1
00:00:00,000 --> 00:00:04,920
 We should start with that actually. Android, you were just shitting on

2
00:00:04,920 --> 00:00:09,260
 Android, so how come you're you're now in the business of building an Android

3
00:00:09,260 --> 00:00:14,000
 operating system? Yeah I mean I shit on Android mostly as a joke like I do I do

4
00:00:14,000 --> 00:00:19,800
 like it and I think I use iOS mostly just out of habit and for simplicity you

5
00:00:19,800 --> 00:00:25,840
 know I use the Mac computers so but like I don't know I get FOMO, I shit on

6
00:00:25,840 --> 00:00:32,600
 Android because I get FOMO about you guys with your you know. The control you

7
00:00:32,600 --> 00:00:38,460
 have over your phone is like really irritates me you know. Not everyone is

8
00:00:38,460 --> 00:00:46,460
 ass cucked. Yeah no no no no yeah yeah yeah yeah so I yeah I got a pixel and set

9
00:00:46,460 --> 00:00:51,160
 it up with graphene and everything and I've been slowly learning about it but uh

10
00:00:51,160 --> 00:00:56,400
 yeah I don't know my my approach to the vibe one of the ways I've been sort of

11
00:00:56,400 --> 00:00:59,160
 navigating the whole vibe coding thing and I'm not I don't know if it's a success

12
00:00:59,160 --> 00:01:04,540
 or not is uh I try to like really simplify things like like remove layers of

13
00:01:04,540 --> 00:01:09,860
 abstraction between me and whatever I'm trying to do and uh and so yeah. Going

14
00:01:09,860 --> 00:01:12,660
 down to the bare metal basically. Yeah I mean so yeah going down to the bare metal that's

15
00:01:12,660 --> 00:01:17,820
 like one of my heuristics so like I've had a lot of these projects like I for my you

16
00:01:17,820 --> 00:01:21,700
 know I've sort of moved myself off github onto my own little thing I'm on my own

17
00:01:21,700 --> 00:01:27,780
 little CI system I'm on my own little cloud where I can spawn VMs and this hasn't

18
00:01:27,780 --> 00:01:33,200
 really been a productive thing to do yet because I'm wasting a lot of time and yeah

19
00:01:33,200 --> 00:01:39,060
 it is AI psychotic of me to do these things but uh it's nice to get close to the

20
00:01:39,060 --> 00:01:44,000
 bare metal and have just a few tools that you can really leverage well and and then to

21
00:01:44,000 --> 00:01:47,880
 also to have a system that you can understand you know and so like Android

22
00:01:47,880 --> 00:01:52,160
 and iOS for example I mean at least Android is open source but it's like I

23
00:01:52,160 --> 00:01:54,820
 think there's like a hundred million lines of code you know I tried to build it

24
00:01:54,820 --> 00:01:56,760
 once and the build process took like six hours

25
00:01:56,760 --> 00:02:00,520
 it's just like such a complex thing so for me it's like an engineer

26
00:02:00,520 --> 00:02:05,920
 vibe coder I'm like uh yeah we need a new word I think

27
00:02:05,920 --> 00:02:09,740
 yeah I mean agentic engineering is the closest thing but it sounds so pretentious

28
00:02:09,740 --> 00:02:19,080
 as an agentic engineer you know it's it's not but yeah so like uh I don't know I I uh well

29
00:02:19,080 --> 00:02:23,880
 there's two two things like I remember like five years ago uh when I first it was me six years

30
00:02:23,880 --> 00:02:27,920
 ago when I first got like a spiral grant I was trying to convince Steve Lee that a

31
00:02:27,920 --> 00:02:33,720
 block should make a phone actually regarding your spiral grant I just read through that all

32
00:02:33,720 --> 00:02:39,460
 of it uh recently and you're listed as Justin's son which made me crack up yeah yeah they they always

33
00:02:39,460 --> 00:02:44,500
 talk to me as mentioned me as Justin's son for some reason I was just far to that yeah I won't go

34
00:02:44,500 --> 00:02:51,500
 there but uh yeah I can't speak too earnestly here Justin Moon Justin Sun for a second I thought we

35
00:02:51,500 --> 00:02:59,000
 were having a conversation not uh the entire 10 people the entire 10 people that like and and zap my

36
00:02:59,000 --> 00:03:04,840
 GM and GN posts we we have at least five listeners yeah like I I made everyone say like who actually

37
00:03:04,840 --> 00:03:10,660
 listens to no solutions that like five people put their hands up so yeah given given that we don't

38
00:03:10,660 --> 00:03:15,400
 actually listen to these back ourselves it stays with five probably for the foreseeable future yeah

39
00:03:15,400 --> 00:03:24,220
 so uh but yeah so I I uh so that's one thing I remember trying to get Steve Lee to make uh

40
00:03:24,220 --> 00:03:27,860
 like an android phone that like put a lightning note at the operating system level

41
00:03:27,860 --> 00:03:33,120
 like I thought that would have been cool back in the day you you you tried to put the note in the

42
00:03:33,120 --> 00:03:38,100
 phone before it was cool yeah I think you know and we still are struggling with this right like

43
00:03:38,100 --> 00:03:43,960
 everyone you know every you know you have lightning apps and in a sense like your phone or ideally

44
00:03:43,960 --> 00:03:49,420
 like you have like one server somewhere yeah that like everyone has way more lightning waltz than

45
00:03:49,420 --> 00:03:53,700
 they wish they had like I have oh absolutely so many and so it's still a non-self problem and having

46
00:03:53,700 --> 00:03:58,260
 one that runs in your phone or something would be kind of neat and then it makes it much easier for

47
00:03:58,260 --> 00:04:03,220
 other apps to integrate right like you know that's one theme of sovereign engineering demos is like

48
00:04:03,220 --> 00:04:07,320
 you know okay now this app has a cashew wallet which makes it kind of unusable because now you have

49
00:04:07,320 --> 00:04:14,000
 to load tokens into it yeah yeah sure you know so it's like it's like everything not only do you

50
00:04:14,000 --> 00:04:19,000
 have to have bitcoin but you have to get this type of bitcoin in yeah yeah yeah yeah I mean cashew I

51
00:04:19,000 --> 00:04:24,380
 still think cashew is fantastic but it has a different set of well trade-offs and ux hurdles

52
00:04:24,380 --> 00:04:28,920
 I'm complimenting cashew because it's the easiest to integrate right so it's like on that front it's

53
00:04:28,920 --> 00:04:33,220
 funny but it's also you know it's it would be nice to make the problem what you're describing is

54
00:04:33,220 --> 00:04:38,320
 basically what greenlight did but in the phone put it on the phone like like have a have a have a light

55
00:04:38,320 --> 00:04:43,880
 node and every wallet plugs into the node that's directly on the phone and every app becomes a

56
00:04:43,880 --> 00:04:48,660
 wallet right like look at how much like one of the things that primal did really well is that they

57
00:04:48,660 --> 00:04:53,420
 built the phone the wallet into the into the app and yeah made it really clean and look at how much

58
00:04:53,420 --> 00:04:57,580
 yeah yeah they're still working on this right oh absolutely now they're just pulling away their

59
00:04:57,580 --> 00:05:01,720
 custodial system and to be fair it's all kind of trending in that direction also with nostal wallet

60
00:05:01,720 --> 00:05:04,820
 connect and those kind of things where you can just connect things together and it doesn't really

61
00:05:04,820 --> 00:05:10,980
 matter yeah where the wallet runs in some sense yeah but even me it's like i'm i'm i'm not a power user we

62
00:05:10,980 --> 00:05:13,760
 talked about this before so with many of these things i don't i just refuse to

63
00:05:13,760 --> 00:05:20,320
 set it up like i'm just not gonna do it but yeah so that was one one to your point i am a power user

64
00:05:20,320 --> 00:05:25,940
 yeah this is one of our yeah okay yeah so that's that's one angles like i've had this idea for a

65
00:05:25,940 --> 00:05:31,620
 long time i mean that's also you know it's like it would just be so cool if if if every like this

66
00:05:31,620 --> 00:05:36,600
 is so yeah the other angle is like i think our first one of these we did i was talking about this

67
00:05:36,600 --> 00:05:41,120
 idea of like these unix you know the unix tools of nostra right like a bunch of these little apps

68
00:05:41,120 --> 00:05:45,140
 and i pursued that and i made this like nice platform where to write the code wouldn't have

69
00:05:45,140 --> 00:05:51,980
 been too hard but to maintain like to actually ship it it's like so it was rust multi-platform

70
00:05:51,980 --> 00:05:56,120
 right which i have now shipped or i've kind of shipped it and there's like six or seven people

71
00:05:56,120 --> 00:06:00,000
 building on this right now um everyone loves it like there's not a single person who's tried it

72
00:06:00,000 --> 00:06:06,980
 hasn't liked it uh and so like code wise more people are building on on uh rust multi-platform

73
00:06:06,980 --> 00:06:12,560
 that are listening to no solutions probably probably yeah yeah yeah exactly uh better looking people

74
00:06:12,560 --> 00:06:18,480
 more popular also more rich more rich better quads

75
00:06:18,480 --> 00:06:26,840
 more sculpted quads we're sculpting we're sculpting right now everything yeah yeah yeah everything you'd

76
00:06:26,840 --> 00:06:31,000
 really want in a person i was gonna say in a man just assuming that there's no women here but

77
00:06:31,000 --> 00:06:36,900
 there may be a female we've got one we've got one yeah we kind of we 2x our women participation

78
00:06:36,900 --> 00:06:42,720
 with this core yeah yeah exactly yeah it's a lot of big things are happening but uh but yeah so i was

79
00:06:42,720 --> 00:06:47,040
 so so i built this thing where i thought i could have like the little unix tools and i could at the

80
00:06:47,040 --> 00:06:51,340
 code level but just like i don't know the maintenance and distribution of it and having to deal with

81
00:06:51,340 --> 00:06:55,840
 ios and play store and all this stuff just made it not i just didn't want to do it you know like

82
00:06:55,840 --> 00:07:01,120
 it's not i just hate that um yeah to tag on to that i think that's the that's the biggest pain

83
00:07:01,120 --> 00:07:05,840
 point is the the permissioned parts of the system yeah right and it's not like it's too much work

84
00:07:05,840 --> 00:07:12,580
 it's just like i just i i just don't want to live that life you know like and it's also a lot of

85
00:07:12,580 --> 00:07:17,900
 work like getting through the ios review process is hell and google play is much better i was building

86
00:07:17,900 --> 00:07:21,700
 this at peak and i was super excited about it and i was trying to set up and i it's like i've been

87
00:07:21,700 --> 00:07:26,420
 for a month i've been trying to get this play store i had to set up an llc yeah that cost money

88
00:07:26,420 --> 00:07:30,440
 all the stuff and finally i had to get a duns number all stuff and i'm still now i'm

89
00:07:30,440 --> 00:07:34,800
 like they can't send me a text message it takes a month or two yeah and i'm still there and i'm

90
00:07:34,800 --> 00:07:40,400
 like so that was uh so i don't know so i've had this dream of i'm like you know one way of doing

91
00:07:40,400 --> 00:07:45,660
 it is just distributing it all through zap store um which i played around with some and i think that's

92
00:07:45,660 --> 00:07:51,600
 a really great yeah step in that direction but uh but yeah so that's kind of where another of the

93
00:07:51,600 --> 00:07:57,180
 so yeah for the listeners who haven't given up by now i started building like a little android

94
00:07:57,180 --> 00:08:04,380
 like uh i started building like a mobile operating system and and you didn't call it moon os to no

95
00:08:04,380 --> 00:08:08,400
 no no not yet i just called it shadow like for my vibe coding projects what i do is when it comes

96
00:08:08,400 --> 00:08:12,600
 time to name and i just look around the room and there's something i see it so i saw a shadow

97
00:08:12,600 --> 00:08:19,500
 all right so i know you got inspired all right yeah and so see i started doing this and uh part of it

98
00:08:19,500 --> 00:08:24,800
 is like i don't know uh i'm just scratching my own curiosity so like in the second cohort last year

99
00:08:24,800 --> 00:08:30,880
 we built a browser yeah yeah so that was like one of the things

100
00:08:30,880 --> 00:08:35,940
 excuse me that was one of the things that was like the final thing i did

101
00:08:35,940 --> 00:08:41,340
 and you were kind enough to release the recording on that one six months later six months later yeah

102
00:08:41,340 --> 00:08:49,080
 yeah um you need to keep people on the tip of their toes yeah yeah yeah yeah now they think browsers

103
00:08:49,080 --> 00:08:55,180
 are relevant but browsers are very irrelevant right now no mobile operating systems are relevant no so

104
00:08:55,180 --> 00:09:00,540
 yeah so that it was kind of a step in that direction it's like okay how could we have a way of building

105
00:09:00,540 --> 00:09:04,880
 an app that is you can distribute it with less permission and i was experimenting with some of the

106
00:09:04,880 --> 00:09:09,580
 ideas in that cohort about like you know not needing a dns name to load it because that's the one thing

107
00:09:09,580 --> 00:09:13,300
 that's permissioned about the web is you have to go buy this stupid name from some central

108
00:09:13,300 --> 00:09:18,280
 yeah authority and it's just kind of dumb and so i had a way where i could like load you know any

109
00:09:18,280 --> 00:09:22,700
 number of people could claim a different name and if there was conflicts you'd just see a little

110
00:09:22,700 --> 00:09:27,800
 picker like hey here are the people that claim the name here's who you don't resolve it perfectly you

111
00:09:27,800 --> 00:09:32,380
 resolve it in a web of trust yeah exactly so and so i had that and that was kind of cool

112
00:09:32,380 --> 00:09:38,220
 and uh and then so after i left the cohort i'm like i've been experimenting with just like how

113
00:09:38,220 --> 00:09:43,720
 how far can i raise my sights you know like how can i like these two we have these new superpowers

114
00:09:43,720 --> 00:09:48,900
 right so we shouldn't we should set our sights really high i think right we should try to like

115
00:09:48,900 --> 00:09:52,600
 really what's the biggest project push the needle yeah exactly so that was like one of the themes of

116
00:09:52,600 --> 00:09:57,400
 my when i was the captain last week is like how can we slay some dragons like let's try to slay

117
00:09:57,400 --> 00:10:03,040
 dragons that's why phips is cool it's like it's trying to slay dragons yeah uh and so yeah i've been

118
00:10:03,040 --> 00:10:06,360
 trying to think very deeply about this like what's and i'll you know also i vibe code stuff and i just

119
00:10:06,360 --> 00:10:09,700
 don't really want to maintain it right so it's like where can i find a project that i do want to

120
00:10:09,700 --> 00:10:15,000
 maintain for like five years yeah yeah okay where what's big enough that you so that you don't lose

121
00:10:15,000 --> 00:10:18,800
 interest after like three months well and all yeah not just interest but it's like rationally i'm like

122
00:10:18,800 --> 00:10:23,020
 yeah this is going to move the needle right like yeah i still feel like you know i've done a lot of

123
00:10:23,020 --> 00:10:26,540
 stuff in this bitcoin freedom tech stuff and i don't know if it ever really moves the needle

124
00:10:26,540 --> 00:10:31,580
 it's very frustrating to work really hard on something and not really know you know it's like

125
00:10:31,580 --> 00:10:36,300
 i think your impact is way bigger than you think i know but i i want to not have

126
00:10:36,300 --> 00:10:41,720
 i don't want to have the doubt right like you're changing the life of like five people right now

127
00:10:41,720 --> 00:10:44,940
 i know four i'm pretty sure one just dropped off

128
00:10:44,940 --> 00:10:49,140
 because i keep sidetracking you so people just quit

129
00:10:49,140 --> 00:10:54,040
 okay let's make it three that's so amazing you changed you changed the lives of three people

130
00:10:54,040 --> 00:10:55,860
 three people two now it's down to two

131
00:10:55,860 --> 00:11:01,540
 no but seriously i think i think i talked with eric case and actually about this once

132
00:11:01,540 --> 00:11:07,880
 and i i just told him uh because lots of people read his stuff because it's way out there and you

133
00:11:07,880 --> 00:11:12,180
 immediately agree and he doesn't feel that way and i told him like you're probably blowing someone's

134
00:11:12,180 --> 00:11:17,880
 mind right now and yeah and he just was like he never considered that yeah it really changed his

135
00:11:17,880 --> 00:11:22,420
 perspective and i think for for your stuff it's very similar you know like yeah but it's like i i don't

136
00:11:22,420 --> 00:11:26,760
 know i i feel like it could i can imagine it being like a hundred or a thousand times more

137
00:11:26,760 --> 00:11:31,340
 like yeah fair enough okay and so i don't know it's like that's possible you're not raising trillions

138
00:11:31,340 --> 00:11:35,040
 of dollars like oh may i does right but i don't know like that's not i don't know if that's

139
00:11:35,040 --> 00:11:40,320
 how you really make a difference either but you know like i don't know if i could do more it would

140
00:11:40,320 --> 00:11:44,300
 be nice to do more i actually want to loop back to what you said before like you've just been

141
00:11:44,300 --> 00:11:47,920
 wasting your time i don't agree by the way because you never know in the final analysis

142
00:11:47,920 --> 00:11:52,340
 what is actually a waste of time and what isn't yeah so that's so this is one another one of my

143
00:11:52,340 --> 00:11:58,140
 things is like usually i just get really results oriented i just like want to like accomplish

144
00:11:58,140 --> 00:12:02,200
 something yeah and so like after leaving fetti and doing these sovereign engineerings i've just been

145
00:12:02,200 --> 00:12:06,840
 more like playing playing around with stuff and playing and playing and playing to be in like and i

146
00:12:06,840 --> 00:12:10,780
 kind of tried to ship something yeah like a month ago pika and then i hit like some little

147
00:12:10,780 --> 00:12:14,840
 difficulties and i just immediately gave up and i gave myself permission to do that and like

148
00:12:14,840 --> 00:12:20,420
 whatever and now i'm playing again but so so yeah like the the mobile thing is interesting because

149
00:12:20,420 --> 00:12:24,960
 it's the computer interface we all use right like that's how we interact with the thing and that

150
00:12:24,960 --> 00:12:29,140
 was one of my questions in my captain week is like well is this going to stay the case right

151
00:12:29,140 --> 00:12:35,700
 if it remains mobile phones and chat interfaces with ai it's like pika and shadow the mobile os are

152
00:12:35,700 --> 00:12:40,760
 good bets yeah because these are attempts to make good nostril versions of that but if it ends up

153
00:12:40,760 --> 00:12:44,820
 being voice like you kind of interact with your ais more with voice it's like then now maybe

154
00:12:44,840 --> 00:12:51,200
 these are these are less good ideas maybe i still think you need like okay we're still gonna need

155
00:12:51,200 --> 00:12:56,540
 servers we're still gonna need proper compute machines like lots of people are getting mac minis now so

156
00:12:56,540 --> 00:13:01,360
 this is another one we're still gonna we're still gonna need phones as because you're gonna need a

157
00:13:01,360 --> 00:13:06,840
 visual interface for certain things but but to tag on to what you just said i think like i'm i'm gonna

158
00:13:06,840 --> 00:13:12,860
 get the the pebble watch ring you know like the pebble ring yeah because all it has is like it takes

159
00:13:12,860 --> 00:13:17,660
 your vitals and it has a microphone yeah and it's for whatever reason i don't know why but it's like

160
00:13:17,660 --> 00:13:22,940
 okay it's it's awesome to take voice notes like they build it for that and my whole agentic system

161
00:13:22,940 --> 00:13:29,380
 builds upon voice notes so that's gonna i think that's gonna be my main computer input system and i think

162
00:13:29,380 --> 00:13:34,600
 what's what has happened over the last decades is like everyone had a big desktop computer right

163
00:13:34,600 --> 00:13:40,160
 yeah and and i think everyone knows the video where just and everyone had a globe and everyone

164
00:13:40,160 --> 00:13:44,520
 had maps and everyone had those those kind of things and it all went into the internet and dematerialized

165
00:13:44,520 --> 00:13:49,320
 and i think we're in the process now of the computer itself dematerializing yeah and we just have

166
00:13:49,320 --> 00:13:53,120
 certain input systems and certain output systems and certain kind of like your brain is going to be

167
00:13:53,120 --> 00:13:57,980
 your mac mini or whatever at home you know or somewhere in the cloud or what have you and i think

168
00:13:57,980 --> 00:14:06,300
 like i think i think you're onto something in terms of for every single layer of compute yeah you need

169
00:14:06,300 --> 00:14:10,620
 to have a permissionless alternative yeah otherwise you're just going to be yeah and we have that with

170
00:14:10,620 --> 00:14:15,900
 linux desktop right like exactly the story on linux desktop is much better than on mobile yeah

171
00:14:15,900 --> 00:14:21,820
 absolutely i mean you can you can install whatever you can and you can really customize how it feels

172
00:14:21,820 --> 00:14:26,480
 and how it looks and like especially with vibe coding now you can build oh i mean the rising is a whole

173
00:14:26,480 --> 00:14:31,460
 yeah thing yeah just you rice your yeah and i've done that before and like i really enjoyed it at

174
00:14:31,460 --> 00:14:35,700
 times and then i end up back on a mac and i'll go back to it i'll go back to it again someday right

175
00:14:35,700 --> 00:14:41,820
 like i think it's really i don't know i think that's i think it's really cool um and even on a mac you

176
00:14:41,820 --> 00:14:48,540
 have you have more control over it than you do on an iphone for example yeah but you can install a

177
00:14:48,540 --> 00:14:52,260
 custom window manager like you buy yeah things like there's there you have a lot more control

178
00:14:52,260 --> 00:14:55,640
 you can install homebrew you can install all kinds of little snitch all these things right

179
00:14:55,640 --> 00:15:02,180
 so even on desktop across the board yeah i mean and and on windows you can install wsl right like

180
00:15:02,180 --> 00:15:08,960
 yeah why do you think that is why do you think the phones got into the like let's fully call the

181
00:15:08,960 --> 00:15:19,460
 that's a really good question um i mean one one is security right they're more secure

182
00:15:19,460 --> 00:15:26,220
 supposedly i'm not sure if that's true yeah um if you're a power i mean

183
00:15:26,220 --> 00:15:32,280
 that's the thing about security it's like if you if you i made this i made this comparison in the past

184
00:15:32,280 --> 00:15:37,580
 i kind of hate to make it but i can't think of a better one but a weapon uh you know like a gun

185
00:15:37,580 --> 00:15:44,780
 yeah like it can only be so secure otherwise you'll nerf it to the point so that it has no use anymore

186
00:15:44,780 --> 00:15:50,540
 yeah same for a car or what have you you know like a car needs to be powerful enough so that you can

187
00:15:50,540 --> 00:15:58,140
 run over things basically yeah and if it if it won't let you do that then it's not a useful tool

188
00:15:58,140 --> 00:16:02,300
 anymore basically so so so i kind of disagree with the security angle because

189
00:16:02,300 --> 00:16:08,820
 you know guns are secure as well if you know how to use them cars as well yeah some sense and it's

190
00:16:08,820 --> 00:16:15,880
 also i think it's also okay for powerful tools to well be able to do damage i agree with that i mean

191
00:16:15,880 --> 00:16:21,220
 it's just like i mean they're more sandbox it's also it's like it's simpler right like they i think

192
00:16:21,220 --> 00:16:26,120
 with mobile phones they sort of like like i was trying to figure out how to like find a file on my

193
00:16:26,120 --> 00:16:32,540
 iphone the other day you just gave up yeah like i downloaded an account a summary from river and i

194
00:16:32,540 --> 00:16:39,820
 really couldn't figure out like if i actually succeeded in the download or not like really hard

195
00:16:39,820 --> 00:16:43,100
 to tell so you know they hide so much from you so that's i don't know that's one thing one thing

196
00:16:43,100 --> 00:16:48,560
 is it's harder to rice a phone what do you do on your little thing oh like now that's the ios

197
00:16:48,560 --> 00:16:54,020
 user speaking on android you have custom launchers and everything like it's sure it's you can do a

198
00:16:54,020 --> 00:16:58,380
 okay what's your answer then fucking smart ass like what no i don't have an answer like that's

199
00:16:58,380 --> 00:17:03,660
 why i asked you like i i don't know why but it's unequivocally true like yeah i don't i don't know

200
00:17:03,660 --> 00:17:07,720
 why we still have so many freedoms on the desktop side and so little freedoms on the mobile side

201
00:17:07,720 --> 00:17:13,580
 mm-hmm i don't know i'm not sure if there's a grand conspiracy there usually isn't yeah it usually

202
00:17:13,580 --> 00:17:18,460
 is just okay things happened and it evolved that way i mean one one is that uh desktop is more

203
00:17:18,460 --> 00:17:24,700
 server-like because like most of the like where linux one and openbsd one was servers yeah and so in a

204
00:17:24,700 --> 00:17:30,860
 sense desktop is a little more server-like so maybe that i don't know it's we kind of inherit that a

205
00:17:30,860 --> 00:17:38,360
 little bit yeah and so what you want to do now with shadow is to kind of start from scratch and

206
00:17:38,360 --> 00:17:44,380
 rethink how an operating system works to some degree yeah so i'm starting from so the idea is like i so

207
00:17:44,380 --> 00:17:49,160
 yeah after the last cohort i went on a big long road trip and as i was going along i was having

208
00:17:49,160 --> 00:17:52,520
 chad gpt teach me about these things because i got interested so i just do voice mode in the car

209
00:17:52,520 --> 00:17:57,800
 yeah yeah and so i would i would in pro i would leave it a voice big long voice memo and then

210
00:17:57,800 --> 00:18:02,300
 like 10 minutes later i get an answer and i just play yeah that's awesome right hit play right

211
00:18:02,300 --> 00:18:06,620
 yeah and so i was you know i went across like south to north i went from texas to

212
00:18:06,620 --> 00:18:10,160
 color uh whatever what was it called canada

213
00:18:10,160 --> 00:18:13,060
 colorado close enough

214
00:18:13,060 --> 00:18:22,140
 yeah canada um and uh and so yeah i started learning a lot about like how a mobile phone

215
00:18:22,140 --> 00:18:27,720
 works like how how does pixel you know what are the drivers like you know what does how do you draw

216
00:18:27,720 --> 00:18:32,740
 to the screen how do you yeah all these things and then when that was done i tried to vibe code

217
00:18:32,740 --> 00:18:36,580
 this and i was like five months ago and it was really hard work like

218
00:18:36,580 --> 00:18:42,040
 i was having to engineer it and i had you know the lms for some level of autocomplete but they

219
00:18:42,040 --> 00:18:47,540
 couldn't really do it on their own uh and so the idea was like two two folds attack from two angles

220
00:18:47,540 --> 00:18:54,140
 one is try to figure out how to boot an android phone without any with as little android stuff as

221
00:18:54,140 --> 00:18:59,300
 possible yeah right so like no jvm for example jvm is the java is how the whole user space

222
00:18:59,300 --> 00:19:03,520
 exists but you know there's stuff in the kernel like i'm going to build on top of linux i'm not

223
00:19:03,520 --> 00:19:07,600
 going to try to reimplement linux i'm i'm not like that's to me a black box that i can't

224
00:19:07,600 --> 00:19:14,300
 i can't probably reinvent intelligently but uh so i'm not going to try and then like for example

225
00:19:14,300 --> 00:19:17,840
 device drivers i do not want to spend time trying to figure out how to write a device driver

226
00:19:17,840 --> 00:19:23,280
 i think you can solo right like yeah it's yeah i mean it's kind of too big of a task there's too many

227
00:19:23,280 --> 00:19:26,800
 and there's too little upside right so it's okay well i'll take those for granted

228
00:19:26,800 --> 00:19:32,460
 and then how could i boot a phone with just those things yeah and so i would i started you know i

229
00:19:32,460 --> 00:19:37,880
 would set it up so like my code my rust code at the time it was zig i was a younger

230
00:19:37,880 --> 00:19:49,120
 less practical man less experienced agentic engineer yes yes yes um yeah less experienced agentic

231
00:19:49,120 --> 00:19:55,240
 engineer oh we're fucked here so but yeah so now it's rust and so the idea is like okay i would my

232
00:19:55,240 --> 00:20:01,840
 code would be the first program in user space like pid1 yeah and then i would intercept that

233
00:20:01,840 --> 00:20:08,260
 i'd log something and just punt it back to the android one and then so i i i started doing

234
00:20:08,260 --> 00:20:14,140
 more and more like could i mount a drive could i you know could i could i paint to the screen

235
00:20:14,140 --> 00:20:20,520
 like i would paint the screen one color and then wait five seconds and then continue the boot

236
00:20:20,520 --> 00:20:25,840
 i saw i saw you upgrading to painting triangles last week i was painting all kinds of shapes yeah i mean

237
00:20:25,840 --> 00:20:30,120
 you should see what it is now oh shit now i have that i missed like six days or something yeah big

238
00:20:30,120 --> 00:20:35,340
 things happened last week last when you last saw it took 13 seconds to click a button yes i know i

239
00:20:35,340 --> 00:20:40,580
 remember that we all thought we all thought that demo failed but yeah it was just a little laggy no no no

240
00:20:40,580 --> 00:20:47,740
 no no no yeah yeah we had humble origins humble origins but yeah so uh where the hell was i um

241
00:20:47,740 --> 00:20:52,620
 painting screens yeah yeah so so that was one angle to attack it from trying to boot it

242
00:20:52,620 --> 00:20:58,360
 the problem is like it's kind of a boring thing to work on you know at a certain point so then the

243
00:20:58,360 --> 00:21:04,820
 other thing i started to do which i kind of had more luck with was i booted in android and then i wait

244
00:21:04,820 --> 00:21:10,920
 for it to actually boot right so like everything turns on and i'm watching and i have a rooted android

245
00:21:10,920 --> 00:21:18,720
 phone and then once everything's on i start turning stuff off right uh so i turn off the compositor

246
00:21:18,720 --> 00:21:24,560
 that's like the thing that draws to the screen yeah uh i turn off the sound driver i turn off

247
00:21:24,560 --> 00:21:30,480
 i don't turn off wi-fi yet i i so currently i leave wi-fi on and i i can when i try to make like an

248
00:21:30,480 --> 00:21:35,640
 internet request it just uses the jvm's wi-fi i see that's one of the next things so component by

249
00:21:35,640 --> 00:21:41,260
 component you you want to replace that you want to turn them on and replace them with your own stuff

250
00:21:41,260 --> 00:21:47,040
 but on a booted phone that works yeah and so it's more fun like i can make an app so so yeah that's

251
00:21:47,040 --> 00:21:53,820
 what i had uh um that's kind of where i got to last friday when you left is like i was able to boot

252
00:21:53,820 --> 00:22:00,480
 the phone and i had a counter demo that took 13 seconds to yeah to click to add like at plus one

253
00:22:00,480 --> 00:22:05,540
 yeah uh and then so i have two environments i can run these in one of them is like on a pixel 4a

254
00:22:05,540 --> 00:22:10,520
 which chat gpt said is the easiest device to do this stuff on because it's just like very well known

255
00:22:10,520 --> 00:22:14,800
 a lot of you know okay people have been doing this for a long time and they're cheap to get so i have

256
00:22:14,800 --> 00:22:20,540
 like i got like four of these and uh and then the other one is i have a keemu emulator uh so just like a

257
00:22:20,540 --> 00:22:24,000
 linux vm that can run on my mac that i think it's basically the same thing as an android emulator

258
00:22:24,000 --> 00:22:29,220
 if you use an android emulator same thing i think they're both keemu uh and so that's really nice

259
00:22:29,220 --> 00:22:34,300
 to iterate on uh it's just faster than having to deal with a device and i can write automated tests

260
00:22:34,300 --> 00:22:40,260
 with that a little easier talk to me about the the apps uh you've built so far exactly how you're

261
00:22:40,260 --> 00:22:46,280
 thinking about an app's ecosystem yeah so so that's where i was last week and so this or when you left

262
00:22:46,280 --> 00:22:54,640
 so over the last week i spent a little time trying to make an app model so um i wanted to make it as

263
00:22:54,640 --> 00:23:00,180
 easy i wanted to make it very easy to make an app and i want to be something that llms would be very

264
00:23:00,180 --> 00:23:07,620
 good at and so the obvious choice is like javascript or typescript even better because that's just what

265
00:23:07,620 --> 00:23:14,040
 it's like by far the most popular language yeah um and so yeah that's i decided to make the apps

266
00:23:14,040 --> 00:23:17,380
 oh jesus christ it's really hard to navigate

267
00:23:17,380 --> 00:23:25,340
 i'm gonna go i'm gonna you take the left path i take the right path yeah yeah and let's see if one

268
00:23:25,340 --> 00:23:32,060
 of us ends up in the mud well you see the state of my shoes already um it's it's definitely the most

269
00:23:32,060 --> 00:23:40,200
 interesting recording and environment we had so far yeah so it's uh you know one day i'm gonna strap a

270
00:23:40,200 --> 00:23:44,540
 gopro to my ankle or something so just the people can see what we're going through literally

271
00:23:44,540 --> 00:23:52,400
 yeah so uh all right typescript and very viable so the goal is that the apps that are in the store

272
00:23:52,400 --> 00:23:58,240
 are incredibly easily yeah so i want it to be basically i don't want there to almost be a store

273
00:23:58,240 --> 00:24:05,460
 i want it to be so like viral sure it's almost like a dm yeah the apps are like a dm yeah and on

274
00:24:05,460 --> 00:24:10,980
 any app you can press and hold yeah and like voice vibe it like change the color change the layout like

275
00:24:10,980 --> 00:24:15,140
 yeah pablo had the same idea like a year ago or something where you're just every app comes with

276
00:24:15,140 --> 00:24:19,460
 a chat box but the problem you can you can change it on demand yeah so like the problem though is like

277
00:24:19,460 --> 00:24:23,480
 okay you can do that locally but you can do that with a web app yeah but even with the web app it's

278
00:24:23,480 --> 00:24:28,380
 like okay you need to vibe the source code yeah and then you need a build step yeah and so you need

279
00:24:28,380 --> 00:24:32,780
 a server now yeah and so now like that's another thing it's like i think i mean this is what another

280
00:24:32,780 --> 00:24:37,260
 one like i built this whole little cloud platform where i could like spawn vms on these bare metal

281
00:24:37,260 --> 00:24:41,540
 hedger machines yeah and like i use it to run ci and i use it to run like a couple open claws

282
00:24:41,540 --> 00:24:46,840
 and like i think i could run this for like 10k users and just just i could really undercut amazon

283
00:24:46,840 --> 00:24:51,280
 on price like it wouldn't have as many nines of reliability but it would be way way cheaper

284
00:24:51,280 --> 00:24:57,420
 and dimmy for example this guy who at the cohort who's running trying to run vps's or try to run

285
00:24:57,420 --> 00:25:03,480
 like simple cloud infrastructure attached to mining mining uh facilities it's doing the same thing

286
00:25:03,480 --> 00:25:06,920
 but he actually has control the hardware so i think this is like a really interesting primitive that

287
00:25:06,920 --> 00:25:12,800
 if we added like if if you could assume that like any user in in their daily course of their using

288
00:25:12,800 --> 00:25:19,160
 stuff made a couple payments to like rent vps's for various amounts of time if it became much much

289
00:25:19,160 --> 00:25:24,180
 easier to like almost like ephemeral things because you just need to run one test pipeline

290
00:25:24,180 --> 00:25:27,740
 as a matter of course yeah right as a matter of course you could spawn

291
00:25:27,740 --> 00:25:34,580
 like servers or docker containers or processes yeah or all this stuff and if the average user did

292
00:25:34,580 --> 00:25:37,980
 throughout the day and they spent a couple dollars on this a month or something yeah that would be a

293
00:25:37,980 --> 00:25:42,380
 fundamentally different nostril i mean in a fully vibed world that's where things are going anyway in

294
00:25:42,380 --> 00:25:48,800
 some sense right because people are fine with paying yeah on demand for and if you if you explain

295
00:25:48,800 --> 00:25:55,500
 this if like if you say like hey this money is going to self-hosting most people are like fuck that you

296
00:25:55,500 --> 00:25:59,320
 know this is too complicated i don't understand it i don't want to do that me included to some degree

297
00:25:59,320 --> 00:26:06,140
 yeah but if you're like hey this is feeding your agent your agent needs to eat like yeah yeah i see

298
00:26:06,140 --> 00:26:11,240
 a lot of people will feed their agent yeah all kinds of things yeah that they wouldn't consume or

299
00:26:11,240 --> 00:26:16,900
 purchase solo so that's kind of how i think about it it's like if we can get agents to be the thing is

300
00:26:16,900 --> 00:26:22,040
 the thing with self-hosting is it comes with lots of maintenance yeah and the i think that what makes

301
00:26:22,040 --> 00:26:28,180
 the gentix stuff so powerful is because if it works exactly there's zero maintenance basically the agent

302
00:26:28,180 --> 00:26:32,500
 figures everything out on their own i think that's why people are willing to throw so much money at it

303
00:26:32,500 --> 00:26:39,080
 because it literally is like a mini employee um that figures stuff out on on his own yeah just get

304
00:26:39,080 --> 00:26:45,320
 out of jail free cards yeah and just you just throw money at it and and and you know things will be

305
00:26:45,320 --> 00:26:52,760
 figured out yeah so that's but that's one so i was saying like how yeah of vibe coding uh apps

306
00:26:52,760 --> 00:26:58,400
 is hard because like okay you can't distribute the actual infrastructure to build the you need

307
00:26:58,400 --> 00:27:02,660
 infrastructure to build the apps and so that's a good nice thing about like if you i mean it's still

308
00:27:02,660 --> 00:27:08,020
 true with my phone but like so the way it works is like you just distribute typescript files yeah and

309
00:27:08,020 --> 00:27:15,460
 like i have dino running on this like dino runs the app so the the apps are just .tsx files or .ts

310
00:27:15,460 --> 00:27:21,260
 files and uh so you can have like one so all my apps are about 500 lines i have a cash wallet an

311
00:27:21,260 --> 00:27:29,400
 officer client a photos app uh these oh wow you really you really transcend the triangles i have a podcast

312
00:27:29,400 --> 00:27:35,900
 app that plays one podcast holy these are half your users i've been producing it can play no solutions

313
00:27:35,900 --> 00:27:42,060
 podcasts actually right now that doesn't work but it was working at a cafe and it was really loud

314
00:27:42,060 --> 00:27:48,340
 and i couldn't turn it off so i just had to turn the phone off but they were yeah i did disrupt one

315
00:27:48,340 --> 00:27:55,260
 cafe um what do you mean disrupt you mean enlighten yeah enlighten excuse me excuse me yeah everyone

316
00:27:55,260 --> 00:28:00,900
 was staring at me i do a lot of things just subtly to offend the european sensibilities like

317
00:28:00,900 --> 00:28:06,400
 we were we did this hike and there's this group of like germans and there was one american with

318
00:28:06,400 --> 00:28:11,040
 them and she goes like oh you're america too where are you from are you from massachusetts

319
00:28:11,040 --> 00:28:17,160
 which is like the worst place to be thought of as being from like massachusetts are the worst

320
00:28:17,160 --> 00:28:23,380
 they're the worst of us they're the true fucking assholes uh they're the worst thing about america

321
00:28:23,380 --> 00:28:31,660
 worse than california and so even her american her views the american view of americans is bad

322
00:28:31,660 --> 00:28:37,840
 yeah so imagine the european view yeah so i yeah i just i i don't know what are we talking about

323
00:28:37,840 --> 00:28:44,240
 oh yeah i was in the cafe offending people with the podcast app yeah and i think i have one more app

324
00:28:44,240 --> 00:28:50,680
 or counter app and so i in the last week now it's instant because i'm i now i'm i i took over the gpu

325
00:28:50,680 --> 00:28:57,700
 all right i can write to the gpu don't ask me how but i can and it went from like 13 second button

326
00:28:57,700 --> 00:29:03,760
 clicks to like instant uh so yeah i'm slowly crawling down the stack what's still missing

327
00:29:03,760 --> 00:29:12,840
 um you think in terms of primitives you mentioned wi-fi wi-fi so networking is still missing in some

328
00:29:12,840 --> 00:29:20,420
 sense yeah wi-fi cellular um buttons i don't think i can do the volume buttons okay so certain hardware

329
00:29:20,420 --> 00:29:25,520
 buttons yeah certain i assume nfc bluetooth all those kind of things yeah those things but i think

330
00:29:25,520 --> 00:29:31,000
 these are all relatively doable like there's drivers and then there's like a language agnostic way called

331
00:29:31,000 --> 00:29:36,700
 the hall hardware hardware extraction abstraction layer well actually so most of the apis you know

332
00:29:36,700 --> 00:29:42,660
 there's only one so you have linux apis and then you have android apis and so for touch i'm actually

333
00:29:42,660 --> 00:29:47,060
 and drawing i'm actually using the pure linux apis i always try to do the linux apis if i can

334
00:29:47,060 --> 00:29:51,800
 even if it's a little more complicated because because it's closer to the bare metal it's closer to the

335
00:29:51,800 --> 00:29:56,360
 bare metal and it's simpler and it's like there's just fewer layers so i've been able to succeed

336
00:29:56,360 --> 00:30:01,120
 everywhere there except for the camera for the camera for some reason it really wanted to use it

337
00:30:01,120 --> 00:30:05,880
 was very funny i had to really fight the agents on this like they all kept trying to get me to use

338
00:30:05,880 --> 00:30:11,080
 the java code like i'd point them at a lower level and they always steer like this is one where i really

339
00:30:11,080 --> 00:30:15,400
 felt the bias of the training data like very few people had attempted this yeah and it kept trying

340
00:30:15,400 --> 00:30:20,020
 to steer me away from it and finally i'm like you know like laying down the law on a new a new agent

341
00:30:20,020 --> 00:30:23,220
 session we are going to do it this way i've been i've been waiting for them to try to figure out

342
00:30:23,220 --> 00:30:27,780
 some java version for the camera just so we could have a proof of concept and then go bare lower

343
00:30:27,780 --> 00:30:33,260
 bare metal and so it wasn't working and i finally i created a new session like i am we are going to

344
00:30:33,260 --> 00:30:37,340
 do it this way and we're going to succeed or fail and there are no other options like burn the boats

345
00:30:37,340 --> 00:30:43,380
 yeah and we were done in like an hour yeah so it's very interesting like the that was one area where

346
00:30:43,380 --> 00:30:50,220
 like the that was one area where i could really feel like the agent or yeah they always try to drag

347
00:30:50,220 --> 00:30:54,040
 you to the middle of the bell curve they didn't yeah exactly like i could really feel the pull of the

348
00:30:54,040 --> 00:30:59,260
 middle of the bell curve there uh even though i and and so but there we use the android c++ code

349
00:30:59,260 --> 00:31:06,520
 um hal's uh hardware extraction abstraction layers and so the the the next lower one is that there are

350
00:31:06,520 --> 00:31:11,920
 like linux drivers for this too and i might try to go to at that level too uh but yeah so for

351
00:31:11,920 --> 00:31:14,880
 everything that's the only one that the camera is the only one where i'm actually using the android

352
00:31:14,880 --> 00:31:21,920
 c++ code so so yeah like i don't know what else is there on a phone um i don't know flashlight i

353
00:31:21,920 --> 00:31:26,200
 probably can't control the flashlight accelerometer but a lot of these things it's like you know like

354
00:31:26,200 --> 00:31:31,600
 i'm trying to i'm trying to basically have something that is like an alternative to pwas on nostre yeah

355
00:31:31,600 --> 00:31:37,080
 and like i want the developer experience to be about the same but i want you to not have to write

356
00:31:37,080 --> 00:31:44,060
 like i want you to have like one wallet one nostre like one one connection per relay across the whole

357
00:31:44,060 --> 00:31:49,280
 device right explain if you have 10 if you have 10 apps all talking to domus relay

358
00:31:49,280 --> 00:31:55,500
 that should be one web okay okay okay i see and you should only ever get any okay that's the same

359
00:31:55,500 --> 00:32:00,740
 idea as having one lightning thing on your exactly and you should have one cashew wallet you should be

360
00:32:00,740 --> 00:32:06,120
 able to have one spark wallet like and it should be implemented once in rust in a disciplined manner

361
00:32:06,120 --> 00:32:10,980
 with a lot of testing and you shouldn't vibe that like i want to build the things that you shouldn't

362
00:32:10,980 --> 00:32:14,760
 vibe and i want users to be able to just vibe code the shit out of everything just plug into

363
00:32:14,760 --> 00:32:20,100
 those yeah yeah and of course natively running you have the services so to speak you have the freedom

364
00:32:20,100 --> 00:32:23,780
 yeah the platform apis you have the freedom to ignore the platform apis and do it yourself

365
00:32:23,780 --> 00:32:28,640
 but it's probably going to be less efficient and it's probably you probably don't need to right

366
00:32:28,640 --> 00:32:33,700
 and so you know like phips i'm going to have phips running in uh at the at that level i'll probably

367
00:32:33,700 --> 00:32:38,240
 have some sort of like tail scale type you know like maybe not for vpn like any all these cool

368
00:32:38,240 --> 00:32:44,380
 things that we have all these nice little primitives i want to have that like this all available to

369
00:32:44,380 --> 00:32:48,720
 any app at the operating system level with like simple apis that you can vibe code and experience

370
00:32:48,720 --> 00:32:55,560
 on top of so that's the goal and uh in theory if someone of the like three people that two people

371
00:32:55,560 --> 00:33:01,520
 let's be generous if the one guy that's still listening is interested can he flash it on his

372
00:33:01,520 --> 00:33:07,100
 pixel 4a yeah i mean if you go to justinmoon slash shadow and you install nix uh because this is a man

373
00:33:07,100 --> 00:33:12,580
 real man a real man's operating system a real man is listening here so these are no children listening

374
00:33:12,580 --> 00:33:20,160
 at this point um surely already nix is the only nix is the only nix installed yeah surely uh or else he

375
00:33:20,160 --> 00:33:25,660
 wouldn't be listening exactly right so nix nix is so i couldn't do every one of my projects uses nix

376
00:33:25,660 --> 00:33:33,320
 nix is like uh it's like a it's like a uh a way of defining a software behave uh environment in like

377
00:33:33,320 --> 00:33:42,580
 a declarative way so it and and uh so you can you can describe inputs and outputs to like a software

378
00:33:42,580 --> 00:33:49,820
 build see it's very easy to define one package this way to the entire uh dependencies for a whole

379
00:33:49,820 --> 00:33:55,600
 project or you can define a whole operating system and it's really really good for build environments

380
00:33:55,600 --> 00:34:01,360
 so like i have a nix environment where i can totally determine what goes on to the rom on this phone

381
00:34:01,360 --> 00:34:07,600
 you know and it's trivial to get new new to cross compile my rust code and all of its dependencies

382
00:34:07,600 --> 00:34:11,680
 to this android phone it's like solves all that for me so that's why if you want to clone this you need

383
00:34:11,680 --> 00:34:17,800
 nix because nix is what the entire build system is built on and so you could run i think just run

384
00:34:17,800 --> 00:34:24,300
 if you have a pixel plugged in pixel 4a plugged in and you root it i think this i think my script roots

385
00:34:24,300 --> 00:34:30,220
 it uses like magisk or whatever uh yeah you know enable developer debugging and all that stuff

386
00:34:30,220 --> 00:34:38,200
 and you run just run and then the the serial number that you see in adb devices it should flash it and

387
00:34:38,200 --> 00:34:43,940
 run it all right awesome uh so yeah yeah in the nix environment so yeah if someone does it let us know

388
00:34:43,940 --> 00:34:49,100
 tag us on roster and if you don't you can also do just run vm and that will run the kumu vm

389
00:34:49,100 --> 00:34:53,640
 so that that should basically one that should be more reliable because you don't have to prepare

390
00:34:53,640 --> 00:34:58,960
 a device in any way and again i'm the only i we did have one guy uh coasty got it to run on

391
00:34:58,960 --> 00:35:05,080
 linux and autumn also got it to run oh wow okay so we have had three people do it coasty made some

392
00:35:05,080 --> 00:35:12,480
 some prs but yeah i'm at this point i'm it's mostly a solo like i'm still trailblazing sure but if

393
00:35:12,480 --> 00:35:16,600
 somebody is like uniquely interested in this if somebody really wants to like this is a project

394
00:35:16,600 --> 00:35:23,780
 that could justify 10 or 20 developers yeah like we could like if this is something where if we did

395
00:35:23,780 --> 00:35:28,960
 really go for it it could make a big impact yeah which is why i'm interested yeah you know but i

396
00:35:28,960 --> 00:35:33,340
 think it's still a little early to attempt that yeah yeah yeah i see i see here's the wind's back

397
00:35:33,340 --> 00:35:39,640
 the wind is back the microphones are usually really really good at filtering it out but the cows are

398
00:35:39,640 --> 00:35:44,100
 here as well so i just took a picture i'll actually add them to the show notes if i manage yeah like

399
00:35:44,100 --> 00:35:48,480
 the people don't know like what kind of recording studio we have yeah it's awesome i'll have to take

400
00:35:48,480 --> 00:35:52,840
 another one it's like it's super misty and like bulls all around and all those kind of things

401
00:35:52,840 --> 00:36:00,000
 yeah all right so in terms of trailblazing what do you think is still left to do oh oh yeah so so one

402
00:36:00,000 --> 00:36:04,700
 other thing typescript so the the difference with this is

403
00:36:04,700 --> 00:36:13,460
 jeez the difference with this is like so i don't like i think the browser environment's not a very

404
00:36:13,460 --> 00:36:18,100
 great developer environment for example that you can't write a file very easily there's all kinds

405
00:36:18,100 --> 00:36:23,200
 of limits on writing to like local sure the browser is its own sandbox yeah like

406
00:36:23,200 --> 00:36:29,960
 it sucks in a lot of ways uh and again there's another one for security yeah yeah yeah that's

407
00:36:29,960 --> 00:36:34,020
 that's fine that's fair i think that's understandable you don't you don't want random code from the

408
00:36:34,020 --> 00:36:38,600
 internet executing on your bare metal all the time and so on yeah you want some guardrails well

409
00:36:38,600 --> 00:36:44,040
 actually i do want this right now sure right right now i want this i want full yolo like yeah i want

410
00:36:44,040 --> 00:36:49,100
 like no fucking guardrails yeah you trust that end pub yeah it's remote code execution from that end

411
00:36:49,100 --> 00:36:55,460
 pub and i completely agree that we should have that and do that and that looping back to the

412
00:36:55,460 --> 00:36:59,860
 comparison with guns or cars it's like i want something that just like if i use it wrong i shoot

413
00:36:59,860 --> 00:37:03,580
 myself in the foot i want a 3d i want a 3d printed gun of phones

414
00:37:03,580 --> 00:37:12,640
 and and then that's someone's gonna clip that you'll land on some sort of list just for that

415
00:37:12,640 --> 00:37:17,440
 sentence alone this is why i hate going on podcasts if i ever start speaking honestly i'm like yeah i

416
00:37:17,440 --> 00:37:25,180
 want a 3d printed gun of phones every it's just like yeah the most unhinged shit yeah yeah no but

417
00:37:25,180 --> 00:37:29,920
 you you want to have a powerful tool i want to max it out on that and then and then on every just to

418
00:37:29,920 --> 00:37:34,280
 to close the thought for real like on every level it should be permissionless yes you should not

419
00:37:34,280 --> 00:37:40,020
 be asking for permission to do whatever you want on your hardware and i want i want no solutions listener

420
00:37:40,020 --> 00:37:45,960
 to run it that's the only person and sec you know and maybe a couple people in austin these are the

421
00:37:45,960 --> 00:37:49,760
 only people i care about running it but i think yeah what's interesting about austin it's like what do we

422
00:37:49,760 --> 00:37:55,160
 have to work with right like there's a lot of people who've built linux phones and they've all

423
00:37:55,160 --> 00:38:00,200
 failed mozilla tried it everyone tried it yeah yeah i had them all they were all horrible they're

424
00:38:00,200 --> 00:38:03,460
 all horrible they failed for i think they fall fail for kind of two reasons i mean one of them is

425
00:38:03,460 --> 00:38:11,780
 like they don't really have this like um really really uh ideologically motivated and like technically

426
00:38:11,780 --> 00:38:17,820
 confident i think in part i think in part shut up shut up god damn it god damn it he finishes my

427
00:38:17,820 --> 00:38:23,200
 thoughts for me he interrupts me no the people that were building it they are not phone users yeah

428
00:38:23,200 --> 00:38:27,760
 they're server users yeah like that was obvious like when you got one of those phones they didn't

429
00:38:27,760 --> 00:38:32,600
 they didn't they didn't use it themselves yeah so that's one thing and but we have like power users

430
00:38:32,600 --> 00:38:37,320
 and we have like tinkerers like that's one thing nostr has like there was a so the guy who made

431
00:38:37,320 --> 00:38:42,460
 wisp i forget what his name is but he you take so yeah so he had an interesting uh poll the other day

432
00:38:42,460 --> 00:38:46,200
 where he asked what phone people use and it was a third ios a third android and a third graph yeah

433
00:38:46,200 --> 00:38:50,880
 that's awesome that's an unusual yeah it's a very unusual thing yeah uh that a third of the people

434
00:38:50,880 --> 00:38:54,500
 would flash their own phones yeah and so i think that's like a primitive we have to work with in

435
00:38:54,500 --> 00:38:58,800
 terms of like building projects and communities around nostr that's very exciting yeah yeah we have

436
00:38:58,800 --> 00:39:04,120
 all the insane people yeah people will do this right and so and the other thing is like so what

437
00:39:04,120 --> 00:39:08,780
 what what were they trying to install on these linux phones like an email client yeah like they had no

438
00:39:08,780 --> 00:39:15,640
 ideas for apps yeah like they had no idea they had no reason to like like oh you can write your app

439
00:39:15,640 --> 00:39:21,780
 right like you you had the ubuntu phone that like ran just ubuntu and then you have ubuntu apps on

440
00:39:21,780 --> 00:39:26,480
 your phone yeah don't work on the touch screen at all yeah it was all terrible i feel like none of

441
00:39:26,480 --> 00:39:31,320
 them actually had a vision for like what a phone could be or should be yeah and so i had some from

442
00:39:31,320 --> 00:39:35,180
 nostr like that i thought of this a year ago where you could like have kind of like these

443
00:39:35,180 --> 00:39:41,980
 really small understandable unix tools and it would be very easy to uh share them or write them

444
00:39:41,980 --> 00:39:46,240
 back then like write them by hand and i thought that was somewhat exciting but it wasn't exciting

445
00:39:46,240 --> 00:39:51,020
 enough i think ai sort of makes brings it to the next level where like anybody could participate in

446
00:39:51,020 --> 00:39:54,780
 that right yeah like you don't have to you know the last two courts we've had people that weren't

447
00:39:54,780 --> 00:39:59,820
 programmers before come yeah and give good demos so it's like a dude like that's something we've seen

448
00:39:59,820 --> 00:40:04,800
 at sec yeah yeah we've we've hit the threshold now because in the first two cohorts that didn't work

449
00:40:04,800 --> 00:40:10,760
 at all and ever ever since last fall last winter that started to work that project managers and

450
00:40:10,760 --> 00:40:15,560
 designers can produce something really good yeah so so yeah so that's another thing it's like

451
00:40:15,560 --> 00:40:22,660
 now you you as the software developer or the os creator don't have to like actually

452
00:40:22,660 --> 00:40:28,480
 figure out what the app should be yes right like you just users can and and you don't need app

453
00:40:28,480 --> 00:40:32,380
 developers like that was what ios did is like they're they figured out how to create an app developer

454
00:40:32,380 --> 00:40:38,560
 ecosystem yeah and that was a new thing and they still get fucking 30 percent you know um and so like

455
00:40:38,560 --> 00:40:42,740
 maybe what is the next thing is like what if it's the user ecosystem and it's i don't know

456
00:40:42,740 --> 00:40:48,080
 like myspace is an inspiration to me or it's like a lot of the actual vibe was from the users tumblr

457
00:40:48,080 --> 00:40:52,360
 or something like these things like i don't know that that is something that platform stepping back

458
00:40:52,360 --> 00:40:56,740
 and then so so yeah we have so that's an appeal like why i have a new phone well so that the user

459
00:40:56,740 --> 00:41:04,160
 can actually fully like own anything about how the the uh apps work or the operating system like hey

460
00:41:04,160 --> 00:41:11,420
 if you want to vibe code how your cellular driver interactions work yeah fucking go for it here's uh

461
00:41:11,420 --> 00:41:17,640
 you know one one dollar a minute i'll i'll rent you a thousand core machine and you know whatever right

462
00:41:17,640 --> 00:41:23,020
 like but that's something i want to be able to do it's like okay uh like so the other type of app would

463
00:41:23,020 --> 00:41:27,960
 be like a bare metal app like written in rust or zig or something that like draws 60 frames a second it's

464
00:41:27,960 --> 00:41:31,600
 like it's like note deck right yeah you could do that but it's like you can't really vibe that on your

465
00:41:31,600 --> 00:41:36,920
 phone but hey rent me a build rent a build server from me and i will i will build it for you and it

466
00:41:36,920 --> 00:41:42,660
 will be relatively quick i mean hopefully if that stuff works well enough an ecosystem around this

467
00:41:42,660 --> 00:41:47,200
 will spring up you won't be the only yeah like wipe this but i feel like i have to with a lot of

468
00:41:47,200 --> 00:41:50,000
 you have to kick start it you have to kick start you have to show people what's possible

469
00:41:50,000 --> 00:41:56,800
 and so so yeah i don't know like that so so yeah but but yeah and i i think the other thing is

470
00:41:56,800 --> 00:41:59,680
 like i've already kind of said this but i think we actually have ideas of what should be on an

471
00:41:59,680 --> 00:42:05,060
 operating system that is different right like we want the money on there we want the how to

472
00:42:05,060 --> 00:42:09,040
 communicate with relays you know mock clients like be able to do audio calls video calls we want

473
00:42:09,040 --> 00:42:12,620
 like some way of doing push notifications that you don't have to bend over and get on your

474
00:42:12,620 --> 00:42:16,600
 fucking knees you know from apple like that's one of the things that's so frustrating it's like

475
00:42:16,600 --> 00:42:22,860
 yeah it's insane it's so frustrating it's like i can't my my i pay a thousand dollars for this phone

476
00:42:22,860 --> 00:42:28,120
 even on android it's not great right like uh i don't know like pokey does are these good like

477
00:42:28,120 --> 00:42:34,220
 there's a few of them for yeah pokey is but pokey is purely like it's purely for noster basically

478
00:42:34,220 --> 00:42:38,940
 yeah it's just one one app that does all noster notifications for all your if you if you're a

479
00:42:38,940 --> 00:42:43,200
 power it works well okay if you are a power user and you have multiple noster accounts yeah like and

480
00:42:43,200 --> 00:42:48,520
 you just want to get spent with notifications pokey is the way to go for sure but i feel like we could

481
00:42:48,520 --> 00:42:54,620
 you know yeah we could it would be cool if like the platform itself had a solution here and it worked

482
00:42:54,620 --> 00:42:59,900
 out of the box and apps could integrate and be a little smarter with how it works and just uh so

483
00:42:59,900 --> 00:43:04,700
 there's so many of these things that i think we could we could do in a different way yeah and uh so

484
00:43:04,700 --> 00:43:08,820
 yeah that's kind of the vision and the crazy thing is like i know how to do it and i think i can

485
00:43:08,820 --> 00:43:13,640
 accomplish it all by myself on the pixel 4a like i can make you buy a hundred dollar fifty dollar

486
00:43:13,640 --> 00:43:18,900
 whatever 75 pixel 4a and you can boot this thing and it it can do everything you can do on other

487
00:43:18,900 --> 00:43:24,200
 noster apps maybe there's a calendar maybe i make a little email client like i think i could get this

488
00:43:24,200 --> 00:43:29,920
 thing where it's like it could replace your phone uh the security properties are not going to be great

489
00:43:29,920 --> 00:43:36,680
 at the beginning i think the most important thing to kickstart is replace it for yourself yeah and

490
00:43:36,680 --> 00:43:42,740
 whatever use you have because people use phones in all kinds of different ways obviously but if you're

491
00:43:42,740 --> 00:43:48,700
 if you're able to basically get it to a point where you can more or less use it as a daily driver

492
00:43:48,700 --> 00:43:53,500
 yeah that would be awesome and like i don't know i don't really like my relationship with my phone

493
00:43:53,500 --> 00:44:00,280
 to be honest like i feel like my my phone kind of it's like a whatever like a siren like sure yeah

494
00:44:00,280 --> 00:44:06,080
 it is yeah i mean it's designed to be it yeah way and so that's another thing like yeah i don't know

495
00:44:06,080 --> 00:44:12,740
 i'd like to be i don't know i'd like i'd like it if like people tried to text message me and it's

496
00:44:12,740 --> 00:44:20,120
 like justin is not texting over the next 12 hours but you can call him you know or like justin is not

497
00:44:20,120 --> 00:44:26,260
 accepting calls right now but if you give him 50 dollars in sats he will refund it if this is a

498
00:44:26,260 --> 00:44:32,900
 valid call not if you're just being a bitch he won't refund it you know like i'd like to be able

499
00:44:32,900 --> 00:44:37,960
 to play around with these things to make like uh you know make it and it's like hey if you want to

500
00:44:37,960 --> 00:44:43,380
 play around with it on the operating system level everyone at href if you want me to give input on

501
00:44:43,380 --> 00:44:51,040
 your google doc you gotta go through just you need to download it you need to you know there's an

502
00:44:51,040 --> 00:44:57,220
 order of operations you know alex like you gotta flash it get a fix of 4a and justin will do his

503
00:44:57,220 --> 00:45:03,600
 work you know i think that's what's awesome about those kind of things once you once you let people

504
00:45:03,600 --> 00:45:08,340
 actually experiment and build stuff like once you actually move into the permissionless realm yeah

505
00:45:08,340 --> 00:45:13,360
 then all kinds of things surprising things will happen exactly and i think that's that's why

506
00:45:13,360 --> 00:45:19,960
 you know like lots of people here are so excited around um you know going permissionless across the

507
00:45:19,960 --> 00:45:27,200
 stack because then you can actually you you you i think lots of people are holding back in some

508
00:45:27,200 --> 00:45:32,500
 sense right now or just in general because it's just okay you're gonna build it out and you have

509
00:45:32,500 --> 00:45:37,560
 no chance of getting into the app stores yeah for example yeah no this is this is exactly that my

510
00:45:37,560 --> 00:45:42,300
 feeling is like i just i kind of gave up on that and i'm like i'm just gonna build my own parallel

511
00:45:42,300 --> 00:45:46,580
 thing here yeah and i'm gonna do it for at least right this is the first thing i've been working

512
00:45:46,580 --> 00:45:51,140
 on since like fetty where i was like i might just trudge away on this one for a few years

513
00:45:51,140 --> 00:45:56,300
 yeah fuck it go really hard on it just yeah why not i mean you're definitely not alone i know i know

514
00:45:56,300 --> 00:46:00,280
 personally of like many many dozens of people that are in this position where it's like

515
00:46:00,280 --> 00:46:05,140
 what's the point of going hard on it because it's you're just going to be deplatformed anyway i know

516
00:46:05,140 --> 00:46:09,500
 you know so yeah that's i want something where i can go hard on yeah and this one it's like i can

517
00:46:09,500 --> 00:46:16,960
 i can really go to town yeah you know um so yeah and i mean you know it's not i don't you know i

518
00:46:16,960 --> 00:46:22,160
 think like graphene is an amazing option and this is like a different thing though yeah like graphene

519
00:46:22,160 --> 00:46:29,040
 is optimizing for security yes and they're trying to they're you know you want to optimize for shooting

520
00:46:29,040 --> 00:46:36,620
 yourself in the foot yeah i want to optimize for hackability and like customization and and you know

521
00:46:36,620 --> 00:46:42,180
 and i think we could get over time we could get the the the security to be pretty good you know

522
00:46:42,180 --> 00:46:46,080
 because it's like all the code all the apps run in v8 right now which is so it's essentially a browser

523
00:46:46,080 --> 00:46:50,740
 sandbox you know and like i'll have to do more stuff they can do file io so we'll need to like

524
00:46:50,740 --> 00:46:55,100
 be smarter i wouldn't be surprised if one app can write into another apps or read from another app

525
00:46:55,100 --> 00:46:59,180
 right now right like i'm sure stuff like that's possible but i think we could lock all that down

526
00:46:59,180 --> 00:47:04,380
 and uh i think we could get security to be you know if we do the newest pixels it's like we do the

527
00:47:04,380 --> 00:47:10,060
 secure boot and some of the stuff that they that they pioneer like i think we could maybe some of

528
00:47:10,060 --> 00:47:15,260
 that like port some of that to rust and run it here or i just run the jvm code whatever uh you know

529
00:47:15,260 --> 00:47:19,040
 like i think over time that could be possible but it's like it's a different trade-off it's a

530
00:47:19,040 --> 00:47:24,180
 different set of trade-offs and goals than graphene and so it's not like uh i think that's a great

531
00:47:24,180 --> 00:47:28,280
 problem i don't want to throw any shade at that yeah no not at all graphene is fantastic obviously

532
00:47:28,280 --> 00:47:33,120
 i mean very much needed right now yeah like to be clear shadow is like in like pre-alpha

533
00:47:33,120 --> 00:47:43,480
 i mean as i said 13 second button clicks last friday yeah yeah yeah yeah but we we transcended that so

534
00:47:43,480 --> 00:47:48,600
 yeah yeah so uh but yeah so that's so i don't know i think i think i think it could be very

535
00:47:48,600 --> 00:47:55,280
 very fun and uh yeah start with the pixel 4as and then maybe try to do like a more flagship one like

536
00:47:55,280 --> 00:48:00,840
 you know nine or ten uh once that works one step at a time yeah talk to me about the network stack a

537
00:48:00,840 --> 00:48:06,620
 little bit i mean the cohort was dominated by phips so yeah one of one of the permission layers of

538
00:48:06,620 --> 00:48:13,000
 course is dns and just ipv4 addresses in general yeah and how are you thinking about that right now

539
00:48:13,000 --> 00:48:19,560
 after the last like two and a half weeks and just kind of in in in light of building a new mobile

540
00:48:19,560 --> 00:48:26,700
 operating system yeah um or maybe before that like in terms of app distribution for example how are you

541
00:48:26,700 --> 00:48:32,520
 thinking about that like you said um it's all typescript files it's all tsx files and in the

542
00:48:32,520 --> 00:48:38,340
 end it should just be like receiving a dm and you you have the app so what are you using right now and

543
00:48:38,340 --> 00:48:45,120
 i assume certain other kinds or or how how do you distribute so far it's one mono repo and everything

544
00:48:45,120 --> 00:48:49,860
 gets installed at once okay uh so i haven't attempted distribution yet i think it would be something

545
00:48:49,860 --> 00:49:00,660
 like you know an app kind and uh signed that contains a signed hash of a blossom or hash tree

546
00:49:00,660 --> 00:49:07,660
 like you know uh thing that actually contains the source code same as app store works right now

547
00:49:07,660 --> 00:49:11,980
 basically exactly same principle yeah i mean it's absolutely yeah the reason why i i hang those

548
00:49:11,980 --> 00:49:16,140
 two questions together by the way because that's obviously the our kill our kill is heel of app

549
00:49:16,140 --> 00:49:21,160
 distribution and so on like if you have an app repository that runs on a domain or just a centralized

550
00:49:21,160 --> 00:49:27,300
 server somewhere if it's spicy enough it will just get shut down sure so yeah i mean the other interesting

551
00:49:27,300 --> 00:49:31,540
 thing is just like having the option to just distribute the source code right and so like

552
00:49:31,540 --> 00:49:36,340
 i can you know dino can run i mean it has to do like a build step it has to download dependencies

553
00:49:36,340 --> 00:49:42,100
 uh but it only takes like a couple seconds yeah you know just being able to just distribute the source

554
00:49:42,100 --> 00:49:46,440
 code makes it much smaller right yeah also much more censorship resistant because you can literally send

555
00:49:46,440 --> 00:49:53,940
 a guy a book yeah so so yeah like that that is you know you're talking about like kilobytes instead of

556
00:49:53,940 --> 00:50:01,860
 megabytes yeah and uh but yeah so i want it to be very i don't i'm not gonna have like an app uh any

557
00:50:01,860 --> 00:50:06,900
 sort of like a store or like a central thing i don't know think about how to monetize app developers

558
00:50:06,900 --> 00:50:14,180
 will be kind of interesting you know um it would be cool if if you know i mean i one of my complaints

559
00:50:14,180 --> 00:50:18,900
 about nostril is like nobody's good at getting money out of the users yeah you know and so i want to

560
00:50:18,900 --> 00:50:27,140
 uh i might try that uh uh a lot more it's like it would be cool if you know if the app developers

561
00:50:27,140 --> 00:50:33,860
 were actually making money like if if you like if somebody made a popular app that they could they

562
00:50:33,860 --> 00:50:38,820
 could get a cut you know um so that's going to be an interesting thing to play with uh but i think my

563
00:50:38,820 --> 00:50:44,500
 feeling there is like the way to do it is going to be lots of tools for agents right yeah like you're

564
00:50:44,500 --> 00:50:50,340
 going to have an agent and uh the phone is a part of the picture yeah i think i think it's like in

565
00:50:50,340 --> 00:50:56,180
 cyberspace it's very hard to figure out what to charge for and the obvious thing is compute yeah

566
00:50:56,180 --> 00:51:02,660
 it's the only thing that's actually scarce yeah and so that makes a lot of sense i like the current

567
00:51:03,380 --> 00:51:10,020
 like we we've come a long way already amethyst for example has the on every app update sap the

568
00:51:10,020 --> 00:51:16,500
 developers quite prominent yeah and people do yeah and sets are flowing to the developers it's like a

569
00:51:16,500 --> 00:51:23,220
 massive massive sap split yeah to everyone who contributed yeah um so i actually wouldn't focus

570
00:51:23,220 --> 00:51:28,580
 on that too hard in some sense because i think something will like that will be figured out yeah

571
00:51:28,580 --> 00:51:34,740
 alongside of you building a new os like you're not going to fix all the issues at once yeah i mean

572
00:51:34,740 --> 00:51:38,980
 i don't know i kind of disagree in the sense that like i want to really play with that a lot from

573
00:51:38,980 --> 00:51:46,980
 the very beginning okay like find good ways to get an economy going you know like the problem with uh

574
00:51:46,980 --> 00:51:52,820
 you know economy in cyberspace is that like a jpeg isn't scarce you know what i mean yeah and an app

575
00:51:52,820 --> 00:51:57,140
 is also not scarce well and that's kind of what the app stores actually do is they enforce and that's

576
00:51:57,140 --> 00:52:01,620
 and that's why it's so permissioned right and that's why you take a 30 cut it's like it's it all hangs

577
00:52:01,620 --> 00:52:07,220
 together that's why i'll make an app store and take 30 cut sounds nice now that you bring it up

578
00:52:07,220 --> 00:52:12,980
 now that you really put it into context what are we maybe like i think i'm gonna do 45 minutes in

579
00:52:12,980 --> 00:52:19,940
 and you're already on the dark side i'm gonna do that i think yeah take 30 yeah yeah 29 i'll call up tsmc

580
00:52:20,660 --> 00:52:31,380
 get in that next two nanometer batch yeah yeah and i'm apple basically but i think the the compute

581
00:52:31,380 --> 00:52:36,740
 angle or the agentic angle is really really good because people are willing to pay for useful

582
00:52:36,740 --> 00:52:42,660
 services period and that might be like in the pre-agentic world it was kind of service quality

583
00:52:42,660 --> 00:52:48,500
 and uptime and just data guarantees and those kind of things and now it's just straight up compute in

584
00:52:48,500 --> 00:52:54,820
 some sense yeah and the a lot of the old services i mean it's interesting so with hrf we use some

585
00:52:54,820 --> 00:52:59,380
 trainings where we like bring a couple activists in and we teach them how to vibe code teach them how

586
00:52:59,380 --> 00:53:05,460
 to uh play with a claw you know set up a claw yeah and you know what the risks are hatch a lobster

587
00:53:05,460 --> 00:53:10,020
 as they say yeah hatch a lobster yeah and what the what the risks are what what you should how to think

588
00:53:10,020 --> 00:53:14,660
 about it how to succeed we try to get them a win or two you know where they have some idea they've always

589
00:53:14,660 --> 00:53:22,980
 wanted to do and now they actually do it and uh and the hardest thing about this is creating accounts

590
00:53:22,980 --> 00:53:30,820
 it's like half the obviously most of these um trainings we do like one of the lobsters gets banned

591
00:53:30,820 --> 00:53:36,660
 on github yeah or one of the lobsters gets you know we got couldn't log into chat gpt or you know

592
00:53:36,660 --> 00:53:42,260
 to for for codex on one of them and like every time there's some really really bad authentication issue

593
00:53:42,260 --> 00:53:46,660
 so yeah um so this is where an ostrich could like really shine is if you could have

594
00:53:46,660 --> 00:53:50,340
 you know a set of tools so that's another kind of little

595
00:53:53,700 --> 00:54:09,300
 are we lost are we lost no but we don't know where everyone went left to right left or right

596
00:54:16,340 --> 00:54:28,180
 okay all good that would have been fun just getting lost in the woods yeah completely where

597
00:54:28,180 --> 00:54:36,180
 were we no exits podcast yeah exactly oh yeah um uh the friction of permission systems and getting

598
00:54:36,180 --> 00:54:42,660
 accounts banned and so that's that's um i don't know that's it's a bigger problem for agents than it is

599
00:54:42,660 --> 00:54:48,180
 for humans i think and also agents are like a little less long lived also you know new password

600
00:54:48,180 --> 00:54:53,380
 managers and stuff are going to come popping up like stripe also right now all the services hate

601
00:54:53,380 --> 00:54:57,860
 agents basically they want to ban them and just they're all overwhelming they're all you know like

602
00:54:57,860 --> 00:55:01,940
 if you look at the github repo creation it's gone just absolutely exponential the last year yeah

603
00:55:01,940 --> 00:55:09,140
 absolutely it is understandable but it's also it's it's it's also a new reality yeah and so yeah i don't

604
00:55:09,140 --> 00:55:13,540
 know like that's another interesting thing like this is some this is what i one of my other projects

605
00:55:13,540 --> 00:55:16,980
 like i've kind of have all these ideas now and now i just need to start like violently shipping

606
00:55:16,980 --> 00:55:23,140
 uh so violently shipping that's like yeah there's a there's a diarrhea joke exactly

607
00:55:23,140 --> 00:55:33,220
 we're not gonna make the joke of course not at all we'll laugh hope the listener is laughing

608
00:55:36,500 --> 00:55:40,500
 it's like half a listen though it's just like the dolphins you know switching switching off half

609
00:55:40,500 --> 00:55:49,780
 the brain yeah exactly yeah i won't say it uh yeah so that's one thing i want i was making is like uh

610
00:55:49,780 --> 00:55:55,620
 you know so i don't like a bare metal machine and i could spawn a vm on it and like what i want is

611
00:55:55,620 --> 00:55:59,860
 like my like i did i have this a little bit i have an open claw that has it's got root on a

612
00:56:00,740 --> 00:56:06,900
 what 64 gig ram you know like 16 thread hetzner machine and it can do whatever the it wants there

613
00:56:06,900 --> 00:56:11,700
 yeah as an xconfig i can tell it's like set up electrum server it sets up an electrum server

614
00:56:11,700 --> 00:56:17,860
 everything gets committed to get to get um the the back of the data the there's like a separate it

615
00:56:17,860 --> 00:56:21,540
 doesn't have access to everything so like the backup there's a backup service that's written by another

616
00:56:21,540 --> 00:56:28,180
 user that it can't up it can up the paths to some extent like so i need a little bit better system

617
00:56:28,180 --> 00:56:33,300
 to make sure that like this one agent user like if it creates data it does actually get

618
00:56:33,300 --> 00:56:40,180
 backed up yeah like what i want is like give it full control over the user space you can do whatever

619
00:56:40,180 --> 00:56:45,620
 it wants if it needs a place to hit set up a git repo install git to you just vibe code your own who

620
00:56:45,620 --> 00:56:51,700
 cares yeah whatever do whatever you want you know yeah and uh so kind of like that idea is like

621
00:56:51,700 --> 00:56:57,380
 like give the agent a whole computer and and like nix is another good tool here because

622
00:56:58,420 --> 00:57:03,140
 it doesn't make i mean one thing agents love to do is make mutations and then leave them

623
00:57:03,140 --> 00:57:08,580
 yeah and forget right there these things are really good at forgetting what they've done i mean that's

624
00:57:08,580 --> 00:57:14,180
 yeah they they are memory is the main problem right now yeah and uh and in a sense i think it always

625
00:57:14,180 --> 00:57:19,380
 will be because it's or maybe not always but yeah they don't have like a sense of experience and like

626
00:57:20,500 --> 00:57:25,300
 you know that like oh it's starting to get messy here like they haven't demonstrated that yet

627
00:57:25,300 --> 00:57:31,220
 uh so yeah like if you could if you could have it all live on like all the configuration live

628
00:57:31,220 --> 00:57:38,020
 and a git repo and then all the data be just synced at all times uh that would be cool and it can just

629
00:57:38,020 --> 00:57:44,740
 like sort of install everything it needs and so you know as long as it doesn't have to do permission

630
00:57:44,740 --> 00:57:50,500
 systems like email or something it can just set everything up it needs so i think that could be

631
00:57:50,500 --> 00:57:56,660
 kind of cool um so this is kind of like what i would want to uh i want to make like a little service for

632
00:57:56,660 --> 00:58:02,340
 people to like set up an agent in this environment and then you know this little thing i set up it's

633
00:58:02,340 --> 00:58:08,100
 like you know if somebody wants to you know do a rust build i could like sell them more compute just

634
00:58:08,100 --> 00:58:15,700
 right then like it's cached on that machine right and uh you know if you need a little more disc like

635
00:58:15,700 --> 00:58:21,700
 just these are all just like you know i think that's where it's heading anyway i mean dimmy the

636
00:58:21,700 --> 00:58:27,700
 the guy you mentioned just before he that's how he is operating as well just giving the agents more or

637
00:58:27,700 --> 00:58:33,940
 less full access to certain parts of the system yeah and for my setup as well like i i have multiple

638
00:58:33,940 --> 00:58:41,460
 machines uh with multiple claws on them multiple lobsters on them and i have one like god-like

639
00:58:41,460 --> 00:58:46,580
 machine that just can ssh into everything and just fix stuff where it's similar to your backup idea

640
00:58:46,580 --> 00:58:52,900
 where it's just like if the thing destroys itself yeah i go one level higher and and say like okay this

641
00:58:52,900 --> 00:58:59,460
 one's down just fix it yeah and it so far it has always worked so it's like the uh yeah it's like

642
00:58:59,460 --> 00:59:08,100
 god's uh god's angels like his uh hierarchy of angels exactly exactly yeah yeah an angel falls and then

643
00:59:08,100 --> 00:59:12,420
 an archangel blows a little kiss exactly he stands back up

644
00:59:12,420 --> 00:59:20,660
 yeah so that's that's i don't know that's another thing where uh the reason why you went into all that

645
00:59:20,660 --> 00:59:26,180
 is in part because of the monetization yeah i mean i think i think the best way to monetize stuff is

646
00:59:26,180 --> 00:59:30,420
 like people will pay for their tamagotchi their intelligent tamagotchi that's what an agent is

647
00:59:30,420 --> 00:59:39,140
 oh absolutely yeah the verbs people use like hatch and grow yeah uh betrays more of a more interest

648
00:59:39,140 --> 00:59:43,460
 than in a normal computer right and so oh absolutely from a sales point of view we should leverage that

649
00:59:43,460 --> 00:59:47,060
 like this is something that people are going to pay money for this will just we're going to be able to

650
00:59:47,060 --> 00:59:51,380
 sell them what we've always wanted to sell them yeah but they'll be willing to the the memory part

651
00:59:51,380 --> 00:59:56,580
 will be solved eventually that's that's i think in part why open claw became so successful because

652
00:59:56,580 --> 01:00:01,780
 because it allows to shoot yourself in the foot and it provides memory yeah yeah as rudimentary as it

653
01:00:01,780 --> 01:00:08,740
 is yeah and then you just the thing learned and it's yours now and it's useful for you but and yeah

654
01:00:09,460 --> 01:00:14,900
 yeah i'm still a bit of a like i've had open claws since like december but i've never really like gotten

655
01:00:14,900 --> 01:00:19,860
 into it partially because like i'm like just a hardcore vibe coder like like

656
01:00:26,100 --> 01:00:34,260
 that's not something you see every day i'll have to i'll just

657
01:00:34,260 --> 01:00:44,980
 yeah we just went to walk by some acrobats what is it is it like that's that's acro yoga right i

658
01:00:44,980 --> 01:00:58,260
 don't know uh strange um yeah so i keep side like i haven't sidetracked you this hard ever

659
01:00:58,260 --> 01:01:05,220
 i keep sidetracking you with everything yeah it's okay it's just who you are you can't help it

660
01:01:06,420 --> 01:01:13,460
 you're past you're past the age where you can change that's very true but in my defense i hatched

661
01:01:13,460 --> 01:01:17,940
 two new ones you know and they'll be different yeah oh so i was saying like i'm i'm a little bit

662
01:01:17,940 --> 01:01:22,500
 of a boomer on this still because i don't really i mean in a sense i try to you know design my whole

663
01:01:22,500 --> 01:01:26,900
 life so i don't have to do a bunch of tasks yeah you know and like open claws are really good at task

664
01:01:26,900 --> 01:01:31,780
 completion like just doing tasks and you don't have any tasks free calendar always well i'm

665
01:01:32,980 --> 01:01:36,260
 the ones that i do have i'm procrastinating so heavily that i forget they're there

666
01:01:36,260 --> 01:01:43,540
 no uh i want to i want to get i want to set up a hermes uh or you know try a few of these a little

667
01:01:43,540 --> 01:01:50,820
 more seriously um yeah i need to like i need to like have more tasks i was going to try to create

668
01:01:50,820 --> 01:01:54,500
 more content and like if i was doing that there's all kinds of things i could have the open claw be

669
01:01:54,500 --> 01:02:01,700
 in charge of you know um so i have to like engineer a bit of an excuse to need an open claw yeah

670
01:02:02,500 --> 01:02:11,780
 what what i landed on now and i'm sure it will kind of evolve again but one agent per project yeah

671
01:02:11,780 --> 01:02:17,220
 and it just learns everything about the project and it also runs its own nostr account and so i just dm

672
01:02:17,220 --> 01:02:22,900
 the project yeah and so i'm just constantly in conversations your vibe they're vibe coding for you

673
01:02:22,900 --> 01:02:32,180
 yeah i'm just in conversations with my projects basically um very simple setup it's it's kind of like

674
01:02:32,980 --> 01:02:39,140
 let's say you're elon musk and you have like 20 different companies you just talk to the ceo so

675
01:02:39,140 --> 01:02:45,700
 you're talking to hr2 now yeah it's i don't have an abstraction hierarchy below that yeah so it's just

676
01:02:45,700 --> 01:02:52,180
 one one one open claw yeah it can spawn its own up sub agents and it will but it's just very simple because

677
01:02:52,180 --> 01:02:58,260
 then i you know like that sounds nice especially because like i go like i really like having the

678
01:02:58,260 --> 01:03:03,140
 fine-grained control of like actually using the like the codex itself yeah for something like shadow

679
01:03:03,140 --> 01:03:08,980
 like you can't you can't communicate high level requirements and not look at the session and not look at the code on that

680
01:03:08,980 --> 01:03:13,700
 yeah but for a lot of other things like that i would like i think that is good enough you know yeah for

681
01:03:13,700 --> 01:03:18,020
 no solutions for example that's how i do it yeah and just everything it needs in terms of memory is

682
01:03:18,020 --> 01:03:23,460
 either in in github or in the chat history yeah or in the in the raw files themselves in the you know

683
01:03:23,460 --> 01:03:29,860
 transcripts and yeah show notes and what have you and yeah it's just like most of the stuff is

684
01:03:29,860 --> 01:03:36,180
 almost fully automated and then i just follow up with some chat messages or voice notes voice memos

685
01:03:36,180 --> 01:03:43,300
 and that's that's where i landed on right now for having like 10 15 different little projects that i just

686
01:03:43,300 --> 01:03:49,460
 that's my that's my you know like waste of time yeah playing mode i just want to play around with the

687
01:03:49,460 --> 01:03:56,740
 things to learn how it behaves learn kind of what various models are good at what like when is it

688
01:03:56,740 --> 01:04:03,460
 actually kind of sucking you in and leading you to ai psychosis and you're just like fully wasting your

689
01:04:03,460 --> 01:04:09,700
 time and going insane yeah and when is it actually helpful yeah and replaces for me like for no solutions

690
01:04:09,700 --> 01:04:14,580
 for example it it replaces a a production team yeah to some degree i mean not that there is a lot of

691
01:04:14,580 --> 01:04:24,740
 production but it's just continue your speech continue your speech no but it's just kind of you know like

692
01:04:24,740 --> 01:04:31,460
 all that's it's it's just busy work and like you said before open again open claw open claw is very

693
01:04:31,460 --> 01:04:36,500
 good at task completion which is kind of doing busy work yeah and just you know like linking stuff and

694
01:04:36,500 --> 01:04:42,340
 doing transcripts and doing like a cover image and just uploading things and blah and updating the feed

695
01:04:42,340 --> 01:04:47,140
 it's just i don't have to do that anymore it's just like here is the mp3 file do everything and that's

696
01:04:47,140 --> 01:04:51,860
 it like i don't do more than that and yep and that that can obviously be automated because i have

697
01:04:51,860 --> 01:04:56,020
 file watches and if an mp3 file lands in that folder then it just does things and yeah that's it

698
01:04:56,020 --> 01:05:00,420
 so i'm trying to think of your last question your last question is about networks

699
01:05:00,420 --> 01:05:06,420
 yes that's how we yes you hijacked no i was i was trying to pick your brain on on fips because this

700
01:05:06,420 --> 01:05:11,620
 cohort like as you know like i don't really follow the rules very well uh so of course this the first

701
01:05:11,620 --> 01:05:17,700
 even as captain of week two you were just like everything yeah uh my my week was really good

702
01:05:17,700 --> 01:05:24,580
 we had really good demos i haven't made the captain's notes yet but uh how dare you justin how dare

703
01:05:24,580 --> 01:05:30,500
 you but uh give me the markdown file i need to feed my lobsters i haven't really uh complete i haven't

704
01:05:30,500 --> 01:05:36,420
 really dived into fips as much as i wanted yet but i have a few like high level thoughts and like one

705
01:05:36,420 --> 01:05:42,500
 of them is that you know i think it will be really neat i mean the basic the pitch is basically that

706
01:05:42,500 --> 01:05:53,780
 like um everything becomes like bit chat in a sense where like like all internet like every network

707
01:05:53,780 --> 01:05:58,580
 device can have those sorts of functionality right so it's like i could basically build

708
01:05:59,460 --> 01:06:03,860
 bit chat into my operating system but it's like kind of a more standard protocol that was written

709
01:06:03,860 --> 01:06:08,500
 protocol first from somebody who has like 50 years or 40 years of experience doing this type of stuff

710
01:06:08,500 --> 01:06:13,220
 you know like jonathan was like he's told us a few times i don't think i'm telling tales out of

711
01:06:13,220 --> 01:06:18,820
 school where he's like you know uh he's kind of ready to retire and uh and then he came to this little

712
01:06:18,820 --> 01:06:22,980
 meetup in costa rica yeah he's like you know i've only felt there's like 30 of us in a room

713
01:06:22,980 --> 01:06:27,700
 sponsored by primal and nvk and he's like you know i've only felt a room like this twice in my life

714
01:06:27,700 --> 01:06:33,220
 one of them was the early 90s with the cypherpunks like the true like the original people which he

715
01:06:33,220 --> 01:06:40,020
 knew because he was there and then early 2010s with uh for the early you know early years of bitcoin

716
01:06:40,020 --> 01:06:45,620
 and he's like that's the only other time i felt this kind of a vibe and uh i just had to get back

717
01:06:45,620 --> 01:06:50,820
 into it right i couldn't i just couldn't watch yeah phipps brought him out of retirement yeah exactly so uh

718
01:06:50,820 --> 01:06:56,340
 uh but yeah so it's like you know so it's it's very good that you know like he it's someone like him

719
01:06:56,340 --> 01:07:00,180
 creating it and it's kind of protocol first whereas bitchat was more like you know starting

720
01:07:00,180 --> 01:07:07,460
 a ct a ceo of a beloved ceo of a fortune 500 company took a weekend to make a ux yeah and they've

721
01:07:07,460 --> 01:07:13,940
 been sort of going from ux to protocol yeah and this is kind of going in the in the opposite

722
01:07:13,940 --> 01:07:19,380
 opposite direction right we still don't even have a decent ux for it yet yeah you know um but it'll get

723
01:07:19,380 --> 01:07:22,900
 there and i think the two will kind of meet in the middle and so that's kind of the idea is that you

724
01:07:22,900 --> 01:07:29,540
 can have any of these little little mini networks uh running over any transport you like you know

725
01:07:29,540 --> 01:07:35,780
 we've been trying to fly you know point antennas at things yeah i think the high level idea is like

726
01:07:35,780 --> 01:07:40,900
 we talked about this before with dns and ip addresses being permissioned yeah and the reason is because

727
01:07:40,900 --> 01:07:47,700
 they are identities in some sense identity systems yeah that don't operate in cryptographic space which means

728
01:07:47,700 --> 01:07:53,540
 they have to be permissioned it's like dns is the clearest example of this but the ip addresses

729
01:07:53,540 --> 01:07:58,740
 because there is so few of them yeah um have the same issue like it's exactly the same issue and so

730
01:07:58,740 --> 01:08:03,060
 you need a consortium of people that comes together and there's like 12 people in the room that you

731
01:08:03,060 --> 01:08:10,420
 know have like ceremonies that yeah will tell the world who is it actually that is in control of google.com

732
01:08:10,420 --> 01:08:17,060
 and all those kind of things and and what fibs does it is it completely removes all that it moves fully

733
01:08:17,060 --> 01:08:23,300
 into the cryptographic space yeah so you're just talking to npubs basically everyone everyone

734
01:08:23,300 --> 01:08:28,500
 every addressable point is in cryptographic space and whether you use

735
01:08:29,780 --> 01:08:36,420
 like the nice thing is that it's both a fundamental protocol right and it can also act as an overlay

736
01:08:36,420 --> 01:08:43,060
 so you can go i can say hey connect me to whatever justin's machines and pub yeah and it can still go

737
01:08:43,060 --> 01:08:52,260
 over dna dns and ip addresses um which i think is what makes it it gives it a shot at actually yeah makes

738
01:08:52,260 --> 01:08:57,940
 it actually useful in some sense like otherwise every other attempt um or many other attempts that

739
01:08:57,940 --> 01:09:03,300
 i've seen so far it's just like okay you have to mesh networkify the world instantly otherwise it just

740
01:09:03,300 --> 01:09:09,460
 won't work yeah and this is a very nice kind of yeah meet in the middle approach attacking it from both

741
01:09:09,460 --> 01:09:16,420
 angles you know like that was another one marty uh malmy you know a legend uh was was here yeah he

742
01:09:16,420 --> 01:09:21,220
 was showing off nostril vpn and i played with it i set it up uh and so like that's another thing that

743
01:09:21,220 --> 01:09:24,980
 is appealing is like hey if i just could have like a little phips overlay just for my own machines my

744
01:09:24,980 --> 01:09:28,340
 own little tail scale yeah absolutely yeah like this would be awesome right and so it's like i don't

745
01:09:28,340 --> 01:09:34,100
 know if i if we could have something like this just like me and my friends in austin doing something

746
01:09:34,100 --> 01:09:38,420
 right like like you know you can you can you can deploy these things in like a small scale and if

747
01:09:38,420 --> 01:09:43,060
 the grid goes down you don't care that much it's similar to the solar and being you know like

748
01:09:43,060 --> 01:09:47,140
 self-sufficient in that way it's self-sufficient networking right that's that's what it should be and i think

749
01:09:47,140 --> 01:09:50,900
 you could always build more centralized things on top of them you know that could maybe get to web

750
01:09:50,900 --> 01:09:56,500
 scale or whatever uh you know but that's kind of a longer term thing and and until then we have our

751
01:09:56,500 --> 01:10:00,740
 own little cottage industry internet that we can play yeah and for certain communities i think it can

752
01:10:00,740 --> 01:10:06,020
 be super helpful like if you're living off grade already in a bigger community yeah um like you know

753
01:10:06,020 --> 01:10:11,540
 cobra door the toll gate guy that's walking in front of us there um you know he he has access to

754
01:10:11,540 --> 01:10:18,340
 these communities and they would that's actually a a big internet yeah i mean i mean i don't know what

755
01:10:18,340 --> 01:10:22,740
 they're doing now but uh this is one of their goals as well yeah to tap internet set these things all

756
01:10:22,740 --> 01:10:29,380
 over up all over africa stuff yeah fatty obviously is on the on the kind of app uh yeah level layer and

757
01:10:29,380 --> 01:10:33,540
 but a lot of times a lot of places where they want to work is like the internet is actually the limiting

758
01:10:33,540 --> 01:10:38,740
 factor yeah so it's like if you could set up a starlink and have like you know their own little network that

759
01:10:38,740 --> 01:10:44,260
 works inside the walls of the village yeah that would be an unlock yeah i mean again back to e-cash being

760
01:10:44,260 --> 01:10:50,100
 offline that's kind of a killer feature for for the yeah the offline offline mostly communities you

761
01:10:50,100 --> 01:10:54,500
 know mostly offline communities and then someday if we're successful enough we can join them and be

762
01:10:54,500 --> 01:11:01,860
 offline mostly and not have to deal with this health care we call the internet ai all these things we

763
01:11:01,860 --> 01:11:06,980
 fucking hate i mean we're pretty offline right now yeah yeah there isn't cell signal here i think yeah

764
01:11:06,980 --> 01:11:12,740
 deep in the woods yeah you wait to see how offline i am when the shadow starts booting on its own

765
01:11:14,340 --> 01:11:20,420
 levels of offline even if the message gets through the fonts won't render

766
01:11:20,420 --> 01:11:29,300
 i mean in some sense you're riding a certain wave and lots of things will play in your favor i think

767
01:11:29,300 --> 01:11:33,620
 the the permissioned internet and the permissioned ecosystems will just get worse and worse and worse

768
01:11:33,620 --> 01:11:40,180
 yeah the models will just get better and better and better yeah vibe coding sorry agentic engineering

769
01:11:40,180 --> 01:11:47,540
 will just get better and better and better and um so yeah with the new sets of models

770
01:11:47,540 --> 01:11:50,900
 right around the corner basically and all all those kind of things wrapping up

771
01:11:50,900 --> 01:11:55,140
 who knows who knows what you can build out in the next 18 months yeah

772
01:11:57,220 --> 01:12:03,940
 yeah we'll see we'll see should we leave it at that sure now we're walking uphill they're huffing and

773
01:12:03,940 --> 01:12:08,180
 puffing at downhill i mean when we're walking perpendicular to where everyone else is walking

774
01:12:08,180 --> 01:12:14,020
 oh yeah that's true we're losing the group yeah but yeah no i mean i'm i feel very blessed that i can

775
01:12:14,020 --> 01:12:19,940
 just you know tap into your brain from time to time yeah and that you're actually you know uh well i feel

776
01:12:19,940 --> 01:12:23,860
 blessed that i could come to sec it really kind of rejuvenated me the last time the first time i came

777
01:12:23,860 --> 01:12:30,580
 and this is i want to no i want to tell the one guy uh half a guy the half a guy who's left uh we'll

778
01:12:30,580 --> 01:12:35,460
 round up you know it's uh it kind of rejuvenated me got me interested again and gave me a bunch of

779
01:12:35,460 --> 01:12:38,660
 different perspectives brought you back from retirement a little bit yeah i mean all the

780
01:12:38,660 --> 01:12:43,140
 things that i've been doing like at this cohort like it's starting to come together like in pika we

781
01:12:43,140 --> 01:12:50,580
 didn't talk about pika um like trying to fix the trying to figure out a good a good actual user usable

782
01:12:50,580 --> 01:12:57,300
 way of doing messaging on nostr still doesn't exist um uh there were all things that i kind of started

783
01:12:57,300 --> 01:13:02,100
 playing around with at the first cohort continued at the second one and so it's been like a very

784
01:13:02,100 --> 01:13:09,460
 linear yeah growth trajectory the whole time yeah uh so a lot of the ideas actually started here uh so

785
01:13:09,460 --> 01:13:13,460
 yeah yeah i'm glad we i'm glad we brought you back not only will i'll be back but i'm probably

786
01:13:13,460 --> 01:13:18,660
 going to try to scam you guys into somehow doing one in the in the u.s or let's go or canada or

787
01:13:18,660 --> 01:13:23,220
 something on some off the grid ranch yeah yeah yeah yeah i know we were talking about ranches we're

788
01:13:23,220 --> 01:13:29,220
 talking about maybe doing one where we like rent a couple rvs moving moving cohort yeah yeah and

789
01:13:29,220 --> 01:13:34,180
 like this is like meet up in the middle of nowhere get some starlinks yeah build some tents yeah we'll

790
01:13:34,180 --> 01:13:38,980
 probably have to call burning man start something else for legal reasons sovereign rving

791
01:13:38,980 --> 01:13:46,820
 but no one will come to that that's the problem no but i'm i'm really glad that you came and and

792
01:13:46,820 --> 01:13:50,820
 stayed and i think we told that this before but you moved your flight like four times or something

793
01:13:50,820 --> 01:13:56,980
 and now you're back again and uh yeah we'll definitely continue and i i love that just bringing

794
01:13:56,980 --> 01:14:01,060
 some like bringing all the insane people together and locking them into a room for

795
01:14:01,780 --> 01:14:06,020
 a couple weeks actually produces stuff it actually produces novel stuff and that's that's kind of the

796
01:14:06,020 --> 01:14:13,780
 goal yeah i know it uh it it uh i don't know it's funny like the i mean the fifth like the for arian

797
01:14:13,780 --> 01:14:16,980
 and cobrador i mean the first time i heard i was joking about this like the first time i heard them

798
01:14:16,980 --> 01:14:21,220
 talking about this stuff i'm like these guys are retarded absolutely hundred percent routers and

799
01:14:21,220 --> 01:14:30,180
 balloons and yeah i mean no thanks for confidentially cobrador still living yeah yeah yeah yeah we all

800
01:14:30,180 --> 01:14:35,940
 agree we all agree but yeah like i mean with a lot of them it's like the ideas the first time i hear

801
01:14:35,940 --> 01:14:42,180
 them are like i don't know about that and then it grows on you a little bit it's like you need to uh

802
01:14:42,180 --> 01:14:47,700
 yeah i don't know it's like uh because then i will talk to some of my normie friends about these things and

803
01:14:47,700 --> 01:14:52,500
 it's like if i in the rare instances i don't heavily self-censor myself they'll be like what

804
01:14:52,500 --> 01:14:56,500
 the are you talking about yeah yeah you know yeah it's just too much and too too out there but but

805
01:14:56,500 --> 01:14:59,380
 that's kind of what you know i've always wondered was like what was it like to be involved in the

806
01:14:59,380 --> 01:15:04,100
 internet in the 80s or the 90s and like complete insane ideas yeah it's like that's so i think it's

807
01:15:04,100 --> 01:15:08,260
 kind of good i mean that's kind of the balance is how do you engage with these kind of insane ideas

808
01:15:08,260 --> 01:15:12,740
 well not going insane yourself yeah which is always feels like i'm always a little close

809
01:15:12,740 --> 01:15:20,500
 like i could psychosis doesn't help i i could i i could go insane i know that about myself oh same

810
01:15:20,500 --> 01:15:25,540
 same yeah yeah so um so that's kind of the thing is like how do you but that also makes me good at

811
01:15:25,540 --> 01:15:30,500
 these sorts of things the the way i see it is like we are like the whole sovereign engineering process

812
01:15:30,500 --> 01:15:36,100
 is basically throwing shit at the wall and and kind of you have to step away and there's just a pile of

813
01:15:36,100 --> 01:15:41,940
 shit forms yeah at the bottom of the wall yeah and then like if you let enough time pass a little

814
01:15:41,940 --> 01:15:46,180
 flower might go out of it and then focus really on the flower yeah yeah yeah you know like we had

815
01:15:46,180 --> 01:15:54,420
 on the flower yeah exactly exactly exactly but we had like uh we're in in cohort number seven now

816
01:15:54,420 --> 01:16:02,260
 and like 21 participants each yeah um and like about 100 demos each cohort like it's 700 demos that's

817
01:16:02,260 --> 01:16:08,820
 700 new ideas and crazy things that people showed and out of those there's maybe like let's be generous

818
01:16:08,820 --> 01:16:15,620
 10 20 good ones something like that yeah and three four five really shine you know and and that's fips

819
01:16:15,620 --> 01:16:20,980
 again like i think everyone agrees that's that's a great one yeah blossom obviously you know insight

820
01:16:20,980 --> 01:16:26,420
 yeah those kind of things um you know sap store came out of first first chord as well and and just

821
01:16:27,540 --> 01:16:31,700
 most of the crazy ideas obviously don't make it but that's just how it goes and we just want to

822
01:16:31,700 --> 01:16:37,460
 encourage people to actually come and actually lock themselves into a room and actually try the crazy

823
01:16:37,460 --> 01:16:41,380
 things it's not a vc thing obviously but it does have the same kind of mechanics where you have to

824
01:16:41,380 --> 01:16:47,460
 just spread a million bets out yeah and it's a numbers game yeah but just and the one thing i observe

825
01:16:47,460 --> 01:16:53,140
 is like i i i don't think i could make it six weeks this time because uh like i'm already overwhelmed on

826
01:16:53,140 --> 01:17:01,620
 week two like i you know yeah this one was like it's like this one felt like like like i'm uh like i'm

827
01:17:01,620 --> 01:17:06,420
 what do you call it free base like injecting it right into the veins or whatever right like it's

828
01:17:06,420 --> 01:17:13,300
 it uh yeah i don't know i mean it's it's a shorter one but it's almost after two weeks i'm almost almost

829
01:17:13,300 --> 01:17:18,340
 brain dead just brain dead yeah yeah i mean and and also i mean my my situation is a little different

830
01:17:18,340 --> 01:17:22,500
 where i was kind of pregnant with two projects when i came here and now i'm like severely pregnant i need to

831
01:17:22,500 --> 01:17:26,580
 ship these uh i actually spent most of the week trying to figure out how to ship pika last week

832
01:17:26,580 --> 01:17:31,620
 because i'm like i shouldn't just start a new big project i need to ship the pika what's holding

833
01:17:31,620 --> 01:17:36,020
 i gave i gave it to my captain's talk was like i was talking about how kind of i have a icicles i'm

834
01:17:36,020 --> 01:17:40,500
 doing all these things and i'm like trying to figure out what direction is up and i'm trying to be really

835
01:17:40,500 --> 01:17:44,980
 ambitious and how do you manage shipping i'm like you know what after the thing is like i had something

836
01:17:44,980 --> 01:17:51,860
 really interesting with pika i had two power users my mom and gg and it's like this is what i've always wanted

837
01:17:51,860 --> 01:17:57,700
 and then i got distracted and we didn't even talk to each other no you were you were just power using

838
01:17:57,700 --> 01:18:05,860
 it and so i don't know i mean uh i need to return to that i think uh i think i don't know if marmot is

839
01:18:05,860 --> 01:18:16,020
 the way um like it's unclear like it's i uh i don't what makes you say that a few things like one is like

840
01:18:16,020 --> 01:18:21,140
 i don't i don't know if for a private encrypted messenger i don't know if we should be uploading

841
01:18:21,140 --> 01:18:26,820
 stuff to public infrastructure this is jesus one of jesus's points yeah you know it's like you're you're

842
01:18:26,820 --> 01:18:31,060
 it's it's it's easily queryable yeah right like i could query every single message you posted from

843
01:18:31,060 --> 01:18:36,340
 marmot marmot app on these nostril relays and it's like so you do leak more metadata in that sense than

844
01:18:36,340 --> 01:18:40,260
 normal the big benefit to me is like as an app developer you don't have to stand up a any sort of

845
01:18:40,260 --> 01:18:47,620
 infrastructure but i don't know like why not just have a centralized chat server that's easily deployable

846
01:18:47,620 --> 01:18:55,220
 um that can be per group and can be easily swapped out and uh and just use nostril identities or why

847
01:18:55,220 --> 01:19:02,100
 not have a server just for signaling um setting up groups and ordering messages have you looked into

848
01:19:02,100 --> 01:19:09,300
 zero zero zero x chat at all uh no what is that i've heard of it yeah that's a that's a um

849
01:19:09,300 --> 01:19:19,220
 basically goes down that path and doesn't have persistent identities at all so every no every

850
01:19:19,220 --> 01:19:29,220
 every kind of chat session you initiate you get a new key pair okay basically yeah um and as far as i

851
01:19:29,220 --> 01:19:34,980
 know they they they provide their own infrastructure as well but they make it very easy to run your own

852
01:19:34,980 --> 01:19:42,740
 and it's kind of uh also relay based and onion routing based so yeah um if you if you kind of want to

853
01:19:42,740 --> 01:19:47,460
 move away from armor that's something i would definitely look into because we can easily over overlay

854
01:19:47,460 --> 01:19:53,380
 identity on it yeah in some sense yeah if you want to i mean the founder doesn't want to have any

855
01:19:53,380 --> 01:19:59,060
 persistent identity in the system itself which i respect in some sense looking looking forward i think

856
01:19:59,060 --> 01:20:04,980
 hosting services is going to get easier like i do agree i have an infrastructure agent it's literally

857
01:20:04,980 --> 01:20:11,220
 like one one phone call away so i think i think we're gonna like the way we use nostr you know like

858
01:20:11,220 --> 01:20:15,780
 if my agent wants to do something one option is to figure out how to do the kinds and everything on

859
01:20:15,780 --> 01:20:21,460
 nostril it could also just like write a little http server and deploy it yeah right like so i think

860
01:20:21,460 --> 01:20:27,060
 that the the days of like nostril as firebase where it's like oh this is the only place my backend can live

861
01:20:27,060 --> 01:20:35,540
 is going to slowly go away maybe um and uh yeah so the other thing with the problem with marmot is like

862
01:20:35,540 --> 01:20:41,860
 so mls is like this you know this uh really complicated but really has you know nice properties

863
01:20:41,860 --> 01:20:48,660
 uh specification for how to do encrypted messaging and so it's designed with two central services in

864
01:20:48,660 --> 01:20:55,300
 you know in mind an authentication service and like a uh coordinator or ordering service uh and so we

865
01:20:55,300 --> 01:21:02,820
 already have the authentication service it's just not uh you know the cryptography of of of nostril or

866
01:21:02,820 --> 01:21:09,220
 keys but we don't have the ordering service and so uh this is like ordering the the the messages and

867
01:21:09,220 --> 01:21:13,940
 the commits and you know people getting added to a group removed to a group rotating the keys all that

868
01:21:13,940 --> 01:21:18,260
 stuff like if you're ever in a marmot chat and it just kind of dies that's what's happened is like

869
01:21:18,260 --> 01:21:22,900
 they're trying to do this they're trying to do this ordering stuff client side when the protocol was

870
01:21:22,900 --> 01:21:28,260
 designed for a server and you know in the spec they say it's possible but extremely extremely difficult

871
01:21:28,260 --> 01:21:35,940
 and you know so they they you know marmot is forging ahead with lots of features you know like

872
01:21:35,940 --> 01:21:39,060
 they're figuring how to do multi-device and push notifications all these other things they still

873
01:21:39,060 --> 01:21:45,140
 haven't figured out how to have the chats not fall apart um and they don't have very paranoid testing

874
01:21:45,140 --> 01:21:50,260
 of any of this you know like in the last cohort i i understand that they tried to set up a big chat

875
01:21:50,260 --> 01:21:54,180
 with everyone in in the group with one of the official implementations and it just it fell over when

876
01:21:54,180 --> 01:21:59,300
 i had like 15 people and that's not good because the entire point of marmot of mls is that it scales

877
01:21:59,300 --> 01:22:03,940
 better than double ratchet so like you can have a 2 000 person group with mls you can't with double ratchet

878
01:22:03,940 --> 01:22:08,980
 which is what signal uses yeah uh and so that's another uh direction is like marty malmy and

879
01:22:08,980 --> 01:22:16,180
 his team are trying to make a double reaction right uh and so it doesn't have as good of like

880
01:22:16,180 --> 01:22:20,180
 theoretical properties but i think you know right now it may be the case that the double ratchet

881
01:22:20,180 --> 01:22:24,820
 scales better than mls because the mls falls apart for coordinate coordination reasons i keep coming

882
01:22:24,820 --> 01:22:31,460
 back to the same point like an encrypted group chat with 2 000 people is nonsensical anyway it's just you

883
01:22:31,460 --> 01:22:36,340
 you have to make it work for small groups it has to work in practice yeah because like it's a 2 000

884
01:22:36,340 --> 01:22:40,980
 people group chat is public by default yeah like someone someone is screencasting the whole thing i

885
01:22:40,980 --> 01:22:45,780
 think it's it's always a spectrum though like yeah it is you know you can say stuff on discord you know

886
01:22:45,780 --> 01:22:49,540
 there's a lot of stuff that gets said on discord that people don't lose their jobs over that they

887
01:22:49,540 --> 01:22:53,540
 would lose their jobs over if they said it on twitter yeah fair enough yeah that's that is a good

888
01:22:53,540 --> 01:22:57,700
 point a lot of people have lost a lot of stuff yeah i'm saying stuff on no no that that's that is a very

889
01:22:57,700 --> 01:23:02,420
 good point that's why everyone is in yeah yeah but but but it's still you know kind of you know like

890
01:23:02,420 --> 01:23:06,580
 people say it depends on the people because it's just like screenshots and and sure but does your

891
01:23:06,580 --> 01:23:14,340
 employer find out about that yeah no i i fully agree with that it's a spectrum um it's a spectrum

892
01:23:14,340 --> 01:23:22,980
 assessment like i i i fully did it 30 seconds ago no no i do fully like i i but it's it is still a spectrum

893
01:23:23,700 --> 01:23:30,020
 you know i get you to reverse back you know i'm not an llm i don't go down the path and never go

894
01:23:30,020 --> 01:23:40,660
 back no you just you're absolutely right no i agree with the i i i think we we i agree with your

895
01:23:40,660 --> 01:23:43,780
 point too i'll just interrupt you here because i haven't you i want to give you a little bit of your

896
01:23:43,780 --> 01:23:48,260
 own medicine yeah uh i agree that what we need is groups that can work for five or ten people i

897
01:23:48,260 --> 01:23:52,420
 mean we really need something that works really well for a clanker and has backwards compatibility and

898
01:23:52,420 --> 01:23:57,300
 doesn't because that was the other issue with mine i don't think we need to theory craft over like

899
01:23:57,300 --> 01:24:01,780
 five thousand people group chats and have it theoretically be perfect but it doesn't work in

900
01:24:01,780 --> 01:24:05,540
 practice for five people yeah one it's also like why do we need to like why do we need to send the

901
01:24:05,540 --> 01:24:11,300
 messages over nostrilay it's like yeah you ended up like that's another thing about phips is like

902
01:24:11,300 --> 01:24:15,860
 phips is like no it's the key that the key is the more important thing than the relay like the relay was

903
01:24:15,860 --> 01:24:23,140
 like maybe it's not you know maybe that goes away by the way that's what makes snip 17 for me like

904
01:24:23,140 --> 01:24:28,500
 unreliable to a degree because it just needs to you need to connect to all the relays to get all the

905
01:24:28,500 --> 01:24:34,820
 messages to get all the order ordering to get all the you know like dm inbox yeah relays connections up

906
01:24:34,820 --> 01:24:40,740
 and so on yeah and it's just it's kind of a mess like for for my clients i need to initiate a chat

907
01:24:40,740 --> 01:24:45,700
 session load the profile manually half the time that connects to all the relays and then i can

908
01:24:45,700 --> 01:24:50,340
 start chatting after like 30 seconds or whatever yeah and there's lots of stuff that that still

909
01:24:50,340 --> 01:24:54,660
 doesn't work right yeah so that's what i what i want to do is is is relax a little bit of the

910
01:24:54,660 --> 01:25:01,140
 maybe the decentralization that marmot strives for in addition to taking a 30 cut you want to

911
01:25:01,140 --> 01:25:10,100
 relax the decentralization minimum that's the that's the os cut the app cut oh wow oh we were i've been

912
01:25:10,100 --> 01:25:13,300
 here a few times i've never actually seen the water i always thought we were higher up i'm a little

913
01:25:13,300 --> 01:25:20,580
 disappointed to be honest it's it's sometimes it's like uh crystal clear blue and then it's just

914
01:25:20,580 --> 01:25:26,020
 breathtaking that would be amazing but right right now we're in the clouds a bit um but relaxing the

915
01:25:26,020 --> 01:25:30,340
 decentralization requirement in the sense of you initiate a chat session and you just go over there's one

916
01:25:30,340 --> 01:25:36,660
 server yeah central server uh and you know i think that's fine for a while uh i want something that's

917
01:25:36,660 --> 01:25:43,860
 really reliable and i mean to some degree that's what chat relays are supposed to be for you just

918
01:25:43,860 --> 01:25:49,780
 run your own and what are chat relays it's basically special purpose relays that are only four dms and

919
01:25:49,780 --> 01:25:54,820
 you just yeah you announce which ones you use and then it routes over this kind of central point where

920
01:25:54,820 --> 01:26:00,180
 i'm torn here is like okay so now we're just building a chat server so like why do we patch it on

921
01:26:00,180 --> 01:26:04,900
 top of nostril yeah i do agree like that's well i don't know if i agree or not but it's just like

922
01:26:04,900 --> 01:26:11,140
 um to me nostril is a broadcasting system and we shouldn't we shouldn't like you know monkey patch

923
01:26:11,140 --> 01:26:16,260
 and we should so you broadcast you're like hey this is how to find me on this this is the relay i'm on

924
01:26:16,260 --> 01:26:20,740
 this is the chat server i use right so use it for those types of things if you're doing an audio call

925
01:26:20,740 --> 01:26:25,460
 like yeah look there for you we're not going to do audio and video calls over relays right like we're

926
01:26:25,460 --> 01:26:30,340
 gonna use we're gonna use mock obviously i have a half implementation of that too you have lots of

927
01:26:30,340 --> 01:26:39,220
 half implemented things yeah so that's why i gotta get the hell out of here i gotta get the hell out of

928
01:26:39,220 --> 01:26:44,420
 software engineering i have to ship all right you know all right violent yeah violent you have to ship

929
01:26:44,420 --> 01:26:51,060
 violent violent shipping like uh yeah there's gonna be blood everywhere no that's oh my god oh my god

930
01:26:51,060 --> 01:26:59,380
 shooting yourself in the foot shipping violently that's not only new new os coming i'm still

931
01:26:59,380 --> 01:27:04,180
 disappointed that it isn't called moon os i'm probably i could call it moon os but that's kind

932
01:27:04,180 --> 01:27:09,460
 of a pretentious shadow shadow it is no well these are all provisional i was calling it fundroid for a

933
01:27:09,460 --> 01:27:15,380
 while but that sounded not serious no that's not that's not good no shadow is fine and how does how

934
01:27:15,380 --> 01:27:21,940
 is how is sovereign engineering from your point of view now i'm glad that i don't have to do absolutely

935
01:27:21,940 --> 01:27:27,700
 everything myself anymore yeah that's good because that literally killed me yeah i was kind of like

936
01:27:27,700 --> 01:27:33,780
 in the fetal position crying on the floor most days what was hard about it just too much just running

937
01:27:33,780 --> 01:27:40,900
 open sets and sovereign engineering at the same time it just killed me um yeah and uh how it is for me i

938
01:27:40,900 --> 01:27:46,420
 i'm glad it's still working i'm glad it's still working in the sense that uh people are actually

939
01:27:46,420 --> 01:27:53,860
 fired up and excited that was kind of the goal from from the get-go and turning that excitement into

940
01:27:53,860 --> 01:28:00,340
 actually shipped software violently shipped software um that's the the whole point of the thing yeah and i'm

941
01:28:00,340 --> 01:28:06,740
 i'm i was in the same boat as you like you mentioned you know um having the opinion that all the ideas

942
01:28:06,740 --> 01:28:11,940
 are super crazy that's never going to work there was my opinion i was the first user i think of tollgate

943
01:28:11,940 --> 01:28:15,140
 like they demoed it and i was the first guy that actually

944
01:28:15,140 --> 01:28:21,140
 managed to send e-cash and actually got a connection and so on and i was like this is awesome

945
01:28:21,140 --> 01:28:25,860
 but it's also very retarded no one's going to use it yeah and just that the project is still alive and

946
01:28:25,860 --> 01:28:32,100
 from that came phips yeah that that makes me insanely bullish on sovereign engineering yeah like that's

947
01:28:32,100 --> 01:28:37,060
 i think that's where it's like you want to i mean the the retard max idea you know it's like you want

948
01:28:37,060 --> 01:28:42,260
 to yeah yeah like it's it's oftentimes so like i don't know sovereign engineering is retired where

949
01:28:42,260 --> 01:28:48,020
 the good where the good ideas come from yeah where do you get your ideas no no no exactly this is actually

950
01:28:48,020 --> 01:28:52,340
 a profound point though it's like it's like how do i'm trying to think of like examples uh

951
01:28:53,300 --> 01:29:00,020
 like you know i don't know like airbnb for example was like you know there's oh wow like some of these

952
01:29:00,020 --> 01:29:04,740
 things that have become very popular or uber right like you had to be kind of in a crazy weird position

953
01:29:04,740 --> 01:29:10,660
 to think that's a good idea right yeah like the people who were there were like you know it also

954
01:29:10,660 --> 01:29:15,540
 it came up at the time where the go-to wisdom was like don't go into strangers cars yeah exactly

955
01:29:15,540 --> 01:29:24,180
 like these were crazy you know you had to like be out on the frontier being in a very different

956
01:29:24,180 --> 01:29:30,100
 situation to to to ever have that idea and so that's what i think what's interesting because

957
01:29:30,100 --> 01:29:33,620
 like tollgate was the other one where i'm i mean i'm still basically like this like no one's ever

958
01:29:33,620 --> 01:29:40,180
 going to use this but last friday they were fucking showing me how they do their ci how they build

959
01:29:40,180 --> 01:29:45,860
 the releases and i'm like oh you guys this is you can do so much better and so by the end of the

960
01:29:45,860 --> 01:29:52,100
 night drinking beers i i'd figured out how to flash one of the routers with rust code and nix and like

961
01:29:52,100 --> 01:29:55,860
 so you're a tollgate contributor i'm i'm i'm gonna write the next version of it i think

962
01:29:55,860 --> 01:29:57,780
 i don't even think it's a good idea

963
01:29:57,780 --> 01:30:07,460
 yet but also to be fair i mean um uh we mentioned copper door a couple of times already he comes from a

964
01:30:07,460 --> 01:30:14,340
 very strange community of people right like no one basically has his background yeah of coming from a

965
01:30:14,340 --> 01:30:20,660
 you know like free state hippie community in india yeah that gets basically bulldozed by the government

966
01:30:20,660 --> 01:30:25,940
 right now yeah and and just like 2 000 people growing up in a commune that's completely off the grid

967
01:30:25,940 --> 01:30:30,900
 similar to the amish right yeah the same idea yeah and it's like obviously he comes up more hippie-ish

968
01:30:30,900 --> 01:30:36,100
 and more yeah and they are technologically savvy yeah and they are they are not and like hindu

969
01:30:36,100 --> 01:30:44,820
 spirituality like exactly like all like very like in that context the project makes much more sense

970
01:30:44,820 --> 01:30:50,820
 yeah he also he has people that that are begging to get something like that so that they can use it

971
01:30:50,820 --> 01:30:54,580
 and deploy it in their communities you know one in some sense like i kind of hope to see

972
01:30:56,100 --> 01:31:00,820
 the world become a little more like this right like i do really like it's coming i really like

973
01:31:00,820 --> 01:31:06,260
 the amish and i really hate the uh i really hate living like feeling like we live in like a shopping

974
01:31:06,260 --> 01:31:09,700
 mall like in america i feel like a lot of times i like live in a fucking shop i've been trying to

975
01:31:09,700 --> 01:31:16,820
 rent an apartment the last couple weeks and like for two weeks they have not accepted any proof of income

976
01:31:16,820 --> 01:31:24,100
 i have because i don't have like a w like an actual formal employer but like i you know have a few

977
01:31:24,100 --> 01:31:30,500
 income streams that is plenty for this apartment yeah they won't acknowledge any of them uh i've been

978
01:31:30,500 --> 01:31:35,540
 trying to i have like a little bit quite enough bitcoin on an exchange that would like be like this

979
01:31:35,540 --> 01:31:40,980
 is enough proof that i think they won't take self-custody bitcoin obviously like it's just a screenshot

980
01:31:40,980 --> 01:31:44,500
 and like like but they wouldn't on the exchange they wouldn't take the screenshot

981
01:31:45,620 --> 01:31:51,060
 you know yeah it's like or not screenshot the the uh the the statements yeah the statements wouldn't

982
01:31:51,060 --> 01:31:55,540
 they they they wouldn't upload to their system yeah according to them you have no money according

983
01:31:55,540 --> 01:32:00,820
 to them i have no money yeah uh and like i can't literally like i can't figure out how to rent an

984
01:32:00,820 --> 01:32:07,540
 apartment right now it's like i just want to kill myself like i want to be amish i i i know a commune

985
01:32:07,540 --> 01:32:11,940
 in india yeah i know so like i don't know and some i mean then that's the fips thing right like

986
01:32:11,940 --> 01:32:15,860
 fips is probably never going to be the you're not going to have one mesh on the entire internet like

987
01:32:15,860 --> 01:32:20,580
 it's probably not going to be fast enough or this or that who knows maybe maybe maybe not but like

988
01:32:20,580 --> 01:32:26,180
 i don't know like this more you know going a little little bit away from the having one one market

989
01:32:26,180 --> 01:32:32,740
 system that you plug into yeah absolutely uh towards something that's more like amish or yeah i mean

990
01:32:32,740 --> 01:32:36,980
 that's what's already happening and if you know lots of bitcoiners then you know like people just buying

991
01:32:36,980 --> 01:32:40,340
 properties and houses with bitcoin outright circumventing the whole system yeah you know

992
01:32:40,340 --> 01:32:44,980
 yeah it's it's we're at the point now where you can do that and it's basically you're living in a

993
01:32:44,980 --> 01:32:52,660
 parallel economic system and i think the fips you live just in a parallel internet system yeah in a

994
01:32:52,660 --> 01:32:57,860
 parallel communication layer in some sense and i i definitely see where that's going and it's it's

995
01:32:57,860 --> 01:33:04,100
 funny that you brought up the the amish and and kind of like uh um you know our friend bitstein he

996
01:33:04,100 --> 01:33:09,460
 has this mission as well yeah he wants to orange pull the amish yeah and his head is in a similar

997
01:33:09,460 --> 01:33:15,380
 place yeah and i talked with shadrach as well and yeah he actually moved he's doing it yeah he moved

998
01:33:15,380 --> 01:33:19,700
 to amish country to orange build the amish no it was interesting i was talking to him and it was like

999
01:33:19,700 --> 01:33:24,500
 i was like so i always had the sense that they were against technology and no they're not against

1000
01:33:24,500 --> 01:33:31,220
 technology per se they're against dependence exactly yeah dependence is the issue yeah so like um yeah

1001
01:33:31,220 --> 01:33:36,260
 rack pull is already issue they won't buy a car because it's unclear yeah you can't really maintain

1002
01:33:36,260 --> 01:33:41,380
 it yeah in the way that you can maintain a buggy yeah you can always repair that buggy right yeah

1003
01:33:41,380 --> 01:33:46,660
 they won't build the buggy with purely imported leather because you may not be able to import the

1004
01:33:46,660 --> 01:33:52,580
 leather in the future right and so now they have like leather works and i don't know like i did a road

1005
01:33:52,580 --> 01:33:57,620
 trip a couple years ago across the u.s and like the most charming place in the entire country was the amish

1006
01:33:57,620 --> 01:34:03,540
 country for me because like every little you know turn of the road every little gate you could tell

1007
01:34:03,540 --> 01:34:09,060
 that there was like care in all these things in a way that there is in like japan or parts of europe

1008
01:34:09,060 --> 01:34:13,940
 right there's just there's you can tell there's care put into the little things yeah and there's a pride

1009
01:34:13,940 --> 01:34:18,660
 across the community you know there's a communal pride not just like on an individual or family basis

1010
01:34:18,660 --> 01:34:23,540
 like most of it is in the u.s and that's the close you know the best the closest thing we have to

1011
01:34:23,540 --> 01:34:28,660
 that is amish country or maybe uh some of the like mormon places in utah or something although yeah

1012
01:34:28,660 --> 01:34:35,220
 amish country is just so pretty uh it's interesting that you start from basically religious ideas

1013
01:34:35,220 --> 01:34:41,060
 fundamentally and that's the end result right i think you have to by the by the fruits uh you

1014
01:34:41,060 --> 01:34:46,180
 shall know that where you get your ideas from yeah but that's uh looping it back hey jesus where do you

1015
01:34:46,180 --> 01:34:54,340
 get your ideas from exactly i'll record with jesus and i'll ask him yeah yeah yeah but to to the point

1016
01:34:54,340 --> 01:35:01,540
 of being um resistant to rug pulls i think we are still at a point technologically where people have

1017
01:35:01,540 --> 01:35:08,100
 just no idea how stuff works because it's just it's it's very easy to understand okay here is your horse and

1018
01:35:08,100 --> 01:35:13,300
 your buggy yeah and you can repair the wagon at any point in time with just very simple tools because

1019
01:35:13,300 --> 01:35:18,020
 everything is so simple but to me that's what bitcoin is as well bitcoin is very complex but but

1020
01:35:18,020 --> 01:35:23,140
 breaking it down it's just like once you understand cryptography and once you understand proof of work

1021
01:35:23,140 --> 01:35:27,940
 yeah there is no dependency on anything yeah and i think that's what phipps and nostr and all those

1022
01:35:27,940 --> 01:35:32,020
 that's what nostr is yeah that's what nostr is that's what i like about nostr it's it's just it's

1023
01:35:32,020 --> 01:35:37,860
 all so very simple you can run it yourself forever yeah and that's just the yeah again i mean

1024
01:35:37,860 --> 01:35:41,540
 that's why people want to orange pull the amish because it's the same idea right like you

1025
01:35:41,540 --> 01:35:46,740
 of course you're not going to produce a mining machine yourself like you're not going to produce

1026
01:35:46,740 --> 01:35:52,900
 a fab yourself and so on but once you have a computing device you can keep the things intact

1027
01:35:52,900 --> 01:35:59,620
 like almost indefinitely yeah the that's what's so appealing to me it's is that's why i moved

1028
01:35:59,620 --> 01:36:05,060
 fully to nostr like i don't know how long ago actually like three and a half years ago or something

1029
01:36:05,060 --> 01:36:10,500
 like i i nuked all my other stuff quite instantly and just went yeah all in nostr when are you going to

1030
01:36:10,500 --> 01:36:18,580
 nuke dns i'm getting there i'm getting there i'm getting there i am running phipps already but i

1031
01:36:18,580 --> 01:36:23,860
 haven't moved everything yeah yeah and we'll allow it i think that's the i think that's the point right

1032
01:36:23,860 --> 01:36:29,860
 like you you want to you don't want to build on quicksand you want to build on a rock yeah and in like

1033
01:36:29,860 --> 01:36:34,340
 bitcoin and nostr allows you to do that yeah of course you have to run your own infrastructure you

1034
01:36:34,340 --> 01:36:38,500
 have to be like fun now that like that's the one of the biggest things with agents is like

1035
01:36:38,500 --> 01:36:45,300
 i've always hated devops yeah i just really don't like it uh and over the last year now it's like

1036
01:36:45,300 --> 01:36:50,580
 one of my favorite things i love it when i get to set up a new service and like you know connect things

1037
01:36:50,580 --> 01:36:57,860
 and you know it's fun like you get a real sense i don't know it's like fun now you think we can do

1038
01:36:57,860 --> 01:37:05,700
 all that with uh purely local models local compute soon i think so um because that's still a heavy

1039
01:37:05,700 --> 01:37:11,540
 dependence we have right now as agentic i mean this is one of the reasons why i don't uh well i'm an i'm

1040
01:37:11,540 --> 01:37:17,140
 an agentic engineer i don't know about you okay i heard i heard i heard i'm not there yet i heard that

1041
01:37:17,140 --> 01:37:23,140
 you set up a separate open claw for every one of your projects and inside i laughed a little inside

1042
01:37:23,140 --> 01:37:28,900
 i did laugh i mean i'm not setting it up my main agent is setting it up okay yeah if that if that's

1043
01:37:28,900 --> 01:37:40,260
 your point of criticism yeah yeah you're like a agentic manager or you're uh fair enough yeah

1044
01:37:40,260 --> 01:37:49,300
 an agentic middle manager but that's that's kind of what what what what i'm eyeing now the the next

1045
01:37:49,300 --> 01:37:54,980
 like i don't know 18 months or so yeah i want to go fully like my own hardware no i i do too and

1046
01:37:54,980 --> 01:38:00,820
 like this is one of the reasons so i have some ideas for how i would like to use open claws and

1047
01:38:00,820 --> 01:38:05,940
 you know i have a lot of information that i've collected over the course of my life that i'm not

1048
01:38:05,940 --> 01:38:13,940
 going to put into anthropic yeah you know and i tried a little bit with uh maple and stuff like that

1049
01:38:13,940 --> 01:38:19,460
 but it's just not quite there you know yeah it's getting there right like these some of these newer

1050
01:38:19,460 --> 01:38:23,940
 ones are getting close for probably what i want but i feel like over the you know in the next six

1051
01:38:23,940 --> 01:38:31,060
 months we'll be able to have stuff that you know a you know a a reasonably successful person from the

1052
01:38:31,060 --> 01:38:36,340
 western world can do everything that the people have been doing on their open claws on their own

1053
01:38:36,340 --> 01:38:40,980
 but on their own hardware you know using anthropic and chate but on their own hardware and i think

1054
01:38:40,980 --> 01:38:45,460
 that's coming in like next 12 months for sure maybe the next six months and it's going to take a big

1055
01:38:45,460 --> 01:38:49,460
 upfront like a decent upfront payment you know you're gonna have to basically like a new car you know

1056
01:38:49,460 --> 01:38:53,540
 yeah yeah yeah but i think that's it's gonna be worth it yeah sure yeah i mean like if you do the

1057
01:38:53,940 --> 01:38:59,460
 the application adjusted uh like what a mac why do people buy a car because it brings you freedom

1058
01:38:59,460 --> 01:39:05,700
 the initial macintoshes were like 10k yeah 7k in today's money and that's if you price it in gold i

1059
01:39:05,700 --> 01:39:11,220
 think it was like yeah it's like five or six k fiat but like in gold it's like more like 10 20k yeah

1060
01:39:11,220 --> 01:39:18,820
 like it was really expensive uh at the time for like the apple 2. so i don't know they look at it

1061
01:39:18,820 --> 01:39:24,420
 like that and yeah the hard thing is me is i just really don't know what to buy yeah like i need to

1062
01:39:25,620 --> 01:39:32,260
 i need to find somebody to make these decisions for me yeah kind of same i mean you know i i asked my

1063
01:39:32,260 --> 01:39:37,060
 agents what to buy yeah but i mean we're gonna try to grant some of these people who are who do these

1064
01:39:37,060 --> 01:39:42,500
 explorations and so that they can help make recommendations across different budgets and

1065
01:39:42,500 --> 01:39:47,700
 and goals yeah stuff like that we said half an hour ago or longer than that that we're gonna stop

1066
01:39:47,700 --> 01:39:52,340
 recording but we didn't are you coming to the barbecue um i'll try to i'll have to figure that out now

1067
01:39:52,340 --> 01:40:00,260
 but let's leave it at that for real we always end on a hug a pleasure to have you as always and we'll

1068
01:40:00,260 --> 01:40:08,340
 see where shadow is and everything else soon i'll ship violently that's a promise awesome wait a second