791 --> 00:08:37,791
I wanted to do a backwards
interest rate calculation.

155
00:08:38,000 --> 00:08:41,000
So I had, I had the,

156
00:08:41,125 --> 00:08:44,166
I had a value, let's say,
at the end of the year.

157
00:08:44,541 --> 00:08:48,875
And I knew that we,
that it got an interest rate of 3%.

158
00:08:49,291 --> 00:08:52,375
And what I wanted to do is
I wanted to calculate backwards

159
00:08:52,708 --> 00:08:56,166
or like a daily interest rate
that I wouldn't would have gotten

160
00:08:56,166 --> 00:09:00,500
for every day of the previous
year and for stuff like that.

161
00:09:01,666 --> 00:09:03,875
This this
just simply takes 2 or 3 minutes.

162
00:09:03,875 --> 00:09:06,875
Just describe the problem. Try it out.

163
00:09:07,166 --> 00:09:08,875
And then do some. Do some testing.

164
00:09:08,875 --> 00:09:12,875
And Claude Cloud was able to solve this,
pretty much instantly.

165
00:09:13,500 --> 00:09:17,000
So right now, Claude is your developer.

166
00:09:17,250 --> 00:09:20,583
So you want to basically introduce
a new future

167
00:09:20,958 --> 00:09:24,375
capable or add to the algorithm.

168
00:09:24,750 --> 00:09:27,125
You said this is my problem.
What do you think?

169
00:09:27,125 --> 00:09:30,583
Do you have to sort of like,
showcase your version

170
00:09:30,583 --> 00:09:33,583
of code base to Claude or Claude?

171
00:09:33,583 --> 00:09:37,750
You act upon, you work with Claude
as if it's your concept

172
00:09:37,791 --> 00:09:41,666
or or developer
that gives you the concept.

173
00:09:41,666 --> 00:09:42,958
Then you bring it.

174
00:09:42,958 --> 00:09:46,666
So how do you like,
let's say, Claude offer you.

175
00:09:46,750 --> 00:09:49,083
Okay, this is
this is how you can approach it.

176
00:09:49,083 --> 00:09:51,500
Here is a code snippet.

177
00:09:51,500 --> 00:09:54,041
How do you then migrated versus.

178
00:09:54,041 --> 00:09:56,583
Yeah. So it's a it's a that's that

179
00:09:56,583 --> 00:10:00,083
that's already
kind of like in how I, how I give the

180
00:10:01,291 --> 00:10:02,833
give Claude the, the exercise.

181
00:10:02,833 --> 00:10:06,791
So obviously with Claude,
I just ask for a specific function.

182
00:10:06,791 --> 00:10:08,750
So here I say, okay, I have a function.

183
00:10:08,750 --> 00:10:10,333
This is my input.

184
00:10:10,333 --> 00:10:13,333
I have a I have a value,
I have an interest rate.

185
00:10:13,333 --> 00:10:15,666
I want you to give me an output
that looks like this.

186
00:10:15,666 --> 00:10:17,750
And then it does it does that.

187
00:10:17,750 --> 00:10:19,500
And then so it simply does it.

188
00:10:19,500 --> 00:10:22,000
And then I just simply copy and paste it
traversal.

189
00:10:22,000 --> 00:10:24,750
And and from there it's
where I have my code architecture.

190
00:10:24,750 --> 00:10:28,166
And I know what, what values I'm passing
in and what values I need to get out.

191
00:10:28,500 --> 00:10:30,958
And also the structure of both of these.

192
00:10:30,958 --> 00:10:34,625
And so so I think once you know that you
can, you can kind of you can combine them.

193
00:10:35,541 --> 00:10:38,666
And then on the other hand
versus it's just much better in terms of

194
00:10:38,666 --> 00:10:42,250
like just knowing
by knowing global variables that I have,

195
00:10:43,708 --> 00:10:46,625
knowing, knowing API calls and

196
00:10:46,625 --> 00:10:49,375
authentication
tokens and, and stuff like that.

197
00:10:49,375 --> 00:10:52,875
So when I write, I want to do stuff, stuff
that has especially to do

198
00:10:52,875 --> 00:10:56,666
with external APIs, that I use, I always,

199
00:10:56,666 --> 00:10:59,666
I always do it in straight in there.

200
00:11:00,541 --> 00:11:03,666
And to be honest, I'm not sure if the
if my setup currently is

201
00:11:03,666 --> 00:11:04,583
the it's the best one.

202
00:11:04,583 --> 00:11:07,583
It's just I kind of like got used to it.

203
00:11:07,666 --> 00:11:08,416
We all know it.

204
00:11:08,416 --> 00:11:11,416
We get used to kind of
like certain workflows,

205
00:11:11,416 --> 00:11:15,250
but that there might be optimization
potential there as well.

206
00:11:16,500 --> 00:11:19,041
And how do you go about sort of like,

207
00:11:19,041 --> 00:11:22,041
testing what you have, you know, built

208
00:11:22,583 --> 00:11:26,750
via help of cloud, move it to Vercel.

209
00:11:26,958 --> 00:11:30,625
So for the QM bug
fixing of that function, what do you do?

210
00:11:31,958 --> 00:11:34,416
Yeah,
I mean, that's at least at least for me.

211
00:11:34,416 --> 00:11:37,791
And in my workflow, that's, that's really
where the where most of the work

212
00:11:38,333 --> 00:11:42,625
still lies and where
most of my personal time goes into it.

213
00:11:42,916 --> 00:11:46,500
Because at the end of the day, like,
as we know, like sometimes

214
00:11:47,541 --> 00:11:51,333
they still do weird stuff
and they don't know the entire context.

215
00:11:52,041 --> 00:11:53,583
So you just have to grind it out.

216
00:11:53,583 --> 00:11:57,375
You just have to try everything
and and all

217
00:11:57,375 --> 00:12:00,375
possible options
and just make sure that your,

218
00:12:01,041 --> 00:12:04,250
that your,
your program doesn't do any weird stuff.

219
00:12:04,583 --> 00:12:08,958
I feel I but I think, I think this is,
this is where it kind of like gets to the,

220
00:12:09,375 --> 00:12:12,375
let's say more advanced or late stage

221
00:12:12,833 --> 00:12:15,166
normal product building processes,

222
00:12:15,166 --> 00:12:17,791
which I haven't seen impacted from.

223
00:12:17,791 --> 00:12:21,791
I, at least in my personal
workflow, as heavily as,

224
00:12:22,750 --> 00:12:24,791
as before,

225
00:12:24,791 --> 00:12:25,333
potentially.

226
00:12:25,333 --> 00:12:29,250
One thing Brad Rasberry is amazing
is just creating test data.

227
00:12:30,250 --> 00:12:32,000
I absolutely love doing that.

228
00:12:32,000 --> 00:12:35,125
It's, it's, it used to be such a pain

229
00:12:35,541 --> 00:12:38,541
to create more
or less realistic test data.

230
00:12:38,875 --> 00:12:42,958
And this is also something that you can
you can do it on any on any tool.

231
00:12:42,958 --> 00:12:45,958
ChatGPT, cloud, etc..

232
00:12:46,500 --> 00:12:49,500
Give it to structure,
give it what it should be.

233
00:12:49,750 --> 00:12:52,750
And it's actually really good to,
to just develop it.

234
00:12:52,791 --> 00:12:54,708
I even do it
sometimes. You can do it on notion.

235
00:12:54,708 --> 00:12:58,166
I, it's it's even even there.

236
00:12:58,166 --> 00:13:01,500
It can give you like good example data
when you for example,

237
00:13:01,500 --> 00:13:04,500
when you write down documentation.

238
00:13:04,500 --> 00:13:06,416
So that's,

239
00:13:06,416 --> 00:13:09,708
it's just, it's just a huge time, time

240
00:13:09,708 --> 00:13:13,375
saving that you, that you have to have,
like super manual task that is.

241
00:13:13,375 --> 00:13:14,916
Oh, that's a really good point.

242
00:13:14,916 --> 00:13:16,000
Yeah, that's a really good point.

243
00:13:16,000 --> 00:13:21,416
So basically you as chat were tell them
also to create the test data for you.

244
00:13:21,416 --> 00:13:23,375
You can test what you have built.

245
00:13:23,375 --> 00:13:24,500
Oh exactly.

246
00:13:24,500 --> 00:13:27,500
But the actual testing
I didn't have to do myself.

247
00:13:27,708 --> 00:13:28,916
But it's just much easier.

248
00:13:28,916 --> 00:13:30,958
I can I can populate a database with

249
00:13:31,958 --> 00:13:33,583
with like

250
00:13:33,583 --> 00:13:38,125
synthetically generated test data and,
then just do

251
00:13:38,125 --> 00:13:41,125
the normal testing, processes and,

252
00:13:42,500 --> 00:13:42,833
yeah.

253
00:13:42,833 --> 00:13:46,125
So one of my curiosity here, Carl, is that

254
00:13:46,166 --> 00:13:49,583
you have a background in computer
engineering, right?

255
00:13:50,000 --> 00:13:53,000
So you are no, you know, you're

256
00:13:53,083 --> 00:13:56,875
no stranger to to programing.

257
00:13:56,875 --> 00:13:59,875
At least you have certain
understanding of it.

258
00:14:00,083 --> 00:14:03,833
Can a founder
that is absolutely non-technical,

259
00:14:03,833 --> 00:14:07,541
let's say they studied psychology
or architecture

260
00:14:07,541 --> 00:14:10,708
or the MBA.

261
00:14:11,000 --> 00:14:12,541
Right.

262
00:14:12,541 --> 00:14:17,958
Can they do what you do right now

263
00:14:17,958 --> 00:14:23,708
with a handful of tools and build MVP or

264
00:14:24,958 --> 00:14:27,541
do they need to involve someone else

265
00:14:27,541 --> 00:14:30,541
to a certain degree?

266
00:14:30,625 --> 00:14:33,416
And so obviously the
the answer is it depends.

267
00:14:33,416 --> 00:14:36,166
But I think in, in, in general,

268
00:14:36,166 --> 00:14:40,791
if you are a smart person with some vague
understanding in your quick

269
00:14:41,916 --> 00:14:43,583
I think, I think there's two variables

270
00:14:43,583 --> 00:14:47,500
that that I think now
have significantly changed is one, you can

271
00:14:47,500 --> 00:14:50,500
you can get to a higher level,
much quicker.

272
00:14:51,833 --> 00:14:53,708
Also just on your own.

273
00:14:53,708 --> 00:14:57,125
And second,
you can go 2 or 3 steps further than you

274
00:14:57,125 --> 00:14:58,333
that you would have otherwise.

275
00:14:58,333 --> 00:15:01,708
And I think I would say typically
I would actually say yes

276
00:15:01,708 --> 00:15:04,958
and maybe even have like a vaguely
complex,

277
00:15:06,416 --> 00:15:09,750
product can be built
by an absolutely non-technical person.

278
00:15:10,375 --> 00:15:14,708
And then obviously, as you scale,
and as you actually start

279
00:15:14,708 --> 00:15:18,000
to create an engineering,
you need a veneering function.

280
00:15:18,000 --> 00:15:21,958
That's simply because the volume is
getting, is getting bigger.

281
00:15:21,958 --> 00:15:25,583
I think that's when you
then need to involve definitely also hire

282
00:15:25,583 --> 00:15:27,958
a technical team.

283
00:15:27,958 --> 00:15:30,541
I don't see it yet
that you can have absolutely

284
00:15:30,541 --> 00:15:33,333
no technical team in particular.

285
00:15:33,333 --> 00:15:37,083
Once you get to a certain scale
or once you start

286
00:15:37,083 --> 00:15:41,166
to sell to certain people who have an A,
in particular in the enterprise sector.

287
00:15:44,125 --> 00:15:44,875
But we'll get there.

288
00:15:44,875 --> 00:15:45,541
Maybe.

289
00:15:45,541 --> 00:15:49,416
Let's see I and it's we are
we are at the very beginning.

290
00:15:49,416 --> 00:15:52,166
So I think that's always
the most important thing to remember.

291
00:15:52,166 --> 00:15:54,375
This is so new.

292
00:15:54,375 --> 00:15:58,166
We are in the early 2000
of that of the computing

293
00:15:58,166 --> 00:16:02,666
era is like right now
it's only a ticking bomb, in my opinion.

294
00:16:02,666 --> 00:16:07,208
Maybe like if we recorded the same episode
six months down the road,

295
00:16:07,208 --> 00:16:10,208
we would laugh at exactly, exactly.

296
00:16:11,250 --> 00:16:14,041
What have we been thinking?

297
00:16:14,041 --> 00:16:16,250
Six months ago?

298
00:16:16,250 --> 00:16:19,333
And. Yeah, but I think there is married to

299
00:16:19,333 --> 00:16:22,750
what you said is that I think, well, it

300
00:16:24,208 --> 00:16:27,416
compared to two, three years ago
that the founders,

301
00:16:27,666 --> 00:16:30,666
they really needed.

302
00:16:31,625 --> 00:16:34,791
When they needed an MVP,
they needed to at least

303
00:16:34,791 --> 00:16:38,416
have one developer
to do something here to get something out.

304
00:16:38,708 --> 00:16:41,958
I think that initial phase right now
with the current

305
00:16:42,291 --> 00:16:45,583
AI abilities,
if a founder really wants it,

306
00:16:46,083 --> 00:16:50,125
they could sit basically, to your point,
grind it out.

307
00:16:50,958 --> 00:16:53,791
I agree, and I think maybe,
maybe they can onboard

308
00:16:53,791 --> 00:16:57,125
the first ten,
20, 50, maybe even a hundred users.

309
00:16:58,708 --> 00:16:59,875
And then simply

310
00:16:59,875 --> 00:17:02,875
depending on what,
what exactly they're doing,

311
00:17:03,041 --> 00:17:06,041
that's where I currently see I

312
00:17:06,125 --> 00:17:08,750
to at least stop working.

313
00:17:08,750 --> 00:17:10,625
But but it doesn't,
it doesn't really matter

314
00:17:10,625 --> 00:17:15,291
because obviously at this stage
at the company or at the at a project,

315
00:17:15,291 --> 00:17:18,000
you would
you would have to hire people anyways

316
00:17:18,000 --> 00:17:22,041
because the entire point is to just prove,
product market fit.

317
00:17:22,041 --> 00:17:22,916
It proves that you have

318
00:17:22,916 --> 00:17:25,958
a value proposition, that is,
that people are willing to pay for.

319
00:17:26,708 --> 00:17:29,541
And going from that,

320
00:17:29,541 --> 00:17:32,666
coming to that point, at the moment,
it's just has just become

321
00:17:32,666 --> 00:17:34,250
so much cheaper and faster.

322
00:17:34,250 --> 00:17:39,458
I think, you know, the, the way I'm
thinking about it, I the way we are seeing

323
00:17:39,625 --> 00:17:43,916
you said very nicely that just
the beginning of the AI phase, you know,

324
00:17:44,333 --> 00:17:49,041
it's been it's been less than a year
that for all chat to be true for all this.

325
00:17:49,041 --> 00:17:52,750
I think that was a hallmark
for the AI application

326
00:17:52,750 --> 00:17:56,541
and all the versions before was nice.

327
00:17:56,541 --> 00:18:00,083
It could, you know,
get copy and do certain things.

328
00:18:00,083 --> 00:18:03,166
But for all was actually, in my opinion,

329
00:18:03,458 --> 00:18:06,333
that you could really build
application on top of it.

330
00:18:08,000 --> 00:18:10,291
And it's been around less than a year,
I would say.

331
00:18:10,291 --> 00:18:12,416
I don't even know this.

332
00:18:12,416 --> 00:18:15,416
This version's coming so fast,

333
00:18:16,166 --> 00:18:19,375
but given the speed,
given the progress that we are seeing,

334
00:18:20,041 --> 00:18:23,250
I think even 4 or 5 months down

335
00:18:23,250 --> 00:18:28,708
the road card
that the MVP that the process

336
00:18:28,708 --> 00:18:31,708
of the outline that you used to

337
00:18:31,916 --> 00:18:34,916
to to develop the MVP,

338
00:18:35,458 --> 00:18:39,916
I think that's going to even
be more straightforward than today.

339
00:18:40,791 --> 00:18:45,291
So basically,
all you need to do as a founder

340
00:18:45,291 --> 00:18:48,291
to have a really detailed

341
00:18:48,500 --> 00:18:52,541
a specification for your MVP
literally uploaded

342
00:18:53,166 --> 00:18:57,458
and be 1 to 1 day to one week of tweaking.

343
00:18:57,458 --> 00:18:58,833
You have your MVP out.

344
00:18:58,833 --> 00:19:00,291
I mean, it could be

345
00:19:00,291 --> 00:19:03,625
so, so I think I think there's kind of
like this in my way of thinking.

346
00:19:03,625 --> 00:19:04,666
There's three big steps.

347
00:19:04,666 --> 00:19:10,041
So there's there's the idea,
the user feedback, the user value.

348
00:19:10,041 --> 00:19:15,625
And then you move that towards a, towards
essentially a design, a software.

349
00:19:15,625 --> 00:19:18,625
But and let, let's,
let's make it super two specific.

350
00:19:18,625 --> 00:19:22,208
You think of the of the idea,
of the value proposition.

351
00:19:22,208 --> 00:19:24,083
You move it to Figma.

352
00:19:24,083 --> 00:19:25,708
So that's one step.

353
00:19:25,708 --> 00:19:28,708
And then you move it from Figma to code.

354
00:19:28,708 --> 00:19:30,166
That's the that's the second.

355
00:19:30,166 --> 00:19:31,750
That's the second step.

356
00:19:31,750 --> 00:19:35,625
And then you obviously do do lots of stuff
and kind of like grow it there.

357
00:19:36,000 --> 00:19:39,625
But if you, if you have these super
simplistic, two intersections

358
00:19:40,458 --> 00:19:43,416
where I think I'm going to see it,
we are going to see

359
00:19:43,416 --> 00:19:47,333
the biggest progress in the next year
is actually the step from Figma to code.

360
00:19:48,416 --> 00:19:51,416
So I'm, I've seen some stuff.

361
00:19:51,625 --> 00:19:56,166
I wasn't super convinced yet, but really,
this, this idea to actually go from

362
00:19:57,041 --> 00:19:59,416
from designs
from a design value proposition

363
00:19:59,416 --> 00:20:02,875
towards code, is what I think
and what I hope.

364
00:20:02,875 --> 00:20:03,750
Where is this?

365
00:20:03,750 --> 00:20:06,750
Where are we going to see
a big, step change,

366
00:20:08,000 --> 00:20:12,000
in hopefully the next year years,
who knows. But

367
00:20:13,208 --> 00:20:15,833
and then the,
the obviously the step before

368
00:20:15,833 --> 00:20:20,000
I think that's where I'm, I'm very curious
how far we're going to get there.

369
00:20:20,000 --> 00:20:24,000
You can think of how do we get like
user feedback, but also like

370
00:20:24,000 --> 00:20:28,500
a founder's idea into a, into good design,

371
00:20:28,500 --> 00:20:31,500
maybe with the help of like, design,

372
00:20:31,500 --> 00:20:34,291
design bots, copilot or whatever

373
00:20:34,291 --> 00:20:37,291
you want an agent,
whatever you want to call it.

374
00:20:37,333 --> 00:20:41,333
So, so I think I think kind of like
and especially these two intersections

375
00:20:41,875 --> 00:20:45,125
and, and major steps
is where I'm hoping to

376
00:20:45,583 --> 00:20:48,583
to see a lot more innovation
and a lot more work to be done.

377
00:20:48,791 --> 00:20:49,250
Oh, wow.

378
00:20:49,250 --> 00:20:51,375
So you're not even worried
about the other side.

379
00:20:51,375 --> 00:20:55,250
You're more worried about,
bringing the idea to the design that's

380
00:20:55,791 --> 00:20:58,541
that's one, one big hurdle, you see.

381
00:20:58,541 --> 00:21:01,541
And then I think after.

382
00:21:01,666 --> 00:21:02,375
Yeah.

383
00:21:02,375 --> 00:21:06,083
Because I think that's that's
also bottleneck because from the figma

384
00:21:06,083 --> 00:21:09,750
to the final, ready to be serve,

385
00:21:10,083 --> 00:21:13,458
what you need to do, you need to take care
of your servers and hosting.

386
00:21:13,458 --> 00:21:15,250
You need to take care of the data
based model.

387
00:21:15,250 --> 00:21:19,208
You need to take care of your backend,
and you need to take care of the frontend.

388
00:21:19,208 --> 00:21:21,000
So the input

389
00:21:22,083 --> 00:21:23,791
signal, Figma as an input

390
00:21:23,791 --> 00:21:26,791
to that system would.

391
00:21:27,833 --> 00:21:30,291
It so Figma is the output.

392
00:21:30,291 --> 00:21:32,750
And we need to basically make sense

393
00:21:32,750 --> 00:21:35,916
of this output
and pass it into four streams.

394
00:21:36,708 --> 00:21:37,000
Right.

395
00:21:37,000 --> 00:21:40,000
So basically the Figma design

396
00:21:40,000 --> 00:21:43,000
as output needs to fit,

397
00:21:43,791 --> 00:21:46,791
needs to be used as input for

398
00:21:46,916 --> 00:21:49,583
for these streams, hosting servers,

399
00:21:49,583 --> 00:21:52,583
database model, backend and frontend.

400
00:21:53,250 --> 00:21:55,375
Right. Absolutely. Exactly.

401
00:21:55,375 --> 00:21:59,541
And but I think what's important here
is that the Figma,

402
00:21:59,541 --> 00:22:01,083
because the figma is kind of fluid.

403
00:22:01,083 --> 00:22:02,375
It's the value proposition.

404
00:22:02,375 --> 00:22:06,291
It would really solve the,
the customer, customer's problem.

405
00:22:06,291 --> 00:22:09,291
And obviously at the moment
this is what you have like huge

406
00:22:09,291 --> 00:22:12,625
product management,
design and engineering teams for to

407
00:22:13,083 --> 00:22:15,750
to to supervise this

408
00:22:15,750 --> 00:22:18,666
and, and to coordinate this,

409
00:22:18,666 --> 00:22:21,416
these tasks and

410
00:22:21,416 --> 00:22:24,416
let's see if we're going to live
in a world where all of this,

411
00:22:24,916 --> 00:22:28,500
can be understood
and, or can be streamlined,

412
00:22:28,500 --> 00:22:32,333
at least to have at least partly by,
by AI tools.

413
00:22:32,333 --> 00:22:36,500
So let's say, sort of work
as a sort of closing question is that

414
00:22:37,083 --> 00:22:40,083
let's say you're working on a startup,

415
00:22:42,041 --> 00:22:45,041
initial validation successful,
you get some,

416
00:22:46,541 --> 00:22:49,375
you know, you raise

417
00:22:49,375 --> 00:22:50,541
you raise some

418
00:22:50,541 --> 00:22:53,541
so you can put together a team,

419
00:22:53,791 --> 00:22:56,083
for the next year. What?

420
00:22:56,083 --> 00:22:59,708
Who would be if you are the manager
or manager

421
00:23:00,500 --> 00:23:03,750
manager of that team,
who would you want to hire?

422
00:23:04,875 --> 00:23:07,125
I mean, it's a it's a difficult question.

423
00:23:07,125 --> 00:23:10,500
It very much depends
on, on the type of application

424
00:23:10,500 --> 00:23:13,500
that you're, you're building.

425
00:23:13,708 --> 00:23:16,708
Let's say for a B2B SAS that,

426
00:23:18,625 --> 00:23:21,750
either
in, in short tech, let's say like this

427
00:23:22,166 --> 00:23:25,458
or fintech,
let's, let's, let's go into fintech.

428
00:23:25,708 --> 00:23:27,625
I mean, I think so.

429
00:23:27,625 --> 00:23:29,625
Absolutely. So

430
00:23:29,625 --> 00:23:32,125
probably for the, for the first,

431
00:23:32,125 --> 00:23:34,250
for the first iteration,

432
00:23:34,250 --> 00:23:37,250
I think the way I would go is.

433
00:23:38,791 --> 00:23:41,875
Work a lot with like, internally

434
00:23:41,875 --> 00:23:46,291
with, business people, but also customer
support, customer success people.

435
00:23:46,625 --> 00:23:49,916
So these are really the the people
that ensure your value proposition.

436
00:23:50,666 --> 00:23:53,791
And I would probably outsource,

437
00:23:55,458 --> 00:23:58,375
a little bit of the engineering and,

438
00:23:58,375 --> 00:24:00,750
and the design in the beginning

439
00:24:00,750 --> 00:24:03,750
before obviously then taking all of that

440
00:24:04,250 --> 00:24:06,333
internally once you're ready to scale.

441
00:24:06,333 --> 00:24:08,666
So I think,

442
00:24:08,666 --> 00:24:11,666
proving the value proposition
proven to go to market,

443
00:24:12,375 --> 00:24:13,083
etc..

444
00:24:13,083 --> 00:24:16,083
This is, this is, this is before you,

445
00:24:16,375 --> 00:24:19,375
prove product market fit or even like,

446
00:24:19,625 --> 00:24:22,083
the fit that you can actually scale that.

447
00:24:22,083 --> 00:24:24,666
And this is the time when I would say,

448
00:24:24,666 --> 00:24:27,208
design and engineering can be.

449
00:24:27,208 --> 00:24:31,000
It doesn't have to be,
but it can be, outsource to it

450
00:24:31,000 --> 00:24:34,000
to, to various degrees.

451
00:24:34,083 --> 00:24:34,708
Yeah.

452
00:24:34,708 --> 00:24:37,500
And then
and then it needs to go, go inside

453
00:24:37,500 --> 00:24:40,625
and become a part of your,
of your identity.

454
00:24:40,625 --> 00:24:41,333
Obviously.

455
00:24:41,333 --> 00:24:44,333
Would you think that

456
00:24:44,750 --> 00:24:47,666
startups and the founding team,

457
00:24:47,666 --> 00:24:51,041
obviously the startup teams

458
00:24:51,500 --> 00:24:54,791
are going to shrink drastically
from now on.

459
00:24:54,875 --> 00:24:57,875
So I think I think if you were,

460
00:24:58,041 --> 00:25:02,041
if you needed it,
ten people as sort of like a

461
00:25:03,333 --> 00:25:06,333
seed stage, team formation,

462
00:25:07,333 --> 00:25:11,541
to do fintech, I think right now
you could pretty much do it

463
00:25:11,541 --> 00:25:14,583
with like 3 to 5, probably probably two

464
00:25:14,875 --> 00:25:17,833
plus obviously external, help.

465
00:25:17,833 --> 00:25:22,250
So I think I think the external help here
is you'll definitely need this,

466
00:25:22,583 --> 00:25:26,666
but you likely don't need them
as full time employees.

467
00:25:26,708 --> 00:25:29,708
But I think part time capacity

468
00:25:29,708 --> 00:25:33,083
can for some for
some of these applications be enough.

469
00:25:33,208 --> 00:25:35,125
That's where I'm really thinking.

470
00:25:35,125 --> 00:25:39,416
I had this conversation yesterday
with another founder.

471
00:25:39,416 --> 00:25:42,416
I think right now
is the time for startups.

472
00:25:43,083 --> 00:25:46,083
We we've seen many booms
in the startup scene.

473
00:25:47,208 --> 00:25:50,166
But again, like five years ago,

474
00:25:50,166 --> 00:25:54,250
if you if you were to go
to, investor pitch and you say, like,

475
00:25:54,250 --> 00:25:57,375
I want to build Salesforce,
I want to compete with Salesforce,

476
00:25:57,750 --> 00:25:59,458
they would laugh you out of the.

477
00:26:01,291 --> 00:26:02,083
But right

478
00:26:02,083 --> 00:26:05,083
now, if you go to the investor pitch and

479
00:26:06,291 --> 00:26:08,583
say something along the line of,

480
00:26:08,583 --> 00:26:12,208
I want to compete with Salesforce,
they would take you

481
00:26:13,375 --> 00:26:16,125
and they would at least
they would see your pitch

482
00:26:16,125 --> 00:26:18,875
because it's possible.

483
00:26:18,875 --> 00:26:20,083
It's exactly.

484
00:26:20,083 --> 00:26:22,625
It seems like maybe
I should give that a go.

485
00:26:22,625 --> 00:26:25,541
Actually, Carl.

486
00:26:25,541 --> 00:26:29,250
Yeah, I know you're busy,
and I know the 2025

487
00:26:29,250 --> 00:26:30,916
is going to be a busy year for you.

488
00:26:30,916 --> 00:26:35,166
Any last words or,
do you see yourself working as a startup

489
00:26:35,166 --> 00:26:41,000
founders in 2026 or are
we'll be taking your work as well, or.

490
00:26:41,791 --> 00:26:44,333
Yeah. Closing marks. Oh, hopefully.

491
00:26:44,333 --> 00:26:46,125
I think I'll keep working on that.

492
00:26:46,125 --> 00:26:49,125
But, yeah, you'll hear more soon.

493
00:26:50,750 --> 00:26:52,750
Pleasure.

494
00:26:52,750 --> 00:26:54,291
Thank you for listening to UX

495
00:26:54,291 --> 00:26:55,375
For AI.

496
00:26:55,375 --> 00:26:59,458
Join us next week for more insightful
conversations about the impact

497
00:26:59,458 --> 00:27:04,416
of artificial intelligence in development,
design, and user experience.