1
00:00:00,000 --> 00:00:08,520
 there we go we're basically life not really life now I get it so Jonathan I

2
00:00:08,520 --> 00:00:15,240
 want to start with saying thanks two times three times actually so first of

3
00:00:15,240 --> 00:00:21,380
 all thanks for being insane enough to come here on you know like with very

4
00:00:21,380 --> 00:00:25,280
 little knowledge about what it is yeah it was a it was a challenging trip to

5
00:00:25,280 --> 00:00:29,640
 get here but it's been a very good trip yeah it took a couple landing attempts

6
00:00:29,640 --> 00:00:33,880
 and a couple days delay but that's that's how Madeira is it's a crazy

7
00:00:33,880 --> 00:00:39,000
 airport the second thanks goes to agreeing to do this go on the record to

8
00:00:39,000 --> 00:00:42,420
 talk FIPS and I'm gonna grill you about FIPS a little bit and about your

9
00:00:42,420 --> 00:00:47,880
 Sovereign Engineering experience and third of all thanks for building FIPS in the

10
00:00:47,880 --> 00:00:53,280
 first place it was very much needed I think yeah the response has been very

11
00:00:53,280 --> 00:00:57,780
 gratifying you know it's I mean I've been in the open source community for a very

12
00:00:57,780 --> 00:01:02,580
 long time but I had kind of forgotten what it was like you know when you produce

13
00:01:02,580 --> 00:01:06,180
 something and other people pick it up and start doing things on their own with

14
00:01:06,180 --> 00:01:10,380
 it that like you didn't tell them to go do yeah you know and you realize like wow you

15
00:01:10,380 --> 00:01:13,140
 know other people actually think this is worth investing in their own time and

16
00:01:13,140 --> 00:01:20,020
 energy into and so you know the response here at SEC has been incredible the demo

17
00:01:20,020 --> 00:01:24,100
 days so far you know the two weeks that we've had you know and the one that we have

18
00:01:24,100 --> 00:01:29,620
 tomorrow have been mostly FIPS oriented and you know I wasn't expecting that when

19
00:01:29,620 --> 00:01:36,520
 I did it but the the pent-up demand and the you know the need that FIPS fills with

20
00:01:36,520 --> 00:01:40,180
 this group has been pretty high what were you expecting when you built it why did you

21
00:01:40,180 --> 00:01:46,180
 build it in the first place well it was it was a challenge to solve which kind of

22
00:01:46,180 --> 00:01:54,580
 sounds funny but you know when I was at the Nostr event in Costa Rica Arjen you

23
00:01:54,580 --> 00:01:57,540
 know mentioned that hey here's a problem we we need to solve and here's kind of

24
00:01:57,540 --> 00:02:01,300
 the outlines of it and you know we want to have a way to have networking that

25
00:02:01,300 --> 00:02:07,300
 doesn't require permission from central authorities that doesn't have you know

26
00:02:07,300 --> 00:02:11,700
 the problems where someone can just like yank a domain name or whatever and I

27
00:02:11,700 --> 00:02:16,260
 thought well you know I think I know how to solve this but I'm not sure and so I

28
00:02:16,260 --> 00:02:21,140
 went away for like a month after after Costa Rica and I think Arjen thought that

29
00:02:21,140 --> 00:02:27,540
 like I had forgotten all about it and and I thought well I think I know how to do

30
00:02:27,540 --> 00:02:32,980
 this and it was just you know like an intense amount of thinking and design and

31
00:02:32,980 --> 00:02:37,300
 then programming you know to come out with it and you know he kept telling me

32
00:02:37,300 --> 00:02:42,820
 they're gonna love this and like yeah well you know that's good to hear I just

33
00:02:42,820 --> 00:02:47,860
 didn't know sort of the breadth of interest yeah and that you know you know

34
00:02:47,860 --> 00:02:51,860
 Cobrador told me that you know I I didn't think this would exist for 10 more

35
00:02:51,860 --> 00:02:55,860
 years you know and now I have it in my hands

36
00:02:55,860 --> 00:03:04,580
 yeah like you say there was a lot of pent-up demand and it kind of it's it's we went full circle

37
00:03:04,580 --> 00:03:08,180
 because and and it's what you you mentioned also a couple days ago

38
00:03:08,180 --> 00:03:12,980
 basically Cobrador by implementing Tollgate and and he has a weird way of

39
00:03:12,980 --> 00:03:15,780
 thinking about these things because he has a very weird background right it's

40
00:03:15,780 --> 00:03:19,380
 like how can you how can you make this work in like an Indian hippie commune of

41
00:03:19,380 --> 00:03:23,620
 2000 people and sell internet by the packet you know like that's yeah that's his

42
00:03:23,620 --> 00:03:24,580
 big mission basically

43
00:03:24,580 --> 00:03:33,540
 he may see it and and he he kind of got Arjen in the mix and then um Arjen and you had this

44
00:03:33,540 --> 00:03:38,980
 interaction and then you locked yourself in the in the cellar for a month like that's how all good

45
00:03:38,980 --> 00:03:44,020
 software projects are born right like that's how git came to be and so on and and then you came back

46
00:03:44,020 --> 00:03:49,620
 and and again like you actually joined for Sovereign Engineering which is amazing to to bring it to the

47
00:03:49,620 --> 00:03:56,420
 group and um to me what's interesting is that

48
00:03:56,420 --> 00:04:00,020
 FIPS kind of in some way turns networking

49
00:04:00,020 --> 00:04:04,020
 on its head and and now we are now we are back to the early days of networking

50
00:04:04,020 --> 00:04:06,980
 again sure yeah we're trying to figure out

51
00:04:06,980 --> 00:04:09,860
 you know how does the scale how you know what are the problems with it you

52
00:04:09,860 --> 00:04:13,700
 know how do you break it how how do you respond to bad actors trying to

53
00:04:13,700 --> 00:04:18,100
 sabotage it you know there's all these things that are just sort of swirling in the air

54
00:04:18,100 --> 00:04:22,500
 that the regular internet you know was dealing with 30 or 40 years ago

55
00:04:22,500 --> 00:04:27,620
 yeah and and that solution was mostly authority based right like oh yeah yeah

56
00:04:27,620 --> 00:04:32,900
 and and i think that's the challenge at hand how do we how do we solve the same problems it's still

57
00:04:32,900 --> 00:04:38,500
 the same problems but without reintroducing a central authority that can yeah i mean so from a purely

58
00:04:38,500 --> 00:04:44,900
 engineering point of view you know FIPS nodes you know are you know not to be too cliche but they're

59
00:04:44,900 --> 00:04:50,180
 they're completely peers with each other so there's no way to enforce anything except what you enforce

60
00:04:50,180 --> 00:04:56,980
 on your node so you have to set the software up in a way that if they all cooperate it works great

61
00:04:56,980 --> 00:05:00,660
 but if someone isn't cooperating by doing like the same protocol

62
00:05:00,660 --> 00:05:05,700
 the impact of it is limited to their you know the blast radius is limited to their small region

63
00:05:06,660 --> 00:05:12,500
 and that's kind of a difficult tension there are a lot of things in FIPS that would be a lot easier

64
00:05:12,500 --> 00:05:18,580
 from a technical point of view if there was a top-down yeah sure you know it's always easier you know

65
00:05:18,580 --> 00:05:24,820
 but that would just be you know yeah following the same path we we followed 30 years ago yeah yeah

66
00:05:24,820 --> 00:05:32,660
 what is the best demo you saw building on FIPS in the last two weeks what surprised you most well

67
00:05:35,380 --> 00:05:44,420
 so this this one's kind of funny um satsonsports designed a transport for FIPS and in FIPS the

68
00:05:44,420 --> 00:05:49,860
 transport is a way for two nodes to communicate it could be over the internet it could be directly

69
00:05:49,860 --> 00:05:57,620
 over ethernet or bluetooth or you know tor or over a serial connection you know the the node abstraction

70
00:05:58,660 --> 00:06:06,980
 is um independent independent of all those so so he actually built a transport that could talk over

71
00:06:06,980 --> 00:06:14,180
 Nostr nice so it was literally sending packets up to relays down from the relay to another node

72
00:06:14,180 --> 00:06:21,780
 as if it was an ip network and like you know his ping times were measured in seconds yeah but it worked

73
00:06:21,780 --> 00:06:27,380
 yeah and and that's you know wildly impractical and kind of crazy but but it just shows that FIPS

74
00:06:27,380 --> 00:06:32,100
 doesn't care about the underlying it it does and actually the thing that it proved out was that

75
00:06:32,100 --> 00:06:38,900
 with these long delays it just works still yeah you know like the everything works on the happy path

76
00:06:38,900 --> 00:06:47,380
 and you've got two milliseconds latency and all that um that um that you know if that's all you touch

77
00:06:47,380 --> 00:06:53,780
 then when you try to do something more strenuous it breaks yeah and you know his little demo was

78
00:06:53,780 --> 00:06:59,860
 something that uh was like a existence proof that the protocol was a little bit more robust yeah that's

79
00:06:59,860 --> 00:07:03,220
 awesome that's awesome what's still on the roadmap for you let's walk up there a little bit and then

80
00:07:03,220 --> 00:07:08,180
 then there's there's a nice bench where we can sure sure so what's still on the roadmap for you more

81
00:07:08,180 --> 00:07:16,180
 battle testing yeah well so in this round with FIPS you know it's like what i call a dancing bear

82
00:07:16,180 --> 00:07:22,740
 it's not that the dare the bear dances very well it said he can dance at all

83
00:07:22,740 --> 00:07:30,580
 and you know FIPS is working great you know we have 150 nodes connected right now on the on the test net

84
00:07:31,380 --> 00:07:36,100
 and which is quite amazing in and of itself like that's way way more than i would have guessed

85
00:07:36,100 --> 00:07:44,820
 yeah i was surprised to see the growth um but you know we're this was a way of learning what worked

86
00:07:44,820 --> 00:07:52,020
 and what didn't i had put uh quite a few features in this first release that i was i was not sure that

87
00:07:52,020 --> 00:07:58,820
 they were necessary you know things to do with like measuring protocol statistics and whatnot and so

88
00:07:59,780 --> 00:08:06,180
 now that you know i have a better sense of what you know where the load bearing parts of FIPS are

89
00:08:06,180 --> 00:08:14,500
 i'm going through not a rewrite but i'm going through a a next stage where you know the the

90
00:08:14,500 --> 00:08:19,860
 protocol negotiation between nodes doesn't just assume like it's you know fixed format and if you

91
00:08:19,860 --> 00:08:24,260
 want to change it it breaks everybody and all of that so i'm going through it like a redesign where

92
00:08:25,380 --> 00:08:31,540
 i can get rid of the pieces that don't uh that don't matter you know or and hardening some things

93
00:08:31,540 --> 00:08:39,380
 that needed hardening or like anticipating okay here's we know how things operate at 150 nodes

94
00:08:39,380 --> 00:08:45,860
 what do i need to put it into it to operate at 1500 or 15 000 nodes yeah yeah so that's

95
00:08:47,300 --> 00:08:55,620
 we're going to be coming out with a uh release maybe in a couple of weeks that's the first version of

96
00:08:55,620 --> 00:09:01,780
 that and you don't really you you won't really see any changes in features but it's kind of like the

97
00:09:01,780 --> 00:09:06,580
 the foundation is now solid yeah i see instead of being like

98
00:09:07,860 --> 00:09:14,180
 there and working but like rickety yeah in in terms of contributions what would be the most helpful

99
00:09:14,180 --> 00:09:22,340
 you think actually testing so just running it and and connecting to stuff and just playing around

100
00:09:22,340 --> 00:09:27,380
 with it basically yeah i mean you know i'm a network engineer if i can get pinged to work you know my

101
00:09:27,380 --> 00:09:35,460
 job is done but the you know the guys at SEC here i mean we've got a port to windows 32 we've got a

102
00:09:35,460 --> 00:09:43,380
 port to mac we've got a port to uh freebsd somebody has got uh it working in web assembly inside a

103
00:09:43,380 --> 00:09:49,380
 browser yeah i saw that and you know it's just like these are all things that prove out you know

104
00:09:49,380 --> 00:09:53,060
 it can operate in different environments and there aren't like fundamental new features

105
00:09:53,060 --> 00:10:04,580
 you know what i i guess what i really want is try to break it so that we can fix it because if we

106
00:10:04,580 --> 00:10:09,460
 don't break it somebody else is going to and they may not have the same intentions that we have

107
00:10:09,460 --> 00:10:10,580
 yeah i see you know

108
00:10:10,580 --> 00:10:13,220
 yeah

109
00:10:13,220 --> 00:10:20,100
 and how was no it won't sound so out of breath

110
00:10:20,100 --> 00:10:28,020
 this is madera after all everything's uphill everything's uphill everything's you know it's

111
00:10:28,020 --> 00:10:33,940
 like the Bitcoin roller coaster guy he goes like this exactly but it we have nice views like that's the

112
00:10:33,940 --> 00:10:39,620
 that's the nice thing about climbing a hill like you get nice views oh it's uh it's fantastic here

113
00:10:39,620 --> 00:10:46,580
 yeah how how how were the last like almost three weeks for you um uh what what did how what did you

114
00:10:46,580 --> 00:10:51,700
 expect coming here and how how was it different from what you expected well i think i think the the

115
00:10:51,700 --> 00:10:58,260
 costa rica event sort of like paved the way because it was the first conference there was like 40

116
00:10:58,260 --> 00:11:06,420
 Nostr developers in one space and i had not encountered that kind of vibe in like three

117
00:11:06,420 --> 00:11:10,980
 decades you know when i was originally involved with the cypherpunk group in silicon valley in the

118
00:11:10,980 --> 00:11:23,140
 early 90s it was that same freedom tech um self-sovereign uh you know don't try to fight the system bypass it

119
00:11:23,140 --> 00:11:30,580
 you know yeah build a new one and leave the old one behind and let the old one do what it it's going

120
00:11:30,580 --> 00:11:37,380
 to do for whatever you know better for worse but you know opt out and build your own alternative um system

121
00:11:37,380 --> 00:11:43,220
 and a lot of stuff came out of that group you know in the early days and then in the Bitcoin days

122
00:11:43,780 --> 00:11:50,980
 uh in the early you know 2010 when i was involved uh it's the same idea you know we were focusing

123
00:11:50,980 --> 00:11:58,340
 almost entirely on the payment rails you know economic part of the problem uh but you know all

124
00:11:58,340 --> 00:12:07,300
 the other privacy issues still existed and they were even worse um the Nostr community and then now here

125
00:12:07,300 --> 00:12:16,900
 the Sovereign Engineering subset of that um has been very refreshing to see that young people today

126
00:12:16,900 --> 00:12:25,700
 are picking up the mantle of you know this kind of fight of not changing the system but building a new

127
00:12:25,700 --> 00:12:30,980
 one and opting out of the old one yeah and doing it in ways that really can't be stopped

128
00:12:32,020 --> 00:12:39,220
 the tools today to accomplish that are much better yeah i mean especially now with agentic tooling but

129
00:12:39,220 --> 00:12:45,220
 you know i don't think that Nostr itself would have been possible 30 years ago yeah

130
00:12:45,220 --> 00:12:52,980
 or even in the 2010s of Bitcoin uh era um it's a uh it's been

131
00:12:55,220 --> 00:12:59,700
 rejuvenating for an old guy like me yeah to interact with you know all these young people that

132
00:12:59,700 --> 00:13:04,420
 are earnest and trying to do the right thing and maybe they don't have all the history or maybe they

133
00:13:04,420 --> 00:13:10,660
 don't have all of the experience or or knowledge but you know that's what they'll get by doing all

134
00:13:10,660 --> 00:13:16,260
 this yeah yeah yeah yeah no i think you're right i mean what what we have on the not it's it's to me

135
00:13:16,260 --> 00:13:22,260
 it's not very surprising that like Nostr is full of Bitcoin people because you need to be able to

136
00:13:22,260 --> 00:13:26,420
 handle keys for example sure and Bitcoin prepares you for that most people are not

137
00:13:26,420 --> 00:13:32,980
 fine with handling key material like yeah well it's a hard thing to do and some people will never be

138
00:13:32,980 --> 00:13:39,780
 fine with that yeah and you know custodial solutions that provide the solution to that yeah will always

139
00:13:39,780 --> 00:13:45,700
 exist yeah and you know people will decide their their risk tolerance for how they want to handle their

140
00:13:45,700 --> 00:13:50,900
 money yeah absolutely you know the consequences of of losing control of your input are a little bit less

141
00:13:50,900 --> 00:13:57,220
 than losing control of your utxos very true but still it exists and in fact i think in the last cohort

142
00:13:57,220 --> 00:14:04,820
 the the issue of you know identity and um you know key revocation and key rotation was the subject of the

143
00:14:04,820 --> 00:14:10,980
 last one yeah um and i have some thoughts on that but i'm not well educated in like what everyone has

144
00:14:10,980 --> 00:14:16,740
 already tried and decided doesn't work so i i've got to go look at the you know the captain's reports from the

145
00:14:16,740 --> 00:14:20,740
 last one and and read all about that that also came up in costa rica there was a big group that was

146
00:14:20,740 --> 00:14:28,020
 interested in the in the uh dealing with in pubs and secret keys and all of that in a more robust way

147
00:14:28,020 --> 00:14:32,260
 i'm quite confident that this will be solved it's just going to take time and in terms of like you

148
00:14:32,260 --> 00:14:36,260
 know there's no real top town structure so you can't force it on all the clients and all those kind

149
00:14:36,260 --> 00:14:42,740
 of things it just have has to emerge organically and that's the way everything is solved nobody comes

150
00:14:42,740 --> 00:14:47,780
 out with solutions that work the first time you know it's a darwinian process you know you'll get

151
00:14:47,780 --> 00:14:53,540
 a hundred different variations on a theme and maybe five of them will work well and the rest will go in

152
00:14:53,540 --> 00:15:02,100
 the dustbin and um you know you'll you'll have a set of solutions and then people will pick like one

153
00:15:02,100 --> 00:15:07,860
 that they like and the network effect will take over from there so i don't think anybody even you know

154
00:15:07,860 --> 00:15:14,740
 you think of the modern internet and you know how rigid it is you know it didn't become this whole

155
00:15:14,740 --> 00:15:20,980
 you know fully formed you know in in the 80s and 90s yeah uh you know built in incremental steps to

156
00:15:20,980 --> 00:15:27,540
 deal with problems that emerged as it got bigger and bigger um and and some solutions you know the

157
00:15:27,540 --> 00:15:33,780
 classic one is the is the seven layer osi stack yeah and it's like that was a hallucination basically

158
00:15:33,780 --> 00:15:38,100
 well yeah i mean you know the standards body imposed you know this is how it's going to be

159
00:15:38,100 --> 00:15:43,700
 and everyone just kind of thumb you know thumb their nose at them and said no yeah i learned that in

160
00:15:43,700 --> 00:15:49,700
 school by the way like i i had to i had to learn the whole seven layer stack by heart and and the

161
00:15:49,700 --> 00:15:54,580
 first three layers made sense and yeah and you know it doesn't matter if it matches reality all protocols

162
00:15:54,580 --> 00:15:59,140
 you know have physical layers and mac layers and network transport layers but then above that it

163
00:15:59,140 --> 00:16:03,140
 gets really murky and and the rest of that was just a fantasy of some people on the committee yeah

164
00:16:03,140 --> 00:16:07,860
 yeah yeah and you know that'll happen with Nostr as well yeah no i agree you mentioned that the

165
00:16:07,860 --> 00:16:15,220
 tooling is better today um what can you say about kind of what other stuff is different when you compare

166
00:16:15,220 --> 00:16:20,580
 it to the early cypherpunk movement or the early Bitcoin movement in terms of i don't know like culture

167
00:16:20,580 --> 00:16:26,260
 or i don't know like state of the world or you you you mentioned you see lots of similarities but i i

168
00:16:26,260 --> 00:16:32,100
 would be interested like how is it different well you probably won't like my answer that's fine so so

169
00:16:32,100 --> 00:16:40,100
 with enthusiasm comes sort of short attention spans yeah and uh you know this just may be the old guy

170
00:16:40,100 --> 00:16:50,180
 and me talking again but in earlier um epics of of this movement there was a lot more academic structure

171
00:16:50,180 --> 00:16:55,220
 behind it yeah you know so all the early cryptography they all came from academia right yeah the early

172
00:16:55,220 --> 00:17:02,980
 70s or the late 70s with rsa and diffe hellman and you know the um hardening of uh symmetric

173
00:17:03,620 --> 00:17:09,860
 cryptography you know from like triple des to you know newer and newer versions that work better you

174
00:17:09,860 --> 00:17:18,420
 know that was strongly academically focused and um people in groups like the cypherpunks you know had

175
00:17:18,420 --> 00:17:25,780
 computer science backgrounds you know they were not necessarily academics from a you know their career

176
00:17:25,780 --> 00:17:32,500
 was you know working in academia at a university publishing papers but more like they were coming at it from

177
00:17:33,860 --> 00:17:40,020
 having been educated in you know what today we would just call stem yeah rigorous academic training but

178
00:17:40,020 --> 00:17:47,700
 but much more rigor involved in pursuing ideas and writing them up and um you know the critiques and the

179
00:17:47,700 --> 00:17:52,900
 you know fortunately you know the expression that's still well known today is you know rough consensus and

180
00:17:52,900 --> 00:17:58,420
 running code yeah you know so code trumped everything if you had something that you know worked and you had

181
00:17:58,420 --> 00:18:05,300
 code that did it then a lot of the fluff behind it sort of didn't matter you know so you could you

182
00:18:05,300 --> 00:18:08,980
 could talk about a protocol all day long but if you had something working

183
00:18:08,980 --> 00:18:12,900
 you know that that lent it credibility because people could you know pick them pick the code apart

184
00:18:12,900 --> 00:18:20,740
 and find out what was uh what was real and what wasn't in Nostr and SEC what i have seen

185
00:18:20,740 --> 00:18:27,220
 is very earnest well-meaning people trying to do the right thing but not necessarily taking the time to

186
00:18:27,220 --> 00:18:34,980
 do you know what i still think is the necessary step of engineering design and review and um

187
00:18:34,980 --> 00:18:42,420
 making sure the design is right yeah before you even attempt to implement it and uh you know

188
00:18:43,380 --> 00:18:50,660
 my my snarky joke of shower thought to zap store in six hours yeah you know is is is meant to be

189
00:18:50,660 --> 00:18:58,340
 triggering to people but um i no it's very true i mean i i think it's a good thing and a bad thing

190
00:18:58,340 --> 00:19:03,540
 you know it's like oh yeah it's like for some things you might strike gold with a shower thought

191
00:19:03,540 --> 00:19:09,060
 oh yeah sure well i mean there's nothing stopping you from doing that unless you know that when i say

192
00:19:09,060 --> 00:19:14,580
 the tools i mean the the ability to disseminate an idea today is vastly better than it was back then

193
00:19:14,580 --> 00:19:19,860
 you had mailing lists you know we didn't have you know discord everything was slower we didn't have

194
00:19:19,860 --> 00:19:25,700
 nostalgia um everything was slower and so you had time to be more deliberate about things yeah and now

195
00:19:25,700 --> 00:19:31,300
 which is also a feature in some sense right like you because if things are slow and require friction

196
00:19:31,300 --> 00:19:37,860
 you really need to want it right like like the idea it's a filter in in some sense so so so only

197
00:19:37,860 --> 00:19:43,780
 the most stubborn people and the best ideas come through in some sense sure uh so you don't drown

198
00:19:43,780 --> 00:19:48,500
 in in noise as much i think that's that's part of the issue today that we're drowning in noise in

199
00:19:48,500 --> 00:19:55,940
 general sure i so i agree with that and that's not just a an SEC or a Nostr thing you know so i say

200
00:19:55,940 --> 00:20:02,100
 those things you know in a kind of a snarky way um you know because i i earnestly do think that

201
00:20:03,460 --> 00:20:08,660
 Nostr could benefit from a little bit more slower pace of deliberate development and design you know

202
00:20:08,660 --> 00:20:17,860
 ahead of time but that doesn't mean that i mean you can just do things as the yeah you know we have

203
00:20:17,860 --> 00:20:23,060
 thousands of people working on stuff that they think is important and you know they're putting it out

204
00:20:23,060 --> 00:20:29,300
 there for other people and they're doing it for free it's a gift of their time and energy you know to the

205
00:20:29,300 --> 00:20:34,740
 world uh and you can't say there's anything wrong with that yeah you know and if you look at the

206
00:20:34,740 --> 00:20:41,860
 volume today versus 30 or 40 years ago it's vastly larger yeah you know so the output of new things

207
00:20:41,860 --> 00:20:48,100
 today is tremendous um you know it also means that the tail extends very far yeah so you get some

208
00:20:48,100 --> 00:20:55,300
 really bad stuff and you get some really good stuff yeah um yeah so i don't want that to stop yeah i just

209
00:20:55,300 --> 00:21:03,460
 sometimes get the feeling especially when i talk to you know normies so to speak who look at Nostr for

210
00:21:03,460 --> 00:21:09,700
 the first time and and the common thing that i hear is that it looks half baked it looks half finished

211
00:21:09,700 --> 00:21:15,140
 it looks like things are kind of slow and mostly work but i'm just going to go back to twitter because

212
00:21:15,140 --> 00:21:20,900
 you know it shows up instantly on my screen and then you have to like explain why you know yeah

213
00:21:20,900 --> 00:21:25,140
 maybe the stuff that's slow is actually all the things that you really care about

214
00:21:25,140 --> 00:21:33,380
 just working the way that they do so i don't know i i i'm gonna upset people by the comments i've made

215
00:21:33,380 --> 00:21:41,860
 no no no don't worry about that the uh and so i'm extremely pleased to see you know what has been in

216
00:21:41,860 --> 00:21:46,420
 front of everyone else you know for the last few years and i've only recently sort of rejoined this

217
00:21:46,420 --> 00:21:53,060
 community after a while to me it's amazing that it works at all Nostr and that it exists at all in

218
00:21:53,060 --> 00:21:58,900
 that sense like because not too long ago we didn't have anything we didn't have a you know like a

219
00:21:58,900 --> 00:22:05,780
 public square form twitter was the best but yeah owned by uh owned by a private company yeah basically

220
00:22:05,780 --> 00:22:10,740
 with all the issues to come with it yeah and all the good issues and the bad issues you know you

221
00:22:10,740 --> 00:22:16,260
 could you could use twitter for free yeah because somebody else is paying the bills yeah um and now

222
00:22:16,260 --> 00:22:19,780
 you can use Nostr for free because somebody else is paying the bills but the bills are really small

223
00:22:19,780 --> 00:22:24,180
 yeah it's like running a Nostr relay is not that expensive yeah um

224
00:22:24,180 --> 00:22:31,460
 i think to your comment about the long tail we went through that with the printing press as well

225
00:22:31,460 --> 00:22:35,540
 and books and music and so on right like it's a similar issue where it's just like anyone can

226
00:22:35,540 --> 00:22:40,580
 produce anything and then you just have to make sense of it after the fact now we just have it with

227
00:22:40,580 --> 00:22:44,580
 software and those kind of things as well it is funny because you bring up the music analogy

228
00:22:44,580 --> 00:22:49,780
 and you think back well you know all the great classical music that we uh that we remember

229
00:22:49,780 --> 00:22:56,100
 you know we only remember the really good ones that survived 200 years yeah you know there was just as

230
00:22:56,100 --> 00:23:02,100
 much um you know what we would call vibe slop today yeah yeah there was just a guy at the corner with

231
00:23:02,100 --> 00:23:07,140
 like a fiddle yeah just doing whatever and and you just don't know none of that went down on history

232
00:23:07,140 --> 00:23:11,700
 it's kind of like you look at these buildings that are you know still standing after 2000 years and

233
00:23:11,700 --> 00:23:15,380
 you're like wow what kind of architecture is that but you don't see all the other buildings

234
00:23:15,380 --> 00:23:20,100
 that collapsed yeah of course you know back then so you know it's kind of a survivor bias yeah

235
00:23:20,900 --> 00:23:26,340
 and it's the same in Nostr you know you you see the sausage in the making yeah but do you remember

236
00:23:26,340 --> 00:23:30,740
 the stuff that was you know at the beginning of most i guess it's only been four years but you know

237
00:23:30,740 --> 00:23:36,500
 the stuff four years ago that like was a flash in the pan and was never seen yeah yeah yeah the early

238
00:23:36,500 --> 00:23:40,980
 clients there yeah i mean you know for all i know in two years people are going to say what was that

239
00:23:40,980 --> 00:23:47,460
 FIPS thing again you know we'll see we'll see we'll see about it yeah i really didn't do FIPS with like

240
00:23:47,460 --> 00:23:52,260
 the strong idea that i was going to go build something that would last for generations you

241
00:23:52,260 --> 00:23:57,220
 know it was you know to scratch an itch yeah yeah see if you can solve and see if i can solve it and

242
00:23:57,220 --> 00:24:02,340
 i mean i'll i think i solved it pretty well yeah you know i'll go on record saying that

243
00:24:02,340 --> 00:24:08,420
 but you know i could be proven wrong we have this demo day tomorrow the last one of the of the cohort

244
00:24:08,420 --> 00:24:14,020
 and i have a feeling that several of the demos demos are going to be people breaking FIPS

245
00:24:14,020 --> 00:24:17,860
 okay we'll see what happens we'll see what's that you know there's just been like these little rumor

246
00:24:17,860 --> 00:24:21,940
 mills like yeah he's working on something but he's not going to tell me until he gets to show off

247
00:24:21,940 --> 00:24:29,300
 you know in front of the group that hey i broke FIPS you know and uh i'll be humbled i have one last

248
00:24:29,300 --> 00:24:34,820
 let's see what happens i have one last question for you um regarding the tools that you mentioned you

249
00:24:34,820 --> 00:24:40,660
 mentioned the agentic tools as well how do you think about it right now and what did you use to

250
00:24:40,660 --> 00:24:46,420
 to build FIPS in terms of agentic tooling and other lands and so on so a couple there's a couple of

251
00:24:46,420 --> 00:24:52,420
 questions are obviously related uh so for FIPS i mean i'm a i'm a huge Claude Code user and and i will

252
00:24:52,420 --> 00:24:59,300
 freely say that Claude Code has been part of the both the design and the implementation of FIPS from the

253
00:24:59,300 --> 00:25:08,180
 beginning i've been using agentic tools for about a year um they're quite good uh especially in the last

254
00:25:08,180 --> 00:25:14,020
 like four months even yeah you know you can see the the increase in productivity they give you

255
00:25:14,020 --> 00:25:22,180
 my main issue with uh agentic coding is that it doesn't replace the need for engineering design

256
00:25:22,180 --> 00:25:33,140
 and people tend to blame their tool instead of themselves and uh i mean i'll say like it's easier to

257
00:25:33,140 --> 00:25:39,780
 play claude wrote 99 of FIPS yeah but i've reviewed every single file every line of code

258
00:25:39,780 --> 00:25:47,220
 i spent about two months iterating with claude on a design for it about how the protocol worked the

259
00:25:47,220 --> 00:25:54,180
 different layers how they interact the corner cases and all of that and i had you know 30 000 words of

260
00:25:54,180 --> 00:26:01,060
 documentation about the design before i wrote any code and then writing the code became easy because all the

261
00:26:01,060 --> 00:26:05,700
 ambiguity was gone you know it was like it had all clear instructions on how to do everything so you

262
00:26:05,700 --> 00:26:11,300
 did lots of rubber duck debugging with claude oh yeah leading up to writing code you know and

263
00:26:11,300 --> 00:26:17,380
 that's the same thing that traditional software developers do so that hasn't changed it's just

264
00:26:17,380 --> 00:26:23,780
 become much easier much faster to do and i have people you know i hear over here people say you know i use

265
00:26:23,780 --> 00:26:30,580
 claude for this or or my agent broke this or whatever and you know you don't blame your spreadsheet you

266
00:26:30,580 --> 00:26:36,020
 don't blame your your compiler you don't say you know my linter did this you just say i did this yeah

267
00:26:36,020 --> 00:26:42,340
 so why are you blaming the random number generator yeah you know so so lms i think are extremely powerful

268
00:26:42,340 --> 00:26:48,020
 tool obviously um and they've sort of unlocked a huge amount of you know like every shower thought

269
00:26:48,020 --> 00:26:51,940
 you've had for the last 30 years you're gonna go try to make it work at least that's the case for me yeah

270
00:26:51,940 --> 00:27:03,380
 but they're a tool in your toolbox or tool chest of things to use to solve a problem but it doesn't

271
00:27:03,380 --> 00:27:10,660
 change the fundamental calculus of know the problem you're trying to solve design a solution remove all

272
00:27:10,660 --> 00:27:15,220
 the ambiguity from it to get you know an implementation plan that you know is going to make it do this

273
00:27:15,220 --> 00:27:22,820
 and then at that point the coding just writes itself it's like um it's like compiling code now you're

274
00:27:22,820 --> 00:27:29,220
 compiling ideas to source code but you got to have all the ideas right first and then the source code

275
00:27:29,220 --> 00:27:34,900
 gets compiled down to machine code and all that we've gone up a layer yeah you know so yeah you're not

276
00:27:34,900 --> 00:27:43,220
 writing the code cloud is or codex or whatever you happen to be using but if you want it to work well

277
00:27:43,220 --> 00:27:49,140
 you have to write it correct you have to write the idea correctly now just like with a compiler you have

278
00:27:49,140 --> 00:27:53,620
 to write the code correctly for it to compile successfully and i guess that's the analogy

279
00:27:53,620 --> 00:28:00,260
 i think pretty much all of the pull requests i've gotten in FIPS

280
00:28:01,220 --> 00:28:05,700
 which have all been well mostly have been pretty good i think you've all been written by cloud yeah

281
00:28:05,700 --> 00:28:10,260
 yeah you know and some of my reviews have been written by cloud yeah you know well i mean yeah

282
00:28:10,260 --> 00:28:14,740
 yeah you mentioned that because we talked about drive-by pull requests and how do you deal with

283
00:28:14,740 --> 00:28:19,700
 well and so on and you were like you know just you just sent the slop back you match effort for effort

284
00:28:19,700 --> 00:28:27,620
 yeah if somebody throws a pull request that was clearly the output of their tooling with no you see no

285
00:28:27,620 --> 00:28:35,060
 evidence of like review or editing or or their own words anywhere in it i'm just going to say hey claude

286
00:28:35,060 --> 00:28:39,380
 do a review and post it i'm not even going to look at it and i'm not going to look at what claude says

287
00:28:39,380 --> 00:28:44,580
 you know just say you know it's going to address you know a bunch of stuff that you'll go back and

288
00:28:44,580 --> 00:28:50,580
 have your cloud address and all that if someone takes the time to use it as a tool and say here's my

289
00:28:50,580 --> 00:28:55,220
 pull request here's what i'm trying to solve here's the problem you know and you can tell that

290
00:28:55,220 --> 00:29:01,620
 they are involved then that's a high value pull request and you might figure out a bug and a

291
00:29:01,620 --> 00:29:06,260
 solution to it in a few hours that might have taken weeks and that's the value of the llm

292
00:29:06,260 --> 00:29:14,100
 but we have a little we have had a little bit of a problem with you know just agents like seemingly

293
00:29:14,100 --> 00:29:22,260
 unknowing to their owner posting issues on the repo and you're like well this is completely out of

294
00:29:22,260 --> 00:29:27,060
 context sometimes even on unrelated uh issues you know they'll post a comment that's like it has

295
00:29:27,060 --> 00:29:34,580
 nothing to do with you know yeah in any case yeah so to answer your question um i'm very pro agentic

296
00:29:34,580 --> 00:29:41,060
 coding uh but you have to own it and it doesn't replace actual engineering yeah it doesn't replace

297
00:29:41,060 --> 00:29:46,580
 human thought yeah yet at least we'll see well yeah now maybe a year from now it really will be you

298
00:29:46,580 --> 00:29:51,860
 know claude make me a sandwich yeah yeah you know and there's a sandwich you know but um

299
00:29:51,860 --> 00:30:00,660
 uh the the the agentic tools let you talk to the tool at the speed of your own thought

300
00:30:00,660 --> 00:30:04,980
 which is really good because you can iterate quickly and you don't have to spend

301
00:30:04,980 --> 00:30:12,740
 a week translating a single thousand years into code to test it you know you can you can think like you're

302
00:30:12,740 --> 00:30:16,580
 talking to another person like we're talking i think i think that's also what makes it so addicting

303
00:30:16,580 --> 00:30:21,220
 because you can stay in the in the flow in the zone basically and just go back and forth you don't have

304
00:30:21,220 --> 00:30:26,580
 to wait for whatever another team to come back to you with feedback or what have you and all those kind

305
00:30:26,580 --> 00:30:31,060
 of things but yeah i think that's what's interesting it's also i mean you know it's a double-edged sword

306
00:30:31,060 --> 00:30:35,700
 because they kind of work like slot machines as well where you you prompt something in and it's very

307
00:30:35,700 --> 00:30:40,180
 exciting what will come back and then you just you stay on it forever and you just it can be very

308
00:30:40,180 --> 00:30:46,580
 addicting i mean you can tell some people are like that you know yeah it's where ai psychosis comes

309
00:30:46,580 --> 00:30:51,060
 from you just you're up all night prompting and and if you ask them how does this work they're like

310
00:30:51,060 --> 00:30:56,180
 well let me go ask claude what he did yeah you know it's like well if you haven't reviewed the output

311
00:30:56,180 --> 00:31:02,180
 then yeah yeah it could be anything it's definitely a double-edged sword because it i mean the evidence

312
00:31:02,180 --> 00:31:06,580
 is clear that that llm use for students for example it will make you on average it will make you a

313
00:31:06,580 --> 00:31:11,780
 number um i think there is a way to use them that actually helps and assists in learning and makes

314
00:31:11,780 --> 00:31:18,100
 you learn faster i yeah i mean that takes but it takes discipline i think it takes a teacher you know

315
00:31:18,100 --> 00:31:23,060
 a human teacher you know in the case of students to help direct them on how to use the tooling but

316
00:31:23,060 --> 00:31:28,180
 the teachers are learning at the same time and this is new to everybody yeah it's it's absolutely but

317
00:31:28,180 --> 00:31:35,940
 my final snarky comment on it is uh agentic tools will make bad programmers worse

318
00:31:35,940 --> 00:31:42,980
 it'll make good programmers non-productive yeah or a good program oh that sounded bad it'll make good

319
00:31:42,980 --> 00:31:48,820
 programmers more productive and it will allow non-programmers to actually program for the first

320
00:31:48,820 --> 00:31:55,060
 time yeah and we've had one attendee here at SEC who has never programmed before but they were able to

321
00:31:55,780 --> 00:32:00,980
 uh design a solution for something that you know was within the spirit of how SEC you know works

322
00:32:00,980 --> 00:32:07,860
 and you know it worked well and everything and it was just an incredible use of or extension of

323
00:32:07,860 --> 00:32:11,940
 somebody i want a solution to exist i don't know how to do it because i've never coded before

324
00:32:11,940 --> 00:32:17,780
 but this person was able to use the tool to create the solution and it actually worked

325
00:32:18,500 --> 00:32:22,100
 and you know there you could see that you know as iterative and you know they fixed things that

326
00:32:22,100 --> 00:32:27,140
 they didn't like and you know so it wasn't just random output it was yeah you have to sit with

327
00:32:27,140 --> 00:32:33,300
 it and and iterate you know and and that has never been possible before you know to to have never

328
00:32:33,300 --> 00:32:39,060
 coded but have a crystal clear idea of of what you want to exist and to be able to translate that

329
00:32:39,060 --> 00:32:46,020
 through uh lms yeah no awesome i promised you to keep it short so i've won one final what may you

330
00:32:46,020 --> 00:32:50,420
 give one final piece of advice to the younglings out there what have you learned from your early

331
00:32:50,420 --> 00:32:55,700
 cypherpunk days and your early Bitcoin days that might be useful for the you know 20-somethings that

332
00:32:55,700 --> 00:33:01,140
 might be listening to this that don't have your experience and your wisdom let me think about

333
00:33:01,140 --> 00:33:08,340
 this for a minute before it gets committed to uh to tape um take all the time you need i'd say don't

334
00:33:08,340 --> 00:33:20,100
 don't be afraid of going your own way there's a tremendous um inertia in the world that is pushing

335
00:33:20,100 --> 00:33:30,580
 you to turn the crank to conform to um act within the system and you know i'm not talking about

336
00:33:31,140 --> 00:33:35,220
 um you know going off and doing crazy things i'm talking about

337
00:33:35,220 --> 00:33:43,220
 your own mind and your own ability to think and to make judgment calls about how the world works

338
00:33:43,220 --> 00:33:51,700
 will lead you to more success than just believing what you've been taught and you know your parents

339
00:33:51,700 --> 00:33:55,620
 raised you in a way that was the best way they knew how but they filled you with information that will

340
00:33:55,620 --> 00:34:02,820
 only get you so far and then you know it'll run out and eventually you get to the point where you're

341
00:34:02,820 --> 00:34:08,340
 like wait a minute that's not how the world works and the ability to have confidence in your own

342
00:34:08,340 --> 00:34:14,020
 ability to look take a fresh look at how things are will go a very long way in helping you not to

343
00:34:14,020 --> 00:34:16,180
 you know struggle against

344
00:34:16,180 --> 00:34:22,900
 that feeling like i know it's not right but i'm gonna like do things the way i'm supposed to and

345
00:34:22,900 --> 00:34:29,780
 i'm gonna you know the fiat world so to speak you know when you're that age you know you can do

346
00:34:29,780 --> 00:34:34,420
 anything you get my age and you've kind of set your place in the world and it's kind of hard to

347
00:34:34,420 --> 00:34:37,300
 to change you know so don't be afraid to

348
00:34:37,300 --> 00:34:45,140
 don't be afraid to dissent and don't be afraid to say you know what i'm going to go try doing it

349
00:34:45,140 --> 00:34:51,380
 this way instead of the way that you know and walk your own way yeah make your own path it's very cliche

350
00:34:51,380 --> 00:34:58,740
 but it's it's very real and it's probably more relevant today than ever is um uh there's a there's

351
00:34:58,740 --> 00:35:06,420
 a layer of conformity that's pushed upon you constantly every day and uh you know stop scrolling

352
00:35:06,420 --> 00:35:13,140
 and start doing stuff and and that's the thing about cliches right the older you get the truer they

353
00:35:13,140 --> 00:35:20,900
 become yeah that's true i should take my own advice all right that was awesome i hope you had fun as

354
00:35:20,900 --> 00:35:28,180
 well we always end on the hug may i sure thanks so much for doing all this building all of it being

355
00:35:28,180 --> 00:35:36,260
 here coming here it's been wonderful and i'm glad we pulled you out of retirement so now i have 10

356
00:35:36,260 --> 00:35:41,700
 more years of supporting FIPS before i can retire again it's it's gonna be fine the young links will

357
00:35:41,700 --> 00:35:42,620
 We'll step up eventually.

358
00:35:42,620 --> 00:35:43,780
 They absolutely will.

359
00:35:43,780 --> 00:35:44,800
 All right.

360
00:35:44,800 --> 00:35:45,420
 Thanks so much.