01 | Cooking Up Apps with Jacquelyn Halpern

1
00:00:00,000 --> 00:00:15,440
Welcome to Conjuring Code. My name is Ian Varley. I am so excited to share this first
2
00:00:15,440 --> 00:00:20,720
episode of the podcast with you all. Our guest today is Jacqueline Halpern and in this episode
3
00:00:20,720 --> 00:00:25,920
we're going to talk about how she got started building software with AI help and the amazing
4
00:00:25,920 --> 00:00:30,480
app that she built for herself and her family called Lil Chef. So let's dive in.
5
00:00:37,200 --> 00:00:41,760
So Jacqueline, welcome to the Conjuring Code podcast. I'm so excited to have you on.
6
00:00:41,760 --> 00:00:43,680
Super excited to be here. Waited all week for this.
7
00:00:45,120 --> 00:00:49,280
So, you know, Jacqueline and I have known each other for a few years. We have worked at the
8
00:00:49,280 --> 00:00:55,680
same company for a while. And, you know, when we first met and first started talking,
9
00:00:55,680 --> 00:01:02,640
I remember you were saying things like, hey, I'm, you know, I'm not an engineer by training,
10
00:01:02,640 --> 00:01:06,320
but I've been doing a lot of engineering work with AI. And we started talking about it and I
11
00:01:06,320 --> 00:01:11,440
was like, wow, this is really fascinating. And you were doing this well ahead of, I feel like,
12
00:01:11,440 --> 00:01:15,280
when most other people were doing it. So I think, I think it'd be probably a pretty interesting
13
00:01:15,280 --> 00:01:18,400
story for people to hear. First of all, just tell me a little bit about yourself. Like,
14
00:01:18,400 --> 00:01:21,360
what do you do for a living? Where do you live? That kind of stuff.
15
00:01:22,080 --> 00:01:27,760
Well, currently for a living product designer, which actually I just went and spoke at one of
16
00:01:27,760 --> 00:01:33,600
our local universities and the professor had introduced me and then nobody knew what a product
17
00:01:33,600 --> 00:01:37,440
designer was. So what can you explain to our audience? What is a product designer?
18
00:01:37,440 --> 00:01:42,720
Yeah, I think when I would be like UI, UX, products and apps, then people kind of bought it. So
19
00:01:43,680 --> 00:01:50,720
I think the term has evolved. Product designer really just means I work on apps, my specific
20
00:01:50,720 --> 00:01:56,080
case, web apps, right? You could have product designers that are native iOS and Android, but
21
00:01:56,800 --> 00:02:02,720
pretty much designing end to end experience for products. In this case, a very specific product
22
00:02:02,720 --> 00:02:07,120
called Trailhead at Salesforce. Okay, awesome. And anyone who's, you know,
23
00:02:07,120 --> 00:02:12,480
ever used Trailhead knows it's a pretty delightful experience, actually. Like it's fun. It looks
24
00:02:12,480 --> 00:02:17,600
inviting. It feels nice. So y'all do a good job. I mean, when you first said product designer,
25
00:02:18,160 --> 00:02:22,400
like the image that popped into my head was like, oh, like there's lots of products in the world,
26
00:02:22,400 --> 00:02:26,640
like, you know, toothbrushes and tires and whatever. Like, do you design those? But you mean
27
00:02:26,640 --> 00:02:31,600
like you were a software product designer? Software, yeah. Yeah, awesome. And where are
28
00:02:31,600 --> 00:02:38,160
you based? Indianapolis, Indiana. Oh, very nice. Currently, I think two degrees. Two, that's not
29
00:02:38,160 --> 00:02:44,000
very many. That's a small number of degrees. We're recording this in January of 2026 and I'm in
30
00:02:44,000 --> 00:02:49,840
Austin and it's cold by our standards here. It was below freezing this morning, but not quite
31
00:02:49,840 --> 00:02:55,680
the same as Indy. You know, so you've been in the software business, but not a software engineer.
32
00:02:55,680 --> 00:03:02,640
What got you interested in trying to actually make some software? So it's probably the same thing
33
00:03:02,640 --> 00:03:08,480
where like if I ever do talks and stuff, I always just kind of label myself as a curious creative,
34
00:03:09,120 --> 00:03:15,120
which also really just means I just want to do a lot of things. So before I was a designer,
35
00:03:15,120 --> 00:03:20,640
I wanted to be an architect. At one point I did want to know how to code, but it's important to
36
00:03:20,640 --> 00:03:26,000
know that it wasn't until my 30s I found out I'm dyslexic, which makes a lot of sense why I couldn't
37
00:03:26,000 --> 00:03:31,200
pursue some of these things. But again, I always wanted to be an animator. I want to make a short
38
00:03:31,280 --> 00:03:38,080
film. I wanted to write children's books. And as a product designer, I think all of us
39
00:03:38,080 --> 00:03:41,040
would love to know how to code. It's one of those things where it's like,
40
00:03:41,840 --> 00:03:48,320
we can only take our ideas so far, right? And so for me, code was always the blocker.
41
00:03:48,960 --> 00:03:54,720
I would try, I would try to learn code, but for someone who's dyslexic, it just like didn't go
42
00:03:54,720 --> 00:04:01,680
good. I mean, even without the challenge of dyslexia, coding is very fidgety. It's very
43
00:04:01,680 --> 00:04:08,160
difficult for most, like in my experience, kind of most people find it to be off-putting. And they
44
00:04:08,160 --> 00:04:13,440
find it to be off-putting in part because it's like, oh, it's very specific. It's very arcane.
45
00:04:13,440 --> 00:04:19,840
And there's so many details. It's just not how most of the work in your life works.
46
00:04:19,840 --> 00:04:23,680
And then you get into this domain and it's extremely exacting. It has to be perfect.
47
00:04:23,680 --> 00:04:26,160
And you're like, oh, no, no, this, I don't, I don't like this.
48
00:04:26,160 --> 00:04:30,640
Yeah. I mean, it's very intimidating too, because you have to think like prior to
49
00:04:31,600 --> 00:04:36,800
products like Replet and Bolt and Lovable, it's, you're still dealing with something like the
50
00:04:36,800 --> 00:04:44,000
terminal, which in general is not designed for non-technical, non-engineering people.
51
00:04:44,000 --> 00:04:48,240
And so I would try and it's like, okay, like it was one of those things I was like,
52
00:04:48,320 --> 00:04:54,080
I accept that like, this is not probably in the cards for me if I wanted to, maybe.
53
00:04:54,080 --> 00:04:59,120
But yeah, it was always like, I get to a point or I'd have all these ideas that could only
54
00:04:59,120 --> 00:05:04,560
come to life if like you're really coding because up to that point before vibe coding,
55
00:05:04,560 --> 00:05:10,000
static prototypes, we used to design in Photoshop, but, and then, you know, sketch came out and Figma
56
00:05:10,000 --> 00:05:16,320
came out and the prototyping advanced, but you're still not able to get like to that final place.
57
00:05:16,320 --> 00:05:20,880
Right. That's why when we work with engineers, it's always back and forth. It's never just a
58
00:05:20,880 --> 00:05:26,800
clean handoff. It's like, oh, okay. Now that I see it live, now that I can interact, now that it's
59
00:05:26,800 --> 00:05:30,880
like, let's change some things. I've had that experience myself a bunch of times in the sense
60
00:05:30,880 --> 00:05:36,320
of like, it's almost like you're moving from 2D to 3D in a certain sense when it goes from,
61
00:05:36,320 --> 00:05:40,960
you know, a prototype to the real thing. I used to use all kinds of tools like Balsamic and,
62
00:05:40,960 --> 00:05:45,840
you know, Acture and things like that for doing more interactive prototyping, right? Where you
63
00:05:45,840 --> 00:05:50,400
could be like, oh, you can click the thing and it changes screens and some of that stuff,
64
00:05:50,400 --> 00:05:55,520
but it's still not quite all the way to the real app because when you get different kinds of data
65
00:05:55,520 --> 00:05:58,960
in there or like this, you know, the name of the thing is longer than you ever thought it would be
66
00:05:58,960 --> 00:06:02,640
what happens and there's just like a whole bunch of stuff that's like not real until it's real.
67
00:06:02,640 --> 00:06:07,040
Oh yeah. Especially like, you know, you forget about things like localization and translation
68
00:06:07,040 --> 00:06:13,040
and to your point, like, oh, this whole thing's in German now or, you know, even things like Arabic
69
00:06:13,040 --> 00:06:18,480
or when it switches from right to left, there's just like so many things. And then just like,
70
00:06:18,480 --> 00:06:24,480
personally, I'm a creative person in general, so I would have ideas outside of work that I just
71
00:06:24,480 --> 00:06:30,960
wanted to do or experiment with. I always have been fascinated with creative coding or generative
72
00:06:30,960 --> 00:06:37,040
code. Like I got a pen plotter. Like that stuff to me was always cool and it goes back forever,
73
00:06:37,040 --> 00:06:42,960
this like collaboration between human and machine of like creating something that is artistic or
74
00:06:42,960 --> 00:06:48,560
creative out of code. That was kind of like my first attempt into like this like vibe coding,
75
00:06:48,560 --> 00:06:54,080
but I don't remember the exact date, February something when Karthi coined the term. So there
76
00:06:54,080 --> 00:07:04,080
was no term before that, but it was researching like JavaScript and I think it's like p.j5 and
77
00:07:04,080 --> 00:07:10,400
whatever. And I would try to go to ChatGPT. This was when ChatGPT was like just starting to be like,
78
00:07:10,400 --> 00:07:15,360
oh, it's like really good at coding. It's really good at like this. And so I would start there and
79
00:07:15,360 --> 00:07:21,920
I would get some interesting things. And so I think that was like the spark. I'd have an idea.
80
00:07:21,920 --> 00:07:25,840
I would like try and push the limits and boundaries of what these things could do.
81
00:07:25,840 --> 00:07:31,760
And then I would kind of get something that I couldn't have done before. And that whenever
82
00:07:31,760 --> 00:07:39,280
that happens for me, I think I kind of go all in. I'm like, oh, like this is, this was never possible
83
00:07:39,360 --> 00:07:45,040
before. Like, I wonder what I can do with this. That's like a turbo boost to your excitement and
84
00:07:45,040 --> 00:07:48,960
energy on the thing. It's like, whoa, whoa, wait, that really that actually worked. What else could
85
00:07:48,960 --> 00:07:57,600
I do? Yeah. And so that's how it started. So I built my first like prototype for work, which
86
00:07:57,600 --> 00:08:04,400
it was basically our character in space as a floppy bird game. Right. And I wanted to do it
87
00:08:04,480 --> 00:08:10,640
as an Easter egg on our, on our website, because on site now, if you triple click our background,
88
00:08:10,640 --> 00:08:15,280
you send Astro into space and I was like, well, what if when you're in space, you could play this
89
00:08:15,280 --> 00:08:21,760
game and then you could, but you know, I was like, I built this whole thing. Like I didn't get any
90
00:08:21,760 --> 00:08:27,760
help. I just spent, you know, an hour or two with chat GBT and it was copy paste, copy paste. This
91
00:08:27,760 --> 00:08:34,320
didn't work. And eventually I got a, a game that like, it worked, looked pretty cool. Yeah,
92
00:08:34,320 --> 00:08:41,680
it worked. And I was like, wow. Okay. Yeah. But like, what else can I do? And then I think I like
93
00:08:41,680 --> 00:08:49,280
coded a, it was like this idea of a, like a 3d trail of like trailhead, if you like zoomed out
94
00:08:49,920 --> 00:08:54,560
and you had different like places you could go, like a video game and you would click it and had
95
00:08:54,560 --> 00:08:58,720
all the associated like trails and learning. I love that as a dimension for the first kind of
96
00:08:58,720 --> 00:09:02,240
stuff you're working on because like, first of all, it's playful. Like you're making games. It's,
97
00:09:02,240 --> 00:09:05,840
it's, it's exciting and fun and like kind of low stakes. Like if this game doesn't work,
98
00:09:05,840 --> 00:09:10,720
like nobody loses their job, nobody dies. Like it's not just not a huge deal if it doesn't come
99
00:09:10,720 --> 00:09:16,240
out right. But then it's also like kind of pulling you in. I recently, so there's a book that came
100
00:09:16,240 --> 00:09:25,200
out in October by, by Jean Kim and Steve Yeggy called Vibe Coding. And their audience with this
101
00:09:25,200 --> 00:09:28,320
book is like professional engineers. And they're trying to kind of bring them down the road of
102
00:09:28,320 --> 00:09:32,800
being like, Hey, you should try this AI assisted engineering vibe coding thing. But one of the
103
00:09:32,800 --> 00:09:37,440
first things in their book is they get you out there on cloud code or one of these things,
104
00:09:37,440 --> 00:09:42,000
and they have you build a Flappy Bird game. Like that is their prompt is like, Hey, build a Flappy
105
00:09:42,000 --> 00:09:44,960
Bird game. And I was reading this book and I was like doing my best to play along, even though I
106
00:09:44,960 --> 00:09:48,560
kind of do this a lot. I was like, I want to see how, how their example works. And I made a little
107
00:09:48,560 --> 00:09:54,160
Flappy Bird game. And I couldn't get past level one. First of all, I was, I was like, this game is
108
00:09:54,160 --> 00:09:57,920
too hard. And then I went, you know, I went to, you know, back to the prompt and I was like, this game
109
00:09:57,920 --> 00:10:01,360
is too hard. And it was like, well, how about I slow it down for you? And I eventually like dialed
110
00:10:01,360 --> 00:10:06,720
it into where the little Flappy Bird was okay. But, but like, there is a magic to that kind of
111
00:10:06,720 --> 00:10:12,160
experience of like, Oh, I made a game and isn't that neat? Isn't that fun? And I made it mostly
112
00:10:12,240 --> 00:10:17,680
using natural language. Like that's the big thing, right? Yeah. And you said, no, you just use the
113
00:10:17,680 --> 00:10:21,760
word mostly. And I think we should get in to talk about that, right? Because that's the, that's the
114
00:10:21,760 --> 00:10:25,760
rub, isn't it? Like if it were just natural language and it just does the magic thing,
115
00:10:25,760 --> 00:10:30,000
that's one thing. But then, you know, there's a little bit of a slippery slope there or something
116
00:10:30,000 --> 00:10:33,840
where it's like, Oh, well, sometimes you have to go a little beyond just natural language.
117
00:10:34,400 --> 00:10:39,600
Sometimes you have to go a lot beyond natural language. Yeah. Especially like early on,
118
00:10:39,600 --> 00:10:47,520
because I think I got into Cursor when Sonnet 3.5 wasn't out yet. But like shortly after I
119
00:10:47,520 --> 00:10:54,480
gone in, that model came out, which drastically, you know, upleveled me into, into what I was
120
00:10:54,480 --> 00:10:59,520
thinking. And so I started out in Cursor and I started with a Figma plugin. I was like,
121
00:10:59,520 --> 00:11:03,680
I've always wanted to build a Figma plugin. Oh, you, you were building a plugin for Figma.
122
00:11:03,680 --> 00:11:07,200
And just for folks who are not familiar with Figma, Figma is like a prototyping,
123
00:11:08,000 --> 00:11:13,120
visual prototyping program where you can drag buttons and lines and boxes and all that stuff
124
00:11:13,120 --> 00:11:18,400
together. And of course, as many, you know, modern software programs do, it has a plugin
125
00:11:18,400 --> 00:11:22,400
system where you can build little custom things that work for your own workflow or whatever. Is
126
00:11:22,400 --> 00:11:26,080
that, is that about how it works? I mean, yeah, there's, there's plugins for like everything
127
00:11:26,080 --> 00:11:31,840
imaginable. And so this was more like, let's see what I can do. Let's see if I can build a
128
00:11:31,840 --> 00:11:37,200
working plugin. And so I went with like, it's, it's called ColorGen and it'll generate
129
00:11:37,200 --> 00:11:44,880
color palettes, gradients, scales. And then I put like a fun HSL game in the style of like old Mac
130
00:11:44,880 --> 00:11:50,160
or old windows. And then you can add those and they'll add them as frames into Figma.
131
00:11:50,160 --> 00:11:52,320
So wait, you built a game in your Figma plugin?
132
00:11:53,280 --> 00:12:01,680
Yeah. Yeah. Cause I just like to create and I like to explore and have fun mostly.
133
00:12:02,000 --> 00:12:07,200
It's like, yeah, this is so fun. And so after that, I was like, why I just built a Figma
134
00:12:07,200 --> 00:12:11,680
plugin. So I could probably build a whole app. And that's kind of how.
135
00:12:13,120 --> 00:12:17,840
And, and thus the, the, the, uh, the hubris of success, right. Turns into,
136
00:12:19,440 --> 00:12:25,680
yeah, took it a little extreme. Um, cause I was like, I don't know why not. But again, I,
137
00:12:25,680 --> 00:12:31,040
I was like, I'm always about trying to push these things to their boundaries to like,
138
00:12:31,760 --> 00:12:36,880
understand what's possible, what's not possible, what they're good at, what's, what's not, which,
139
00:12:36,880 --> 00:12:42,160
you know, obviously if you're in this space that allows you to look ahead and kind of be able to
140
00:12:42,160 --> 00:12:47,840
say, oh, okay, like sure. In two years, I think this definitely will be possible. Like, and so
141
00:12:47,840 --> 00:12:54,960
going back to like the mostly natural language, this app that I built, which I do have the working
142
00:12:54,960 --> 00:13:00,960
app. I like, I have authentication. I have a backend database, uh, all kinds of stuff.
143
00:13:01,520 --> 00:13:07,760
And I, it is not mostly natural language. It is a lot of domain expertise that goes into it,
144
00:13:07,760 --> 00:13:15,920
proper product development and a lot of research and learning. And mostly like you have to think
145
00:13:15,920 --> 00:13:22,080
too, there's so many programming languages. There's so many building blocks and there's so
146
00:13:22,080 --> 00:13:29,040
many tech stacks. It's kind of bewildering, right? Yeah. And so, yeah, we're talking hours.
147
00:13:29,040 --> 00:13:32,960
So, so, so let me, let me, let me make sure I'm following you here. So basically you were like,
148
00:13:32,960 --> 00:13:36,320
okay, I had these early successes. It was really fun. Let's build something real.
149
00:13:36,320 --> 00:13:42,480
And then you do, you, you kind of, you, you dove into it and then all of a sudden a raid before
150
00:13:42,480 --> 00:13:48,080
you was just the vast, endless complexity of the tech landscape of like, should this be in Python
151
00:13:48,080 --> 00:13:52,240
or Java? Should this be on, you know, web based or should it be this thing or what kind of database
152
00:13:52,240 --> 00:13:56,640
do you use? What even is a database, et cetera, just basically all right in your face right away.
153
00:13:57,200 --> 00:14:02,240
Oh yeah. And like you get it wrong, not saying you can't fix it, but for a non-technical person,
154
00:14:02,240 --> 00:14:06,080
you don't, I don't really want to get it wrong from the beginning. Right. Cause you, you,
155
00:14:06,080 --> 00:14:09,520
you don't even know what the downsides of getting it wrong would be. Like if you're like,
156
00:14:10,080 --> 00:14:14,400
Hey, you know, like, so any, any programmers listening to this will understand the joke
157
00:14:14,400 --> 00:14:17,120
I'm making here. If you're like, Hey, database, I want to use a database. Like, okay, we'll use
158
00:14:17,120 --> 00:14:22,640
Fox pro. So Fox pro is like this butt of jokes database from, you know, the nineties, uh,
159
00:14:22,640 --> 00:14:26,080
it was a Microsoft thing that like no one would use cause it has a whole bunch of problems with
160
00:14:26,080 --> 00:14:30,400
it. Right. But you don't know that. Right. And, and someone who's like, you know, not a technology
161
00:14:30,400 --> 00:14:34,240
kind of career person wouldn't know that, but that's the fear you're like, what if I pick the
162
00:14:34,240 --> 00:14:38,320
Fox pro? What if I pick the thing that every programmer would deride and laugh at as a dumb
163
00:14:38,320 --> 00:14:46,720
choice? How would I even know? Yeah, exactly. And so a lot of my research went into that.
164
00:14:47,120 --> 00:14:54,560
And not only that, but I was taking big bets on these companies. Right. And so I would look at
165
00:14:54,560 --> 00:15:00,160
how much staff they have, how much funding they got. Like, that's partially why I stuck with
166
00:15:00,160 --> 00:15:06,960
cursor and never fully jumped to windsurf or never fully switched to rep wet. Right. Like I put my
167
00:15:06,960 --> 00:15:12,800
bet on cursor. I have like a huge design crush on their like head of design. And I was like,
168
00:15:12,800 --> 00:15:17,680
I don't think they're going away. And based off of like their funding and like how quickly they're
169
00:15:17,680 --> 00:15:23,600
shipping features and how fast I'm seeing this product, I'm using advanced like week after week,
170
00:15:24,320 --> 00:15:29,760
I'm going to bet my chips on this one, which I don't think was a bad idea. I still use other
171
00:15:29,760 --> 00:15:36,560
ones in collaboration, but the same thing with like Vercel huge deep dive on them. I was in the
172
00:15:36,560 --> 00:15:42,160
beta for like their V zero, which was, they were kind of one of the first to come to like
173
00:15:42,160 --> 00:15:46,800
generative UI and generative components, which for me, that was something I was interested in.
174
00:15:47,520 --> 00:15:54,240
Super base was a big one. And so I'm good. Like it was between Firebase and super base and like,
175
00:15:54,240 --> 00:16:00,240
and I went super base because for me, the UI was easier to understand. I thought they were cool.
176
00:16:00,240 --> 00:16:03,920
They were open source. And so, yeah, it's interesting the way that you talk about this,
177
00:16:03,920 --> 00:16:07,360
because you kind of came to this and we'll talk about what your app is in just a minute,
178
00:16:07,360 --> 00:16:13,040
but you kind of came to this with, you know, an underlying desire to build a thing and sort of
179
00:16:13,040 --> 00:16:16,880
kind of kit and caboodle bundled in with that was a need to make a bunch of these decisions,
180
00:16:16,880 --> 00:16:20,960
right? Where, and you, I think we're kind of coming at a lot of these decisions from a,
181
00:16:20,960 --> 00:16:24,640
you know, comparatively speaking to the most of the population, pretty informed perspective of
182
00:16:24,640 --> 00:16:27,920
like, you even knew to ask what is the funding of this company and like, are they going to be
183
00:16:27,920 --> 00:16:32,320
around in three years and that kind of thing. But then even there, like the sense I'm getting
184
00:16:32,320 --> 00:16:35,600
from the way you're describing it is you were kind of just sort of like groping around a little
185
00:16:35,600 --> 00:16:40,720
bit in the dark, trying to find pieces that would both work and that gave you the right feel as far
186
00:16:40,720 --> 00:16:45,200
as like, it's going to keep working some something along those lines. Yeah. Yeah. Oh, and another
187
00:16:45,200 --> 00:16:51,760
thing I looked for was companies that had really, really, really good documentation. Yeah. Which is
188
00:16:51,760 --> 00:16:58,080
not most of them. Which is not most of them. And these like, I call them like AI first companies,
189
00:16:58,080 --> 00:17:02,960
like that, that stack, because if you go to a lot of their like documentation websites,
190
00:17:02,960 --> 00:17:06,960
you can literally copy the whole thing and mark down. So they're, they're very much forward
191
00:17:06,960 --> 00:17:13,440
thinking of like, we are writing this documentation for you, but also for LLM, which is big for me,
192
00:17:13,440 --> 00:17:20,080
because I can't rely on the documentation by myself. I would have no idea. And just to give
193
00:17:20,080 --> 00:17:25,600
you an example of like how I started, which if, yeah, if you're an engineer, this will probably
194
00:17:25,600 --> 00:17:29,680
be really funny. It's, but it's just like, I didn't know I never worked with this stuff. And so
195
00:17:30,320 --> 00:17:38,720
at first I was using Docker and like auth.js for authentication. And I found this guy on YouTube
196
00:17:39,440 --> 00:17:45,920
and I can't tell you how many hours I spent watching his YouTube video of everything he was
197
00:17:45,920 --> 00:17:52,160
typing and then pausing it and then typing it. And it wasn't until like, I don't even know
198
00:17:52,800 --> 00:17:58,320
how I realized that he just linked his GitHub repository in the comments.
199
00:17:58,320 --> 00:18:00,720
And you didn't have to do any of that. You could have just gone over there and copy and
200
00:18:00,720 --> 00:18:07,680
pasted it. Yeah. I didn't have to do any of it, which also I wasn't version controlling anything.
201
00:18:07,680 --> 00:18:12,640
I didn't really understand GitHub. I, yeah. So there's a lot of like,
202
00:18:13,280 --> 00:18:17,440
well, what's, I mean, I love the way that you're telling the story. And because like, you know,
203
00:18:17,440 --> 00:18:20,720
now of course you have some hindsight, you figured out some of those things and why they matter,
204
00:18:20,720 --> 00:18:25,120
but actually like the path that you were taking, there's like literally nothing wrong with that
205
00:18:25,120 --> 00:18:28,560
path because you were doing it out of excitement. You were doing it out of exploration. Like that's
206
00:18:28,560 --> 00:18:33,680
how we all learn stuff. Right. And if it's, if it's wrong, if it's wrong to do that, I don't
207
00:18:33,680 --> 00:18:38,480
want to be right. Like that's how you are supposed to go into some new area. It was with curiosity
208
00:18:38,480 --> 00:18:43,120
and trying things. Right. So I think that's like, it's super delightful. Like before we go any
209
00:18:43,120 --> 00:18:47,360
further, tell me about this app. What is the app that you actually like wanted to build and why
210
00:18:47,360 --> 00:18:52,800
did you want to build it? Yeah, it's so I called it little chef, little chef. Okay. Yeah. And
211
00:18:53,440 --> 00:18:59,600
we love to cook. I love to cook and early, early on we like the, this whole thing worked with this
212
00:18:59,600 --> 00:19:05,760
holistic doctor. The point being we have now set like our dietary restrictions and like what we
213
00:19:05,760 --> 00:19:12,240
buy and what we cook with as a, as a, as a family. Yeah. So early, early on I was using chat GPT for
214
00:19:13,120 --> 00:19:20,960
recipes or again, dyslexia. Like you've been to a recipe website and I understand the ads
215
00:19:20,960 --> 00:19:26,640
and I understand the preamble, but like I could not follow, like I could not
216
00:19:28,160 --> 00:19:32,240
recipe websites. Very harsh. I have noticed the same thing. It's like, oh,
217
00:19:32,240 --> 00:19:35,360
oh guys, you're making this so much harder than it needs to be. Just tell me what to cook.
218
00:19:36,720 --> 00:19:41,120
Right. And then I'd have to figure out like the conversions or the substitutes or
219
00:19:41,840 --> 00:19:47,840
whatever. Right. And like, I love cooking from cookbooks, but, but again, even for me there
220
00:19:47,840 --> 00:19:53,680
would be times where I have to like cover things sometimes with paper or I'll accidentally jump
221
00:19:53,680 --> 00:20:00,240
or skip and then I'll miss like a step. Oh, I mean, I, I am not dyslexic and I do exactly the
222
00:20:00,240 --> 00:20:03,040
same thing. Like I'm cooking, I'll be cooking mac and cheese and it's like, oh, you were supposed
223
00:20:03,040 --> 00:20:06,640
to do this thing before that thing. And your, your, I just jumped down the page and you did the wrong
224
00:20:06,640 --> 00:20:12,960
thing at the wrong time. Yeah. Shoot. Yeah. We were using chat GPT and I made a custom GPT.
225
00:20:12,960 --> 00:20:18,080
I went a little extreme. So obviously in the custom instructions, I laid out all of our
226
00:20:18,080 --> 00:20:24,000
preferences, which was like dietary restrictions, preferences of what we like. I even uploaded like
227
00:20:24,000 --> 00:20:28,400
an Excel sheet of our kitchen. I'm like, so I'm like, don't tell me to cook things with the
228
00:20:28,400 --> 00:20:32,720
utensils that I don't have just in case people aren't familiar. So a custom GPT, if people
229
00:20:32,720 --> 00:20:37,920
haven't heard of that, it's a feature in chat GPT and other language models have the same feature
230
00:20:37,920 --> 00:20:42,080
where you can essentially kind of package together a whole bunch of stuff that you,
231
00:20:42,080 --> 00:20:46,640
so you don't have to paste it in or type it in every time you make a request, you put it into
232
00:20:46,640 --> 00:20:50,880
kind of this special place. And then when you make requests, all of that stuff is already in there.
233
00:20:50,880 --> 00:20:55,680
Is that, that's the basic idea behind a custom GPT? Yeah. And I think chat GPT was probably the
234
00:20:55,680 --> 00:21:02,960
first to maybe launch that if, you know, Google user, you got gems, you know, and now at this
235
00:21:02,960 --> 00:21:08,720
point, like perplexity in projects and anthropic projects. So yeah, it's really just like, I don't
236
00:21:08,720 --> 00:21:14,960
want to have to repeat myself. Just give me a recipe that works off like immediately and stop
237
00:21:14,960 --> 00:21:18,880
putting fricking gluten in it. Cause I can't eat that or whatever it is. Yeah. Yeah. Like one of
238
00:21:18,880 --> 00:21:24,480
ours is soy, right. And even knowing like I started with like common pantry staples,
239
00:21:24,480 --> 00:21:29,920
cause I like to bake. So it's like, typically we have all of these things on hand. So just know
240
00:21:29,920 --> 00:21:35,760
this. Like, I don't have to tell you. And like I said, down to like, it knows the type of oven
241
00:21:35,760 --> 00:21:40,480
that we have. It knows we have a gas stove top. It knows we have a dual oven. It knows I have a
242
00:21:40,480 --> 00:21:47,600
kitchen aid mixer. It knows I have a gourmet pasta attachment and the uni oven. And so a lot of our
243
00:21:47,600 --> 00:21:54,160
recipes all of a sudden started coming from chat GPT and you know, they were pretty good.
244
00:21:54,240 --> 00:21:57,440
Like it wasn't, it wasn't making dubious food. It was actually like, Hey, that recipe came out
245
00:21:57,440 --> 00:22:01,520
pretty good. And, and, and some people, some people might at this moment be thinking like,
246
00:22:01,520 --> 00:22:05,760
how would an AI be making good recipes? Like it doesn't cook. It doesn't taste. How does it
247
00:22:05,760 --> 00:22:09,600
doing that? And in case, in case you audience member are asking that question, the answer to
248
00:22:09,600 --> 00:22:13,600
that is large language models are basically a zip file of the whole internet. And so it's just sort
249
00:22:13,600 --> 00:22:18,080
of an average of all the things that all the people have said. And on average, if you take
250
00:22:18,080 --> 00:22:21,840
the average of most recipes, you get a pretty good recipe, right? Like this recipe might say
251
00:22:21,840 --> 00:22:25,200
you cook the salmon at 425 and this one might say you cook it at 400 and it's going to pick
252
00:22:25,200 --> 00:22:30,240
one of those two and then you're good. And it's going to be fine. Right. And so you should expect
253
00:22:30,240 --> 00:22:34,240
that dumping the requirements into an AI would yield a good recipe. Like that is a normal thing
254
00:22:34,240 --> 00:22:40,080
to expect. Probably one of the things it's like better at, right? Cause yeah, right. I just like,
255
00:22:40,080 --> 00:22:44,960
don't think there's thousands and hundreds of thousands of bad recipes on the internet.
256
00:22:44,960 --> 00:22:47,680
Right. You know what I mean? You wouldn't publish it on the internet if it was really bad, right?
257
00:22:48,480 --> 00:22:56,160
Yeah. So in that case, yeah, to your point, this was like an area where I had a little more trust
258
00:22:56,160 --> 00:23:03,280
in the LLM, right? Like telling me there would be times though, and there were recipes we'd cook
259
00:23:03,280 --> 00:23:09,920
where like, something's not right. Like this didn't go right. Like, I don't know what went wrong.
260
00:23:09,920 --> 00:23:13,440
Well, and which to be fair, that happens without AI too.
261
00:23:14,400 --> 00:23:21,200
Yeah. And the other thing that happened, which I told people was like, it made me a better cook.
262
00:23:21,200 --> 00:23:27,280
Like I was learning the way I was interacting with it too, because I wouldn't just get it with
263
00:23:27,280 --> 00:23:32,880
the recipe. I was then taking that recipe and like going into voice mode when I was cooking.
264
00:23:32,880 --> 00:23:38,240
And basically that allowed me to say, okay, I'm doing something I've never done before.
265
00:23:38,960 --> 00:23:44,320
How do I know when the oil is ready to fry the eggplant parmesan? And it's like, well,
266
00:23:44,320 --> 00:23:47,440
you do the wooden spoon test. And I'm like, I don't know what you're talking about.
267
00:23:48,160 --> 00:23:53,440
And it was just like small things like that. I'm like, oh, okay, sure. And I do it and it worked.
268
00:23:53,440 --> 00:23:58,720
And so I learned a lot and I realized I was like learning as I was cooking.
269
00:23:58,720 --> 00:23:59,200
Yeah.
270
00:23:59,200 --> 00:24:05,520
And it's because I'm able to ask questions in the moment to ensure that I'm going on the right track
271
00:24:05,520 --> 00:24:10,480
or if something I didn't think seemed right, how do I pivot or whatever.
272
00:24:10,480 --> 00:24:15,200
That feels really exciting to me in the sense that like, for me, the times I get most excited
273
00:24:15,200 --> 00:24:19,760
about AI are not when it just does it for me. It's when it makes me better.
274
00:24:19,760 --> 00:24:20,240
Yeah.
275
00:24:20,240 --> 00:24:23,440
And like, and so this is what you're, what's happening when you're doing that is like,
276
00:24:23,440 --> 00:24:27,680
oh, it's taking the collective wisdom of the internet and the world about cooking
277
00:24:27,680 --> 00:24:32,000
and turning it into a little chef coach for you. That's being like, oh, okay. You've never
278
00:24:32,000 --> 00:24:34,720
folded a souffle before or whatever. I don't know if it folds souffles, but
279
00:24:36,880 --> 00:24:38,240
I'm not a great cook myself.
280
00:24:38,240 --> 00:24:39,600
But you got to ask little chef.
281
00:24:39,600 --> 00:24:43,600
Yeah. Little chef. Can you fold a souffle? You can, but you shouldn't. So
282
00:24:46,000 --> 00:24:48,320
I think that's really exciting. That's really cool.
283
00:24:48,320 --> 00:24:55,600
Yeah. And so I basically took that because the problem was, you know, if you use something like
284
00:24:55,600 --> 00:25:02,400
chat GPT, your conversations can live in a project, but it was still hard for me to like
285
00:25:02,400 --> 00:25:08,960
find. And again, hard for me to read and whatever, or save or have a database of my recipes.
286
00:25:08,960 --> 00:25:14,480
And so basically I was like, I want that and I want it in my own app. And so that's where
287
00:25:15,280 --> 00:25:22,400
little chef came from was that whole experience. So in the app, you can basically do everything I
288
00:25:22,400 --> 00:25:28,480
described. You can describe what you want to eat. You'll go into a conversation with little chef.
289
00:25:28,480 --> 00:25:34,880
You can preview the recipe, save that recipe, and then it puts it in like a really nice layout.
290
00:25:35,680 --> 00:25:41,840
I have an image generator. Sometimes those images are kind of funny. Mostly they're good.
291
00:25:43,920 --> 00:25:48,640
But the most exciting piece for me, which is the one that I'm currently trying to finish
292
00:25:48,640 --> 00:25:53,920
the feature right now is cook mode for the dyslexia. So it's not anything crazy,
293
00:25:53,920 --> 00:25:59,280
but if you go into cook mode on little chef, typically cook mode on a website is a focus state,
294
00:25:59,280 --> 00:26:05,840
except in my focus state, it's the step with the ingredients, but you have little chef here.
295
00:26:07,360 --> 00:26:14,880
So you have, you can chat. And then I'm basically trying to add in voice, so go hands free.
296
00:26:15,600 --> 00:26:21,920
And that's been super fun because now when I'm cooking, I've been testing it out and it'll read
297
00:26:21,920 --> 00:26:27,280
the step to me out loud or I could say, oh, sorry, how much flour did I need in this step?
298
00:26:27,280 --> 00:26:31,920
And then I can say, can you set a 10 minute timer for me? I'm ready to move on to step two.
299
00:26:31,920 --> 00:26:37,520
It'll literally like navigate me through the UI. It'll set timers for me. I could exit cook mode.
300
00:26:37,520 --> 00:26:42,000
And now I'm trying to bring in the learning aspect, which I think is most fun of like,
301
00:26:42,800 --> 00:26:47,600
okay, I'm at this step, but like, what's the best way to do it? And it could tell me,
302
00:26:47,600 --> 00:26:51,280
or it can go fetch, you know, a video from YouTube and then play it.
303
00:26:51,280 --> 00:26:54,880
Oh, nice. If it's some complicated technique or like you need to julienne the carrots or
304
00:26:54,880 --> 00:26:58,400
whatever it is, then it's like, I don't know what that means. Will you show me? And then poof,
305
00:26:58,400 --> 00:27:02,560
there's a little video that shows you. Yeah. And so, I mean, I do have an MVP.
306
00:27:02,560 --> 00:27:07,200
Obviously I'm still, it's hard for me because like, oh, people are like, you need to launch this.
307
00:27:07,200 --> 00:27:11,200
I think I have like 30 users in the app. Oh, wow. So you have like, this is, this has gone
308
00:27:11,200 --> 00:27:14,560
beyond just you using it. You've actually opened up a few other people. Well, yes,
309
00:27:14,560 --> 00:27:19,200
it's opened also, right? Because you have to think about when you're doing these fun things,
310
00:27:19,200 --> 00:27:25,760
you're also funding them. And so, right. You're paying for the server space. You're paying for
311
00:27:25,760 --> 00:27:31,040
the AI tokens, et cetera. Right. My research was like, what is the cheapest way to do this?
312
00:27:31,040 --> 00:27:37,040
And I will tell you that I got to my MVP basically on the $20 a month cursor plan,
313
00:27:37,040 --> 00:27:44,080
because a part of that research was understanding the free tiers and what came with all of the free
314
00:27:44,080 --> 00:27:48,560
stuff and all of these things. And I will say like, right now it's at a point where I am getting
315
00:27:48,560 --> 00:27:55,440
like blocked and like rate limited and super base, but I'm reaching like approaches on Versailles.
316
00:27:55,440 --> 00:28:01,920
And so even now I'm like, oh my God, it's a decision of like, do I ditch Versailles,
317
00:28:01,920 --> 00:28:08,080
which is amazing. It's like one quick deployment. It's so easy to use. The interface is so nice,
318
00:28:08,080 --> 00:28:13,600
but you're going from like zero to either zero to 30 or 50 bucks a month, like that.
319
00:28:13,600 --> 00:28:16,800
Right. At which point then you have to start asking questions like, okay, well, if I had
320
00:28:16,800 --> 00:28:23,440
a hundred users, either I would go bankrupt or I would be now needing to charge them money.
321
00:28:23,440 --> 00:28:27,280
But then how much do I charge? Oh, people aren't going to pay a hundred bucks a month for a little
322
00:28:27,280 --> 00:28:32,880
chef, but would they pay $3 a month or et cetera? All those questions come flooding in all of a
323
00:28:32,880 --> 00:28:38,160
sudden. Oh yeah. And you know, I was like, haven't done it because again, I struggle with like,
324
00:28:38,160 --> 00:28:42,880
am I going and trying to be a solo crooner or like, am I still trying to like have fun?
325
00:28:42,880 --> 00:28:48,480
And I have this product that like, I can allow people to use, but we use it every single week.
326
00:28:48,480 --> 00:28:54,720
And you know, I think I just looked and I have almost 400 recipes that I've created. I built
327
00:28:54,720 --> 00:29:00,240
that little scraper. So if we do recipes, I can pull those in. It'll, you know, keep all the steps,
328
00:29:00,240 --> 00:29:04,720
pull the image, pull the title. I link back, you know, to the original recipe and then
329
00:29:04,720 --> 00:29:11,440
little chef enhances it with like tips and storage and, you know, notes. And then, yeah,
330
00:29:11,440 --> 00:29:16,560
I have like a whole backlog of features and an easy one will be a remix. So if you come in,
331
00:29:16,560 --> 00:29:22,080
you could publish your recipes. So there's like a public place you can go to view all the public
332
00:29:22,080 --> 00:29:30,080
ones. But if you like one, but you don't eat meat or you're vegan, but you like the recipe,
333
00:29:30,080 --> 00:29:35,760
you could just remix it and it'll remix it based off of your culinary profile. So that again,
334
00:29:35,760 --> 00:29:42,000
you don't have to say, it'll just ground it in your preferences and then remix that for you.
335
00:29:42,000 --> 00:29:46,480
That is really cool. I feel so excited about this. I want to try this myself. And I promise I will
336
00:29:46,480 --> 00:29:52,160
put in my $3 a month for you. But the like, just that notion of like, you're taking all these
337
00:29:52,160 --> 00:29:56,080
things that like are, you know, in this particular case, it's totally possible for the vegan to take
338
00:29:56,080 --> 00:29:59,680
this, you know, this pork recipe and go, I'm going to make a vegan version of it and I got to change
339
00:29:59,680 --> 00:30:04,800
all these different things. And that's a whole bunch of work for them. That is, you know, I got
340
00:30:04,800 --> 00:30:10,560
to, you know, spend five minutes doing this thing and converting this. And like, that's not fun
341
00:30:10,560 --> 00:30:14,800
work. And it's not useful work. Just that's a great thing for an AI to just do. And especially
342
00:30:14,800 --> 00:30:19,920
if you already did the work of like telling it what you want, one button later, now you have a
343
00:30:19,920 --> 00:30:25,680
remixed, you know, pork recipe that is totally vegan and you're good to go. That sounds amazing.
344
00:30:25,680 --> 00:30:30,800
Yeah. Well, and so I, in the app, it's the culinary profile, which is basically everything
345
00:30:30,800 --> 00:30:36,240
I put in that custom GPT. So when you log in, again, all these things you have to fix. Also,
346
00:30:36,240 --> 00:30:43,040
it's like, wow, web is so complicated. I was like, should I just like off the bat on like
347
00:30:43,680 --> 00:30:48,800
native app? Anyways, it's and, and, and so, so that's, that's, I mean, that itself is a really
348
00:30:48,800 --> 00:30:52,240
interesting question that you have a ton of background on in that you're a professional
349
00:30:52,240 --> 00:30:56,080
software designer, but other people might not realize like, Hey, there's, you know, you could
350
00:30:56,080 --> 00:31:00,080
do things on a webpage, kind of like as if you went to google.com and you could do really complicated
351
00:31:00,080 --> 00:31:03,520
stuff in there. And, you know, people who have used Gmail probably know like you do that in a
352
00:31:03,520 --> 00:31:07,760
web browser and yet it's this complicated app with like different screens and you can scroll
353
00:31:07,760 --> 00:31:11,360
and select things. And it's like quite impressive what it could do right inside a web browser
354
00:31:11,360 --> 00:31:16,800
versus native, which is like, Hey, this is an Android app or an iOS app or a Mac app, something
355
00:31:16,800 --> 00:31:21,760
like that, where it uses all those features, but it's not running inside a web browser. That's a
356
00:31:21,760 --> 00:31:26,960
tough choice to be able to make. And, and at this point, both of those options are really complicated.
357
00:31:26,960 --> 00:31:31,760
Like what is possible? You could do really complex stuff in both of those. I know that. Yeah. And
358
00:31:31,760 --> 00:31:38,240
now like, you know, adding, I'm like, Oh my God, I have to test Safari Firefox from, and now I'm
359
00:31:38,240 --> 00:31:48,080
testing in like Dia and Arc and Atlas and Comet. And I'm like, Oh, yeah. Well, it's funny what
360
00:31:48,080 --> 00:31:52,800
you're talking about here is like an interesting step that you have made. So, you know, you made
361
00:31:52,800 --> 00:31:57,600
this thing, you started using it, you had fun, you started building features. And then there's this
362
00:31:57,600 --> 00:32:02,880
like big step of like sharing with others, because as soon as you share it with others, as soon as
363
00:32:02,880 --> 00:32:06,800
other people start to use it, you're going to get someone going like, Hey man, I use opera browser
364
00:32:06,800 --> 00:32:10,800
and this feature you have doesn't work. And now that's your problem. Now you got to go like,
365
00:32:10,800 --> 00:32:15,840
okay, I guess I'll download opera browser and figure out why this thing doesn't work. And kind
366
00:32:15,840 --> 00:32:21,840
of a little bit in the old days, like pre AI days, that was real hairy because you got to download
367
00:32:21,840 --> 00:32:25,520
the thing. And then you're going to dive in there and be like, Oh, the, the, the, the, you know,
368
00:32:25,520 --> 00:32:29,680
the document object model or Dom in the web browser on opera has this one little tiny thing
369
00:32:29,680 --> 00:32:33,680
that's slightly different from the standard. And they say it's better. And so they're not going to
370
00:32:33,680 --> 00:32:37,920
change it. But now I have to do some complicated thing and I got to figure out how to do that.
371
00:32:39,040 --> 00:32:44,160
And now, you know, it's a little easier now to just go, Hey cursor, uh, this thing isn't working
372
00:32:44,160 --> 00:32:48,960
in, you know, in, in opera and it'll go, Oh yeah, I see that, you know, here's this thing. They do
373
00:32:48,960 --> 00:32:52,400
this differently. Let me do it in a way that works for both. Yeah. There's just so many,
374
00:32:52,400 --> 00:32:58,240
and especially if you're one person, right? Like this is such an issue when you have a team and
375
00:32:58,240 --> 00:33:04,960
like also part of my hesitancy to open it up, especially to paying customers is it's a very AI
376
00:33:06,240 --> 00:33:17,760
first focused app. And it's like, I'm, uh, kind of teetering on the line of health, right? Um,
377
00:33:17,760 --> 00:33:23,280
I think there was a guy who like, I just remember like almost died or something because he followed,
378
00:33:24,160 --> 00:33:27,920
uh, no, no, no, no, I do remember seeing something like this, followed an AI recipe
379
00:33:27,920 --> 00:33:31,120
and then like got food poisoning or something like this, right? Yeah. Yeah. And I'm like,
380
00:33:31,120 --> 00:33:37,760
I can't be like for a little chef, like going off the rails. I don't know. You know what I mean?
381
00:33:37,760 --> 00:33:44,960
Like, and, and AI is so fun, but especially in like software and doing something like that,
382
00:33:44,960 --> 00:33:50,800
that is so scary because you just can't, you can't know what people are going to put into it. And
383
00:33:50,800 --> 00:33:55,200
you hope that the guard rails are there, but I had a friend who tested it and he was like,
384
00:33:55,360 --> 00:34:01,440
can I, how do I cook pizza with lava? And it gave him a recipe. And I was like, what?
385
00:34:02,000 --> 00:34:09,040
But then I Googled it and people cook pizza with lava in like certain states.
386
00:34:11,040 --> 00:34:16,640
Wow. Well, okay. Things I learned today, uh, include that, but, uh, but yeah, right. Then,
387
00:34:16,640 --> 00:34:22,080
then it absolutely is a Pandora's box as soon as you open it, because then you get somebody who
388
00:34:22,240 --> 00:34:25,600
comes along and they're like, how do I cook something? How do I cook meth? Can you make me
389
00:34:25,600 --> 00:34:30,240
a meth, you know, recipe in your app? And it happily makes a meth recipe. And you're like,
390
00:34:30,240 --> 00:34:34,800
Oh wait, yeah, right. That's illegal. And I shouldn't have, right. So whose problem is that?
391
00:34:34,800 --> 00:34:38,640
And so forth, you know, I have this mental model and tell me if this, if this seems right to you,
392
00:34:38,640 --> 00:34:44,880
where there's sort of like three distinct levels or maybe four distinct levels of like how serious
393
00:34:44,880 --> 00:34:49,520
you are with like making a piece of software. And the first one is like, I just have a thing I need
394
00:34:50,160 --> 00:34:57,440
to do right now. And I need a one-off single use script that just does that thing. Like I have these
395
00:34:57,440 --> 00:35:00,560
data over here and I need to move it over there. And it's going to be a big hassle. Can you write
396
00:35:00,560 --> 00:35:04,320
a thing that does that one off? You throw it away in the garbage can when you're done, you had a
397
00:35:04,320 --> 00:35:08,480
nice day. The second level is like, I would think, and this is kind of where you were for a long time
398
00:35:08,480 --> 00:35:12,240
with this app. I have a thing where it's like, I want to do this regularly. I want to cook for my
399
00:35:12,240 --> 00:35:15,680
family. I want to use this. I want to have this thing that I did with a custom GPT, like in an
400
00:35:15,680 --> 00:35:21,280
app, I want cook mode. I want these things just for me. And that level, that's the level that
401
00:35:21,280 --> 00:35:25,760
you've been at kind of with, with a little chef for a long time. And that's a, that's really fun.
402
00:35:25,760 --> 00:35:29,120
That's a like super fun place to be in because you don't have to worry about any of those things.
403
00:35:29,120 --> 00:35:33,920
You're the only user or you and your family or you and your friends or whatever. And so, but then
404
00:35:33,920 --> 00:35:38,320
there's that level of beyond that, where you open it up to others and then you're kind of beholden
405
00:35:38,320 --> 00:35:43,280
to them and you have to, you know, worry about them. You have to fix things for them. And then
406
00:35:43,280 --> 00:35:46,000
maybe the fourth level is then they start paying you money and then it's real serious, right?
407
00:35:46,640 --> 00:35:54,960
It's real serious. I'm like, I'm not trying to spend weekends fixing bugs. And I get it. And
408
00:35:54,960 --> 00:36:01,040
to your point, like that's the decision, right? Of like, I mean, if maybe all of a sudden overnight,
409
00:36:01,040 --> 00:36:07,680
I got like 200, 300,000 users and it's like, Oh my gosh, I'd have to like hire a team. So that
410
00:36:07,680 --> 00:36:14,640
piece, I'm like, this is still fun. It's like when you are a designer or an artist and then you,
411
00:36:14,640 --> 00:36:20,240
you know, you start taking commissions or you just start doing things that like you didn't want to
412
00:36:20,240 --> 00:36:25,520
do. And again, I'm not in a space. Like I like my job. I love Salesforce. Like I love Trailhead.
413
00:36:26,400 --> 00:36:30,640
I'm not trying to, you're not trying to be an entrepreneur. Like that's not your point.
414
00:36:30,640 --> 00:36:38,560
Yeah, that's not, is it appealing? Yes. Am I, am I, cause I'm also plugged into like the AI
415
00:36:39,120 --> 00:36:43,840
creative community. And now I'm really plugged into like the vibe coding community, specifically
416
00:36:43,840 --> 00:36:49,440
with product designers. And I'm seeing some who that's what they did. They had the idea,
417
00:36:49,440 --> 00:36:54,480
they made the app and now they've quit their job and, or they're still just doing it for fun,
418
00:36:54,480 --> 00:37:00,560
but they make like such a simple, productive product with like a one-time payment. There's
419
00:37:00,560 --> 00:37:06,640
not like complexity of AI involved or whatever. And so, yeah, I think once you, you know, add
420
00:37:07,280 --> 00:37:14,240
super rich AI features where I just think in general, we're still figuring a lot of stuff out.
421
00:37:14,240 --> 00:37:18,880
Oh, heck yeah. Well, and that's, that's, I mean, that's another interesting distinction that I
422
00:37:18,880 --> 00:37:23,760
think like some people that I've, I've talked to, when we, when I, when I start talking about
423
00:37:23,760 --> 00:37:30,640
making software with AI help, they assume all the way to like, Oh, the software uses AI to do cool
424
00:37:30,640 --> 00:37:36,160
stuff. And I'm like, Oh no, no, those are separate questions. Like you can just build Flappy Bird,
425
00:37:36,960 --> 00:37:42,800
you know, and put it out on the app store and you're done. Flappy Bird doesn't need AI. Flappy Bird
426
00:37:42,800 --> 00:37:48,160
just a regular piece of software, or you can build Little Chef and Little Chef. And actually,
427
00:37:48,160 --> 00:37:53,920
you know, I feel like Little Chef, even without the AI part would actually be pretty cool. Just
428
00:37:53,920 --> 00:37:57,920
the pieces that are like, you can save recipes and have cook mode, like even without the AI
429
00:37:57,920 --> 00:38:01,520
parts, it would probably still be a neat app. And I bet there's a few out there. Yeah. Of course,
430
00:38:01,520 --> 00:38:06,240
what you've then done is you've gone and put AI in it in a way that makes it like, Oh, really quite
431
00:38:06,240 --> 00:38:11,520
magical as an app experience. And then it's like, but I guess I just want to say like,
432
00:38:12,080 --> 00:38:16,080
if you're hearing this and thinking like, Oh, I have to make AI apps. No, no, no.
433
00:38:16,640 --> 00:38:21,120
You can make software apps. And to your point, like that might actually be like a better first
434
00:38:21,120 --> 00:38:26,160
step as far as an app, because then you can build this thing and it doesn't, it's not quite so risky.
435
00:38:26,960 --> 00:38:32,640
Yeah. Like a few that stick out, I called them, they're called Threnz. That's what they're called
436
00:38:32,640 --> 00:38:38,080
on threads. If you make friends, Threnz. Threnz. Okay. That's a new one. One of the designers I
437
00:38:38,080 --> 00:38:44,480
follow, I think he works at YouTube, but he loves hiking. And so he launched the iOS app called
438
00:38:44,480 --> 00:38:51,920
Spru like T H R U. And it lets you see how far you've hiked certain famous trails in a year,
439
00:38:51,920 --> 00:38:56,320
like how far you've gone. Like, like I'm 10% of the way through the Appalachian trail or whatever.
440
00:38:56,320 --> 00:39:02,960
Yeah. And it's like, wow, so cool. And it's like fun. And there's like animation and right,
441
00:39:02,960 --> 00:39:07,920
there's no AI, but like, again, that was a simple one. I think like
442
00:39:09,040 --> 00:39:15,280
prior to this technology just like, wasn't possible for designers like us. Right. And
443
00:39:15,280 --> 00:39:21,920
then another one is Bobbin. I think it's the first person I've seen to use the Threads API
444
00:39:21,920 --> 00:39:27,200
in a really cool way. And what a cool name for an app that does that. Of course, if anyone,
445
00:39:27,200 --> 00:39:30,240
if you're familiar with sewing, I don't know what a Bobbin is, but I know it's part of a sewing
446
00:39:30,240 --> 00:39:36,560
machine. Yeah. And it's like all your analytics, you know, you log in and you have this beautiful
447
00:39:36,560 --> 00:39:41,200
dash and it's like, cool. And, and the other thing is like, obviously I follow these people
448
00:39:41,200 --> 00:39:48,880
and literally every week, like the rate at which they're able to launch new features
449
00:39:49,440 --> 00:39:54,560
is just like incredible. It's bonkers. I I'm, I'm experiencing it myself with a lot of the things
450
00:39:54,560 --> 00:39:58,400
that I'm working on, but you know, something you just said, you know, in this designer community
451
00:39:58,400 --> 00:40:04,480
of Threads, I, I realized that something that you particularly are bringing to this, you know,
452
00:40:04,480 --> 00:40:09,680
arena here is you're bringing the curiosity and the ability to build stuff with AI. Sure, sure.
453
00:40:09,680 --> 00:40:12,480
Right. Maybe that's the thing that's common across all of us, but you're also bringing
454
00:40:12,480 --> 00:40:15,920
something really different, which is you are a professional designer. You have this thing
455
00:40:15,920 --> 00:40:20,800
that is like, you're really good at, and you've spent decades getting good at, but now when you
456
00:40:20,800 --> 00:40:25,120
bring that in, like I, you know, I'll, I'll, I'll try to, I'll put, put some pictures of little chef,
457
00:40:25,120 --> 00:40:29,280
you know, in the, in the show notes for this episode so people can see it. But it's like,
458
00:40:29,280 --> 00:40:34,160
it's gorgeous. I mean, it is like, when I look at it, I'm like, whoa, this is like a real person
459
00:40:34,160 --> 00:40:37,600
made like this, who knew what they were doing. And I look at the apps that I make and I'm like,
460
00:40:37,600 --> 00:40:41,920
this looks like the Power Rangers, you know, it's like, it's not, it's not that same level of, of,
461
00:40:41,920 --> 00:40:48,480
of polish and quality and, and, and intent and layout and design and golden mean ratio and all,
462
00:40:48,480 --> 00:40:54,800
you know, the stuff that designers know. And, and so like that Venn diagram that you have, right,
463
00:40:54,800 --> 00:41:00,240
where you're like, oh, I have, you know, I have these two things that I'm doing. You're, you and
464
00:41:00,240 --> 00:41:04,480
these other designers are building these things that like part of their value isn't just that they
465
00:41:04,480 --> 00:41:10,320
work because AI made them work. Part of the value is a human who cared, made this thing beautiful
466
00:41:10,320 --> 00:41:15,600
and made it usable and made it obvious how to, how to use it. And that like, it doesn't have to be
467
00:41:15,600 --> 00:41:19,840
designed, but people who are coming to this from other places are going to bring whatever their
468
00:41:19,840 --> 00:41:24,240
Venn diagram is. If I'm a scientist, I'm going to bring in like a really rigorous understanding
469
00:41:24,240 --> 00:41:27,360
of this scientific discipline into what I'm doing. If I'm a lawyer, I'm going to bring this
470
00:41:27,360 --> 00:41:31,440
really, you know, my perspective as a lawyer, whatever that is. And I think that's a really
471
00:41:31,440 --> 00:41:36,480
cool thing. The, the huge diversity of what we're going to get here from a software perspective is,
472
00:41:37,200 --> 00:41:40,080
is, you know, we're going to get apps like little chef that are just like, oh, wow,
473
00:41:40,080 --> 00:41:45,120
just drop dead gorgeous, you know? I mean, that's a good point too, because I think I wrote like a
474
00:41:45,120 --> 00:41:53,520
blog article about like, oh, AI has an aesthetic, like, like all of a sudden I'm seeing all these
475
00:41:53,520 --> 00:41:58,400
landing pages and like, and I was like, okay, well, I'm pretty sure that was made in lovable,
476
00:41:58,400 --> 00:42:05,440
pretty sure that was like bold. And there's just like this aesthetic and style that is fine. It's
477
00:42:05,440 --> 00:42:12,640
not like bad actually in general, like if the entire internet, especially think like government
478
00:42:12,640 --> 00:42:17,440
websites or something were to even just get to that level, right? That would be a delightful.
479
00:42:18,400 --> 00:42:24,000
But to your point, yeah, like, okay, cool. AI probably, you know, helped you with this. And
480
00:42:24,000 --> 00:42:30,640
so that's where I think people are saying, like, to your point, aesthetic and taste and like really
481
00:42:30,640 --> 00:42:37,920
thoughtful user experience and user flows that AI is just not great at, like.
482
00:42:37,920 --> 00:42:40,640
Well, or the way I would say it is a little different. I would say it's not necessarily
483
00:42:40,640 --> 00:42:45,760
that AI is not not great at it per se. It's that by default, like what comes out if you don't know
484
00:42:45,760 --> 00:42:51,520
what to ask for is this median thing. Whereas what you're saying is like, well, but I know what to
485
00:42:51,520 --> 00:42:56,880
ask for in this domain. Like I know to say like, hey, those rounded corners with the shadow, like
486
00:42:56,880 --> 00:43:01,120
that's very 1990s. Don't do that. Do it this way. And so whatever that whatever those things that
487
00:43:01,120 --> 00:43:08,480
you're able to say are you bring the taste to plot the AI then through some particular part of
488
00:43:08,480 --> 00:43:13,520
latent space that actually does have the design aesthetic that you're talking about and can
489
00:43:13,520 --> 00:43:19,200
reproduce it and do it. Yes, and it can too. Like I'm at a point where, you know, you have skills
490
00:43:19,200 --> 00:43:25,440
now. You have, you know, your rules and commands and you can get a lot of that stuff in markdown
491
00:43:25,440 --> 00:43:32,320
files for it to follow in a design system manner. I guess, though, I'm also thinking of, like,
492
00:43:32,320 --> 00:43:37,920
the user experience and a lot of things. If I'm working on a new feature, it depends what it is.
493
00:43:38,000 --> 00:43:43,920
Sometimes I'll start in Figma. Sometimes I won't. And usually I'll just let the AI take the first
494
00:43:43,920 --> 00:43:52,480
pass at that user experience. And I'd say like 90% of the time, I'm like, wow, like, why?
495
00:43:55,280 --> 00:43:59,280
Interesting choice. Which kind of makes sense, right? You think like I talked to engineers and
496
00:43:59,280 --> 00:44:04,480
it's like, wow, it's like having like, a little better than like a junior level engineer helping
497
00:44:04,480 --> 00:44:10,720
and it's same thing with design, right? Junior level designer for UX. Cool. It's just like,
498
00:44:10,720 --> 00:44:17,360
to your point, like I have the years of experience of like crafting user experience of like how that
499
00:44:17,360 --> 00:44:22,320
should work. And obviously that does, you know, make a big difference. I mean, the other thing
500
00:44:22,320 --> 00:44:27,680
that that triggers for me is that in this space of kind of the level two apps, you know, not the
501
00:44:27,680 --> 00:44:32,240
one-off scripts, but the thing that's still just for you, you have this incredible space to iterate
502
00:44:32,240 --> 00:44:36,480
and try things. And then you have your own feedback mechanisms. And I'm sure you've probably
503
00:44:36,480 --> 00:44:40,480
encountered a bunch of these things while making Little Chef where you were like, okay, I made it
504
00:44:40,480 --> 00:44:45,120
this way, but then you know what? I tried it and it's like, oh, that really shouldn't be on the
505
00:44:45,120 --> 00:44:48,720
left there because I'm always looking on the right side of the screen at that moment. And I should
506
00:44:48,720 --> 00:44:52,560
just try it on the other side. But because it's just for you and your family and whatever, you can
507
00:44:52,560 --> 00:44:56,880
just experiment. You can just iterate really fast in a way that once you open it to other people,
508
00:44:56,880 --> 00:45:00,160
it becomes hard because people are like, hey, why'd you move the thing? I liked it on the left.
509
00:45:00,160 --> 00:45:08,400
Yeah. And that's also part of what's fun too. I'll go from maybe 15 to 30 iterations
510
00:45:08,400 --> 00:45:14,080
of what it should look like, right? Or how it reacts. And that part's fun. You can do that
511
00:45:14,080 --> 00:45:19,040
in Figma, but you end up with like, and it's like, I'm able to do the thing. I'm able to see it. I'm
512
00:45:19,040 --> 00:45:25,280
able to test it live. I'm able to go back. And then even in Cursor, I think they're just a really
513
00:45:25,280 --> 00:45:31,040
forward thinking company. And so they've launched all these new features. One of them being the,
514
00:45:31,040 --> 00:45:38,160
obviously the web browser inside of Cursor with the very detailed editing capabilities, right?
515
00:45:38,160 --> 00:45:45,360
So you can actually just now pick all the elements that you want and do all of the CSS UI changes
516
00:45:45,360 --> 00:45:50,640
right there and see it immediately. Which is a huge advance. If you've tried doing it without,
517
00:45:50,640 --> 00:45:54,160
if you're trying to do that thing without it, there's a lot of back and forth with the AI.
518
00:45:54,160 --> 00:45:57,600
We're like, no, no, I mean like a little more rounded. No, no, like two pixels up,
519
00:45:57,600 --> 00:46:04,000
two more pixels up. Okay. Right. Yeah. And so again, as it advances, and I think it's all
520
00:46:04,560 --> 00:46:10,320
credit to the models, right? Obviously user interface makes a big difference. Like those
521
00:46:10,320 --> 00:46:16,160
features with Cursor or plan mode and that type of stuff, but it's the models. Yeah. What model
522
00:46:16,160 --> 00:46:22,880
do you usually use with Cursor? It depends. I switch depending on what I'm doing. Opus 4.5,
523
00:46:23,760 --> 00:46:29,520
usually if it's like a big thing, I'll do that to implement. Sometimes I'll go between like
524
00:46:30,960 --> 00:46:38,720
Codex 5.2 or Gemini 3 Pro for planning, but often this is also just because I just have like a
525
00:46:38,720 --> 00:46:45,280
common layer of mistrust for all of them. I'll have one do the plan and then I'll take that plan
526
00:46:45,280 --> 00:46:51,520
and run it through another one and be like, what do you think of this plan? Or like throw it into
527
00:46:51,520 --> 00:46:57,040
perplexity and like do a deep dive. Cause like, I'm not a hundred percent sure. What a fascinating
528
00:46:57,040 --> 00:47:00,480
way. I mean, what you just said, I absolutely love, which is like this, this kind of common
529
00:47:00,480 --> 00:47:04,800
layer of mistrust that is absolutely the right way to hold these things. No matter how good they
530
00:47:04,800 --> 00:47:11,680
get, like they, you can't trust them. You shouldn't trust them. That's not a dig. That's not saying
531
00:47:11,680 --> 00:47:14,800
they're not good and not useful. Obviously they're letting you and me and other people do these
532
00:47:14,800 --> 00:47:20,560
incredible things we couldn't have done six months ago. And yet the failure mode, the common failure
533
00:47:20,560 --> 00:47:24,000
mode is you just go like, I don't know, I just put it in there and whatever it came out, I was like,
534
00:47:24,000 --> 00:47:29,360
I guess that's fine. As opposed to like, hmm, maybe by default I should be like, no, I bet it's not
535
00:47:29,360 --> 00:47:34,480
fine. Um, let me see how it's not fine. Right. And then once I've had a few viewpoints on it, like,
536
00:47:34,480 --> 00:47:39,840
Hey, Gemini said this, but then Opus said this, right. Then your critical thinking engages as a
537
00:47:39,840 --> 00:47:45,520
human and you're like, okay, I see where that came from over there, but that objection is a really
538
00:47:45,520 --> 00:47:50,160
good point. Let's roll that into what we're asking for. I think posted this the other day,
539
00:47:50,400 --> 00:47:54,880
I was like, if you're using AI and not using any of your critical thinking skills, then I think
540
00:47:54,880 --> 00:48:00,240
you're using AI wrong. Yes. Yeah. Which is, I mean, that's the danger, right? Like the danger for us
541
00:48:00,240 --> 00:48:06,800
as, uh, you know, society is like, uh, you know, you get so used to using the GPS that you forget
542
00:48:06,800 --> 00:48:11,440
how to navigate without a GPS and then, you know, extend that out into the thinking about anything
543
00:48:11,440 --> 00:48:15,920
world. You get so used to just pressing a button, like, you know, a monkey with a smartphone and all
544
00:48:15,920 --> 00:48:22,080
of a sudden you have this app and like, no, I'm sorry. That's, that's slop. Uh, we are here.
545
00:48:22,080 --> 00:48:25,760
We're part of this because we are the critical thinkers and we are the wanters. We're the ones
546
00:48:25,760 --> 00:48:31,440
who want some cool thing. And in particular, Jacqueline, I think the, the, the thing that
547
00:48:31,440 --> 00:48:35,120
like you come into this with, when you started at the beginning of the conversation, we were like,
548
00:48:35,120 --> 00:48:40,320
I came into this because I wanted to play and I was curious and I want, I like building stuff like,
549
00:48:40,320 --> 00:48:44,240
yes, yes, yes. That's the right way to approach this. So I guess I would say to anybody listening,
550
00:48:44,240 --> 00:48:50,480
like if you feel that way, even a tiny bit, like use the bellows of AI to stoke that flame
551
00:48:50,480 --> 00:48:54,800
rather than put it out. Right. That's the, that's the, the, the, the real thing here.
552
00:48:54,800 --> 00:49:01,840
Yeah. Just have fun. Like if you're like, just getting into this in your life, I have this idea
553
00:49:01,840 --> 00:49:07,360
and I'm going to make this app and this is going to make money. And, and it's like, cool. But like,
554
00:49:07,360 --> 00:49:12,400
you can't throw product development life cycle out the window. Right. Did you do your research?
555
00:49:12,400 --> 00:49:16,800
What's your PRD? Do you have your user stories and your job stories? How are you going to chunk
556
00:49:16,800 --> 00:49:22,640
this out? Like what's your backend? And that automatically becomes like a little overwhelming
557
00:49:22,640 --> 00:49:27,040
versus to your point, if you get in and you're like, I want to make a Flappy Bird game in the
558
00:49:27,040 --> 00:49:35,200
style of Star Wars, or I want to make a website that recommends movies to me based off whatever.
559
00:49:36,480 --> 00:49:41,760
This is the big lie at the heart of modern software development. Something like,
560
00:49:41,760 --> 00:49:45,520
you know what you want and then you make what you want. And that's just not how the world works.
561
00:49:45,520 --> 00:49:49,760
Right. The way things actually work is you make a thing and then you realize whether it was what
562
00:49:49,760 --> 00:49:54,080
you wanted or not. And in what way it differs from what you wanted or what you could have.
563
00:49:54,080 --> 00:49:57,200
Because now that you've seen this, you're like, oh, what would be even cooler is this other thing.
564
00:49:58,000 --> 00:50:03,600
That's real software development, like PRDs and waterfall and all of these ways of chunking things
565
00:50:03,600 --> 00:50:08,480
down and making user stories and estimates like that is almost the antithesis of how to best make
566
00:50:08,480 --> 00:50:14,320
software because of this facet. And so the play oriented approach of just like, I'm just playing,
567
00:50:14,320 --> 00:50:19,920
I'm trying things, I'm iterating, I'm seeing what I like, I'm seeing what feels good. Once I feel it,
568
00:50:19,920 --> 00:50:24,800
then I feel, yes, that's the right thing. Man, how much better software we're all going to get if we
569
00:50:24,800 --> 00:50:29,920
could do more of that. Well, and there's also like, depending on where you're coming in,
570
00:50:30,720 --> 00:50:37,600
in that journey, I kind of also have to remember that I've been, you know, talking to LLMs
571
00:50:37,600 --> 00:50:44,320
consistently now for whatever, three, four years, right? And so there's also that of like,
572
00:50:45,760 --> 00:50:49,200
you just have fun, but like, you're going to have to like,
573
00:50:50,480 --> 00:50:54,480
almost learn how to work with these things. Yeah, you got to get the hang of it. Because
574
00:50:54,480 --> 00:51:00,320
yeah, they're all different. All the models are different and they're rapidly advancing.
575
00:51:00,320 --> 00:51:04,800
And that's right. They're different, different last week to this week, let alone from model to model.
576
00:51:04,800 --> 00:51:11,840
Yeah. And like, there's still things and skill sets to be able to achieve those types of like,
577
00:51:11,840 --> 00:51:15,680
high level things. Yeah. And I think there's, I think there's actually like, also you mentioned
578
00:51:15,680 --> 00:51:19,200
earlier, just some basic software engineering stuff, where it's like, you started and you
579
00:51:19,200 --> 00:51:22,800
weren't using source control, because why would you use source control? What even is it? Who cares?
580
00:51:23,840 --> 00:51:27,520
And, you know, most software engineers would hear that and cringe maybe and go like, of course,
581
00:51:27,520 --> 00:51:31,680
you got to use, you know, source control. But it's not obvious if you're not in that profession,
582
00:51:31,680 --> 00:51:35,040
why it matters. Someone has to tell you that where you have to have a personal experience,
583
00:51:35,040 --> 00:51:37,920
where you lose some important thing and then realize, oh, I wish I had had that in source
584
00:51:37,920 --> 00:51:44,240
control. Right. Or like, I mean, I know it's important, but like, GitHub seems like overwhelming.
585
00:51:44,880 --> 00:51:52,000
I don't know. Like, I don't know these commands, like in the terminal, whatever, like, get commit
586
00:51:52,000 --> 00:51:58,240
or new branch, like all the things I had to learn. Right. Well, and I will say just as speaking as a
587
00:51:58,240 --> 00:52:02,080
professional software engineer, we don't understand that either. Git is a complete mystery to
588
00:52:02,080 --> 00:52:06,640
everybody. But that, like, but that said, like, that's just one among many. How do you use a
589
00:52:06,640 --> 00:52:10,640
database? What is it? Why do I, why would I care about what there's so many of these things that
590
00:52:10,640 --> 00:52:15,600
you kind of explored your way into? And there's like, I was about to say there's no shortcuts.
591
00:52:15,600 --> 00:52:19,520
I think maybe there are shortcuts in the sense of like, maybe just a good explanation of some
592
00:52:19,520 --> 00:52:22,880
of these things of why you would need them would probably have been helpful at the right moment
593
00:52:22,960 --> 00:52:29,360
for you. But that's no substitute at all for just trying it and learning firsthand what it is that
594
00:52:29,360 --> 00:52:34,960
you, why you care about these pieces. Yeah. I'm like, what's SQL? Right. That's what I had to
595
00:52:34,960 --> 00:52:40,800
learn. What's RLS? Row Level Security Policies? What? You got to write those if you have like a
596
00:52:40,800 --> 00:52:46,480
database. Yeah. There's the, and there's, and the rabbit hole goes really deep, which is in a sense
597
00:52:46,480 --> 00:52:50,080
why it's a career, like why it's a profession. Software engineering is a whole profession.
598
00:52:50,080 --> 00:52:55,040
Um, but I mean, I think what we're experiencing, what you have personally experienced is that like,
599
00:52:55,040 --> 00:53:03,680
oh, some, maybe 40, 50, 60% of the work people do to get good at that career is now no longer needed.
600
00:53:03,680 --> 00:53:08,000
Like you just don't need to know what the particular command in Ruby is for doing this particular
601
00:53:08,000 --> 00:53:12,000
thing. You just don't need to know that anymore, right? But you still do need to know what's the
602
00:53:12,000 --> 00:53:16,000
database and why do I care? You still need to know conceptually what source control, why do I care?
603
00:53:16,640 --> 00:53:23,120
And that's the piece that I think I want to open the doors to the world to be like, hey, this whole
604
00:53:23,120 --> 00:53:27,040
thing, you thought this whole career was scary. And now actually a bunch of the scary stuff has
605
00:53:27,040 --> 00:53:31,280
gone. And what's left is actually essential and interesting. If you could just warm up to it a
606
00:53:31,280 --> 00:53:36,320
little bit with some help and explanations. Yeah. And yeah. And like I've said this about a lot of
607
00:53:36,320 --> 00:53:44,240
stuff, but I think it's probably most drastic and dramatic in software is that the barrier to entry
608
00:53:44,960 --> 00:53:53,120
for this has like hit the floor, right? Like, and you could say that with like video and
609
00:53:53,120 --> 00:53:57,760
vote, but like with, with code, like, I think this is probably one of the biggest
610
00:53:57,760 --> 00:54:03,520
ways in which I've seen that gap not disappear, but like really, really, really shrink.
611
00:54:03,520 --> 00:54:07,280
I think you're exactly right. I mean, that's why I'm doing this podcast. That's why I'm doing all of
612
00:54:07,280 --> 00:54:14,240
this stuff is literally, I like, here it is right in front of us. And, you know, I just, I want,
613
00:54:14,240 --> 00:54:19,520
I want people to get past the, the fear, the, the being scared part. And I think, you know, you were
614
00:54:19,520 --> 00:54:26,960
here, you know, early on in doing this, because my sense is you have very little fear about this or
615
00:54:26,960 --> 00:54:30,800
maybe anything else. I don't know. I don't, you know, like you were just like, well, you know,
616
00:54:30,800 --> 00:54:36,000
let's try it, right? That's that, that is fun to you. And like, yeah, that's the spirit. That's the
617
00:54:36,000 --> 00:54:40,960
spirit to get into this with Jacqueline. This has been such a fun conversation. Thank you so much
618
00:54:40,960 --> 00:54:46,720
for this deep dive in what you're doing. I'm going to put links in, you know, in the, the show notes
619
00:54:46,720 --> 00:54:52,960
to information, your blog and your blog posts and your, your friends group and, you know, anything
620
00:54:52,960 --> 00:54:58,080
you want to share so that people can, can, can check out what you've done and, you know, and can,
621
00:54:58,080 --> 00:55:03,040
can watch for the fateful moment when you, you know, leave it all behind and go pro with a little
622
00:55:03,040 --> 00:55:08,080
chef or, or, or not. Or not. We gotta be careful. You know, it can't be opening up too much. I'll
623
00:55:08,080 --> 00:55:12,720
hit up my own. That's right. That's right. Yeah. Well, thank you. Thank you so much for, for being
624
00:55:12,720 --> 00:55:19,840
here. Yeah, this is awesome. Yeah. Thanks so much. Okay. That's a wrap for today's episode. If you
625
00:55:19,840 --> 00:55:25,120
liked it, please do me two quick favors. First, think of exactly one person, you know, who would
626
00:55:25,120 --> 00:55:29,520
like this and share it with them. If you go to conjuringcode.com slash podcast, you could find
627
00:55:29,600 --> 00:55:34,160
all of the share links and all that good stuff. And second, give a quick rating or thumbs up or
628
00:55:34,160 --> 00:55:39,280
subscribe on your podcast player of choice. That really helps other people find the podcast. If
629
00:55:39,280 --> 00:55:44,160
you have comments or suggestions or questions, just swing by conjuringcode.com and say hello
630
00:55:44,160 --> 00:55:56,800
or comment on this episode. Thanks for listening and we'll see you next time.