CD174: ALEX GLEASON - VIBE CODING UPDATE

WEBVTT

NOTE
Transcription provided by Podhome.fm
Created: 08/15/2025 21:14:23
Duration: 4261.739
Channels: 1

1
00:00:01.839 --> 00:00:03.060
You're talking about

2
00:00:03.919 --> 00:00:06.640
a number way above 750,000,000,000.

3
00:00:06.640 --> 00:00:08.020
What's the story on gold?

4
00:00:08.960 --> 00:00:09.460
Look,

5
00:00:09.760 --> 00:00:18.904
we we've got this the large gold holding. I I doubt we're gonna revalue it, but we we are going to keep it there, again, as a store of value

6
00:00:19.205 --> 00:00:21.865
for the American people. We've also started,

7
00:00:22.325 --> 00:00:26.920
a to get into the twenty first century, a Bitcoin strategic reserve.

8
00:00:27.300 --> 00:00:30.039
We're not gonna be buying that, but we are going to,

9
00:00:30.900 --> 00:00:31.800
use confiscated

10
00:00:32.340 --> 00:00:40.280
assets and continue to build that up. We're gonna stop selling that. You know, I I believe that Bitcoin reserve at today's prices

11
00:00:40.905 --> 00:00:43.405
is somewhere between 15 and 20,000,000,000.

12
00:00:46.425 --> 00:00:50.285
And and, mister secretary, all of these ideas to monetize the assets.

13
00:01:29.030 --> 00:01:34.409
Happy Bitcoin Friday, freaks. It's your host, Odell, here for another civil dispatch,

14
00:01:34.710 --> 00:01:35.530
the interactive

15
00:01:36.550 --> 00:01:41.690
live show focused on actionable Bitcoin and Freedom Tech discussion.

16
00:01:43.165 --> 00:01:44.384
That intro clip

17
00:01:44.685 --> 00:01:47.025
was treasury secretary percent.

18
00:01:47.645 --> 00:01:49.025
I think it was a Fox

19
00:01:49.884 --> 00:01:51.185
Fox News interview yesterday

20
00:01:52.284 --> 00:01:54.305
that sent a bunch of people into a tizzy,

21
00:01:55.480 --> 00:01:59.020
because he said he wasn't gonna buy Bitcoin or the government wasn't gonna buy Bitcoin,

22
00:01:59.720 --> 00:02:03.740
and that they were gonna confiscate Bitcoin to add to the strategic Bitcoin reserve.

23
00:02:04.360 --> 00:02:06.780
The second part was not really news, but

24
00:02:07.985 --> 00:02:09.605
really sucks to hear it out loud.

25
00:02:10.785 --> 00:02:16.325
Probably will always suck to hear it out loud. He did say that they have about a 125,000.

26
00:02:17.745 --> 00:02:22.885
Kind of a questionable number considering I'm pretty sure they still have a 100,000 Bitcoin that they

27
00:02:23.379 --> 00:02:27.319
recovered from the Bitfinex hack, which should be returned to Bitfinex and Tether.

28
00:02:28.500 --> 00:02:31.879
So they might have way less Bitcoin than they think, or they might

29
00:02:32.260 --> 00:02:33.000
be playing

30
00:02:35.575 --> 00:02:40.555
a pretty fucked up game and not giving that a 100,000 Bitcoin back to Bitfinex. Either way,

31
00:02:41.814 --> 00:02:49.194
it seems like percent clarified his comments later and said that they still intend to acquire Bitcoin through budget neutral ways, whatever that means.

32
00:02:50.200 --> 00:02:56.220
This is completely unrelated to the show, but it felt like an important time stamp to put here. I would just remind people.

33
00:02:58.599 --> 00:03:03.180
I actually it actually is kind of relevant to our show. We have Alex Gleason joining us again,

34
00:03:03.995 --> 00:03:07.375
for we're gonna do updates every two to three months. How's it going, Alex?

35
00:03:07.675 --> 00:03:08.495
Hello. Hello.

36
00:03:09.275 --> 00:03:11.375
I did your intro a little bit early because,

37
00:03:12.075 --> 00:03:16.015
you know, as you might have remember from our last conversation with Alex,

38
00:03:16.670 --> 00:03:20.290
which, like I said, happened about two months ago. I think that was episode

39
00:03:20.910 --> 00:03:23.890
one sixty four. Give that a listen if you haven't yet.

40
00:03:25.630 --> 00:03:31.310
He helped build Truth Social, and the public company that is Truth Social now has 15,000

41
00:03:31.310 --> 00:03:31.810
Bitcoin.

42
00:03:32.965 --> 00:03:37.065
So Trump has packed his bags. Him and his cronies have a shit ton of Bitcoin.

43
00:03:37.845 --> 00:03:44.425
I think that's really the only bullet point you need in terms of wondering whether or not they're gonna try and pump the

44
00:03:44.870 --> 00:03:48.650
living daylights out of Bitcoin. I mean, the self interest is clearly there.

45
00:03:49.269 --> 00:03:50.569
But I skipped a step.

46
00:03:51.829 --> 00:03:53.450
Freaks, as always, dispatch

47
00:03:54.069 --> 00:04:01.925
is brought to brought to us by viewers like you. We have no ads or sponsors. Thank you to everyone who supports the show by donating Bitcoin.

48
00:04:02.305 --> 00:04:04.965
The most fun way to do that is through Nostr,

49
00:04:05.265 --> 00:04:06.644
either in our live chat

50
00:04:07.025 --> 00:04:09.045
or through your favorite Nostr app.

51
00:04:09.905 --> 00:04:12.885
If you've never used Nostr, I suggest try and download Primal.

52
00:04:13.580 --> 00:04:15.680
Download the Primal app in your favorite app store.

53
00:04:16.060 --> 00:04:17.840
Give it a spin. Give your commentary.

54
00:04:18.780 --> 00:04:19.280
Interact

55
00:04:20.140 --> 00:04:21.920
with the show. It's a great way

56
00:04:22.300 --> 00:04:22.960
to participate,

57
00:04:23.260 --> 00:04:25.600
and I just love kicking it with you freaks.

58
00:04:26.565 --> 00:04:31.625
Another great way to support the show is fountain podcast or any other podcasting two point o app.

59
00:04:32.565 --> 00:04:35.925
The largest app from last week was Nostra Kaye with 21,000

60
00:04:35.925 --> 00:04:36.425
sats

61
00:04:36.885 --> 00:04:38.105
saying highly speculative

62
00:04:38.740 --> 00:04:42.600
and degen, but obviously enjoyed stay humble and stack sats. Great advice.

63
00:04:43.540 --> 00:04:45.480
And our highest sap on

64
00:04:46.100 --> 00:04:46.600
Noster

65
00:04:47.940 --> 00:04:49.960
was Mav 21

66
00:04:50.695 --> 00:04:52.455
with 10,031

67
00:04:52.455 --> 00:04:54.075
sets. Thank you, Mav 21.

68
00:04:55.255 --> 00:04:58.795
Anyway, guys, all links to the show is at civildispatch.com.

69
00:04:58.935 --> 00:05:04.475
Everything relevant is there. Share with your friends and family. It is helpful. Alex, good to have you back.

70
00:05:05.600 --> 00:05:06.740
Happy to be here.

71
00:05:10.240 --> 00:05:11.460
How's everything been?

72
00:05:12.320 --> 00:05:13.620
It's been good. So

73
00:05:14.000 --> 00:05:15.300
when we last talked,

74
00:05:15.600 --> 00:05:18.740
you know, we were talking about stacks, and we were talking about

75
00:05:19.465 --> 00:05:21.485
people vibe coding stuff on Noister

76
00:05:22.025 --> 00:05:25.405
and just, like, the potential that that can have for the future.

77
00:05:26.105 --> 00:05:32.605
I think it's one of the it's like the frontier of freedom online right now is AI.

78
00:05:33.900 --> 00:05:34.400
And

79
00:05:35.420 --> 00:05:36.560
and so I'm

80
00:05:37.100 --> 00:05:40.000
I've been building tools to make it easier for people to

81
00:05:40.460 --> 00:05:43.840
use AI to build their own sort of, like,

82
00:05:44.380 --> 00:05:47.200
you know, worlds online to be able to

83
00:05:47.855 --> 00:05:48.835
have that freedom.

84
00:05:51.455 --> 00:05:51.955
But

85
00:05:52.815 --> 00:05:53.635
I'm also

86
00:05:54.575 --> 00:06:00.035
I've also been feeling concerned about big AI and about how we can fight against

87
00:06:00.575 --> 00:06:01.075
that.

88
00:06:02.560 --> 00:06:06.580
And so these are kind of some of the things I've been working on and and thinking about.

89
00:06:07.120 --> 00:06:07.620
And

90
00:06:08.320 --> 00:06:09.060
most notably,

91
00:06:10.720 --> 00:06:13.380
since our last talk, I have launched Shakespeare,

92
00:06:14.215 --> 00:06:16.295
and that's shakespeare.diy.

93
00:06:16.295 --> 00:06:17.355
And that is a

94
00:06:17.735 --> 00:06:21.995
a vibe coding tool that people can use to develop Noister clients.

95
00:06:24.375 --> 00:06:25.035
I'm listening.

96
00:06:25.735 --> 00:06:32.800
Yep. And so that's been that's been really exciting. We've had people building, you know, just dozens of different Noister clients

97
00:06:33.260 --> 00:06:34.800
with this with this tool.

98
00:06:35.260 --> 00:06:46.525
And what's great about it is you don't have to have any technical knowledge or experience. You can just you know it's similar to other solutions like Lovable, VeoDuro. You can kinda just tell it what you want, and then it will build it.

99
00:06:48.345 --> 00:06:48.845
And

100
00:06:49.784 --> 00:06:50.604
and so

101
00:06:51.305 --> 00:06:59.630
I have a v one of this right now, which is what is live, and I'm currently working on a v two of Shakespeare that is gonna give people even more freedom and control,

102
00:07:00.250 --> 00:07:02.990
because the new version is gonna run entirely in the browser.

103
00:07:04.330 --> 00:07:05.150
That's badass.

104
00:07:06.705 --> 00:07:07.205
Yep.

105
00:07:07.905 --> 00:07:08.145
So

106
00:07:08.785 --> 00:07:09.285
Yeah.

107
00:07:09.985 --> 00:07:13.365
Sorry. I'm having technical difficulties over here that I'm trying to Okay.

108
00:07:16.225 --> 00:07:22.324
Troubleshoot live on air. I think I got us to a decent spot. Let's see if this works.

109
00:07:24.040 --> 00:07:27.900
I lost my live chat freaks. It's all about the live chat on dispatch.

110
00:07:33.640 --> 00:07:34.140
Inspect.

111
00:07:42.265 --> 00:07:42.765
Delete.

112
00:07:46.265 --> 00:07:46.765
Okay.

113
00:07:47.625 --> 00:07:48.285
I guess

114
00:07:50.060 --> 00:07:51.500
I got a Vibe code, a new,

115
00:07:52.220 --> 00:07:53.520
live chat set up.

116
00:07:54.700 --> 00:07:59.120
I wanna talk about, I wanna talk about the concerns about big AI real quick.

117
00:07:59.580 --> 00:08:06.435
Sure. Because this is something that I'm intimately concerned with. Look. Look at that. There we go. We got our live chat back.

118
00:08:08.655 --> 00:08:09.155
I,

119
00:08:10.335 --> 00:08:14.675
so it's it's an interesting phenomenon, twofold. So I'll I'll start the

120
00:08:15.055 --> 00:08:16.530
I'll start the discussion here.

121
00:08:17.490 --> 00:08:19.110
First off, it's interesting

122
00:08:19.570 --> 00:08:20.070
that,

123
00:08:21.490 --> 00:08:26.310
in The US, we probably are in the lead in terms of models and tools,

124
00:08:27.250 --> 00:08:28.470
but it's all proprietary

125
00:08:29.169 --> 00:08:30.630
and closed source

126
00:08:31.155 --> 00:08:31.975
except for

127
00:08:32.755 --> 00:08:33.255
recently,

128
00:08:34.355 --> 00:08:37.015
OpenAI finally released an open source model.

129
00:08:37.875 --> 00:08:41.495
That's decent, but not nearly as good as the proprietary closed models.

130
00:08:42.115 --> 00:08:43.895
And then China, on the other hand,

131
00:08:44.279 --> 00:08:51.019
is, like, just aping into open source. They're kinda also playing catch up, but, like, all the top open source models are really out of China.

132
00:08:51.399 --> 00:08:56.540
Meta had an open source model. Llama, that's just complete shit. Doesn't really is not really competitive.

133
00:08:59.075 --> 00:09:08.775
The one caveat we have on the American side is, you know, Block came out with Goose, but that's on, like, the agent side. It's more of, like, a front end. It's not the actual model itself.

134
00:09:09.955 --> 00:09:11.655
And then to top it all off,

135
00:09:15.090 --> 00:09:23.510
like, there's a lot of speculation on, like, okay. What will LLMs be capable in the future? What kind of new functionality will they provide us? But in the short term,

136
00:09:23.890 --> 00:09:26.850
the single thing that they're very, very good at is,

137
00:09:28.185 --> 00:09:28.685
analyzing,

138
00:09:29.145 --> 00:09:38.125
like, mass hoards of harvested data. It's, like, perfect for the surveilling like, levels up their surveillance state almost instantly without any improvements really needing to be made.

139
00:09:39.720 --> 00:09:45.100
So how do you think about that? Just like the general I mean, I kinda gave you two different things, but

140
00:09:45.560 --> 00:09:53.464
Oh, I I I have, like, ten ten different things I wanna say. Hopefully, I can keep it all in my in my head and say it all. Go for it. On GPT's

141
00:09:53.765 --> 00:09:55.065
open source models,

142
00:09:56.485 --> 00:10:05.545
I noticed that they don't dog food them at all. You can't even use the their open source models through their web UI, which is really annoying. Like, it's it's almost just like a,

143
00:10:06.930 --> 00:10:11.670
you know, just like a bone that they're throwing to somebody, maybe to Trump. Because

144
00:10:12.210 --> 00:10:12.710
in,

145
00:10:13.250 --> 00:10:21.830
in Trump's AI, like, for America document, it actually listed open source as one of the main ways to make America AI win.

146
00:10:22.394 --> 00:10:23.935
And I thought that was really interesting,

147
00:10:25.115 --> 00:10:26.014
rare w.

148
00:10:26.475 --> 00:10:26.975
But

149
00:10:27.355 --> 00:10:29.615
so maybe it's, like, just trying to appease that.

150
00:10:30.714 --> 00:10:32.975
In terms of China leading

151
00:10:33.675 --> 00:10:37.774
open source models, yes. And also, I think that this kind of proves

152
00:10:39.010 --> 00:10:40.550
that, like, the the

153
00:10:40.930 --> 00:10:46.790
how open source can be a thing that gets your foot in the door to make it something that's relevant.

154
00:10:47.250 --> 00:10:58.005
Right? Like, they don't have the best models, but they have open source models. So, like, people are seeing what they're doing. Right? And I'm sure that in China, there's all kinds of organizations making

155
00:10:58.545 --> 00:11:05.445
models that are not open source. But we we only hear about the ones that are open source because open source is what allows them to win.

156
00:11:05.829 --> 00:11:16.570
Right. And and this kind of overlap Like, I would never use it. I would never use a Chinese closed source model. I mean, I'm probably also one of the few that would never I I don't use the proprietary

157
00:11:17.029 --> 00:11:24.395
US models either, but I definitely wouldn't use a Chinese closed model. And then I end up using the Chinese open models because it's the best open models that are available.

158
00:11:25.255 --> 00:11:35.675
Yeah. And and so I've been experimenting with some of those. I tried Kimmy k two for, like, a week, and I was really impressed with it at first, but then I ended up going back to Claude because it just wasn't there.

159
00:11:36.260 --> 00:11:43.720
And I've been researching, like, how do you run this thing? And it's like, okay. Well, you need a million dollars worth of graphics cards.

160
00:11:44.100 --> 00:11:59.065
So and and that's just that's just the initial cost of the hardware. That's not even the continual running cost of it. So, like, that's a barrier. And and I and then I'm saying, like, so can you tell me about Tiananmen Square? And it's like, Tiananmen Square? I don't know what you're talking about.

161
00:11:59.685 --> 00:12:16.050
And and so it just felt to me like it even though it's open source, it violates the principles of open source by being something that's literally impossible for normal people to run while also having censorship built in. But then again, the American models also have censorship built in on on all kinds of different things too.

162
00:12:17.245 --> 00:12:21.505
And so okay. So Kimi k two, it wasn't really it. But then recently,

163
00:12:22.205 --> 00:12:23.985
z AI came out with GLM.

164
00:12:24.365 --> 00:12:35.829
And this is the one that's that has been finally the one to push me over onto onto the Chinese AI side. And now I'm, like, I'm gonna be speaking Mandarin in in a year. I'm gonna be, like, in Taiwan.

165
00:12:37.570 --> 00:12:43.829
And I feel like China has entered into my life, and I'm really curious and interested in thinking about

166
00:12:44.255 --> 00:12:45.715
China a lot. Probably

167
00:12:46.815 --> 00:12:51.315
I'm thinking about China more now than I was when I was working with Donald Trump.

168
00:12:51.935 --> 00:12:52.435
Well

169
00:12:53.295 --> 00:12:57.395
well, let's let's just pull on that thread for a second, this thought on,

170
00:12:59.220 --> 00:13:00.920
because it's not really censorship.

171
00:13:01.380 --> 00:13:03.960
Censorship's the wrong word. We almost need a new word.

172
00:13:04.580 --> 00:13:08.440
But when you're using these models, like, the way they're trained or what?

173
00:13:08.900 --> 00:13:16.995
I think it kind of is censorship because they they train it to specifically not respond to certain requests. Or respond a certain way.

174
00:13:17.615 --> 00:13:22.355
Right. And and I think, like, in Tiananmen Square, there's training data that instructs it.

175
00:13:23.135 --> 00:13:35.250
I mean, it it's it's self censorship in a way in the same way that you you would tell, like, a person that they're not really censoring us. Yeah. They're not censoring you. Right. It's true. They're blocking our access to the information.

176
00:13:35.870 --> 00:13:36.690
You're right.

177
00:13:37.070 --> 00:13:37.570
But

178
00:13:38.985 --> 00:13:41.165
it's just so, I mean,

179
00:13:41.785 --> 00:13:43.385
Tiananmen Square is a good,

180
00:13:43.945 --> 00:13:45.725
example, and I guess

181
00:13:46.025 --> 00:13:47.485
I'm sorry for the

182
00:13:48.105 --> 00:13:50.845
Chinese listeners we have because this episode is probably

183
00:13:51.500 --> 00:13:54.540
just gonna get, like, blackballed all over there right now. But,

184
00:13:55.500 --> 00:13:56.960
it's a very extreme example.

185
00:13:57.340 --> 00:14:02.400
But the the crazier thing to me is, like, the more nuanced examples that are happening on in the West

186
00:14:03.065 --> 00:14:12.204
that are harder to pinpoint where they're happening. And sometimes it's probably even unintentional. Right? It's like, where are they getting the training data from? Like, they could get the training data from

187
00:14:12.665 --> 00:14:18.330
Wikipedia, and Wikipedia could have a slant or a specific Wikipedia entry could have a slant, and then the

188
00:14:18.810 --> 00:14:20.110
model has a slant.

189
00:14:20.490 --> 00:14:24.510
Yep. But that's pretty dark. It's because it's so hard. It's, like, not even book burning.

190
00:14:25.610 --> 00:14:26.110
It's,

191
00:14:26.890 --> 00:14:28.430
like, the book didn't exist.

192
00:14:28.810 --> 00:14:44.385
I you know what I mean? It's, like, kind of a weird thing to wrap your head around. Yeah. It's funny because if China would just give up on this one thing and just be like, yeah. Yeah. Tiananmen Square happened, then, like, it would be so much easier to just ignore all of the little things that Right. That they do that are still That's the western way.

193
00:14:44.925 --> 00:14:45.425
Exactly.

194
00:14:45.965 --> 00:14:46.465
Yeah.

195
00:14:47.550 --> 00:14:51.810
Yeah. I wonder, like I mean, to me, like, there's something there about so

196
00:14:52.830 --> 00:14:59.010
I first went down, like, the and I don't really use I've I I know I promised you that I was gonna

197
00:14:59.550 --> 00:15:02.350
start vibe coding by the time I had you back on, but,

198
00:15:03.225 --> 00:15:14.285
fortunately, we've had you on sooner. And second of all, I haven't actually held up onto my side of the bargain. So I don't really use I don't I don't use these tools for that. I use them more,

199
00:15:15.065 --> 00:15:16.925
like, chatbot as, like, chatbots

200
00:15:17.310 --> 00:15:22.770
and and, you know, like, virtual assistant type of scenarios. And I mostly use Maple.

201
00:15:24.430 --> 00:15:26.530
So I first started by self hosting.

202
00:15:26.910 --> 00:15:31.330
Right? And I would run these models, and I'd have to run, like, the really gimped models because

203
00:15:32.024 --> 00:15:34.204
my computer is not $5,000.

204
00:15:34.264 --> 00:15:41.485
You know? And I and then I switched to Maple because they give me pretty strong privacy and security guarantees. It's, like, kind of a similar model to signal.

205
00:15:42.345 --> 00:15:45.165
But I can use, like, the latest and greatest open models.

206
00:15:46.310 --> 00:15:48.730
And the cool part there with Maple is,

207
00:15:49.190 --> 00:15:55.770
like, I can pick which model I use. Right? It's got, like, a drop down of, like like, seven of the top ones or whatever that they support.

208
00:15:57.750 --> 00:16:02.135
And I think there's something there, and we're not quite there yet, but there's something there about

209
00:16:02.915 --> 00:16:09.015
kind of running them in that environment. It's harder to do it in a self hosted environment, but in that environment, kind of running them simultaneously

210
00:16:10.035 --> 00:16:11.495
to check each other's work

211
00:16:12.035 --> 00:16:12.775
might be

212
00:16:14.100 --> 00:16:17.800
it's not really a solution, but it might be the mitigation to that kind of foundational

213
00:16:19.139 --> 00:16:19.639
level

214
00:16:20.019 --> 00:16:23.160
bias or slant or censorship, whatever you wanna call it.

215
00:16:23.540 --> 00:16:37.735
Yeah. I've been just trying to figure out how how to self host because all of these services people are using are are through some, you know, API that's run by somebody else. And I really just I really just wanna host it my myself. Like, you know,

216
00:16:38.675 --> 00:16:41.574
I feel like ten years ago, the fight was

217
00:16:42.069 --> 00:16:42.569
Twitter,

218
00:16:43.029 --> 00:16:44.970
and we need to self host Twitter.

219
00:16:46.149 --> 00:16:55.850
And so, like, Mastodon was sort of the biggest first introduction to that, which is when I got started in this movement. And now there's just so many ways to self host Twitter

220
00:16:57.075 --> 00:17:03.815
that it it's become this saturated market. And now with Shakespeare, you could just, like, vibe code a kind one client

221
00:17:04.434 --> 00:17:06.135
Right. However you want. So

222
00:17:06.755 --> 00:17:08.775
now it feels like the big

223
00:17:09.090 --> 00:17:12.549
impossible thing that I'm reaching for is

224
00:17:13.330 --> 00:17:14.950
self hosting AI.

225
00:17:15.809 --> 00:17:21.030
And how can we do this in a way to where people can actually have control over this revolutionary

226
00:17:21.890 --> 00:17:23.190
and potentially dangerous

227
00:17:23.585 --> 00:17:28.485
technology, dangerous if they can't if normal people don't have the means to control it themselves.

228
00:17:28.865 --> 00:17:31.445
So how do you I mean, how do you think that looks besides

229
00:17:32.785 --> 00:17:34.005
waiting ten years

230
00:17:34.945 --> 00:17:35.445
for,

231
00:17:36.010 --> 00:17:36.830
like, hardware

232
00:17:37.370 --> 00:17:42.730
hardware costs to be lower. Like, if everyone's if if the requirement is a $5,000

233
00:17:42.730 --> 00:17:43.230
computer,

234
00:17:45.210 --> 00:17:48.010
there's gonna be ten years much higher than a $5,000

235
00:17:48.010 --> 00:17:51.710
computer. But even just I mean, even to be, like, in relative competency,

236
00:17:52.755 --> 00:17:54.695
like, to to even be close to

237
00:17:55.155 --> 00:17:55.815
the completely

238
00:17:56.835 --> 00:18:01.655
you know, not only hosted, but, like, controlled walled garden KYC walled garden type of stuff,

239
00:18:02.755 --> 00:18:05.575
you need to have a very expensive piece of hardware.

240
00:18:06.120 --> 00:18:06.860
Yes. Exactly.

241
00:18:08.040 --> 00:18:11.900
And and so, I mean, there are advancements that people are making.

242
00:18:12.600 --> 00:18:13.340
In particular,

243
00:18:13.720 --> 00:18:16.940
Chinese open source models are making these advancements.

244
00:18:17.480 --> 00:18:19.100
So all of the main,

245
00:18:20.305 --> 00:18:27.685
like, big tech big AI companies are using dense transformer models, which is where, basically, whenever you ask it a question,

246
00:18:28.065 --> 00:18:34.725
it's responding to you with all of the world's information in that one response, and it has to activate all of the world's information,

247
00:18:35.950 --> 00:18:38.850
which requires electricity, which comes with a cost.

248
00:18:39.309 --> 00:18:44.289
Right. But these open source Chinese AI models are using a mixture of experts,

249
00:18:45.389 --> 00:18:53.265
approach in which only part of the AI model needs to be activated to to respond to your query. So if you're asking it a question about football,

250
00:18:53.645 --> 00:18:58.225
then maybe it only needs the sport expert to respond to that query. And the results is that

251
00:18:58.765 --> 00:19:01.800
it can be a lot cheaper and use a lot less

252
00:19:02.200 --> 00:19:06.300
electricity to respond to that particular query because only part of the model is activated.

253
00:19:06.680 --> 00:19:07.180
And,

254
00:19:08.280 --> 00:19:13.100
big AI companies are afraid to do this approach because they are all racing for AGI.

255
00:19:13.560 --> 00:19:14.620
And it's harder

256
00:19:14.945 --> 00:19:22.245
it's harder to to have a model that is that works really well when you have to balance this mixture of experts

257
00:19:22.625 --> 00:19:23.125
thing,

258
00:19:23.905 --> 00:19:30.485
as opposed to just having having it respond with all the world's information on each request. So this is just an example of,

259
00:19:31.770 --> 00:19:37.470
like, ways that the model itself can be structured differently so that it be can become more accessible

260
00:19:37.850 --> 00:19:39.070
to people. But

261
00:19:39.450 --> 00:19:45.174
the accessible version of this, after all of this is done right, requires $1,000,000

262
00:19:45.174 --> 00:19:46.715
worth of GPU hardware.

263
00:19:47.335 --> 00:19:47.835
So

264
00:19:48.294 --> 00:19:52.235
so, like, this is this is step one here and, like, how do we get

265
00:19:52.615 --> 00:19:53.755
to a step where

266
00:19:54.135 --> 00:19:55.674
even $5,000

267
00:19:55.735 --> 00:20:06.210
is good enough? I don't know. And it's even worse than, like, it requires $1,000,000 of the GPUs, but it to me I mean, I look at everything from a Bitcoin perspective. It reminds me of ASICs.

268
00:20:06.510 --> 00:20:08.529
Like, you pay a million dollars of the GPUs,

269
00:20:08.909 --> 00:20:10.370
and then in, like, four years,

270
00:20:10.924 --> 00:20:15.105
they're worth, like, a 100 k. Like, it's you're just, like, completely burning money.

271
00:20:15.485 --> 00:20:20.544
It's not like it's like you're investing for your family or your compute community and you're, like, good for a decade.

272
00:20:21.804 --> 00:20:23.265
Yeah. It's it's unfortunate.

273
00:20:25.610 --> 00:20:28.669
But my I my hope is that the models are gonna get cheaper.

274
00:20:29.210 --> 00:20:34.990
Another thing also that I've I was reading about this morning is that in in China,

275
00:20:35.370 --> 00:20:35.870
they're

276
00:20:36.410 --> 00:20:37.549
they have, like,

277
00:20:38.054 --> 00:20:41.674
so much energy that they don't know what to do with. Apparently

278
00:20:41.975 --> 00:20:45.355
Yeah. They they are running on 80%

279
00:20:45.414 --> 00:20:45.914
reserve

280
00:20:46.375 --> 00:20:55.220
at all times or higher. And in The US, like, in Texas here, twice a year, the power grid fails and we lose power for, like, a week because it rained a little bit.

281
00:20:55.600 --> 00:20:56.340
And so

282
00:20:57.360 --> 00:20:59.059
it's also cheaper to just

283
00:20:59.600 --> 00:21:03.299
consume this electricity in China because they have so much energy

284
00:21:03.920 --> 00:21:05.140
availability there.

285
00:21:07.255 --> 00:21:14.475
Yeah. I mean, that's a battle that's a battle that big corners have been fighting for a while. We need more energy production. It leads to human flourishing.

286
00:21:15.095 --> 00:21:23.149
China's way ahead of us in that queue. And this the stat is I might be I I'm definitely wrong, but it's something like China's adding, like, a full American

287
00:21:23.529 --> 00:21:27.549
America worth of capacity of energy every year or something insane like that.

288
00:21:28.409 --> 00:21:29.149
Yeah. Crazy.

289
00:21:29.450 --> 00:21:35.070
I I mean, it's interesting. Right? Because, like, on our side in America, we're constrained on the energy side,

290
00:21:35.524 --> 00:21:38.985
And that should probably change now that big tech is finally not

291
00:21:39.764 --> 00:21:46.105
concerned trolling Bitcoiners on energy usage and instead encouraging more energy production. So we're, like, finally aligned

292
00:21:46.485 --> 00:21:47.784
with, like, the Mag seven,

293
00:21:48.540 --> 00:21:50.720
for, like, the first time in over a decade,

294
00:21:51.580 --> 00:21:54.400
on that point. I mean, they were all the proof of stake maxis

295
00:21:54.700 --> 00:21:59.520
for a while. So that should hopefully at least start trending in a better direction.

296
00:22:00.154 --> 00:22:06.174
But so, like, we're energy restricted, and then they're chip restricted because, like, the export controls and whatnot.

297
00:22:06.635 --> 00:22:12.495
So we're they're on both sides, we're kinda operating under different and so then as a result,

298
00:22:12.910 --> 00:22:13.810
on on different,

299
00:22:14.350 --> 00:22:17.330
constraints. And so as a result, the model's kind of

300
00:22:17.710 --> 00:22:18.610
and, like, the

301
00:22:19.630 --> 00:22:23.730
the strategy is kind of modeled basically off of those constraints. Right?

302
00:22:24.190 --> 00:22:26.530
Right. But and and this this is why, ironically,

303
00:22:27.085 --> 00:22:33.505
China's models they're developing are more efficient than our models. Even though they have access to more energy, their chip restrictions

304
00:22:33.805 --> 00:22:37.105
are preventing them from being able to have enough GPU

305
00:22:37.565 --> 00:22:43.780
power and to run these bigger models efficiently. So they're kind of being forced to do these mixture of expert

306
00:22:44.080 --> 00:22:50.980
things, but then that ends up being better for us because now, like I mean, a million dollars is not, you know, inaccessible

307
00:22:51.600 --> 00:22:52.100
here.

308
00:22:52.480 --> 00:22:53.460
It's just like

309
00:22:54.305 --> 00:23:00.165
it it it it's much lower of a barrier than it it would be to run GPT five.

310
00:23:00.625 --> 00:23:01.125
Right.

311
00:23:02.145 --> 00:23:04.885
Okay. All this is fascinating. So let's talk about

312
00:23:05.745 --> 00:23:07.285
Shakespeare. I see someone

313
00:23:08.870 --> 00:23:16.330
is trying to use it and says that their lightning path is failing. But, so you might wanna check your lightning note. I don't know how you're processing payments.

314
00:23:16.790 --> 00:23:17.850
It's through Alby.

315
00:23:18.550 --> 00:23:20.250
Well, that there's your issue.

316
00:23:20.630 --> 00:23:24.090
Alright. Maybe someone can someone in the community can help me out.

317
00:23:26.554 --> 00:23:35.455
Yeah. We'll figure I mean, you're you're running your own node? Are you self hosting your own node? It's it's AlbiHub. So, technically, yes, but I'm not really doing anything.

318
00:23:35.914 --> 00:23:38.414
Right. But it so it is your own node with your own channels?

319
00:23:38.910 --> 00:23:39.650
Yeah. It is.

320
00:23:40.190 --> 00:23:41.730
Okay. For example, here's here's

321
00:23:42.430 --> 00:23:46.530
a A bigger one? Yeah. Here's here's a call to action.

322
00:23:48.270 --> 00:23:49.810
Either look up his node

323
00:23:50.670 --> 00:23:59.805
or I see Ditto in the comments. Either look up his node or I'm gonna put his node in the show notes. Let's, open a bunch of liquidity to his node so he doesn't have payment failures anymore.

324
00:24:01.065 --> 00:24:04.925
You understand, like, the concept of lightning. Right? Like, you need to have inbound liquidity

325
00:24:05.450 --> 00:24:09.790
and outbound liquidity. Inbound's much harder to get. But if you're if you're a good product or service,

326
00:24:10.330 --> 00:24:12.910
people generally will open inbound to you

327
00:24:13.290 --> 00:24:14.170
just so that

328
00:24:15.930 --> 00:24:19.640
no. I'll be I see decoy. I'll I'll be I'll be I'll be I'll be I'll be I'll be I'll be I'll be I'll be I'll be I'll be I'll be I'll be I'll be I'll be I'll be I

329
00:24:21.255 --> 00:24:24.955
but it's for managing your own node. So at the end of the day, your issue is,

330
00:24:26.375 --> 00:24:37.115
I thought at first, I thought he meant he was using Albi's old custodial wallet, which they've been phasing out. It might be completely phased out. AlbiHub just at the end of the day, you're managing your own lightning node. So

331
00:24:37.470 --> 00:24:42.929
anyone who's managed a lightning node knows it's a mental burden, and inbound liquidity can be a pain in the ass.

332
00:24:43.950 --> 00:24:47.730
And that's clearly where we're at. But, anyway, the good part is in this thing.

333
00:24:48.030 --> 00:24:48.530
Because,

334
00:24:49.565 --> 00:24:58.305
you know, I have to pay the AI cost from the upstream providers in US dollars, but the majority of people using Shakespeare are paying me in Lightning.

335
00:24:58.765 --> 00:24:59.585
And so

336
00:25:00.365 --> 00:25:07.529
all of you people who are using Shakespeare and paying with Lightning, you're converting my USD into Bitcoin for me. So thank you.

337
00:25:07.830 --> 00:25:10.169
That's a great way to get an okay way to see Bitcoin.

338
00:25:10.950 --> 00:25:17.769
There you go. It's a great way to stack sets. I've always said that the way we see merchant adoption happen is actually from

339
00:25:18.285 --> 00:25:20.865
that direction. It's merchants who want Bitcoin

340
00:25:21.165 --> 00:25:33.940
rather than users that want to spend Bitcoin. You kinda can't force it. But if the merchant wants Bitcoin, then it makes a lot more sense. Okay. So let's talk the rise of the the vlog. First of all, do you know how to pronounce

341
00:25:34.799 --> 00:25:36.019
v l o g?

342
00:25:38.000 --> 00:25:39.380
I believe it's vlog.

343
00:25:40.240 --> 00:25:44.500
Yeah. But that doesn't make any sense. It should be video log. Right? So it's vlog.

344
00:25:46.375 --> 00:25:57.995
So I'm like you over this on the Internet? Yeah. Suppose I've I've Sorry. I I learned how the pronunciation happened on my other show yesterday. So I've been I've been getting bullied. Do you know what the b in blog stands for?

345
00:25:59.500 --> 00:26:00.640
I didn't know that

346
00:26:00.940 --> 00:26:02.000
it was an acronym.

347
00:26:02.700 --> 00:26:06.000
The b in blog stands for web. It's a web log.

348
00:26:07.020 --> 00:26:10.800
Oh. So if you go down that route, it's a soft b,

349
00:26:11.420 --> 00:26:12.640
web blog,

350
00:26:13.255 --> 00:26:13.755
and

351
00:26:14.294 --> 00:26:16.155
then video log is a hard v,

352
00:26:16.615 --> 00:26:17.434
so it should

353
00:26:17.975 --> 00:26:18.955
be v log.

354
00:26:19.495 --> 00:26:23.914
Video. Video log. I don't know. I'm wrong. So blog should be blog,

355
00:26:25.174 --> 00:26:25.674
basically.

356
00:26:25.975 --> 00:26:26.875
Web. Blog.

357
00:26:27.575 --> 00:26:28.554
Yeah. Blog.

358
00:26:29.309 --> 00:26:35.490
Blog. Yeah. I don't know. Well, anyway, the rise of the vlogs. Anyway, this is our our live vlog or

359
00:26:36.429 --> 00:26:36.929
vlog.

360
00:26:38.269 --> 00:26:42.690
And then what's the difference between a vlog and TikTok? Isn't TikTok just like the modern vlog?

361
00:26:46.375 --> 00:26:50.075
I I hadn't I had not really thought about that before, but I guess so.

362
00:26:50.695 --> 00:26:54.394
Maybe I mean, I'm sorry. It's you it has to be, like,

363
00:26:55.414 --> 00:26:59.080
you have to be describing something. Right? Like like, it has to be, like, autobiographical.

364
00:26:59.460 --> 00:27:09.240
I don't know. It doesn't matter. I thought you were gonna ask me about Nostra versus Nostra again. And then But we also I mean, we can disagree. We can agree to disagree on that one because you did pronounce it wrong.

365
00:27:10.804 --> 00:27:14.345
Okay. So Shakespeare, what does that look like right now? How does

366
00:27:15.924 --> 00:27:20.424
what what has changed this got released since our last episode. Right? It did.

367
00:27:20.804 --> 00:27:25.705
Yeah. So this is this is the way that people previously, you had to use a terminal to use MKStack.

368
00:27:26.320 --> 00:27:29.620
Now Right. People can just go on the web. You can do it on your phone.

369
00:27:30.240 --> 00:27:34.580
And and it looks like, you know, a chat g b t UI. And you can say,

370
00:27:35.600 --> 00:27:40.420
build me a site like Twitter on Noister, or build me a site like Instagram on Noister.

371
00:27:40.960 --> 00:27:41.460
Or

372
00:27:42.355 --> 00:27:44.215
ideally, you'd say something like

373
00:27:44.755 --> 00:27:46.054
build me a,

374
00:27:47.075 --> 00:27:48.215
Gnostr community

375
00:27:48.755 --> 00:27:49.255
for

376
00:27:49.635 --> 00:27:50.455
my Pokemon,

377
00:27:52.434 --> 00:27:52.934
group,

378
00:27:53.235 --> 00:27:56.850
my Pokemon club, and then add a Pokedex to it.

379
00:27:57.570 --> 00:28:00.789
Or, like, build me a nosier commute community for farmers

380
00:28:01.410 --> 00:28:02.549
and add a marketplace

381
00:28:02.850 --> 00:28:05.350
to be able to buy and sell tractors.

382
00:28:06.370 --> 00:28:07.350
And so, like,

383
00:28:07.809 --> 00:28:11.225
it's just gonna allow us to create these extremely personalized,

384
00:28:11.605 --> 00:28:13.705
customized experiences with AI.

385
00:28:14.805 --> 00:28:16.345
And it's kinda just waiting

386
00:28:16.965 --> 00:28:17.465
for

387
00:28:18.245 --> 00:28:18.745
artists

388
00:28:19.685 --> 00:28:20.185
and,

389
00:28:20.565 --> 00:28:22.825
you know, product people to come in

390
00:28:23.330 --> 00:28:23.830
and

391
00:28:24.370 --> 00:28:28.310
just write the the right terms to make Noister win.

392
00:28:29.970 --> 00:28:31.110
Oh, this is awesome.

393
00:28:32.770 --> 00:28:40.005
So, like, everything can just happen right in browser. Like, I don't need to download anything, or I just pay you and just in browser, I just start vibing?

394
00:28:40.545 --> 00:28:41.045
Yep.

395
00:28:42.865 --> 00:28:45.685
And so then what's going on in the back end?

396
00:28:46.385 --> 00:28:52.200
So right now, I I have a server rack in my living room, and that's where all of the files are stored.

397
00:28:53.460 --> 00:28:59.320
And so and and that thing is talking to the AI behind the scenes. It's just connecting to to,

398
00:29:00.740 --> 00:29:03.160
you know, some upstream AI provider.

399
00:29:03.885 --> 00:29:09.425
I've been switching it between Claude and GLM and, you know, testing different providers. But

400
00:29:10.285 --> 00:29:11.505
so your your,

401
00:29:12.365 --> 00:29:15.505
browser is going over Nostra relays to communicate

402
00:29:15.805 --> 00:29:16.625
with this

403
00:29:17.190 --> 00:29:20.490
service provider that is then doing all the AI stuff and managing your files.

404
00:29:22.470 --> 00:29:29.145
But I've been this is act one, and I'm currently developing act two, which changes the architecture so that the so that

405
00:29:29.705 --> 00:29:34.925
the files will actually be in your browser instead of in my server. I'm, like, working on decloudifying

406
00:29:35.305 --> 00:29:37.405
it right now, if that makes sense.

407
00:29:38.985 --> 00:29:39.805
Got it.

408
00:29:40.265 --> 00:29:42.125
So what's my NSP?

409
00:29:43.639 --> 00:29:45.340
My master service provider?

410
00:29:45.880 --> 00:29:47.179
Yes. I'm

411
00:29:47.480 --> 00:29:47.980
logged

412
00:29:48.840 --> 00:29:50.299
I'm logged in right now.

413
00:29:51.240 --> 00:29:51.740
Cool.

414
00:29:52.200 --> 00:29:58.539
So now you should be able to click on the NSP drop down and add funds to it, and then you should be able to prompt it.

415
00:29:59.855 --> 00:30:06.115
Just blindly authorizing things for my extension while I'm live on the show is probably not the best idea.

416
00:30:07.615 --> 00:30:08.355
And it's,

417
00:30:09.135 --> 00:30:13.395
it'll just start running once you give it a prompt, and it usually takes about ten minutes,

418
00:30:14.190 --> 00:30:17.010
maybe fifteen, depending on the complexity of the prompts

419
00:30:17.549 --> 00:30:23.169
to finish this because it's kind of like you're just telling a developer to do something, and it's the fastest possible developer,

420
00:30:23.630 --> 00:30:25.010
you know, known to humankind.

421
00:30:25.645 --> 00:30:28.385
Okay. So I see in my NSPs,

422
00:30:28.845 --> 00:30:29.345
there's

423
00:30:29.965 --> 00:30:31.505
Gemini Flash Playground,

424
00:30:32.445 --> 00:30:33.424
there's MKStack,

425
00:30:34.445 --> 00:30:37.184
there's Soapbox Playground, and there's MKStack

426
00:30:37.485 --> 00:30:37.985
dev.

427
00:30:38.780 --> 00:30:39.679
So if you

428
00:30:39.980 --> 00:30:45.679
the Gemini Flash Playground is a free one that you can use to test things, but the experience is

429
00:30:46.059 --> 00:30:48.240
what you would expect from something that's free.

430
00:30:48.620 --> 00:30:53.385
Okay. So the MKStack is the best one to use? Yes. And that one requires

431
00:30:54.085 --> 00:30:54.825
you to

432
00:30:55.365 --> 00:30:57.625
add funds to it before you can use it.

433
00:30:58.645 --> 00:31:01.065
Open payment page. It brings me to Stripe.

434
00:31:01.684 --> 00:31:05.785
There's a selector box there, which you can click to do lightning.

435
00:31:07.130 --> 00:31:10.350
And I've been thinking about switching the order code tabs.

436
00:31:10.730 --> 00:31:19.390
I see. So the the Bitcoin one is actually the default tab. It'd be, like, the lamest thing ever if I paid with credit card. Continue to pay. Well,

437
00:31:19.765 --> 00:31:20.825
yeah. I've enabled,

438
00:31:21.845 --> 00:31:22.345
Fiat

439
00:31:22.645 --> 00:31:24.105
over Nostr, unfortunately

440
00:31:24.485 --> 00:31:25.225
or fortunately.

441
00:31:25.765 --> 00:31:31.365
Let's see. You got I mean, look, at the end of the day, you gotta accept Fiat. I like the idea of,

442
00:31:31.765 --> 00:31:32.825
Kinda my thinking.

443
00:31:33.370 --> 00:31:38.590
You should I mean, first of all, prioritize the Bitcoin side, but give people a discount if they pay in Bitcoin.

444
00:31:39.610 --> 00:31:40.429
I like that.

445
00:31:42.169 --> 00:31:44.809
Alright. You don't have liquidity. Right? So I can't,

446
00:31:45.530 --> 00:31:48.725
someone says they're were Ditto says they're working on it. And,

447
00:31:49.985 --> 00:31:51.684
that's you. Right? Is that your team?

448
00:31:52.145 --> 00:31:53.205
Yeah. It is.

449
00:31:54.065 --> 00:31:58.164
Or you just in the live chat typing? I guess that's your team. Right? That's my team.

450
00:31:58.625 --> 00:32:01.924
Okay. I can't pay for it. But, so so presumably

451
00:32:02.465 --> 00:32:02.625
in

452
00:32:04.180 --> 00:32:06.200
so I can just so so what

453
00:32:06.820 --> 00:32:07.320
what

454
00:32:08.100 --> 00:32:14.520
you're connecting to some kind of API provider for the actual LLM on the back end. Right? And so is that Yes. Exactly.

455
00:32:15.460 --> 00:32:16.840
So it's configurable.

456
00:32:17.380 --> 00:32:18.680
Something that's really

457
00:32:19.384 --> 00:32:24.284
cool is that most of the AI providers now have just standardized on OpenAI's

458
00:32:24.585 --> 00:32:25.085
API.

459
00:32:25.544 --> 00:32:29.404
So they're you can just drop and replace them. And so,

460
00:32:29.865 --> 00:32:34.520
like, two days ago, it was clogged. Right now, it's GLM. It's z a

461
00:32:35.900 --> 00:32:36.400
I.

462
00:32:37.780 --> 00:32:41.640
Make me a Nasr app that just

463
00:32:42.500 --> 00:32:43.000
shows

464
00:32:43.380 --> 00:32:43.880
images.

465
00:32:44.660 --> 00:32:53.835
Do I have to be, like, more descriptive than that? Nope. Because Let's see what happens. What you'll see is it it actually knows to read the nips. And the first thing it'll do is it will

466
00:32:54.295 --> 00:32:59.275
check the nips to see if there are any that could pertain to your request.

467
00:33:00.080 --> 00:33:03.780
And then if and then if there are, then it it will read those specific nips.

468
00:33:04.480 --> 00:33:04.980
Interesting.

469
00:33:05.520 --> 00:33:07.940
And you mean, we should be able to see it doing it.

470
00:33:08.480 --> 00:33:10.980
Yeah. I'm watching it. Well, I'm in the free one because

471
00:33:11.360 --> 00:33:13.060
you wanna take my money right now.

472
00:33:13.475 --> 00:33:14.295
That's okay. It

473
00:33:14.675 --> 00:33:17.575
should mostly But, yeah, it's gone. Well, let's see if,

474
00:33:20.275 --> 00:33:22.215
replace screen sharing window.

475
00:33:28.550 --> 00:33:32.490
Okay. So the first thing it did was it checked the project structure, right, at the very top?

476
00:33:34.470 --> 00:33:34.970
Okay.

477
00:33:37.030 --> 00:33:39.850
Yep. And so now it's looking at the package of JSON.

478
00:33:43.355 --> 00:33:44.975
Dude, this is pretty fucking awesome.

479
00:33:46.715 --> 00:33:50.495
This one? Alright. I guess it decided not to read the nips. The thing about

480
00:33:51.115 --> 00:33:54.495
Flash, Gemini Flash, is that it has been

481
00:33:55.340 --> 00:33:58.720
optimized to do things quickly and to ignore instructions.

482
00:34:00.700 --> 00:34:01.840
Well, that's not good.

483
00:34:02.140 --> 00:34:02.640
Yeah.

484
00:34:03.500 --> 00:34:05.820
Why does this say $999?

485
00:34:05.820 --> 00:34:08.640
Is that just the way you hacked it to be free? Yes.

486
00:34:09.245 --> 00:34:11.105
And does that have a way to work with it?

487
00:34:12.285 --> 00:34:15.265
Oh, okay. Now it's checking the notes. It is reading the notes. Good. Alright.

488
00:34:16.125 --> 00:34:18.945
Is and so this is saying it's 21%

489
00:34:19.005 --> 00:34:20.705
along the progress right now?

490
00:34:21.410 --> 00:34:24.230
That's the context window. And so each

491
00:34:25.090 --> 00:34:26.950
AI model has a limit

492
00:34:27.410 --> 00:34:27.910
to

493
00:34:28.690 --> 00:34:29.190
the

494
00:34:29.650 --> 00:34:32.790
size of the chat history before it can't

495
00:34:33.665 --> 00:34:35.605
continue in that chat anymore.

496
00:34:36.465 --> 00:34:44.565
And so once that hits a it's kind of like a reverse health bar I've heard it described as to where once it hits 100, then, like, this robot must die.

497
00:34:45.265 --> 00:34:50.160
Got it. So it's actually not a progress bar. It is just telling me Hopefully,

498
00:34:50.780 --> 00:34:53.680
it doesn't hit a 100% because if it does, then it's

499
00:34:54.220 --> 00:34:55.680
you're gonna have to start over.

500
00:34:57.020 --> 00:34:59.600
Well, I gave it, like, the simplest fucking prompt ever.

501
00:34:59.900 --> 00:35:00.400
Yep.

502
00:35:01.580 --> 00:35:03.600
So which which NIPS did it read?

503
00:35:05.474 --> 00:35:06.775
NIPS 68.

504
00:35:07.954 --> 00:35:10.775
And then you should be able to click and see. Your first feeds.

505
00:35:11.234 --> 00:35:13.815
Okay. Great. So it's gonna be it's gonna be like, like, Instagram.

506
00:35:16.194 --> 00:35:17.734
View full NIPS document.

507
00:35:18.540 --> 00:35:19.580
Yeah. That's the,

508
00:35:21.100 --> 00:35:22.800
I forget. What's the name of Pablo's?

509
00:35:23.980 --> 00:35:24.480
Olas.

510
00:35:24.860 --> 00:35:27.360
Yeah. Olas. It should be Olas compatible.

511
00:35:27.820 --> 00:35:29.680
Olas is NIP 68?

512
00:35:30.300 --> 00:35:31.360
I believe so.

513
00:35:34.694 --> 00:35:35.755
This is pretty cool.

514
00:35:36.135 --> 00:35:37.595
Yeah. And so

515
00:35:38.295 --> 00:35:39.755
this could take ten

516
00:35:40.055 --> 00:35:43.890
to fifteen minutes probably to finish, and then once it does, it'll pop up

517
00:35:45.410 --> 00:35:47.910
Okay. Your team just messaged me in

518
00:35:48.529 --> 00:35:49.910
the in the chat.

519
00:35:50.529 --> 00:35:54.549
They just gave me credits because they're trying to figure out the lightning node. Oh, awesome.

520
00:35:56.130 --> 00:35:59.190
It's not loading now. I had this happen to me before.

521
00:35:59.655 --> 00:36:00.955
Okay. Let's stop

522
00:36:01.855 --> 00:36:02.355
screen.

523
00:36:03.255 --> 00:36:04.475
Stop screen sharing.

524
00:36:06.295 --> 00:36:07.595
Have you seen that error?

525
00:36:08.535 --> 00:36:11.035
Where it goes white? When you refresh?

526
00:36:16.650 --> 00:36:18.750
It's working it's working on my end.

527
00:36:19.050 --> 00:36:22.910
Well, I'm gonna show well, it's only when I it happened to me when I clicked refresh.

528
00:36:23.930 --> 00:36:25.070
Yeah. I

529
00:36:25.610 --> 00:36:30.405
I'm gonna show Just, like, login with extension is so easy. So I I wanted to

530
00:36:31.125 --> 00:36:32.105
okay. I'm back.

531
00:36:32.725 --> 00:36:34.425
I'm back. MK stack.

532
00:36:35.205 --> 00:36:40.905
Nice. I got a $100. $100 in there. Great. Okay. What should my prompt be? Build me

533
00:36:41.205 --> 00:36:43.465
an do I have to say a Nostra web app?

534
00:36:45.340 --> 00:36:48.720
No. Because it already knows that. So It's always a web app. Right?

535
00:36:49.260 --> 00:36:49.760
Yeah.

536
00:36:50.060 --> 00:36:51.040
It doesn't hurt.

537
00:36:52.380 --> 00:36:53.119
That shows

538
00:36:55.500 --> 00:36:56.320
full screen

539
00:36:57.414 --> 00:36:57.914
images

540
00:36:59.414 --> 00:37:01.595
as they are published

541
00:37:06.055 --> 00:37:07.755
only from people

542
00:37:08.694 --> 00:37:09.434
I follow.

543
00:37:11.510 --> 00:37:12.329
Does that work?

544
00:37:12.790 --> 00:37:13.290
Yeah.

545
00:37:15.510 --> 00:37:16.010
Okay.

546
00:37:18.790 --> 00:37:25.130
Send vibe. I've been I've been doing a lot weirder stuff, like, you know, build me a Nostril site where I can trade Pokemon

547
00:37:26.535 --> 00:37:32.075
and, like, build me a Nostril badges site and things like that. Do people still use

548
00:37:32.535 --> 00:37:33.595
Nostril badges?

549
00:37:34.214 --> 00:37:35.275
Is that the thing?

550
00:37:35.974 --> 00:37:41.115
I really like the idea of it, but I think that it's one of those things that's, like, well, it's kind of not

551
00:37:42.030 --> 00:37:42.770
a priority

552
00:37:43.550 --> 00:37:45.410
for a lot of projects. But

553
00:37:46.030 --> 00:37:49.810
if you can just vibe code it in five minutes, then why not?

554
00:37:51.390 --> 00:37:51.890
Right.

555
00:37:52.670 --> 00:37:55.170
Fair enough. I mean, I like the concept of badges.

556
00:37:55.635 --> 00:37:58.135
So do you think that, like, people are just gonna

557
00:37:58.755 --> 00:38:01.735
vibe their own apps and that's how they're gonna interact with Nostr?

558
00:38:02.835 --> 00:38:04.215
I do. I mean,

559
00:38:04.835 --> 00:38:08.695
I I kinda feel like Nostr is at this turning point right now,

560
00:38:13.970 --> 00:38:20.710
and I I think that, like, people just aren't buying what we're selling. You know? Like, the world just isn't

561
00:38:21.970 --> 00:38:23.430
the world doesn't want it.

562
00:38:23.730 --> 00:38:29.395
I want it. We want it. Our community wants it. But in terms of mainstream adoption,

563
00:38:30.255 --> 00:38:32.355
like, something's gonna have to happen.

564
00:38:33.135 --> 00:38:33.635
And

565
00:38:34.175 --> 00:38:47.550
I'm hopeful that if we just wait long enough, then some major world event would occur that would cause people to flood to to it because of, like, you know, some big censorship issue. It's kinda happening right now, right, with, like,

566
00:38:49.210 --> 00:38:52.349
the Online Safety Act or whatever out of The UK? Yeah.

567
00:38:53.445 --> 00:38:55.305
But it just feels like we've kind of

568
00:38:55.605 --> 00:38:57.705
passed that point. Like,

569
00:38:58.645 --> 00:38:59.785
we've already had

570
00:39:00.485 --> 00:39:04.265
these major events that have happened, and we've already had mass migrations.

571
00:39:05.060 --> 00:39:06.119
And I'm not sure

572
00:39:07.700 --> 00:39:15.240
that the same type of thing is gonna happen in the same way again. And and so I've been thinking about, okay, maybe I need to change my

573
00:39:15.940 --> 00:39:17.240
point of view here.

574
00:39:18.765 --> 00:39:22.625
Instead of being this sort of, like, war hero type of person,

575
00:39:24.444 --> 00:39:30.145
that I've perceived myself as being before where I'm I'm, like, going in and and it's this big political battle.

576
00:39:30.630 --> 00:39:34.089
Maybe I need to position myself more as a product

577
00:39:34.630 --> 00:39:39.210
person who is building something that is actually interesting that people

578
00:39:39.589 --> 00:39:41.130
wanna use in and of itself.

579
00:39:42.549 --> 00:39:43.289
And then,

580
00:39:44.645 --> 00:39:49.545
like and the great thing is that this only really works because of Noister. So

581
00:39:50.005 --> 00:39:50.505
Right.

582
00:39:51.445 --> 00:39:54.005
It's not like Noister is just a side effect, although it

583
00:39:54.805 --> 00:39:56.105
in some ways, it is.

584
00:39:56.565 --> 00:39:57.065
But

585
00:39:58.710 --> 00:40:05.210
in other solutions, people are struggling right now to get sign in to work on their Vibe coded apps

586
00:40:05.750 --> 00:40:08.810
because because they need, like, a Twitter API and shit. Right?

587
00:40:09.670 --> 00:40:10.170
Right.

588
00:40:10.645 --> 00:40:14.745
They like, or like email. Right? Like, you need to you need a back end, basically.

589
00:40:15.685 --> 00:40:17.705
And that's something that AI

590
00:40:18.005 --> 00:40:24.905
tools are struggle to build is a full stack application with the back ends. And the people who are building those things, I feel

591
00:40:25.285 --> 00:40:26.730
they are fighting

592
00:40:27.350 --> 00:40:27.850
hard

593
00:40:28.710 --> 00:40:30.250
against a strong current.

594
00:40:30.870 --> 00:40:33.770
And it's, like, super insecure. Right? Like, they had, like,

595
00:40:34.310 --> 00:40:42.085
that app where, like, the women were raiding the men or whatever, and then, like, all the KYC docs were just in in, like, an open Firebase

596
00:40:43.025 --> 00:40:44.724
folder or whatever. A great example.

597
00:40:45.665 --> 00:40:49.444
But with NoSir, like, NoSir flips the authentication, and it flips,

598
00:40:49.825 --> 00:40:51.684
like, where the protected resource

599
00:40:52.464 --> 00:40:53.365
actually is.

600
00:40:53.665 --> 00:40:54.964
So so authorization

601
00:40:56.600 --> 00:40:59.100
happens from you as the user rather

602
00:40:59.560 --> 00:41:01.260
than, like, from the server.

603
00:41:01.880 --> 00:41:03.580
And the servers are

604
00:41:04.200 --> 00:41:04.700
permissionless.

605
00:41:05.160 --> 00:41:05.980
They're relays,

606
00:41:07.080 --> 00:41:09.900
as long as you can identify yourself using your key.

607
00:41:10.755 --> 00:41:12.855
And so because of the design of Noester,

608
00:41:13.474 --> 00:41:19.974
it allows us to build Vibe coded apps where sign in just works the first time.

609
00:41:20.994 --> 00:41:21.734
And it's

610
00:41:22.115 --> 00:41:23.974
not very hard. It's easy.

611
00:41:24.914 --> 00:41:25.154
So

612
00:41:25.930 --> 00:41:29.390
And then you get the social graph. You don't need to get any APIs.

613
00:41:30.330 --> 00:41:36.350
Exactly. I mean, it's really I and then you mix Bitcoin into that puzzle, and then all of a sudden you have payments.

614
00:41:36.890 --> 00:41:39.870
Yep. So you don't have to worry about integrating Stripe or whatever

615
00:41:40.625 --> 00:41:42.244
and everything that that entails.

616
00:41:42.785 --> 00:41:45.525
So it just makes so, like, maybe that is

617
00:41:45.905 --> 00:41:47.365
the biggest selling point

618
00:41:47.984 --> 00:41:49.845
of Nasr. Maybe that is the

619
00:41:50.145 --> 00:41:55.685
the actual Lindy selling point of Nasr and not censorship resistant. The censorship resistant part is just

620
00:41:56.040 --> 00:41:59.020
a side effect. It's just being able to vibe without permission.

621
00:41:59.880 --> 00:42:03.820
That's kind of what I've been thinking is is, like Oh, I think it's done.

622
00:42:04.840 --> 00:42:05.740
Sorry. Continue.

623
00:42:06.440 --> 00:42:08.460
This is the future of technology.

624
00:42:10.405 --> 00:42:14.184
And so if we can get in on on that,

625
00:42:15.684 --> 00:42:16.184
then,

626
00:42:16.805 --> 00:42:18.105
you know, it's gonna matter.

627
00:42:20.484 --> 00:42:22.585
Is this my app on the left on the right?

628
00:42:27.250 --> 00:42:27.990
Yes. I

629
00:42:28.770 --> 00:42:33.430
did. No. It's two x isn't gonna show my private key up here. Right? Oh, you can't even see it. Nice.

630
00:42:35.890 --> 00:42:37.910
Let's see. Oh, okay. Let's see if it worked.

631
00:42:38.835 --> 00:42:42.695
Oh, alright. I was a little bit worried at first. That's pretty cool.

632
00:42:48.115 --> 00:42:48.855
That's dope.

633
00:42:50.755 --> 00:42:51.255
Damn.

634
00:42:52.180 --> 00:42:54.360
You made it a lot easier than it was before.

635
00:42:56.980 --> 00:42:57.480
Whew.

636
00:42:58.020 --> 00:42:59.080
Shout out to you, Glenn.

637
00:42:59.940 --> 00:43:04.420
On the live demo. So that only cost me $2.26.

638
00:43:04.420 --> 00:43:07.160
Well, technically, it cost your team $2.26.

639
00:43:07.775 --> 00:43:08.895
I'll I'll I'll,

640
00:43:09.695 --> 00:43:10.835
reimburse you guys.

641
00:43:11.455 --> 00:43:12.435
That's pretty crazy.

642
00:43:12.975 --> 00:43:15.155
I think you probably sent me more Zaps,

643
00:43:15.535 --> 00:43:16.915
you know, over the years.

644
00:43:17.615 --> 00:43:20.370
Well, we'll keep we'll keep the balance correct. I'll make sure.

645
00:43:21.410 --> 00:43:23.510
So so okay. So I have this.

646
00:43:24.450 --> 00:43:24.950
Is

647
00:43:25.970 --> 00:43:27.750
this just a regular web link?

648
00:43:28.690 --> 00:43:31.670
Yes. It is. So I could share that link

649
00:43:32.525 --> 00:43:34.705
to this new web app that I just created.

650
00:43:35.085 --> 00:43:37.345
I can share that with, like, friends and family, whoever.

651
00:43:37.725 --> 00:43:40.305
I could post it, and then anyone could use this app.

652
00:43:40.685 --> 00:43:41.185
Exactly.

653
00:43:42.765 --> 00:43:44.145
Now how does that work

654
00:43:45.750 --> 00:43:48.730
in terms of your hosting cost? Are you just eating those right now?

655
00:43:49.110 --> 00:43:49.430
Or

656
00:43:49.990 --> 00:43:53.130
Yes. But because it is a static,

657
00:43:54.230 --> 00:43:54.730
website,

658
00:43:56.070 --> 00:43:56.570
it

659
00:43:57.110 --> 00:44:04.755
it has not really cost me anything in terms of of CPU. It's only costing me in term or RAM, really. It's only costing me in terms of disk.

660
00:44:05.214 --> 00:44:05.714
So,

661
00:44:07.055 --> 00:44:08.335
you know, I need to have

662
00:44:09.535 --> 00:44:11.395
but I'm running this all on my

663
00:44:12.174 --> 00:44:14.355
on my server rack in my living room.

664
00:44:14.770 --> 00:44:15.590
So, like,

665
00:44:16.210 --> 00:44:25.910
if I run out of disk space, I don't have recurring monthly cost for disk. I just need to buy have a one time cost of buying a new disk and putting it in there.

666
00:44:26.530 --> 00:44:27.270
Got it.

667
00:44:27.705 --> 00:44:28.205
So

668
00:44:28.985 --> 00:44:30.685
the cost is not great.

669
00:44:31.865 --> 00:44:32.365
Probably

670
00:44:33.145 --> 00:44:34.125
more than anything,

671
00:44:34.585 --> 00:44:35.485
there's a

672
00:44:36.345 --> 00:44:56.240
sort of ongoing maintenance burden of this thing of just ensuring that it's constantly running and working correctly. It's like your Trump bot. It's like you close your laptop, and then we don't get Trump Exactly. Trump post. And so this for that reason, that's why the new version of Shakespeare I'm developing tries to move as much of the application logic as possible

673
00:44:56.724 --> 00:44:58.105
to run within

674
00:44:58.885 --> 00:45:02.345
your browser instead of having it run on my server.

675
00:45:02.964 --> 00:45:03.785
Makes sense.

676
00:45:04.484 --> 00:45:04.984
Yeah.

677
00:45:05.444 --> 00:45:09.704
So, like, it's really running you're trying to offload as much of it locally in browser.

678
00:45:10.150 --> 00:45:11.690
Exactly. And the the

679
00:45:12.150 --> 00:45:15.930
the web page is just serving me that local code or whatever.

680
00:45:16.310 --> 00:45:16.810
Yep.

681
00:45:18.310 --> 00:45:23.590
And and something I'm really excited about of the sort of But you could also charge for that. Right? Like

682
00:45:24.825 --> 00:45:25.325
Yeah.

683
00:45:25.865 --> 00:45:35.165
And, like, I would happily pay to have it. I mean, because you it's an interesting little problem. It's an interesting problem, right, because you add so much friction if you were making me

684
00:45:35.625 --> 00:45:39.484
actually go and set up hosting myself to host the web app.

685
00:45:40.020 --> 00:45:48.280
Even though if if it is any good, like, eventually, someone would. But, like, to just have it running automatically is fucking great. I think people would pay for that privilege.

686
00:45:48.900 --> 00:45:50.360
It's true. It's a good point.

687
00:45:53.335 --> 00:45:55.355
You can make a decent margin on it too.

688
00:45:56.375 --> 00:46:08.640
Yeah. We've been we've been kind of struggling with the with the finances of it because of the fact that Claude is so expensive, and we're hoping that with by changing the model to something that's a little bit more efficient, it's gonna allow us to

689
00:46:09.580 --> 00:46:11.520
not be negative on this, basically.

690
00:46:11.980 --> 00:46:15.920
But that's what I'm saying. Like Yeah. The margins are tight on the actual

691
00:46:16.460 --> 00:46:18.000
Right. LLM side.

692
00:46:19.180 --> 00:46:20.160
But on the,

693
00:46:20.765 --> 00:46:26.465
like, the web hosting side, the, like, the short link that you a good idea. Is actually probably relatively cheap.

694
00:46:26.925 --> 00:46:28.065
But to the user,

695
00:46:28.845 --> 00:46:29.905
it's quite valuable

696
00:46:30.605 --> 00:46:34.785
because it's just awesome to just immediately have a URL you can point people to.

697
00:46:36.400 --> 00:46:38.500
I think you have a point. Yeah.

698
00:46:39.360 --> 00:46:40.420
Business on air.

699
00:46:40.880 --> 00:46:41.400
Thank you, man.

700
00:46:41.840 --> 00:46:42.340
Update.

701
00:46:43.360 --> 00:46:44.500
Financial advice.

702
00:46:45.120 --> 00:46:47.380
I'm trying to have it update my app now.

703
00:46:48.055 --> 00:46:49.915
I want four images in it.

704
00:46:50.215 --> 00:46:52.315
I should have done something more crazy.

705
00:46:52.775 --> 00:46:55.515
Like, what I just have I have such a simple mind.

706
00:46:57.095 --> 00:47:02.235
Okay. Here's the app. Someone asked me to You can always just make new prompts and new,

707
00:47:02.619 --> 00:47:03.839
you know, new tabs.

708
00:47:04.540 --> 00:47:05.040
And,

709
00:47:05.820 --> 00:47:06.480
you know,

710
00:47:08.380 --> 00:47:16.395
it's good to just to challenge the AI to see if there's things that you think that maybe it can't do and just to see see if it can do it. You know?

711
00:47:17.115 --> 00:47:18.815
You know, I feel like this is

712
00:47:19.115 --> 00:47:20.655
a kind of

713
00:47:22.075 --> 00:47:23.535
I've been, like, a little bit

714
00:47:23.835 --> 00:47:24.335
distracted

715
00:47:24.875 --> 00:47:25.615
this episode

716
00:47:28.315 --> 00:47:30.815
because we had the live chat go down.

717
00:47:31.650 --> 00:47:34.230
But this might be one of the first times

718
00:47:34.770 --> 00:47:39.670
someone's vibe coded an app on a live show. So I apologize with the audio

719
00:47:40.130 --> 00:47:43.829
to the audio listeners, but we might have just made history here.

720
00:47:45.805 --> 00:47:46.545
Hell, yeah.

721
00:47:48.205 --> 00:47:51.425
Oh, okay. Look. I got my new my new app just loaded.

722
00:47:51.885 --> 00:47:53.665
So four images is better.

723
00:47:54.445 --> 00:47:56.625
It was two it was two full screen before.

724
00:47:57.085 --> 00:47:58.065
Now you're addicted.

725
00:47:59.450 --> 00:48:05.390
You know it was the pro move, by the way? On my second time I did the prompt, I said only people I follow because we're live,

726
00:48:06.089 --> 00:48:09.289
and we didn't want any accidental dick pics. I think,

727
00:48:11.049 --> 00:48:13.450
it's a good tool because I've seen some

728
00:48:14.904 --> 00:48:24.444
I was a bit worried because I've seen some some content in these Instagram style feeds that Yeah. It's like what Dorsey what Dorsey made or whatever, which was just like the

729
00:48:24.984 --> 00:48:26.765
Fire hose? Like the schizophrenic

730
00:48:27.145 --> 00:48:31.040
fire hose where it's just, like, constantly loading new images from everybody?

731
00:48:32.700 --> 00:48:35.040
There was a lot of questionable stuff in there.

732
00:48:35.340 --> 00:48:35.840
Yep.

733
00:48:36.620 --> 00:48:39.600
But, yeah, I I saved myself by doing the

734
00:48:39.980 --> 00:48:41.340
I saved us by doing the

735
00:48:42.145 --> 00:48:43.685
let's see. What happens when I click?

736
00:48:43.985 --> 00:48:45.365
You can click full screen.

737
00:48:46.305 --> 00:48:48.245
Wow. That's pretty fucking cool, dude.

738
00:48:48.865 --> 00:48:51.125
Well done. Shout out to you and your team.

739
00:48:53.105 --> 00:48:57.125
Thank you. And shout out to my followers for having good images.

740
00:48:59.560 --> 00:49:01.660
But this this is the thing now. Like,

741
00:49:03.240 --> 00:49:04.220
back in the day,

742
00:49:05.160 --> 00:49:06.780
it was so hard to build

743
00:49:07.160 --> 00:49:07.660
things.

744
00:49:07.960 --> 00:49:15.335
I mean, it kind of still is. Be people haven't really realized this is something that you can do completely yet. I think that we're still early on this. But

745
00:49:15.875 --> 00:49:16.775
now, like,

746
00:49:18.515 --> 00:49:20.615
now ideas matter more than ever.

747
00:49:21.715 --> 00:49:33.800
It used to be execution was all all that matters. Ideas are a dime a dozen. Oh, you have some great brilliant idea. Who cares? It only matters if you implement it. Now it's becoming a thing where implementing it is becoming a given that you can just expect

748
00:49:34.180 --> 00:49:35.320
because an AI

749
00:49:36.005 --> 00:49:37.465
can do it. So

750
00:49:38.085 --> 00:49:38.905
it's really

751
00:49:39.365 --> 00:49:45.065
moving to a world where the people with the best ideas are the ones who can do the most.

752
00:49:48.670 --> 00:49:49.109
I,

753
00:49:49.550 --> 00:49:55.330
I agree with that. And, also, yeah, I mean, I'm full of shit. Like, there's no way I'm the first person to vibe live.

754
00:49:56.350 --> 00:49:56.850
So,

755
00:49:58.030 --> 00:50:05.885
you couldn't let me have it. Ski, I see you. Or Skye, I don't know how to pronounce your name. It's like it's like vlog. It's too confusing.

756
00:50:07.385 --> 00:50:09.325
Freaks, do you have any questions for Alex?

757
00:50:10.265 --> 00:50:16.200
Because I'm a little bit lost here. I'm tired, and I'm now my head's just running on what I'm gonna vibe next.

758
00:50:17.240 --> 00:50:19.660
But if you have any questions, put it in the live chat.

759
00:50:20.200 --> 00:50:21.500
Welcome to the party.

760
00:50:22.520 --> 00:50:23.240
It does,

761
00:50:25.240 --> 00:50:28.380
but then once you actually so, like, on that thread,

762
00:50:29.800 --> 00:50:30.520
once you

763
00:50:32.015 --> 00:50:34.515
you can always get, like, the MVP executed

764
00:50:35.055 --> 00:50:37.715
on an idea. But then once that happens,

765
00:50:38.655 --> 00:50:44.515
you either need to be a developer, like, bring in someone who knows what the fuck they're doing if you actually want, like, to scale it. Right?

766
00:50:46.575 --> 00:50:47.075
So

767
00:50:47.619 --> 00:50:48.359
it depends.

768
00:50:49.540 --> 00:50:50.040
Like,

769
00:50:51.460 --> 00:50:55.480
it it depends on the scope of the project. So, yeah, scale is the right word because

770
00:50:55.780 --> 00:50:56.599
right now,

771
00:50:57.619 --> 00:50:59.800
there are limits to this for sure.

772
00:51:00.545 --> 00:51:01.445
But, like,

773
00:51:02.065 --> 00:51:05.525
the limit is a lot further than you might think.

774
00:51:06.545 --> 00:51:07.525
Like, for example,

775
00:51:08.225 --> 00:51:08.725
BitChat,

776
00:51:09.825 --> 00:51:19.100
I'm pretty sure Jack just completely vibe coded that. I mean, I'm I think there's also a lot of back and forth of talking to the agent and,

777
00:51:19.400 --> 00:51:23.580
like, understanding things about how, you know, Bluetooth mesh networks actually work.

778
00:51:25.160 --> 00:51:25.660
But

779
00:51:27.585 --> 00:51:33.285
if you if you're willing to learn along the way and you're willing to have some technical understanding of what's happening,

780
00:51:33.905 --> 00:51:36.405
I don't think that you necessarily have to touch code.

781
00:51:36.785 --> 00:51:39.365
I mean, the coolest part about BitChat to me was

782
00:51:39.780 --> 00:51:41.160
he vibed a native

783
00:51:41.540 --> 00:51:42.600
iOS app,

784
00:51:43.220 --> 00:51:44.520
and then Cali

785
00:51:44.900 --> 00:51:48.360
vibed a native Yeah. Android app from that.

786
00:51:48.740 --> 00:51:50.760
It's so It's like the cross platform

787
00:51:51.060 --> 00:51:51.560
capability.

788
00:51:51.860 --> 00:51:54.600
I think that is, like, one of the single coolest examples.

789
00:51:57.355 --> 00:51:58.734
I mean, how many times

790
00:51:59.515 --> 00:52:02.815
it's always been such a frustrating thing for me as someone

791
00:52:03.275 --> 00:52:05.694
who has done a lot of work on the education side

792
00:52:06.714 --> 00:52:06.875
is,

793
00:52:07.950 --> 00:52:12.130
apps that are just stuck on a single platform. Like, it's really hard to recommend,

794
00:52:12.910 --> 00:52:13.890
for instance,

795
00:52:14.510 --> 00:52:16.050
if you're, like, use

796
00:52:16.670 --> 00:52:27.415
on if oh, like, first, you have to ask, like, what device you're using. It's It's like, oh, you're on iPhone? Like, use Nasdr or use Domus or whatever. Oh, you're on Android? You have Amethyst available to you. And then

797
00:52:27.715 --> 00:52:36.520
and we see this in the Bitcoin world a million times over. And when you just have, like, one app, when you can just say use Phoenix Wallet, and it doesn't matter if they're

798
00:52:37.300 --> 00:52:40.200
if they're on Android or iOS, it's such an easier recommendation.

799
00:52:41.460 --> 00:52:41.960
Yeah.

800
00:52:43.220 --> 00:52:47.160
So, I mean, it's gonna change it's gonna change everything. It's changing the whole game.

801
00:52:48.285 --> 00:52:49.265
I love it.

802
00:52:50.605 --> 00:53:01.025
I see a question. I saw a question, and I got distraction. Yeah. I was there's there's a few people I wanna shout out, actually. Okay. Go for it. Doing stuff. So so Fran's app

803
00:53:01.450 --> 00:53:12.910
has been building Purple Stack. And Purple Stack is similar to MK Stack, but it's for Android. And that's something that is really exciting too. It's kind of related to what we're talking about with, you know,

804
00:53:13.770 --> 00:53:15.470
Cali being able to Vibe code

805
00:53:16.025 --> 00:53:17.725
an Android app. But

806
00:53:18.265 --> 00:53:30.285
but having all of that context there for no store clients, it just makes it a lot easier to do things on Android. And so, you know, he's kind of followed my methodology that I've set out for for building MKStack for the web and then applied that to Android.

807
00:53:30.770 --> 00:53:48.325
And and there's there's opportunities for other spaces too like iOS or, you know, maybe just something that's a more specific use case that we may wish for it to be customized in various different ways, like a kind one, you know, sort of client stack that then could be customized various different ways. So

808
00:53:48.625 --> 00:53:55.765
what he's doing is really cool. Check out Purple Stack as well to see, you know, other ways that you can build stuff with a similar methodology.

809
00:53:56.785 --> 00:53:57.525
And then,

810
00:53:58.490 --> 00:54:00.109
I saw Hazard earlier.

811
00:54:01.849 --> 00:54:09.069
He he had mentioned something about a service worker thing I could reuse. But in in Shakespeare, in the original version of it,

812
00:54:09.690 --> 00:54:10.430
it's using

813
00:54:10.730 --> 00:54:11.869
it was using Nsight.

814
00:54:12.490 --> 00:54:12.990
And

815
00:54:13.835 --> 00:54:16.255
I'm gonna be moving back to that in the v two

816
00:54:16.634 --> 00:54:17.134
where,

817
00:54:19.355 --> 00:54:20.174
people can

818
00:54:20.634 --> 00:54:24.894
deploy stuff directly from their browser, and then it it gets uploaded to Blossom servers.

819
00:54:25.275 --> 00:54:38.220
And so that means that you can Vibe code like a web app in one of these AI things, and then you can deploy it to Blossom servers. And then a Blossom gateway would be able to serve it from in a censorship resistant way by

820
00:54:38.765 --> 00:54:40.785
scanning across multiple Blossom

821
00:54:41.165 --> 00:54:41.665
servers,

822
00:54:41.965 --> 00:54:45.185
which is, like, one of the that actual hosting aspect.

823
00:54:45.805 --> 00:54:46.305
Yeah.

824
00:54:47.005 --> 00:54:48.785
But what do you do about the actual,

825
00:54:49.565 --> 00:54:50.625
like, web URL?

826
00:54:51.680 --> 00:54:57.380
Yeah. So that would be that would be like a you'd have an NPUB and the subdomain is pretty ugly, but then I've been thinking

827
00:54:57.920 --> 00:55:00.020
about designing some sort of system

828
00:55:00.560 --> 00:55:01.700
where you can

829
00:55:02.720 --> 00:55:09.135
create, like, a a DNS over no sort of thing where you can then map subdomain labels to,

830
00:55:09.595 --> 00:55:11.055
end site end pubs,

831
00:55:11.995 --> 00:55:15.615
and then you'd be able to just you'd be able to support custom domains that way.

832
00:55:18.329 --> 00:55:25.790
But still, like, the proxy goes down, then you can't access the app anyway even though it's all hosted. There are multiple Nsight proxies.

833
00:55:26.490 --> 00:55:31.309
Got it. Gateways gateway is the correct term. And then with the, like, a NNS

834
00:55:32.385 --> 00:55:32.885
thing,

835
00:55:33.265 --> 00:55:34.325
then you could

836
00:55:34.785 --> 00:55:35.845
map to multiple

837
00:55:36.545 --> 00:55:37.605
domains to multiple,

838
00:55:37.985 --> 00:55:38.885
insight gateways.

839
00:55:39.425 --> 00:55:39.925
So

840
00:55:40.305 --> 00:55:45.765
that's really cool of a thing that exists that we're gonna be leaning into a little bit more in the future.

841
00:55:46.230 --> 00:55:46.730
That's

842
00:55:47.110 --> 00:55:50.170
kind of hard to imagine, but seems

843
00:55:50.950 --> 00:55:52.170
like it could be massive.

844
00:55:53.030 --> 00:55:53.530
Yep.

845
00:55:53.910 --> 00:56:01.770
And so we've already been experimenting with that. If you run NPM run deploy in an MK Stack site, it will upload it to an N site,

846
00:56:02.295 --> 00:56:06.395
and you'll be able to just share an end site link. And it's completely permissionless. It's amazing.

847
00:56:08.775 --> 00:56:09.675
That's awesome.

848
00:56:10.295 --> 00:56:14.635
We also need more Blossom servers. Right? Like, we just don't have that money right now.

849
00:56:15.359 --> 00:56:19.220
So I was thinking maybe in terms of a name. Like, a paid subscription

850
00:56:19.680 --> 00:56:22.980
type of model, maybe on a specific Blossom server

851
00:56:23.519 --> 00:56:26.579
is a good place to put that. I mean, that could actually

852
00:56:28.875 --> 00:56:32.095
I mean, I know Blossom was, like, originally envisioned for,

853
00:56:32.875 --> 00:56:34.815
like, video and photo.

854
00:56:35.755 --> 00:56:36.255
But

855
00:56:36.875 --> 00:56:43.535
and, I mean, one of the big issues with video So Yeah. Any file. But, like, the one of the issues with video is, like, videos are so large.

856
00:56:44.220 --> 00:56:46.480
Like, this payload would be significantly smaller.

857
00:56:46.860 --> 00:56:47.360
Absolutely.

858
00:56:47.740 --> 00:56:51.360
Like, how big is that app I created, like, how big is that app?

859
00:56:52.540 --> 00:56:54.800
I would say two megabytes max.

860
00:56:55.260 --> 00:56:58.240
Right. Probably closer to one megabyte, but also

861
00:56:59.625 --> 00:57:01.805
it's split across three or four

862
00:57:02.105 --> 00:57:02.605
files.

863
00:57:03.225 --> 00:57:07.405
So probably each file is, like, 500 kilobytes or less. So,

864
00:57:08.025 --> 00:57:10.605
like, this you know, the size of a photo, size of a JPEG.

865
00:57:11.550 --> 00:57:31.194
Yeah. That's that's pretty fucking cool. And so then it's and the payload's so lightweight because most things are just running inside the browser locally anyway. Yep. Right. And so then from so if we move all the files in the browser, which, by the way, this brings me to I see Dan Conway, dev, is in this chat too. And he's been working on amazing things with decentralized Git,

866
00:57:31.654 --> 00:57:35.515
and that's that's the thing I'm most excited about of the next version of Shakespeare

867
00:57:35.815 --> 00:57:40.770
is the deep Gnoster Git integrations because here's how I imagine the user experience being.

868
00:57:41.570 --> 00:57:51.430
You go on to Shakespeare. You sign in with your Noester account. You Vibe code something, and then you click, like, okay. I'm done. And then it pushes that to a

869
00:57:51.730 --> 00:57:52.950
Noester Git repository

870
00:57:53.330 --> 00:57:53.830
completely

871
00:57:54.130 --> 00:57:58.445
permissionlessly. You don't have to do anything else. All you need is a Noester account

872
00:57:58.745 --> 00:58:01.405
because that's how grasp works is that

873
00:58:02.025 --> 00:58:05.325
you have a Noester account, and then you can immediately push code to get.

874
00:58:05.705 --> 00:58:08.845
And so now other people can collaborate with you on that code.

875
00:58:09.150 --> 00:58:15.410
Now when people go to your VibeCoded website, imagine there's a edit with Shakespeare button on that website.

876
00:58:15.710 --> 00:58:17.730
They click on that button. It opens Shakespeare.

877
00:58:18.030 --> 00:58:19.089
It then clones

878
00:58:19.630 --> 00:58:24.130
that Gnost or Git repository into your browser, and then you can immediately start,

879
00:58:25.025 --> 00:58:33.205
like, writing a patch for it. And then there can be a button that you click again that says contribute this patch back, and then it makes, you know, a Noester

880
00:58:33.825 --> 00:58:34.645
merge request.

881
00:58:35.665 --> 00:58:39.205
And then then, like, the original person can review it and merge it in,

882
00:58:39.690 --> 00:58:45.390
and then deploy it. And so I kinda like, this is potentially a replacement of GitHub,

883
00:58:46.250 --> 00:58:53.150
you know, in a way. It's like it's an but it's starting with the AI rather than starting with the repositories. And and so

884
00:58:53.535 --> 00:58:54.595
between this

885
00:58:55.055 --> 00:58:57.714
and Gnostr Hub and Git workshop,

886
00:58:58.494 --> 00:58:58.994
like,

887
00:58:59.454 --> 00:59:04.835
you you have this this full spectrum being covered of, you know, vibe coding

888
00:59:05.295 --> 00:59:06.035
to traditional

889
00:59:06.494 --> 00:59:06.994
development

890
00:59:07.530 --> 00:59:08.270
and collaboration

891
00:59:08.650 --> 00:59:15.470
over Gnostr because Gnostr supplies the social component and the social layer that is just so powerful.

892
00:59:18.650 --> 00:59:20.750
And then you have payments built in too.

893
00:59:21.365 --> 00:59:21.865
Exactly.

894
00:59:23.125 --> 00:59:28.424
And and we have we have Lightning integration directly into MKStack, thanks to Chad.

895
00:59:28.805 --> 00:59:29.625
And so

896
00:59:30.405 --> 00:59:35.704
now when you just if you say, like, add Lightning to this, then it's gonna be able to just work.

897
00:59:36.380 --> 00:59:37.680
Well, what does that add?

898
00:59:38.940 --> 00:59:48.560
So it it has a, like, additions to the context that explains to the AI how to implement these features, and it also has files that are baked into the stack

899
00:59:48.915 --> 00:59:49.415
that

900
00:59:49.875 --> 00:59:51.495
just make it really easy

901
00:59:52.995 --> 01:00:01.315
to have, like, a Zap button. And so then if you you can just click a Zap button, and it'll open like a modal window that'll say it just looks like primal Zap,

902
01:00:02.289 --> 01:00:05.829
window where you can just, you know, zap an amount on a particular event.

903
01:00:06.930 --> 01:00:09.490
Oh, that's awesome. Oh, yeah. So I could have just been, like, add

904
01:00:10.130 --> 01:00:12.789
do I say add lightning or do I say add zaps?

905
01:00:13.569 --> 01:00:17.865
I think either should work. I'll play around with it. I'm making a new app right now.

906
01:00:19.125 --> 01:00:22.265
And this one's a more difficult prompt. I don't know if it's gonna work or not.

907
01:00:23.444 --> 01:00:27.545
Okay. This is awesome. I feel like we're at, like, the precipice here or something

908
01:00:28.480 --> 01:00:31.060
that could be just a complete game changer.

909
01:00:33.040 --> 01:00:35.460
I feel like my mind has been blown a little bit.

910
01:00:36.880 --> 01:00:40.340
It feels like it's kinda all coming together. Good. It'd be like

911
01:00:41.085 --> 01:00:42.625
open source, open protocol,

912
01:00:42.925 --> 01:00:43.425
renaissance.

913
01:00:44.205 --> 01:00:48.224
Yeah. Exactly. Renaissance is the word I've been thinking of a lot lately.

914
01:00:49.085 --> 01:00:50.865
Especially when you add the git piece

915
01:00:51.325 --> 01:01:01.109
because then everyone can collaborate with each other. Yeah. Like like, everyone who's not a developer can just be like, oh, I don't like this of this particular client. Like, edit with Shakespeare. Boom.

916
01:01:01.570 --> 01:01:02.050
And

917
01:01:02.530 --> 01:01:05.910
yeah. It's gonna be kind of annoying to the to the actual developers

918
01:01:06.494 --> 01:01:20.275
to a little bit. But But also, it'll be so easy to fork and run your own because you can just ignore it if they want to. And then you'll have your own custom URL. And and Hazard says c name resolution is already a thing in some end site gateway. So we're already doing it. Amazing.

919
01:01:20.600 --> 01:01:23.660
Thank you. I'm glad someone else is doing it so I don't have to do it.

920
01:01:24.200 --> 01:01:25.020
So that,

921
01:01:29.080 --> 01:01:35.900
yes. Yeah. I agree with none of your business. Nobody can come close to figuring out the implications of what was just said. My head hurts already.

922
01:01:37.535 --> 01:01:40.595
And I feel like it, like, really flips the script a bit

923
01:01:41.055 --> 01:01:43.395
in terms of it. It puts a lot of

924
01:01:44.335 --> 01:01:46.355
advantage to open source over

925
01:01:46.895 --> 01:01:51.555
closed stuff, to go all the way back to our earlier conversation about closed proprietary shit.

926
01:01:54.740 --> 01:01:58.280
I mean, you could just have, like, millions of people working off of each other

927
01:01:59.220 --> 01:02:01.320
and just shipping shit live

928
01:02:01.860 --> 01:02:07.560
with very little security implications because you're not handling private key material or anything sensitive.

929
01:02:08.954 --> 01:02:11.855
No API calls. You don't have to deal with any of that shit.

930
01:02:12.875 --> 01:02:14.414
You don't have to deal with hosting.

931
01:02:15.434 --> 01:02:17.855
You just reckon Vibe and just see what happens.

932
01:02:18.555 --> 01:02:23.454
I think this is the way Nostra wins, and so we've gotta embrace this this path hard.

933
01:02:24.140 --> 01:02:26.400
I think you're right. It's either that or v logs.

934
01:02:27.660 --> 01:02:28.960
It's one or the other.

935
01:02:29.660 --> 01:02:30.800
We'll see what happens.

936
01:02:31.260 --> 01:02:32.480
Okay. This is awesome.

937
01:02:33.500 --> 01:02:34.220
I'm just gonna

938
01:02:34.940 --> 01:02:37.040
I'm I now I'll actually,

939
01:02:37.935 --> 01:02:46.835
I'm guilty of I tried to do it the other way, and I hit, like, a single piece of friction. And then, like, my kid was crying or some shit, and I just never I had never continued.

940
01:02:47.135 --> 01:02:48.835
But this is, like, super addicting

941
01:02:49.215 --> 01:02:51.475
that I could just do it in the fucking browser window.

942
01:02:51.980 --> 01:02:53.839
I'm just gonna spend so much money

943
01:02:54.940 --> 01:02:55.260
on,

944
01:02:55.819 --> 01:02:57.200
once you get your lightning

945
01:02:58.779 --> 01:03:06.915
lightning up. I feel like you're the greatest person to be doing it too. I'm really excited to see what you build. Well, I don't I'm not sure about that, but we'll see what happens.

946
01:03:07.855 --> 01:03:08.355
I

947
01:03:10.175 --> 01:03:12.835
well, while I have you here, I'm trying to

948
01:03:13.215 --> 01:03:14.755
this is, like, kind of an evil

949
01:03:15.375 --> 01:03:18.115
I gave it, like, kind of a semi evil prompt.

950
01:03:21.190 --> 01:03:26.329
Build me a Nostril web app that shows the location of photos for my followers on a map view.

951
01:03:27.109 --> 01:03:30.490
I feel like the map part might be the most difficult for it to handle.

952
01:03:34.335 --> 01:03:42.755
Well, like, I know that it's been done before. For example, the Treasures app by Chad has, like, a geocaching map in their user. Use, like, OpenStreetMaps

953
01:03:43.215 --> 01:03:43.875
or something?

954
01:03:44.335 --> 01:03:49.609
Yeah. And so it should be possible. I I don't know where it's gonna get the location data, though.

955
01:03:50.070 --> 01:03:53.210
From the photos because they didn't strip their metadata. Alright.

956
01:03:53.589 --> 01:03:57.050
That I think that step may actually be the hardest part.

957
01:03:57.510 --> 01:04:00.345
But if it can get the data, it can put it on a map.

958
01:04:00.825 --> 01:04:04.605
Everyone strip the metadata from your photos before you upload them anywhere.

959
01:04:06.425 --> 01:04:10.765
If you're if you're uploading on Primal, it should strip the photo metadata for you.

960
01:04:14.185 --> 01:04:15.645
It definitely used to.

961
01:04:17.740 --> 01:04:19.760
Primal's open source. You can check the code,

962
01:04:20.299 --> 01:04:22.720
but it's supposed to strip the data.

963
01:04:23.099 --> 01:04:31.280
An easy way to make sure you have all data stripped is if you paste it in signal and then download from signal, it strips all metadata automatically.

964
01:04:31.635 --> 01:04:35.335
It's also a very easy way to block out faces and stuff before you post.

965
01:04:35.955 --> 01:04:37.015
Signal has, like,

966
01:04:37.875 --> 01:04:40.375
some limited photo editing tools in there that

967
01:04:40.915 --> 01:04:42.455
I like because it's just convenient.

968
01:04:44.355 --> 01:04:46.135
I'll put in, like, IDs to

969
01:04:46.515 --> 01:04:47.095
to Noister.

970
01:04:48.680 --> 01:04:53.160
I dude, I I'll tell you, like, that's probably one of the scariest parts, especially with

971
01:04:54.279 --> 01:04:55.020
so, like,

972
01:04:55.640 --> 01:04:58.700
now that a lot of these apps have Blossom integration,

973
01:04:59.655 --> 01:05:02.795
and, like, do you have I, like, have, like, Blossom Mirror setup or whatever,

974
01:05:03.095 --> 01:05:10.235
it's, like, as soon as I upload something, it just, like, immediately gets mirrored to, like, a bunch of different servers where I can't remove and is, like, signed, hashed,

975
01:05:10.695 --> 01:05:13.755
and I can't easily remove it. It's, like, pretty scary.

976
01:05:14.859 --> 01:05:15.359
Yeah.

977
01:05:15.740 --> 01:05:17.200
Kinda set up. Feel you.

978
01:05:18.060 --> 01:05:23.359
But I don't know. I I've been thinking about, like, do you have any thoughts on that? Like, from the primal side,

979
01:05:24.140 --> 01:05:33.265
we wanted to just add Blossom and offer that and all the people always say delete like, deletions are not real deletions on NoServ, but but, like,

980
01:05:33.965 --> 01:05:34.945
prove it. Because

981
01:05:35.645 --> 01:05:39.265
anytime we've tested whether deletions work, they do work,

982
01:05:39.645 --> 01:05:50.089
at least for notes. So I don't know I don't know if anyone has tested it on Blossom servers, but in terms of, like, is it possible in a decentralized way to do deletions, the answer is definitely yes.

983
01:05:50.390 --> 01:05:57.065
Well, it's just the question is if they are the question is if the servers honor it or not. Right. Exactly. It would be hazardous.

984
01:05:57.445 --> 01:05:59.945
I know the I know prime primal servers,

985
01:06:00.245 --> 01:06:02.105
like, primal's blossom and,

986
01:06:03.845 --> 01:06:06.025
the relay and the caching service

987
01:06:07.125 --> 01:06:08.585
all honor deletes.

988
01:06:09.520 --> 01:06:12.900
Yeah. I mean But I also know, like, UTXO relays.

989
01:06:13.680 --> 01:06:14.180
UTXO,

990
01:06:14.480 --> 01:06:15.220
the webmaster,

991
01:06:15.600 --> 01:06:28.194
he supposedly made a relay that, like, only saves things that have delete requests. So, like, if you send out a delete request, like, he automatically saves it. Alright. And it's one of the few things he hasn't released open source because he's, like,

992
01:06:28.575 --> 01:06:31.714
doesn't wanna be the dude that does it, but someone's gonna do it. Right?

993
01:06:32.255 --> 01:06:37.075
Well, don't stop listening to that relay and don't read from that relay, I guess. And then

994
01:06:37.619 --> 01:06:38.519
everything's good.

995
01:06:38.980 --> 01:06:46.599
Okay. Fair enough. Well, I'm gonna who knows how long this app's gonna take to Vibe? I'm glad we got a successful demo. I'm glad you got to be here for my first.

996
01:06:48.579 --> 01:06:51.380
It was a pleasure having you back. I will say that,

997
01:06:52.545 --> 01:06:53.684
the top zap

998
01:06:53.984 --> 01:06:55.845
from the last rip we did,

999
01:06:58.065 --> 01:07:04.005
what did I say? CD one sixty four, which, by the way, freaks, I think, was a better conversation than this

1000
01:07:07.500 --> 01:07:12.560
because we started from the top Yeah. Rather than being, like, an update episode.

1001
01:07:12.860 --> 01:07:21.835
Beholder said one of the most interesting guests on dispatch would love to see him join more often, and I second that. So let's try and do this every two and three two to three months.

1002
01:07:22.215 --> 01:07:27.275
We'll do, AI vibe, Nostra update. I love hearing from you, so it's a win win.

1003
01:07:28.055 --> 01:07:34.750
Next, we'll be hearing from you because you're gonna build the future of Nostra with this. Yeah. I'm gonna get you some inbound liquidity.

1004
01:07:35.610 --> 01:07:43.070
Let's get your lightning node up and going. I will say, I don't know if you want like, this is also kinda like evil empire shit.

1005
01:07:44.090 --> 01:07:46.510
It's great that you're running your own node, and

1006
01:07:48.245 --> 01:07:52.345
you're getting sats. If you wanna just hold on to the sats, then that's probably the best path.

1007
01:07:52.645 --> 01:07:53.705
But if you do

1008
01:07:54.005 --> 01:07:57.945
if you are trying to convert it all into dollars to pay your bills and shit,

1009
01:07:58.325 --> 01:08:00.585
like, the strike API is a good path.

1010
01:08:02.100 --> 01:08:03.320
Very easy to implement.

1011
01:08:04.100 --> 01:08:05.480
Yeah. But disclosure,

1012
01:08:05.940 --> 01:08:07.400
largest investor in strike,

1013
01:08:08.100 --> 01:08:08.600
and,

1014
01:08:09.780 --> 01:08:11.320
I sit on the board. So,

1015
01:08:12.340 --> 01:08:19.925
but we intentionally made it's if it's if you want dollars. If you want you have to KYC, but if you want dollars, it can That's good to know. May maybe in a certain point,

1016
01:08:20.305 --> 01:08:20.805
like,

1017
01:08:21.425 --> 01:08:23.045
maybe maybe yeah.

1018
01:08:23.665 --> 01:08:28.085
Maybe it makes sense. You don't have and you don't have to worry about liquidity or anything like that.

1019
01:08:28.705 --> 01:08:29.825
Yeah. Because converting,

1020
01:08:31.130 --> 01:08:35.230
is a pain at a certain point when when I do have to keep paying the,

1021
01:08:35.610 --> 01:08:37.389
you know, AI providers in dollars.

1022
01:08:38.409 --> 01:08:40.030
Let's see if my evil app worked.

1023
01:08:41.690 --> 01:08:43.690
I spent $8.94

1024
01:08:43.690 --> 01:08:44.349
for this.

1025
01:08:46.665 --> 01:08:47.645
View photo

1026
01:08:50.985 --> 01:08:51.485
map.

1027
01:08:52.345 --> 01:08:55.165
I think there's, like I I don't think this is gonna work.

1028
01:08:56.265 --> 01:08:58.845
I think it's gonna display a map, but

1029
01:08:59.305 --> 01:09:02.205
I wonder if it's gonna be able to

1030
01:09:02.890 --> 01:09:03.950
extract any

1031
01:09:04.810 --> 01:09:06.990
data from any of these images.

1032
01:09:07.690 --> 01:09:09.070
Like, these exact coordinates.

1033
01:09:09.450 --> 01:09:09.950
Right?

1034
01:09:10.650 --> 01:09:12.190
Oh, shit. Okay.

1035
01:09:14.570 --> 01:09:16.350
Now this now I kinda feel bad.

1036
01:09:20.324 --> 01:09:24.505
Photo do lunchtime, and I have exact coordinates from lunchtime. Delaware.

1037
01:09:26.724 --> 01:09:28.344
For strip your metadata.

1038
01:09:28.885 --> 01:09:31.065
Anyway, this is fucking cool, though.

1039
01:09:31.764 --> 01:09:33.540
We live in a brave new world. Alex,

1040
01:09:34.260 --> 01:09:36.920
I'll post a link to your notes so people give inbound liquidity.

1041
01:09:37.700 --> 01:09:40.760
Shakespeare is at shakespeare.diy.

1042
01:09:41.220 --> 01:09:42.040
We have,

1043
01:09:42.420 --> 01:09:45.620
what is it, soapbox.pub/tools

1044
01:09:45.620 --> 01:09:47.960
is all your other tools that you guys have been building.

1045
01:09:49.565 --> 01:09:53.824
I'll link those all in the show notes. You have any final thoughts for the Freaks before we wrap?

1046
01:09:55.804 --> 01:09:58.145
Build the future with AI on Noister.

1047
01:09:59.485 --> 01:10:00.385
I love it.

1048
01:10:01.060 --> 01:10:04.520
Freaks, I hope you enjoyed this rip. I know it's a little bit different.

1049
01:10:05.300 --> 01:10:12.120
We're just doing things live. I was a little bit distracted, but I hope you enjoyed it. If you like the show, as always, best way to support it,

1050
01:10:12.659 --> 01:10:15.480
is actually not with Bitcoin. It's sharing with your friends and family.

1051
01:10:15.985 --> 01:10:21.045
But Bitcoin donations are appreciated. Interacting with the show, whether that's on Noster or

1052
01:10:21.505 --> 01:10:24.405
podcasting two point o apps, leaving comments, leaving feedback,

1053
01:10:24.785 --> 01:10:25.685
leaving suggestions,

1054
01:10:26.945 --> 01:10:28.245
all that is awesome.

1055
01:10:28.625 --> 01:10:30.725
All links at syllabusbatch.com.

1056
01:10:31.010 --> 01:10:33.350
All my links are at odell.xyz.

1057
01:10:33.730 --> 01:10:35.510
I'm taking next week off.

1058
01:10:39.090 --> 01:10:39.830
I'm traveling

1059
01:10:40.130 --> 01:10:40.630
for

1060
01:10:42.610 --> 01:10:45.350
this work thing, or this Bitcoin thing.

1061
01:10:47.635 --> 01:10:50.135
That should be pretty good. But as a result,

1062
01:10:50.755 --> 01:10:53.015
I'm skipping a live show next week,

1063
01:10:53.554 --> 01:10:55.415
but we'll be back the week after that.

1064
01:10:55.875 --> 01:10:59.415
Alex, thank you again. Thank you. Freaks.

1065
01:10:59.795 --> 01:11:01.739
Stay on the Stack Sats. Love you all. Peace.