1
00:00:00,085 --> 00:00:01,835
Leo Dion (host): Hey everybody,
I hope you're doing well.

2
00:00:01,835 --> 00:00:05,323
Thank you for joining us for
another episode of Empower Apps.

3
00:00:05,323 --> 00:00:07,538
Before we get to my interview with Jared,

4
00:00:07,538 --> 00:00:10,196
just wanted to give you an
update on a few things.

5
00:00:10,196 --> 00:00:15,149
I will be speaking at SwiftCon,
October 7th through 9th, in Berlin.

6
00:00:15,149 --> 00:00:17,758
So if you're there, let me know.

7
00:00:17,758 --> 00:00:26,182
I'll be doing a workshop on Swift automation,
so from Xcode project to fastlane.

8
00:00:26,182 --> 00:00:31,749
So check that out. I just got back
from iOSDevUK. It was amazing.

9
00:00:31,749 --> 00:00:33,697
I think it's one of my
favorite conferences.

10
00:00:33,697 --> 00:00:36,923
I did not pick up Welsh,
but I did pick up a lot

11
00:00:36,923 --> 00:00:40,518
of great tips on development
in the Apple space.

12
00:00:40,589 --> 00:00:43,627
Paul Hudson, of course, had
a great talk on marketing.

13
00:00:43,627 --> 00:00:48,241
And Daniel Steinberg had a really
great talk on Foundation Models,

14
00:00:48,241 --> 00:00:51,162
which I continue to want
to spend some time with.

15
00:00:51,162 --> 00:00:53,791
What else? Philly's
talk on sketchboarding.

16
00:00:53,791 --> 00:00:57,118
I definitely want to get into that.
And there's so many great talks.

17
00:00:57,118 --> 00:01:00,834
I highly recommend, if
they do it again, to go.

18
00:01:00,834 --> 00:01:05,822
I know it's a lot for us Americans, but
if you're in Europe, you have no excuse.

19
00:01:05,822 --> 00:01:08,001
So definitely check that conference out.

20
00:01:08,001 --> 00:01:10,583
I also want to mention
the next two episodes.

21
00:01:10,583 --> 00:01:14,163
After this one with Jared, I'll
be speaking with Stewart Lynch.

22
00:01:14,163 --> 00:01:18,485
He's coming back on to talk
about how to learn using AI,

23
00:01:18,485 --> 00:01:20,914
which I think is gonna be great.

24
00:01:20,914 --> 00:01:22,127
Definitely check that out.

25
00:01:22,127 --> 00:01:30,106
And then the one after that will be Tim
Condon on Vapor 5 and its upcoming release.

26
00:01:30,106 --> 00:01:32,035
So check that out as well.

27
00:01:32,035 --> 00:01:35,121
If you have any thoughts or things
you want to talk about, let me know.

28
00:01:35,121 --> 00:01:37,700
I'll be sure to ask them those questions.

29
00:01:37,700 --> 00:01:42,510
And then lastly, if you have not
had a chance, download the new beta.

30
00:01:42,510 --> 00:01:46,859
At least, I am trying to get
it out in time for Shipaton.

31
00:01:47,270 --> 00:01:51,561
So go ahead and go to the website and
get the new TestFlight beta there.

32
00:01:51,561 --> 00:01:53,525
I'm hoping to submit by
the end of the month.

33
00:01:53,525 --> 00:01:57,314
We'll see while I get ready
for SwiftCon. That's about it.

34
00:01:57,314 --> 00:02:01,983
If you're not a subscriber to the newsletter
or the YouTube channel or the podcast,

35
00:02:01,983 --> 00:02:04,762
definitely choose one and
take advantage of that,

36
00:02:04,762 --> 00:02:07,692
especially when these new episodes
and announcements come out.

37
00:02:07,692 --> 00:02:11,069
And then, I don't know, that's about it.

38
00:02:11,069 --> 00:02:12,949
If you're looking for a
Swift developer, reach out.

39
00:02:12,949 --> 00:02:14,731
All right, let's get on
with the rest of our show.

40
00:02:14,731 --> 00:02:18,735
Welcome to another episode of Empower
Apps. I'm your host, Leo Dion.

41
00:02:18,735 --> 00:02:22,007
Today I'm joined once
again by Jared Sorge.

42
00:02:22,007 --> 00:02:24,227
Jared, thank you so much
for coming on the show.

43
00:02:24,227 --> 00:02:26,878
Jared Sorge (guest): Thanks for having me.
It's great to be back. It's been a while.

44
00:02:26,878 --> 00:02:30,005
Leo Dion (host): It has been a while.
Our last episode was on automation.

45
00:02:30,005 --> 00:02:33,807
And I don't know if you've followed
technology in the last few years,

46
00:02:33,807 --> 00:02:36,003
but a lot has changed in that space.

47
00:02:36,003 --> 00:02:41,730
So, I know, really. Skynet
came and automated everything.

48
00:02:43,215 --> 00:02:46,668
Before we get into that, I'll let
you go ahead and introduce yourself.

49
00:02:46,668 --> 00:02:52,171
Jared Sorge (guest): Yeah, I'm Jared Sorge. I'm
a developer just outside of Seattle, Washington.

50
00:02:52,171 --> 00:02:57,399
My day job, I work for Adobe on our
iOS apps for Premiere and Firefly.

51
00:02:57,399 --> 00:03:00,961
And on the side, I have my own
business called Taphouse Software.

52
00:03:00,961 --> 00:03:03,394
I released an app called
Arborist earlier this year,

53
00:03:03,394 --> 00:03:06,835
which is a developer tool, which I'm
sure we'll talk about in a little bit.

54
00:03:06,835 --> 00:03:10,444
And actually just yesterday I released
a big update for my Lego collecting

55
00:03:10,444 --> 00:03:13,788
app called Baseplate, which I'm very
happy with how that turned out.

56
00:03:13,788 --> 00:03:17,558
Leo Dion (host): Yeah, that is
very cool. I guess we'll start.

57
00:03:17,558 --> 00:03:21,390
Before we get into Arborist
and why you built it,

58
00:03:21,390 --> 00:03:26,796
I think we should maybe do a little
explanation about work trees.

59
00:03:29,208 --> 00:03:35,253
Because in my world, work trees,
I didn't use them until AI.

60
00:03:35,253 --> 00:03:39,154
When did you start using work trees?
And then how are they useful?

61
00:03:39,154 --> 00:03:42,221
How are they especially useful
for agentic development?

62
00:03:42,221 --> 00:03:44,949
Jared Sorge (guest): We'll come back to
the agentic part, because I think that adds

63
00:03:44,949 --> 00:03:48,024
a whole other wrinkle in how
people are starting to use them.

64
00:03:48,024 --> 00:03:50,680
I first heard about them
when I worked at Lyft,

65
00:03:50,680 --> 00:03:53,559
I think it was, like
six or seven years ago.

66
00:03:53,656 --> 00:03:57,058
Because we would, you know, when
you're working on a feature,

67
00:03:57,058 --> 00:04:00,843
in our case a large monorepo, we
had a repo that had both the driver

68
00:04:00,843 --> 00:04:04,719
app and the passenger app and
all of our shared Swift packages.

69
00:04:04,719 --> 00:04:07,927
And so I'd be working on a feature
and then I'd need to fix a bug,

70
00:04:07,927 --> 00:04:09,506
or something else would come up,

71
00:04:09,506 --> 00:04:13,162
or something would divert my attention
from the thing I'm working on.

72
00:04:13,162 --> 00:04:16,928
And a lot of times the first instinct
is to have another checkout.

73
00:04:16,928 --> 00:04:19,898
So you would go git
clone and redo your repo.

74
00:04:19,898 --> 00:04:21,606
So you'd have two copies of that repo.

75
00:04:21,606 --> 00:04:25,558
So that I'm not context switching
and having to push and pop

76
00:04:25,558 --> 00:04:29,584
stashes and all that stuff, or have
so many commits that are working,

77
00:04:29,584 --> 00:04:32,360
so I can, you know, pause and
move on to something else.

78
00:04:32,360 --> 00:04:36,031
And then I heard about work trees
as a nice alternative for that,

79
00:04:36,031 --> 00:04:42,300
because you don't have to check out again,
and it shares the same .git root on your disk.

80
00:04:42,300 --> 00:04:45,229
And so it's a really
fast way to get another

81
00:04:45,229 --> 00:04:47,704
checkout of your sources up and running.

82
00:04:47,704 --> 00:04:50,981
And so I would have, you know,
the feature that I'm working on,

83
00:04:50,981 --> 00:04:55,999
or like our main branch, and I would have the bugs
that I need to work on all in separate places.

84
00:04:55,999 --> 00:05:00,456
And so it was easier for me to not have
to worry about context switching as much.

85
00:05:00,456 --> 00:05:04,057
And so now, fast forward,
you know, six years,

86
00:05:04,057 --> 00:05:09,497
and we have these AI agents, and at
work we're starting to use Copilot.

87
00:05:09,497 --> 00:05:12,786
And by default, Copilot spins up
a new work tree for everything,

88
00:05:12,786 --> 00:05:15,064
which I'm not necessarily a fan of.

89
00:05:15,260 --> 00:05:20,529
Leo Dion (host): Yeah, I
didn't either. Yeah, I know.

90
00:05:20,529 --> 00:05:25,966
As the audience probably knows, I use a
lot of Claude and I use a lot of Cursor.

91
00:05:25,966 --> 00:05:29,128
And I got into work trees
while getting into it.

92
00:05:29,128 --> 00:05:31,910
It wasn't like these
agents use work trees.

93
00:05:31,910 --> 00:05:34,170
It wasn't until later
that I was like, wait,

94
00:05:34,170 --> 00:05:37,332
if I'm gonna do parallel stuff,
this is where it's super helpful.

95
00:05:37,332 --> 00:05:43,184
And recloning is stupid to do.
So that's when I was like, okay.

96
00:05:43,795 --> 00:05:48,198
Jared Sorge (guest): Being able to have
all of your branches in one view as well.

97
00:05:48,198 --> 00:05:51,360
So I use Tower. Tower is
how my mind works with Git.

98
00:05:51,360 --> 00:05:56,242
And when you have separate clones, you can't
cherry pick from one branch to another.

99
00:05:56,242 --> 00:05:59,072
You'd have to export a patch
and then reapply the patch.

100
00:05:59,072 --> 00:06:00,818
And that sucks, right?

101
00:06:00,818 --> 00:06:05,509
But with work trees in Tower, I can
just drag from one branch to another.

102
00:06:05,509 --> 00:06:08,272
And because it has access
to all of my branches,

103
00:06:08,272 --> 00:06:11,098
I don't need to worry
about that patch exporting.

104
00:06:11,098 --> 00:06:12,672
It's a much easier workflow for me.

105
00:06:12,672 --> 00:06:13,550
Leo Dion (host): Yep.

106
00:06:13,897 --> 00:06:19,223
So what I found really useful about work
trees, and the way I've been doing it,

107
00:06:19,223 --> 00:06:23,816
is kind of creating a plan for
several GitHub issues slash features,

108
00:06:23,816 --> 00:06:27,846
and then having them work on
it in parallel, basically.

109
00:06:27,846 --> 00:06:33,182
And that's really where I was able to use
work trees and take advantage of them.

110
00:06:33,182 --> 00:06:37,309
Jared Sorge (guest): Yeah, I'm the
same. It's a similar workflow, right?

111
00:06:37,309 --> 00:06:42,104
I would have my feature that I'm working
on, and then I need to fix a bug or do

112
00:06:42,104 --> 00:06:46,839
a quick update or whatever the case may
be, and I don't have to pause entirely

113
00:06:46,839 --> 00:06:51,356
what I'm doing in the one spot to save
that state and then come back to it.

114
00:06:51,356 --> 00:06:55,187
It's really easy to switch between
things in our modern tooling.

115
00:06:55,187 --> 00:06:56,902
Leo Dion (host): So let's get into it.

116
00:06:56,902 --> 00:06:59,449
How did you come up with
the idea for Arborist?

117
00:06:59,449 --> 00:07:02,856
Jared Sorge (guest): It's something
I've had in my head for a while,

118
00:07:02,856 --> 00:07:07,156
because in various projects I've
been a fan of, first, make.

119
00:07:07,156 --> 00:07:09,534
I learned about Makefiles
when I was at Lyft,

120
00:07:09,534 --> 00:07:14,288
and then a few years later I learned about
mise, and they're both like task runners.

121
00:07:14,288 --> 00:07:19,786
Leo Dion (host): I'll put a link to my
article on that, so people can know.

122
00:07:19,786 --> 00:07:23,238
Jared Sorge (guest): Yeah. But at
Adobe, we don't use either of them.

123
00:07:23,238 --> 00:07:25,406
We've started to use make a little bit.

124
00:07:25,406 --> 00:07:27,878
But we have these scripts that we need to

125
00:07:27,878 --> 00:07:30,584
run that are in arbitrary
places in the repo.

126
00:07:30,584 --> 00:07:34,409
But when you have multiple work
trees, you can't just say, okay,

127
00:07:34,409 --> 00:07:39,086
run this script and have it know which
work tree you're in without doing a whole

128
00:07:39,086 --> 00:07:45,997
lot of git rev-parse kind of things to get to
the absolute path of the script you're running.

129
00:07:45,997 --> 00:07:48,902
So if we have like
tools/sync.py, for instance,

130
00:07:48,902 --> 00:07:51,990
I can't hard-code a shell
command to run that one,

131
00:07:51,990 --> 00:07:57,037
because if I've got multiple checkouts, it's
gonna be at any number of locations, right?

132
00:07:58,334 --> 00:08:02,343
So the idea for Arborist came
about as, one, I love Mac apps.

133
00:08:02,343 --> 00:08:03,492
Mac apps are great.

134
00:08:03,492 --> 00:08:07,434
And I wanted something where
I could see all of the work

135
00:08:07,434 --> 00:08:11,187
trees that I had listed, and
then I could combine that with

136
00:08:11,187 --> 00:08:14,749
the arbitrary commands that
we use in every work tree.

137
00:08:14,749 --> 00:08:18,297
So no matter which checkout I'm
in, I can click the run button and

138
00:08:18,297 --> 00:08:21,258
the right script is gonna run
because I know where it's running.

139
00:08:21,258 --> 00:08:26,168
And then combine that with: I
use the same tools everywhere.

140
00:08:26,168 --> 00:08:31,561
I use Xcode, I use Tower.
Nova is my IDE of choice.

141
00:08:31,561 --> 00:08:35,748
I know people love VS Code. I love
Mac apps, not cross-platform whatever.

142
00:08:35,748 --> 00:08:36,612
I love Nova.

143
00:08:36,612 --> 00:08:41,281
And oftentimes I would get myself
confused, because I would have Nova

144
00:08:41,281 --> 00:08:44,082
pointed to one work tree,
Tower pointed to another,

145
00:08:44,082 --> 00:08:48,616
and then maybe Codex pointed to another,
and my workspace would get out of sync.

146
00:08:48,616 --> 00:08:53,294
And so I wanted a way to easily launch
any of those tools in the right work tree.

147
00:08:53,294 --> 00:08:54,586
And so I did that.

148
00:08:54,586 --> 00:08:59,201
At the top of the main window, you
can set up whatever tools you want.

149
00:08:59,201 --> 00:09:02,570
And then by clicking on them, that
will open them in the right work tree.

150
00:09:02,570 --> 00:09:05,340
So your stuff is always in the right spot.

151
00:09:06,383 --> 00:09:10,294
It's been really nice for me. It's an
app that I really enjoyed working on.

152
00:09:10,294 --> 00:09:13,963
I use it all the time at
Adobe on my work machine.

153
00:09:13,963 --> 00:09:18,290
And yeah, it's been a
really nice addition as

154
00:09:18,290 --> 00:09:22,792
like the command center
for all of my projects.

155
00:09:22,792 --> 00:09:25,749
Leo Dion (host): Yeah, and I
gotta say it's well designed.

156
00:09:25,749 --> 00:09:27,067
I'm really impressed with it.

157
00:09:27,067 --> 00:09:30,135
People should definitely
check it out and try it.

158
00:09:30,135 --> 00:09:33,833
And as a big fan of work trees,
and we've talked offline about

159
00:09:33,833 --> 00:09:37,928
Grove and all that stuff, and I built
my own little command to do stuff,

160
00:09:37,928 --> 00:09:41,182
and it seems like every other
developer in the world has.

161
00:09:42,928 --> 00:09:44,848
But yeah, it's really great.

162
00:09:44,848 --> 00:09:47,122
I guess we'll get into, like,

163
00:09:47,122 --> 00:09:51,050
what did you learn along the way
while you were building Arborist?

164
00:09:51,050 --> 00:09:53,192
What are some APIs that you didn't know?

165
00:09:53,192 --> 00:09:58,372
And I should amend that, because
you probably used a coding agent.

166
00:09:58,372 --> 00:10:01,017
So maybe you didn't learn and
you let the coding agent do it.

167
00:10:01,017 --> 00:10:05,438
But what did you pick up from the
work that you had done with Arborist?

168
00:10:05,438 --> 00:10:11,659
Jared Sorge (guest): I do lean heavily on
Codex. That's my coding agent of choice.

169
00:10:11,659 --> 00:10:15,849
I'd never made a Mac-assed
Mac app before, you know,

170
00:10:15,849 --> 00:10:21,516
like with a good settings window or
being able to change the Dock icon.

171
00:10:21,516 --> 00:10:24,241
I learned a little bit
about Dock tile plugins.

172
00:10:24,241 --> 00:10:27,337
Leo Dion (host): Have you built
a Mac app that's not Mac-assed?

173
00:10:27,337 --> 00:10:29,930
Like, did you build an
Electron app, or what?

174
00:10:29,945 --> 00:10:34,522
Jared Sorge (guest): So Mac-assed, to me, isn't
necessarily the technology that you use to build it.

175
00:10:34,522 --> 00:10:38,294
It's how well does it integrate with
the platform that it's running in.

176
00:10:38,294 --> 00:10:41,688
So how well does it fit in with
the overall Mac ecosystem?

177
00:10:42,497 --> 00:10:47,032
I think some people probably think
Catalyst apps couldn't be Mac-assed,

178
00:10:47,032 --> 00:10:49,735
but Baseplate has a Mac version as well.

179
00:10:49,735 --> 00:10:52,790
And I've tried to make it
fit in with the Mac rather

180
00:10:52,790 --> 00:10:55,500
than like just being a
port of the iPhone app.

181
00:10:55,500 --> 00:10:57,687
It actually has Mac paradigms to it.

182
00:10:58,950 --> 00:11:02,007
And you could make a
non-Mac-assed AppKit app, right?

183
00:11:02,007 --> 00:11:05,967
I think it's more about the attention
to detail that you put into it,

184
00:11:05,967 --> 00:11:09,206
like how much keyboard
shortcut support is there,

185
00:11:09,206 --> 00:11:12,905
how much customization is
there, and that kind of thing.

186
00:11:14,081 --> 00:11:17,134
I started the app out mostly, or entirely,

187
00:11:17,134 --> 00:11:20,097
in SwiftUI, and I backed out some of that.

188
00:11:20,097 --> 00:11:21,577
I went into AppKit more,

189
00:11:21,577 --> 00:11:25,072
especially when it came to like
sidebar and inspector views.

190
00:11:25,072 --> 00:11:31,477
So I learned more about not necessarily
like the specific APIs that I was using,

191
00:11:31,477 --> 00:11:37,987
but the thought behind when to use the
right frameworks to build the Mac app.

192
00:11:38,742 --> 00:11:41,822
So for instance, I had a feature request:

193
00:11:41,822 --> 00:11:45,058
my inspector view was
not super adjustable.

194
00:11:45,058 --> 00:11:49,380
Like in Xcode, you can make that
inspector view as wide as you want to.

195
00:11:49,380 --> 00:11:54,707
I would get crashes, because even
though my app was completely SwiftUI,

196
00:11:54,707 --> 00:11:59,740
I would get AppKit constraint
errors and runtime crashes when my

197
00:11:59,740 --> 00:12:03,504
views would get too wide with my
sidebar and my inspector view.

198
00:12:03,504 --> 00:12:07,567
And so I ended up backing most
of that out and going with an

199
00:12:07,567 --> 00:12:12,113
NSSplitView wrapper that, admittedly,
Codex helped me put together.

200
00:12:12,113 --> 00:12:16,300
But I learned more about the
interop between SwiftUI and AppKit,

201
00:12:16,300 --> 00:12:21,131
and when to use one versus the other,
and to not necessarily lean entirely

202
00:12:21,131 --> 00:12:26,606
on SwiftUI, because maybe it's not
the best fit for that exact moment.

203
00:12:26,912 --> 00:12:31,046
Leo Dion (host): Quick question, what
was your target OS for Arborist?

204
00:12:31,046 --> 00:12:31,706
Jared Sorge (guest): 15.

205
00:12:31,706 --> 00:12:36,471
Leo Dion (host): Wow,
okay. Do you think there—

206
00:12:36,471 --> 00:12:39,054
Jared Sorge (guest): There were Tahoe
holdouts, so I was holding back,

207
00:12:39,054 --> 00:12:42,144
trying to help them out to expand
the audience a little bit.

208
00:12:42,144 --> 00:12:48,655
Leo Dion (host): Okay. Do you think you would have
gained anything as far as moving over to 26 or 27?

209
00:12:48,655 --> 00:12:50,441
Jared Sorge (guest):
Well, 27 just came out,

210
00:12:50,441 --> 00:12:53,837
so I couldn't have released it
when I did if I had targeted 27.

211
00:12:53,837 --> 00:12:56,582
I don't think I lost
anything by not going.

212
00:12:56,582 --> 00:13:00,699
So I still, like, my work machine
was Tahoe, so I still tested it

213
00:13:00,699 --> 00:13:05,548
thoroughly on Tahoe and made it feel
at home there as much as I could.

214
00:13:05,548 --> 00:13:07,342
I'd built it on Golden Gate,

215
00:13:07,342 --> 00:13:11,617
'cause I had Golden Gate on my
personal machine from beta one.

216
00:13:13,111 --> 00:13:16,771
You know, I'm not usually the guy
who puts beta one on everything,

217
00:13:16,771 --> 00:13:21,081
but I kept hearing how good it was,
and it was an amazing beta season.

218
00:13:21,081 --> 00:13:25,523
I don't know if you ran any of
them, but from like day one,

219
00:13:25,523 --> 00:13:32,005
Golden Gate was better than Tahoe all
summer in most ways, which shocked me.

220
00:13:32,005 --> 00:13:33,291
I've never seen that happen before.

221
00:13:33,291 --> 00:13:33,824
Leo Dion (host): Yeah.

222
00:13:33,824 --> 00:13:38,890
So as someone who was also trying
to build a Mac-assed Mac app,

223
00:13:38,890 --> 00:13:43,933
where do you see the line where
SwiftUI just doesn't deliver?

224
00:13:43,933 --> 00:13:52,134
Jared Sorge (guest): I was happy to start with SwiftUI,
and when I saw problems, to then go into AppKit.

225
00:13:52,916 --> 00:13:56,839
So in the crashing case, I tried
a few different workarounds.

226
00:13:56,839 --> 00:14:01,153
I had this odd thing where I would
try to hide the sidebar and then

227
00:14:01,153 --> 00:14:04,451
show the inspector and then
show the sidebar again,

228
00:14:04,451 --> 00:14:09,174
and it was kind of hacky, and especially
that was for macOS 26 workarounds.

229
00:14:09,174 --> 00:14:12,347
Didn't happen on 15. It
did not happen on 27.

230
00:14:12,347 --> 00:14:16,333
So I think my general thought
is to start with SwiftUI,

231
00:14:16,333 --> 00:14:20,390
and then when I see issues,
see if there's an AppKit way

232
00:14:20,390 --> 00:14:24,251
to do it that might make the
solution a bit more elegant.

233
00:14:24,251 --> 00:14:26,418
I found that a little bit
with my sidebar as well.

234
00:14:26,418 --> 00:14:28,795
Like, I kept fighting
with selection state,

235
00:14:28,795 --> 00:14:33,406
especially as the app would go into
the background or become inactive.

236
00:14:35,937 --> 00:14:40,405
And so the sidebar is an outline
view instead of a SwiftUI list.

237
00:14:40,405 --> 00:14:44,627
SwiftUI lists seem to present some
performance issues in Arborist

238
00:14:44,627 --> 00:14:48,487
that I was surprised by, but was
happy to be able to work around.

239
00:14:48,487 --> 00:14:49,274
Leo Dion (host): Okay.

240
00:14:49,274 --> 00:14:55,202
One thing I've seen a lot on
social media is that SwiftUI

241
00:14:55,202 --> 00:14:59,499
has become less useful in
an age of agentic coding.

242
00:14:59,499 --> 00:15:05,010
And why not just have the
agent do UIKit or AppKit too?

243
00:15:05,010 --> 00:15:06,037
What do you think of that?

244
00:15:06,037 --> 00:15:10,209
Jared Sorge (guest): Everyone's experience
is different, and I think it depends.

245
00:15:10,209 --> 00:15:11,750
It depends on what your goal is.

246
00:15:11,750 --> 00:15:13,704
It depends on where your experience lies.

247
00:15:13,704 --> 00:15:17,182
I looked at all the code
that the agent was writing

248
00:15:17,182 --> 00:15:20,622
and I also had some pretty
specific instructions.

249
00:15:20,622 --> 00:15:26,628
Like, I make fairly heavy use of some
of the Point-Free libraries, not TCA,

250
00:15:26,628 --> 00:15:32,397
but I use their SQLiteData, I use
their Sharing, I use their Navigation.

251
00:15:32,397 --> 00:15:38,619
And so the instructions that I give to
the agents is to use these tools heavily.

252
00:15:38,619 --> 00:15:41,658
Don't use @AppStorage,
use @Shared, for instance.

253
00:15:42,244 --> 00:15:45,264
Put most of the state
in observable models.

254
00:15:45,264 --> 00:15:48,390
And abstracting the state and the logic to

255
00:15:48,390 --> 00:15:51,556
those models puts less
burden on the views.

256
00:15:51,556 --> 00:15:55,213
It makes it easier for me to
say, no, let's just experiment

257
00:15:55,213 --> 00:15:57,772
and see what this looks like with AppKit,

258
00:15:57,772 --> 00:16:02,568
and it doesn't have to put a bunch of state
in the SwiftUI view or the AppKit view.

259
00:16:02,568 --> 00:16:04,349
'Cause it's already in the model.

260
00:16:04,349 --> 00:16:10,536
So I think a lot of it depends on how
you're prompting and what you're after.

261
00:16:10,536 --> 00:16:14,406
And, you know, if your end goal
is more the app that you're

262
00:16:14,406 --> 00:16:18,404
getting at the other end of it,
probably how the agent builds

263
00:16:18,404 --> 00:16:22,918
it is less important than the
output that you're getting.

264
00:16:22,918 --> 00:16:27,148
I remember when SwiftUI first came
out, or even Swift first came out,

265
00:16:27,148 --> 00:16:32,042
and apps would put in their change notes,
rewrote this part of the app in Swift,

266
00:16:32,042 --> 00:16:36,922
or rebuilt this part of the UI in
SwiftUI as a release note for users.

267
00:16:36,922 --> 00:16:39,701
And something that I find
all the time is that

268
00:16:39,701 --> 00:16:42,303
users don't really care about that stuff.

269
00:16:42,303 --> 00:16:45,376
Users care about how the app
works, not how you built the app.

270
00:16:45,376 --> 00:16:48,365
And so are you building it as
an exercise as an engineer?

271
00:16:48,365 --> 00:16:49,747
I love doing that stuff.

272
00:16:49,747 --> 00:16:51,311
But that doesn't
necessarily help the product

273
00:16:51,311 --> 00:16:52,722
that you're trying to get out the door.

274
00:16:52,722 --> 00:16:56,457
And that's where I'm trying to get
myself better and actually ship stuff.

275
00:16:56,457 --> 00:16:58,508
Because I've gone a long
time without shipping stuff.

276
00:16:58,508 --> 00:17:01,055
And then this year has been
a lot of shipping stuff.

277
00:17:01,055 --> 00:17:03,104
And the shipping feels really good too.

278
00:17:03,104 --> 00:17:05,110
Leo Dion (host): Yeah. Exactly.

279
00:17:05,297 --> 00:17:08,764
God, I forgot what I was gonna
ask. I hate when that happens.

280
00:17:08,764 --> 00:17:09,495
Jared Sorge (guest): Same.

281
00:17:10,334 --> 00:17:13,157
Leo Dion (host): I know what I'm gonna ask.
I'm just trying to phrase it correctly.

282
00:17:13,157 --> 00:17:13,457
Jared Sorge (guest): Sure.

283
00:17:13,457 --> 00:17:14,972
Leo Dion (host): So
there's a couple of things.

284
00:17:14,972 --> 00:17:19,133
Agentic development makes you care
less about the code in some ways,

285
00:17:19,133 --> 00:17:20,957
because you're not writing it.

286
00:17:20,957 --> 00:17:26,459
I think, maybe not the code,
but how you get from A to B.

287
00:17:26,459 --> 00:17:29,071
I mean, I care about the code
a lot. I look at every PR.

288
00:17:29,071 --> 00:17:35,674
But as far as, okay, you
wanna use AppKit or SwiftUI?

289
00:17:35,674 --> 00:17:41,666
I don't really care. Just make one
a preference and the other one...

290
00:17:41,666 --> 00:17:46,863
I guess, first of all, did
you use NSViewRepresentable,

291
00:17:46,863 --> 00:17:49,826
or what was the top of the app?

292
00:17:49,826 --> 00:17:51,661
Was it SwiftUI?

293
00:17:51,661 --> 00:17:53,253
Jared Sorge (guest): Representable, yeah.

294
00:17:53,323 --> 00:17:56,059
Leo Dion (host): Okay, so
basically the whole app is...

295
00:17:56,059 --> 00:17:58,929
Jared Sorge (guest):
SwiftUI. SwiftUI lifecycle.

296
00:17:58,929 --> 00:18:03,554
Leo Dion (host): Yeah. And then you
made AppKit stuff work in SwiftUI.

297
00:18:03,554 --> 00:18:05,818
Jared Sorge (guest): Okay. Yep.

298
00:18:05,818 --> 00:18:09,219
Leo Dion (host): Yeah. And
that's like totally legit.

299
00:18:09,219 --> 00:18:14,360
And yeah, like you said, users aren't
gonna care what technology you used.

300
00:18:14,360 --> 00:18:18,244
It really comes down to support
and maintenance, really.

301
00:18:18,244 --> 00:18:23,137
It's like, okay, if I'm gonna
maintain this old Objective-C code,

302
00:18:23,137 --> 00:18:29,082
and it's not really supported actively,
do I wanna do Objective-C, right?

303
00:18:29,131 --> 00:18:32,019
Or do I wanna move the
Objective-C code to Swift?

304
00:18:32,019 --> 00:18:33,704
It's like that's kind
of the question, right?

305
00:18:33,704 --> 00:18:34,419
Jared Sorge (guest): Yeah.

306
00:18:34,419 --> 00:18:39,225
I mean, the way that you could phrase
stuff like that as important to users is

307
00:18:39,225 --> 00:18:44,289
stabilizing the foundations for the future
to bring you more cool stuff, you know.

308
00:18:44,289 --> 00:18:45,029
Leo Dion (host): Yeah.

309
00:18:45,029 --> 00:18:50,246
Jared Sorge (guest): It's the same technical thing,
but pitched as something that is user-impacting,

310
00:18:50,246 --> 00:18:52,789
or something they can look forward to,

311
00:18:52,789 --> 00:18:57,205
or enables things that they're
gonna get in later releases, right?

312
00:18:57,205 --> 00:19:05,287
Leo Dion (host): Yeah. I think there's a strong case
for migrating from Objective-C to Swift in 2026.

313
00:19:05,287 --> 00:19:11,074
I still think it's not super necessary to
move your stuff from UIKit to SwiftUI,

314
00:19:11,074 --> 00:19:15,075
or AppKit to SwiftUI, and
it's even less so with macOS,

315
00:19:15,075 --> 00:19:20,422
because SwiftUI is still
not fully in, I don't know.

316
00:19:22,051 --> 00:19:22,360
It is from your
experience, it sounds like.

317
00:19:22,360 --> 00:19:26,402
At least if you're gonna go with
macOS 15, it's not quite there yet.

318
00:19:26,402 --> 00:19:31,001
Jared Sorge (guest): Yeah, and I think the
migration in the UI frameworks comes down to:

319
00:19:31,001 --> 00:19:36,845
are you seeing issues that you're gonna try
to mitigate by going from one to the other?

320
00:19:36,845 --> 00:19:42,376
Like there's known problems with
large data sets in SwiftUI lists.

321
00:19:42,376 --> 00:19:48,597
So maybe the right choice there is to go
to an NSCollectionView or UICollectionView

322
00:19:48,597 --> 00:19:53,745
and have that power your collection, and
then put a SwiftUI content configuration

323
00:19:53,745 --> 00:19:58,097
as the view itself, because that's
going to be lighter weight and easier to

324
00:19:58,097 --> 00:20:01,561
maintain than all the constraints and
everything you have to do with Auto Layout.

325
00:20:01,561 --> 00:20:04,751
But it's really about how do you
pick the right tool for the job.

326
00:20:04,751 --> 00:20:10,366
At the end of the day, work trees,
Swift, SwiftUI, UIKit, AppKit,

327
00:20:10,366 --> 00:20:15,599
those are all tools, and how do
you best use the tools for the job?

328
00:20:15,599 --> 00:20:18,152
What's the best tool for the
thing you're trying to accomplish?

329
00:20:18,152 --> 00:20:21,969
Leo Dion (host): The other use case would be a
redesign, where you can assess that at redesign,

330
00:20:21,969 --> 00:20:26,487
'cause if you're gonna be doing a new UI,
you don't really care what you build it in.

331
00:20:26,487 --> 00:20:30,195
Jared Sorge (guest): Yeah, and that's just
what I did with the Lego app, Baseplate.

332
00:20:30,195 --> 00:20:36,570
It had a lot of custom pieces in it before,
and separate iPad and iPhone layouts,

333
00:20:36,570 --> 00:20:41,502
which is now a problem with
something like the iPhone Duo.

334
00:20:42,458 --> 00:20:46,907
And so even before, I wanted
to make it a bit more unified.

335
00:20:46,907 --> 00:20:50,290
And so I still have a
fair amount of the custom

336
00:20:50,290 --> 00:20:53,683
pieces to give it the
personality that it had.

337
00:20:53,683 --> 00:20:57,485
But my goal was to improve
the information architecture

338
00:20:57,485 --> 00:21:01,377
and what things are emphasized
from the jump in the app.

339
00:21:01,377 --> 00:21:06,231
But I chose, because some Lego
collections have hundreds or thousands of

340
00:21:06,231 --> 00:21:11,140
sets or pieces, and I didn't want
that to be a performance problem,

341
00:21:11,140 --> 00:21:17,202
so I chose to keep that in a collection
view for the data itself and the management

342
00:21:17,202 --> 00:21:21,840
of the cells and reuse and all that,
but the views themselves are SwiftUI.

343
00:21:21,840 --> 00:21:25,461
And that was a really good
balance for my particular app.

344
00:21:25,461 --> 00:21:27,004
Leo Dion (host): Yeah,
and it makes total sense.

345
00:21:27,004 --> 00:21:30,210
We had one question from
the audience, from Caleb.

346
00:21:30,823 --> 00:21:35,834
He asks, "I'd be curious to hear about
selling Mac apps outside the App Store.

347
00:21:35,834 --> 00:21:39,529
And while he's developing
apps such as Arborist,

348
00:21:39,529 --> 00:21:44,520
does he start out with a solid idea
on how he wants the app to be,

349
00:21:44,520 --> 00:21:49,045
or does he start out with a problem
he is having, and he builds the app

350
00:21:49,045 --> 00:21:52,943
and he uses it to flesh out ideas
and add features before shipping it?"

351
00:21:52,943 --> 00:21:54,714
Jared Sorge (guest): I'll
take the second part first.

352
00:21:54,714 --> 00:21:58,678
I started Arborist with the
problem that I wanted to solve,

353
00:21:58,678 --> 00:22:03,826
and saw that there was maybe a thread
of an idea that could become a product.

354
00:22:03,826 --> 00:22:06,387
And so my initial thought was,

355
00:22:06,387 --> 00:22:12,177
I want something that can be a command
runner across my Git work trees.

356
00:22:12,177 --> 00:22:16,253
And the thing, you know, I put it
into Codex, it gave me an app,

357
00:22:16,253 --> 00:22:18,342
and I was like, huh,
that could be something.

358
00:22:18,342 --> 00:22:23,630
And to actually help flesh the ideas
out, back in, what, May or June,

359
00:22:23,630 --> 00:22:31,635
or whenever there was Fable included in
the monthly Claude plans, I used Fable in

360
00:22:31,635 --> 00:22:38,101
Claude Design to help me build the UI and figure
out how, like, what might a good layout be.

361
00:22:38,101 --> 00:22:41,258
And Claude Design, actually, I've
been meaning to write about this.

362
00:22:41,258 --> 00:22:42,414
It's really good.

363
00:22:42,414 --> 00:22:44,242
Leo Dion (host): It's really good.

364
00:22:44,242 --> 00:22:46,130
Jared Sorge (guest): And so I'm using Cl—

365
00:22:47,349 --> 00:22:49,063
Leo Dion (host): Go ahead,
I'll let you finish.

366
00:22:49,063 --> 00:22:53,053
Jared Sorge (guest): Yeah, you can give it
prompts of things that you want to emphasize,

367
00:22:53,053 --> 00:22:58,745
or how you want it to fit in, or what platform
you're building for, and fine-tune along the way.

368
00:22:58,745 --> 00:23:01,964
And then that workflow of, okay,
prepare that for engineering,

369
00:23:01,964 --> 00:23:06,535
and it builds out all these documents or
whatever, and then I give that over to Codex.

370
00:23:06,535 --> 00:23:10,699
And Codex then helps me put the actual
code together and get the app running.

371
00:23:10,699 --> 00:23:15,180
And it's been a really nice
workflow for me as someone who

372
00:23:15,180 --> 00:23:18,968
is completely intimidated
by Sketch and Figma,

373
00:23:18,968 --> 00:23:25,246
'cause I'm not a designer, and I have
very much designer insecurities.

374
00:23:25,246 --> 00:23:26,993
I wouldn't try to do it myself.

375
00:23:26,993 --> 00:23:31,961
Leo Dion (host): Yeah, I mean, I could do
it, but I don't feel like I have experience,

376
00:23:31,961 --> 00:23:35,443
especially when it comes to, like
you said, Mac-assed Mac apps.

377
00:23:35,443 --> 00:23:39,874
But yeah, I've been using Claude Design
and Claude Code and it's been awesome.

378
00:23:39,874 --> 00:23:44,340
I should do a video on it,
honestly, and show how I did it.

379
00:23:45,627 --> 00:23:48,625
I have everything for at
least my new watch app,

380
00:23:48,625 --> 00:23:51,933
and I have everything I want
for the Bushel redesign.

381
00:23:51,933 --> 00:23:58,075
And I've been dialoguing between the
two, Claude Design and Claude Code,

382
00:23:58,075 --> 00:24:00,572
and be like, hey, take
a look at this design.

383
00:24:00,572 --> 00:24:02,529
You can export the HTML, right?

384
00:24:02,529 --> 00:24:07,766
And then it'll be like, well, based on our
specs, you need to change this and this.

385
00:24:07,766 --> 00:24:09,468
And then they go back and forth.

386
00:24:09,468 --> 00:24:12,175
And like now I'm at a good
point where I know exactly

387
00:24:12,175 --> 00:24:14,864
what to do for the next
beta for the redesign.

388
00:24:14,864 --> 00:24:16,946
Or not a redesign, but
finishing up the design.

389
00:24:16,946 --> 00:24:18,181
And it's really good.

390
00:24:18,181 --> 00:24:23,004
Jared Sorge (guest): Yeah. I love it. It's
been a wonderful addition to my workflow.

391
00:24:23,004 --> 00:24:27,771
So then, as I'm using the app, I'll
see like, this needs to be fixed

392
00:24:27,771 --> 00:24:32,872
over in this part of the project, or
wouldn't it be cool if it did this?

393
00:24:32,872 --> 00:24:36,105
Like I wanted to add drag
and drop support for repos.

394
00:24:36,410 --> 00:24:43,706
And drag and drop is somewhat tricky
sometimes, but Codex was able to nail it.

395
00:24:43,706 --> 00:24:47,824
And the initial implementation was
like you dragged it onto the sidebar.

396
00:24:47,824 --> 00:24:49,540
And I was like, no, no, no.

397
00:24:49,540 --> 00:24:53,865
I want to be able to drag a repo or a
work tree to any part of the window,

398
00:24:53,865 --> 00:24:57,647
and the app will just ingest it
and pop it right on the sidebar.

399
00:24:57,647 --> 00:25:00,497
And it did it. And it's so good.

400
00:25:00,497 --> 00:25:01,890
Leo Dion (host): It's so good.

401
00:25:01,916 --> 00:25:05,433
Jared Sorge (guest): I love the
drag and drop first-run experience,

402
00:25:05,433 --> 00:25:09,789
when you have a clean window
and you drag a work tree in.

403
00:25:09,789 --> 00:25:15,399
It crawls from the work tree to
the root .git and then finds all

404
00:25:15,399 --> 00:25:18,656
the spawned work trees, no matter
where they are on your system.

405
00:25:19,177 --> 00:25:24,508
It's really cool to see it work.
And then... Sorry, go ahead.

406
00:25:24,508 --> 00:25:25,375
Leo Dion (host): No, you go ahead.

407
00:25:25,375 --> 00:25:27,209
Jared Sorge (guest): I didn't know
if you had something that you want—

408
00:25:27,209 --> 00:25:28,642
I was gonna go back to the first part.

409
00:25:28,642 --> 00:25:29,974
Leo Dion (host): Yeah, yeah. Go ahead.

410
00:25:29,974 --> 00:25:30,649
Jared Sorge (guest): Okay.

411
00:25:30,649 --> 00:25:33,497
Then when it comes to building
for outside the Mac App Store,

412
00:25:33,497 --> 00:25:35,232
this is the first time I've done that.

413
00:25:35,232 --> 00:25:38,830
Because Arborist runs arbitrary shell
commands, it can't be sandboxed.

414
00:25:38,830 --> 00:25:41,345
So I can't send it through the app store.

415
00:25:41,345 --> 00:25:46,461
Which meant that I had to build out
the entire, like, delivery system,

416
00:25:46,461 --> 00:25:51,550
licensing, and beta channels
and Sparkle and all of that.

417
00:25:51,550 --> 00:25:56,087
And I use Cloudflare to
host all of the binaries.

418
00:25:56,087 --> 00:25:58,817
I've paid them nothing
so far, which is great.

419
00:25:58,817 --> 00:26:00,430
That's exactly what I'm after.

420
00:26:00,430 --> 00:26:05,650
But I've got a workflow that's plugged
into my fastlane infrastructure.

421
00:26:05,650 --> 00:26:09,636
So I can run my Arborist build
command, and it will build a beta,

422
00:26:09,636 --> 00:26:14,304
send it up to Cloudflare, and update
the Sparkle appcast and all that.

423
00:26:15,176 --> 00:26:19,952
And I'm trying to generalize it so
that I could do this in the future

424
00:26:19,952 --> 00:26:25,365
with other apps if I want to, but,
you know, that's all for future Jared.

425
00:26:25,370 --> 00:26:28,937
Leo Dion (host): So yeah, I use
fastlane for App Store stuff.

426
00:26:28,937 --> 00:26:35,081
I've never used fastlane for non-App
Store stuff, but I'm happy that it works

427
00:26:35,081 --> 00:26:40,337
perfectly fine with Sparkle or Cloudflare
or wherever it's uploading stuff.

428
00:26:40,337 --> 00:26:45,494
Jared Sorge (guest): Yeah, with fastlane,
it's just Ruby at the end of the day.

429
00:26:45,494 --> 00:26:46,027
So I have my own custom lanes.

430
00:26:46,027 --> 00:26:53,115
And my usual fastlane architecture
is that I have one main Fastfile,

431
00:26:53,115 --> 00:27:00,174
but that's actually a very small file, and
I have sub-Fastfiles for each of my apps.

432
00:27:00,174 --> 00:27:02,517
And then I have environment
configurations for each of my apps.

433
00:27:02,517 --> 00:27:09,654
I don't use Appfile or Matchfile or any
of those bespoke files from fastlane,

434
00:27:09,654 --> 00:27:11,690
and I use environments all the time.

435
00:27:11,690 --> 00:27:14,005
I'm a very heavy environment user.

436
00:27:14,167 --> 00:27:18,747
Leo Dion (host): And I have thoughts
why, but I want to hear your opinion.

437
00:27:18,747 --> 00:27:21,589
Jared Sorge (guest): So
I'm in my own monorepo.

438
00:27:21,589 --> 00:27:24,815
I have Arborist and Baseplate
and all my apps that

439
00:27:24,815 --> 00:27:28,166
are on the store and that
are my own internal things.

440
00:27:28,166 --> 00:27:31,497
They're all just in one spot.
So it makes it really easy.

441
00:27:33,568 --> 00:27:35,362
Leo Dion (host): Sounds like a lot. Why?

442
00:27:35,362 --> 00:27:40,512
I mean, so does the workflow run
if you make a change to any app,

443
00:27:40,512 --> 00:27:44,366
rather than just putting
it in its own repo?

444
00:27:45,219 --> 00:27:46,750
Yeah, that's what I meant, sorry.

445
00:27:46,750 --> 00:27:51,871
Jared Sorge (guest): So for CI, I use GitHub
Actions, and then I use path filters on the actions.

446
00:27:51,871 --> 00:27:55,575
So if an app's content
changes, the workflow runs.

447
00:27:55,575 --> 00:27:56,900
If it doesn't, it doesn't.

448
00:27:56,900 --> 00:27:58,326
Leo Dion (host): That's interesting.

449
00:27:58,326 --> 00:28:03,270
So your commit history for an app will
have code changes for another app?

450
00:28:03,270 --> 00:28:04,040
Jared Sorge (guest): Yes.

451
00:28:04,904 --> 00:28:07,262
The upshot is that I have one licensing

452
00:28:07,262 --> 00:28:09,969
package that handles
all my RevenueCat stuff.

453
00:28:09,969 --> 00:28:14,344
And it's used by both Arborist
and Baseplate and it works great.

454
00:28:15,833 --> 00:28:19,399
Leo Dion (host): Well, you don't
use match. Do you use match at all?

455
00:28:19,399 --> 00:28:23,658
So I have my app certificates in
their own repo that's shared.

456
00:28:23,658 --> 00:28:26,754
But I assume you put the app
certificates in that repo?

457
00:28:26,754 --> 00:28:31,724
Jared Sorge (guest): I have one repo for my signing
credentials that has the certificates and profiles.

458
00:28:31,724 --> 00:28:32,708
Leo Dion (host): Okay, cool.

459
00:28:32,708 --> 00:28:33,522
Jared Sorge (guest): Yeah.

460
00:28:33,522 --> 00:28:36,405
And when I decide to
have to add a new app,

461
00:28:36,405 --> 00:28:42,079
the process for that is actually really
straightforward to get it hooked into my CI.

462
00:28:42,079 --> 00:28:44,159
I create the new folder for the app.

463
00:28:44,159 --> 00:28:48,449
I create its environment and fill in its
variables, and I create a quick little

464
00:28:48,449 --> 00:28:53,479
build lane for that app that plugs into my
regular infrastructure, and it just works.

465
00:28:54,806 --> 00:28:55,880
Leo Dion (host): Yeah.

466
00:28:56,190 --> 00:29:04,347
I will say, I shiver a little bit at
the monorepo idea, but I will say what I

467
00:29:04,347 --> 00:29:08,891
need to do is take a lot of my BrightDigit
stuff and put it in its own repo for,

468
00:29:08,891 --> 00:29:13,376
like, a template, and then set up a
script for each other project, and, hey,

469
00:29:13,376 --> 00:29:16,536
you need to keep in sync with the new
script updates and stuff like that.

470
00:29:16,536 --> 00:29:21,278
That part I can totally see is nice about
having all your projects in one repo.

471
00:29:21,278 --> 00:29:25,040
Jared Sorge (guest): Yeah.
So I have two monorepos.

472
00:29:25,040 --> 00:29:29,547
I have one for all my app code
and one for my back-end website.

473
00:29:29,547 --> 00:29:34,209
The Taphouse website and Taphouse API
service are just a couple Vapor things.

474
00:29:34,209 --> 00:29:37,588
Those are live in separate repos.

475
00:29:37,588 --> 00:29:43,156
I may eventually make them one giant
monorepo, because the Baseplate

476
00:29:43,156 --> 00:29:47,912
API does share some code with the
app, so I'll have to figure that out.

477
00:29:47,912 --> 00:29:50,091
But it's an evolving process.

478
00:29:50,091 --> 00:29:52,764
Leo Dion (host): And
they're all just tools.

479
00:29:52,764 --> 00:29:56,854
Honestly, not that big of a deal if you
ever wanted to go the other way, right?

480
00:29:58,116 --> 00:29:59,183
Jared Sorge (guest): Yeah.

481
00:29:59,809 --> 00:30:03,821
As for the licensing part,
that was a fun little journey,

482
00:30:03,821 --> 00:30:11,193
because I didn't want to spin up a license
key service or a license validation something.

483
00:30:11,193 --> 00:30:14,867
I wanted it to be easy
for me to implement.

484
00:30:14,867 --> 00:30:19,340
And what I ended up doing was I
went with Stripe and RevenueCat.

485
00:30:19,340 --> 00:30:22,297
And Stripe, I use their
managed payments service.

486
00:30:22,297 --> 00:30:26,546
Because I don't want to be
the merchant of record on

487
00:30:26,546 --> 00:30:29,236
these things and handle like
taxes and all that stuff.

488
00:30:29,236 --> 00:30:32,483
RevenueCat, I initially was going
to use their web billing product,

489
00:30:32,483 --> 00:30:34,628
which would make me
the merchant of record.

490
00:30:34,628 --> 00:30:36,967
And I didn't know what that
term actually meant when

491
00:30:36,967 --> 00:30:39,202
I first started this, and
I ended up learning that.

492
00:30:39,202 --> 00:30:42,595
So one of the APIs I
learned was about taxes.

493
00:30:42,956 --> 00:30:43,376
Leo Dion (host): Right?

494
00:30:43,376 --> 00:30:47,118
Jared Sorge (guest): The tax code
of the world is one of the APIs.

495
00:30:47,118 --> 00:30:49,661
But then I found out about
Stripe's Managed Payments service

496
00:30:49,661 --> 00:30:52,498
and ended up using them, and
they plug right into RevenueCat.

497
00:30:52,498 --> 00:30:55,687
And RevenueCat, you give an identifier,

498
00:30:55,687 --> 00:31:01,189
and the identifier I ended up using
was the user's iCloud identifier

499
00:31:01,189 --> 00:31:06,096
for a CloudKit container that I
spun up that has no data in it.

500
00:31:06,096 --> 00:31:09,389
It'll never have data in it.
I'm only using it for licensing.

501
00:31:09,389 --> 00:31:12,035
But that's worked really well. And...

502
00:31:12,035 --> 00:31:16,571
Leo Dion (host): There's the thing in the
app where you can switch the Apple ID, right?

503
00:31:16,571 --> 00:31:18,872
If you want to purchase it
through a different Apple ID.

504
00:31:18,872 --> 00:31:19,885
Yeah.

505
00:31:20,446 --> 00:31:22,541
Jared Sorge (guest): Exactly.
So I use this myself.

506
00:31:22,541 --> 00:31:26,021
On my work machine, I don't
use my personal Apple ID.

507
00:31:26,021 --> 00:31:28,934
I have an Apple ID tied
to my work email address.

508
00:31:30,206 --> 00:31:34,726
Yeah, so I logged in on my work
laptop with my personal email,

509
00:31:34,726 --> 00:31:39,970
to just the website, to get my container
identifier, and then it checked with

510
00:31:39,970 --> 00:31:44,040
RevenueCat, said, yep, you're licensed,
and it transferred automatically.

511
00:31:44,234 --> 00:31:44,715
Leo Dion (host): Awesome.

512
00:31:44,715 --> 00:31:47,526
Jared Sorge (guest): The idea is that
I wanted to enable people to not

513
00:31:47,526 --> 00:31:50,288
just buy a license and
it's good on one machine.

514
00:31:50,288 --> 00:31:52,789
I wanted it to be transferable
to other of their machines.

515
00:31:52,789 --> 00:31:53,605
Leo Dion (host): Yeah.

516
00:31:53,991 --> 00:31:56,633
Jared Sorge (guest): And their iCloud
identifier is one that naturally

517
00:31:56,633 --> 00:31:59,258
they carry along with
them and can go anywhere.

518
00:31:59,258 --> 00:32:01,699
There's no restore button
once you're logged in.

519
00:32:01,699 --> 00:32:04,996
And if you got a matching identifier,
you're automatically licensed.

520
00:32:04,996 --> 00:32:07,393
It's really cool how fast it works.

521
00:32:07,393 --> 00:32:12,383
And it's been a really nice process,
and the RevenueCat folks are fantastic.

522
00:32:12,383 --> 00:32:13,883
Leo Dion (host): Yes, I agree.

523
00:32:13,883 --> 00:32:19,775
So my live streaming app, Heartwitch,
I ended up only doing Stripe,

524
00:32:19,775 --> 00:32:26,125
because there was no real StoreKit
for the watch at the time.

525
00:32:26,125 --> 00:32:28,880
And so there's a website for it,

526
00:32:28,880 --> 00:32:32,750
so you can then upgrade on
the website, not on the watch.

527
00:32:32,750 --> 00:32:36,124
And I ended up doing all the
Stripe stuff myself with Vapor.

528
00:32:36,124 --> 00:32:39,129
There's a Stripe library
in Vapor that you can use.

529
00:32:39,129 --> 00:32:45,606
It worked out well. This is pre-AI,
by the way, but it worked out well.

530
00:32:45,616 --> 00:32:46,895
I know.

531
00:32:46,895 --> 00:32:52,714
This was back then, like before the
pandemic killed humanity and AI took over.

532
00:32:53,905 --> 00:32:58,100
Yeah, and it was a good experience, but
going forward, when I rewrite that app

533
00:32:58,100 --> 00:33:01,647
someday, I'd totally go with RevenueCat
on it, because that's a great example

534
00:33:01,647 --> 00:33:06,153
of something where it both has a website
component and an App Store component.

535
00:33:06,153 --> 00:33:10,002
And then with Bushel, I probably
would stick with the Mac App Store,

536
00:33:10,002 --> 00:33:12,511
I think, because it's Apple-targeted.

537
00:33:18,178 --> 00:33:23,209
And then, at least since
I'm doing it for Shipaton,

538
00:33:23,209 --> 00:33:25,641
I have to put it through RevenueCat.

539
00:33:25,641 --> 00:33:30,923
So yeah, I'm working with RevenueCat
on that and it's pretty nice.

540
00:33:30,923 --> 00:33:32,794
So, yeah, I wanna talk about that.

541
00:33:32,794 --> 00:33:37,370
And this gets into another topic I wanted
to talk about: the reason you couldn't

542
00:33:37,370 --> 00:33:44,074
put it in the App Store is because it's not
sandboxed, because you're running commands.

543
00:33:44,074 --> 00:33:46,225
Is that correct? Let's talk
about that a little bit.

544
00:33:46,225 --> 00:33:50,933
What commands are you running?
And it takes user input?

545
00:33:51,278 --> 00:33:51,750
Jared Sorge (guest): Okay.

546
00:33:51,750 --> 00:33:56,664
For any Git repository, there's
a command center in the middle,

547
00:33:56,664 --> 00:34:02,205
and there's a section at the top
where your user-entered commands live.

548
00:34:02,205 --> 00:34:06,843
And you can click edit and add a
new one, and it will run whatever.

549
00:34:06,843 --> 00:34:08,282
It's like Anchorman, right?

550
00:34:08,282 --> 00:34:10,619
It'll read whatever the teleprompter says,

551
00:34:10,619 --> 00:34:14,246
and it'll run that command from
within the selected work tree.

552
00:34:14,246 --> 00:34:15,868
So there's not a limit.

553
00:34:15,868 --> 00:34:20,969
Like, you could do rm -rf /
and it would try to run that.

554
00:34:22,715 --> 00:34:25,985
So there's definitely some
danger on the user side.

555
00:34:25,985 --> 00:34:29,549
You need to be able to trust
what you're typing into the box.

556
00:34:29,549 --> 00:34:33,159
And then a feature that
I added after, in 1.2,

557
00:34:33,159 --> 00:34:37,068
was it picks up mise tasks
and make tasks as well.

558
00:34:37,068 --> 00:34:39,178
Leo Dion (host): So if you drop in a repo—

559
00:34:39,178 --> 00:34:41,790
Jared Sorge (guest): Yeah, I'm so
happy with how that turned out.

560
00:34:41,790 --> 00:34:44,017
If you drop in a repo
that has either of those,

561
00:34:44,017 --> 00:34:46,365
it'll populate it with
those tasks automatically.

562
00:34:46,365 --> 00:34:51,293
And in the 1.3 beta, I've added
support for just and npm as well.

563
00:34:51,293 --> 00:34:55,392
So it'll add those as
supported task runners too.

564
00:34:55,392 --> 00:35:03,367
Leo Dion (host): Nice. Yeah. So you recently
migrated to the Subprocess library, is that correct?

565
00:35:03,367 --> 00:35:09,718
Jared Sorge (guest): I was using it from the
start. I've built my own process runners before.

566
00:35:09,718 --> 00:35:13,686
I do some client work, and
for them we have a CLI,

567
00:35:13,686 --> 00:35:19,872
and I built out a wrapper around the Process
API that was usable but kind of clunky.

568
00:35:20,263 --> 00:35:26,491
And then I think it was a year ago
that Subprocess came out, pre-1.0.

569
00:35:27,719 --> 00:35:34,399
And so Arborist 1.0 and 1.1 used the
early pre-release versions of Subprocess,

570
00:35:34,399 --> 00:35:39,118
and then in 1.2 I migrated to
the Subprocess 1.0 package.

571
00:35:39,118 --> 00:35:41,923
Leo Dion (host): Okay. Were
there any hiccups along the way?

572
00:35:41,923 --> 00:35:46,146
Jared Sorge (guest): For the pre-1.0, I had to
have some pieces in place around cancellation.

573
00:35:46,146 --> 00:35:51,827
They really beefed up how cancellation
propagates through processes,

574
00:35:51,827 --> 00:35:53,735
and so I was able to
clean up a lot of that.

575
00:35:53,735 --> 00:35:57,153
But overall, no, it's been very smooth.

576
00:35:57,153 --> 00:35:59,264
It's been a really nice
little library to work with.

577
00:35:59,264 --> 00:36:06,179
Leo Dion (host): Yeah, I had
a library that used Process.

578
00:36:06,502 --> 00:36:08,918
And yeah, I moved to
Subprocess as soon as I could.

579
00:36:08,918 --> 00:36:11,878
I know one, and I assume this is the case,

580
00:36:11,878 --> 00:36:17,266
but one big headache was what OSes it
supports, as far as what versions.

581
00:36:18,767 --> 00:36:19,867
Did you run into any issues with that?

582
00:36:19,867 --> 00:36:25,065
Jared Sorge (guest): No. I don't know how far
back it goes, but it supports 15 at least.

583
00:36:25,065 --> 00:36:28,043
I would have gotten build
failures if it didn't.

584
00:36:28,814 --> 00:36:30,843
So that's kind of what
I'm leaning on for that.

585
00:36:30,843 --> 00:36:35,327
Leo Dion (host): So I wanna talk a little
bit more about your development workflow,

586
00:36:35,327 --> 00:36:40,295
especially with Taphouse stuff like
Arborist, and maybe we could talk about your

587
00:36:40,295 --> 00:36:46,035
development flow at Adobe, but how has that
changed with the advent of agentic coding?

588
00:36:46,035 --> 00:36:50,383
Jared Sorge (guest): I think when I'm
trying to form my prompts, you know,

589
00:36:50,383 --> 00:36:56,289
I've gotten better over the last year and
a half, because I think about the details

590
00:36:56,289 --> 00:37:02,033
that are needed for a good prompt more,
not just, hey, I want this thing, build it.

591
00:37:02,260 --> 00:37:05,508
And if you don't give it as much detail...

592
00:37:05,508 --> 00:37:10,781
I think about it more like I'm writing
a product requirement document.

593
00:37:10,781 --> 00:37:15,419
And a really thorough prompt reads
a lot like those that I've read

594
00:37:15,419 --> 00:37:19,849
over the years at Adobe or Lyft
or Zulily, where we're building

595
00:37:19,849 --> 00:37:22,799
out a feature and it's
got all these details,

596
00:37:22,799 --> 00:37:27,902
because the tool that you're typing into
doesn't have the context that's in your head.

597
00:37:27,902 --> 00:37:31,920
And by putting the context that's
in your head out into the prompt,

598
00:37:31,920 --> 00:37:37,167
or give it a Markdown document, or put it
in your own wiki and point it at the wiki

599
00:37:37,167 --> 00:37:41,850
to go read this and then ask questions,
however it is that you decide to do it,

600
00:37:41,850 --> 00:37:47,098
the more detail and instruction you give the
agent, the better the output you're gonna get.

601
00:37:47,098 --> 00:37:50,619
Not just in the code itself,
but how does the code run?

602
00:37:50,619 --> 00:37:53,861
What's the experience when you're
actually running what it outputs?

603
00:37:53,861 --> 00:37:58,745
And so, there's a new term
that we're starting to

604
00:37:58,745 --> 00:38:01,743
use at Adobe when we hire
people or bring people in.

605
00:38:01,743 --> 00:38:03,167
It's product engineer.

606
00:38:03,167 --> 00:38:09,080
And I think I'm more understanding what
that term means than when I first heard it.

607
00:38:09,080 --> 00:38:13,374
Because there's software engineer, and
that's a title that I've had before.

608
00:38:13,374 --> 00:38:16,673
I think I'm actually a
computer scientist at Adobe.

609
00:38:16,673 --> 00:38:19,937
I have no CS degree, so I don't know why
I'm a computer scientist, but whatever.

610
00:38:19,937 --> 00:38:24,191
But the idea of being a product engineer,
about thinking about it from the

611
00:38:24,191 --> 00:38:28,558
product side and not just how are we
building it, if you can blend those two

612
00:38:28,558 --> 00:38:33,403
together, you're gonna get better output
from the tools that you're working with.

613
00:38:33,403 --> 00:38:37,539
And some of that might just be, the starting
prompt is what needs all that context,

614
00:38:37,539 --> 00:38:41,594
because once you've got your long chat
history, I don't need to tell it every time,

615
00:38:41,594 --> 00:38:44,211
hey, I'm working in this app or
hey I'm working in that app.

616
00:38:44,211 --> 00:38:46,759
It knows from the context that it's in.

617
00:38:46,759 --> 00:38:51,903
But to start it and seed it with that
thorough, well-thought-out idea,

618
00:38:51,903 --> 00:38:55,743
I think gives you a lot of
benefit at the end of it.

619
00:38:55,743 --> 00:38:57,998
You're gonna get better output.

620
00:38:57,998 --> 00:38:59,672
Leo Dion (host): I like the
idea of product engineer.

621
00:38:59,672 --> 00:39:03,833
That makes a lot of sense,
because the thing is,

622
00:39:03,833 --> 00:39:08,876
you know, I was talking about
this at iOSDevUK last week.

623
00:39:08,876 --> 00:39:11,232
It's at a university, right?

624
00:39:11,232 --> 00:39:17,230
And the thing I was thinking about is, I
have a computer science degree as well,

625
00:39:17,230 --> 00:39:21,852
and computer science is really one
abstraction layer above computer engineering,

626
00:39:21,852 --> 00:39:25,621
which is like one abstraction
layer above electrical engineering.

627
00:39:25,621 --> 00:39:28,602
And it almost feels like
what we're doing is the

628
00:39:28,602 --> 00:39:32,159
next abstraction level
above computer science.

629
00:39:32,857 --> 00:39:35,772
You don't have to get into
the nitty-gritty of the code.

630
00:39:35,772 --> 00:39:40,667
It's good to know how it works,
but you need to know how to talk to

631
00:39:40,667 --> 00:39:45,631
an AI and give it the right context
in order to move forward with it.

632
00:39:46,315 --> 00:39:48,239
That's what it kind of feels
like. Does that make sense?

633
00:39:48,239 --> 00:39:49,243
Jared Sorge (guest): Absolutely.

634
00:39:49,243 --> 00:39:52,272
Leo Dion (host): Yeah, 'cause it's like, if
you're a computer scientist, you don't care.

635
00:39:52,272 --> 00:39:56,051
The great thing about Swift and
a lot of programming languages,

636
00:39:56,051 --> 00:39:59,357
unless you do C, is you don't
have to care about memory

637
00:39:59,357 --> 00:40:02,978
management or which register
you're gonna put a number in.

638
00:40:02,978 --> 00:40:05,528
You don't care. But there
are people who know.

639
00:40:05,528 --> 00:40:12,604
Jared Sorge (guest): 'Cause you're using
the unsafe APIs, or it's things where

640
00:40:12,604 --> 00:40:17,213
you're on purpose doing this
thing that could be dangerous.

641
00:40:17,213 --> 00:40:18,305
Leo Dion (host): Right.

642
00:40:18,305 --> 00:40:24,774
Yeah, and it's like, it's good to know
how the code works, but to be productive,

643
00:40:24,774 --> 00:40:30,436
you can go a lot faster by using an
AI, I guess, and it does that for you.

644
00:40:31,389 --> 00:40:35,648
Jared Sorge (guest): I go back and
forth on all of this all the time.

645
00:40:35,648 --> 00:40:41,076
It feels like AI has created a fairly
consistent existential crisis for me,

646
00:40:41,076 --> 00:40:46,290
because there are some people who are
"this was not created with AI," or

647
00:40:46,290 --> 00:40:50,813
fully anti-AI, and that's kind of
the purity test that they might use.

648
00:40:50,813 --> 00:40:54,686
My 10-year-old is kind of like
that. He's like, it's just AI slop.

649
00:40:54,686 --> 00:41:00,208
AI slop. It's like, come
on, man. It's a tool.

650
00:41:00,208 --> 00:41:04,224
How are you using the tool to accomplish
the end that you're happy with?

651
00:41:04,224 --> 00:41:06,821
That's really what's important, anyways.

652
00:41:07,303 --> 00:41:11,783
Leo Dion (host): My 10-year-old has been
using my Claude Code license to build games.

653
00:41:11,783 --> 00:41:14,942
But unfortunately, most of
his games just come from,

654
00:41:14,942 --> 00:41:17,849
"I just played Pokémon,
I wanna make a Pokémon."

655
00:41:17,849 --> 00:41:21,874
And it's like, and I don't know
how to teach a 10-year-old this.

656
00:41:21,874 --> 00:41:24,809
I think it's impossible, but
it's like, and it's boring,

657
00:41:24,809 --> 00:41:28,300
it's like, you gotta, buddy,
you gotta make a PRD document.

658
00:41:28,300 --> 00:41:30,286
And yeah, he doesn't want to do that.

659
00:41:30,286 --> 00:41:34,316
But one of the things I'm saying is,
and we talked about this with Joe when

660
00:41:34,316 --> 00:41:38,967
he was on, part of it is being able to
put into words exactly what you want.

661
00:41:38,967 --> 00:41:42,228
And I think that's really the skill.

662
00:41:42,228 --> 00:41:46,975
That's really the skill that's part
of the product engineer skill, right?

663
00:41:46,975 --> 00:41:47,841
Jared Sorge (guest): Yeah.

664
00:41:47,841 --> 00:41:49,890
My 10-year-old, a couple of years ago,

665
00:41:49,890 --> 00:41:53,433
wanted to do YouTube videos, and
I would say, okay, that's great.

666
00:41:53,433 --> 00:41:56,947
What are you making? And he's like, I
don't know, just turn the camera on.

667
00:41:56,947 --> 00:41:58,219
Like, you don't know?

668
00:41:58,219 --> 00:42:04,192
We need to have a plan for this, because
it's all about how is it planned out?

669
00:42:04,192 --> 00:42:05,076
What are you gonna be doing?

670
00:42:05,076 --> 00:42:07,781
And then a little bit later,
you know, that fizzled out.

671
00:42:07,781 --> 00:42:10,855
But he wanted to make
video games, similar thing.

672
00:42:10,855 --> 00:42:16,011
And he has some great ideas for
storytelling, and he could be a great

673
00:42:16,011 --> 00:42:21,446
D&D dungeon master someday, 'cause
he's made these tabletop games with

674
00:42:21,446 --> 00:42:26,720
different quests that you go on to
have a full story and everything.

675
00:42:26,720 --> 00:42:29,702
But he wants to make a video
game and he doesn't know

676
00:42:29,702 --> 00:42:32,824
what he wants and he doesn't
know how to plan it out.

677
00:42:32,824 --> 00:42:34,360
And I'm like, well, you're not gonna—

678
00:42:34,360 --> 00:42:36,390
if you don't know what you're going for,

679
00:42:36,390 --> 00:42:40,100
you're not gonna get what you want
because you don't know what you want.

680
00:42:40,100 --> 00:42:46,432
And it's wild how much pre-production
and planning needs to go into a movie,

681
00:42:46,432 --> 00:42:52,417
a TV show, an app, or a feature of an
app in order to make it successful.

682
00:42:52,417 --> 00:42:55,959
And all of that is gonna make
for better work, better output,

683
00:42:55,959 --> 00:42:58,284
when you're using especially
these coding agents.

684
00:42:58,284 --> 00:43:02,765
Leo Dion (host): With any project, we always
think about the fun stuff we get to do,

685
00:43:02,765 --> 00:43:06,151
and never the paperwork
and planning and all that.

686
00:43:06,914 --> 00:43:08,730
No one wants to think
about the non-fun stuff.

687
00:43:08,730 --> 00:43:12,508
And I don't know, you know, going
back to AI, I don't know yet,

688
00:43:12,508 --> 00:43:16,995
and maybe there is a skill for this,
where it'll ask you the questions that

689
00:43:16,995 --> 00:43:21,466
need to be asked in order to know how to
build that PRD, you know what I mean?

690
00:43:21,466 --> 00:43:24,414
Jared Sorge (guest): If there's not
a skill, you could make a skill.

691
00:43:26,302 --> 00:43:27,949
Leo Dion (host): It's— we
did the episode with Donny.

692
00:43:27,949 --> 00:43:30,990
There's that grill-me skill that's
useful, and I've used it and I love it.

693
00:43:30,990 --> 00:43:37,173
But that's not exactly a
planning skill per se. Did you—

694
00:43:37,173 --> 00:43:42,542
what do you use, sorry, for Arborist, what
do you use for your like, work items?

695
00:43:42,542 --> 00:43:47,057
Jared Sorge (guest): My work items?
Like a to-do checklist, or...

696
00:43:47,057 --> 00:43:49,432
I've gone back and forth.

697
00:43:50,496 --> 00:43:54,674
I want to use something like GitHub
issues and have milestones and

698
00:43:54,674 --> 00:43:57,791
have things easily accessible
to an agent to say,

699
00:43:57,791 --> 00:44:02,710
read this GitHub issue and ask me questions
about it, and then go build the thing.

700
00:44:02,710 --> 00:44:05,391
What I really use is a list in Things.

701
00:44:05,391 --> 00:44:11,223
I have a project for an app release.
I'll have each item be an idea.

702
00:44:11,223 --> 00:44:15,118
I've started using sections, section
headers, so that I can know what are

703
00:44:15,118 --> 00:44:19,512
my ideas versus what are things I've
heard from customers or user feedback.

704
00:44:20,643 --> 00:44:24,178
But it's the thing that provides
me the least amount of friction,

705
00:44:24,178 --> 00:44:26,177
and that I just fall back to.

706
00:44:29,481 --> 00:44:33,970
I would like to have something that's
a bit more connected and feels more

707
00:44:33,970 --> 00:44:38,664
professional, but whenever I try that,
it doesn't stick for whatever reason.

708
00:44:38,664 --> 00:44:40,033
Leo Dion (host): I like GitHub issues.

709
00:44:40,033 --> 00:44:43,842
That's a good compromise
instead of going down the

710
00:44:43,842 --> 00:44:47,452
rabbit hole of Kanban
boards and stuff like that.

711
00:44:47,452 --> 00:44:50,202
But I could see that
would be an issue for you.

712
00:44:50,254 --> 00:44:54,129
No pun intended, because
you have a monorepo, right?

713
00:44:54,129 --> 00:44:57,941
So you're gonna have all your
issues in one repo, right?

714
00:44:57,941 --> 00:45:01,021
Jared Sorge (guest): I mean, I've got the
different tags for the different apps,

715
00:45:01,021 --> 00:45:04,634
so I could filter based on
the app that I'm working on.

716
00:45:04,700 --> 00:45:08,201
You know, monorepos are
not the be-all end-all,

717
00:45:08,201 --> 00:45:11,354
and I understand why
people don't like them.

718
00:45:11,354 --> 00:45:14,637
For me, it's the tool that does
the job that I want it to do,

719
00:45:14,637 --> 00:45:17,288
so it works for me, but
your mileage may vary.

720
00:45:17,288 --> 00:45:21,711
Leo Dion (host): We had a question from the
audience about work trees and where they fit.

721
00:45:21,711 --> 00:45:22,869
It's from Twitter.

722
00:45:22,869 --> 00:45:25,931
The question is, "With agents
doing most of the typing,

723
00:45:25,931 --> 00:45:28,932
what does the work tree
layout look like in practice?

724
00:45:28,932 --> 00:45:31,579
One per agent, one per feature?

725
00:45:31,579 --> 00:45:34,634
I ended up running everything
serially on a single tree,

726
00:45:34,634 --> 00:45:37,037
and I suspect that was
the wrong instinct."

727
00:45:37,037 --> 00:45:44,094
So I'll just say, I have one agent
per work tree, per issue, per agent.

728
00:45:44,094 --> 00:45:46,012
So that's the way I would do it.

729
00:45:46,012 --> 00:45:49,188
But I also ask the agent
the best way to do that.

730
00:45:49,188 --> 00:45:51,528
Because there may be parts that need to be

731
00:45:51,528 --> 00:45:53,864
done serially or where there's dependence.

732
00:45:53,864 --> 00:45:54,602
But yeah, that's it.

733
00:45:54,684 --> 00:45:56,565
Jared Sorge (guest): How many work
trees do you have at a given moment?

734
00:45:57,275 --> 00:45:58,313
Leo Dion (host): Five at most.

735
00:45:58,313 --> 00:45:59,077
Jared Sorge (guest): Okay.

736
00:45:59,077 --> 00:46:02,203
Leo Dion (host): So basically I
have a GitHub issue. Excuse me.

737
00:46:02,203 --> 00:46:06,461
So basically I have a GitHub issue, and
then I'll have a branch attached to that

738
00:46:06,461 --> 00:46:11,366
issue, and then I'll have an agent attached
to that work tree for that branch.

739
00:46:11,366 --> 00:46:12,379
Jared Sorge (guest): Okay.

740
00:46:12,379 --> 00:46:13,207
Leo Dion (host): Does that make sense?

741
00:46:13,207 --> 00:46:19,131
Jared Sorge (guest): Yeah. The way that I do
stuff is I generally have a work tree per app.

742
00:46:19,131 --> 00:46:24,414
So I'll have an Arborist dev work
tree, a Baseplate dev work tree.

743
00:46:24,414 --> 00:46:29,229
I might have a work tree that I spin
up temporarily to do one thing,

744
00:46:29,229 --> 00:46:33,079
like make a release or work
on a bug or what have you.

745
00:46:33,079 --> 00:46:36,627
But one of the things that I like
about Arborist is that it shows

746
00:46:36,627 --> 00:46:40,621
me in the inspector view all the
chats I've had in that work tree.

747
00:46:41,506 --> 00:46:46,171
Which, if you're in a one-work-tree-per-issue
or one-work-tree-per-chat,

748
00:46:46,171 --> 00:46:49,930
that's gonna be a list of
one, which is fine, right?

749
00:46:50,833 --> 00:46:52,843
There's no right or wrong way.

750
00:46:52,843 --> 00:46:58,221
It's really about how is the workflow
working for you and the tools that you use.

751
00:46:58,221 --> 00:47:02,533
At Adobe, we have Premiere
and Firefly in one repo,

752
00:47:02,533 --> 00:47:06,081
and sharing a bunch of code with
each other through Swift packages.

753
00:47:06,081 --> 00:47:08,144
And I'm much more loosey goosey there.

754
00:47:08,144 --> 00:47:11,808
Like, I've got a work tree for
the Xcode 27 work I'm doing.

755
00:47:11,808 --> 00:47:14,209
I've got a work tree for a
bug that I might be trying

756
00:47:14,209 --> 00:47:16,260
to fix or this feature
that I'm trying to add.

757
00:47:16,260 --> 00:47:24,792
And it's less regimented in a way, but
the work in my day job feels a lot more

758
00:47:24,792 --> 00:47:29,732
chaotic than the side stuff because
I'm not in control over there, right?

759
00:47:29,732 --> 00:47:34,556
I'm one member of the team, whereas
with my own stuff, I am the team.

760
00:47:34,556 --> 00:47:37,928
So I get to run things how I want to.

761
00:47:37,928 --> 00:47:41,665
And that's one of the nice things
about having side projects.

762
00:47:41,665 --> 00:47:47,151
Leo Dion (host): 100%, yes. What
are your future plans for Arborist?

763
00:47:47,151 --> 00:47:49,336
What is the future you
really want to get in?

764
00:47:49,336 --> 00:47:50,856
Jared Sorge (guest): I've
got a long list of ideas.

765
00:47:50,856 --> 00:47:53,110
One of the things that I want to do,

766
00:47:53,110 --> 00:47:57,207
that I hope is lowish-hanging
fruit, is stuff like command input.

767
00:47:57,207 --> 00:48:00,565
So right now it's a non-interactive shell.

768
00:48:00,565 --> 00:48:04,325
You type in the command you want it to run
and it will run precisely that command.

769
00:48:04,325 --> 00:48:09,504
But for instance, I have a
mise task on my blog repo

770
00:48:09,504 --> 00:48:13,110
for making a new post that
could take in a post name.

771
00:48:13,110 --> 00:48:16,574
And right now I have to run
that from a command prompt,

772
00:48:16,574 --> 00:48:19,047
because Arborist will just no-op on it.

773
00:48:19,047 --> 00:48:21,189
It'll say this exited with code one.

774
00:48:21,189 --> 00:48:25,341
So I'd like to support things like
that. I also would like to...

775
00:48:25,341 --> 00:48:29,350
So you introduced me to the concept
of Grove, or these tools where,

776
00:48:29,350 --> 00:48:32,977
when you create a new repo, it
runs a certain set of tasks.

777
00:48:32,977 --> 00:48:37,841
I'd like to have some sort of way to
say, on a new work tree creation,

778
00:48:37,841 --> 00:48:43,825
automatically run these commands
or these tasks from Arborist.

779
00:48:43,825 --> 00:48:45,487
I think that would be really nice to have.

780
00:48:45,487 --> 00:48:50,999
Leo Dion (host): And then, a .arborist file or
something, so you can put your preferences in there?

781
00:48:51,152 --> 00:48:56,701
Jared Sorge (guest): Having a .arborist file
would then depend on knowing how the...

782
00:48:56,701 --> 00:49:03,357
Like, then I would almost make Arborist
my own command runner, in a way.

783
00:49:04,422 --> 00:49:07,167
I'm thinking through this in real time,

784
00:49:07,167 --> 00:49:09,208
because I hadn't considered
a .arborist file.

785
00:49:09,208 --> 00:49:13,502
I was thinking about it more in
how can I do this in the UI?

786
00:49:14,530 --> 00:49:17,537
And maybe that's the right
way to go. I don't know.

787
00:49:17,537 --> 00:49:19,745
I don't know how that's
actually gonna work out yet.

788
00:49:23,965 --> 00:49:25,913
And then the other thing is,

789
00:49:25,913 --> 00:49:30,279
I've got a client who has two
repos that I work in all the time.

790
00:49:30,279 --> 00:49:33,705
One is their apps monorepo,
and another is the CLI

791
00:49:33,705 --> 00:49:36,925
that makes the tools that
support that apps repo.

792
00:49:38,331 --> 00:49:41,683
I would like to have a
higher-level collection,

793
00:49:41,683 --> 00:49:45,244
so that Arborist can run
commands from not just...

794
00:49:45,244 --> 00:49:48,526
Or accept in the sidebar
not just a .git repo,

795
00:49:48,526 --> 00:49:52,531
but some other collection of
repos to run commands in.

796
00:49:52,531 --> 00:49:54,429
Leo Dion (host): Okay, that makes sense.

797
00:49:56,365 --> 00:49:57,694
Jared Sorge (guest): That's
all down the road, though.

798
00:49:57,694 --> 00:50:00,371
Leo Dion (host): Was there anything else
you wanted to mention about Arborist?

799
00:50:00,371 --> 00:50:01,694
Jared Sorge (guest): I think
we've pretty well covered it.

800
00:50:01,694 --> 00:50:04,856
Leo Dion (host): Okay, cool. How are you
doing on time? I wanna respect your time.

801
00:50:04,856 --> 00:50:07,771
Jared Sorge (guest): I could do a couple more
minutes. I need to wrap up pretty soon though.

802
00:50:07,771 --> 00:50:12,583
Leo Dion (host): Okay. Before we close out,
do you have any thoughts on the iPhone event?

803
00:50:12,716 --> 00:50:15,640
Jared Sorge (guest): Oh man, so I'm
excited to get to see the 18 Pro.

804
00:50:15,640 --> 00:50:17,435
My wife's got one coming today.

805
00:50:17,435 --> 00:50:21,025
She's on like we kind of alternate
iPhone years, so this is her year.

806
00:50:21,025 --> 00:50:23,246
I'm also... It's strange.

807
00:50:23,246 --> 00:50:29,333
I'm not as tempted by the 18
Pro as I thought I might be.

808
00:50:29,333 --> 00:50:30,903
I have a 17 Pro.

809
00:50:31,216 --> 00:50:39,730
Leo Dion (host): I like the 17 Pro. Yeah,
I am envious of you. I want the orange.

810
00:50:42,314 --> 00:50:45,882
Jared Sorge (guest): Orange has been my favorite
color for a long time, and so I don't know that

811
00:50:45,882 --> 00:50:49,278
I would have bought one, just because
I don't want the orange to go away.

812
00:50:49,278 --> 00:50:50,894
And then they brought out the Duo.

813
00:50:50,894 --> 00:50:54,163
And I've been so torn on
what to do for the Duo.

814
00:50:54,163 --> 00:50:57,280
Because it's the most
intriguing iPhone that they

815
00:50:57,280 --> 00:51:00,089
brought out since at
least the 10, but maybe...

816
00:51:00,089 --> 00:51:01,364
Leo Dion (host): Since the first one?

817
00:51:01,364 --> 00:51:08,130
Jared Sorge (guest): Just because
of that giant screen on the inside.

818
00:51:08,130 --> 00:51:10,989
It's so thin when it's unfolded.

819
00:51:10,989 --> 00:51:11,373
Leo Dion (host): Yeah.

820
00:51:11,373 --> 00:51:16,849
Jared Sorge (guest): And then software-wise, as soon
as they showed that book cover expanding effect

821
00:51:16,849 --> 00:51:21,544
when they opened it up, oh my goodness,
I want to play with that so badly.

822
00:51:21,544 --> 00:51:24,510
Leo Dion (host): Yeah.

823
00:51:26,457 --> 00:51:27,953
Jared Sorge (guest):
And then as a developer,

824
00:51:27,953 --> 00:51:31,997
I really want to see how my apps
will work with the screen open.

825
00:51:31,997 --> 00:51:34,920
I thankfully don't use a
lot of custom stuff anymore,

826
00:51:34,920 --> 00:51:37,950
especially in Baseplate, so
that should work just fine.

827
00:51:37,950 --> 00:51:41,394
I've been working really hard
to get the adaptability pieces.

828
00:51:41,394 --> 00:51:43,340
Leo Dion (host): Yeah, I think so.

829
00:51:43,392 --> 00:51:44,413
Jared Sorge (guest): Yeah.

830
00:51:46,559 --> 00:51:50,448
Leo Dion (host): But I
know... I wanna see it.

831
00:51:50,448 --> 00:51:54,035
I don't have I don't have
the budget for an iPhone Duo.

832
00:51:54,035 --> 00:51:58,823
By the way, if people are interested,
join my Patreon or sponsor me on GitHub.

833
00:51:58,823 --> 00:52:01,617
Just saying. No, that's totally unrelated.

834
00:52:01,617 --> 00:52:08,443
But yeah, hopefully by the time we release
this episode, we have a beta for 27.1.

835
00:52:08,443 --> 00:52:10,681
We have the beta for 27.2 for some reason,

836
00:52:10,681 --> 00:52:14,544
which doesn't include the
simulator for Duo, but...

837
00:52:14,544 --> 00:52:18,258
Jared Sorge (guest): Yeah. Just how locked
down the device was from everybody.

838
00:52:18,276 --> 00:52:22,301
Like people knew it was coming,
but no one had access to any of it

839
00:52:22,301 --> 00:52:26,595
internally because it was like a totally
separate release train internally.

840
00:52:26,595 --> 00:52:32,047
Leo Dion (host): It's kinda wild. That's
what you've heard from the grapevine?

841
00:52:32,216 --> 00:52:36,037
Yeah. Oh, well. We'll see.

842
00:52:36,875 --> 00:52:38,055
Jared Sorge (guest): 27.2
is interesting, though.

843
00:52:38,055 --> 00:52:38,913
Leo Dion (host): In what way?

844
00:52:38,913 --> 00:52:41,591
Jared Sorge (guest): At least on the Xcode
side, we have a whole new project format.

845
00:52:41,591 --> 00:52:43,337
Leo Dion (host): Oh yeah, that's amazing.

846
00:52:43,494 --> 00:52:46,122
Jared Sorge (guest): It's like it's 2019.

847
00:52:46,122 --> 00:52:49,323
Leo Dion (host): It's
wonderful. Glad we're there.

848
00:52:50,577 --> 00:52:53,813
I'm glad I was able to get all my content
about Xcode project generators out

849
00:52:53,813 --> 00:52:57,326
like six years ago, before they clobbered
it all with this new JSON format.

850
00:52:57,326 --> 00:53:00,340
I mean, I think there's
still a big use case

851
00:53:00,340 --> 00:53:03,095
for Tuist and XcodeGen in a lot of ways.

852
00:53:03,095 --> 00:53:05,773
'Cause that JSON format, I
haven't looked at it yet,

853
00:53:05,773 --> 00:53:07,841
but I can't imagine it's pretty.

854
00:53:07,841 --> 00:53:10,402
Jared Sorge (guest): Yeah, I haven't
looked at it yet either, very thoroughly.

855
00:53:10,402 --> 00:53:12,127
It only came out like a day or two ago.

856
00:53:12,127 --> 00:53:14,663
Leo Dion (host): Jared, thank you
so much for coming on the show.

857
00:53:14,663 --> 00:53:16,500
I'm really happy to have you back on.

858
00:53:16,655 --> 00:53:19,735
We'll be talking more
about AI and how to learn

859
00:53:19,735 --> 00:53:22,479
AI in the next episode with Stewart Lynch.

860
00:53:22,479 --> 00:53:24,382
So be on the lookout for that.

861
00:53:24,382 --> 00:53:28,637
Subscribe if you aren't a subscriber
on the podcast or YouTube channel.

862
00:53:28,637 --> 00:53:32,815
And then, if you build your website
using Swift and you use Vapor,

863
00:53:32,815 --> 00:53:36,257
well, we're gonna have Tim
Condon on to talk about Vapor 5.

864
00:53:36,257 --> 00:53:38,782
So be on the lookout.

865
00:53:38,782 --> 00:53:43,664
Yeah, we'll post a link to the blog
posts just to get you interested.

866
00:53:43,664 --> 00:53:46,154
But yeah, so be a subscriber.

867
00:53:46,154 --> 00:53:49,532
Check it out so you can check
these episodes out when they come.

868
00:53:49,590 --> 00:53:53,241
Jared, do you wanna plug
where you're at and Arborist?

869
00:53:53,241 --> 00:53:57,287
Jared Sorge (guest): Yeah,
I'm @jsorge on Mastodon.

870
00:53:57,287 --> 00:54:02,872
My website is jsorge.net, and you
can check out my apps at taphouse.io,

871
00:54:02,872 --> 00:54:04,993
or Arborist is at macarborist.com.

872
00:54:04,993 --> 00:54:07,070
Leo Dion (host): Awesome.
Thank you so much.

873
00:54:07,070 --> 00:54:11,427
And I look forward to talking
to everybody next time. Bye.

874
00:54:11,427 --> 00:54:12,274
Jared Sorge (guest): See you later.