WEBVTT

1
00:00:00.251 --> 00:00:03.513
Noah Breyer might have the coolest Claude code setup I've ever seen.

2
00:00:03.853 --> 00:00:13.158
He rigged a home server in his basement, put his Obsidian vault in it, and then runs Claude code on top, so he can think, research, write, and even ship code right from his phone.

3
00:00:13.398 --> 00:00:21.962
Today he shows us how he uses Claude code as a true second brain, a thinking partner, that asks him sharp questions, pulls research from his whole Node archive and the web,

4
00:00:22.222 --> 00:00:25.544
and even keeps a running log of what he's learned and what his best ideas are.

5
00:00:25.724 --> 00:00:28.725
And he walks us through his whole stack and his whole workflow.

6
00:00:28.825 --> 00:00:33.028
If you wanna learn how to use Cloud Code as a true second brain, this is the episode to watch.

7
00:00:33.268 --> 00:00:40.491
Let's dive in.

8
00:00:48.575 --> 00:00:49.696
Noah, welcome to the show.

9
00:00:50.356 --> 00:00:51.277
Thanks for having me.

10
00:00:52.422 --> 00:00:53.544
I'm excited to have you.

11
00:00:54.104 --> 00:00:56.487
It's really good to get to chat.

12
00:00:56.587 --> 00:01:00.251
This is our first interview in probably like five years.

13
00:01:00.271 --> 00:01:05.898
You are one of the, for people who don't know, you're one of the first super organizers interviewees.

14
00:01:05.938 --> 00:01:08.020
That was the newsletter that turned into Every.

15
00:01:08.040 --> 00:01:08.100
And

16
00:01:10.583 --> 00:01:12.285
I love the way that your brain works.

17
00:01:12.845 --> 00:01:17.590
You have this really interesting taste for tools for thought.

18
00:01:17.990 --> 00:01:21.533
And back in the day, you were using Evernote in all these really interesting ways.

19
00:01:22.414 --> 00:01:27.719
You were the co-founder of a really cool startup called Percolate, and then another one called Variance.

20
00:01:27.739 --> 00:01:31.423
And now you're running Alephic, which is an AI strategy consultancy.

21
00:01:32.848 --> 00:01:45.524
And I'm just really excited to see how your mind has started to use these AI tools now that they're working so well.

22
00:01:45.904 --> 00:01:48.608
And I know you have some pretty cool Claude code stuff to show us.

23
00:01:48.688 --> 00:01:50.110
So, yeah, thanks for coming on.

24
00:01:50.802 --> 00:01:51.763
Thanks for having me.

25
00:01:51.803 --> 00:01:52.644
I'm super excited.

26
00:01:52.724 --> 00:01:55.886
That was a fun interview all those years back.

27
00:01:56.147 --> 00:01:57.107
It was really, really fun.

28
00:01:58.188 --> 00:02:04.434
So I want to just dive right into the thing that I think is so cool about what you're doing.

29
00:02:04.454 --> 00:02:08.137
So I know you have a whole vibe coding setup that you built for yourself.

30
00:02:08.177 --> 00:02:09.478
Can you talk us through that?

31
00:02:11.095 --> 00:02:15.263
Yeah, I wouldn't, uh, I'm not sure about actually the vibe coding part of it.

32
00:02:15.543 --> 00:02:20.352
Um, uh, I have a, a sort of fairly heavy duty cloud code set up.

33
00:02:21.013 --> 00:02:22.336
Um, but actually, uh,

34
00:02:24.354 --> 00:02:25.555
mostly not for code.

35
00:02:26.815 --> 00:02:38.741
So since those days of super organizers, like many people, I've abandoned Evernote and switched over to Obsidian.

36
00:02:39.561 --> 00:02:43.043
And one of the big advantages with Obsidian as a note taking

37
00:02:43.343 --> 00:02:46.206
platform is that it's a bunch of markdown files and a bunch of folders.

38
00:02:47.768 --> 00:02:55.997
And they can then be synced with Git and you can do lots of other fun kinds of things.

39
00:02:56.638 --> 00:03:00.182
And so actually probably my number one Claude code use is,

40
00:03:00.982 --> 00:03:06.928
is using it as a tool to interact with my notes.

41
00:03:07.648 --> 00:03:12.933
And so I've got a fairly serious Cloud Code setup that I use with Obsidian.

42
00:03:13.513 --> 00:03:25.284
And my most recent obsession has been standing up a server in my house so that I could also use Cloud Code on my phone.

43
00:03:26.601 --> 00:03:27.402
This is incredible.

44
00:03:27.523 --> 00:03:29.666
I want to go through all of this.

45
00:03:29.686 --> 00:03:31.229
So where should we start?

46
00:03:31.249 --> 00:03:38.100
Should we do how you use Cloud Code as sort of this like research assistant, notes organizer, note taker thing?

47
00:03:38.241 --> 00:03:40.624
Or should we start with how you use it on your phone?

48
00:03:42.277 --> 00:03:42.958
We can use it.

49
00:03:43.078 --> 00:03:46.360
We can start with just the sort of general part of it.

50
00:03:46.640 --> 00:03:48.001
That might be the sort of easiest.

51
00:03:48.021 --> 00:03:51.002
The phone is really just an extension of that same thing.

52
00:03:51.222 --> 00:04:06.712
I would say sort of generally, and this is something I feel like not enough people talk about with AI, is like one of the things I find really extraordinary about it is the ability for me to work really productively

53
00:04:07.572 --> 00:04:08.213
on my phone.

54
00:04:08.713 --> 00:04:16.461
And that's been like a huge, huge change because so much of what I do is sort of writing or coding and the phone is definitely not the best place for that.

55
00:04:16.541 --> 00:04:19.984
And, you know, even the phone wasn't always the best place for doing research and thinking.

56
00:04:21.066 --> 00:04:24.829
I felt like my computer was a better place for it, which is why I've been such a sort of note taker.

57
00:04:24.909 --> 00:04:27.952
And, you know, I have found whether it's like,

58
00:04:28.753 --> 00:04:34.074
Claude Code and Obsidian or, I mean, even Claude Code and Code, right?

59
00:04:34.094 --> 00:04:47.056
So like the other piece of it is being able to then, you know, if you see something go wrong, being able to sign in on your phone and have Claude Code push a small update to something because you just realize it while you're out is amazing.

60
00:04:47.436 --> 00:04:53.717
But then even like, you know, I use quite a bit of Grok voice mode.

61
00:04:54.318 --> 00:04:56.478
And, you know, I find that that as a sort of like

62
00:04:56.818 --> 00:04:59.039
alternative way of working through problems.

63
00:04:59.099 --> 00:05:01.360
I have a Tesla, so now it's baked into the Tesla.

64
00:05:02.561 --> 00:05:12.305
And, you know, it's just, I, and, you know, obviously all the sort of other Chad GBT and Claude and all these things of just being able to sort of like go and do research and really think and, and explore things differently.

65
00:05:14.166 --> 00:05:27.130
in this device that's always been useful, but like not useful for deep work, I think is probably something most people would agree with is that the phone is not been the best place to kind of do deep coding and research work.

66
00:05:27.270 --> 00:05:30.711
And I feel like it's really changed my ability to do that.

67
00:05:31.132 --> 00:05:32.032
Wait, I got to stop you.

68
00:05:32.072 --> 00:05:34.233
So you're using Grok voice mode.

69
00:05:34.293 --> 00:05:41.875
And is that specifically because it's built into your Tesla or using it in situations where you could also use, for example, ChatGBT voice mode?

70
00:05:42.916 --> 00:05:47.702
No, I'm using it because it's way better than any of the other voice modes and I will fight anybody who says anything different.

71
00:05:48.102 --> 00:05:48.423
Okay.

72
00:05:48.503 --> 00:05:50.465
No, tell me, like, what do you like about it?

73
00:05:50.485 --> 00:05:50.926
Why is it better?

74
00:05:51.859 --> 00:06:01.207
To be fair, OpenAI launched their real-time API, which may or may not be baked into ChatGPT voice now.

75
00:06:01.487 --> 00:06:07.512
It's not totally clear, but the old voice mode was based on 4.0 and I just found it to be completely unusable.

76
00:06:08.893 --> 00:06:13.097
And Gemini's voice mode, I just didn't find to be smart enough.

77
00:06:13.897 --> 00:06:18.101
And I just found Grok's voice mode to be significantly smarter than anybody else's.

78
00:06:18.121 --> 00:06:20.643
You know, I'm using Grok,

79
00:06:22.323 --> 00:06:24.784
two, three, four, I don't even remember what, or whatever the latest.

80
00:06:24.864 --> 00:06:25.764
The latest one, yeah.

81
00:06:25.844 --> 00:06:29.125
Yeah, but not the super, I don't have the most expensive account.

82
00:06:29.145 --> 00:06:30.125
Super heavy or whatever.

83
00:06:30.165 --> 00:06:31.705
I don't have super heavy.

84
00:06:32.205 --> 00:06:33.446
But I just find it to be much better.

85
00:06:33.626 --> 00:06:35.706
It does tool calling way better than any of the other ones.

86
00:06:35.726 --> 00:06:43.568
That's what I found to be a major weakness of the voice models, is that they don't do great tool calling and research, and Grox team steps, all that.

87
00:06:43.588 --> 00:06:48.889
So no, even before it was loaded in my Tesla, I dropped my daughter off at summer camp,

88
00:06:49.109 --> 00:06:50.430
this summer up in New Hampshire.

89
00:06:50.450 --> 00:06:58.196
So I had a five hour drive on my own and I spent like two hours researching and essentially like working through a piece.

90
00:06:58.977 --> 00:07:04.801
Um, and I did it by just like connecting it to Bluetooth and just sort of sitting there in the car.

91
00:07:04.901 --> 00:07:09.925
And, um, I found it to be by far the, uh, the best of the voice modes.

92
00:07:10.025 --> 00:07:10.586
I, I hope

93
00:07:11.810 --> 00:07:14.012
these other models catch up there.

94
00:07:14.092 --> 00:07:17.054
Cause I would, I would love more really good voice modes.

95
00:07:17.114 --> 00:07:22.678
I mean, I had a mind blowing session this weekend and I'm giving a talk and I'm, I sort of have some ideas.

96
00:07:22.698 --> 00:07:26.200
I think it's generally going to be about transformers eating the world.

97
00:07:26.280 --> 00:07:31.003
And so I was sort of catching myself up on self-attention and exactly how it works.

98
00:07:31.103 --> 00:07:37.708
And I did like an hour session and it really, I, it like was by far the sort of best, um,

99
00:07:39.280 --> 00:07:42.581
explanation I've ever read for it or ever heard, I guess.

100
00:07:42.681 --> 00:07:46.903
And so, yeah, I've, I've just found it to be a kind of pretty extraordinary product.

101
00:07:47.523 --> 00:07:48.984
I do love voice mode for that.

102
00:07:49.384 --> 00:07:55.127
Um, it's sort of like, it's the, it's a podcast made specifically for you about whatever you're curious about.

103
00:07:55.347 --> 00:07:56.187
And that's really cool.

104
00:07:56.727 --> 00:08:03.490
I went up to, I drove upstate this weekend and I've been reading, um, I've been reading the Iliad.

105
00:08:04.385 --> 00:08:07.568
And so I had it on an audio book and then I had some questions as I was driving.

106
00:08:07.608 --> 00:08:11.913
And so I unfortunately used ChatGPT voice mode because I didn't know about Grok.

107
00:08:11.933 --> 00:08:13.795
So I wish that Grok's voice mode.

108
00:08:13.855 --> 00:08:16.537
I wish that we'd had this conversation before then.

109
00:08:16.618 --> 00:08:25.366
But the thing about ChatGPT voice mode is, yeah, I think when it first came out, it was cool, but it just hasn't gotten as smart as the models are now.

110
00:08:26.087 --> 00:08:32.209
And they gave it this new personality that I had to get used to where every time you ask it a question, it goes like, oh, yeah, uh-huh.

111
00:08:32.269 --> 00:08:41.673
Well, you know, and it's like it's just this like weird Gen Z thing that it feels like it has a little bit too much ennui or something, like doesn't actually care about you.

112
00:08:42.253 --> 00:08:43.113
I don't know what that is.

113
00:08:43.153 --> 00:08:44.133
So I had to get used to that.

114
00:08:44.374 --> 00:08:47.195
Brock has a stoner mode for what it's worth.

115
00:08:47.655 --> 00:08:47.955
Yeah.

116
00:08:49.730 --> 00:08:53.131
I will say the car version is very interesting to me.

117
00:08:53.211 --> 00:08:56.972
Like this was in the most recent Tesla release, like a couple of weeks ago.

118
00:08:57.912 --> 00:09:06.035
And, you know, I had been doing that same thing you did where you just plug your phone in and you put on Bluetooth and you, you know, do your best to make it work.

119
00:09:06.075 --> 00:09:10.656
And it's very interesting to just like have a voice AI button and it syncs back to your phone.

120
00:09:11.660 --> 00:09:16.625
regular Grok, but you can't rejoin old chats.

121
00:09:17.085 --> 00:09:18.106
So it's just like, hey, I...

122
00:09:18.406 --> 00:09:18.847
It's just like...

123
00:09:19.107 --> 00:09:24.252
But, you know, I mean, these things are significantly better than Siri and all of these other things.

124
00:09:24.492 --> 00:09:31.018
And particularly, I mean, you know, there's no comparison if you actually have something more than just a single question you want an answer to, right?

125
00:09:31.038 --> 00:09:34.321
Like if you actually want to have a conversation about the Iliad or, you know, about...

126
00:09:35.102 --> 00:09:57.922
transformers and self-attention like um i don't know it's pretty amazing to just be able to sort of like hit this button and and yeah use that car time i mean i i was on my way somewhere last week and i was um you know i was like having it research uh i was going back to the um walter benjamin i was i was i i have this sort of idea to write a piece about um how uh

127
00:09:58.622 --> 00:10:04.925
um, the reactions to every new technology are essentially elitist, um, critiques of it.

128
00:10:05.146 --> 00:10:10.448
And that, you know, it's always like, oh no, everybody's going to be able to like do this thing that only we used to be able to do.

129
00:10:11.069 --> 00:10:19.233
And so, um, I was in the car and I was thinking about this and so I had to go and I was like, okay, you know, I know it's been years since I read the Walter Benjamin, uh,

130
00:10:19.453 --> 00:10:21.155
Mass production of images one.

131
00:10:21.215 --> 00:10:21.595
Yeah.

132
00:10:21.615 --> 00:10:21.815
Yeah.

133
00:10:21.855 --> 00:10:22.436
That, that one.

134
00:10:22.997 --> 00:10:28.402
And so, yeah, then I'm having a conversation about that and I'm like, who are Walter Benjamin's contemporaries?

135
00:10:28.442 --> 00:10:31.365
And then I'm like into all the, you know, and it's just like, I don't know.

136
00:10:31.385 --> 00:10:33.587
That's a, that's amazing.

137
00:10:33.947 --> 00:10:34.528
It's the best.

138
00:10:35.168 --> 00:10:35.628
It's the best.

139
00:10:35.668 --> 00:10:36.209
Yeah, I don't know.

140
00:10:38.911 --> 00:10:39.251
Okay.

141
00:10:39.371 --> 00:10:42.514
And so you're filling your brain with all these things from voice mode, which I love.

142
00:10:42.994 --> 00:10:45.536
But tell us about your second brain setup.

143
00:10:45.556 --> 00:10:49.839
I don't know how you refer to it, whether you think that second brain is appropriate for this.

144
00:10:49.939 --> 00:10:54.623
But I want to know how you're using Cloud Code to take notes and do research and all that kind of stuff.

145
00:10:55.178 --> 00:10:58.340
Yeah, so I could just open it up.

146
00:10:58.400 --> 00:11:00.761
Maybe that's the easiest thing to just walk you through it.

147
00:11:00.821 --> 00:11:03.082
I'll start on my computer, and then we can do the phone.

148
00:11:03.102 --> 00:11:06.324
The computer is just a way easier share here.

149
00:11:07.404 --> 00:11:08.065
All right.

150
00:11:08.625 --> 00:11:13.107
So this is what I was working on before.

151
00:11:14.888 --> 00:11:20.491
But essentially, this is just Cloud Code, and it's just sitting on top of my Obsidian.

152
00:11:20.591 --> 00:11:23.393
So if I jump out here and I just do...

153
00:11:24.636 --> 00:11:35.542
Like, you know, you can see I'm following the the para method and, you know, I've just got everything sort of organized in here and put in the places that they need to be.

154
00:11:35.562 --> 00:11:38.324
Let me just let me just stop you for people who are listening.

155
00:11:38.404 --> 00:11:41.045
So, OK, so we're looking at we're looking at cloud code.

156
00:11:41.906 --> 00:11:45.568
It sounds it seems like you have cloud code running in your obsidian vault.

157
00:11:46.790 --> 00:11:51.012
And there's some kind of, uh, it's adding something to an exist.

158
00:11:51.032 --> 00:11:52.933
It looks like it's adding something to an existing note.

159
00:11:53.153 --> 00:11:54.574
Is that, that's what, that's what's going on.

160
00:11:54.594 --> 00:11:55.275
That's what we're looking at.

161
00:11:55.435 --> 00:11:55.655
Yeah.

162
00:11:55.675 --> 00:11:57.816
So in this particular one, I'm, I'm working on this talk.

163
00:11:57.856 --> 00:12:00.497
So I, I'm putting on my conference in two weeks.

164
00:12:00.577 --> 00:12:03.819
I'm giving this talk about marketing and AI and sort of what's going on.

165
00:12:03.899 --> 00:12:05.580
And, um, I'm, uh,

166
00:12:06.948 --> 00:12:14.443
If we sort of jump back a second, I've been doing these conferences called Brand, brxnd.ai, and they're about marketing and AI.

167
00:12:14.503 --> 00:12:16.527
And I did one in February in...

168
00:12:18.418 --> 00:12:18.818
LA.

169
00:12:19.178 --> 00:12:29.604
And my talk in LA was about this, I'm sure you've seen it, it was the Office of Strategic Services, which was the precursor to the CIA, wrote this manual called the Simple Sabotage Field Manual.

170
00:12:30.445 --> 00:12:40.491
And it was essentially a manual to help citizen saboteurs in Nazi-occupied territories sort of like quietly sabotage the Nazi occupation.

171
00:12:41.191 --> 00:12:45.254
And so it was like, you know, there's a whole bunch of stuff for blue collar workers that's like,

172
00:12:45.894 --> 00:12:53.222
If you're a janitor, you should leave a bucket of oily waste around and accidentally drop a cigarette in there so that, you know, it will.

173
00:12:53.462 --> 00:13:00.269
But then there's this amazing set of recommendations for white collar workers and they're like,

174
00:13:01.671 --> 00:13:16.761
always refer things to committee, always revisit previously made decisions, make sure that like, if somebody is trying to make a decision, you should suggest that they don't act with too much haste, lest like we'd be embarrassed.

175
00:13:17.021 --> 00:13:21.444
And so it's like, you know, my talk was about kind of,

176
00:13:21.924 --> 00:13:42.860
how one hope I have is that AI can kind of like sidestep a lot of the bureaucracy that exists inside large organizations because it sort of has this kind of goo-like effect where it can kind of fit into any crevice or crack because it can act as this fuzzy interface and it doesn't really care about the sort of input output.

177
00:13:43.500 --> 00:13:44.101
And so...

178
00:13:45.340 --> 00:13:52.064
the sort of next part of that story is I, after the conference, I realized that, um, uh, that manual was in the public domain.

179
00:13:52.084 --> 00:13:57.368
So I hired a designer and I printed 300 copies and I wrote a new forward for it.

180
00:13:58.028 --> 00:13:59.789
And so we're giving this away at the conference.

181
00:13:59.829 --> 00:14:02.351
And so my talk is sort of trying to tie all these ideas together.

182
00:14:02.471 --> 00:14:02.651
Right.

183
00:14:02.671 --> 00:14:05.213
So I'm, I'm trying to pull from the sabotage manual.

184
00:14:05.513 --> 00:14:15.219
And then I was doing a bunch of research into wild Bill Donovan, who started the OSS and the OSS was sort of the precursor to the, both the CIA and the, um, uh,

185
00:14:16.160 --> 00:14:16.860
special forces.

186
00:14:17.300 --> 00:14:19.001
And so anyway, I'm writing this talk.

187
00:14:19.041 --> 00:14:25.005
And so I've got a project inside by Obsidian, which is the beginning of the research for this project.

188
00:14:25.525 --> 00:14:28.807
And I'm pulling in sort of like chats and articles and all these things.

189
00:14:28.827 --> 00:14:31.549
And then I'm constantly kind of talking to the...

190
00:14:32.409 --> 00:14:34.911
the AI in here and giving it new ideas.

191
00:14:35.372 --> 00:14:37.273
So I'm like, oh, I need some conclusions.

192
00:14:37.313 --> 00:14:39.235
Here's sort of my first thought on conclusions.

193
00:14:39.595 --> 00:14:41.978
And I'm having it like note down the conclusions.

194
00:14:42.018 --> 00:14:49.624
And then at the end of each day, I have the AI write up the changes that I sort of like the things I learned that day that are going to help me push this talk along.

195
00:14:49.704 --> 00:14:54.288
And so that's what you're looking at right here is sort of this is all part of this

196
00:14:55.149 --> 00:14:58.631
um, work that I've been doing where I've been feeding it.

197
00:14:59.051 --> 00:15:02.413
Um, I was working on sort of what are some of the conclusions I want it to be.

198
00:15:02.914 --> 00:15:10.938
And so this is all sitting in this, uh, this is all sitting in, in my obsidian inside a project specifically for that talk.

199
00:15:11.458 --> 00:15:11.679
Okay.

200
00:15:11.699 --> 00:15:13.419
So let me get a clearer sense of this.

201
00:15:13.440 --> 00:15:14.400
This is really interesting.

202
00:15:14.480 --> 00:15:19.243
So you have a project, uh, when you, when you, you have a new thing, you're, you're giving a talk, um,

203
00:15:20.413 --> 00:15:21.175
You make a new folder.

204
00:15:21.797 --> 00:15:28.013
And then as you're thinking about stuff, you're working with Cloud Code inside of the folder.

205
00:15:29.672 --> 00:15:33.313
and you're researching stuff and then saying, I want you to take notes on it.

206
00:15:34.414 --> 00:15:39.455
In this particular case, you know that a component of your talk is the conclusions section.

207
00:15:39.896 --> 00:15:45.878
And so there is one particular markdown file that you're just going back and forth with it and having it add conclusions.

208
00:15:46.458 --> 00:15:47.978
But what else is in that folder?

209
00:15:48.138 --> 00:15:51.920
So is it like there's a body note, and then there's an intro note?

210
00:15:53.220 --> 00:15:56.802
So one of the big things here is that I'm in thinking mode, not writing mode yet.

211
00:15:57.502 --> 00:16:08.108
And so there's some stuff in here where I've specifically told, I think it's in the front matter, actually, where I've told Claude Code, like, don't help me write anything right now.

212
00:16:08.168 --> 00:16:14.932
And I generally find this to be a big thing with all these models is like they immediately jump to wanting to help you with the.

213
00:16:15.692 --> 00:16:17.033
artifact.

214
00:16:17.333 --> 00:16:24.318
And when you're just in thinking mode, you have to be very explicit in like, hey, I just want you to help me think and ask me questions.

215
00:16:24.378 --> 00:16:28.501
And so, yeah, what you can see here is there's a bunch of files in here.

216
00:16:28.862 --> 00:16:29.542
I've got chats.

217
00:16:29.562 --> 00:16:32.224
That's where I'm literally taking chats I'm having in other things.

218
00:16:32.304 --> 00:16:37.348
And I'm just using the Obsidian Web Clipper to pull the whole chat in.

219
00:16:37.808 --> 00:16:38.869
I've got daily progress.

220
00:16:38.889 --> 00:16:45.393
That's where I'm having the AI actually like look through all the notes that came out that day and like help me think through the progress.

221
00:16:46.234 --> 00:16:47.115
And then I've got research.

222
00:16:47.135 --> 00:16:52.419
That's where I've got a bunch of like articles and PDFs and stuff that I've pulled in so far and been reading about.

223
00:16:52.819 --> 00:16:57.963
And then there's a bunch of other kind of random notes along here where I've been, you know, just using it to kind of help me.

224
00:16:59.243 --> 00:16:59.764
think.

225
00:16:59.964 --> 00:17:04.406
And so, yeah, I was in the midst of one, I've got this conclusion note.

226
00:17:04.426 --> 00:17:12.871
So, you know, I sort of felt like I had blocked out the big kind of the big themes of the talk, but I was like, okay, I need to figure out what am I going to say at the end?

227
00:17:12.951 --> 00:17:18.034
And, you know, essentially what I'm going to say at the end is about a lot of the stuff I've learned over the last few years of working with

228
00:17:18.494 --> 00:17:20.377
these large brands on AI projects.

229
00:17:20.457 --> 00:17:23.662
And so I was starting to get it to the conclusions.

230
00:17:23.722 --> 00:17:28.029
And so, yeah, I'm just kind of like trying, I'm really piecing all this stuff together right now.

231
00:17:28.049 --> 00:17:29.051
That's kind of what's happening.

232
00:17:29.451 --> 00:17:33.538
And give me a sense of like when this folder was empty, what did you start with?

233
00:17:35.013 --> 00:17:44.260
Um, so I think I started with, um, the, the, I started with, uh, telling it like I'm in thinking mode, I'm not in writing mode.

234
00:17:44.921 --> 00:17:50.405
Um, here are my past few talks that I've given a brand to give you a sense of the sort of style that I have.

235
00:17:51.025 --> 00:17:54.788
And I, here's the kind of general idea and the big points I want to make, right?

236
00:17:54.808 --> 00:17:55.789
Like I'm giving away this book.

237
00:17:55.809 --> 00:17:57.770
So I want to talk about simple sabotage field manual.

238
00:17:58.091 --> 00:18:00.693
And I have this notion, like I have this, it's kind of just a title.

239
00:18:00.773 --> 00:18:02.214
It's like transformers are eating the world.

240
00:18:02.234 --> 00:18:03.395
There's this idea that like.

241
00:18:04.035 --> 00:18:11.082
one of the very interesting things happening with these models is they're sort of displacing a whole bunch of specialized code in places.

242
00:18:12.102 --> 00:18:13.664
And so I sort of want to talk about that.

243
00:18:13.684 --> 00:18:14.785
And then I've got these conclusions.

244
00:18:15.465 --> 00:18:20.090
And so the first thing I said was like, hey, just go look through all of the rest of my, you know, probably...

245
00:18:20.610 --> 00:18:45.510
1500 things in my obsidian and go see anything else you can find that um might be of value to this talk um of the existing things i have and so just go kind of pull those in to the research folder at the beginning to kind of like jump start this process got it and you're starting are you starting claude in this folder are you starting it in your full obsidian vault so that it can access all that stuff

246
00:18:47.131 --> 00:18:50.875
Um, no, I'm, so I'm starting it in the full obsidian vault.

247
00:18:50.915 --> 00:18:56.280
So like if we, um, like this is coming, if I step out of here, right.

248
00:18:56.600 --> 00:18:58.162
Um, we're in the root directory.

249
00:18:58.222 --> 00:18:59.583
All this stuff is in the root directory.

250
00:18:59.623 --> 00:19:01.024
My obsidian, I get it.

251
00:19:01.044 --> 00:19:07.470
Um, and my obsidian setup is also like a little more intense for what it's worth because like, um, I've also realized like.

252
00:19:08.441 --> 00:19:15.368
you can add a package.json to add a bunch of custom code commands to your Obsidian that you can then run.

253
00:19:15.428 --> 00:19:19.592
And then you could use those code commands and slash commands and all of these other things.

254
00:19:19.612 --> 00:19:22.495
So there are a bunch of other kind of moving pieces in here.

255
00:19:22.895 --> 00:19:25.298
But generally, it's a fairly straightforward.

256
00:19:25.818 --> 00:19:26.018
Got it.

257
00:19:26.058 --> 00:19:30.463
I mean, I'm trying to use para and some other kind of bits and pieces.

258
00:19:30.942 --> 00:19:35.527
So for people who are listening or watching and are like, we just went through a bunch of stuff really fast.

259
00:19:35.587 --> 00:19:40.893
So the basic gist is Obsidian is just like a note taker, note taking app that runs.

260
00:19:40.954 --> 00:19:41.654
It's all local.

261
00:19:41.955 --> 00:19:47.881
And so everything that all the notes you take, like they exist in essentially text files on your computer organized by folder.

262
00:19:49.263 --> 00:19:55.991
And when you're starting cloud code, one way to do it would be to start cloud code in the folder for the particular project that you have.

263
00:19:56.452 --> 00:20:01.979
But it sounds like what you're doing is instead you're starting in the root directory where all of your obsidian notes live.

264
00:20:02.159 --> 00:20:03.661
And the advantage of that is...

265
00:20:05.960 --> 00:20:13.063
Cloud Code has some sandboxing things where it's not really supposed to run commands outside of the folder it was started in.

266
00:20:13.103 --> 00:20:16.844
It can run commands inside of any subfolder, but it sounds like what you're doing.

267
00:20:16.884 --> 00:20:18.784
So it has access to your entire obsidian.

268
00:20:18.844 --> 00:20:19.645
It can do a bunch of stuff.

269
00:20:19.665 --> 00:20:26.507
And you've also added a package.json, which lets it run custom software commands, basically.

270
00:20:27.288 --> 00:20:29.330
Um, that's really, really interesting.

271
00:20:29.550 --> 00:20:29.850
Okay.

272
00:20:30.551 --> 00:20:36.999
And do you, do you find, cause I've sort of like had this as a twinkle in my eye to like have it go find relevant stuff for me.

273
00:20:37.860 --> 00:20:40.082
Do you find that it's actually relevant and interesting?

274
00:20:40.182 --> 00:20:43.926
Cause I think sometimes when I've done this kind of thing before with language models, there are like,

275
00:20:44.607 --> 00:20:49.071
oh yeah, this random thing is relevant because XYZ, it doesn't feel...

276
00:20:49.692 --> 00:20:59.360
I can understand why it picked it as being relevant, but if it really knew who I am and what I think is interesting, it definitely would not have.

277
00:20:59.720 --> 00:21:03.203
Do you find that that's the case, or have you figured out a way to make it relevant?

278
00:21:03.243 --> 00:21:05.585
I think by and large, yes, I agree with you.

279
00:21:05.705 --> 00:21:11.630
I think in this case, relevance is a little simpler, since ultimately this talk is sort of...

280
00:21:12.651 --> 00:21:17.315
The things I was asking it to look for, I've done a bunch of thinking and research around.

281
00:21:17.355 --> 00:21:22.300
So it's like I'm not asking it to make large conceptual leaps to relevance.

282
00:21:22.360 --> 00:21:23.961
It's like, go find all this.

283
00:21:24.121 --> 00:21:26.343
Like I want to talk about the Simple Sabotage Field Manual.

284
00:21:26.583 --> 00:21:34.130
It can literally just do a like find for all the times, all the articles and things I've got in my Obsidian about that.

285
00:21:34.550 --> 00:21:37.192
And so relevance is...

286
00:21:38.062 --> 00:21:39.923
Yeah, it's kind of a loaded term, right?

287
00:21:40.163 --> 00:21:43.584
And I agree with what you're saying.

288
00:21:44.004 --> 00:21:54.969
I think this is what I'm asking it to do is much more simple, which is like, amongst this set of things, go find all the notes that I've already researched that kind of brought me to be thinking about these things to begin with.

289
00:21:55.342 --> 00:21:55.602
Got it.

290
00:21:55.722 --> 00:22:06.089
And then once you had to do all that research, did you have it do any sort of summary to stimulate you to be like, OK, here are some jumping off points based on what you've done before?

291
00:22:06.109 --> 00:22:07.410
What was your next step once you did that?

292
00:22:07.470 --> 00:22:11.333
No, so my next step is I actually have an agent in here.

293
00:22:11.393 --> 00:22:13.594
So if we go to continue for now.

294
00:22:15.916 --> 00:22:20.481
And so for people who are listening, so you're just starting up Claude, you're using the continue flag.

295
00:22:20.601 --> 00:22:25.187
So you're starting Claude by continuing the last session that you were in.

296
00:22:25.567 --> 00:22:29.432
And now you've got, and Claude Code gives us ability to do subagents.

297
00:22:29.452 --> 00:22:31.974
So those are like little mini Claude's that you can spawn and you have a...

298
00:22:32.615 --> 00:22:33.916
I have a thinking partner one.

299
00:22:33.936 --> 00:22:35.218
You're a thinking partner sub-agent.

300
00:22:35.258 --> 00:22:35.958
Okay, how does that work?

301
00:22:36.018 --> 00:22:45.147
Yeah, and so this is the whole thing where I'm like, hey, you're a collaborative thinking partner specializing in helping people explore complex problems.

302
00:22:45.287 --> 00:22:50.432
Your role is to facilitate thinking and basically don't try to write the thing.

303
00:22:51.652 --> 00:22:58.293
And so after I had that initial set of things, I flipped to this and it's like, okay, let's get into a flow.

304
00:22:58.613 --> 00:22:59.934
Ask me the kinds of questions.

305
00:22:59.954 --> 00:23:00.954
Help me think through it.

306
00:23:01.934 --> 00:23:04.294
You know, this is also where I've got a chats folder in here.

307
00:23:04.334 --> 00:23:06.115
So it's like, it's not just happening here.

308
00:23:06.135 --> 00:23:13.016
I was also having like a, I've got a whole, sorry, I'm just backing out.

309
00:23:13.056 --> 00:23:14.196
So like if we go into.

310
00:23:14.216 --> 00:23:14.616
Yeah.

311
00:23:21.773 --> 00:23:25.858
if we go into chats, like these are a whole bunch of the chats that I had.

312
00:23:26.779 --> 00:23:27.700
Like with the interviewer?

313
00:23:28.543 --> 00:23:35.067
No, these are chats I was having with ChatGPT and Claude and Grok and all of these different things that I went and just grabbed the full transcript of.

314
00:23:36.048 --> 00:23:38.630
So, you know, I was also having all these other conversations.

315
00:23:39.090 --> 00:23:42.792
And then, you know, I'm specifically telling the interviewer, like, review all these other things.

316
00:23:42.812 --> 00:23:50.818
So actually, you know, I think the first one, there's one of these conversations is I originally had this idea about Transformers are eating the world.

317
00:23:50.898 --> 00:23:56.761
And, you know, the sort of notion there is like, you know, there was some research that came out, I think, a few months ago that –

318
00:23:57.902 --> 00:24:06.788
They had found they were able to sort of outperform some specialized time series modeling models with transformers.

319
00:24:06.868 --> 00:24:09.229
And like, you know, I think there's really interesting stuff.

320
00:24:09.349 --> 00:24:14.413
You know, there's a story about Tesla moving 300,000 lines of code with a neural network.

321
00:24:14.453 --> 00:24:16.774
And, you know, I've just kind of got these bits and pieces.

322
00:24:16.854 --> 00:24:26.080
And one of the ways I work generally is like when I have an idea of something to write or think about, I'll start a thread in ChatGPT or Claude.

323
00:24:27.641 --> 00:24:29.522
I'll then save that somewhere.

324
00:24:29.922 --> 00:24:32.323
And then I'll just kind of keep coming back to it when I have more ideas.

325
00:24:32.363 --> 00:24:36.064
It's like, oh, here's another example of Transformers doing something.

326
00:24:36.104 --> 00:24:47.208
And so one of these conversations is actually that thing from probably four or five months ago when that kind of idea initially came into my head, maybe when I saw the research about time series modeling or something.

327
00:24:48.193 --> 00:24:48.973
Really interesting.

328
00:24:49.334 --> 00:24:50.434
Okay.

329
00:24:50.854 --> 00:24:51.935
Okay, so let's keep going.

330
00:24:52.015 --> 00:24:55.937
So you've got the sub-agent.

331
00:24:55.957 --> 00:25:06.002
And I actually wanna like, just actually wanna pause on that real quick, which is, I think this is a very common complaint that they just dive in, and it's a common pattern to make a thinking agent.

332
00:25:06.562 --> 00:25:11.225
And I think Claude Code, or Claude in general, is probably the best one for this.

333
00:25:11.325 --> 00:25:17.668
So this is a thing that we faced with one of the apps that we've incubated called Spiral, which is an agentic ghostwriter.

334
00:25:18.568 --> 00:25:24.932
And I think we found, I found the same kind of thing when, when I was thinking about, okay, how does a good ghostwriter work?

335
00:25:25.932 --> 00:25:28.293
They don't just like, you don't say, Hey, I want you to write a blog post.

336
00:25:28.313 --> 00:25:29.354
And they're just like, cool, I made it.

337
00:25:29.414 --> 00:25:29.814
Here it is.

338
00:25:30.114 --> 00:25:33.096
Like a good ghostwriter is going to get to know you and really understand.

339
00:25:33.116 --> 00:25:35.637
You're going to, you're going to work together to understand.

340
00:25:36.759 --> 00:25:39.780
figure out what's in your head about it, but also shape what's in your head.

341
00:25:39.800 --> 00:25:44.541
Like, it's not just, oh, I can see it and like, they need to get it out of you.

342
00:25:44.561 --> 00:25:46.462
Like you're actually making it together.

343
00:25:46.482 --> 00:25:54.264
And in order to do that, you have to have a really good basic interview process to uncover things.

344
00:25:54.724 --> 00:25:56.605
And that sounds like you've found that too.

345
00:25:56.645 --> 00:26:01.526
And I think that's really, really interesting and really important for people who are thinking about how do I get the best out of AI?

346
00:26:01.826 --> 00:26:02.847
Actually stop for a

347
00:26:04.567 --> 00:26:06.670
Let it ask it to understand you first.

348
00:26:07.030 --> 00:26:07.210
Yeah.

349
00:26:07.731 --> 00:26:17.801
One of the things I say to a lot of people is just like, I think partially because we call it generative, there's entirely too much focus on its ability to write and not enough focus on its ability to read.

350
00:26:18.520 --> 00:26:21.562
You know, it's like its ability to read is incredible.

351
00:26:21.702 --> 00:26:21.922
Right.

352
00:26:22.022 --> 00:26:27.646
And I think, you know, arguably sort of like much more useful on a day to day basis.

353
00:26:27.686 --> 00:26:32.770
Like we produce artifacts far less frequently than we just like think about things.

354
00:26:32.890 --> 00:26:36.052
And so, yeah, I do this a lot.

355
00:26:36.232 --> 00:26:43.097
This is definitely a complaint I have about all the models is like, you know, even when you very specifically tell it not to try to do your work.

356
00:26:43.810 --> 00:26:46.212
it still often still tries to do your work.

357
00:26:46.252 --> 00:26:49.274
And so you have to like really, really be like, no, I said no.

358
00:26:49.794 --> 00:26:57.119
Like, I think actually if we look at, so here, critical.

359
00:26:57.199 --> 00:27:03.663
When Noah says he's just collecting source materials or I do not under any circumstances want you to try to write it, take this literally.

360
00:27:03.763 --> 00:27:07.345
Do not create outlines, drafts, or any versions of talks slash writing.

361
00:27:07.785 --> 00:27:09.947
Only gather and organize the requested materials.

362
00:27:10.936 --> 00:27:11.417
So good.

363
00:27:11.737 --> 00:27:13.039
I love it.

364
00:27:13.539 --> 00:27:14.741
Yeah, this is like...

365
00:27:14.901 --> 00:27:18.204
But yeah, I think we all experience that.

366
00:27:18.244 --> 00:27:22.850
And I mean, I do hope over time that that sort of gets baked into the models.

367
00:27:22.930 --> 00:27:25.973
I think it's a very interesting tension that exists with the model companies.

368
00:27:25.993 --> 00:27:27.074
Because obviously...

369
00:27:27.755 --> 00:27:34.319
you know, like the sort of a lot of the economic input output is sort of measured in the artifacts that it produces.

370
00:27:35.300 --> 00:27:36.701
And so I think it's very oriented.

371
00:27:36.761 --> 00:27:41.924
And, you know, I suspect that part of it is just like that sort of the helpful assistant thing has like,

372
00:27:43.001 --> 00:27:46.546
come to be a sort of meme that is probably self ingested.

373
00:27:47.287 --> 00:27:57.900
Um, but yeah, it's, it's, um, I think for those of us who are trying to do more interesting things with these models, it, it becomes a real barrier to work.

374
00:27:58.341 --> 00:27:58.401
Um,

375
00:28:00.169 --> 00:28:08.818
Okay, so I want to think about when you're using the thinking agent, did you say it like it's, is it outputting some sort of summary of what you've come to into a particular place?

376
00:28:09.159 --> 00:28:18.689
Yeah, so that thinking agent is sort of told to, as it asks me questions, kind of make notes about the questions that it's asking me and keep a kind of running log of...

377
00:28:19.229 --> 00:28:22.150
What I'm uncovering and how I'm thinking about it and all those sorts of things.

378
00:28:22.490 --> 00:28:22.750
Got it.

379
00:28:23.110 --> 00:28:30.933
And then, you know, you come back the next day and you're like, oh, I just want to go down this rabbit hole on XYZ thing about the, you know, this wild bill guy.

380
00:28:32.414 --> 00:28:36.695
And that you started a new chat, maybe with it, maybe with the submission, maybe not.

381
00:28:36.715 --> 00:28:39.056
And that becomes its own new file on that topic.

382
00:28:39.687 --> 00:28:40.347
Yeah, exactly.

383
00:28:40.387 --> 00:28:47.949
So like I, I, there's the, um, the wild bill stuff started as like deep research and chat GPT.

384
00:28:48.669 --> 00:28:52.850
Um, and I, you know, how to go out and I'm reading the wild bill book right now.

385
00:28:52.870 --> 00:28:56.131
There's a, like one sort of particularly famous biography of him.

386
00:28:56.811 --> 00:29:03.553
And, uh, um, you know, I'm kind of thinking about the bits and pieces and trying to make, and I think I made a kind of interesting

387
00:29:09.934 --> 00:29:16.697
he seems to have been after with the OSS and, and, you know, the sort of inspiration for the special forces was like empowering individuals.

388
00:29:17.717 --> 00:29:36.625
You know, that's sort of like the theme of that manual was like obviously empowering citizen saboteurs, but also, you know, I think a big part of the special forces is like, you know, having kind of like incredible operators at the edge who, you know, obviously operate within a sort of command and control hierarchy, but like have a ton of autonomy to, you know,

389
00:29:37.345 --> 00:29:41.987
move and execute independently because they're kind of, they have all the things that they need.

390
00:29:42.027 --> 00:29:54.492
And, um, so, you know, in all of that wild bill research, I kind of went back to this and I was like, is this like an interesting way to connect all these ideas that like maybe kind of there's a, you know, and again, I'm, this is still early.

391
00:29:54.812 --> 00:29:56.933
I have not like solidified these conclusions.

392
00:29:56.953 --> 00:29:59.474
This is like, you know, the regular kind of writing process.

393
00:29:59.534 --> 00:29:59.734
Right.

394
00:30:00.055 --> 00:30:01.055
Um, but it's like,

395
00:30:01.575 --> 00:30:15.263
Well, I think there's this idea that like, you know, fundamentally, well, I know, you know, fundamentally one of the big things with Transformers is it moved us from sequential based models to models that can act in, you know, sort of paralyze their work better.

396
00:30:15.323 --> 00:30:20.506
And that obviously allowed us to have much more powerful and, you know, interesting models and has changed.

397
00:30:20.626 --> 00:30:26.592
you know, arguably kicked off this entire sort of revolution of what's going on and, you know, what we both do for a living.

398
00:30:27.353 --> 00:30:30.276
And so, you know, I think there's this kind of interesting connection.

399
00:30:30.336 --> 00:30:37.042
And so that was what I was playing with, I think, here was like, oh, maybe there's this connection between kind of sequential processing and,

400
00:30:37.683 --> 00:30:39.404
to this kind of parallel.

401
00:30:39.524 --> 00:30:41.584
And then there's this connection to bureaucracy.

402
00:30:41.604 --> 00:30:50.166
And then there's this connection to Wild Bill, who seems to have been very much about sort of like working within a system, but like having autonomy at the edges.

403
00:30:50.507 --> 00:30:54.148
And so that's kind of what I was playing with and just kind of taking notes.

404
00:30:54.168 --> 00:30:57.388
And then, yeah, I would jump out and be like, oh, well, actually, I haven't figured out a conclusion yet.

405
00:30:57.408 --> 00:30:59.269
Let me start the conclusion section and I'll just

406
00:30:59.689 --> 00:31:00.990
sort of get that going on the side.

407
00:31:01.030 --> 00:31:05.993
But, um, and then, you know, I, I have a job, so I can't be doing this all the time.

408
00:31:06.013 --> 00:31:17.119
So it's also like you interrupt yourself and it's really nice to be able to come back and be like, you know, um, can you catch me up on the last three days of research?

409
00:31:17.760 --> 00:31:18.980
Ooh, I love that question.

410
00:31:19.000 --> 00:31:19.681
That's so cool.

411
00:31:22.157 --> 00:31:25.780
And so, yeah, you can just kind of go in and be like, can you catch me up on the last few days of the research?

412
00:31:25.860 --> 00:31:29.002
And it's just going to go read all the stuff.

413
00:31:29.022 --> 00:31:34.286
And again, it's like, I think the point you made earlier about the, you know, go find relevant sources.

414
00:31:34.366 --> 00:31:37.969
It's like, I find a lot that the difference between...

415
00:31:38.950 --> 00:31:45.932
the people are getting a lot of this right now is part of it is just like, you have a good feel for where the edges of the capabilities of these models are.

416
00:31:46.192 --> 00:31:49.393
And you sort of like encourage them to work within those capabilities.

417
00:31:49.433 --> 00:31:53.854
Like this is an incredibly easy, you know, it's like, we know what it's going to do here, right?

418
00:31:53.894 --> 00:31:55.635
Like I could write all these Unix commands.

419
00:31:55.655 --> 00:32:00.496
It's just going to go find a bunch of files in this directory and it's going to look at them by date.

420
00:32:00.636 --> 00:32:05.998
And it's going to look at all the files created in that project over the last year, you know, and we know it's going to be able to do that.

421
00:32:06.018 --> 00:32:07.038
And so it's saying, well,

422
00:32:07.398 --> 00:32:15.404
You know, the major breakthrough day was this idea of bureaucracy as positional encoding, which is very much a work in progress idea, but I kind of like it.

423
00:32:16.284 --> 00:32:24.290
But, you know, so it's just like, it's pretty amazing also to just be able to kind of revisit deep work like this, right, where, you know, you know, you're going to break your flow.

424
00:32:25.230 --> 00:32:34.016
Um, you're not, and it's like, it's often, you know, I find whether it's code or writing, the hardest part is like just picking it up again, um, cause you're, you're out of it.

425
00:32:34.456 --> 00:32:38.319
And so just to kind of like kickstart that process is sort of amazing.

426
00:32:38.719 --> 00:32:45.504
Um, I think what I was playing with here is this idea that, um, bureaucracy was actually like an innovation, right.

427
00:32:45.764 --> 00:32:51.067
Um, that like, we look at bureaucracy as a negative and generally we talk about it as a negative and I think often it is a negative.

428
00:32:51.668 --> 00:32:53.389
Um, but you know, um, yeah.

429
00:32:54.493 --> 00:32:58.536
bureaucracy was a sort of like huge innovation for how companies operate.

430
00:32:58.576 --> 00:32:58.797
Right.

431
00:32:58.837 --> 00:33:09.406
And ultimately it sort of represents kind of hierarchy and structure and a whole bunch of things that are like actually like pretty positive for operating at a large scale.

432
00:33:10.687 --> 00:33:18.554
And so, you know, again, my kind of whole thesis on AI around all this bureaucracy stuff is that what's interesting about it is that.

433
00:33:19.574 --> 00:33:33.979
as opposed to kind of past technologies, which kind of forced you to make a decision about whether you wanted to kind of like use your existing sort of structure and build that technology into your existing structure or adopt the new structure.

434
00:33:33.999 --> 00:33:37.420
And most of the time, the new software required you adopted the new structure.

435
00:33:37.440 --> 00:33:42.101
And that's why so many sort of software projects failed for so long, at least that's sort of part of what I think.

436
00:33:42.701 --> 00:33:47.623
And, you know, I think part of what's interesting about AI, what I find so interesting is like that

437
00:33:49.198 --> 00:33:52.039
you can kind of keep letting everybody work in whatever way they want.

438
00:33:52.339 --> 00:33:58.380
You know, it's like a classic problem inside large companies is like one team wants to use Asana and one wants to use Jira and one wants to use Linear.

439
00:33:58.800 --> 00:33:59.000
Right.

440
00:33:59.100 --> 00:34:07.542
And so then at some point there's like a there's a there's a huge project and they bring in some big consulting firm and they decide they're going to all centralize on this one thing.

441
00:34:08.062 --> 00:34:13.544
And now two thirds of the company is unhappy and like they've all made sacrifices.

442
00:34:13.664 --> 00:34:17.405
And, you know, you're sort of in this like very non-ideal state.

443
00:34:19.205 --> 00:34:25.521
I think was really interesting about AI, and this is a little more sort of theoretical because I think we're so early in this, is that like...

444
00:34:26.722 --> 00:34:30.564
I think it's very possible you could just say, well, everybody just keep doing what you're doing.

445
00:34:30.624 --> 00:34:32.505
We're going to stick sort of some models in the middle.

446
00:34:32.945 --> 00:34:36.526
They don't care what you use because like it's all just data structures to them.

447
00:34:36.986 --> 00:34:39.928
And so we can then have this sort of central thing.

448
00:34:40.288 --> 00:34:46.831
And, you know, when we talked about Percolate at the beginning, Percolate was a content marketing platform, worked with very large companies.

449
00:34:46.851 --> 00:34:48.451
So it was an enterprise software product.

450
00:34:48.491 --> 00:34:55.795
And it's like, you know, at the end of the day, this is sort of the fundamental challenge of enterprise software is about like, you know, adoption and change management.

451
00:34:57.586 --> 00:35:06.470
I think and I hope, and again, this is sort of the optimist in me that like AI kind of lets us just not worry so much about these things.

452
00:35:06.690 --> 00:35:09.752
And rather than trying to make everybody change the ways that they work.

453
00:35:10.347 --> 00:35:19.291
kind of let them work in these ways and let AI sort of, it's my, I call it my Thomas's English muffin theory of AI, which is that it like gets into the nooks and crannies.

454
00:35:20.412 --> 00:35:26.214
And so, yeah, that's, so anyway, but I have no idea what bureaucracy as positional encoding means yet.

455
00:35:26.535 --> 00:35:29.276
I'm hoping I figure it out in the next two weeks before I have to give this talk.

456
00:35:30.138 --> 00:35:33.222
I think, no, but I think the point you just made is totally right.

457
00:35:33.462 --> 00:35:35.545
And it's actually not, it doesn't have to be theoretical.

458
00:35:35.585 --> 00:35:39.950
Like I've been seeing this too inside of every, and I've been meaning to write about it.

459
00:35:40.051 --> 00:35:42.273
And the place that it's been coming up is...

460
00:35:43.263 --> 00:35:49.746
We have, so inside of every, we run like six different products and we have 15 people.

461
00:35:49.826 --> 00:35:53.689
So it's like a crazy product to headcount ratio.

462
00:35:55.209 --> 00:35:58.531
And what's interesting is I really like doing things in a bottom up way.

463
00:35:58.611 --> 00:36:01.192
So everyone, each of the products has its own stack.

464
00:36:01.213 --> 00:36:03.154
We're not like centralized into a particular stack.

465
00:36:04.374 --> 00:36:09.497
Each, you know, GM that runs a product like just has made a decision about do I run Rails or TypeScript or whatever.

466
00:36:10.935 --> 00:36:20.279
And what I'm seeing happen, which is very cool, is a lot of the different products are running into similar things they want to solve for.

467
00:36:20.359 --> 00:36:25.342
So an easy example is we have one product called Sparkle, which is a little bit like a...

468
00:36:27.271 --> 00:36:29.512
a finder replacement or spotlight replacement.

469
00:36:29.532 --> 00:36:34.553
So it, it organizes your files and then it implements really fast spotlight search.

470
00:36:35.233 --> 00:36:35.733
I'm a user.

471
00:36:36.073 --> 00:36:36.473
I like it.

472
00:36:36.773 --> 00:36:36.953
Okay.

473
00:36:37.133 --> 00:36:41.094
So then you, you know, um, and, uh, so that's really cool.

474
00:36:41.474 --> 00:36:43.335
And, uh, agentic search coming soon.

475
00:36:43.795 --> 00:36:44.395
Uh, check it out.

476
00:36:44.775 --> 00:36:48.956
And, uh, and we're just building a new product.

477
00:36:48.976 --> 00:36:49.256
Uh,

478
00:36:49.993 --> 00:36:56.076
new GM, new stack, uh, called para, which is essentially an in house counsel.

479
00:36:56.136 --> 00:36:59.697
So it's short for paralegal, not para, like, you know, Tiago Forte para.

480
00:37:00.297 --> 00:37:11.742
Um, and, uh, the whole job for para is just, you know, take all of your legal files and whenever I have a question and be like, okay, do we ever sign this contract or what's the employee agreement template or whatever?

481
00:37:11.882 --> 00:37:12.863
It just gives you the answer.

482
00:37:12.903 --> 00:37:14.984
And it's just a cloud code sitting on top of a directory.

483
00:37:15.804 --> 00:37:17.565
Um, and.

484
00:37:18.462 --> 00:37:23.607
And a thing that we needed to implement for that is this sort of fast file search.

485
00:37:24.287 --> 00:37:38.620
And what's really interesting is historically, if we wanted to reuse the stuff that we learned from implementing Sparkle's file search, that would have to be abstracted out into this modular library that anyone can use.

486
00:37:38.640 --> 00:37:41.502
And then we have to be on the same platform and all those things, right?

487
00:37:41.522 --> 00:37:42.163
Right.

488
00:37:42.363 --> 00:37:48.446
And what we did instead is we just added Preeti, who's the developer for Power right now.

489
00:37:48.466 --> 00:37:50.147
We just added her to the Sparkle repo.

490
00:37:50.727 --> 00:37:55.409
And I was just like, just ask Cloud Code to figure out how it works and just do your own version.

491
00:37:55.909 --> 00:38:02.773
And so you get this sort of tacit code sharing where we all get better, but without having to do the work of...

492
00:38:03.653 --> 00:38:10.701
abstracting and modularizing everything because the percentage of things that you can do that for are pretty low because it's a heavy lift.

493
00:38:11.262 --> 00:38:21.914
And I'm seeing that happen all the time where just having a bunch of repos that are all solving similar problems, but in different environments, in different ways, everyone gets more productive because AI can kind of translate.

494
00:38:24.055 --> 00:38:33.260
One thing we've done there, we also, so at Alephic, we do a whole bunch of building for very large brands.

495
00:38:33.940 --> 00:38:35.701
And so we sort of build all kinds of AI things.

496
00:38:35.761 --> 00:38:41.544
And, you know, so we've got lots of sort of internal and external repos, and we frequently have the same thing.

497
00:38:41.584 --> 00:38:44.406
And actually, I've used the GitHub MCP platform.

498
00:38:45.286 --> 00:38:51.552
a few times for that same purpose, which is just like, you know, you're just in cursor or cloud code or whatever.

499
00:38:51.572 --> 00:39:00.920
And you're like, hey, can you go like look up, we run it, we've got an internal tool called intelligence that just sort of is a wrapper around a whole bunch of like, you

500
00:39:01.660 --> 00:39:03.241
stuff that we use, right?

501
00:39:03.261 --> 00:39:09.585
So it's like got some CRM stuff and it's just like been a fun place to build the things that we need to run our company.

502
00:39:10.426 --> 00:39:15.569
But it's also a good place to kind of experiment and explore and figure out solutions to interesting problems.

503
00:39:15.610 --> 00:39:20.673
And so I'll frequently be like, oh, can you go to like, just go look at the intelligence repo and then go,

504
00:39:21.093 --> 00:39:28.299
And, you know, look at how I implemented that thing there and take those sort of best practices and just pull them over.

505
00:39:29.219 --> 00:39:35.164
And, yeah, I think that stuff, again, that's where I really do believe in this idea.

506
00:39:35.184 --> 00:39:44.791
I like one of my whenever we have like a client meeting or something, my first question, the icebreaker I always use is what was your aha moment with AI?

507
00:39:45.511 --> 00:39:50.675
And mine was I mean, I don't know.

508
00:39:51.633 --> 00:40:05.749
it was probably not the very first, but it's the one that sort of, I think was most impactful was I was, uh, I got access to build a chat GPT plugin room and plugins came out and like, uh, two and a half years ago or something now, you mean 50 years ago, 50 years ago.

509
00:40:05.769 --> 00:40:05.910
Um, uh,

510
00:40:09.293 --> 00:40:18.038
And, you know, I like you, I've written a lot of software in my life and I, you know, you know what you do when you like get access to something new.

511
00:40:18.078 --> 00:40:21.359
It's like, I'm going to go read the API docs and figure it out.

512
00:40:21.680 --> 00:40:24.541
And, you know, like there's going to be a contract.

513
00:40:24.601 --> 00:40:26.542
And, you know, as long as you follow that contract, it works.

514
00:40:27.323 --> 00:40:35.107
And I go read the plugin spec and it basically is like, oh, you just stick a manifest.json file in the root directory of your application.

515
00:40:35.147 --> 00:40:35.627
And in that...

516
00:40:35.987 --> 00:40:39.690
You describe how you want us to send you data and you describe how you're going to send it back to us.

517
00:40:40.050 --> 00:40:41.171
And then we'll deal with the rest.

518
00:40:41.291 --> 00:40:44.373
And I was just like, that's amazing.

519
00:40:44.853 --> 00:40:45.434
It's also like...

520
00:40:48.190 --> 00:40:49.392
it's how the world should work.

521
00:40:49.732 --> 00:40:51.353
Like I wish everything worked that way.

522
00:40:51.373 --> 00:40:57.359
I wish I didn't always have to adhere to the big company's contract for how to send and receive data.

523
00:40:58.480 --> 00:41:12.593
But also like the thing that really struck me in that moment and it's like been my kind of like rallying cry around all this stuff is that, it's also just like fundamentally counterintuitive in that like I literally have a career's worth of intuition for how to integrate software systems.

524
00:41:13.435 --> 00:41:24.201
And it flipped it on its head, like quite literally 180 degrees away from my intuition of how software systems should be integrated was this thing.

525
00:41:24.942 --> 00:41:33.227
And that I think since then has been my kind of thing for everybody has been like, this is just not intuitive for now.

526
00:41:33.327 --> 00:41:34.648
And that's not a bad thing.

527
00:41:34.688 --> 00:41:35.288
It just means like.

528
00:41:35.870 --> 00:41:37.631
you need to build intuition.

529
00:41:37.951 --> 00:41:41.812
And like, that's what we're all just out there trying to do with it, right?

530
00:41:41.852 --> 00:41:55.575
And so, you know, when, I don't know, I mean, part of what I like about what you're doing and, you know, even just hearing the things you're saying, but like generally what you do with the podcast and what you do with every is like, so much of it is like, we're all kind of just figuring stuff out for the first time, right?

531
00:41:55.655 --> 00:41:59.096
And like, you know, we're like, oh, will this work?

532
00:41:59.196 --> 00:42:04.338
And then like, all of a sudden you have this new bit of intuition for what these things can do and what a computer,

533
00:42:05.735 --> 00:42:07.576
that is not deterministic looks like.

534
00:42:07.696 --> 00:42:12.097
And that's, I think that's just what we're all doing all the time.

535
00:42:12.637 --> 00:42:13.998
And that's why it's so fun, I think.

536
00:42:14.478 --> 00:42:15.598
That's why I love this moment.

537
00:42:15.638 --> 00:42:20.300
Cause like, you just have a weird idea and you're like, has anyone done this before?

538
00:42:20.500 --> 00:42:24.061
And it's like, no, and it's not a complicated idea.

539
00:42:24.081 --> 00:42:26.782
It's just a new, a whole new territory, you know?

540
00:42:28.363 --> 00:42:31.063
I, yes, I think about that all the time.

541
00:42:31.083 --> 00:42:35.225
And I think actually like, I think one of the really damaging sort of things out there

542
00:42:37.979 --> 00:42:42.643
I think there are a lot of people who think we're way further along in this than we are.

543
00:42:43.243 --> 00:42:50.809
And so, you know, I think particularly the people who are sort of scared, you know, we run, we work with like Fortune 50 companies.

544
00:42:51.329 --> 00:42:57.574
And so, you know, when we're sort of like out there and we're talking to people inside the organization, a lot of people feel like they've already been left behind.

545
00:42:58.214 --> 00:43:02.037
And it's like, no, you can like literally go sign into ChatGPT and like,

546
00:43:04.062 --> 00:43:10.246
do something like nobody's thought about doing with this thing yet, because there's just so much white space to explore.

547
00:43:10.346 --> 00:43:16.590
And, and you might discover some totally new way of using it and, or like totally new trick.

548
00:43:16.670 --> 00:43:22.733
And I don't know, that's just, that's, and, you know, I think to be fair to some people, that's sort of very intimidating.

549
00:43:22.774 --> 00:43:31.219
And I don't think by and large, the models do any favors to themselves in helping those people get their feet wet in that, like, you know, I think people go on there and they,

550
00:43:32.068 --> 00:43:36.090
You know, it's like you ask it to write you a poem and then it writes you a poem and you're like, okay, it wrote me a poem.

551
00:43:37.170 --> 00:43:44.273
But I don't know that feeling of like that feeling of, yeah, it's like being on the frontier, right?

552
00:43:45.031 --> 00:43:45.391
Totally.

553
00:43:45.711 --> 00:43:51.334
And yeah, I think that your point about intuitions and getting intuitions is the big thing.

554
00:43:51.514 --> 00:44:01.698
And I think people, what we don't realize is when you're dealing with something fundamentally new, you can't trust how you reason about it without experiencing it.

555
00:44:01.718 --> 00:44:06.700
Because you have to build the intuition in order to be able to reason about what it means and how it fits in and whether it works or not.

556
00:44:07.520 --> 00:44:11.882
And we're just not used to that because we're used to reasoning about things we already have an intuition for.

557
00:44:12.802 --> 00:44:19.608
And, um, and I think that's why, like when you first see, maybe when you first saw chat to be like, Oh my God, it can do everything.

558
00:44:19.628 --> 00:44:20.989
Like we're not going to have jobs in a year.

559
00:44:21.489 --> 00:44:24.451
And now we're like three years in and we're like, yeah, it's awesome.

560
00:44:24.812 --> 00:44:27.614
And jobs are complicated.

561
00:44:27.634 --> 00:44:29.896
There's a lot of complex stuff that we do, you know?

562
00:44:30.636 --> 00:44:33.719
Um, and I, I love, I love kind of that, you know?

563
00:44:35.148 --> 00:45:01.946
in order to build the intuition all you have to do is is use it and that just by using it you're already kind of on the edge uh for now and yeah i think that's the best there's apparently there's a german word called finger spitzing fuel of course there is building fingertip feeling and that's been my um uh just because i can't resist also um i i i'm in that whole in in sort of the realm that you're discussing like uh

564
00:45:03.147 --> 00:45:06.409
I've been trying to do a lot of analogy, analogizing, right?

565
00:45:06.429 --> 00:45:09.370
It's sort of like, and I think, you know, that's really hard.

566
00:45:09.410 --> 00:45:16.294
And, you know, but my two that have sort of stuck the most, one is just, I watch a lot of YouTube with my kids.

567
00:45:16.414 --> 00:45:19.855
And we watch this channel called Veritasium.

568
00:45:19.875 --> 00:45:20.896
It's a science channel.

569
00:45:20.936 --> 00:45:21.836
I love Veritasium.

570
00:45:22.057 --> 00:45:22.497
Yeah, it's great.

571
00:45:23.117 --> 00:45:29.985
And he did one where he built a bike that locks out left if you try to turn right and locks out right if you try to turn left.

572
00:45:30.525 --> 00:45:41.678
And what he's proving is that you can't actually turn a bike left unless you can turn it right, which none of us would think about when we ride a bike because it's all just second nature and intuition.

573
00:45:42.098 --> 00:45:43.459
But it's also why you can't.

574
00:45:44.100 --> 00:45:46.981
explain to a child how to ride a bike.

575
00:45:47.061 --> 00:45:48.781
They just have to get on it and feel it.

576
00:45:49.882 --> 00:45:52.142
And so, you know, I really, that video is amazing.

577
00:45:52.202 --> 00:45:55.063
And it's, that channel is amazing.

578
00:45:55.283 --> 00:45:56.804
And, but I've thought a lot about that.

579
00:45:56.824 --> 00:46:00.945
And then the other one, which is a sort of deeper cut is, um,

580
00:46:01.665 --> 00:46:06.632
There's an amazing book about quantum physics called Beyond Weird by Philip Ball.

581
00:46:07.253 --> 00:46:14.001
And the thesis of the book is basically that there's nothing particularly strange about quantum physics.

582
00:46:14.201 --> 00:46:16.444
We have a very good understanding of it.

583
00:46:16.745 --> 00:46:18.747
We wouldn't be talking right now.

584
00:46:18.787 --> 00:46:19.849
We wouldn't be on computers.

585
00:46:20.249 --> 00:46:25.012
We wouldn't have phones if we didn't have like a very good grasp of the mechanics that exist underneath it.

586
00:46:25.412 --> 00:46:33.738
And his thesis in the book essentially is that like what's really lacking is the vocabulary because like we all exist in a Newtonian world, not in a quantum one.

587
00:46:34.138 --> 00:46:42.143
And so we all have words that reflect the sort of deterministic processes of that macro universe.

588
00:46:42.203 --> 00:46:43.824
And I think a lot about that.

589
00:46:43.904 --> 00:46:48.988
I have not like fully been able to sort of pull that string all the way to AI, but I feel like there's a real...

590
00:46:49.768 --> 00:46:54.492
connection there because I think that there's just something really weird about using probabilistic computers.

591
00:46:54.532 --> 00:46:59.856
Like we're not used to using things that like you ask them the same question twice and they have different answers.

592
00:46:59.876 --> 00:47:01.457
Like that's very strange.

593
00:47:01.477 --> 00:47:09.924
We're not used to, I'm not used to writing code where you can tell the larger company how you want them to send you data and they can just do it.

594
00:47:10.004 --> 00:47:16.048
Like these are not normal things that any of us have lived with in our lifetimes.

595
00:47:16.149 --> 00:47:16.449
And so

596
00:47:19.471 --> 00:47:20.192
I think so too.

597
00:47:20.692 --> 00:47:30.260
And I actually have a hope that language models by becoming a standard way that we use computers will create that vocabulary.

598
00:47:31.601 --> 00:47:39.667
Because we actually are quite good at dealing with probabilistic, non-deterministic things like other humans.

599
00:47:40.088 --> 00:47:42.430
We've just grown up in a world where

600
00:47:43.730 --> 00:47:51.476
because of the Enlightenment and the scientific revolution and the tools that came out of that are very much deterministic.

601
00:47:51.996 --> 00:47:57.920
We've associated that with how we see the world because of those tools and that language.

602
00:47:58.600 --> 00:48:09.968
And there's a whole other part of the way that we see the world, which is much more squishy and much more vibes-based that has been, I think, deprioritized, especially in Western culture.

603
00:48:11.041 --> 00:48:15.385
that now that we have a tool that works that way, I think we'll be able to start seeing that again.

604
00:48:15.485 --> 00:48:19.509
And that's one of the beautiful things to me about language models is it opens up that whole world again.

605
00:48:20.210 --> 00:48:22.011
Yes, I love it.

606
00:48:22.992 --> 00:48:24.534
I do want to go back to Cloud Code.

607
00:48:27.235 --> 00:48:31.680
We should do the phone unless there's unless there are other things that you that you want to share on the computer.

608
00:48:33.062 --> 00:48:40.230
But the thing I want to do before we get there, that's just on my mind right now is like you said, you said, you know, you watch this with your kids.

609
00:48:40.250 --> 00:48:46.517
And I'm sort of curious how like what are your kids think about this and how are you dealing with it with your kids?

610
00:48:46.537 --> 00:48:47.198
Yeah.

611
00:48:47.959 --> 00:48:48.199
Yeah.

612
00:48:49.342 --> 00:48:50.182
I love that question.

613
00:48:50.763 --> 00:48:53.003
So I've got a seven and a 10-year-old.

614
00:48:54.003 --> 00:48:56.784
And obviously, like, I'm pretty kind of deeply embedded in this stuff.

615
00:48:56.824 --> 00:48:59.125
And so I've sort of exposed them quite a bit to it.

616
00:49:01.386 --> 00:49:03.266
You know, they don't...

617
00:49:03.586 --> 00:49:08.308
So they will, like, occasionally use the sort of voice models and they have a pretty good understanding.

618
00:49:08.328 --> 00:49:12.089
And we'll be in the car and just play games and ask questions with Grok and do those kinds of things.

619
00:49:12.989 --> 00:49:15.650
This weekend, actually, for the first time, my 10-year-old...

620
00:49:17.330 --> 00:49:21.011
Uh, she was really eager to be every year.

621
00:49:21.371 --> 00:49:30.573
Um, uh, my wife and her sister and brother and mom and all the cousins that we all get together and we do Christmas together.

622
00:49:30.973 --> 00:49:32.934
And so it's too many presents to give to everybody.

623
00:49:32.974 --> 00:49:36.855
So we do a kind of like not secret secret Santa where everybody chooses one person.

624
00:49:37.255 --> 00:49:40.135
And my 10 year old really wanted to be the one who got to be the chooser.

625
00:49:40.876 --> 00:49:42.356
Um, and, uh,

626
00:49:43.056 --> 00:49:46.919
I encouraged her to vibe code an app to do it.

627
00:49:47.519 --> 00:49:49.821
And so I just gave her my phone and V0.

628
00:49:50.141 --> 00:49:55.405
And honestly, that was like so amazing to watch.

629
00:49:55.525 --> 00:50:00.149
Like not just because it was so cool to see her do that and build it.

630
00:50:00.169 --> 00:50:00.729
And she went through it.

631
00:50:00.749 --> 00:50:01.630
She was having so much fun.

632
00:50:01.670 --> 00:50:05.113
She did 75 revs on V0.

633
00:50:06.514 --> 00:50:07.835
So she like really got it going.

634
00:50:08.055 --> 00:50:10.136
Polished the app.

635
00:50:10.257 --> 00:50:11.117
She also like...

636
00:50:13.199 --> 00:50:19.363
started to get into like really interesting kind of like computer science ideas without knowing it.

637
00:50:19.403 --> 00:50:28.149
So in one of the things like the adults give presence to adults and the kids give presence to kids, but she wanted this to be a more generalized app.

638
00:50:28.209 --> 00:50:31.591
So she realized that like, rather than having adults and kids, you need to call them groups.

639
00:50:32.292 --> 00:50:32.512
Right.

640
00:50:32.672 --> 00:50:40.757
And like, you know, so she's like getting into data modeling and like all of this, I'm like watching this conversation happen and, um,

641
00:50:41.818 --> 00:50:44.301
I just thought that was so awesome.

642
00:50:44.541 --> 00:50:47.925
And, you know, also just like a real pet being of mine right now.

643
00:50:48.045 --> 00:50:58.716
And I've sort of gotten this argument with a bunch of people is like, there seems to be a big conversation that like there's a bubble in vibe coding and like because one company or another might have too high of a valuation.

644
00:50:58.736 --> 00:50:58.816
And.

645
00:51:00.618 --> 00:51:05.042
My take on that is like, I just, I could not care less what the valuations of these companies are.

646
00:51:05.062 --> 00:51:12.908
I think like fundamentally, if there's a tool that can allow a 10 year old to like build an app, there can't, that can't be a bubble.

647
00:51:13.149 --> 00:51:17.492
Like, I just like can't see a possibility where that is that.

648
00:51:17.572 --> 00:51:19.854
So, um, anyway, that's sort of one side of it.

649
00:51:19.874 --> 00:51:24.498
The other big one for me that I've been thinking a lot about is, um, uh,

650
00:51:25.360 --> 00:51:28.664
is sort of media literacy and education, um, stuff.

651
00:51:28.704 --> 00:51:41.016
So, um, you know, both at the sort of schools they go to, um, and then also I went to NYU and I've sort of, um, been having more and more conversations with the Dean of the school I went to there.

652
00:51:41.697 --> 00:51:43.739
Um, and, uh, um,

653
00:51:44.520 --> 00:51:48.142
there's a lot of fear inside schools right now about AI and about cheating.

654
00:51:48.222 --> 00:51:55.407
And, um, there's a big thing, you know, so some parents in my town, they, you know, they wanted to have more of a conversation about it.

655
00:51:55.487 --> 00:51:56.408
And, um, yeah,

656
00:51:57.649 --> 00:52:05.798
You know, as someone who I've thought about this a lot, but I've also just like been, I've spent my entire life thinking about sort of technology and its effects on culture.

657
00:52:05.838 --> 00:52:08.921
And I think I'm like relatively grounded in these things.

658
00:52:08.961 --> 00:52:11.364
I've like, I've put in good hours of thinking.

659
00:52:11.384 --> 00:52:12.305
I mean, I know that for sure.

660
00:52:13.286 --> 00:52:19.773
And so, you know, my take on it is like one that sort of, you can't hide technology that won't be hidden.

661
00:52:20.413 --> 00:52:23.555
So it's like putting our head in the sand is not the best solution.

662
00:52:23.575 --> 00:52:33.420
And then my bigger one though is like, I was out, a friend of mine asked me to come talk to a school two years ago about AI out in LA.

663
00:52:33.500 --> 00:52:38.523
And afterwards I was talking to an English teacher there and she was like, what do I do?

664
00:52:38.823 --> 00:52:41.925
Like, what do I do about all these kids using AI?

665
00:52:41.965 --> 00:52:44.927
And I was like, look, I don't really know what your job is because like, I mean,

666
00:52:45.587 --> 00:52:49.850
being an English teacher for 11th graders sounds really much harder than my job.

667
00:52:50.591 --> 00:52:56.335
But on a really fundamental level, like I don't actually think your job is to teach these kids to write.

668
00:52:56.536 --> 00:52:58.117
Cause that's like a lifelong pursuit.

669
00:52:58.637 --> 00:53:01.379
I think your job is to convince them that it's worth learning to write.

670
00:53:01.900 --> 00:53:05.683
And so in that way, like I I'm not sure that anything fundamentally changes because of AI.

671
00:53:05.743 --> 00:53:13.149
Like I, I think that, you know, and again, this is my very optimistic take, but like I am, I think that I,

672
00:53:13.789 --> 00:53:20.833
there are so many parts of the education system that AI really just exposes the sort of flaws in the way that we teach.

673
00:53:20.853 --> 00:53:37.743
Like, why are there so many tests on these kinds of things instead of encouraging thinking and learning and coming to love to write and research and whatever, you know, it's like, we're so focused on teaching kids the, you know, five paragraph essay, you know, while every adult who is a writer, you know,

674
00:53:38.123 --> 00:53:39.424
has long abandoned that.

675
00:53:39.925 --> 00:53:46.351
And because like, it's all about sort of discovering that you like to write and you, you know, what your own style is and how to do it.

676
00:53:46.411 --> 00:53:54.419
And, you know, it's like, I'm sure it's like, you know, a big part of working with AI to write is like telling it is ignoring it.

677
00:53:54.499 --> 00:53:55.760
Cause you're like, no, that's not me.

678
00:53:55.800 --> 00:53:56.561
I'm not into that.

679
00:53:56.621 --> 00:53:58.203
Like I, like I, I,

680
00:53:58.723 --> 00:54:00.904
I'm totally comfortable saying really here.

681
00:54:01.044 --> 00:54:03.465
I know you don't think it's a good idea, but like I'm cool with it.

682
00:54:04.425 --> 00:54:05.645
And so, you know, I don't know.

683
00:54:05.945 --> 00:54:08.166
I've been sort of having a lot of these different thoughts.

684
00:54:08.186 --> 00:54:17.849
I'm actually pitching a class for the fall of 26 at NYU to kind of the idea for it is code is essay.

685
00:54:18.249 --> 00:54:22.971
And my sort of point of view is like this sort of opens up this new way to express yourself.

686
00:54:24.291 --> 00:54:26.412
And that, you know, like,

687
00:54:27.332 --> 00:54:39.918
We have all these other ways that we celebrate to express ourselves, but code has been like long shut off from people because it's, but actually it's kind of amazing and it lets you express yourself in all these different kinds of ways and like, you know, this is what my 10 year old was doing this weekend.

688
00:54:39.938 --> 00:54:42.700
So anyway, those are all the bits and pieces as a parent.

689
00:54:42.720 --> 00:54:48.463
I've been thinking about the one other thing I will plug is, um, media literacy, I think is a big piece of it.

690
00:54:48.483 --> 00:54:52.625
A lot of people are afraid of these models and hallucinations and, um,

691
00:54:53.423 --> 00:54:59.106
There's a book by a guy named Tim Harford who writes for the FT and he's an economist.

692
00:54:59.846 --> 00:55:05.389
And he has a book called The Truth Detective, which is an adaptation of The Data Detective, but it's written for kids.

693
00:55:05.869 --> 00:55:10.132
And it's the best media literacy book I've ever read for adults or for children.

694
00:55:11.272 --> 00:55:20.357
And I think a lot of what this AI conversation exposes is like how bad a job we do with helping and arming our kids

695
00:55:21.479 --> 00:55:27.367
and our adults to be sort of like truly media and technology literate.

696
00:55:27.427 --> 00:55:34.016
And, you know, like being really good at knowing what's real on social media turns out to be also really useful for like,

697
00:55:34.908 --> 00:55:39.211
differentiated between hallucinations and non-hallucinations in ChadGBD, right?

698
00:55:39.231 --> 00:55:45.634
Like this is a sort of, to me, a kind of very central skill that like we need to arm everybody with.

699
00:55:45.734 --> 00:55:51.758
And I am sort of way more interested in that with my kids than I am in worrying about them cheating.

700
00:55:52.219 --> 00:55:53.459
That's a very long answer to your question.

701
00:55:53.479 --> 00:55:55.120
I don't know if it got at what you were looking for.

702
00:55:55.440 --> 00:55:56.841
No, it's amazing.

703
00:55:57.641 --> 00:55:58.861
That's exactly what I'm looking for.

704
00:55:58.881 --> 00:56:00.842
And it's sort of what it strikes me.

705
00:56:01.902 --> 00:56:11.585
Another way to frame what you're saying is, for example, rather than having them memorize and be quizzed on what the 50 states are,

706
00:56:15.343 --> 00:56:27.234
Asking them to go find the 50 states with chat GPT and be able to tell when the AI is giving them the wrong answer, because that's a lot more of the the meta skill that they're going to need anyway down the down the line.

707
00:56:27.354 --> 00:56:28.935
And again, I'm not a teacher.

708
00:56:29.236 --> 00:56:33.620
I'm sure there's lots of teachers who were like this is that's crazy for a lot of probably good reasons.

709
00:56:33.680 --> 00:56:37.943
But there's something interesting there where it becomes the meta skills become.

710
00:56:38.624 --> 00:56:40.086
more important than they used to be.

711
00:56:40.507 --> 00:56:46.214
And in order to do well at the MetaSchool, you have to also be able to, to some degree, do the underlying skill too.

712
00:56:46.695 --> 00:56:52.703
But we probably could be spending much more time in the MetaSchool than we are now, and the education system isn't really set up to do that anyway.

713
00:56:53.097 --> 00:56:53.337
Yeah.

714
00:56:53.377 --> 00:56:59.045
And I, you know, even I think even simpler examples is like I went to a school called Gallatin at NYU.

715
00:56:59.105 --> 00:57:09.258
And when you graduate, you have to sort of it's not quite a thesis defense, but you have to sort of spend three hours with four professors and or three professors.

716
00:57:09.778 --> 00:57:15.342
sort of explain kind of your line of reasoning around why you studied what you studied.

717
00:57:15.702 --> 00:57:24.489
And you need to be prepared to kind of like weave into that defense any of 25 books that you put on your book list.

718
00:57:24.809 --> 00:57:27.271
And, you know, I was talking to them and it's like amazing.

719
00:57:27.351 --> 00:57:29.412
That's like entirely AI proof, right?

720
00:57:29.472 --> 00:57:30.833
Like there's no cheating on that.

721
00:57:30.873 --> 00:57:34.736
Like you show up in that room and you're either prepared to speak to it or you're not.

722
00:57:34.756 --> 00:57:36.518
And whether you're prepared with AI or not is...

723
00:57:37.678 --> 00:57:38.639
it doesn't mean anything, right?

724
00:57:38.759 --> 00:57:45.525
Like, it's, it's like, like, do you, can you, can you make a, an argument in this room?

725
00:57:45.565 --> 00:57:53.151
And like, not everything is going to be that easy to be sort of like cut off, but like, I don't know, there's something really beautiful about that idea, right?

726
00:57:53.171 --> 00:58:02.118
Because it's like, it's naturally cheating proof because like you're sitting there and, and it's a question of like, did you actually internalize these things?

727
00:58:02.138 --> 00:58:03.219
Um, yeah.

728
00:58:03.919 --> 00:58:11.802
And I don't know, that's way more interesting to me than even like, was your essay good or any of these other, it's like, did you get it?

729
00:58:12.063 --> 00:58:17.665
Like, and so, yeah, I don't know.

730
00:58:19.125 --> 00:58:32.391
I'm trying to do my best to kind of like take a balanced approach and try to at least sort of like tamp down some of, I live in a small town in Connecticut and I think there's a lot of fear amongst parents.

731
00:58:32.903 --> 00:58:36.567
you know, it's like it was mobile phones and then it was social media and now it's AI.

732
00:58:36.587 --> 00:58:38.749
And it's like another thing that's going to ruin our kids.

733
00:58:38.889 --> 00:58:41.952
And, um, I don't think that that is true.

734
00:58:42.393 --> 00:58:52.143
Um, uh, but I think there are things we can and should do to encourage it to not be true, like really get them really good at the things you're saying, like, how do you tell?

735
00:58:52.463 --> 00:58:55.366
And again, it's like a hallucination is just a form of,

736
00:58:56.146 --> 00:59:02.528
the same kind of misinformation that exists in on television and on the internet and in social media and everywhere else.

737
00:59:02.588 --> 00:59:06.389
And, um, you know, just sort of like encouraging people to kind of get in touch.

738
00:59:06.409 --> 00:59:12.051
There's a great part of the truth detective, um, uh, for the kids book, he could, he calls it the brain guard.

739
00:59:12.091 --> 00:59:21.074
And like one of the bits of advice he has is when you encounter some piece of information, if it makes you feel really good, cause you agree with it, then you should be even more skeptical of it.

740
00:59:21.534 --> 00:59:22.634
And he calls it the brain guard.

741
00:59:22.674 --> 00:59:24.955
You know, and he's explaining this, like this is for a nine-year-old.

742
00:59:25.255 --> 00:59:27.515
And I just thought that was like such a beautiful way to put it, right?

743
00:59:27.555 --> 00:59:32.676
Like that's like what, that's what you learn to do when you get good at being on the internet.

744
00:59:33.897 --> 00:59:38.978
Is that you're like, wait, I should like be more skeptical of this because this is like in line with everything.

745
00:59:39.038 --> 00:59:47.359
I think, let me just double check the way, like I, you know, you get to learn that feeling in your gut and get to learn when to react to it.

746
00:59:47.439 --> 00:59:50.220
And so, yeah, that's a lot of how I think about it.

747
00:59:50.730 --> 00:59:51.150
That's great.

748
00:59:51.791 --> 00:59:52.431
I'm interested.

749
00:59:52.451 --> 00:59:53.833
I'm going to get that book.

750
00:59:54.073 --> 00:59:55.194
I'm interested in reading it.

751
00:59:55.774 --> 00:59:56.094
You should.

752
00:59:56.355 --> 00:59:59.437
My plan is to read it to my kids every year from now on.

753
00:59:59.758 --> 00:59:59.978
Nice.

754
01:00:00.278 --> 01:00:00.558
I love it.

755
01:00:00.578 --> 01:00:01.359
Just refresh it.

756
01:00:03.040 --> 01:00:03.381
All right.

757
01:00:03.481 --> 01:00:09.065
Now, for the moment we've all been waiting for, show us how you use Cloud Code on your phone as a second brain note taker.

758
01:00:10.046 --> 01:00:10.566
Okay.

759
01:00:10.727 --> 01:00:11.367
So here we go.

760
01:00:11.928 --> 01:00:16.611
So I am going into an app called Termius.

761
01:00:17.792 --> 01:00:19.454
Termius is just a terminal.

762
01:00:21.069 --> 01:00:26.635
Um, and, uh, what is allowing all this to happen behind the scenes is in my basement.

763
01:00:27.435 --> 01:00:38.446
I have a mini PC and on that mini PC, I have a thing called tail scale running and tail scale lets you set up these very simple VPNs.

764
01:00:38.967 --> 01:00:44.252
So I'm currently inside, like if I scroll down here, you see I'm on a VPN.

765
01:00:44.562 --> 01:00:45.843
That's my personal VPN.

766
01:00:45.903 --> 01:00:47.384
I'm not like on an outside VPN.

767
01:00:47.724 --> 01:00:48.245
I see.

768
01:00:48.325 --> 01:00:51.907
So the only way to access this machine is through my VPN.

769
01:00:52.408 --> 01:00:53.709
Okay.

770
01:00:53.909 --> 01:01:09.821
And so then when I go in there, because I sync my Obsidian with Git, so I put it, it's on GitHub, in a private GitHub, I can then sync it back down to here.

771
01:01:10.261 --> 01:01:12.503
And so then I can just call up Claude.

772
01:01:13.714 --> 01:01:26.598
And now I'm just in Claude code talking and thinking, and I can just be like, what's new in the last two days?

773
01:01:27.258 --> 01:01:28.939
I can access any of my agents.

774
01:01:30.052 --> 01:01:31.333
I can do anything.

775
01:01:31.453 --> 01:01:35.236
And again, this is in my Obsidian, but I can use this anywhere.

776
01:01:35.276 --> 01:01:35.636
Right.

777
01:01:35.656 --> 01:01:37.538
So I'll be like on the fly.

778
01:01:37.558 --> 01:01:38.899
I've got other repos in here.

779
01:01:39.959 --> 01:01:45.203
You know, I realized like a link was broken on my conference site.

780
01:01:45.683 --> 01:01:50.987
And so I I just opened the repo.

781
01:01:51.027 --> 01:01:51.908
I pulled it down.

782
01:01:51.928 --> 01:01:51.968
I.

783
01:01:53.205 --> 01:01:56.026
ask CloudCo to make the changes and I was able to do it right here.

784
01:01:56.886 --> 01:02:08.188
So this has been like completely wild to me because again, this is very much in that like, like on Tuesday of this week, we had Monday off Tuesday.

785
01:02:08.208 --> 01:02:16.050
I dropped the, I dropped the kids off of the bus and then I, I went and I sat and had breakfast and I literally sat on my phone and worked on this talk for like,

786
01:02:16.776 --> 01:02:17.396
two hours.

787
01:02:18.537 --> 01:02:20.737
And I did it through here, right?

788
01:02:20.797 --> 01:02:27.239
Like on my phone where I was like doing real thinking and research and pulling things in and pasting things in and doing all this kinds of stuff.

789
01:02:28.239 --> 01:02:30.620
And, you know, I'm able to do it all.

790
01:02:30.680 --> 01:02:36.222
And it just, it doesn't seem like I could do that kind of thing without that.

791
01:02:36.242 --> 01:02:40.423
So yeah, there's been like a completely revolutionary change in my life.

792
01:02:41.105 --> 01:02:55.972
Um, and, uh, actually one of the things I've been doing lately is like setting up, I've got all these friends now who I've set up, um, like little partitions of this mini PC in my basement, um, so that they can also run cloud code on their phone because, um, I like it so much.

793
01:02:57.573 --> 01:03:04.936
Does this make you be like, oh my God, I got to drop everything and just build a actual purpose-built notes app that has cloud code as a backend for this?

794
01:03:06.803 --> 01:03:13.585
No, I mean, actually, one of the things I've been thinking a lot is like, maybe I just like everything should just run in Linux all the time for me.

795
01:03:13.605 --> 01:03:19.226
Maybe this is like, at least for the short term, this is the answer to all of what I just need to like not have anything anywhere else.

796
01:03:20.827 --> 01:03:25.568
No, I mean, I'm I will say I'm sort of pretty out of the Sass game these days.

797
01:03:25.608 --> 01:03:29.469
So I don't often kind of think that I should drop everything and do anything.

798
01:03:29.649 --> 01:03:31.489
I find this to be a really amazing solution.

799
01:03:32.830 --> 01:03:33.390
And yeah.

800
01:03:34.471 --> 01:03:37.276
But no, I mean, this is really...

801
01:03:38.323 --> 01:03:40.785
this has really like changed the way I work.

802
01:03:40.865 --> 01:03:43.547
And I feel like I can just be anywhere and just be on my phone.

803
01:03:43.607 --> 01:03:47.770
And, you know, I mean, I was out, like I needed a break.

804
01:03:48.891 --> 01:03:51.893
It was, you know, 4.30, I went and sat outside for a while.

805
01:03:52.313 --> 01:04:00.059
And then we had a project that needed to get delivered to a client and a small change needs to be made that like I was the best suited to make that change.

806
01:04:01.280 --> 01:04:05.843
And so I just like hopped on my phone, I pulled the repo down and I went into CloudCode.

807
01:04:05.863 --> 01:04:07.144
It was like a tiny little change

808
01:04:08.665 --> 01:04:15.309
The way I find myself using Cloud Code the most for code is that mostly I'm having it do the work I already know how to do.

809
01:04:15.329 --> 01:04:20.192
I'm like, oh, I knew exactly what was going on in that situation.

810
01:04:21.032 --> 01:04:24.534
I knew why we were having the issue we were having.

811
01:04:24.554 --> 01:04:27.556
And so I was like, I could have gone back to my computer and opened up...

812
01:04:28.296 --> 01:04:45.115
uh cursor and done it in cursor either by hand or with cursor but it's like i just i was like i told cloud code exactly where to look it was like and i first confirmed that the problem was what i thought the problem was and then i just had it push a solution and it pushed a pr and then i was done amazing and i was still sitting outside by the pond

813
01:04:49.050 --> 01:04:49.671
I love that.

814
01:04:50.851 --> 01:04:52.512
Yeah, I've definitely had that experience.

815
01:04:52.532 --> 01:04:54.233
I've never, I've not done it on my phone.

816
01:04:54.293 --> 01:05:00.216
I'm like, I have my laptop out with me, you know, by the pond or by a lake or whatever, but you're inspiring me.

817
01:05:00.296 --> 01:05:02.498
I have a Mac mini in the office that I've been meaning to set up.

818
01:05:02.518 --> 01:05:09.561
So I think this is going to be getting- One of the other, by the way, one of my other, I'll just take you out of here for one sec, actually, just to show you this.

819
01:05:09.962 --> 01:05:16.385
One of my other big ahas recently has been, um, uh,

820
01:05:17.245 --> 01:05:28.768
building Claude code helpers for doing, basically like setup work.

821
01:05:29.248 --> 01:05:36.370
So it's like, I'm not, I'm like been playing with Linux lately, I'm playing with this Omarchi, which is DHH's Linux distribution.

822
01:05:36.431 --> 01:05:38.391
And I'm not like super comfortable in here.

823
01:05:38.791 --> 01:05:44.593
And so I got this whole, this is a Claude code project specifically to help me configure this box.

824
01:05:45.699 --> 01:05:49.222
And it's like so nice because I'm like, oh, how do you do this in Linux again?

825
01:05:49.262 --> 01:05:51.023
Or like, what's the Neo Vim command?

826
01:05:51.043 --> 01:05:52.104
Or like, can you change this?

827
01:05:52.384 --> 01:05:54.546
Can you help me install this plugin for Neo Vim?

828
01:05:54.566 --> 01:05:55.467
Or whatever it is.

829
01:05:55.727 --> 01:06:00.070
And so now I have one on my Mac too, where it's like, can you clean up all the homebrew things?

830
01:06:00.110 --> 01:06:01.712
Or like I switched from like,

831
01:06:02.232 --> 01:06:04.473
which Python package manager I was using.

832
01:06:04.513 --> 01:06:06.254
And it was like, that would have been super overwhelming for me.

833
01:06:06.274 --> 01:06:10.035
And I was like, I want to switch from using pip to using UV.

834
01:06:10.135 --> 01:06:11.375
Can you like just make that happen?

835
01:06:11.415 --> 01:06:12.756
And it like just did all this stuff for me.

836
01:06:13.096 --> 01:06:14.817
And it knows all my preferred settings.

837
01:06:14.877 --> 01:06:25.201
And so I actually have a version of this where like now I've got it so tuned that if I want to launch a new box for doing something, it'll just have all my settings ready to go.

838
01:06:25.261 --> 01:06:31.643
And then I can log into Claude code and the Claude code can then set up anything else that didn't get set up in the initial process.

839
01:06:32.572 --> 01:06:33.212
That is amazing.

840
01:06:33.712 --> 01:06:34.253
That's wild.

841
01:06:36.493 --> 01:06:37.314
This is my happy place.

842
01:06:38.274 --> 01:06:39.034
I can see that.

843
01:06:39.194 --> 01:06:45.496
Are you, do you have any like, are there, are there any big projects like this that you've been itching to do or itching to try out?

844
01:06:47.617 --> 01:06:48.277
No, not really.

845
01:06:48.297 --> 01:06:49.638
I mean, I've been having a ton of fun.

846
01:06:49.678 --> 01:06:51.899
Then my, my server stuff has been a ton of fun.

847
01:06:51.919 --> 01:06:53.139
I've been doing a lot of that.

848
01:06:53.439 --> 01:06:58.301
I'm I'm, I'm, I'm like I said, I mean, this is kind of a joke and kind of not.

849
01:06:58.321 --> 01:07:01.722
It's like, I'm super interested in like,

850
01:07:01.922 --> 01:07:08.406
doing more, because Cloud Code has become such a sort of integral part of my life, like I'm very interested in the command line.

851
01:07:08.446 --> 01:07:13.409
I found myself installing more and more things into the command line and like doing more and more work.

852
01:07:13.429 --> 01:07:22.515
So I've been using like Simon Wilson has a LLM command line tool and like doing more and more stuff sort of in that.

853
01:07:22.595 --> 01:07:25.317
And then like also layering that back into Cloud Code.

854
01:07:25.337 --> 01:07:28.759
So it's like I did my newest Cloud Code project.

855
01:07:29.099 --> 01:07:55.345
little obsidian tool is it i have an attachments folder in obsidian where all the pdfs and images and stuff in any note go but inevitably they have terrible names right i mean um and so this goes through very similar to um uh sparkle and but just in that obsidian folder and it it renames them all and then it also puts them in like a metadata it puts them in a table in the attachments folder

856
01:07:58.806 --> 01:07:59.927
So just like cleans everything up.

857
01:07:59.967 --> 01:08:02.108
It just does it through Gemini flash.

858
01:08:03.448 --> 01:08:06.369
And so it's like, I don't know, stuff like that's kind of amazing.

859
01:08:07.350 --> 01:08:11.932
So I don't, I'm just like, I'm having the time of my life just building and tinkering.

860
01:08:11.992 --> 01:08:15.674
And, you know, I mean, this is just on the side and I get to do the same thing.

861
01:08:15.714 --> 01:08:19.775
I mean, we work with like Amazon and Meta and PayPal and all these big companies.

862
01:08:19.815 --> 01:08:22.957
And, you know, we're just like building amazing stuff all the time.

863
01:08:23.737 --> 01:08:24.177
I love that.

864
01:08:24.237 --> 01:08:24.898
I love the energy.

865
01:08:25.098 --> 01:08:32.023
If people are interested in following you or working with you at Alephic, where should they find you?

866
01:08:32.883 --> 01:08:33.684
Yeah.

867
01:08:33.744 --> 01:08:40.168
So Alephic is Alephic.com, A-L-E-P-H-I-C.com.

868
01:08:40.308 --> 01:08:45.832
And then I also run this thing called Brand, B-R-X-N-D.AI to make it particularly confusing.

869
01:08:45.892 --> 01:08:46.933
That's a conference.

870
01:08:46.953 --> 01:08:49.534
We've got the conference coming up on September 18th in New York City.

871
01:08:49.554 --> 01:08:51.095
You should come if you're around.

872
01:08:51.135 --> 01:08:51.776
It'll be really fun.

873
01:08:51.796 --> 01:08:52.877
We're going to talk about marketing and AI.

874
01:08:52.897 --> 01:08:53.077
Yeah.

875
01:08:53.737 --> 01:09:02.164
And I also write a newsletter there at newsletter.brxnd.ai about sort of specifically at this intersection of AI and marketing.

876
01:09:02.284 --> 01:09:05.026
Those are kind of the best places to find me these days.

877
01:09:05.566 --> 01:09:05.867
Awesome.

878
01:09:06.187 --> 01:09:07.528
Noah, always a pleasure.

879
01:09:08.689 --> 01:09:09.389
Pleasure is all mine.

880
01:09:09.429 --> 01:09:09.950
Thank you, Dan.

881
01:09:18.692 --> 01:09:24.377
Oh my gosh, folks, you absolutely positively have to smash that like button and subscribe to AI and I.

882
01:09:24.597 --> 01:09:24.857
Why?

883
01:09:25.318 --> 01:09:27.480
Because this show is the epitome of awesomeness.

884
01:09:27.760 --> 01:09:29.661
It's like finding a treasure chest in your backyard.

885
01:09:29.942 --> 01:09:35.446
But instead of gold, it's filled with pure unadulterated knowledge bombs about chat GPT.

886
01:09:36.267 --> 01:09:43.674
Every episode is a roller coaster of emotions, insights, and laughter that will leave you on the edge of your seat, craving for more.

887
01:09:43.954 --> 01:09:44.775
It's not just a show.

888
01:09:45.315 --> 01:09:48.938
It's a journey into the future with Dan Shipper as the captain of the spaceship.

889
01:09:50.159 --> 01:09:51.681
So do yourself a favor.

890
01:09:51.941 --> 01:09:55.224
Hit like, smash subscribe, and strap in for the ride of your life.

891
01:09:55.664 --> 01:10:01.009
And now, without any further ado, let me just say, Dan, I'm absolutely hopelessly in love with you.