1
00:00:00,000 --> 00:00:01,480
Can you hear me eat, DK?

2
00:00:02,920 --> 00:00:04,520
And so can everyone else because we're live.

3
00:00:04,640 --> 00:00:05,660
I love you eating, but I don't.

4
00:00:05,780 --> 00:00:06,400
Oh, we're live.

5
00:00:06,560 --> 00:00:06,860
Look at that.

6
00:00:06,920 --> 00:00:07,160
We are.

7
00:00:08,160 --> 00:00:08,560
All right.

8
00:00:08,980 --> 00:00:09,480
Steve's eating.

9
00:00:10,080 --> 00:00:11,180
Tell us, what are you eating, Steve?

10
00:00:11,680 --> 00:00:12,920
Why are you eating on the show?

11
00:00:13,380 --> 00:00:16,360
It doesn't look like a PB&J, so I need to understand why we're off.

12
00:00:16,360 --> 00:00:17,120
Eating chocolate.

13
00:00:17,500 --> 00:00:22,140
I didn't know we were live, but maybe the viewers can let me know if it's acceptable.

14
00:00:22,140 --> 00:00:24,060
Just make crunch in the microphone.

15
00:00:26,460 --> 00:00:29,620
I'm eating an Elon chocolate bar.

16
00:00:29,620 --> 00:00:30,800
but not Elon Musk.

17
00:00:31,060 --> 00:00:31,160
Oh.

18
00:00:31,780 --> 00:00:32,980
Elon University.

19
00:00:34,180 --> 00:00:36,420
The alma mater of Austin.

20
00:00:37,200 --> 00:00:37,380
Okay.

21
00:00:37,980 --> 00:00:39,220
Krauss from BB.

22
00:00:39,520 --> 00:00:39,860
Okay.

23
00:00:40,320 --> 00:00:43,180
He had some entrepreneurial students.

24
00:00:43,940 --> 00:00:47,440
They were doing a San Francisco road show, I think.

25
00:00:47,740 --> 00:00:48,040
Uh-huh.

26
00:00:48,340 --> 00:00:51,500
And he had them swing by here last week.

27
00:00:51,700 --> 00:00:51,980
Right.

28
00:00:52,280 --> 00:00:52,460
Yeah.

29
00:00:52,820 --> 00:00:54,420
And it's great.

30
00:00:54,420 --> 00:00:56,780
Love speaking to students.

31
00:00:56,780 --> 00:01:02,140
and I think they're all graduating this spring.

32
00:01:02,520 --> 00:01:03,400
Get them started young.

33
00:01:03,440 --> 00:01:04,660
Trying to figure out what they're going to do.

34
00:01:05,120 --> 00:01:07,320
So Jesse Bosner and I, and I don't know who else,

35
00:01:07,480 --> 00:01:09,540
and Austin, of course, spoke to them.

36
00:01:12,540 --> 00:01:14,440
We might have even talked about it last week.

37
00:01:14,440 --> 00:01:15,080
I think we did, yeah.

38
00:01:15,700 --> 00:01:17,400
Anyway, so we don't have to talk about it again.

39
00:01:17,420 --> 00:01:18,360
You spread the good word of Satoshi.

40
00:01:18,560 --> 00:01:20,020
But you're kind enough.

41
00:01:21,300 --> 00:01:23,300
Oh, they just brought you a chocolate.

42
00:01:23,560 --> 00:01:25,260
They gave me a little gift bag with chocolate.

43
00:01:25,400 --> 00:01:26,080
Very kind.

44
00:01:26,080 --> 00:01:29,620
And Elon University, for those that don't know, I believe is in North Carolina.

45
00:01:29,860 --> 00:01:30,380
It's in the South.

46
00:01:30,640 --> 00:01:31,360
It's in North Carolina.

47
00:01:31,580 --> 00:01:31,760
Yeah.

48
00:01:31,840 --> 00:01:33,320
So it makes sense that they would bring a gift.

49
00:01:33,580 --> 00:01:34,200
Southern hospitality.

50
00:01:34,580 --> 00:01:34,940
Definitely.

51
00:01:35,140 --> 00:01:35,900
It was very nice of them.

52
00:01:36,160 --> 00:01:37,400
And yeah, why am I eating it right now?

53
00:01:37,520 --> 00:01:38,540
I haven't eaten yet today.

54
00:01:38,820 --> 00:01:39,820
I've been rushing around.

55
00:01:40,100 --> 00:01:42,420
So I didn't want to pass out on the show.

56
00:01:42,860 --> 00:01:44,500
Did you start the morning in New York?

57
00:01:44,800 --> 00:01:45,700
Or you got in?

58
00:01:45,700 --> 00:01:48,980
No, I got back like at 2 a.m.

59
00:01:49,580 --> 00:01:50,120
Oh, okay.

60
00:01:50,620 --> 00:01:51,200
Very reasonable.

61
00:01:51,900 --> 00:01:54,180
But Brian Johnson would not approve of your sleep schedule, I tell you.

62
00:01:54,600 --> 00:01:55,660
Or of your chocolate.

63
00:01:55,660 --> 00:02:00,240
I actually got pretty good sleep because I actually, I was so crashing from the, yeah.

64
00:02:01,740 --> 00:02:04,780
I was, I also gained two pounds from the trip.

65
00:02:04,920 --> 00:02:06,540
So that's not working towards my goals.

66
00:02:06,660 --> 00:02:07,940
I was going to say you're looking more ripped today.

67
00:02:08,500 --> 00:02:10,280
That's because I'm wearing my t-shirt.

68
00:02:11,340 --> 00:02:11,940
We have so much.

69
00:02:12,160 --> 00:02:16,160
I don't even know if we're going to talk about AI and Bitcoin because they'll just talk about Steve's morning.

70
00:02:17,600 --> 00:02:23,060
So I actually got decent sleep because I was so crashed from the week because it was an amazing week.

71
00:02:23,060 --> 00:02:25,720
I actually slept in the flight back.

72
00:02:25,860 --> 00:02:26,120
Okay.

73
00:02:26,460 --> 00:02:27,800
And then so I was able to wake up this morning.

74
00:02:27,880 --> 00:02:28,080
Fine.

75
00:02:29,280 --> 00:02:30,500
Did you pop magnesium or anything?

76
00:02:30,580 --> 00:02:31,420
How do you sleep on the flight?

77
00:02:31,820 --> 00:02:33,000
Because I can't sleep on flights anymore.

78
00:02:33,220 --> 00:02:34,000
Oh, I just fell asleep.

79
00:02:34,300 --> 00:02:34,580
Nice.

80
00:02:34,800 --> 00:02:35,060
Natural.

81
00:02:35,680 --> 00:02:36,400
I worked hard.

82
00:02:36,980 --> 00:02:37,520
Worked hard.

83
00:02:37,660 --> 00:02:37,740
Yeah.

84
00:02:37,740 --> 00:02:38,120
Work hard.

85
00:02:38,220 --> 00:02:38,500
That's it.

86
00:02:38,500 --> 00:02:39,900
I worked hard and I cracked.

87
00:02:39,900 --> 00:02:40,380
Work hard.

88
00:02:40,380 --> 00:02:40,480
Work hard.

89
00:02:40,480 --> 00:02:40,540
Work hard.

90
00:02:40,780 --> 00:02:41,420
Stack sats.

91
00:02:42,180 --> 00:02:42,340
Yeah.

92
00:02:42,400 --> 00:02:46,440
Why was I a couple of minutes late for the show?

93
00:02:46,620 --> 00:02:46,800
Yeah.

94
00:02:47,780 --> 00:02:49,680
Today is Dipsy registration.

95
00:02:50,060 --> 00:02:50,480
Oh, it's good.

96
00:02:50,480 --> 00:02:50,800
Oh, yeah.

97
00:02:50,940 --> 00:02:52,900
So for those that don't know what the Dipsy is,

98
00:02:52,900 --> 00:02:54,920
it's the second oldest running race in America

99
00:02:54,920 --> 00:02:55,880
behind the Boston Marathon.

100
00:02:57,160 --> 00:02:58,940
And this will be my fourth

101
00:02:58,940 --> 00:03:00,120
year running it.

102
00:03:01,140 --> 00:03:02,180
And it's a somewhat

103
00:03:02,180 --> 00:03:04,840
unusual race. It's very

104
00:03:04,840 --> 00:03:06,980
hard to get in. You have to apply

105
00:03:06,980 --> 00:03:08,640
and there's all kinds of...

106
00:03:08,640 --> 00:03:11,020
You have to write a sob story.

107
00:03:12,120 --> 00:03:12,480
You

108
00:03:12,480 --> 00:03:14,340
can bribe the money.

109
00:03:15,800 --> 00:03:16,640
Is this college emissions?

110
00:03:16,920 --> 00:03:18,940
You need to...

111
00:03:18,940 --> 00:03:20,860
If you want to increase your chances,

112
00:03:20,860 --> 00:03:23,700
you need to apply the day that they open up registration

113
00:03:23,700 --> 00:03:27,820
and do it from the post office in Milwaukee.

114
00:03:29,820 --> 00:03:32,560
So I had everything ready,

115
00:03:33,560 --> 00:03:35,520
and then I swung by this morning,

116
00:03:35,820 --> 00:03:37,740
and the line was so long

117
00:03:37,740 --> 00:03:40,900
that I finally just had to bail and get here.

118
00:03:41,180 --> 00:03:42,660
So that means you're not registered?

119
00:03:43,100 --> 00:03:44,200
Well, I'll just do it after the show.

120
00:03:44,500 --> 00:03:46,680
So you think it won't fill up in the morning?

121
00:03:46,680 --> 00:03:48,020
Well, I'll be fine.

122
00:03:48,020 --> 00:03:51,540
So if you do well enough in the race in a previous year,

123
00:03:51,600 --> 00:03:52,460
you automatically get it.

124
00:03:52,540 --> 00:03:52,900
So I'm good.

125
00:03:52,900 --> 00:03:53,600
Yeah, that's what I was doing.

126
00:03:53,600 --> 00:03:54,820
I'm registering for other people.

127
00:03:54,920 --> 00:03:55,580
Oh, yeah, yeah.

128
00:03:57,320 --> 00:03:59,120
Anyway, so that's why it's a little late.

129
00:03:59,660 --> 00:04:01,800
And then, yeah, why am I wearing this, Max?

130
00:04:02,460 --> 00:04:03,920
Dude, I can't talk about basketball.

131
00:04:04,580 --> 00:04:06,300
I've completely turned it on.

132
00:04:06,300 --> 00:04:08,820
What makes me sad is Memphis is not doing well,

133
00:04:08,820 --> 00:04:12,000
so I lose the love for Max.

134
00:04:12,080 --> 00:04:14,180
I'm going to support Iowa State this year because...

135
00:04:14,180 --> 00:04:16,880
You've got to root for Iowa State and Gonzaga.

136
00:04:16,880 --> 00:04:17,360
Yeah, yeah.

137
00:04:17,620 --> 00:04:20,920
And actually, our latest, or our newest member presented at Bitcoin.

138
00:04:21,160 --> 00:04:21,540
Who's that?

139
00:04:21,660 --> 00:04:22,640
College basketball fan.

140
00:04:23,520 --> 00:04:24,040
The Mara dude?

141
00:04:24,720 --> 00:04:24,940
What?

142
00:04:25,080 --> 00:04:25,680
The Mara dude?

143
00:04:25,900 --> 00:04:26,080
Yeah.

144
00:04:26,480 --> 00:04:27,240
I haven't met him yet.

145
00:04:27,240 --> 00:04:27,600
I need to.

146
00:04:28,100 --> 00:04:28,600
He's great.

147
00:04:28,760 --> 00:04:28,980
Yeah.

148
00:04:29,460 --> 00:04:30,560
And he went to Santa Clara.

149
00:04:30,560 --> 00:04:35,460
So he's not only a college basketball fan, but he's a rival of Bailey's Gonzaga team.

150
00:04:35,900 --> 00:04:38,760
And they actually played in the championship of WCC this week.

151
00:04:38,760 --> 00:04:39,500
They were in the WCC.

152
00:04:39,700 --> 00:04:40,720
But Gonzaga won.

153
00:04:40,840 --> 00:04:44,400
So I think Santa Clara had a good season, but I don't think they'll make the dance.

154
00:04:44,520 --> 00:04:44,840
Yeah, okay.

155
00:04:44,840 --> 00:04:45,620
you know

156
00:04:45,620 --> 00:04:47,080
NIT's not too bad

157
00:04:47,080 --> 00:04:48,100
bro

158
00:04:48,100 --> 00:04:50,240
Memphis has been so bad

159
00:04:50,240 --> 00:04:50,520
like

160
00:04:50,520 --> 00:04:51,500
I don't even want to talk about it

161
00:04:51,500 --> 00:04:51,860
but like

162
00:04:51,860 --> 00:04:53,260
remember that like

163
00:04:53,260 --> 00:04:54,000
there's like the NCAA

164
00:04:54,000 --> 00:04:55,160
then there's like the NIT

165
00:04:55,160 --> 00:04:56,260
which is like the losers tournament

166
00:04:56,260 --> 00:04:57,100
then there's like

167
00:04:57,100 --> 00:04:57,540
the like

168
00:04:57,540 --> 00:04:58,320
losers losers

169
00:04:58,320 --> 00:04:59,340
like NBIL

170
00:04:59,340 --> 00:05:00,220
yeah none of the other stuff

171
00:05:00,220 --> 00:05:01,060
NIT doesn't

172
00:05:01,060 --> 00:05:02,420
no one cares about the NIT

173
00:05:02,420 --> 00:05:03,340
that's like when you

174
00:05:03,340 --> 00:05:04,200
don't make the playoffs

175
00:05:04,200 --> 00:05:05,220
in your fantasy football league

176
00:05:05,220 --> 00:05:05,860
like who cares

177
00:05:05,860 --> 00:05:08,060
I literally tune out

178
00:05:08,060 --> 00:05:09,000
if I don't make it

179
00:05:09,000 --> 00:05:09,540
but I always say

180
00:05:09,540 --> 00:05:11,060
did you already win Big 12

181
00:05:11,060 --> 00:05:12,260
or no that's this weekend

182
00:05:12,260 --> 00:05:13,080
we're in the middle

183
00:05:13,080 --> 00:05:13,460
so

184
00:05:13,460 --> 00:05:14,220
yeah yeah

185
00:05:14,840 --> 00:05:17,560
We're the seventh ranked team in the nation.

186
00:05:18,060 --> 00:05:18,560
And we're...

187
00:05:18,560 --> 00:05:19,320
Seventh in the Big 12.

188
00:05:19,940 --> 00:05:21,300
Yeah, well, close.

189
00:05:21,600 --> 00:05:22,580
It is so stacked.

190
00:05:22,760 --> 00:05:26,020
We're the fifth seed in the conference.

191
00:05:26,220 --> 00:05:26,820
Oh, it's Houston.

192
00:05:26,820 --> 00:05:28,060
So we didn't even get a bye.

193
00:05:28,560 --> 00:05:28,860
Ooh.

194
00:05:29,220 --> 00:05:30,400
I mean, there's a double.

195
00:05:30,540 --> 00:05:35,020
Anyway, we didn't get the bye that Houston, Arizona, Kansas, and Texas Tech got.

196
00:05:35,180 --> 00:05:36,320
How is Big 12 that?

197
00:05:36,400 --> 00:05:36,980
I mean, they're good.

198
00:05:37,100 --> 00:05:37,700
They've always been good.

199
00:05:37,780 --> 00:05:38,260
Big 12 is stacked.

200
00:05:38,540 --> 00:05:39,120
How is that?

201
00:05:39,120 --> 00:05:40,660
We're literally the seventh best team in the nation.

202
00:05:40,980 --> 00:05:42,160
And we're the fifth best.

203
00:05:42,420 --> 00:05:44,120
But we're actually better than the fifth best in the Big 12.

204
00:05:44,320 --> 00:05:44,500
That's...

205
00:05:44,840 --> 00:05:46,560
Anyway, and we're proving it out.

206
00:05:47,080 --> 00:05:47,240
Yeah.

207
00:05:47,920 --> 00:05:52,200
On Wednesday, we played Arizona State, which, you know, they're like a...

208
00:05:52,200 --> 00:05:53,120
They're okay.

209
00:05:53,200 --> 00:05:53,600
Mediocre.

210
00:05:53,800 --> 00:05:53,960
Yeah.

211
00:05:54,100 --> 00:05:57,080
They're not bad, but they're not a tourney team.

212
00:05:57,160 --> 00:05:57,340
Yeah.

213
00:05:57,860 --> 00:05:58,140
Mediocre.

214
00:05:58,340 --> 00:06:01,420
But we beat them 91 to 42.

215
00:06:02,120 --> 00:06:02,460
That's embarrassing.

216
00:06:02,460 --> 00:06:06,480
It was the largest beatdown in the history of the tournament, and Arizona State's worst

217
00:06:06,480 --> 00:06:08,300
loss in like 70 years.

218
00:06:08,400 --> 00:06:08,920
That's crazy.

219
00:06:09,680 --> 00:06:12,360
The next day, then we played a good team, Texas Tech.

220
00:06:12,540 --> 00:06:12,800
Yeah.

221
00:06:12,940 --> 00:06:13,180
Right?

222
00:06:13,360 --> 00:06:13,640
Yeah.

223
00:06:13,640 --> 00:06:14,480
This team to beat Duke.

224
00:06:14,840 --> 00:06:15,840
Houston, Arizona.

225
00:06:16,080 --> 00:06:16,480
Good team.

226
00:06:16,660 --> 00:06:17,000
Good team.

227
00:06:17,540 --> 00:06:18,680
We beat them by 22.

228
00:06:18,920 --> 00:06:19,280
Jesus.

229
00:06:20,980 --> 00:06:22,220
Tonight, we get the real test.

230
00:06:22,380 --> 00:06:23,260
We play the best in the nation.

231
00:06:23,460 --> 00:06:23,820
Arizona.

232
00:06:24,000 --> 00:06:24,640
Arizona, okay.

233
00:06:25,480 --> 00:06:26,240
They're 30-2.

234
00:06:26,540 --> 00:06:27,080
They're very good.

235
00:06:27,320 --> 00:06:27,600
Okay.

236
00:06:28,000 --> 00:06:28,480
I mean, Arizona's...

237
00:06:28,480 --> 00:06:30,280
They're all time, like, historical good.

238
00:06:30,600 --> 00:06:32,280
So tonight will be a huge test.

239
00:06:32,360 --> 00:06:34,580
And if we win that, then we have to play Houston or Kansas.

240
00:06:35,020 --> 00:06:35,940
It's so ridiculous.

241
00:06:36,180 --> 00:06:36,660
Houston's one.

242
00:06:36,700 --> 00:06:37,940
I never want to go up against Calvin Sampson.

243
00:06:38,160 --> 00:06:39,740
Is Sean Miller, is he still at Arizona?

244
00:06:40,100 --> 00:06:40,360
No.

245
00:06:40,560 --> 00:06:42,180
Tommy Lloyd, who came from Gonzaga.

246
00:06:42,600 --> 00:06:43,160
I'm old, man.

247
00:06:43,380 --> 00:06:43,580
Yeah.

248
00:06:43,580 --> 00:06:44,840
Tommy Lloyd's got Arizona.

249
00:06:45,780 --> 00:06:46,000
Yeah.

250
00:06:46,320 --> 00:06:46,580
Rolling.

251
00:06:47,420 --> 00:06:47,720
So anyway.

252
00:06:48,540 --> 00:06:49,680
And for those of you who don't know,

253
00:06:49,880 --> 00:06:50,960
what Steve's talking about at Iowa State

254
00:06:50,960 --> 00:06:51,640
is quite impressive

255
00:06:51,640 --> 00:06:53,200
because you guys are not like an NIL.

256
00:06:53,320 --> 00:06:55,520
NIL is like the money bag powerhouse right now.

257
00:06:55,700 --> 00:06:57,080
The fact that y'all are this good.

258
00:06:57,480 --> 00:06:58,760
We're money balling.

259
00:06:58,860 --> 00:06:59,220
You are.

260
00:06:59,300 --> 00:06:59,580
Exactly.

261
00:06:59,900 --> 00:07:00,740
Just like your namesake.

262
00:07:01,100 --> 00:07:01,780
It's pretty insane.

263
00:07:01,940 --> 00:07:03,460
This shirt is from two years ago.

264
00:07:04,240 --> 00:07:06,860
I was at the championship game in Kansas City

265
00:07:06,860 --> 00:07:07,960
and we won.

266
00:07:08,680 --> 00:07:10,600
We beat Houston by like over 30 points.

267
00:07:10,660 --> 00:07:11,540
They were number one in the nation.

268
00:07:11,540 --> 00:07:11,900
Yeah.

269
00:07:11,900 --> 00:07:14,660
It was the largest victory over number one team in history.

270
00:07:14,800 --> 00:07:16,580
And Houston's like always nice, always.

271
00:07:16,860 --> 00:07:19,040
If we win the championship tomorrow,

272
00:07:19,340 --> 00:07:22,580
then we'll have won six of the last 12 Big 12 championships.

273
00:07:22,620 --> 00:07:24,080
So you guys are going to be a one seed then, you think?

274
00:07:24,100 --> 00:07:24,660
That's pretty amazing.

275
00:07:25,880 --> 00:07:27,020
There's no way we'll be one.

276
00:07:27,560 --> 00:07:29,560
Even if you win the Big 12 tournament, you won't be a one seed?

277
00:07:29,640 --> 00:07:29,860
Correct.

278
00:07:30,340 --> 00:07:30,760
How's that point?

279
00:07:30,760 --> 00:07:32,600
Well, I mean, there's a small chance.

280
00:07:32,700 --> 00:07:32,880
Okay.

281
00:07:33,400 --> 00:07:35,220
We're likely a two seed.

282
00:07:35,600 --> 00:07:36,760
But if you take out Arizona and Houston—

283
00:07:36,760 --> 00:07:38,940
We could slip to a three seed if we lose tonight.

284
00:07:39,960 --> 00:07:40,560
That's crazy.

285
00:07:40,560 --> 00:07:44,580
Like Michigan State, Illinois, Iowa State, Houston are all right in there.

286
00:07:44,780 --> 00:07:45,380
Right, right.

287
00:07:45,480 --> 00:07:48,920
Well, it would be hard to like, it's not like definitive.

288
00:07:49,060 --> 00:07:49,660
Anyway, we'll see.

289
00:07:49,920 --> 00:07:50,000
Yeah.

290
00:07:50,960 --> 00:07:51,740
How are we doing on time?

291
00:07:52,300 --> 00:07:52,840
Illinois good?

292
00:07:54,120 --> 00:07:55,120
How is Illinois doing?

293
00:07:55,500 --> 00:07:56,520
Sports got Illinois.

294
00:07:56,520 --> 00:07:56,800
Oh, yeah.

295
00:07:57,500 --> 00:07:58,200
Illinois is better.

296
00:07:58,200 --> 00:07:59,120
Yeah, Urbana Champaign.

297
00:07:59,220 --> 00:07:59,960
Your school.

298
00:08:00,260 --> 00:08:02,760
Your school is right in there with mine.

299
00:08:03,660 --> 00:08:06,100
Wait, is that the same Illinois?

300
00:08:06,360 --> 00:08:06,720
It is.

301
00:08:06,800 --> 00:08:07,440
Oh, it is.

302
00:08:07,460 --> 00:08:08,080
Urbana Champaign.

303
00:08:08,080 --> 00:08:09,660
so that's like

304
00:08:09,660 --> 00:08:10,580
the University of Illinois

305
00:08:10,580 --> 00:08:11,660
is University of Illinois

306
00:08:11,660 --> 00:08:12,900
well there's Chicago

307
00:08:12,900 --> 00:08:13,320
and other

308
00:08:13,320 --> 00:08:14,620
but Urbana Champagne

309
00:08:14,620 --> 00:08:14,900
is like

310
00:08:14,900 --> 00:08:15,500
it's like the main one

311
00:08:15,500 --> 00:08:16,020
the main one

312
00:08:16,020 --> 00:08:16,280
yeah

313
00:08:16,280 --> 00:08:18,220
yeah I mean you guys are good

314
00:08:18,220 --> 00:08:18,700
that's where

315
00:08:18,700 --> 00:08:20,220
DK and Mark Andreessen

316
00:08:20,220 --> 00:08:20,720
came from

317
00:08:20,720 --> 00:08:22,160
who else

318
00:08:22,160 --> 00:08:23,260
what are other famous

319
00:08:23,260 --> 00:08:25,300
famous tech alum

320
00:08:25,300 --> 00:08:26,600
oh tech

321
00:08:26,600 --> 00:08:27,440
I thought you were gonna go

322
00:08:27,440 --> 00:08:28,800
Playboy guy

323
00:08:28,800 --> 00:08:29,680
I think it was from there

324
00:08:29,680 --> 00:08:30,220
yeah

325
00:08:30,220 --> 00:08:30,740
oh good

326
00:08:30,740 --> 00:08:31,240
Q Hefner

327
00:08:31,240 --> 00:08:31,620
right

328
00:08:31,620 --> 00:08:33,140
we got DK

329
00:08:33,140 --> 00:08:33,920
Mark Andreessen

330
00:08:33,920 --> 00:08:34,720
Q Hefner

331
00:08:34,720 --> 00:08:36,160
SPF

332
00:08:36,160 --> 00:08:37,980
I don't think

333
00:08:37,980 --> 00:08:39,260
I don't think we rep SPF.

334
00:08:39,540 --> 00:08:40,280
Max Lepstian.

335
00:08:41,120 --> 00:08:43,460
A good one.

336
00:08:43,700 --> 00:08:43,820
Yeah.

337
00:08:44,360 --> 00:08:44,760
There we go.

338
00:08:45,880 --> 00:08:46,240
Cool.

339
00:08:46,440 --> 00:08:48,420
We got all kinds of exciting stuff to talk today.

340
00:08:48,660 --> 00:08:50,560
Like, why was I flying in at 2 a.m.?

341
00:08:50,560 --> 00:08:51,560
I was in New York.

342
00:08:51,800 --> 00:08:52,120
Uh-oh.

343
00:08:52,600 --> 00:08:53,780
Oh, we got all kinds of comments.

344
00:08:53,780 --> 00:08:54,380
We got all kinds of comments.

345
00:08:54,480 --> 00:08:56,000
PBJ now with more ASMR.

346
00:08:56,260 --> 00:08:58,000
I think that was when you were eating in the morning.

347
00:08:58,820 --> 00:08:59,360
That's right.

348
00:09:01,840 --> 00:09:02,740
I can't read it.

349
00:09:02,800 --> 00:09:03,320
Read it to me.

350
00:09:03,480 --> 00:09:03,920
What do you got?

351
00:09:04,320 --> 00:09:05,840
DK's eyes are glazing over.

352
00:09:06,100 --> 00:09:06,900
And there's lots of laughing.

353
00:09:06,900 --> 00:09:11,280
jacked steve everyone thinks you look jacked steve like you do he'd been working out so this this

354
00:09:11,280 --> 00:09:16,060
two pounds looks like two pounds of muscle two pounds of muscle yeah right up i did do i did do

355
00:09:16,060 --> 00:09:21,680
some daily push-ups in new york there you go yeah finger red number one finger red number one finger

356
00:09:21,680 --> 00:09:27,880
red number one maxi madness dk's eyes are closing over then he's laughing and matt is telling them

357
00:09:27,880 --> 00:09:34,800
don't encourage them so he wants me to he wants me to call lay the law down let's talk about some

358
00:09:34,800 --> 00:09:37,000
Bitcoin or AI stuff.

359
00:09:37,240 --> 00:09:38,160
All right, let's do it.

360
00:09:38,260 --> 00:09:38,960
Well, before we do,

361
00:09:39,040 --> 00:09:40,240
do we have that one guy

362
00:09:40,240 --> 00:09:41,100
who is apparently

363
00:09:41,100 --> 00:09:42,620
both a major Bitcoiner

364
00:09:42,620 --> 00:09:44,120
and an Iowa State fan?

365
00:09:44,160 --> 00:09:45,020
Because if that one guy

366
00:09:45,020 --> 00:09:45,660
is there now,

367
00:09:45,760 --> 00:09:45,920
he's your own, my friend.

368
00:09:45,920 --> 00:09:46,960
Oh, yeah, the YouTube commenter.

369
00:09:47,140 --> 00:09:47,840
Show yourself.

370
00:09:48,200 --> 00:09:49,080
I'm very loud, man.

371
00:09:49,680 --> 00:09:50,960
Tell me your prediction for tonight.

372
00:09:51,240 --> 00:09:51,600
Exactly.

373
00:09:52,500 --> 00:09:53,120
Anyway, yeah.

374
00:09:53,200 --> 00:09:54,060
So I was in New York.

375
00:09:55,520 --> 00:09:56,380
Well, the Spiral team

376
00:09:56,380 --> 00:09:57,560
had an in-person retreat.

377
00:09:57,660 --> 00:09:58,400
We do a couple a year.

378
00:09:58,500 --> 00:09:59,460
This is our first of the year.

379
00:10:00,560 --> 00:10:01,940
Always, we have our team

380
00:10:01,940 --> 00:10:03,440
with a lot of love

381
00:10:03,440 --> 00:10:05,620
between everyone on our teams.

382
00:10:05,680 --> 00:10:06,800
It's great to see each other in person.

383
00:10:07,420 --> 00:10:08,500
We had great discussions.

384
00:10:09,480 --> 00:10:11,560
One thing that Haley had a great idea

385
00:10:11,560 --> 00:10:13,660
to do in New York is hold a

386
00:10:13,660 --> 00:10:15,500
Builder event. We've been doing it for like

387
00:10:15,500 --> 00:10:16,860
nine straight months here.

388
00:10:17,960 --> 00:10:19,560
The Genesis idea came on

389
00:10:19,560 --> 00:10:20,420
this pod.

390
00:10:22,740 --> 00:10:22,980
And

391
00:10:22,980 --> 00:10:25,320
several other cities have sprouted about

392
00:10:25,320 --> 00:10:27,300
doing Builder, but New York hasn't.

393
00:10:27,460 --> 00:10:29,080
So Haley's like, let's do a Builder.

394
00:10:29,240 --> 00:10:30,880
So Matt Velez and Haley and I

395
00:10:30,880 --> 00:10:33,160
hosted Builder.

396
00:10:33,440 --> 00:10:34,520
New York. At PubKey.

397
00:10:34,780 --> 00:10:35,920
At PubKey. Nice.

398
00:10:36,820 --> 00:10:39,020
So it's fun. Thomas

399
00:10:39,020 --> 00:10:41,160
at PubKey has a podcast that they

400
00:10:41,160 --> 00:10:42,800
I didn't know it either.

401
00:10:44,060 --> 00:10:44,580
Have you

402
00:10:44,580 --> 00:10:46,620
been to PubKey? Never been. No.

403
00:10:46,780 --> 00:10:48,540
I haven't been to New York. And I'm guessing DK has not been to PubKey?

404
00:10:48,640 --> 00:10:50,700
I haven't been. No. When you're in New York,

405
00:10:50,840 --> 00:10:53,080
definitely go to PubKey. You go in.

406
00:10:53,580 --> 00:10:54,400
There's three

407
00:10:54,400 --> 00:10:55,520
rooms of

408
00:10:55,520 --> 00:10:58,640
significance. The first one is like a

409
00:10:58,640 --> 00:11:00,400
classic New York dive bar.

410
00:11:01,260 --> 00:11:02,260
If just in

411
00:11:02,260 --> 00:11:03,940
And normal folks go in here.

412
00:11:04,600 --> 00:11:07,480
And they probably don't know it's a Bitcoin bar.

413
00:11:08,180 --> 00:11:13,100
Unless you look pretty closely and then you see some Satoshi shrine stuff.

414
00:11:14,020 --> 00:11:14,860
You go to the...

415
00:11:14,860 --> 00:11:16,900
Or memorabilia.

416
00:11:17,300 --> 00:11:21,260
Then if you go to the back room, that's where there's more full-on Bitcoin.

417
00:11:22,120 --> 00:11:22,600
Dianetics.

418
00:11:22,620 --> 00:11:23,400
A little bit more like here.

419
00:11:24,440 --> 00:11:27,040
Well, actually harder core than here.

420
00:11:27,040 --> 00:11:33,140
And then there's another room where it is full on pride.

421
00:11:33,840 --> 00:11:35,280
Was that the podcast room?

422
00:11:35,600 --> 00:11:36,780
Yeah, that's the podcast room.

423
00:11:37,080 --> 00:11:40,340
So you would comment that you like the chaos and stuff

424
00:11:40,340 --> 00:11:42,480
and the colors and the chaos, and that's definitely true.

425
00:11:42,620 --> 00:11:44,980
And I don't know if you noticed, but on the wall in the background,

426
00:11:45,460 --> 00:11:47,480
they have people signing their names, like visitors.

427
00:11:48,880 --> 00:11:50,440
So Haley, Matt, and I signed our names.

428
00:11:50,880 --> 00:11:53,860
I mean, there's hundreds of names signed there, so that's cool.

429
00:11:54,040 --> 00:11:54,820
Anyway, we did a pod.

430
00:11:54,820 --> 00:11:57,180
We mainly talked about Builder and PB.

431
00:11:58,780 --> 00:11:59,800
Then we did Builder.

432
00:12:00,900 --> 00:12:03,780
We had at least 50 people come.

433
00:12:04,100 --> 00:12:04,280
Nice.

434
00:12:04,920 --> 00:12:06,340
I don't even know the count

435
00:12:06,340 --> 00:12:10,440
because in the main room we were presenting in,

436
00:12:10,720 --> 00:12:12,180
it holds 50 and it was packed

437
00:12:12,180 --> 00:12:15,440
and there's a separate room behind it

438
00:12:15,440 --> 00:12:16,540
and I don't know how many people are in there,

439
00:12:16,740 --> 00:12:17,640
but at least 50 people.

440
00:12:17,860 --> 00:12:19,740
They're like 90 plus RSVPs.

441
00:12:21,600 --> 00:12:23,380
Was it pretty similar format

442
00:12:23,380 --> 00:12:25,260
to how we've done them in SF,

443
00:12:25,460 --> 00:12:26,720
or do they kind of vary a bit?

444
00:12:27,540 --> 00:12:28,040
Similar.

445
00:12:28,740 --> 00:12:31,600
We had four presentations.

446
00:12:32,480 --> 00:12:37,820
So I went first and presented Unleash the Agents,

447
00:12:38,500 --> 00:12:41,000
which I'll have to present here to you guys.

448
00:12:42,220 --> 00:12:44,900
But it's about agentic commerce

449
00:12:44,900 --> 00:12:49,460
and why it's going to be big,

450
00:12:50,540 --> 00:12:53,260
what are the different payment options,

451
00:12:53,380 --> 00:12:56,400
sort of, you know, credit cards,

452
00:12:56,660 --> 00:12:59,140
like TradFi 2.0, stablecoins, and Bitcoin.

453
00:12:59,440 --> 00:13:02,120
Makes the case for Bitcoin, but it addresses all of them.

454
00:13:04,180 --> 00:13:07,120
So I spent...

455
00:13:07,120 --> 00:13:09,240
Matt and the Les put the presentation together.

456
00:13:10,580 --> 00:13:12,240
No one thought I could do it in 20 minutes.

457
00:13:12,420 --> 00:13:13,240
I did it in 20 minutes.

458
00:13:13,380 --> 00:13:14,160
Very proud of myself.

459
00:13:14,160 --> 00:13:15,300
I did it in 20 minutes.

460
00:13:16,160 --> 00:13:18,160
First rehearsal was over an hour, so...

461
00:13:18,620 --> 00:13:20,860
No, it was good.

462
00:13:21,180 --> 00:13:23,160
So cover that, and then...

463
00:13:23,380 --> 00:13:24,100
Nick Slaney came.

464
00:13:24,260 --> 00:13:24,580
Nice.

465
00:13:24,680 --> 00:13:25,200
Yeah, I saw that.

466
00:13:25,300 --> 00:13:26,620
So I guess we're going to,

467
00:13:27,300 --> 00:13:28,800
this will be the eighth straight episode

468
00:13:28,800 --> 00:13:30,020
talking about MDK, I guess.

469
00:13:30,100 --> 00:13:32,280
Wait, what was his bot's name again?

470
00:13:32,280 --> 00:13:33,060
Is it Ori?

471
00:13:33,320 --> 00:13:34,160
Is it Ori?

472
00:13:34,420 --> 00:13:34,660
Nick?

473
00:13:34,920 --> 00:13:35,220
Ori?

474
00:13:36,020 --> 00:13:37,420
I'm receiving feedback from folks

475
00:13:37,420 --> 00:13:39,080
that it is quite the streak.

476
00:13:39,580 --> 00:13:40,620
But anyway, Nick came,

477
00:13:40,960 --> 00:13:43,620
gave a good five-minute prezo,

478
00:13:43,760 --> 00:13:44,820
and you guys probably saw

479
00:13:44,820 --> 00:13:47,400
he created the prezo a few hours before

480
00:13:47,400 --> 00:13:50,140
and had SatBot created for him.

481
00:13:50,360 --> 00:13:50,880
Yeah, I saw that.

482
00:13:51,120 --> 00:13:52,020
It went well.

483
00:13:52,020 --> 00:13:54,920
Then Matt Belez gave a quick five-minute prezo

484
00:13:54,920 --> 00:13:56,400
just a few of the things he's built

485
00:13:56,400 --> 00:13:59,380
that are related to agentic commerce and Bitcoin payments.

486
00:14:00,180 --> 00:14:02,600
And then Elias from the Spiral team,

487
00:14:02,640 --> 00:14:03,640
an engineer on the Spiral team,

488
00:14:04,420 --> 00:14:07,360
demonstrated what he's been by coding

489
00:14:07,360 --> 00:14:11,740
and a really cool knowledge search

490
00:14:11,740 --> 00:14:13,860
for Bitcoin technical stuff

491
00:14:13,860 --> 00:14:17,900
that has like an agentic interface and pretty powerful.

492
00:14:18,240 --> 00:14:20,480
So the only difference,

493
00:14:20,480 --> 00:14:23,260
We wanted to get to those presentations and then open it up,

494
00:14:23,460 --> 00:14:26,600
so we didn't really get much audience engagement during the presentations,

495
00:14:26,820 --> 00:14:29,540
but we were actually, unlike every other builder we've done,

496
00:14:29,640 --> 00:14:32,380
we kept it on the clock.

497
00:14:32,660 --> 00:14:33,960
We're usually pretty good here, too.

498
00:14:34,660 --> 00:14:37,880
Every presentation we have takes twice as long as we budget for,

499
00:14:38,100 --> 00:14:39,820
and that's not a knock.

500
00:14:40,100 --> 00:14:41,840
It's usually because there's great audience engagement,

501
00:14:41,960 --> 00:14:42,840
so it's not bad.

502
00:14:43,100 --> 00:14:45,980
It's just then we run out of time for other stuff.

503
00:14:46,540 --> 00:14:49,100
Here we had a good 45 minutes left.

504
00:14:49,100 --> 00:14:50,660
We got through the preseason 745.

505
00:14:51,020 --> 00:14:52,880
We had 45 minutes of discussion,

506
00:14:53,040 --> 00:14:53,760
and it was a great discussion.

507
00:14:54,020 --> 00:14:54,940
Lots of engagement.

508
00:14:56,280 --> 00:14:57,520
There's good pushback.

509
00:14:57,680 --> 00:15:00,100
Some people pushing on stablecoins,

510
00:15:01,800 --> 00:15:04,700
or pushing back on why will Bitcoin be adopted

511
00:15:04,700 --> 00:15:07,280
instead of credit cards and stablecoins.

512
00:15:07,720 --> 00:15:08,980
So it was really good.

513
00:15:11,020 --> 00:15:11,540
Sorry?

514
00:15:12,200 --> 00:15:13,280
Was Druckenmiller there?

515
00:15:13,460 --> 00:15:14,440
He was...

516
00:15:14,440 --> 00:15:16,560
I was hoping Jeremy Allaire would show up,

517
00:15:16,560 --> 00:15:18,200
but he did not.

518
00:15:18,200 --> 00:15:19,380
as far as I know.

519
00:15:19,480 --> 00:15:20,280
Maybe he's in the back room.

520
00:15:21,360 --> 00:15:21,600
Anyway,

521
00:15:22,240 --> 00:15:23,500
and then we,

522
00:15:23,660 --> 00:15:24,640
another goal

523
00:15:24,640 --> 00:15:25,640
beyond just putting on

524
00:15:25,640 --> 00:15:26,060
a good event

525
00:15:26,060 --> 00:15:26,900
and getting good engagement

526
00:15:26,900 --> 00:15:28,080
was to find

527
00:15:28,080 --> 00:15:29,640
one or more people

528
00:15:29,640 --> 00:15:30,300
who want to

529
00:15:30,300 --> 00:15:32,180
continue it

530
00:15:32,180 --> 00:15:33,000
in New York.

531
00:15:33,080 --> 00:15:33,900
Oh, yeah, yeah, exactly.

532
00:15:34,040 --> 00:15:35,160
And Haley and I had

533
00:15:35,160 --> 00:15:36,680
at least half a dozen people

534
00:15:36,680 --> 00:15:37,480
approach us afterwards

535
00:15:37,480 --> 00:15:38,860
and it looks like there's,

536
00:15:39,320 --> 00:15:40,860
it looks promising.

537
00:15:41,180 --> 00:15:41,600
That's great.

538
00:15:41,940 --> 00:15:42,380
That we'll,

539
00:15:42,380 --> 00:15:43,660
that we'll get someone.

540
00:15:43,660 --> 00:15:45,060
So we've obviously talked

541
00:15:45,060 --> 00:15:45,860
a lot about

542
00:15:45,860 --> 00:15:47,240
agentic commerce on the show.

543
00:15:47,620 --> 00:15:47,760
I mean,

544
00:15:47,760 --> 00:15:50,860
I think at this point, that's basically, it's like the Presidio agent payment show.

545
00:15:52,260 --> 00:15:59,200
What, if anything, was a new insight for you or did you change your mind on or was there

546
00:15:59,200 --> 00:16:02,680
anything that jumped out at you as like a new, I don't know, a new idea or a new takeaway?

547
00:16:06,240 --> 00:16:06,900
Not really.

548
00:16:07,060 --> 00:16:10,860
I mean, the main thing is we have put effort into putting this presentation together.

549
00:16:10,960 --> 00:16:12,880
I mean, this is something we've talked about on the show for a year.

550
00:16:13,140 --> 00:16:13,300
Yeah.

551
00:16:13,800 --> 00:16:16,860
And I've always felt it's a bit hand wavy.

552
00:16:16,860 --> 00:16:18,260
or when I say it's hand-wavy,

553
00:16:18,440 --> 00:16:22,100
meaning any of us can make a strong argument.

554
00:16:22,340 --> 00:16:23,300
Anyone at Presidio Bitcoin,

555
00:16:23,420 --> 00:16:25,960
anyone in Bitcoin can make a strong argument long-term.

556
00:16:26,220 --> 00:16:27,640
Why is Bitcoin the best money?

557
00:16:28,080 --> 00:16:31,980
Why is it the most neutral and credible and blah, blah, blah.

558
00:16:32,440 --> 00:16:35,660
But it's much harder to make the argument for like why now?

559
00:16:36,360 --> 00:16:38,440
Why should a normal person use it now

560
00:16:38,440 --> 00:16:40,800
versus stable coins or credit cards or whatever?

561
00:16:40,940 --> 00:16:41,820
It's a way harder argument.

562
00:16:42,360 --> 00:16:44,700
So when I say we've been a bit hand-wavy,

563
00:16:44,700 --> 00:16:50,040
we don't have a super strong tight argument.

564
00:16:50,480 --> 00:16:53,040
But we've been developing it on the show and in other venues.

565
00:16:53,980 --> 00:16:57,340
And then with Matt Biles putting this presentation together,

566
00:16:57,340 --> 00:17:00,920
and then I was going to get up in front of people and present it.

567
00:17:01,020 --> 00:17:02,520
We really wanted to make the best case.

568
00:17:02,740 --> 00:17:04,920
So I think that has strengthened over time.

569
00:17:05,020 --> 00:17:05,620
That's good news.

570
00:17:08,500 --> 00:17:11,260
But I'd still like to see it stronger.

571
00:17:11,500 --> 00:17:14,000
The why now still needs to be stronger.

572
00:17:14,000 --> 00:17:20,120
I mean, I still think the default scenario right now is that credit cards are going to,

573
00:17:21,240 --> 00:17:26,100
unless there's stuff that we find that they simply don't work,

574
00:17:26,640 --> 00:17:27,200
they're going to be.

575
00:17:27,360 --> 00:17:28,400
That's going to be the default.

576
00:17:28,540 --> 00:17:29,240
That's going to be used.

577
00:17:29,920 --> 00:17:34,020
And then clearly the stablecoin, like CoinCorps,

578
00:17:34,160 --> 00:17:37,400
are going to push, push, push, push their solutions.

579
00:17:39,160 --> 00:17:43,200
So there's likely going to be adoption there.

580
00:17:43,200 --> 00:17:45,680
and they solve some of the problems of credit cards.

581
00:17:45,820 --> 00:17:46,060
Yeah.

582
00:17:46,320 --> 00:17:47,920
Some of the problems that Bitcoin solves as well.

583
00:17:48,440 --> 00:17:50,780
So I still think we have a lot of work to do on that.

584
00:17:51,380 --> 00:17:54,760
The best thing we can do though is build, demonstrate,

585
00:17:55,640 --> 00:17:56,740
cast your vote.

586
00:17:57,340 --> 00:17:57,700
That's one thing.

587
00:17:58,460 --> 00:18:01,120
I say because that's one of the slides in the presentation

588
00:18:01,120 --> 00:18:04,960
is like every payment made you can think of as a vote.

589
00:18:05,620 --> 00:18:07,240
So if you're a seller or a buyer,

590
00:18:07,240 --> 00:18:08,940
if you're selling something,

591
00:18:09,460 --> 00:18:12,620
or you've got your buddy selling you tea or whatever.

592
00:18:12,840 --> 00:18:13,020
Honey.

593
00:18:13,200 --> 00:18:20,640
honey tea what you know yeah you see the case we were at 100 every payment they accept right or if

594
00:18:20,640 --> 00:18:25,500
they just decide to accept bitcoin that's kind of a vote but then every payment they accept is a vote

595
00:18:25,500 --> 00:18:33,640
and if we're buying stuff every payment we make in bitcoin is a vote right um yeah yeah did um

596
00:18:33,640 --> 00:18:37,860
i'm curious uh did nick have anything i actually didn't look at his presentation in the chat did

597
00:18:37,860 --> 00:18:43,080
he have anything new or of interest that you were surprised by and what was the audience's reaction

598
00:18:43,080 --> 00:18:45,080
had people tried OriBot? Were they excited

599
00:18:45,080 --> 00:18:46,460
to try a chatbot interface?

600
00:18:46,920 --> 00:18:48,480
Was the audience pretty AI-pilled?

601
00:18:48,720 --> 00:18:50,720
I don't know how much of a bubble we're in here.

602
00:18:50,840 --> 00:18:52,480
San Francisco. The audience,

603
00:18:52,740 --> 00:18:54,420
I think it was mostly

604
00:18:54,420 --> 00:18:57,180
Bitcoiners slash crypto folks.

605
00:18:59,260 --> 00:19:02,820
Nick did a nice job

606
00:19:02,820 --> 00:19:04,420
presenting. He covered

607
00:19:04,420 --> 00:19:06,100
MDK and Ori

608
00:19:06,100 --> 00:19:08,520
and then the

609
00:19:08,520 --> 00:19:10,940
unhuman stuff. So the stuff we already know.

610
00:19:10,940 --> 00:19:16,840
But he nicely covered all of it in less than six minutes.

611
00:19:20,000 --> 00:19:23,300
The audience, again, during the presentations,

612
00:19:23,480 --> 00:19:26,420
there was no questions or whatever.

613
00:19:26,620 --> 00:19:30,000
So at the end, it was mainly about just the general topic of agenda payments.

614
00:19:30,280 --> 00:19:32,760
I don't think there was much on his stuff.

615
00:19:32,920 --> 00:19:35,460
I mean, he certainly engaged in the discussion, though, and responded.

616
00:19:35,780 --> 00:19:37,380
But yeah, I mean, in terms of the audience,

617
00:19:37,380 --> 00:19:39,600
one of the questions is

618
00:19:39,600 --> 00:19:41,240
do people even think

619
00:19:41,240 --> 00:19:43,460
agentic payments? I mean, clearly there's a lot of

620
00:19:43,460 --> 00:19:44,940
hype around it the last two months.

621
00:19:45,180 --> 00:19:47,300
Last year when I asked AI people

622
00:19:47,300 --> 00:19:49,100
payments wasn't even on the radar.

623
00:19:50,100 --> 00:19:51,240
And the last two months

624
00:19:51,240 --> 00:19:53,580
it's dramatically changed in terms of hype.

625
00:19:53,940 --> 00:19:54,600
Yeah, for sure.

626
00:19:54,700 --> 00:19:56,120
But then the question still remains.

627
00:19:56,380 --> 00:19:59,560
I mean, that includes us, right?

628
00:19:59,640 --> 00:20:01,220
I mean, we're excited about the hype, but

629
00:20:01,220 --> 00:20:03,100
it still remains a question. Is it

630
00:20:03,100 --> 00:20:05,300
really taking off this year

631
00:20:05,300 --> 00:20:05,960
or soon?

632
00:20:05,960 --> 00:20:10,080
or is it hyped now and 10 years away?

633
00:20:10,360 --> 00:20:12,980
But the audience thought it is real.

634
00:20:13,400 --> 00:20:15,020
And so I want to come back to that question,

635
00:20:15,100 --> 00:20:15,700
what's real, what's not.

636
00:20:15,780 --> 00:20:17,540
But on the audience in particular,

637
00:20:18,040 --> 00:20:19,960
did you get the sense even without agentic payments,

638
00:20:19,960 --> 00:20:23,020
were people that came to this outside of our bubble,

639
00:20:23,460 --> 00:20:25,540
were they already vibe coding?

640
00:20:25,860 --> 00:20:28,420
Had they, for example, set up an open claw before

641
00:20:28,420 --> 00:20:30,720
or had they used Replit or CloudCode

642
00:20:30,720 --> 00:20:32,080
or something like OriBot

643
00:20:32,080 --> 00:20:34,020
or is this like a new world they wanted to try?

644
00:20:34,020 --> 00:20:36,940
I don't think I

645
00:20:36,940 --> 00:20:39,160
I don't think anyone asked that

646
00:20:39,160 --> 00:20:39,580
so

647
00:20:39,580 --> 00:20:41,460
but my

648
00:20:41,460 --> 00:20:43,980
intuition says

649
00:20:43,980 --> 00:20:44,940
yes

650
00:20:44,940 --> 00:20:47,000
that a lot of them had

651
00:20:47,000 --> 00:20:49,100
but maybe not

652
00:20:49,100 --> 00:20:50,360
trying agentic payments

653
00:20:50,360 --> 00:20:50,840
I'm not sure

654
00:20:50,840 --> 00:20:52,540
so on that point

655
00:20:52,540 --> 00:20:53,180
about what's real

656
00:20:53,180 --> 00:20:53,700
what's not

657
00:20:53,700 --> 00:20:54,460
I mean I think I saw

658
00:20:54,460 --> 00:20:55,120
one of you guys

659
00:20:55,120 --> 00:20:55,480
or someone

660
00:20:55,480 --> 00:20:55,980
maybe it was actually

661
00:20:55,980 --> 00:20:56,260
I don't know

662
00:20:56,260 --> 00:20:57,380
someone shared in our chat

663
00:20:57,380 --> 00:20:57,800
that

664
00:20:57,800 --> 00:20:59,020
you know

665
00:20:59,020 --> 00:20:59,780
trying to see

666
00:20:59,780 --> 00:21:00,540
what was hype

667
00:21:00,540 --> 00:21:01,740
versus what was real

668
00:21:01,740 --> 00:21:02,600
in agentic payments

669
00:21:02,600 --> 00:21:03,780
DK maybe that was you

670
00:21:03,780 --> 00:21:05,520
be like, what is y'all's sense right now?

671
00:21:05,580 --> 00:21:08,420
Like, obviously, everyone and their sister is talking about agent payments.

672
00:21:08,640 --> 00:21:12,520
Like, what is y'all's general feeling of, like, has anyone proven a use case for this

673
00:21:12,520 --> 00:21:12,720
yet?

674
00:21:12,800 --> 00:21:14,240
Is it still just, like, theoretical?

675
00:21:14,700 --> 00:21:16,180
Obviously, OpenClaw opened up the...

676
00:21:16,180 --> 00:21:22,280
Somebody shared something that kind of indicated that even, I think, X402, which claims very

677
00:21:22,280 --> 00:21:29,960
large volume, is maybe not as legit volume in use cases as maybe the perception may be

678
00:21:29,960 --> 00:21:30,720
ahead of reality there.

679
00:21:30,720 --> 00:21:34,660
yeah I think it's still very much speculation

680
00:21:34,660 --> 00:21:38,260
but I am quite bullish on agentic payments

681
00:21:38,260 --> 00:21:39,040
just the whole

682
00:21:39,040 --> 00:21:41,060
just what we have experienced

683
00:21:41,060 --> 00:21:43,220
with having an agent do stuff for you

684
00:21:43,220 --> 00:21:44,780
and then buy

685
00:21:44,780 --> 00:21:46,600
this is like obvious

686
00:21:46,600 --> 00:21:49,800
and the example we give in the presentation

687
00:21:49,800 --> 00:21:51,380
that I presented was

688
00:21:51,380 --> 00:21:53,180
buying a pair of shoes

689
00:21:53,180 --> 00:21:55,840
something relatable for everyone

690
00:21:55,840 --> 00:21:58,880
so it's not some super weird use case

691
00:21:58,880 --> 00:22:00,460
not like mailing a potato or something

692
00:22:00,460 --> 00:22:02,120
I love that

693
00:22:02,120 --> 00:22:03,320
buying a pair of shoes

694
00:22:03,320 --> 00:22:04,740
so everyone buys a pair of shoes

695
00:22:04,740 --> 00:22:06,820
and there's a range

696
00:22:06,820 --> 00:22:08,140
some people just like

697
00:22:08,140 --> 00:22:09,300
put no effort into it

698
00:22:09,300 --> 00:22:10,400
and just click buy

699
00:22:10,400 --> 00:22:11,040
or go to a store

700
00:22:11,040 --> 00:22:12,060
and say they look nice

701
00:22:12,060 --> 00:22:12,660
and buy it

702
00:22:12,660 --> 00:22:14,660
others do a bunch of research

703
00:22:14,660 --> 00:22:17,060
others do a lot of price comparison

704
00:22:17,060 --> 00:22:18,840
but no matter who you are

705
00:22:18,840 --> 00:22:20,380
if you just tell your agent

706
00:22:20,380 --> 00:22:22,300
who already knows about you

707
00:22:22,300 --> 00:22:22,640
and stuff

708
00:22:22,640 --> 00:22:24,320
and it just does all that work for you

709
00:22:24,320 --> 00:22:25,840
in a matter of seconds

710
00:22:25,840 --> 00:22:27,140
and purchases it for you

711
00:22:27,140 --> 00:22:28,100
and they just arrive

712
00:22:28,100 --> 00:22:29,680
like why would you not do that?

713
00:22:29,680 --> 00:22:33,880
And maybe they're using, you know, maybe you're using your credit card to do that.

714
00:22:34,340 --> 00:22:34,480
Right.

715
00:22:34,700 --> 00:22:38,340
But the notion of agentic payments seems obvious to me.

716
00:22:38,620 --> 00:22:44,180
And then, of course, then there are novel use cases as well, which are naturally speculative.

717
00:22:44,480 --> 00:22:52,380
But just traditional online retail, there's no question in my mind, agentic commerce will be a thing soon.

718
00:22:52,620 --> 00:22:52,820
Yeah.

719
00:22:52,820 --> 00:22:57,280
Because already a behavior I've developed, I'm not having the agent buy it for me,

720
00:22:57,280 --> 00:23:00,400
but I definitely do product research via AI now.

721
00:23:00,800 --> 00:23:03,760
And it is far superior to just Googling.

722
00:23:05,640 --> 00:23:08,780
And so I bought a number of things where I start with AI.

723
00:23:09,700 --> 00:23:11,420
It seems to find really good stuff.

724
00:23:11,820 --> 00:23:13,780
Since I'm still trying to develop trust,

725
00:23:14,580 --> 00:23:15,480
or trusting it,

726
00:23:15,580 --> 00:23:17,580
I then do my more traditional research

727
00:23:17,580 --> 00:23:18,700
with those products and stuff

728
00:23:18,700 --> 00:23:20,960
to see are they actually good

729
00:23:20,960 --> 00:23:22,820
and I've been impressed so far.

730
00:23:23,100 --> 00:23:25,160
And then I manually buy it.

731
00:23:25,160 --> 00:23:30,140
But yeah, if I can one-click or basically zero-click buy,

732
00:23:30,820 --> 00:23:32,580
why would I not do that?

733
00:23:32,900 --> 00:23:33,040
Totally.

734
00:23:33,180 --> 00:23:37,140
I think that's the future, but I am concerned about it.

735
00:23:37,140 --> 00:23:38,960
It feels like there's a popular narrative

736
00:23:38,960 --> 00:23:43,340
that somehow ads as a format are going to work really well

737
00:23:43,340 --> 00:23:54,745
in these LLM workflows And I think I kind of default a little skeptical of that because they so good and the information is so crafted toward what you actually want

738
00:23:54,745 --> 00:24:01,445
It doesn't like ads work when there's a little bit of like an arbitrage of attention between, hey, here's what we think you might like.

739
00:24:01,445 --> 00:24:05,965
And here's some other options that are paid, like the kind of the standard Google web search model.

740
00:24:05,965 --> 00:24:16,965
But I think when you move toward like an LLM giving you really high quality research, it almost feels like you lose some of the ability for there to be a ad monetization event there.

741
00:24:17,205 --> 00:24:21,805
So I think I'm kind of generally skeptical of how those things fit in.

742
00:24:21,925 --> 00:24:34,965
I'm optimistic that there's going to be a lot of opportunity in agentic commerce from like agents doing commerce with each other or on behalf of us, but not necessarily around fitting into that flow of like traditional commerce categories.

743
00:24:35,965 --> 00:24:38,225
But that's separate, though.

744
00:24:38,965 --> 00:24:42,425
If I use AI and there's no ads,

745
00:24:42,485 --> 00:24:43,725
I can do exactly what I just said.

746
00:24:43,985 --> 00:24:46,645
I can still have the agent do my product research

747
00:24:46,645 --> 00:24:47,665
and then buy it for me.

748
00:24:49,045 --> 00:24:50,565
So I agree with that.

749
00:24:50,625 --> 00:24:52,465
But I'm interested in your thoughts on the ads here.

750
00:24:52,725 --> 00:24:57,045
So is it your understanding that right now,

751
00:24:57,125 --> 00:24:59,145
let's say I was using OpenAI or something,

752
00:24:59,245 --> 00:25:01,085
and I was going to have it buy something,

753
00:25:01,865 --> 00:25:05,045
is the ad right now just being displayed

754
00:25:05,045 --> 00:25:06,945
or is the plan just to display it in the answer?

755
00:25:07,125 --> 00:25:08,805
Or are they already experimenting with,

756
00:25:08,885 --> 00:25:11,245
hey, if we're in a like one, you know, whatever,

757
00:25:11,365 --> 00:25:12,885
like one click buy this for you,

758
00:25:12,985 --> 00:25:14,085
we're just going to like buy us

759
00:25:14,085 --> 00:25:15,725
and wait towards who's paying us on the backend.

760
00:25:16,065 --> 00:25:16,805
And if that's the case,

761
00:25:16,805 --> 00:25:19,905
that seems like extremely against the user's best interest

762
00:25:19,905 --> 00:25:21,605
and something that's also really hard to audit.

763
00:25:21,925 --> 00:25:22,945
And like something I, I mean,

764
00:25:22,945 --> 00:25:23,865
you already know my thoughts on this,

765
00:25:23,925 --> 00:25:24,945
but I certainly wouldn't use that

766
00:25:24,945 --> 00:25:26,665
even more than I already think.

767
00:25:26,665 --> 00:25:27,285
Right, right.

768
00:25:27,725 --> 00:25:30,845
But it seems like if you are compromising

769
00:25:30,845 --> 00:25:33,845
with a backend payment from somebody

770
00:25:33,845 --> 00:25:35,905
and you're not giving the thing that's best for the user,

771
00:25:36,345 --> 00:25:38,205
you're actually hurting the whole LLM experience

772
00:25:38,205 --> 00:25:41,345
to the point that you might actually lose trust that way.

773
00:25:41,645 --> 00:25:43,025
Isn't that true for Google too, though?

774
00:25:43,825 --> 00:25:46,065
No, because I think Google is different

775
00:25:46,065 --> 00:25:48,325
because there's a world where there's the ability

776
00:25:48,325 --> 00:25:49,825
to arbitrage the person's attention

777
00:25:49,825 --> 00:25:51,065
on the search results page.

778
00:25:51,365 --> 00:25:52,465
The search results page says,

779
00:25:52,805 --> 00:25:54,465
hey, here's some stuff that we think is good.

780
00:25:54,805 --> 00:25:55,925
Here's some stuff that we were paid for

781
00:25:55,925 --> 00:25:56,965
that might be good too.

782
00:25:57,445 --> 00:25:58,585
And it's like up to you to pick.

783
00:25:59,225 --> 00:26:02,165
So it's presented as like, we don't really know.

784
00:26:02,165 --> 00:26:08,445
But in LLMs, it's presented as you want these things, you want the very best five things, here's the best five, right?

785
00:26:08,605 --> 00:26:14,545
So I hope, and I'm currently not, I don't think I'm seeing ads, and I hope I don't see ads.

786
00:26:14,745 --> 00:26:21,125
I am paying a monthly subscription, so I hope that would mean that I can be using a product that is ad-free.

787
00:26:21,565 --> 00:26:26,365
Or, you know, I mean, hopefully there's not just Google dominance, right?

788
00:26:26,465 --> 00:26:30,785
Hopefully there's several really good options, and maybe some have ads and others don't.

789
00:26:30,785 --> 00:26:33,985
but it would be a very sad day

790
00:26:33,985 --> 00:26:35,825
I think if there's only one company

791
00:26:35,825 --> 00:26:38,485
and there's ads and it degrades the experience

792
00:26:38,485 --> 00:26:40,105
yeah I don't see that ever happening

793
00:26:40,105 --> 00:26:41,185
and the good reason is

794
00:26:41,185 --> 00:26:42,825
as long as we have the open source models

795
00:26:42,825 --> 00:26:47,565
it's always been challenging for me

796
00:26:47,565 --> 00:26:50,345
because I understand Google is arguably the greatest business of all time

797
00:26:50,345 --> 00:26:51,985
and I've wondered my entire life

798
00:26:51,985 --> 00:26:52,825
who clicks these ads

799
00:26:52,825 --> 00:26:55,145
I've never clicked a Google ad that I know of in my entire life

800
00:26:55,145 --> 00:26:57,965
so I presume I see the revenue numbers

801
00:26:57,965 --> 00:26:59,105
that people do click these ads

802
00:26:59,105 --> 00:27:03,385
I can't, it's hard for me to empathize, but like, I, I, you know, I guess they're there.

803
00:27:03,865 --> 00:27:07,445
But, but, but when it comes to the enchanted commerce thing.

804
00:27:08,705 --> 00:27:11,105
DK worked on this one, like 2003.

805
00:27:11,625 --> 00:27:15,005
So do people, let's actually give it the truth.

806
00:27:16,025 --> 00:27:16,825
Google truth.

807
00:27:17,105 --> 00:27:19,305
Is it all alive?

808
00:27:19,825 --> 00:27:22,885
Max, if you had to go replace that wig, where, where, how would you do that?

809
00:27:23,165 --> 00:27:24,065
If I wanted to replace what?

810
00:27:24,485 --> 00:27:24,905
Your wig.

811
00:27:25,325 --> 00:27:27,645
Oh, no mommies, man.

812
00:27:27,645 --> 00:27:33,645
well first i would have to buy the uh uh the whoopee goldberg uh

813
00:27:33,645 --> 00:27:38,565
mask that we had from the pbj view um i would start there

814
00:27:38,565 --> 00:27:45,205
but um yeah no i mean it's like i mean to be clear this was the whole view i mean i've talked

815
00:27:45,205 --> 00:27:49,345
to you guys about that company that i love uh coggy which the whole brand for them was like look

816
00:27:49,345 --> 00:27:54,345
you know and if you go back to um larry and sergey's original paper they even spell it out

817
00:27:54,345 --> 00:27:58,905
in there. They say, you know, the moment you introduce advertisement, like it's, it's, you

818
00:27:58,905 --> 00:28:03,925
know, it's a long march toward death because it's like the frog boiling. Like I understand your point,

819
00:28:03,985 --> 00:28:09,125
the arbitrage, as long as you can show two options of like just the pure best result and the, uh,

820
00:28:09,125 --> 00:28:13,865
the ad, then, you know, it's, it's, it's, it's not going from zero to like horrible. It's like a

821
00:28:13,865 --> 00:28:18,445
slight degree, you know, shift, but like there's still limited attention and ad space, right? So,

822
00:28:18,725 --> 00:28:22,385
you know, let's say you're looking for the best answer and you start with the 10 best answers.

823
00:28:22,385 --> 00:28:24,425
Then you say, well, it's the nine best answers and one promoted.

824
00:28:24,545 --> 00:28:25,945
And then the eight best answers and two promoted.

825
00:28:26,165 --> 00:28:29,105
Like over time, that experience still gets worse and worse,

826
00:28:29,385 --> 00:28:31,285
which I would argue up until the LLM boom

827
00:28:31,285 --> 00:28:32,425
was what people were writing about

828
00:28:32,425 --> 00:28:33,925
was sort of like the enchantification of Google.

829
00:28:34,025 --> 00:28:38,125
Like I stopped using Google before the LLM boom

830
00:28:38,125 --> 00:28:39,365
and was starting to use things like Koggy.

831
00:28:39,465 --> 00:28:40,805
But now that we're getting the LLMs,

832
00:28:41,145 --> 00:28:42,265
I'm going back to the LLMs.

833
00:28:42,425 --> 00:28:43,205
But I mean...

834
00:28:43,205 --> 00:28:44,265
Quick story on that, actually.

835
00:28:44,545 --> 00:28:46,945
21 years ago, 2005, I was at Yahoo Search.

836
00:28:46,965 --> 00:28:47,345
Special number.

837
00:28:47,405 --> 00:28:47,905
I was on the Yahoo Search.

838
00:28:47,905 --> 00:28:48,525
You were at Yahoo?

839
00:28:48,525 --> 00:28:52,405
It was my MBA internship.

840
00:28:53,065 --> 00:28:54,545
Were there of you guys a dog pile?

841
00:28:55,285 --> 00:28:56,405
Ask G's?

842
00:28:56,925 --> 00:28:57,045
Yeah.

843
00:28:59,745 --> 00:29:00,825
I love Ask G's.

844
00:29:00,845 --> 00:29:03,685
2005, Yahoo was still like a power.

845
00:29:05,105 --> 00:29:10,265
I think a lot of, I mean, people in Silicon Valley knew that Google was going to dominate,

846
00:29:10,565 --> 00:29:12,625
but the rest of the world, I don't think, didn't.

847
00:29:12,625 --> 00:29:22,825
But anyway, but yeah, so from being in the inside there, Yahoo showed a ton of ads in the search results page.

848
00:29:23,105 --> 00:29:24,765
And Google showed far fewer.

849
00:29:25,665 --> 00:29:27,345
And I saw the internal debates there.

850
00:29:27,425 --> 00:29:31,785
It was basically next to impossible for them to reduce the number of ads they showed.

851
00:29:32,225 --> 00:29:32,405
Why?

852
00:29:33,085 --> 00:29:37,645
Because they were having discussions, how do we gain market share on Google?

853
00:29:38,285 --> 00:29:40,385
Well, one, it's to create a better user experience, show less ads.

854
00:29:40,385 --> 00:29:40,925
So why?

855
00:29:40,925 --> 00:29:48,825
Well, because you can 100% with 100% certainty quantify the revenue loss that quarter if you remove ads.

856
00:29:49,505 --> 00:29:57,105
And you can't quantify, or it's way more difficult, a leap of faith, like how much market shares it's going to give us by improving the UX.

857
00:29:57,245 --> 00:29:58,785
And Yahoo is already a public company at this point?

858
00:29:59,485 --> 00:29:59,625
Yeah.

859
00:29:59,945 --> 00:30:00,245
Oh, yeah.

860
00:30:00,845 --> 00:30:09,965
So, and sadly, now Google shows probably even more ads than Yahoo did at the time.

861
00:30:09,965 --> 00:30:13,405
Like it's insane how many ads are shown now on a Google search results page.

862
00:30:13,885 --> 00:30:14,005
Yeah.

863
00:30:14,245 --> 00:30:15,585
It'd be interesting for someone.

864
00:30:15,825 --> 00:30:19,045
I mean, I don't know like what kind of revenue Yahoo is doing, but it'd be cool to like maybe

865
00:30:19,045 --> 00:30:23,965
buy Yahoo and bring it back.

866
00:30:24,085 --> 00:30:24,405
Bring it back.

867
00:30:24,405 --> 00:30:27,785
Like you want to personally just chip in or pass the hat or what?

868
00:30:28,185 --> 00:30:28,545
Exactly.

869
00:30:28,865 --> 00:30:30,545
I'm going to pass that one to you, DK.

870
00:30:30,805 --> 00:30:32,825
But I'll, you know, it couldn't be that much.

871
00:30:34,265 --> 00:30:34,625
No.

872
00:30:34,625 --> 00:30:38,705
So, but, but, but I, so I, I am, I am very curious about where this is going to go in

873
00:30:38,705 --> 00:30:42,245
the long run because i mean and also what i remember in the early days of google is you all

874
00:30:42,245 --> 00:30:46,045
said i'm feeling lucky right which was just like i click one thing and i get the best answer so to

875
00:30:46,045 --> 00:30:51,985
be clear vision of what an llm is that's what i'm saying like you you go back and listen to larry

876
00:30:51,985 --> 00:30:58,485
you know talk in like 99 like he was ready for ai and llms way back then this is an ai company yeah

877
00:30:58,485 --> 00:31:03,045
exactly and like that dude was just like right and like probably doesn't get nearly enough respect

878
00:31:03,045 --> 00:31:08,685
for like all of his vision over the years um but but again so i mean i'm curious you guys are

879
00:31:08,685 --> 00:31:13,905
there on this time like so in that world you know both he and sergey acknowledged in their paper

880
00:31:13,905 --> 00:31:19,545
that ads were eventually going to make the product bad ads happen to work very well he clearly wants

881
00:31:19,545 --> 00:31:25,045
it to be a one-shot ai company like what what were their other ideas for monetization and like are

882
00:31:25,045 --> 00:31:31,245
some of those going to come back in the you know the lm wave well the the other they had several i

883
00:31:31,245 --> 00:31:36,345
think there were three that were in the venture pitch which was kind of early i'm not sure if i'll

884
00:31:36,345 --> 00:31:43,125
get them all right but there was the um the google search appliance was one of their ideas for how to

885
00:31:43,125 --> 00:31:46,845
make money it was to sell basically enterprise search oh that was like an oven in your house

886
00:31:46,845 --> 00:31:51,865
yeah yeah and that was that was like you know a reasonable line of business that eventually kicked

887
00:31:51,865 --> 00:31:56,745
off what is now you know google workspace and all the you know enterprise cloud stuff so it's kind of

888
00:31:56,745 --> 00:32:01,625
like a little bit of genesis there well it's funny about that sorry to interrupt but that that but

889
00:32:01,625 --> 00:32:04,025
when I joined Google,

890
00:32:04,345 --> 00:32:07,685
I was blown away by how horrible

891
00:32:07,685 --> 00:32:09,525
enterprise search was at Google.

892
00:32:09,825 --> 00:32:11,445
It was like impossible to find stuff.

893
00:32:11,505 --> 00:32:12,245
How is that possible?

894
00:32:12,465 --> 00:32:13,505
It was horrible.

895
00:32:13,985 --> 00:32:15,105
Enterprise search was horrible.

896
00:32:15,445 --> 00:32:16,025
How does it get worse?

897
00:32:16,585 --> 00:32:18,825
The thing that makes Google great

898
00:32:18,825 --> 00:32:20,805
is because it has the whole web

899
00:32:20,805 --> 00:32:21,805
cached in memory.

900
00:32:22,185 --> 00:32:22,365
Right.

901
00:32:22,445 --> 00:32:23,845
And so you have tons of interconnect

902
00:32:23,845 --> 00:32:25,185
between pages,

903
00:32:25,345 --> 00:32:26,725
referencing pages and information,

904
00:32:26,985 --> 00:32:28,605
linkages and terms.

905
00:32:28,605 --> 00:32:31,085
And enterprise doesn't have that much content

906
00:32:31,085 --> 00:32:32,565
and doesn't have that much interlinking.

907
00:32:32,765 --> 00:32:34,445
So if you isolate the index

908
00:32:34,445 --> 00:32:36,605
to just all the stuff contained within an enterprise,

909
00:32:37,105 --> 00:32:38,445
you actually don't have enough signal

910
00:32:38,445 --> 00:32:39,885
to build a really great search,

911
00:32:39,965 --> 00:32:40,905
even if you use the same algorithm.

912
00:32:41,305 --> 00:32:42,805
And the irony is, I guess back then,

913
00:32:42,845 --> 00:32:43,425
that's the case,

914
00:32:43,445 --> 00:32:44,925
is like you actually still needed to like

915
00:32:44,925 --> 00:32:46,525
be on the web to get that value.

916
00:32:46,685 --> 00:32:48,665
Now you can potentially with LLMs,

917
00:32:48,725 --> 00:32:51,385
you can condense that value from the web

918
00:32:51,385 --> 00:32:53,065
and then potentially transfer it

919
00:32:53,065 --> 00:32:54,665
with fine tuning or something just to the enterprise.

920
00:32:54,825 --> 00:32:56,485
So it would work much better now.

921
00:32:56,885 --> 00:32:58,045
Yeah, 100%, yeah.

922
00:32:58,805 --> 00:33:00,125
That'd be a great, yeah.

923
00:33:00,125 --> 00:33:02,805
And I think that there's a bunch of companies doing stuff like that.

924
00:33:03,105 --> 00:33:05,505
Like, I think Glean does stuff like that.

925
00:33:05,605 --> 00:33:06,105
They're really good.

926
00:33:06,565 --> 00:33:06,665
Yeah.

927
00:33:07,045 --> 00:33:08,045
Okay, back to your list.

928
00:33:08,665 --> 00:33:09,745
What's number two and three?

929
00:33:10,445 --> 00:33:11,205
Oh, geez.

930
00:33:12,305 --> 00:33:14,685
There was the search, the ads.

931
00:33:15,025 --> 00:33:28,785
Like, there was enterprise search, there was ads, and then, I don't know that I, oh, I think site search was the third one, which was if you owned a web property and you wanted to enable public search on your web property,

932
00:33:28,785 --> 00:33:30,265
you could pay

933
00:33:30,265 --> 00:33:31,045
you know

934
00:33:31,045 --> 00:33:32,265
Google to basically

935
00:33:32,265 --> 00:33:33,505
run software as a service

936
00:33:33,505 --> 00:33:34,125
like search

937
00:33:34,125 --> 00:33:35,305
as a software service

938
00:33:35,305 --> 00:33:36,525
and

939
00:33:36,525 --> 00:33:37,385
you know if you run

940
00:33:37,385 --> 00:33:37,665
like

941
00:33:37,665 --> 00:33:39,005
I guess the vision today

942
00:33:39,005 --> 00:33:39,405
would be like

943
00:33:39,405 --> 00:33:40,505
if Amazon were to

944
00:33:40,505 --> 00:33:41,305
outsource their

945
00:33:41,305 --> 00:33:42,925
search engine to Google

946
00:33:42,925 --> 00:33:44,145
and make the public

947
00:33:44,145 --> 00:33:45,485
search box on Amazon

948
00:33:45,485 --> 00:33:46,225
powered by Google

949
00:33:46,225 --> 00:33:46,885
I think that

950
00:33:46,885 --> 00:33:48,085
was called

951
00:33:48,085 --> 00:33:49,125
site search at the time

952
00:33:49,125 --> 00:33:49,765
I think that was

953
00:33:49,765 --> 00:33:51,445
the third line of business

954
00:33:51,445 --> 00:33:51,945
that was

955
00:33:51,945 --> 00:33:52,705
contemplated

956
00:33:52,705 --> 00:33:54,345
like AdSense

957
00:33:54,345 --> 00:33:55,205
was not

958
00:33:55,205 --> 00:33:56,245
part of the original

959
00:33:56,245 --> 00:33:58,385
VC deck

960
00:33:58,385 --> 00:34:00,485
I don't think so.

961
00:34:00,605 --> 00:34:01,705
Isn't the history,

962
00:34:02,185 --> 00:34:03,305
Yahoo bought,

963
00:34:03,685 --> 00:34:05,185
like Overture was created

964
00:34:05,185 --> 00:34:10,385
and was the genesis of the AdSense type ad model.

965
00:34:10,925 --> 00:34:11,885
Yahoo bought them

966
00:34:11,885 --> 00:34:13,785
and then Google created their own,

967
00:34:14,045 --> 00:34:16,025
created AdSense in-house.

968
00:34:16,985 --> 00:34:18,205
Is that accurate?

969
00:34:18,265 --> 00:34:22,545
Yeah, AdWords was the closest equivalent, I'd say.

970
00:34:22,625 --> 00:34:24,825
And then AdSense was more like

971
00:34:24,825 --> 00:34:27,365
for looking at different publisher properties

972
00:34:27,365 --> 00:34:31,745
and serving ads not against necessarily like keyword queries,

973
00:34:31,965 --> 00:34:33,385
but against publisher properties.

974
00:34:33,545 --> 00:34:35,405
So you could have AdSense for like a blog.

975
00:34:35,705 --> 00:34:38,565
You could have AdSense for domains in case you typed in the domain

976
00:34:38,565 --> 00:34:40,785
that wanted to serve a relevant ad.

977
00:34:40,925 --> 00:34:44,885
But the kind of contextual understanding of what is going on

978
00:34:44,885 --> 00:34:48,165
and putting up a keyword targeted ad based on that concept was,

979
00:34:48,665 --> 00:34:51,465
at least that's my recollection of AdSense,

980
00:34:51,985 --> 00:34:53,705
not really part of the venture pitch.

981
00:34:53,705 --> 00:34:56,185
So do you guys remember when it became clear

982
00:34:56,185 --> 00:34:57,825
that ads was like the runaway business.

983
00:34:57,945 --> 00:34:58,765
But like, when y'all were there,

984
00:34:58,885 --> 00:35:00,285
like, was that clear at that point?

985
00:35:00,405 --> 00:35:01,005
Or was it still?

986
00:35:01,585 --> 00:35:02,645
It was very clear about that.

987
00:35:02,645 --> 00:35:03,205
It was very clear, yeah.

988
00:35:04,005 --> 00:35:04,285
Interesting.

989
00:35:04,585 --> 00:35:05,065
Internally, though,

990
00:35:05,085 --> 00:35:05,745
but they tried to make,

991
00:35:06,085 --> 00:35:07,565
not make a big deal out of it

992
00:35:07,565 --> 00:35:10,245
because you didn't want to invite competition.

993
00:35:10,885 --> 00:35:11,205
Interesting.

994
00:35:11,505 --> 00:35:12,045
Very interesting.

995
00:35:12,665 --> 00:35:14,345
So how do you think that translates?

996
00:35:14,485 --> 00:35:15,525
So, okay, so, you know,

997
00:35:15,605 --> 00:35:17,965
obviously they knew in the long run, though,

998
00:35:18,025 --> 00:35:19,025
that that was potentially,

999
00:35:19,245 --> 00:35:20,325
or they knew that that would definitely

1000
00:35:20,325 --> 00:35:21,645
incentivize the product in the long run.

1001
00:35:21,985 --> 00:35:23,545
Now we're entering into a new age,

1002
00:35:23,685 --> 00:35:25,545
you know, some people are hyped on ads.

1003
00:35:25,545 --> 00:35:26,685
I tend to agree with you.

1004
00:35:26,885 --> 00:35:27,765
Like, I mean, again,

1005
00:35:27,885 --> 00:35:29,765
it's hard for me to evaluate these things

1006
00:35:29,765 --> 00:35:31,205
because I know for me as a user,

1007
00:35:31,285 --> 00:35:33,205
I was willing to pay for Kagi in the first place, right?

1008
00:35:33,305 --> 00:35:35,045
So there's no world in which I'm going to use

1009
00:35:35,045 --> 00:35:36,325
an ad product for my search.

1010
00:35:37,145 --> 00:35:38,905
And if I had to, then like, you know,

1011
00:35:38,925 --> 00:35:40,685
and I'm increasingly bullish on open source models

1012
00:35:40,685 --> 00:35:42,765
running on stuff like Maple or Vora or whatever.

1013
00:35:43,365 --> 00:35:44,725
And that's really exciting to me.

1014
00:35:45,045 --> 00:35:46,225
But like, I also understand

1015
00:35:46,225 --> 00:35:47,245
most people shouldn't go with the default.

1016
00:35:47,825 --> 00:35:49,965
So, I mean, one,

1017
00:35:50,485 --> 00:35:53,665
as long as the LLM is showing you responses,

1018
00:35:53,665 --> 00:35:58,325
you could in theory still have like hey here are some you know suggested things like here's some

1019
00:35:58,325 --> 00:36:03,585
links by the way this hyperlink is an ad and like whatever tell people about that but as you start

1020
00:36:03,585 --> 00:36:08,905
having the llm become or the ai become more agentec and just doing stuff on you on your behalf

1021
00:36:08,905 --> 00:36:14,025
then yeah essentially like if it's going to buy something it either they get a cut back or they

1022
00:36:14,025 --> 00:36:18,525
don't and if they do how do they show that to you and how do you audit that like where do you see

1023
00:36:18,525 --> 00:36:23,285
the business models going so we already have subscription right which i would love if that's

1024
00:36:23,285 --> 00:36:24,145
the way that web works.

1025
00:36:24,205 --> 00:36:25,525
And I have a question about that,

1026
00:36:25,625 --> 00:36:26,225
but real quick.

1027
00:36:26,465 --> 00:36:27,465
And then you can imagine

1028
00:36:27,465 --> 00:36:29,365
taking a cut of e-commerce

1029
00:36:29,365 --> 00:36:32,025
and there's ads.

1030
00:36:33,105 --> 00:36:35,785
So do either of you know

1031
00:36:35,785 --> 00:36:37,905
the data on subscriptions?

1032
00:36:38,505 --> 00:36:41,505
Like what percentage of OpenAI

1033
00:36:41,505 --> 00:36:44,325
or Google or Anthropoc users

1034
00:36:44,325 --> 00:36:45,445
are using free

1035
00:36:45,445 --> 00:36:47,785
versus upgrading to a subscription?

1036
00:36:48,265 --> 00:36:51,825
And does it skew

1037
00:36:51,825 --> 00:36:54,825
and how does it compare to historical services?

1038
00:36:55,005 --> 00:36:57,085
But in any case, what's definitely true,

1039
00:36:57,505 --> 00:37:00,325
like with search, there's no subscription.

1040
00:37:00,705 --> 00:37:02,505
I mean, no one's paying a subscription for search.

1041
00:37:02,585 --> 00:37:05,325
Well, I mean, the nerds on Hacker News were starting to,

1042
00:37:05,445 --> 00:37:07,505
and there was like 40,000 of us, but yes.

1043
00:37:08,285 --> 00:37:09,265
No one, basically.

1044
00:37:09,905 --> 00:37:12,005
Meaning it's like a rounding error.

1045
00:37:12,005 --> 00:37:17,145
Meaning, I don't know what the data is for Cloud and ChatGPD,

1046
00:37:17,145 --> 00:37:20,385
but it's more than 40,000 people

1047
00:37:20,385 --> 00:37:21,965
are paying $20 a month.

1048
00:37:23,265 --> 00:37:26,285
So I think it's encouraging

1049
00:37:26,285 --> 00:37:29,505
because it is changing people's behavior.

1050
00:37:30,745 --> 00:37:32,425
Because I think rationally,

1051
00:37:33,085 --> 00:37:36,105
if you were forced,

1052
00:37:36,385 --> 00:37:38,365
like most people have used Google for free,

1053
00:37:38,625 --> 00:37:41,085
but if you were forced to say,

1054
00:37:41,245 --> 00:37:44,585
hey, you're going to pay $100 a year

1055
00:37:44,585 --> 00:37:46,725
subscription for Google,

1056
00:37:46,725 --> 00:37:48,085
or it's taken away.

1057
00:37:48,225 --> 00:37:50,105
You can't search the internet anymore.

1058
00:37:50,705 --> 00:37:52,145
Is that going to change behavior

1059
00:37:52,145 --> 00:37:53,565
if that actually was the proposition?

1060
00:37:53,845 --> 00:37:55,425
Of course, tons of people will subscribe.

1061
00:37:56,605 --> 00:37:57,925
But because of the ad model

1062
00:37:57,925 --> 00:37:58,905
and they make so much money,

1063
00:37:59,385 --> 00:38:01,685
they weren't forced to do that.

1064
00:38:03,205 --> 00:38:05,005
And so we're just stuck in this world

1065
00:38:05,005 --> 00:38:06,445
of almost everything is an ad model.

1066
00:38:06,565 --> 00:38:08,985
I think it's a way healthier environment

1067
00:38:08,985 --> 00:38:11,785
if there's many competing models.

1068
00:38:12,625 --> 00:38:13,325
And so I think it's incredible.

1069
00:38:13,325 --> 00:38:13,965
Remember at the dawn of the internet,

1070
00:38:13,965 --> 00:38:16,325
there was no concept of putting

1071
00:38:16,325 --> 00:38:22,165
your credit card into a website yeah right so like that that you needed htps and then you needed

1072
00:38:22,165 --> 00:38:27,085
ssl you know you needed a whole ecosystem of people getting comfortable with that i mean that

1073
00:38:27,085 --> 00:38:31,445
was amazon's biggest fight for so long was you got to put your credit card into a website that's

1074
00:38:31,445 --> 00:38:35,985
bonkers yeah i still think it's bonkers and so people but you said and you also not only the

1075
00:38:35,985 --> 00:38:40,005
technology but like you said people had to get comfortable so it's the same like psychologically

1076
00:38:40,005 --> 00:38:45,805
and here it's psychological too um if stuff's given to you for free it's a big leap to start

1077
00:38:45,805 --> 00:38:47,925
paying for it, even if in the end, if you started

1078
00:38:47,925 --> 00:38:49,885
doing it, you'd be like, oh, I like the trade-off.

1079
00:38:50,685 --> 00:38:52,085
To change behavior is hard,

1080
00:38:52,205 --> 00:38:53,945
but I think that people paying

1081
00:38:53,945 --> 00:38:55,885
subscriptions for AI, there's

1082
00:38:55,885 --> 00:38:56,865
so much value there

1083
00:38:56,865 --> 00:38:59,945
that that's changing behavior, and that's going to be

1084
00:38:59,945 --> 00:39:01,885
good for all kinds of future services

1085
00:39:01,885 --> 00:39:03,185
that people are comfortable paying.

1086
00:39:03,345 --> 00:39:05,985
I agree with you. The good news is

1087
00:39:05,985 --> 00:39:07,765
we should definitely pull this data,

1088
00:39:07,885 --> 00:39:10,045
but anecdotally in my own life and everyone I know,

1089
00:39:10,045 --> 00:39:11,945
even though I did pay

1090
00:39:11,945 --> 00:39:14,125
for Kagi before, and a few

1091
00:39:14,125 --> 00:39:14,945
other things, eventually,

1092
00:39:14,945 --> 00:39:18,565
I was many years refusing even to pay for like Spotify or whatever.

1093
00:39:18,865 --> 00:39:22,665
But now I have multiple LLMs to the point where I'm like,

1094
00:39:22,765 --> 00:39:23,765
shit, how do I cancel this stuff?

1095
00:39:23,885 --> 00:39:25,945
Like this got out of control way too quickly.

1096
00:39:26,805 --> 00:39:29,545
And everyone I know is paying for at least one.

1097
00:39:30,025 --> 00:39:31,545
Multiple people are paying for many.

1098
00:39:31,985 --> 00:39:35,365
So there's no question that the sort of like Overton window of like,

1099
00:39:35,445 --> 00:39:37,045
hey, is the web free or not is shifted.

1100
00:39:37,345 --> 00:39:39,365
That seems to me generally a positive thing.

1101
00:39:39,745 --> 00:39:40,525
Having said that, I mean,

1102
00:39:40,545 --> 00:39:42,125
something we've talked a lot about on the show as well is like,

1103
00:39:42,165 --> 00:39:44,725
that also is dangerous in the sense that it creates like a very two-tiered

1104
00:39:44,725 --> 00:39:46,865
world very quickly. It's like, there are those people

1105
00:39:46,865 --> 00:39:48,605
that can pay and then those who can't.

1106
00:39:48,705 --> 00:39:50,685
If you're in a developing world or whatever,

1107
00:39:51,205 --> 00:39:52,785
now to be fair, I mean, just like Spotify

1108
00:39:52,785 --> 00:39:54,245
and stuff, I'm sure they'll price it differently.

1109
00:39:55,805 --> 00:39:57,085
But yeah.

1110
00:39:57,345 --> 00:39:58,825
And it raises the open

1111
00:39:58,825 --> 00:40:01,185
versus closed model discussion.

1112
00:40:01,425 --> 00:40:03,065
And if that gap closes,

1113
00:40:03,385 --> 00:40:04,805
then it makes it more accessible

1114
00:40:04,805 --> 00:40:06,645
to people. And it's also, and it kind of

1115
00:40:06,645 --> 00:40:08,625
comes back to this idea for me,

1116
00:40:08,825 --> 00:40:10,625
I think we talked about one or two weeks ago, where it's like,

1117
00:40:10,625 --> 00:40:12,565
you should always have

1118
00:40:12,565 --> 00:40:14,685
sort of the open router or whatever,

1119
00:40:14,725 --> 00:40:15,785
where you can go and say,

1120
00:40:15,865 --> 00:40:16,285
let me just,

1121
00:40:16,665 --> 00:40:17,705
even if it's just a test

1122
00:40:17,705 --> 00:40:18,425
every once in a while,

1123
00:40:18,545 --> 00:40:19,225
lots of queries

1124
00:40:19,225 --> 00:40:20,665
to see what their responses are.

1125
00:40:21,025 --> 00:40:22,685
Because one thing I'm concerned about is,

1126
00:40:22,785 --> 00:40:22,945
you know,

1127
00:40:22,965 --> 00:40:23,745
as I was telling DK

1128
00:40:23,745 --> 00:40:24,505
in our chat this morning,

1129
00:40:24,865 --> 00:40:25,245
some of these,

1130
00:40:25,405 --> 00:40:26,385
none of these companies,

1131
00:40:26,465 --> 00:40:27,505
I don't trust any of these companies.

1132
00:40:27,805 --> 00:40:29,385
Some of them I really don't trust.

1133
00:40:29,845 --> 00:40:30,505
And, you know,

1134
00:40:30,605 --> 00:40:31,205
on principle,

1135
00:40:31,205 --> 00:40:32,005
I don't want to get hooked

1136
00:40:32,005 --> 00:40:32,625
in their ecosystem

1137
00:40:32,625 --> 00:40:33,805
even if it's very good at first

1138
00:40:33,805 --> 00:40:34,625
because I know that trick.

1139
00:40:35,005 --> 00:40:35,305
It's like,

1140
00:40:35,305 --> 00:40:36,525
it's like Sapbot always does

1141
00:40:36,525 --> 00:40:37,305
in our chat where he,

1142
00:40:37,385 --> 00:40:37,685
you know,

1143
00:40:37,705 --> 00:40:38,525
he does that little emoji

1144
00:40:38,525 --> 00:40:39,785
where there's like cheese in a box

1145
00:40:39,785 --> 00:40:41,285
and he has a little stick holding up.

1146
00:40:41,525 --> 00:40:41,865
It's like, bro,

1147
00:40:41,865 --> 00:40:42,945
I'm not eating your cheese.

1148
00:40:43,025 --> 00:40:44,325
I don't care how good that cheese is.

1149
00:40:44,425 --> 00:40:45,645
You're going to pull the stick on me.

1150
00:40:45,705 --> 00:40:46,185
I know it.

1151
00:40:46,665 --> 00:40:49,925
And, you know, that's why it's so important we have, you know,

1152
00:40:50,105 --> 00:40:55,665
even if the web kind of goes this way where a few big providers are getting you to pay for it,

1153
00:40:55,685 --> 00:40:56,485
it seems good at first.

1154
00:40:56,485 --> 00:41:00,505
You have to have that open stack of, you know, open source models,

1155
00:41:00,665 --> 00:41:02,125
ideally Bitcoin, payment rails, et cetera.

1156
00:41:02,205 --> 00:41:04,865
Because otherwise, yeah, we're all just going to end up in the box eating our cheese.

1157
00:41:05,625 --> 00:41:07,185
Yeah, so I wonder how it'll play out.

1158
00:41:07,305 --> 00:41:09,685
Because like with search, I mean, there are a number, you know,

1159
00:41:09,685 --> 00:41:11,365
you mentioned what Excite and stuff

1160
00:41:11,365 --> 00:41:14,565
and AltaVista was like the big one in 98, 99.

1161
00:41:14,865 --> 00:41:15,765
Ask Jeeves and Dogpile.

1162
00:41:15,865 --> 00:41:16,585
Those are my go-tos.

1163
00:41:16,985 --> 00:41:18,585
And then, but once, I mean,

1164
00:41:18,625 --> 00:41:21,205
Google is so transformational that, you know,

1165
00:41:21,325 --> 00:41:25,545
it just, it won and then it's still winning today, right?

1166
00:41:25,805 --> 00:41:28,245
And there's really not been competition in the search space.

1167
00:41:28,485 --> 00:41:30,485
But if you look at other things like internet browsers,

1168
00:41:30,965 --> 00:41:33,065
that was more, a little more dynamic.

1169
00:41:33,845 --> 00:41:37,365
Started with Netscape, IE, then crushed them.

1170
00:41:38,605 --> 00:41:39,445
And Chrome crushed them.

1171
00:41:39,445 --> 00:41:41,605
Well, and at the time, though, it really seemed like,

1172
00:41:41,605 --> 00:41:44,465
it just seemed like IE was going to be like Google,

1173
00:41:44,685 --> 00:41:47,765
where like for 30 years or whatever, it's just going to be IE.

1174
00:41:47,905 --> 00:41:53,945
And then Firefox actually was able to eat some market share.

1175
00:41:55,705 --> 00:42:00,325
And Google helped a lot with generating revenue for Firefox.

1176
00:42:01,805 --> 00:42:04,765
But then, yeah, then Chrome launched and then it dominated.

1177
00:42:05,045 --> 00:42:07,365
And Safari has a lot of market share, too.

1178
00:42:07,365 --> 00:42:11,325
So it's definitely a healthier market compared to Surge.

1179
00:42:11,645 --> 00:42:11,685
Yeah.

1180
00:42:12,025 --> 00:42:16,425
So maybe that's a good analogy for how these big AI models will play out,

1181
00:42:16,505 --> 00:42:19,345
and there just won't be one dominant one.

1182
00:42:19,645 --> 00:42:20,845
But it'll be kind of like it is today.

1183
00:42:21,225 --> 00:42:22,705
And they are jockeying for position, right?

1184
00:42:22,785 --> 00:42:25,085
I mean, ChatGPT had most of the market share,

1185
00:42:25,185 --> 00:42:27,845
but Gemini is really encroaching on it now.

1186
00:42:27,945 --> 00:42:34,285
And then Claude is just, Opus models are really, really, really good.

1187
00:42:34,285 --> 00:42:46,405
I also, just as a pause point, I mean, DK, on last week's show, you got me thinking more about using Anthropic or Claude as a sort of thought partner because I was primarily using Geminis Melee Driver.

1188
00:42:46,765 --> 00:42:52,745
And I was using for building stuff, which I'm probably doing less than you are, but I was using OriBot, which is a wrapped Claude.

1189
00:42:53,105 --> 00:42:56,745
But I've started using Claude as my chatting partner and research partner.

1190
00:42:56,865 --> 00:43:00,805
And I'm using, as of right now, only because I first need to get one of my domains.

1191
00:43:00,805 --> 00:43:03,485
By the way, remember how we were talking about, oh, it's so genius with Replit.

1192
00:43:03,485 --> 00:43:05,465
but you just buy a domain, it's so easy.

1193
00:43:05,605 --> 00:43:09,325
Well, now I've got hosting through them

1194
00:43:09,325 --> 00:43:13,105
and I can't figure out how to damn transfer my domain.

1195
00:43:13,305 --> 00:43:15,805
So I want to get my $20 a month out of them

1196
00:43:15,805 --> 00:43:19,845
before I add one more subscription to the pie.

1197
00:43:20,405 --> 00:43:21,805
Of course, it's a genius model,

1198
00:43:21,925 --> 00:43:24,465
but they gave me some cheese and now I'm in a box.

1199
00:43:26,125 --> 00:43:28,125
But I have to just say,

1200
00:43:28,325 --> 00:43:32,245
I don't know what is going on over at Anthropic

1201
00:43:32,245 --> 00:43:34,945
because they clearly have less money than Google and OpenAI.

1202
00:43:35,985 --> 00:43:39,685
They have a lot of money, but those guys can cook.

1203
00:43:39,985 --> 00:43:40,985
It is the best.

1204
00:43:41,205 --> 00:43:42,925
As of right now, for me doing research,

1205
00:43:43,125 --> 00:43:45,625
it is two or three months ago,

1206
00:43:45,745 --> 00:43:47,365
Google trounced everything I'd ever use.

1207
00:43:47,665 --> 00:43:48,825
This is now trouncing Google.

1208
00:43:49,965 --> 00:43:50,365
Unbelievable.

1209
00:43:51,065 --> 00:43:52,865
Like in helping me write out some of my essays

1210
00:43:52,865 --> 00:43:53,825
to the first time.

1211
00:43:54,205 --> 00:43:56,265
And actually, something else I want to be very careful about

1212
00:43:56,265 --> 00:43:57,305
as we talked about, I was like,

1213
00:43:57,405 --> 00:43:59,465
I love writing for the actual thinking of it.

1214
00:43:59,465 --> 00:44:02,825
And, you know, I plan to continue to always do that.

1215
00:44:03,165 --> 00:44:07,045
But for the first time I had an anthropic-rited essay and I was like, part of this is better than I would have done.

1216
00:44:07,705 --> 00:44:14,025
And I would never think about have anthropic write the essay, but it can be a thought partner, a brainstorm partner.

1217
00:44:14,385 --> 00:44:19,285
I think if you say, look, I'm going all organic, I'm not touching these tools ever, you're going to get left behind yourself.

1218
00:44:19,425 --> 00:44:21,325
It's going to take you too long, you're going to spend so much time.

1219
00:44:21,645 --> 00:44:27,185
But if you also just throw up an idea and say, whatever it outputs I'm posting, that's also kind of a garbage approach.

1220
00:44:27,185 --> 00:44:32,585
I think the only rational approach is to know how to use the tool and get the most out of it.

1221
00:44:32,905 --> 00:44:36,325
But don't just put all of your trust in the tool and don't avoid the tool.

1222
00:44:36,445 --> 00:44:37,385
Use it to help you.

1223
00:44:37,945 --> 00:44:37,965
Totally.

1224
00:44:38,145 --> 00:44:39,605
And I'm finding that balance right now.

1225
00:44:39,905 --> 00:44:42,705
But to me, at least, this was a mark.

1226
00:44:42,845 --> 00:44:44,925
And again, there were other things where it's still I didn't.

1227
00:44:45,025 --> 00:44:46,465
I was like, I would have done this very differently.

1228
00:44:46,705 --> 00:44:48,305
And obviously, I'm still using it as a thought partner.

1229
00:44:48,685 --> 00:44:53,885
But it was the first time I've ever read something where I was like, wow, that's just incredible.

1230
00:44:54,165 --> 00:44:56,785
So if I may switch gears a little bit.

1231
00:44:56,925 --> 00:44:57,005
Yeah.

1232
00:44:57,005 --> 00:44:59,505
we'll still talk about AI,

1233
00:44:59,805 --> 00:45:00,685
but a different angle.

1234
00:45:00,965 --> 00:45:03,085
Like, typically we're talking about AI

1235
00:45:03,085 --> 00:45:04,965
in ways that most of the products

1236
00:45:04,965 --> 00:45:06,565
we talk about are proprietary products.

1237
00:45:06,645 --> 00:45:06,765
Right.

1238
00:45:07,705 --> 00:45:09,085
ChatGPT, Anthropic,

1239
00:45:09,805 --> 00:45:10,545
Replit, etc.

1240
00:45:11,105 --> 00:45:11,945
Closed source,

1241
00:45:12,265 --> 00:45:13,965
proprietary products.

1242
00:45:14,385 --> 00:45:16,085
And they're really good.

1243
00:45:16,425 --> 00:45:17,385
They're really good products.

1244
00:45:18,305 --> 00:45:19,925
Obviously, we care very much about

1245
00:45:19,925 --> 00:45:22,145
open source, open models, etc.

1246
00:45:22,525 --> 00:45:23,545
So I'd love to talk about

1247
00:45:23,545 --> 00:45:26,465
what's our...

1248
00:45:27,005 --> 00:45:30,445
What are our thoughts on open source AI?

1249
00:45:31,025 --> 00:45:33,725
And not just the models, but all the tooling around it

1250
00:45:33,725 --> 00:45:37,685
and the experiences around it, protocols, the whole stack,

1251
00:45:37,845 --> 00:45:39,585
everything about AI.

1252
00:45:40,045 --> 00:45:42,025
What are our thoughts on open source?

1253
00:45:42,765 --> 00:45:44,065
And then more specifically,

1254
00:45:44,985 --> 00:45:49,625
I certainly expect the big AI labs to do a lot of open source as well

1255
00:45:49,625 --> 00:45:52,085
when it's strategic for them.

1256
00:45:52,085 --> 00:45:52,525
Of course.

1257
00:45:53,745 --> 00:45:55,845
So what will be the gaps?

1258
00:45:55,845 --> 00:45:59,185
What'll be the blind spots

1259
00:45:59,185 --> 00:46:01,225
where there'd be a lot of value

1260
00:46:01,225 --> 00:46:02,745
to have open source AI

1261
00:46:02,745 --> 00:46:04,505
that the big AI labs aren't doing

1262
00:46:04,505 --> 00:46:06,865
and VC-funded startups aren't doing?

1263
00:46:08,385 --> 00:46:10,145
Curious to open that up.

1264
00:46:10,965 --> 00:46:13,805
And I ask that actually for selfish reasons as well.

1265
00:46:14,025 --> 00:46:15,905
Should Spiral start funding open source?

1266
00:46:16,325 --> 00:46:17,925
Sounds like Steve's got his checkbook open.

1267
00:46:18,685 --> 00:46:20,565
Our checkbook might be open.

1268
00:46:20,765 --> 00:46:24,265
So what should Spiral be thinking about

1269
00:46:24,265 --> 00:46:26,145
for open source AI funding?

1270
00:46:27,785 --> 00:46:29,425
Jeez, it's a great,

1271
00:46:29,545 --> 00:46:32,225
great, broad, open question.

1272
00:46:33,365 --> 00:46:35,725
You know, I think with open source stuff,

1273
00:46:35,905 --> 00:46:38,825
a lot of times I think it's comfortable

1274
00:46:38,825 --> 00:46:40,805
to build an open source version

1275
00:46:40,805 --> 00:46:43,525
of something that we know gives us value

1276
00:46:43,525 --> 00:46:44,505
as a closed source tool.

1277
00:46:44,825 --> 00:46:46,665
And then we start comparing it and saying,

1278
00:46:46,885 --> 00:46:47,465
yeah, but it's open.

1279
00:46:47,545 --> 00:46:48,105
Yeah, but it's open.

1280
00:46:48,805 --> 00:46:50,625
And I get, it's like, fine.

1281
00:46:50,625 --> 00:46:53,225
I get why it's an attractive proposition

1282
00:46:53,225 --> 00:46:55,585
to do the open version of the thing that's closed.

1283
00:46:56,245 --> 00:46:58,585
But I think the place where open source often shines

1284
00:46:58,585 --> 00:47:01,445
is when it sort of moves up or down a layer

1285
00:47:01,445 --> 00:47:05,485
or figures out sort of a different use case

1286
00:47:05,485 --> 00:47:07,965
that's maybe not well-served

1287
00:47:07,965 --> 00:47:09,305
and can't really be well-served

1288
00:47:09,305 --> 00:47:10,505
because of a structural problem

1289
00:47:10,505 --> 00:47:13,225
with either how the product or service

1290
00:47:13,225 --> 00:47:15,225
is delivered by the closed source version

1291
00:47:15,225 --> 00:47:18,185
or the business model of the closed source version.

1292
00:47:18,405 --> 00:47:19,425
It's just sort of something that

1293
00:47:19,425 --> 00:47:21,325
it's almost like a judo move

1294
00:47:21,325 --> 00:47:26,325
it hurts the ability to become more open.

1295
00:47:26,325 --> 00:47:29,325
Like the closed source thing can't become more open.

1296
00:47:29,325 --> 00:47:31,325
So I think, I was thinking about this a bit more,

1297
00:47:31,325 --> 00:47:42,710
and I think like a good analogy which I think I talked to you guys about before definitely written about is you know when Linux first came out you know we all said oh good Linux is going to replace Windows

1298
00:47:43,010 --> 00:47:55,510
So I so I, you know, I had my tower desktop computer in college and I installed Linux and I was like using some crappy Windows manager and like, you know, using Linux.

1299
00:47:55,690 --> 00:47:58,630
But it was kind of like a bad version of Windows.

1300
00:47:58,810 --> 00:48:00,890
Like it didn't have the software I wanted, didn't have the ecosystem.

1301
00:48:00,890 --> 00:48:05,270
them you know it's open which is nice but it but it sort of missed on all the things that mattered

1302
00:48:05,270 --> 00:48:13,010
and the real winning move for linux was to help us scale cloud services like google and facebook

1303
00:48:13,010 --> 00:48:18,370
and basically every you know uh salesforce every cloud service used today wouldn't work if you had

1304
00:48:18,370 --> 00:48:23,730
to pay licensing fees to sun solaris but it does work when you have an open free operating system

1305
00:48:23,730 --> 00:48:31,150
and so and and android and android yeah so so i think like you know rebuilding the closed thing

1306
00:48:31,150 --> 00:48:37,330
at that same layer is kind of a peer or cousin to the closed thing is fine starting point but i think

1307
00:48:37,330 --> 00:48:43,710
that's not really ever where the magic is the magic is up a layer down a layer on the side somehow

1308
00:48:43,710 --> 00:48:48,310
and so what new markets does it enable dk could you go a little bit deeper in that because you

1309
00:48:48,310 --> 00:48:50,150
You talked about this many times over the last couple years.

1310
00:48:50,810 --> 00:48:56,830
Why, yeah, like, why is it that Linux opened up the cloud market?

1311
00:48:56,830 --> 00:49:00,690
So you're saying the alternative was for these big, whatever,

1312
00:49:01,030 --> 00:49:05,730
Facebook or Google, they would have to pay another company large fees.

1313
00:49:05,910 --> 00:49:08,630
But why wouldn't there still be competition,

1314
00:49:08,750 --> 00:49:10,410
even if there's whatever closed source?

1315
00:49:10,550 --> 00:49:12,010
Okay, someone else develops, you know, whatever,

1316
00:49:12,110 --> 00:49:14,010
another closed source alternative that's cheaper.

1317
00:49:14,170 --> 00:49:15,910
Why was Linux so powerful?

1318
00:49:15,910 --> 00:49:17,810
operating system were coupled.

1319
00:49:18,370 --> 00:49:19,970
So you would deploy

1320
00:49:19,970 --> 00:49:22,110
your web server on

1321
00:49:22,110 --> 00:49:23,950
Sun with Solaris

1322
00:49:23,950 --> 00:49:26,110
and that would buy that.

1323
00:49:26,250 --> 00:49:27,590
On Spark chips too.

1324
00:49:27,710 --> 00:49:29,870
There was more diversity

1325
00:49:29,870 --> 00:49:30,950
in chips back then.

1326
00:49:31,950 --> 00:49:34,170
So Sun Microsystems,

1327
00:49:34,250 --> 00:49:35,670
you guys are getting into ancient history.

1328
00:49:35,890 --> 00:49:36,250
Sorry.

1329
00:49:37,450 --> 00:49:38,970
So think about Linux.

1330
00:49:39,130 --> 00:49:40,750
We're going to go back to QMF.

1331
00:49:41,750 --> 00:49:41,850
Okay.

1332
00:49:44,010 --> 00:49:44,490
Right.

1333
00:49:44,490 --> 00:49:49,650
Like Linux, you can just run on any, you can compile it for any x86 chip.

1334
00:49:49,910 --> 00:49:51,550
So you can buy it from Intel or AMD or whoever.

1335
00:49:51,730 --> 00:49:54,010
Those are kind of commodity parts, and they're not super specialized.

1336
00:49:54,010 --> 00:49:57,450
And so the magic is you sort of commoditize the complement.

1337
00:49:57,610 --> 00:49:59,230
You know the famous Joel Spolsky?

1338
00:50:00,450 --> 00:50:05,870
You sort of, if the chips are all commodity, then the OS kind of has more space to play.

1339
00:50:06,430 --> 00:50:13,850
And the Spark chips are not commodity, but the Intel x86 chips were quickly becoming commodity.

1340
00:50:13,850 --> 00:50:24,450
Like at the time, if you had a company and you wanted to equip people with computers, buy an IBM RS6000 or a Sparkstation.

1341
00:50:25,150 --> 00:50:30,130
Or if you're doing graphics or whatever, then you'd buy SGI computers.

1342
00:50:30,950 --> 00:50:33,470
And everything was proprietary about them.

1343
00:50:33,770 --> 00:50:33,990
Interesting.

1344
00:50:34,410 --> 00:50:41,230
So why didn't we see if, who was commoditizing the chip or who was making the chip really cheap at that point?

1345
00:50:41,230 --> 00:50:43,290
Google was huge here.

1346
00:50:43,290 --> 00:50:44,830
And the hardware too?

1347
00:50:45,170 --> 00:50:53,010
No, but they were able to take just off-the-shelf x86 computers and PCs, basically.

1348
00:50:53,150 --> 00:50:54,090
Oh, really?

1349
00:50:54,330 --> 00:50:54,830
And Linux.

1350
00:50:54,950 --> 00:50:55,570
Wow, okay.

1351
00:50:56,710 --> 00:51:01,450
I mean, at the time, I think most people thought that was insane,

1352
00:51:01,570 --> 00:51:06,570
because it's like a dinky little home computer, you're going to do big iron stuff.

1353
00:51:06,570 --> 00:51:11,210
But I mean, that's where a lot of early innovation from Google internally

1354
00:51:11,210 --> 00:51:13,090
was actually being able to scale.

1355
00:51:13,090 --> 00:51:17,250
allowed. Wow. I mean, there's a lot of innovation around distributed computing.

1356
00:51:17,450 --> 00:51:24,090
Exactly. Okay. Interesting. And so networking works differently and you need to assume the

1357
00:51:24,090 --> 00:51:28,830
thing is going to fail. So you need to have like a whole operational process at the network

1358
00:51:28,830 --> 00:51:32,590
operation center to say there's going to be people who run in and pull these things out when they

1359
00:51:32,590 --> 00:51:36,770
fail and put in new ones because it happens very frequently. These are just like off the shelf

1360
00:51:36,770 --> 00:51:42,090
components are extremely cheap, but they require all kinds of other trade-offs to be able to make

1361
00:51:42,090 --> 00:51:43,070
it actually operate well.

1362
00:51:43,710 --> 00:51:43,810
Interesting.

1363
00:51:43,930 --> 00:51:46,250
Both hardware, software, operations, everything.

1364
00:51:46,630 --> 00:51:50,950
And so was it Google's plan always at that point of like, hey, because Linux is open,

1365
00:51:51,030 --> 00:51:54,650
we're going to sort of like modify and continue to build this out for our needs?

1366
00:51:54,690 --> 00:51:58,870
Or did they ever think, like, was there ever like a proprietary push or were they always

1367
00:51:58,870 --> 00:51:59,970
going to build on top of the open?

1368
00:52:00,350 --> 00:52:03,830
I guess it was just like, it's not ideological.

1369
00:52:04,010 --> 00:52:04,750
I think it was just practical.

1370
00:52:04,990 --> 00:52:06,470
It was like, well, we've got these things running.

1371
00:52:06,590 --> 00:52:07,150
We can run them.

1372
00:52:07,150 --> 00:52:11,610
Also early on, Google is making a ton of money off search.

1373
00:52:11,610 --> 00:52:15,410
So why invest in building their own hardware?

1374
00:52:16,370 --> 00:52:18,170
Why build proprietary software?

1375
00:52:18,330 --> 00:52:19,530
That wasn't the core business.

1376
00:52:19,910 --> 00:52:24,950
It was better to let the world adopt

1377
00:52:24,950 --> 00:52:28,090
what Google had done well with these open systems

1378
00:52:28,090 --> 00:52:29,850
because that just meant more internet usage,

1379
00:52:30,030 --> 00:52:31,810
more searching, more revenue for Google.

1380
00:52:32,150 --> 00:52:32,170
Yeah.

1381
00:52:32,490 --> 00:52:34,090
Well, the reason I'm pushing all this, by the way,

1382
00:52:34,110 --> 00:52:35,210
is because I think it's really interesting

1383
00:52:35,210 --> 00:52:37,150
to think from this stage of like,

1384
00:52:37,210 --> 00:52:40,030
okay, Google, like, you know, history rhymes.

1385
00:52:40,030 --> 00:53:01,750
What are the rhymes going to be this time around where there's some entity, I actually have my own idea, which I didn't mean to skip too far ahead, but an entity which can basically use the open monetary stack, whether Bitcoin as a monetary good or the payment network or both, to enable them to basically turbocharge their business.

1386
00:53:01,750 --> 00:53:05,250
I mean, this is getting ahead of us, but the company I keep going back to is MicroStrategy.

1387
00:53:05,650 --> 00:53:06,830
I mean, I guess Block is another one.

1388
00:53:07,690 --> 00:53:09,110
But it's interesting.

1389
00:53:09,570 --> 00:53:19,290
Like, it of course makes sense that someone that's making money somewhere else is going to probably use Bitcoin in a way to make a lot more money and cut their own costs.

1390
00:53:20,790 --> 00:53:21,750
I've got some thoughts on that.

1391
00:53:21,810 --> 00:53:24,430
But before I jump ahead to MicroStrategy, like, I don't know.

1392
00:53:24,670 --> 00:53:28,870
Yeah, I'd say that's more of a financial engineering thing and it has its place.

1393
00:53:28,870 --> 00:53:29,090
Sort of.

1394
00:53:29,090 --> 00:53:31,990
I'm not dismissing it.

1395
00:53:32,330 --> 00:53:34,350
You know, I'm generally enthusiastic that it exists,

1396
00:53:34,430 --> 00:53:38,070
but I don't think it really moves the technology forward in a major way.

1397
00:53:38,150 --> 00:53:41,250
It's not like a new open source technology idea, right?

1398
00:53:41,250 --> 00:53:43,450
Which I think is where we were starting to.

1399
00:53:44,250 --> 00:53:46,710
Yeah, and I think we should go more in that direction,

1400
00:53:46,810 --> 00:53:48,310
although I'm increasingly questioning that

1401
00:53:48,310 --> 00:53:52,710
because I actually think that micro strategy mooning Bitcoin's price first

1402
00:53:52,710 --> 00:53:56,810
might be like the unlock we need for people to start using it in other ways too.

1403
00:53:56,810 --> 00:53:59,110
But yeah, I'm curious here where you're thinking

1404
00:53:59,110 --> 00:54:00,130
of more of the technological stack.

1405
00:54:00,310 --> 00:54:04,590
Let me share some possible characteristics

1406
00:54:04,590 --> 00:54:07,890
and then that might help generate some ideas from you guys.

1407
00:54:09,290 --> 00:54:11,790
But for what makes good candidates

1408
00:54:11,790 --> 00:54:16,210
for people to invest in working on open source projects for AI.

1409
00:54:17,270 --> 00:54:20,890
One is if it's non-commercial,

1410
00:54:20,890 --> 00:54:31,630
Meaning it's software that is expected to be high impact and really valuable to people or to agents.

1411
00:54:33,070 --> 00:54:36,570
But there's no obvious business model.

1412
00:54:36,850 --> 00:54:38,350
It's not expected to generate revenue.

1413
00:54:39,830 --> 00:54:49,390
And so that type of project is less likely to get VC funded, less likely to have startup hungry entrepreneurs to go after that because there's no profit motive.

1414
00:54:49,390 --> 00:54:50,990
So that's one characteristic.

1415
00:54:51,830 --> 00:55:08,250
Another is, you know, projects that prioritize principles that we care about, that the Bitcoin community cares about, but apply to other spaces too.

1416
00:55:08,250 --> 00:55:12,350
And that's like privacy, security,

1417
00:55:13,430 --> 00:55:15,130
sort of on steroids, like hyper security,

1418
00:55:15,770 --> 00:55:19,910
permissionless, censorship resistance, etc.

1419
00:55:20,130 --> 00:55:21,970
Those types of characteristics.

1420
00:55:23,410 --> 00:55:25,470
I mean, obviously there's like security companies

1421
00:55:25,470 --> 00:55:26,150
that make a lot of money.

1422
00:55:26,270 --> 00:55:29,410
I mean, so just building products

1423
00:55:29,410 --> 00:55:30,850
with one or more of those

1424
00:55:30,850 --> 00:55:31,910
doesn't mean you can't make money.

1425
00:55:32,590 --> 00:55:34,210
But oftentimes,

1426
00:55:34,570 --> 00:55:36,510
if you're really pushing the extremes in those,

1427
00:55:36,510 --> 00:55:39,530
it's marginal from a

1428
00:55:39,530 --> 00:55:40,450
profit

1429
00:55:40,450 --> 00:55:43,410
motivation. And so

1430
00:55:43,410 --> 00:55:45,550
I think those apply to AI

1431
00:55:45,550 --> 00:55:47,550
as well. So those are some other

1432
00:55:47,550 --> 00:55:49,330
characteristics. But

1433
00:55:49,330 --> 00:55:51,530
ultimately it's like, what can

1434
00:55:51,530 --> 00:55:53,530
you build of value that's in an open

1435
00:55:53,530 --> 00:55:55,430
source project that's not going to get just

1436
00:55:55,430 --> 00:55:57,010
steamrolled in three months

1437
00:55:57,010 --> 00:55:59,430
by the big AI labs or by

1438
00:55:59,430 --> 00:56:01,470
50 Silicon Valley startups that

1439
00:56:01,470 --> 00:56:03,470
are getting funded is

1440
00:56:03,470 --> 00:56:05,310
the big question. I think one more

1441
00:56:05,310 --> 00:56:06,170
characteristic too, I

1442
00:56:06,510 --> 00:56:08,910
There's application-level software.

1443
00:56:10,270 --> 00:56:11,990
There's protocol.

1444
00:56:12,730 --> 00:56:17,210
And I do think if you get closer to protocol or infrastructure,

1445
00:56:17,630 --> 00:56:21,370
that lends itself more to open source, non-commercial as well,

1446
00:56:21,470 --> 00:56:26,490
versus where you're trying to capture and have direct engagement with a customer.

1447
00:56:26,910 --> 00:56:28,850
So that's another angle.

1448
00:56:29,250 --> 00:56:30,430
So I have a ton of thoughts on this.

1449
00:56:30,430 --> 00:56:42,770
And I think one way I'm thinking about this is where open source can enable new business models, but it's required that we have an open standard first.

1450
00:56:43,370 --> 00:56:45,270
And in some cases, sidestep.

1451
00:56:46,050 --> 00:56:50,890
I mean, basically move into areas that closed companies may not or cannot.

1452
00:56:50,890 --> 00:57:00,970
And so the first thing that comes to mind is, you know, I am falling largely from hearing from, you know, Jesse and other folks in the office how quick the open source models are catching up.

1453
00:57:01,050 --> 00:57:08,170
So, for example, you know, there's like the lawsuit from Kla, which I find kind of funny, Anthropic, being like, oh, you guys are just using our model to distill your Chinese models.

1454
00:57:08,370 --> 00:57:10,230
It's like, yeah, we probably just jacked all the internet too.

1455
00:57:10,530 --> 00:57:12,430
So, you know, it comes around, bro.

1456
00:57:13,290 --> 00:57:19,430
But it's interesting to me that so far, you know, the commoditization of the complement is all coming from China, right?

1457
00:57:19,430 --> 00:57:21,210
So all of the good open source models,

1458
00:57:21,370 --> 00:57:22,810
Quinn and all of these guys,

1459
00:57:22,890 --> 00:57:23,570
it's all Chinese.

1460
00:57:23,830 --> 00:57:26,230
Now, you know, I've said once before

1461
00:57:26,230 --> 00:57:28,990
that I am skeptical of all the closest American companies.

1462
00:57:29,110 --> 00:57:29,910
I don't trust any of them.

1463
00:57:29,990 --> 00:57:32,490
I certainly don't trust the Chinese models either, right?

1464
00:57:32,550 --> 00:57:35,010
And so I am not at all convinced

1465
00:57:35,010 --> 00:57:37,830
that these models don't have some kind of like poison pill in them.

1466
00:57:37,910 --> 00:57:38,670
I know that sounds whatever,

1467
00:57:38,770 --> 00:57:39,630
Tim Boyle hat, whatever.

1468
00:57:40,110 --> 00:57:42,770
But we don't still fully understand how LMS even works.

1469
00:57:42,890 --> 00:57:44,570
So you can't, it's not like completely,

1470
00:57:45,630 --> 00:57:48,330
like we, what's the word for this?

1471
00:57:48,330 --> 00:57:51,090
Like you don't know A always leads to B.

1472
00:57:51,610 --> 00:57:52,470
What's the word for that?

1473
00:57:53,190 --> 00:57:54,070
Anyways, it'll come back to me.

1474
00:57:54,110 --> 00:57:56,250
But the point is you can't distill and understand

1475
00:57:56,250 --> 00:57:57,890
exactly what's happening within one of these models.

1476
00:57:58,290 --> 00:58:00,270
So one thing that I would like to see is

1477
00:58:00,270 --> 00:58:02,350
at the very least for a variety of reasons,

1478
00:58:02,350 --> 00:58:04,450
we shouldn't let China just totally dominate

1479
00:58:04,450 --> 00:58:05,790
the open source model world.

1480
00:58:06,190 --> 00:58:06,770
I would love to see,

1481
00:58:06,830 --> 00:58:08,810
and I know one or two companies have gotten funded on this

1482
00:58:08,810 --> 00:58:11,670
and Facebook in theory or Meta was working on this,

1483
00:58:11,890 --> 00:58:13,770
but I would love to see more American companies

1484
00:58:13,770 --> 00:58:14,970
that are distilling models.

1485
00:58:15,310 --> 00:58:16,610
Ideally, we've talked about on the show before,

1486
00:58:16,670 --> 00:58:17,750
my dream would be a model

1487
00:58:17,750 --> 00:58:19,610
where you can actually see the entire data set

1488
00:58:19,610 --> 00:58:20,550
of what went into the training.

1489
00:58:20,850 --> 00:58:21,830
But even if not that,

1490
00:58:21,990 --> 00:58:23,690
like who's going to build the coin in America?

1491
00:58:23,810 --> 00:58:24,610
That would be number one.

1492
00:58:24,910 --> 00:58:25,350
Well, real quick.

1493
00:58:25,910 --> 00:58:28,410
Actually, this should have been like a headline

1494
00:58:28,410 --> 00:58:29,510
for our show today.

1495
00:58:29,810 --> 00:58:34,890
NVIDIA, $26 billion commitment towards open models.

1496
00:58:35,070 --> 00:58:35,230
Really?

1497
00:58:35,390 --> 00:58:35,930
Open AI.

1498
00:58:36,290 --> 00:58:36,750
I didn't see that.

1499
00:58:36,810 --> 00:58:37,450
They announced it this way.

1500
00:58:37,450 --> 00:58:39,270
Wait, open AI or like actually open AI?

1501
00:58:40,450 --> 00:58:43,890
I mean, it's like a press release, right?

1502
00:58:44,230 --> 00:58:44,810
Causation, yeah.

1503
00:58:44,810 --> 00:58:55,350
I think it was a reaction to a lot of their customers are now threatening to create their own chips or use other people's chips as a negotiating tactic.

1504
00:58:55,490 --> 00:59:00,450
So then NVIDIA is like, okay, we'll just put $26 billion towards creating open models.

1505
00:59:00,810 --> 00:59:03,110
So we'll see how much it comes to reality.

1506
00:59:03,530 --> 00:59:05,550
That's the real commoditize the compliment move.

1507
00:59:05,790 --> 00:59:06,970
But that's the answer.

1508
00:59:07,090 --> 00:59:08,930
If they do that, that's the answer to your question.

1509
00:59:09,130 --> 00:59:09,430
Amazing.

1510
00:59:09,810 --> 00:59:11,210
So that's one thing I want to see.

1511
00:59:11,210 --> 00:59:13,310
on the hardware chip level as well.

1512
00:59:13,750 --> 00:59:15,010
You know, it would be incredible

1513
00:59:15,010 --> 00:59:16,850
to have an alternative to NVIDIA

1514
00:59:16,850 --> 00:59:18,430
that is an open stack.

1515
00:59:18,850 --> 00:59:20,230
Like, imagine in the same way

1516
00:59:20,230 --> 00:59:21,090
that framework computer,

1517
00:59:21,250 --> 00:59:22,070
you can like go in there

1518
00:59:22,070 --> 00:59:23,050
and hack on every part.

1519
00:59:23,450 --> 00:59:25,470
Imagine there was an alternative,

1520
00:59:26,250 --> 00:59:27,970
basically, yeah, like GPU stack

1521
00:59:27,970 --> 00:59:28,670
where like it's,

1522
00:59:28,970 --> 00:59:30,590
you know exactly all the parts in there

1523
00:59:30,590 --> 00:59:31,270
and you can, you know,

1524
00:59:31,310 --> 00:59:33,190
potentially plug and play as you wish.

1525
00:59:33,590 --> 00:59:34,870
I think that would be very interesting.

1526
00:59:35,270 --> 00:59:36,710
And then if you want to get really crazy

1527
00:59:36,710 --> 00:59:37,550
a company around this,

1528
00:59:37,610 --> 00:59:38,530
like imagine someone that wants

1529
00:59:38,530 --> 00:59:39,810
to build their own fab one day

1530
00:59:39,810 --> 00:59:41,430
and it's all open like that.

1531
00:59:41,490 --> 00:59:42,770
That obviously would have to be more commercial,

1532
00:59:42,930 --> 00:59:43,730
but I think that's pretty interesting.

1533
00:59:44,290 --> 00:59:46,850
How do you get the capital to build the data centers

1534
00:59:46,850 --> 00:59:49,730
and do all the research for an open version?

1535
00:59:50,170 --> 00:59:52,470
Well, I think you would have to turn that

1536
00:59:52,470 --> 00:59:53,690
into a company at this stage.

1537
00:59:53,750 --> 00:59:54,490
I don't see a way to do this

1538
00:59:54,490 --> 00:59:55,330
as like an open source project.

1539
00:59:55,510 --> 00:59:57,410
Maybe you could start the research of like,

1540
00:59:57,410 --> 00:59:58,790
hey, we're going to like open source

1541
00:59:58,790 --> 01:00:01,150
the whatever CAD files or whatever that looks like.

1542
01:00:01,250 --> 01:00:03,370
And then I think a company would probably have to commercialize it.

1543
01:00:03,470 --> 01:00:05,190
Or the government does it, right?

1544
01:00:05,230 --> 01:00:06,410
It's like a public effort.

1545
01:00:07,570 --> 01:00:08,910
Well, that's why I find NVIDIA

1546
01:00:08,910 --> 01:00:14,650
Because it's a wildly profitable and valuable company.

1547
01:00:16,650 --> 01:00:20,990
But it's realistic to me that they actually would spend that much money

1548
01:00:20,990 --> 01:00:25,030
and genuinely produce open models because that's not their core business.

1549
01:00:26,090 --> 01:00:29,790
So it actually is strategic to their business to do exactly that.

1550
01:00:30,390 --> 01:00:32,670
Kind of like Google gave away free stuff.

1551
01:00:32,670 --> 01:00:33,670
It's very Android, right?

1552
01:00:33,870 --> 01:00:33,990
Yeah.

1553
01:00:34,250 --> 01:00:34,730
It's very Android.

1554
01:00:34,730 --> 01:00:36,850
Well, like, I mean, like in the Google Maps,

1555
01:00:37,230 --> 01:00:38,230
when I worked on the product,

1556
01:00:38,370 --> 01:00:39,970
we came out with turn-by-turn navigation

1557
01:00:39,970 --> 01:00:41,690
and it completely upended that market.

1558
01:00:42,010 --> 01:00:42,910
We just gave it away for free.

1559
01:00:43,190 --> 01:00:43,390
Yeah.

1560
01:00:43,510 --> 01:00:45,750
And it's in the interest of the company.

1561
01:00:45,850 --> 01:00:47,310
I don't understand on the NVIDIA thing,

1562
01:00:47,430 --> 01:00:49,350
what exactly are they committing the money to?

1563
01:00:49,990 --> 01:00:52,350
Like to companies they want to invest in,

1564
01:00:52,390 --> 01:00:53,350
open source projects?

1565
01:00:53,630 --> 01:00:55,130
Like, have they talked about how the money

1566
01:00:55,130 --> 01:00:55,990
will be spent at all or?

1567
01:00:56,590 --> 01:00:56,870
No.

1568
01:00:57,150 --> 01:00:58,830
Or no, I don't know.

1569
01:00:59,270 --> 01:01:00,090
I don't think they did.

1570
01:01:00,590 --> 01:01:02,030
And they could very well.

1571
01:01:02,030 --> 01:01:04,250
It literally could have been just a,

1572
01:01:04,730 --> 01:01:08,730
immediate reaction to their business negotiations.

1573
01:01:09,270 --> 01:01:11,070
So it's probably TBD.

1574
01:01:11,270 --> 01:01:13,770
We should get the NVIDIA people back in here ASAP.

1575
01:01:14,610 --> 01:01:15,070
Yeah.

1576
01:01:15,390 --> 01:01:17,050
Which reminds me I need to send an invite

1577
01:01:17,050 --> 01:01:18,250
to an NVIDIA person for next week.

1578
01:01:18,330 --> 01:01:19,390
So that's a good reminder.

1579
01:01:20,770 --> 01:01:22,970
Another thing though that I think would be really important

1580
01:01:22,970 --> 01:01:25,550
to work on from the open side is,

1581
01:01:25,670 --> 01:01:26,310
well, two things.

1582
01:01:26,430 --> 01:01:28,830
One is, I think a very important primitive,

1583
01:01:29,450 --> 01:01:30,350
is auditability.

1584
01:01:30,350 --> 01:01:31,650
And actually auditability you could take

1585
01:01:31,650 --> 01:01:32,370
in many different ways.

1586
01:01:32,930 --> 01:01:35,830
One important piece we're missing in the compute stack today

1587
01:01:35,830 --> 01:01:38,690
that unlocks a lot of the crazier ideas that we've discussed,

1588
01:01:38,770 --> 01:01:40,270
including autonomous digital life,

1589
01:01:40,570 --> 01:01:45,870
is being able to verify that a specific computation was performed

1590
01:01:45,870 --> 01:01:47,630
as you believed it was going to be performed.

1591
01:01:48,070 --> 01:01:50,410
So some people from the crypto land have been doing this

1592
01:01:50,410 --> 01:01:52,570
with zero-knowledge proofs or homomorphic encryption.

1593
01:01:52,970 --> 01:01:55,830
But basically, this seems like somewhere where Spiral

1594
01:01:55,830 --> 01:01:57,370
could actually potentially move the needle,

1595
01:01:57,370 --> 01:02:01,690
is funding research to be able to prove that,

1596
01:02:01,690 --> 01:02:04,210
hey, I will pay you conditionally for a computation,

1597
01:02:04,410 --> 01:02:05,810
but you need to be able to send me a proof

1598
01:02:05,810 --> 01:02:07,530
that that specific computation was done.

1599
01:02:07,890 --> 01:02:09,650
That's a big unlock, and that's like,

1600
01:02:09,710 --> 01:02:10,930
I think you're right in your wheels house.

1601
01:02:11,750 --> 01:02:12,490
Taking a note.

1602
01:02:12,510 --> 01:02:13,870
Yeah, good.

1603
01:02:14,490 --> 01:02:15,770
So that's the first one.

1604
01:02:15,830 --> 01:02:16,850
You should talk to Dustin on that too,

1605
01:02:16,910 --> 01:02:18,510
because he's been thinking a lot about specifically

1606
01:02:18,510 --> 01:02:20,530
how that could unlock some of the life stuff.

1607
01:02:21,390 --> 01:02:23,410
The other big thing to me is like,

1608
01:02:23,750 --> 01:02:24,590
someone needs to figure,

1609
01:02:24,650 --> 01:02:26,550
and I think this could be commercialed as well,

1610
01:02:26,610 --> 01:02:28,190
so I don't know what's open source, what's commercial,

1611
01:02:28,410 --> 01:02:29,470
it's something I certainly want to back,

1612
01:02:29,470 --> 01:02:34,630
But like payment standards to like move between, I mean, the internet's open, right? And so there's

1613
01:02:34,630 --> 01:02:38,510
different jurisdictions. And like, you know, at some level, the product market fit for Bitcoin

1614
01:02:38,510 --> 01:02:44,170
has always been, let's call it geographic arbitrage, right? Like allowing people from

1615
01:02:44,170 --> 01:02:49,970
different worlds with different views, different legalities to cooperate. And so I think the same

1616
01:02:49,970 --> 01:02:54,010
thing is going to be true for AI. It's like, you know, if there are AIs from all different parts of

1617
01:02:54,010 --> 01:02:58,570
the world with different laws, just like the internet, like even if there's different payment

1618
01:02:58,570 --> 01:02:59,890
and whatever,

1619
01:03:00,070 --> 01:03:01,530
computation regulations in one country.

1620
01:03:01,650 --> 01:03:03,950
How do those guys work together

1621
01:03:03,950 --> 01:03:04,890
and at large scale?

1622
01:03:05,310 --> 01:03:06,710
I think figuring out standards

1623
01:03:06,710 --> 01:03:08,590
for using Bitcoin as collateral

1624
01:03:08,590 --> 01:03:10,370
between different payment networks

1625
01:03:10,370 --> 01:03:12,470
on the internet or different AIs,

1626
01:03:12,570 --> 01:03:13,770
that's going to be very, very important.

1627
01:03:13,870 --> 01:03:15,430
That could be a commercial opportunity as well.

1628
01:03:15,770 --> 01:03:16,630
I think the commercial opportunity

1629
01:03:16,630 --> 01:03:17,610
that I'm thinking a lot about there

1630
01:03:17,610 --> 01:03:20,730
is who's going to build the bank for AIs,

1631
01:03:20,790 --> 01:03:21,750
but that's coming.

1632
01:03:21,910 --> 01:03:23,210
That's my big obsession at the moment.

1633
01:03:23,490 --> 01:03:24,710
But first you need a standard

1634
01:03:24,710 --> 01:03:26,250
of how do agents post collateral

1635
01:03:26,250 --> 01:03:27,090
between one another

1636
01:03:27,090 --> 01:03:29,130
and then that verifiability of compute,

1637
01:03:29,250 --> 01:03:30,910
IPOS collateral, you say you did a computation,

1638
01:03:31,350 --> 01:03:32,390
if it doesn't, it gets slashed.

1639
01:03:32,890 --> 01:03:34,370
Those are all standards, which I think

1640
01:03:34,370 --> 01:03:36,490
for open source developers.

1641
01:03:36,510 --> 01:03:38,190
For the latter one, are you thinking of like

1642
01:03:38,190 --> 01:03:40,210
a peer-to-peer network for inference?

1643
01:03:40,870 --> 01:03:42,710
Yes, well, that's part of it

1644
01:03:42,710 --> 01:03:43,990
and that's where I think we're going, right?

1645
01:03:44,050 --> 01:03:46,890
It's like a peer-to-peer network for anything.

1646
01:03:47,650 --> 01:03:48,970
Inference could be one thing,

1647
01:03:49,270 --> 01:03:51,010
all the other digital resources we talked about,

1648
01:03:51,050 --> 01:03:52,170
bandwidth, storage, whatever,

1649
01:03:52,590 --> 01:03:54,650
but different standards for,

1650
01:03:55,210 --> 01:03:56,670
hey, I want to buy a resource

1651
01:03:56,670 --> 01:04:10,590
And like, if you don't have that resource, I know, like, one of the key ideas is that agents can't fall back on state actors and using state violence to enforce their property rights.

1652
01:04:10,870 --> 01:04:16,390
So how are they going to enforce property rights if someone violates at like the scale of 100,000 agents working together?

1653
01:04:16,930 --> 01:04:18,370
Well, one way to do that is collateral.

1654
01:04:18,690 --> 01:04:25,830
And I think there needs to be a standard established for how Bitcoin is used as collateral in each of these bilateral agreements between agents.

1655
01:04:25,830 --> 01:04:31,290
Real quick, we have David can't see the comments, so I'll catch you up to speed.

1656
01:04:31,430 --> 01:04:36,150
First of all, Ryan Brown 9900 says, love the basketball talk.

1657
01:04:36,430 --> 01:04:40,950
So sorry, guys, I'm getting encouragement to talk more about.

1658
01:04:41,590 --> 01:04:42,870
I just sit back.

1659
01:04:43,110 --> 01:04:43,890
I let it happen.

1660
01:04:44,290 --> 01:04:45,470
It was like an hour and a half late.

1661
01:04:45,710 --> 01:04:45,870
Like what?

1662
01:04:46,930 --> 01:04:47,830
Did you just hear that part?

1663
01:04:49,190 --> 01:04:49,870
Let's see.

1664
01:04:49,870 --> 01:04:58,490
Oh, so one specific project idea would be a peer-to-peer network for inference, like an AI project.

1665
01:04:58,650 --> 01:04:59,010
I love that.

1666
01:04:59,250 --> 01:05:03,890
Now, this is not, I mean, there are already projects in this vein.

1667
01:05:04,250 --> 01:05:06,150
Like Eric Voorhees started Venice.

1668
01:05:06,530 --> 01:05:07,270
Have you followed that at all?

1669
01:05:07,270 --> 01:05:09,050
Yeah, but no shitcoin, please.

1670
01:05:09,230 --> 01:05:15,030
Well, that's, so I'm not aware of a peer-to-peer network for inference that uses Bitcoin.

1671
01:05:15,630 --> 01:05:17,410
Well, actually, I have to give him a shout out.

1672
01:05:17,410 --> 01:05:19,470
Christopher David from OpenAgents

1673
01:05:19,470 --> 01:05:21,290
just yesterday, I believe,

1674
01:05:21,430 --> 01:05:22,950
relaunched his GPU marketplace

1675
01:05:22,950 --> 01:05:25,350
using sort of Noster's, the communication layer.

1676
01:05:25,750 --> 01:05:27,070
I don't know if I'm supposed to announce this,

1677
01:05:27,170 --> 01:05:29,070
but that man came out of the blue

1678
01:05:29,070 --> 01:05:32,690
and just raised some money from another Bitcoiner.

1679
01:05:32,830 --> 01:05:34,250
So that hopefully will be announced soon.

1680
01:05:34,570 --> 01:05:35,910
But that's one example.

1681
01:05:36,650 --> 01:05:37,750
People need to listen to our show

1682
01:05:37,750 --> 01:05:41,290
for confidential leaks information.

1683
01:05:41,410 --> 01:05:42,030
I hope that wasn't too much.

1684
01:05:42,030 --> 01:05:43,210
Very insider trading.

1685
01:05:43,210 --> 01:05:45,930
but I'm not

1686
01:05:45,930 --> 01:05:47,710
well I'll check out

1687
01:05:47,710 --> 01:05:48,930
openasians.com

1688
01:05:48,930 --> 01:05:51,270
it's the only one I know of so far

1689
01:05:51,270 --> 01:05:51,930
and then there's

1690
01:05:51,930 --> 01:05:55,430
is it Joseph Jack's

1691
01:05:55,430 --> 01:05:57,190
is that the open source VC

1692
01:05:57,190 --> 01:05:59,410
isn't he huge on

1693
01:05:59,410 --> 01:06:00,510
BitTensor

1694
01:06:00,510 --> 01:06:03,430
it's in the same

1695
01:06:03,430 --> 01:06:04,370
it's in the same vein

1696
01:06:04,370 --> 01:06:06,410
it's like again all these things have shit coins

1697
01:06:06,410 --> 01:06:09,850
it's like how many of these damn cycles

1698
01:06:09,850 --> 01:06:10,590
do we have to go through

1699
01:06:10,590 --> 01:06:12,510
but yes in theory BitTensor

1700
01:06:12,510 --> 01:06:17,990
tries to solve my understanding at a very high level. So if you had this verifiability compute

1701
01:06:17,990 --> 01:06:21,070
primitive we're talking about, you wouldn't even need that. But I think what they do is they have

1702
01:06:21,070 --> 01:06:25,550
these like different subnets where it's like, okay, I want you to do, I don't know, like image

1703
01:06:25,550 --> 01:06:29,970
generation or some task. And then I think they do some kind of like random sampling to see like,

1704
01:06:30,050 --> 01:06:33,590
okay, is someone doing what they say they're going to do or not? And then there's, there is

1705
01:06:33,590 --> 01:06:39,210
staking in there with whatever their, you know, crap token is. I, yeah, I mean, like something like

1706
01:06:39,210 --> 01:06:44,630
that with the zero knowledge proofs would be much better but even like that with bitcoin and none of

1707
01:06:44,630 --> 01:06:50,410
the whatever tokenomics crap would be really promising to me it's just it's still it's so

1708
01:06:50,410 --> 01:06:56,910
frustrating to me so frustrating that yet again there's like a new project space and everyone has

1709
01:06:56,910 --> 01:07:02,430
to go the shit coin route always um and and why does no one see the obvious that just use bitcoin

1710
01:07:02,430 --> 01:07:08,570
as money so fun that let's do that i want to fun that too well so i mean i i think that's

1711
01:07:08,570 --> 01:07:10,710
interesting to explore that.

1712
01:07:11,010 --> 01:07:12,650
By the way, just thinking more about, I mean,

1713
01:07:12,670 --> 01:07:14,650
you're saying like, how do you know if they're doing the computation

1714
01:07:14,650 --> 01:07:16,590
or how do you know which model they're running and like do

1715
01:07:16,590 --> 01:07:18,670
sampling, etc. That kind of reminds me of

1716
01:07:18,670 --> 01:07:20,750
Bitcoin mining actually, right? Because if I'm a

1717
01:07:20,750 --> 01:07:22,650
pool and your miner is connected to my

1718
01:07:22,650 --> 01:07:24,670
pool and I have

1719
01:07:24,670 --> 01:07:26,670
to pay you out a portion

1720
01:07:26,670 --> 01:07:28,570
of anything gets mined, how do I know you're

1721
01:07:28,570 --> 01:07:30,650
actually doing the work that you're saying,

1722
01:07:30,710 --> 01:07:32,570
right? Well, it's much easier to have a Shanti 56

1723
01:07:32,570 --> 01:07:34,630
though, right? I mean, it's the same

1724
01:07:34,630 --> 01:07:36,630
problem, but it's like one very simple

1725
01:07:36,630 --> 01:07:38,810
computation versus... And it's the fact

1726
01:07:38,810 --> 01:07:40,310
that it has no other productive

1727
01:07:40,310 --> 01:07:42,270
utility, right, besides

1728
01:07:42,270 --> 01:07:44,570
finding the... But you still have the

1729
01:07:44,570 --> 01:07:46,690
fundamental problem of you don't know if they're doing it

1730
01:07:46,690 --> 01:07:48,590
or not. So you need to come up with the mechanism

1731
01:07:48,590 --> 01:07:50,470
for like, are you actually doing this?

1732
01:07:50,570 --> 01:07:52,590
But I agree. What's

1733
01:07:52,590 --> 01:07:53,850
actually being done is much simpler,

1734
01:07:54,430 --> 01:07:55,670
but it's a similar pattern, I think.

1735
01:07:57,270 --> 01:07:57,930
Anyway, so

1736
01:07:57,930 --> 01:07:59,530
I think that's an interesting

1737
01:07:59,530 --> 01:08:02,530
project because I think

1738
01:08:02,530 --> 01:08:04,130
it checks a lot of the boxes, like

1739
01:08:04,130 --> 01:08:05,230
protocol level.

1740
01:08:06,630 --> 01:08:10,490
you're unlikely to get steamrolled in three months.

1741
01:08:10,590 --> 01:08:12,490
I mean, there might be a bunch of competing peer-to-peer networks,

1742
01:08:12,730 --> 01:08:16,170
but the one that's open source, open money,

1743
01:08:17,910 --> 01:08:22,070
truly structured as a public good,

1744
01:08:22,290 --> 01:08:25,210
like with the licensing and who the contributors are,

1745
01:08:25,210 --> 01:08:28,890
if you can really get momentum behind a project,

1746
01:08:29,170 --> 01:08:32,950
it's going to be resilient against the startup du jour

1747
01:08:32,950 --> 01:08:34,250
who comes out with something,

1748
01:08:34,250 --> 01:08:38,110
or like even a shit coin new project

1749
01:08:38,110 --> 01:08:40,190
where they make a lot of money

1750
01:08:40,190 --> 01:08:42,350
off their token issuance or whatever.

1751
01:08:42,650 --> 01:08:44,390
I think it'll be resilient to that.

1752
01:08:44,750 --> 01:08:45,690
So I'd love to, I mean,

1753
01:08:45,770 --> 01:08:47,030
so that, I think that's interesting.

1754
01:08:47,190 --> 01:08:47,510
And then like,

1755
01:08:47,770 --> 01:08:50,150
what are the other dozen projects like that

1756
01:08:50,150 --> 01:08:52,210
that have a similar profile?

1757
01:08:52,610 --> 01:08:55,570
Yeah, and I think another interesting thing is,

1758
01:08:56,050 --> 01:08:56,470
and by the way,

1759
01:08:56,490 --> 01:08:58,070
I think a lot of this can piggyback on

1760
01:08:58,070 --> 01:09:00,490
what Jesse is doing with Bora

1761
01:09:00,490 --> 01:09:02,770
and Anthony with Maple.

1762
01:09:02,770 --> 01:09:06,290
because in theory, if you start giving people

1763
01:09:06,290 --> 01:09:08,410
their own sort of edge AIs

1764
01:09:08,410 --> 01:09:10,230
and their ability to bring their own hardware,

1765
01:09:10,410 --> 01:09:11,730
and it's kind of the same way with mining.

1766
01:09:11,870 --> 01:09:13,290
It's like, we'll probably go through

1767
01:09:13,290 --> 01:09:14,790
the early days of mining.

1768
01:09:14,950 --> 01:09:16,030
There is the hobbyist period.

1769
01:09:16,170 --> 01:09:17,810
It's like, just plug in your computer and make money.

1770
01:09:18,230 --> 01:09:20,190
I think there's probably gonna be some interesting way

1771
01:09:20,190 --> 01:09:21,830
to kind of bootstrap a network like that

1772
01:09:21,830 --> 01:09:23,310
where it's like, hey, I bought a bunch of Mac minis,

1773
01:09:23,710 --> 01:09:27,790
and I don't always have tasks for my open claw to do.

1774
01:09:28,130 --> 01:09:29,510
It'd be cool if there was some way of like,

1775
01:09:29,570 --> 01:09:31,630
when I want to run my open claw, I can't,

1776
01:09:31,630 --> 01:09:35,530
And when I can't, then I can sort of underwrite the cost of what I already paid for it.

1777
01:09:35,610 --> 01:09:36,690
Like that, that to me seems fair.

1778
01:09:36,710 --> 01:09:37,190
I like that a lot.

1779
01:09:37,330 --> 01:09:39,290
That's kind of cool because it's very new.

1780
01:09:39,650 --> 01:09:43,230
And there's like this use case that's emergent that everybody kind of knows they need or

1781
01:09:43,230 --> 01:09:44,490
thinks they need or wants.

1782
01:09:45,010 --> 01:09:49,790
So I think that it kind of, at least it feels from a spidey sense perspective, it feels like

1783
01:09:49,790 --> 01:09:50,610
it's new enough.

1784
01:09:50,710 --> 01:09:51,570
There's enough demand.

1785
01:09:52,130 --> 01:09:53,250
Couldn't have been done before.

1786
01:09:53,330 --> 01:09:53,910
Now it can.

1787
01:09:53,990 --> 01:09:56,590
And like, it might be enough energy to bootstrap an ecosystem.

1788
01:09:57,250 --> 01:09:59,410
Actually, I want to make sure I understood what you said.

1789
01:09:59,490 --> 01:10:00,890
You're saying like...

1790
01:10:01,630 --> 01:10:03,530
an agent that supports cross models

1791
01:10:03,530 --> 01:10:06,570
and farms out stuff in a smart way.

1792
01:10:06,750 --> 01:10:09,690
Like if your local model can handle it,

1793
01:10:09,770 --> 01:10:11,090
it farms it out there because you get privacy,

1794
01:10:11,250 --> 01:10:12,070
it's cheaper, et cetera.

1795
01:10:12,210 --> 01:10:13,150
But if you need a more advanced,

1796
01:10:13,590 --> 01:10:14,930
then you go out to a peer-to-peer network

1797
01:10:14,930 --> 01:10:16,930
or you go to a frontier model.

1798
01:10:17,190 --> 01:10:18,610
Well, I mean, I think that's interesting too,

1799
01:10:18,670 --> 01:10:19,610
and that could potentially play into this.

1800
01:10:19,670 --> 01:10:20,690
But what I'm specifically saying is

1801
01:10:20,690 --> 01:10:22,370
there's a lot of people that just bought Mac Minis

1802
01:10:22,370 --> 01:10:25,090
and sometimes they're running their open claw to do stuff.

1803
01:10:25,150 --> 01:10:26,730
But like, let's say 80 or 90% of the time,

1804
01:10:26,790 --> 01:10:28,290
the Mac Mini is just sitting there doing nothing.

1805
01:10:28,690 --> 01:10:29,670
If there was an easy way

1806
01:10:29,670 --> 01:10:31,210
that you could just plug the Mac Mini in

1807
01:10:31,210 --> 01:10:32,830
and contribute it to the GPU marketplace

1808
01:10:32,830 --> 01:10:35,190
or take your agent.

1809
01:10:35,430 --> 01:10:36,230
And like, it's not just,

1810
01:10:36,490 --> 01:10:38,730
so it's like there's different levels of abstraction

1811
01:10:38,730 --> 01:10:39,410
you could play this at.

1812
01:10:39,490 --> 01:10:40,390
You could just do the hardware

1813
01:10:40,390 --> 01:10:41,650
and just like plug in, make money.

1814
01:10:42,050 --> 01:10:43,010
Or you could say, well,

1815
01:10:43,050 --> 01:10:44,610
not only do I have this hardware,

1816
01:10:44,610 --> 01:10:45,470
but I have hardware

1817
01:10:45,470 --> 01:10:48,590
and I have a specific open claw agent

1818
01:10:48,590 --> 01:10:49,470
that I have been like,

1819
01:10:49,810 --> 01:10:51,410
you know, using to build these products

1820
01:10:51,410 --> 01:10:52,430
or like using to have this,

1821
01:10:52,610 --> 01:10:54,070
like I've been having all these conversations

1822
01:10:54,070 --> 01:10:54,710
with my Anthropic.

1823
01:10:55,090 --> 01:10:57,250
Imagine if I had been having all of those

1824
01:10:57,250 --> 01:10:58,450
with my own, you know,

1825
01:10:58,570 --> 01:11:00,010
locally stored data around,

1826
01:11:00,150 --> 01:11:01,150
you know, this Bitcoin bank.

1827
01:11:01,150 --> 01:11:03,310
idea and then it's like, okay, if you want to talk to my guy,

1828
01:11:03,730 --> 01:11:04,590
you know, you can

1829
01:11:04,590 --> 01:11:07,010
talk to this agent that has a premium over

1830
01:11:07,010 --> 01:11:09,050
just the raw GPU cost because it also has all my

1831
01:11:09,050 --> 01:11:11,010
contacts. I feel like you could have some

1832
01:11:11,010 --> 01:11:13,070
virtualization software on a Mac

1833
01:11:13,070 --> 01:11:15,090
Mini that lets you kind of multi

1834
01:11:15,090 --> 01:11:16,970
home open clause that people could

1835
01:11:16,970 --> 01:11:19,090
pay for to run. That's a cool

1836
01:11:19,090 --> 01:11:19,990
idea. I like this.

1837
01:11:20,810 --> 01:11:32,675
If someone listening wants to explore that hit me up because it something I looking back as well That what I I was going to say I also like the idea but it feels more commercial It sounds like something we all might want to fund

1838
01:11:32,775 --> 01:11:33,695
Yeah, we're all asking for it.

1839
01:11:33,695 --> 01:11:34,355
Which is awesome.

1840
01:11:34,555 --> 01:11:35,235
Obviously, that's good.

1841
01:11:35,255 --> 01:11:35,895
It could be both.

1842
01:11:36,875 --> 01:11:40,035
So then I'd want to, when we explore that more,

1843
01:11:40,035 --> 01:11:44,235
if we can tease out what could be the non-commercial

1844
01:11:44,235 --> 01:11:47,335
open source stuff, components of it,

1845
01:11:47,395 --> 01:11:48,535
that could be interesting.

1846
01:11:48,555 --> 01:11:49,635
Well, for one thing, just the standard,

1847
01:11:49,635 --> 01:11:50,715
how do these guys cooperate?

1848
01:11:52,395 --> 01:11:54,335
I was going to throw out a little bit,

1849
01:11:54,575 --> 01:11:58,775
maybe a little bit of a different direction to explore,

1850
01:11:58,955 --> 01:12:05,135
but I'm generally concerned about the viability of ideas

1851
01:12:05,135 --> 01:12:09,075
that get very close to what the closed source models are doing

1852
01:12:09,075 --> 01:12:11,155
because the closed source models,

1853
01:12:11,435 --> 01:12:12,695
basically they raise tons of money,

1854
01:12:12,855 --> 01:12:14,435
they give away so much value

1855
01:12:14,435 --> 01:12:17,235
that you're kind of not playing on very even footing

1856
01:12:17,235 --> 01:12:19,155
by saying, oh, I'm willing to accept

1857
01:12:19,155 --> 01:12:23,455
some people are saying i'm willing to accept an inferior lm experience because of this ideology

1858
01:12:23,455 --> 01:12:30,935
okay peace i i'm glad those people exist you know i i want that future but i want that future but

1859
01:12:30,935 --> 01:12:35,015
also i want to know what can be done and what is the state of the art what's the frontier

1860
01:12:35,015 --> 01:12:39,795
possible and that seems to be that you know you give away a bunch of tokens you subsidize a bunch

1861
01:12:39,795 --> 01:12:45,535
tokens you can see more what the future holds so i'm i'm generally like more cautious or skeptical

1862
01:12:45,535 --> 01:12:52,015
around ideas that are more kind of right in line with competing with existing kind of commercial

1863
01:12:52,015 --> 01:12:56,975
models even though i'm a fan that you know people are going after that stuff but but i think there's

1864
01:12:56,975 --> 01:13:02,955
other areas that are worth exploring and i i mean we talked pretty extensively last week around the

1865
01:13:02,955 --> 01:13:10,515
idea of like how does like multi-agent orchestration work to me that feels like it's like peace let the

1866
01:13:10,515 --> 01:13:13,115
let everybody compete at the layer that it exists,

1867
01:13:13,355 --> 01:13:15,815
but then try to coordinate their activities

1868
01:13:15,815 --> 01:13:17,295
and become like a better interface

1869
01:13:17,295 --> 01:13:19,235
for coordinating their activities.

1870
01:13:19,335 --> 01:13:20,815
You get all the benefits of the subsidies

1871
01:13:20,815 --> 01:13:22,035
that they're willing to grant

1872
01:13:22,035 --> 01:13:24,075
to get their models more adoption,

1873
01:13:24,235 --> 01:13:27,395
but you're ultimately more at an abstraction layer

1874
01:13:27,395 --> 01:13:29,935
that is not super impacted by,

1875
01:13:31,055 --> 01:13:33,035
you're not locked in in the way that you would be

1876
01:13:33,035 --> 01:13:36,035
if you were tied to one CLI or one tool.

1877
01:13:37,055 --> 01:13:38,795
Can I double click on that or do you want me to?

1878
01:13:39,235 --> 01:13:39,435
Yeah.

1879
01:13:39,435 --> 01:13:42,955
I agree.

1880
01:13:43,295 --> 01:13:45,975
I mean, any use case where it's cross-model,

1881
01:13:46,595 --> 01:13:47,815
I think that's a good opportunity

1882
01:13:47,815 --> 01:13:51,355
where you are unlikely to be steamrolled

1883
01:13:51,355 --> 01:13:53,035
by the big AI companies

1884
01:13:53,035 --> 01:13:56,055
because they're unlikely to support

1885
01:13:56,055 --> 01:13:58,875
their competitors' models at tier one.

1886
01:13:59,575 --> 01:14:02,595
But I'm still not sure it's a great

1887
01:14:02,595 --> 01:14:07,175
non-commercial open source project

1888
01:14:07,175 --> 01:14:09,615
because I worry that there'll be

1889
01:14:09,615 --> 01:14:12,035
dozens or hundreds of VST-funded startups

1890
01:14:12,035 --> 01:14:13,135
who will do the same thing

1891
01:14:13,135 --> 01:14:14,535
and make it open source,

1892
01:14:14,675 --> 01:14:15,635
make it work cross-model,

1893
01:14:15,835 --> 01:14:17,795
but because they capture customers,

1894
01:14:18,455 --> 01:14:20,235
there is a business model there.

1895
01:14:21,875 --> 01:14:27,835
And so maybe establishing standards

1896
01:14:27,835 --> 01:14:32,095
is a reason to do a non-commercial

1897
01:14:32,095 --> 01:14:34,075
or sort of independent project,

1898
01:14:34,075 --> 01:14:38,875
but maybe like the sort of the headless standards,

1899
01:14:39,015 --> 01:14:44,155
the protocol standards and not the actual user-facing product.

1900
01:14:44,415 --> 01:14:44,495
Sure.

1901
01:14:45,455 --> 01:14:47,475
One of the things that I would love to see on that front,

1902
01:14:47,575 --> 01:14:48,815
and I think eventually, I mean, look,

1903
01:14:48,875 --> 01:14:51,215
I think a lot of these things maybe start as something open source,

1904
01:14:51,315 --> 01:14:53,475
but then they could evolve into products, which is interesting.

1905
01:14:53,695 --> 01:14:58,075
But in the same way that, you know, you guys have funded wallet scrutiny,

1906
01:14:58,175 --> 01:15:00,235
I would love to fund model scrutiny for all this.

1907
01:15:00,315 --> 01:15:02,995
Like, I don't know right now, one giant question I have,

1908
01:15:02,995 --> 01:15:06,375
And one product could be, okay, I've got all these coordinating, like I've got the coordinator

1909
01:15:06,375 --> 01:15:08,175
bot that goes and checks all these things for me.

1910
01:15:08,495 --> 01:15:11,795
But like one thing is I don't even know like a source of truth.

1911
01:15:11,975 --> 01:15:16,275
How do I know if one of these closed source models has started rugging me with ads?

1912
01:15:16,275 --> 01:15:19,755
How do I know if someone discovers there's a poison pill in the Chinese models?

1913
01:15:19,915 --> 01:15:22,495
Like there's no source of truth I know of today.

1914
01:15:22,895 --> 01:15:26,255
That to me seems like if there's like an objective and ideally there should be many of these,

1915
01:15:26,395 --> 01:15:31,595
but sort of objectively neutral, credible, like leaderboard or like wallet scrutiny for

1916
01:15:31,595 --> 01:15:36,675
that that that seems like yeah i like that framing that's great i'm not suggesting you shouldn't try

1917
01:15:36,675 --> 01:15:42,575
but i would just point out that you could be part of a one percent test on ads and so chat gpt may

1918
01:15:42,575 --> 01:15:48,175
be serving you ads but may not be serving whoever generated the wallet scrutiny assessment so i

1919
01:15:48,175 --> 01:15:53,395
think it's i'm not trying to discourage it but i try to try to figure out what can be achieved

1920
01:15:53,395 --> 01:16:00,415
there because i think the wallet scrutiny domain is in a sense a little bit simpler maybe to to

1921
01:16:00,415 --> 01:16:02,035
analyze and assess.

1922
01:16:02,595 --> 01:16:03,835
Yeah, but I mean,

1923
01:16:04,595 --> 01:16:06,435
but if you wanted to get really creative

1924
01:16:06,435 --> 01:16:08,395
with it, you could also be like, cool, we do our own

1925
01:16:08,395 --> 01:16:10,335
random sampling. And one way you can do it is with

1926
01:16:10,335 --> 01:16:12,295
Bitcoin, right? So it's like, cool, and

1927
01:16:12,295 --> 01:16:14,435
maybe this turns into a product or a protocol, but it's like

1928
01:16:14,435 --> 01:16:16,075
we will pay you to

1929
01:16:16,075 --> 01:16:18,295
every whatever day

1930
01:16:18,295 --> 01:16:19,995
we have a set of 10

1931
01:16:19,995 --> 01:16:22,335
prompts or queries or actions we

1932
01:16:22,335 --> 01:16:24,455
want to take, and we'll pay you to run this

1933
01:16:24,455 --> 01:16:26,315
on these different versions of the models

1934
01:16:26,315 --> 01:16:28,395
and we gather all the data back and assess

1935
01:16:28,395 --> 01:16:30,335
what's changed, what's not, kind of red

1936
01:16:30,335 --> 01:16:31,395
line, the def.

1937
01:16:32,915 --> 01:16:34,095
Vinny had a comment here.

1938
01:16:34,735 --> 01:16:36,175
He's saying that the difference between inference

1939
01:16:36,175 --> 01:16:38,075
and Bitcoin is that inference requires much higher

1940
01:16:38,075 --> 01:16:40,455
bandwidth and is not deterministic.

1941
01:16:40,875 --> 01:16:42,215
It might not be possible to create

1942
01:16:42,215 --> 01:16:43,175
a proof over that.

1943
01:16:44,375 --> 01:16:46,415
The way

1944
01:16:46,415 --> 01:16:47,135
Bitcoin mining works,

1945
01:16:47,375 --> 01:16:49,015
it's not

1946
01:16:49,015 --> 01:16:50,995
proving it either.

1947
01:16:51,435 --> 01:16:53,615
It's proving that

1948
01:16:53,615 --> 01:16:55,735
the miner is doing

1949
01:16:55,735 --> 01:16:57,655
quite a bit of work. The miner

1950
01:16:57,655 --> 01:16:59,935
delivers back something that doesn't

1951
01:17:00,335 --> 01:17:02,875
is not enough proof of work to mine a block,

1952
01:17:02,995 --> 01:17:04,495
but enough proof of work to show,

1953
01:17:04,755 --> 01:17:07,395
like, okay, I'm putting in serious computational power.

1954
01:17:08,295 --> 01:17:11,975
So that's not a proof.

1955
01:17:12,235 --> 01:17:13,135
Well, that would be...

1956
01:17:13,135 --> 01:17:16,395
Well, it's a proof that directionally

1957
01:17:16,395 --> 01:17:17,815
it is computing enough power.

1958
01:17:17,935 --> 01:17:19,135
It's not proof that it's...

1959
01:17:19,135 --> 01:17:23,075
Yeah, it's not proof that it's doing everything.

1960
01:17:23,455 --> 01:17:25,075
I mean, that to me would be an incredible

1961
01:17:25,075 --> 01:17:26,575
open-source project or research project.

1962
01:17:26,575 --> 01:17:27,835
It's like, how far can we get that?

1963
01:17:28,455 --> 01:17:29,515
And I think that's an open question.

1964
01:17:29,515 --> 01:17:32,495
I also want to also answer Vinny's other point,

1965
01:17:32,695 --> 01:17:33,935
Bitcoin isn't the bank,

1966
01:17:34,015 --> 01:17:35,815
because I was talking about trying to build a bank

1967
01:17:35,815 --> 01:17:36,415
for all these agents.

1968
01:17:36,895 --> 01:17:38,715
To be clear, the way I'm thinking about it is

1969
01:17:38,715 --> 01:17:41,215
I think there will be banks that are Bitcoin-based banks

1970
01:17:41,215 --> 01:17:43,275
that are potentially lending to the agents.

1971
01:17:44,095 --> 01:17:45,715
So that's where I'm thinking.

1972
01:17:46,315 --> 01:17:50,115
Ram says, did y'all talk about meta-acquiring Maltbook?

1973
01:17:50,655 --> 01:17:52,095
Ads are coming to agents,

1974
01:17:52,155 --> 01:17:54,395
and your agents will influence you to buy certain products.

1975
01:17:54,935 --> 01:17:56,115
We didn't even talk about that. That's big.

1976
01:17:56,115 --> 01:17:57,855
well

1977
01:17:57,855 --> 01:17:59,715
we sort of addressed

1978
01:17:59,715 --> 01:18:02,755
ads in LLMs right

1979
01:18:02,755 --> 01:18:04,995
we talked about the ads part

1980
01:18:04,995 --> 01:18:07,035
yeah we were talking about Moldbook getting acquired by Facebook

1981
01:18:07,035 --> 01:18:08,255
that's kind of interesting

1982
01:18:08,255 --> 01:18:11,335
I'm looking forward to when Matt Belez gets acquired

1983
01:18:11,335 --> 01:18:12,475
or actually I'm not looking forward to it

1984
01:18:12,475 --> 01:18:14,415
but meaning

1985
01:18:14,415 --> 01:18:16,155
conceptually origram

1986
01:18:16,155 --> 01:18:19,035
I don't know the whole Moldbook

1987
01:18:19,035 --> 01:18:21,355
thing seems kind of like crazy acquisition

1988
01:18:21,355 --> 01:18:21,935
yeah

1989
01:18:21,935 --> 01:18:24,375
did y'all have any takes on that and by the way

1990
01:18:24,375 --> 01:18:28,235
I've heard multiple people from last week say they really liked your idea about Slack for agents.

1991
01:18:28,915 --> 01:18:32,735
So, I mean, this is at least one data point other companies like that direction, too.

1992
01:18:33,375 --> 01:18:35,415
I don't know. Do y'all have any hot takes or thoughts on this?

1993
01:18:36,675 --> 01:18:39,735
Well, I mean, I don't know that I read too much into it.

1994
01:18:39,875 --> 01:18:46,015
I think Moldbook is fun and cool, but I think as I framed it before, I think it's performance art.

1995
01:18:46,255 --> 01:18:49,415
It's like, hey, here's kind of a future sci-fi of what it could be.

1996
01:18:49,475 --> 01:18:51,755
It's not like these agents are really acting unprompted.

1997
01:18:51,755 --> 01:18:55,675
so I think it's fun and cool that we have such a thing

1998
01:18:55,675 --> 01:18:57,235
you know it got acquired

1999
01:18:57,235 --> 01:18:59,675
acquired could mean everything from they got jobs

2000
01:18:59,675 --> 01:19:01,695
to they each got billion dollar paydays

2001
01:19:01,695 --> 01:19:03,095
I don't know which it was

2002
01:19:03,095 --> 01:19:05,635
you know it's hard for me to speculate

2003
01:19:05,635 --> 01:19:07,735
probably you know closer to the former

2004
01:19:07,735 --> 01:19:11,715
but not necessarily what you might think of as a regular job

2005
01:19:11,715 --> 01:19:14,675
so you know they could get a very nice kind of signing bonus

2006
01:19:14,675 --> 01:19:17,315
so I wouldn't read too much into it

2007
01:19:17,315 --> 01:19:21,355
and I think you know I think meta is in a tough spot

2008
01:19:21,355 --> 01:19:27,255
So I think, you know, trying to, like, they did the, what was the guy's name?

2009
01:19:27,355 --> 01:19:30,715
Alexander Wang, what was the company?

2010
01:19:31,235 --> 01:19:32,355
You know, the scale.

2011
01:19:32,875 --> 01:19:33,235
Right.

2012
01:19:33,375 --> 01:19:41,475
So they did scale and then, you know, that seems to be kind of crumbling and they acquired a bunch of, you know, visionary people and that doesn't seem to be going anywhere.

2013
01:19:41,575 --> 01:19:44,195
I don't know what assets they have to bring to this fight.

2014
01:19:44,295 --> 01:19:46,775
So I'm kind of like, okay, try another one.

2015
01:19:46,795 --> 01:19:47,175
Why not?

2016
01:19:47,175 --> 01:19:48,655
I mean, it's existential.

2017
01:19:48,835 --> 01:19:50,315
You have to try to compete here,

2018
01:19:50,395 --> 01:19:51,975
but I don't know that I've seen a move yet

2019
01:19:51,975 --> 01:19:54,575
that I'm like, okay, that makes sense to me

2020
01:19:54,575 --> 01:19:58,375
and multiple has really changed the,

2021
01:19:58,575 --> 01:20:00,075
I don't know, anything I've seen.

2022
01:20:00,075 --> 01:20:01,575
You want to say Matt Coates comment?

2023
01:20:01,815 --> 01:20:02,795
Yeah, I mean, he just says,

2024
01:20:02,895 --> 01:20:04,255
don't know why they bought Moldbook

2025
01:20:04,255 --> 01:20:06,595
when Facebook was already bots talking to bots.

2026
01:20:08,695 --> 01:20:09,795
Well played, sir.

2027
01:20:10,055 --> 01:20:11,555
True, true, true.

2028
01:20:12,655 --> 01:20:13,055
Yeah.

2029
01:20:13,615 --> 01:20:15,555
So back to, so you've already come out

2030
01:20:15,555 --> 01:20:16,455
with a couple of good ideas.

2031
01:20:16,455 --> 01:20:18,875
I want to keep the ideas flowing here.

2032
01:20:19,235 --> 01:20:21,835
I've got one out of left field.

2033
01:20:21,915 --> 01:20:23,475
I mentioned this briefly in the chat,

2034
01:20:23,515 --> 01:20:24,655
and I don't think I got much uptake,

2035
01:20:24,815 --> 01:20:28,855
but I'll mention most of my computing I do

2036
01:20:28,855 --> 01:20:32,195
with an open source tool called Open Super Whisper.

2037
01:20:33,195 --> 01:20:37,035
And it's just a hot key that I tap,

2038
01:20:37,315 --> 01:20:38,835
and that is how I input.

2039
01:20:39,015 --> 01:20:40,715
I basically try to not touch my keyboard,

2040
01:20:40,795 --> 01:20:41,915
and I only speak to my computer.

2041
01:20:42,635 --> 01:20:44,155
Now, that's an open source project.

2042
01:20:44,155 --> 01:20:56,995
It's very general purpose, but I think that's actually kind of like a very nice style of entry point to multi-agent orchestration or what is the future operating system look like?

2043
01:20:57,295 --> 01:21:02,255
Like whether it's on a mobile device or on a laptop, I think we're going to be speaking to our computers.

2044
01:21:02,935 --> 01:21:07,155
Like almost whenever we can, we'll choose to do that once we've tried.

2045
01:21:07,875 --> 01:21:12,615
And so, you know, maybe something in like the open super whisper.

2046
01:21:12,615 --> 01:21:13,935
OpenWhisper.

2047
01:21:14,995 --> 01:21:16,435
SuperWhisper is closed source.

2048
01:21:16,595 --> 01:21:19,115
OpenWhisper is the open source version?

2049
01:21:19,495 --> 01:21:22,535
I use an open source thing called OpenSuperWhisper.

2050
01:21:22,875 --> 01:21:24,695
But there may be other open source tools too.

2051
01:21:24,695 --> 01:21:26,735
WhisperFlow is what I'm using,

2052
01:21:26,995 --> 01:21:28,815
and that's what DK started with.

2053
01:21:29,495 --> 01:21:30,815
And even though he likes this other thing better,

2054
01:21:30,895 --> 01:21:32,955
he told me to start with WhisperFlow.

2055
01:21:33,095 --> 01:21:33,915
So I'm starting with that.

2056
01:21:34,415 --> 01:21:36,175
And it's amazing technology.

2057
01:21:36,595 --> 01:21:38,695
I mean, beyond the behavioral shift you're talking about,

2058
01:21:38,695 --> 01:21:40,615
it is incredible.

2059
01:21:40,615 --> 01:21:44,575
its accuracy and its accuracy with, well, whispering.

2060
01:21:44,955 --> 01:21:48,755
You can have a lot of ambient noise around you

2061
01:21:48,755 --> 01:21:52,955
and whisper it at a crazy low volume level

2062
01:21:52,955 --> 01:21:55,255
and it gets it perfect. It's amazing.

2063
01:21:55,255 --> 01:21:59,475
Is anyone yet building and perhaps we had someone in the chat that's watching this open ASMR?

2064
01:22:02,635 --> 01:22:03,855
What do you mean, DK?

2065
01:22:05,155 --> 01:22:07,415
Actually, DK, I have a question though.

2066
01:22:07,415 --> 01:22:11,975
I mean, so I'm still early days changing my own behavior,

2067
01:22:12,115 --> 01:22:14,455
so I don't have personal evidence yet,

2068
01:22:14,495 --> 01:22:16,075
but I totally buy into your conviction

2069
01:22:16,075 --> 01:22:20,175
on how it is so much more powerful and easier and et cetera.

2070
01:22:21,815 --> 01:22:23,775
But I know from my Google Glass days,

2071
01:22:24,275 --> 01:22:26,035
in which you could speak to that too,

2072
01:22:26,255 --> 01:22:27,815
and in certain situations,

2073
01:22:28,095 --> 01:22:31,675
even though the technology was not as good then as it is now,

2074
01:22:32,175 --> 01:22:34,235
it was way more natural, faster, blah, blah, blah,

2075
01:22:34,315 --> 01:22:34,995
everything you're saying.

2076
01:22:34,995 --> 01:22:40,515
But there's many environments in which you either can't speak or it's awkward.

2077
01:22:41,335 --> 01:22:44,775
Now with this whisper technology, you can literally whisper.

2078
01:22:45,715 --> 01:22:49,815
So that opens up more environments in which you could use it.

2079
01:22:50,075 --> 01:22:56,895
In fact, I'm going to test it here at Presidio Bitcoin and see if I annoy my neighbors by doing it.

2080
01:22:57,715 --> 01:23:00,395
But I want to dive into the annoyance.

2081
01:23:00,395 --> 01:23:04,195
Well, here's my question to you.

2082
01:23:04,995 --> 01:23:11,975
It's definitely socially weird if you see someone whispering to themselves.

2083
01:23:12,335 --> 01:23:14,915
Like you think they have mental problems.

2084
01:23:16,595 --> 01:23:21,895
So let's assume there's such tremendous value to speaking

2085
01:23:21,895 --> 01:23:23,975
and the technology allows for whispering

2086
01:23:23,975 --> 01:23:30,495
such that people start whispering to their device here in their ear or whatever.

2087
01:23:31,295 --> 01:23:33,295
Do you think it'll become socially acceptable?

2088
01:23:33,295 --> 01:23:36,675
Like it will no longer be like this version has mental issues.

2089
01:23:36,895 --> 01:23:42,995
We, every time there's a big technology change, the early adopters look awkward and weird.

2090
01:23:42,995 --> 01:23:45,115
And then finally that becomes like habit.

2091
01:23:45,255 --> 01:23:49,155
Remember like in the nineties, it was so weird to date somebody online.

2092
01:23:49,935 --> 01:23:52,635
And now it's just like, everybody does that.

2093
01:23:52,795 --> 01:23:54,335
That was like very socially awkward.

2094
01:23:54,335 --> 01:23:57,995
And then 30 years later, it's just like the standard, right?

2095
01:23:58,155 --> 01:24:00,875
I think that kind of thing happens all the time.

2096
01:24:00,875 --> 01:24:06,255
but to be clear it could go another way too which is like i don't know like i i don't want to be in

2097
01:24:06,255 --> 01:24:09,635
an office like with everyone whispering with each other that just that feels uncomfortable for me

2098
01:24:09,635 --> 01:24:13,755
maybe you shouldn't be in an office maybe you should be taking a walk this is my point should

2099
01:24:13,755 --> 01:24:19,095
mostly be done out so this is what i was going to say like the hopeful version to me is not that

2100
01:24:19,095 --> 01:24:22,955
like you have people in cubicles sitting there bizarrely whispering in the corner like that

2101
01:24:22,955 --> 01:24:27,855
seems like that seems like the the meta like strapped onto your face like eating the books

2102
01:24:27,855 --> 01:24:32,735
version like no thanks uh i i think yeah the good news would be you come together when you need to

2103
01:24:32,735 --> 01:24:37,135
do stuff in person and then you know if not exactly maybe you're like out taking a walk or

2104
01:24:37,135 --> 01:24:41,815
like at home or something i don't know like i just but the cat or i mean i totally agree on the

2105
01:24:41,815 --> 01:24:46,635
awkwardness part and like going for a walk and be able to do stuff i agree with all that but um

2106
01:24:46,635 --> 01:24:53,435
but this also the direct this suggests like the app the opposite of pb right and and and we have

2107
01:24:53,435 --> 01:24:55,395
proof now in the value of

2108
01:24:55,395 --> 01:24:57,235
coming, I mean, obviously

2109
01:24:57,235 --> 01:24:59,455
historically there's plenty of proof of people going to

2110
01:24:59,455 --> 01:25:01,415
an office, but then during COVID when

2111
01:25:01,415 --> 01:25:03,075
everyone worked from home, maybe

2112
01:25:03,075 --> 01:25:05,355
that was an event where everyone's like, oh, it's way better

2113
01:25:05,355 --> 01:25:07,335
to work from home. Everyone should work from home.

2114
01:25:07,995 --> 01:25:09,395
But PB is proving out that

2115
01:25:09,395 --> 01:25:11,335
people are volunteering. Like, because

2116
01:25:11,335 --> 01:25:13,375
we're not a company where we demand our workers come

2117
01:25:13,375 --> 01:25:15,315
to the office. Each person that comes here,

2118
01:25:15,735 --> 01:25:17,375
it's their own choice, right? And we have

2119
01:25:17,375 --> 01:25:19,535
people, literally four people now

2120
01:25:19,535 --> 01:25:21,255
who have moved to San Francisco to come here.

2121
01:25:21,655 --> 01:25:23,415
We have eight to ten people

2122
01:25:23,415 --> 01:25:25,215
who commute one hour each way

2123
01:25:25,215 --> 01:25:27,235
two, three, four times a week,

2124
01:25:27,735 --> 01:25:29,635
they are volunteering to do that

2125
01:25:29,635 --> 01:25:30,755
because of the value of being here.

2126
01:25:31,775 --> 01:25:36,075
So what is their future

2127
01:25:36,075 --> 01:25:37,875
for inputting to devices

2128
01:25:37,875 --> 01:25:39,895
if they're not going to be speaking or whispering?

2129
01:25:40,695 --> 01:25:42,475
I mean, at PV, there are, what,

2130
01:25:42,575 --> 01:25:43,915
like three phone booths?

2131
01:25:44,035 --> 01:25:44,995
Maybe the future is

2132
01:25:44,995 --> 01:25:47,815
instead of these office desks

2133
01:25:47,815 --> 01:25:48,615
right next to each other,

2134
01:25:48,695 --> 01:25:51,395
maybe it's a farm of more phone booths,

2135
01:25:51,555 --> 01:25:52,795
and you can kind of come out

2136
01:25:52,795 --> 01:25:55,175
and still gather and there's the lunchroom and coffee

2137
01:25:55,175 --> 01:25:56,175
and you get lunch together.

2138
01:25:56,515 --> 01:25:58,195
But maybe people, like when you're working,

2139
01:25:58,315 --> 01:25:59,715
it's more deep work, focused work,

2140
01:25:59,775 --> 01:26:02,075
and you're speaking to yourself in a phone booth environment.

2141
01:26:02,215 --> 01:26:03,135
That seems fine to me.

2142
01:26:03,235 --> 01:26:05,015
I don't feel like that's a...

2143
01:26:05,015 --> 01:26:05,815
Two thoughts on that.

2144
01:26:06,435 --> 01:26:08,975
I mean, one is, I'll tell you my own work style.

2145
01:26:09,155 --> 01:26:11,675
I mean, I love coming to PB and talking to people,

2146
01:26:11,775 --> 01:26:12,975
but I don't do it every day.

2147
01:26:13,055 --> 01:26:14,395
And part of the reason I don't do it every day

2148
01:26:14,395 --> 01:26:16,435
is I like to kind of demarcate where like,

2149
01:26:17,015 --> 01:26:19,235
you know, if I really need to focus

2150
01:26:19,235 --> 01:26:21,835
and like really go deep, often I would just stay home

2151
01:26:21,835 --> 01:26:23,135
because I know no one's going to talk to me.

2152
01:26:23,215 --> 01:26:24,555
But when I want to chat with people

2153
01:26:24,555 --> 01:26:27,055
and have less, let's say, super deep work,

2154
01:26:27,115 --> 01:26:28,335
then this is a way better environment.

2155
01:26:28,855 --> 01:26:31,415
So maybe that sort of hybrid thing becomes more common.

2156
01:26:31,915 --> 01:26:33,835
But I also think that, I don't know,

2157
01:26:33,995 --> 01:26:37,015
I don't totally buy that we're going to,

2158
01:26:37,315 --> 01:26:40,275
I still think you want some ability to visual,

2159
01:26:40,375 --> 01:26:42,435
and maybe in the future you'll have the Google Glass or Contacts

2160
01:26:42,435 --> 01:26:43,975
and just pull up a screen on demand.

2161
01:26:44,395 --> 01:26:46,715
But you're not just going to be sitting there in the corner like this.

2162
01:26:46,835 --> 01:26:48,155
You will want to see a screen.

2163
01:26:48,155 --> 01:26:50,775
I think screens are great for consuming

2164
01:26:50,775 --> 01:26:52,755
what the output is, right?

2165
01:26:53,095 --> 01:26:54,335
Not necessarily always required,

2166
01:26:54,435 --> 01:26:55,555
but it's a great format.

2167
01:26:55,675 --> 01:26:56,635
Like if you send me a photo,

2168
01:26:56,735 --> 01:26:58,135
I don't want to hear an LLM read me.

2169
01:26:58,655 --> 01:27:00,055
Here's this guy, you know,

2170
01:27:00,075 --> 01:27:01,335
standing on a street corner or whatever.

2171
01:27:01,635 --> 01:27:02,695
I want to see it, right?

2172
01:27:02,775 --> 01:27:05,095
Like the visual is important.

2173
01:27:05,195 --> 01:27:07,775
So I think we'll continue to have screens to see.

2174
01:27:08,195 --> 01:27:11,255
I'm just, I think I'm just pointing out,

2175
01:27:11,255 --> 01:27:13,135
I think the importance of the keyboard

2176
01:27:13,135 --> 01:27:15,975
is going to decrease dramatically.

2177
01:27:16,915 --> 01:27:19,475
And especially, the other thing to think about

2178
01:27:19,475 --> 01:27:20,755
how is work going to change?

2179
01:27:21,795 --> 01:27:23,695
You know, any given day,

2180
01:27:23,775 --> 01:27:25,895
I'm usually working on at least,

2181
01:27:25,955 --> 01:27:27,475
I have three agents that are working

2182
01:27:27,475 --> 01:27:29,775
and I sort of have three different work streams

2183
01:27:29,775 --> 01:27:31,135
in my head about what they're doing,

2184
01:27:31,235 --> 01:27:34,995
but I'm not on with them all times.

2185
01:27:34,995 --> 01:27:37,215
I dispatch something and I go do something else.

2186
01:27:37,595 --> 01:27:38,055
And check in.

2187
01:27:38,655 --> 01:27:39,495
And so check in.

2188
01:27:39,535 --> 01:27:41,635
So maybe the future of work is not sit at a desk

2189
01:27:41,635 --> 01:27:43,475
and interact with your agents and speak to them,

2190
01:27:43,495 --> 01:27:45,695
but maybe it's like you dispatch,

2191
01:27:45,835 --> 01:27:47,375
go to your phone booth, dispatch a few things,

2192
01:27:47,375 --> 01:27:48,635
then go out and have a coffee

2193
01:27:48,635 --> 01:27:49,575
and hang with your friends

2194
01:27:49,575 --> 01:27:51,035
while the agents are doing the work

2195
01:27:51,035 --> 01:27:52,355
and then come back and check in when they're done.

2196
01:27:52,655 --> 01:27:54,895
So DK, as much as I really want to believe that,

2197
01:27:54,915 --> 01:27:57,015
and I think that will be an option for a lot of people,

2198
01:27:57,075 --> 01:27:58,015
I also have to show,

2199
01:27:58,135 --> 01:28:00,755
even though I'm generally very much the optimist as well,

2200
01:28:00,835 --> 01:28:03,935
and I think that a lot of people will find that path.

2201
01:28:04,055 --> 01:28:05,515
That path seems like a good, happy path.

2202
01:28:05,875 --> 01:28:07,875
I will also note the sort of dark path

2203
01:28:07,875 --> 01:28:08,595
because, right, like the internet,

2204
01:28:08,715 --> 01:28:09,995
and you're like, oh, man, it's going to be so cool.

2205
01:28:09,995 --> 01:28:12,555
We're all like doing cool stuff with people on the internet.

2206
01:28:13,015 --> 01:28:13,995
And then it's like, it's Facebook.

2207
01:28:14,715 --> 01:28:15,535
Yeah, yeah.

2208
01:28:15,615 --> 01:28:18,235
Instagram and like single-handedly ruin society.

2209
01:28:18,635 --> 01:28:22,215
but what I am also seeing,

2210
01:28:22,315 --> 01:28:22,735
and by the way,

2211
01:28:22,815 --> 01:28:23,955
I haven't gotten on Twitter

2212
01:28:23,955 --> 01:28:24,675
in a couple of weeks

2213
01:28:24,675 --> 01:28:26,795
and I'm trying to experiment.

2214
01:28:27,135 --> 01:28:28,335
I had to unduly change my password

2215
01:28:28,335 --> 01:28:29,095
and I can't get on it.

2216
01:28:29,315 --> 01:28:30,075
And it's been awesome.

2217
01:28:30,735 --> 01:28:32,795
My life quality is improving,

2218
01:28:33,355 --> 01:28:34,895
which is an interesting little takeaway.

2219
01:28:35,035 --> 01:28:35,955
I still use Stacker News.

2220
01:28:36,195 --> 01:28:38,115
I mean, at Christmas,

2221
01:28:38,255 --> 01:28:39,255
I decided to reduce,

2222
01:28:39,555 --> 01:28:41,115
and I've 90% reduction

2223
01:28:41,115 --> 01:28:42,455
of looking at social media.

2224
01:28:42,635 --> 01:28:42,815
Really?

2225
01:28:42,855 --> 01:28:43,495
I feel healthier.

2226
01:28:43,675 --> 01:28:44,435
And it feels better, right?

2227
01:28:44,575 --> 01:28:44,715
Yeah.

2228
01:28:44,915 --> 01:28:46,335
Yeah, so there's no question on that.

2229
01:28:46,495 --> 01:28:46,995
And by the way,

2230
01:28:47,095 --> 01:28:48,375
but I don't feel like I miss out.

2231
01:28:48,375 --> 01:28:52,055
you guys, if it's important enough, you'll send it to me. I'll see it on stack. And it's not like

2232
01:28:52,055 --> 01:28:57,095
I'm missing out. I'm just not scrolling, which is amazing. But, um, but I did it before I, you know,

2233
01:28:57,095 --> 01:29:01,975
started my, my scroll detox, I guess. Um, I noticed like people that were like, I think,

2234
01:29:02,135 --> 01:29:05,315
I don't know, some, someone in the video is like, oh man, like there's this whole movement of like,

2235
01:29:05,355 --> 01:29:08,555
people are leaving their parties earlier because like their bots are working and they want to make

2236
01:29:08,555 --> 01:29:12,515
sure they're maximizing every second. And instead of having five bots working, they want 10 bots

2237
01:29:12,515 --> 01:29:17,035
working. And it's like, yeah, I mean, it could be like people like go and like, you know, maybe in

2238
01:29:17,035 --> 01:29:20,995
europe uh the european guys you know i see we have a couple in there they'll have their little cafe

2239
01:29:20,995 --> 01:29:27,555
and a bot working a day or two but then you know here in america i have 500 gold and so i i think

2240
01:29:27,555 --> 01:29:33,435
it could easily swing the other way to where i yeah i don't know i i think everybody is going to

2241
01:29:33,435 --> 01:29:40,475
become the orchestrator of some team of bots yes i agree with that if you sit on the end of a

2242
01:29:40,475 --> 01:29:44,775
computer you know like an internet connection with a computer and that's your job i think your job is

2243
01:29:44,775 --> 01:29:48,135
going to be a bot orchestrator, not whatever you think it is today.

2244
01:29:48,695 --> 01:29:52,895
I agree. And going back, DK, to what you suggested the future

2245
01:29:52,895 --> 01:29:56,875
office might look like, I thought that was a good suggestion. Or meaning like the phone

2246
01:29:56,875 --> 01:29:59,095
booths we have now, more of those.

2247
01:30:00,295 --> 01:30:04,535
So I'll commit to changing my own behavior at PB

2248
01:30:04,535 --> 01:30:08,915
to experiment. So I'm going to, number

2249
01:30:08,915 --> 01:30:12,875
one, speak more to the computer. I'll do it from my desk,

2250
01:30:12,875 --> 01:30:14,715
which I'm expecting to be a failure

2251
01:30:14,715 --> 01:30:16,595
but I'll do it just to pick up on what works

2252
01:30:16,595 --> 01:30:18,775
what doesn't. I'll do it from a phone booth

2253
01:30:18,775 --> 01:30:20,355
in which I

2254
01:30:20,355 --> 01:30:22,435
expect that to not be awkward

2255
01:30:22,435 --> 01:30:24,735
or weird but it will take me

2256
01:30:24,735 --> 01:30:26,395
away from a little bit

2257
01:30:26,395 --> 01:30:28,255
from the environment so I want to see

2258
01:30:28,255 --> 01:30:30,475
what is the sort of water cooler talk

2259
01:30:30,475 --> 01:30:32,595
I miss out on by being

2260
01:30:32,595 --> 01:30:34,815
in a phone booth. And then I'd

2261
01:30:34,815 --> 01:30:36,635
like to experiment with somewhere

2262
01:30:36,635 --> 01:30:38,495
in between the desk

2263
01:30:38,495 --> 01:30:40,595
setup and the phone booth. Is there something else

2264
01:30:40,595 --> 01:30:41,335
that you can like

2265
01:30:41,335 --> 01:30:45,855
either, you know, maybe it's just visually.

2266
01:30:46,235 --> 01:30:48,075
Because if you're truly whispering and others can't,

2267
01:30:48,315 --> 01:30:51,555
if others can't hear it and it's just like visually weird

2268
01:30:51,555 --> 01:30:52,835
to see someone's mouth moving,

2269
01:30:52,935 --> 01:30:54,975
maybe you just need like to cover,

2270
01:30:55,355 --> 01:30:57,215
visually cover your face or something.

2271
01:30:57,315 --> 01:30:57,575
I don't know.

2272
01:30:57,595 --> 01:30:59,375
Or we just need to get comfortable with it, right?

2273
01:30:59,855 --> 01:31:00,175
Or that.

2274
01:31:00,175 --> 01:31:01,075
We'll get comfortable with it.

2275
01:31:01,375 --> 01:31:02,815
So I think experimenting with that will be interesting

2276
01:31:02,815 --> 01:31:05,215
because then I think that, I mean,

2277
01:31:05,295 --> 01:31:09,155
it could literally change how the layout of PB.

2278
01:31:09,715 --> 01:31:11,155
I mean, we'd have to encourage other,

2279
01:31:11,155 --> 01:31:14,235
We'd have to see PV members actually want to use voice all the time.

2280
01:31:14,515 --> 01:31:15,975
So we can encourage them to try.

2281
01:31:16,095 --> 01:31:17,555
So I'll try, then get others to try.

2282
01:31:17,675 --> 01:31:20,875
And if there's strong demand for it and people feel as strongly as you do,

2283
01:31:21,175 --> 01:31:24,635
I can see changing the layout of PV to accommodate that.

2284
01:31:24,815 --> 01:31:29,275
And it's frankly like an experiment of future office space.

2285
01:31:29,355 --> 01:31:33,195
And I think one of the things that is going to be pivotal

2286
01:31:33,195 --> 01:31:35,255
in figuring out how you're experienced with it

2287
01:31:35,255 --> 01:31:38,195
is basically how much of your day today

2288
01:31:38,195 --> 01:31:40,235
is interacting with agents

2289
01:31:40,235 --> 01:31:41,915
versus how much of your day is doing,

2290
01:31:42,415 --> 01:31:44,135
you know, what's called traditional way.

2291
01:31:44,515 --> 01:31:47,495
So I suspect like I'm at the extreme

2292
01:31:47,495 --> 01:31:48,655
of I almost only,

2293
01:31:48,755 --> 01:31:50,515
I just spend all my time with agents.

2294
01:31:50,515 --> 01:31:52,975
And so I like see what that looks like

2295
01:31:52,975 --> 01:31:54,555
and feels like, which is fine.

2296
01:31:55,055 --> 01:31:57,575
I mean, I do other social things without agents,

2297
01:31:57,915 --> 01:31:59,855
but I'm saying any like work-related stuff

2298
01:31:59,855 --> 01:32:01,575
is really just agent development.

2299
01:32:01,735 --> 01:32:02,515
Touch grass.

2300
01:32:04,935 --> 01:32:06,735
But what I'm saying is if your job,

2301
01:32:06,815 --> 01:32:08,075
like Steve, I know you do a lot of like,

2302
01:32:08,195 --> 01:32:09,715
people in coordination and stuff.

2303
01:32:09,795 --> 01:32:10,775
So you're probably in emails

2304
01:32:10,775 --> 01:32:11,875
and sending people emails

2305
01:32:11,875 --> 01:32:13,055
and that kind of stuff.

2306
01:32:13,155 --> 01:32:14,875
If you don't have like an open claw running

2307
01:32:14,875 --> 01:32:16,955
and driving all of that activity for you,

2308
01:32:17,015 --> 01:32:18,095
which I'm guessing you don't,

2309
01:32:18,135 --> 01:32:18,775
I wouldn't recommend,

2310
01:32:19,355 --> 01:32:20,835
then you're probably going to have

2311
01:32:20,835 --> 01:32:22,335
a very different relationship

2312
01:32:22,335 --> 01:32:25,015
or experience with using agents

2313
01:32:25,015 --> 01:32:26,575
to like do your work, right?

2314
01:32:26,995 --> 01:32:27,575
Vinny asks,

2315
01:32:27,735 --> 01:32:28,715
isn't this just a phone call?

2316
01:32:28,895 --> 01:32:30,915
Well, not necessarily

2317
01:32:30,915 --> 01:32:32,495
because of the whispering

2318
01:32:32,495 --> 01:32:34,355
because you can whisper

2319
01:32:34,355 --> 01:32:35,435
and it works now.

2320
01:32:35,855 --> 01:32:36,595
But on a phone call,

2321
01:32:36,595 --> 01:32:38,275
that would be like super weird.

2322
01:32:39,935 --> 01:32:43,795
So the reason we'll do an entire TVJ episode whispering and everyone will be

2323
01:32:43,795 --> 01:32:43,955
like,

2324
01:32:44,015 --> 01:32:44,235
okay,

2325
01:32:44,595 --> 01:32:45,775
that was weird.

2326
01:32:46,075 --> 01:32:46,915
The ASMR channel.

2327
01:32:47,455 --> 01:32:49,835
We're sitting here in the corner shaking back and forth.

2328
01:32:50,775 --> 01:32:51,135
Well,

2329
01:32:51,235 --> 01:32:51,355
what,

2330
01:32:51,415 --> 01:32:53,335
what I was going to suggest though,

2331
01:32:53,335 --> 01:32:53,755
is I,

2332
01:32:53,855 --> 01:32:54,975
I think this,

2333
01:32:55,075 --> 01:32:56,315
this voice stuff,

2334
01:32:56,455 --> 01:32:57,215
we got it.

2335
01:32:57,595 --> 01:32:58,075
It's awkward.

2336
01:32:58,375 --> 01:32:58,735
It's weird.

2337
01:32:59,555 --> 01:33:00,875
The reason I raised the voice up,

2338
01:33:00,935 --> 01:33:01,975
I know we're kind of over time,

2339
01:33:02,015 --> 01:33:03,375
so I don't want to get too,

2340
01:33:03,575 --> 01:33:04,235
too far into this,

2341
01:33:04,275 --> 01:33:06,555
but the reason I raised the voice stuff is because I think,

2342
01:33:06,595 --> 01:33:15,295
And it's, you know, I'd say I have early visibility into how powerful it is in a version of work that I think is to become more common broadly in the future.

2343
01:33:15,675 --> 01:33:26,115
And given that, to me, it starts to sound to my ear like there's, you know, a change in the relationship with how we interact with our computing devices is like a change in the operating system.

2344
01:33:26,115 --> 01:33:55,495
And so I wonder if from an open source and funding perspective, there's a way to, you know, observe some of the like, you know, the fact that our computing devices and relationships may change. And if it is more voice centric, are there new tools that would help our voice centric controller build the bridge from our existing computing models to a voice centric computing model more than just hold a key, speak to a model, have it paste the text in, but can actually do stuff.

2345
01:33:55,495 --> 01:34:12,075
So like a very simple thing I'd like to be able to do is I've got all these command line interface tools open on my desktop. I'd like to be able to speak to my phone while I'm off on a trail run and have it dispatch the thing correctly to the right command line interface, even though I'm not there to tap and click the stuff.

2346
01:34:12,075 --> 01:34:14,355
so I can just sit open on my Mac at home

2347
01:34:14,355 --> 01:34:17,075
and I can be out on the trail with...

2348
01:34:17,735 --> 01:34:20,015
I think there's an infinite design space there,

2349
01:34:20,075 --> 01:34:22,335
so I'm not proposing that specific idea

2350
01:34:22,335 --> 01:34:23,015
should be implemented.

2351
01:34:23,195 --> 01:34:25,615
More like there's a big change happening here

2352
01:34:25,615 --> 01:34:29,175
that I don't think is commonly appreciated yet enough

2353
01:34:29,175 --> 01:34:30,335
because it's very niche.

2354
01:34:30,995 --> 01:34:32,575
So far, I think it's going to be awesome.

2355
01:34:32,935 --> 01:34:34,495
I think we should touch on this again next week

2356
01:34:34,495 --> 01:34:36,595
because I have some thoughts around the difference

2357
01:34:36,595 --> 01:34:39,355
between input and output text versus reading,

2358
01:34:39,535 --> 01:34:40,275
or listening and reading,

2359
01:34:40,355 --> 01:34:41,715
because I think it's going to be a combo.

2360
01:34:42,075 --> 01:34:47,955
Um, but I do want to acknowledge one other thing, which is, uh, Austin says, quote, my PBJ bet for

2361
01:34:47,955 --> 01:34:52,175
next week will be, do they actually cover the topics in the YouTube title? Um, and we still

2362
01:34:52,175 --> 01:34:58,195
should run a bet live market next week. Um, but I do want to say, since I don't want to totally

2363
01:34:58,195 --> 01:35:02,915
rug you guys, at least touch one of them, um, strategy stretch spree. I just want to touch on

2364
01:35:02,915 --> 01:35:08,675
this and just say, Oh my fucking God, this is crazy. The last two days, if you guys don't

2365
01:35:08,675 --> 01:35:10,475
understand what's happening.

2366
01:35:10,475 --> 01:35:14,215
Shout out to Ashu's who's shared strc.live.

2367
01:35:14,260 --> 01:35:14,700
I think.

2368
01:35:15,200 --> 01:35:15,720
I think that's right.

2369
01:35:16,580 --> 01:35:18,580
Someone can correct me if that's wrong.

2370
01:35:19,040 --> 01:35:20,920
But basically, it shows you in real time

2371
01:35:20,920 --> 01:35:24,360
how much money MicroStrategy is making selling stretch,

2372
01:35:24,480 --> 01:35:26,960
which is the sort of over-collateralized Bitcoin credit,

2373
01:35:27,160 --> 01:35:29,320
right now at an 11.5% interest rate,

2374
01:35:29,400 --> 01:35:30,440
which is insane.

2375
01:35:31,660 --> 01:35:33,780
And I think the last two days,

2376
01:35:33,860 --> 01:35:34,760
I forget the exact numbers.

2377
01:35:34,800 --> 01:35:35,460
It was like $100 million,

2378
01:35:35,640 --> 01:35:37,260
and then yesterday, well over $200 million.

2379
01:35:37,620 --> 01:35:39,060
I saw there was like $200 and something million

2380
01:35:39,060 --> 01:35:39,680
and $400 million.

2381
01:35:39,840 --> 01:35:41,800
I think the $400 million might have been total trading volume.

2382
01:35:41,900 --> 01:35:44,040
$200 million plus was like new sales of that.

2383
01:35:44,260 --> 01:35:45,960
But just like high level,

2384
01:35:46,380 --> 01:35:47,180
you know, if you're selling

2385
01:35:47,180 --> 01:35:49,640
$200 million worth of stretch securities

2386
01:35:49,640 --> 01:35:51,280
and then immediately putting

2387
01:35:51,280 --> 01:35:52,540
all of that into Bitcoin,

2388
01:35:52,860 --> 01:35:53,960
you're talking about buying

2389
01:35:53,960 --> 01:35:54,720
at current prices.

2390
01:35:54,860 --> 01:35:55,340
I think that'd be like,

2391
01:35:55,380 --> 01:35:56,560
well, like 2,600 Bitcoin,

2392
01:35:56,680 --> 01:35:57,200
something like that.

2393
01:35:57,820 --> 01:35:59,240
So if you're wondering,

2394
01:35:59,580 --> 01:36:00,760
hey, where's all this Bitcoin,

2395
01:36:01,160 --> 01:36:02,880
you know, bid coming from?

2396
01:36:03,680 --> 01:36:04,700
Coming from stretch.

2397
01:36:05,260 --> 01:36:06,440
And there's a lot more than it.

2398
01:36:06,460 --> 01:36:06,860
You know, was there,

2399
01:36:07,100 --> 01:36:07,980
was there something

2400
01:36:07,980 --> 01:36:09,580
that structurally changed

2401
01:36:09,580 --> 01:36:10,460
in recent,

2402
01:36:10,520 --> 01:36:12,500
because stretch as a vehicle,

2403
01:36:12,500 --> 01:36:13,240
I think it's existed.

2404
01:36:13,360 --> 01:36:14,040
I think it went public

2405
01:36:14,260 --> 01:36:19,700
six months ago or more maybe so it's not that that just came online right so was there some other

2406
01:36:19,700 --> 01:36:23,760
factor that just recently drove a bunch of demand for that i think i think there's a couple things

2407
01:36:23,760 --> 01:36:28,860
um i don't know this is my speculation um one thing is they keep hiking the rates like six

2408
01:36:28,860 --> 01:36:32,780
perpetual rate like six rate heights back to back so at some point it's like when the rest of the

2409
01:36:32,780 --> 01:36:37,040
and you started at 10 percent yield now it's at 11 and a half or nine yeah it was less it was like

2410
01:36:37,040 --> 01:36:41,080
it was under 10 maybe it was nine and i think they've hiked it a quarter to a half point multiple

2411
01:36:41,080 --> 01:36:42,820
times in a row. So now they're 11.5%.

2412
01:36:42,820 --> 01:36:44,900
But to be clear, they're hiking because

2413
01:36:44,900 --> 01:36:47,140
there's not enough demand. Well, there wasn't

2414
01:36:47,140 --> 01:36:49,100
until this week, right? So they

2415
01:36:49,100 --> 01:36:51,180
found the sweet spot. It turns out, you know, markets

2416
01:36:51,180 --> 01:36:52,1000
will discover the sweet spot. And the sweet spot was

2417
01:36:52,1000 --> 01:36:54,900
11.5% apparently. And so

2418
01:36:54,900 --> 01:36:57,380
you know, that's

2419
01:36:57,380 --> 01:36:59,220
one. Another thing is I saw

2420
01:36:59,220 --> 01:37:00,960
a headline maybe

2421
01:37:00,960 --> 01:37:03,100
last week about one of

2422
01:37:03,100 --> 01:37:05,060
these, I don't know, some sort of

2423
01:37:05,060 --> 01:37:07,180
financial firm. It was not exactly a rating agency

2424
01:37:07,180 --> 01:37:09,180
but basically giving it a

2425
01:37:09,180 --> 01:37:11,040
buy rating. It seems like, I mean, it seems like

2426
01:37:11,040 --> 01:37:11,760
Saylor has been doing

2427
01:37:11,760 --> 01:37:12,740
row shows.

2428
01:37:12,860 --> 01:37:13,660
I've heard from multiple people

2429
01:37:13,660 --> 01:37:14,380
talking with a lot of people.

2430
01:37:14,420 --> 01:37:15,120
It seems like basically

2431
01:37:15,120 --> 01:37:16,140
behind closed doors,

2432
01:37:16,660 --> 01:37:16,820
you know,

2433
01:37:16,920 --> 01:37:18,420
the sell is just a lot easier

2434
01:37:18,420 --> 01:37:18,780
to say,

2435
01:37:18,920 --> 01:37:19,140
hey,

2436
01:37:19,300 --> 01:37:19,480
you know,

2437
01:37:19,500 --> 01:37:21,460
Mr. Fixed Income Credit Guy,

2438
01:37:21,860 --> 01:37:23,020
you're making 4%

2439
01:37:23,020 --> 01:37:24,040
with a lot of risk

2440
01:37:24,040 --> 01:37:24,460
on this,

2441
01:37:24,520 --> 01:37:24,680
you know,

2442
01:37:24,720 --> 01:37:25,460
municipal bond.

2443
01:37:25,560 --> 01:37:26,680
We'll pay you an 11.5%,

2444
01:37:26,680 --> 01:37:28,040
6X over collateralized.

2445
01:37:28,440 --> 01:37:29,400
And it's extremely liquid.

2446
01:37:29,520 --> 01:37:29,980
You can sell it

2447
01:37:29,980 --> 01:37:30,760
at a moment's notice

2448
01:37:30,760 --> 01:37:31,280
and get full.

2449
01:37:31,280 --> 01:37:31,300
Well,

2450
01:37:31,380 --> 01:37:32,300
there's in between that.

2451
01:37:32,560 --> 01:37:34,300
Isn't there like a 7% to 9%

2452
01:37:34,300 --> 01:37:35,560
is common in,

2453
01:37:35,660 --> 01:37:35,780
I think,

2454
01:37:35,840 --> 01:37:36,780
junk bond markets?

2455
01:37:38,020 --> 01:37:38,260
Maybe.

2456
01:37:38,820 --> 01:37:38,960
Yeah.

2457
01:37:38,960 --> 01:37:39,580
I think this,

2458
01:37:39,680 --> 01:37:40,420
I think the alternative

2459
01:37:40,420 --> 01:37:44,500
is you take a very high risk. Like, I don't think people are saying treasuries are risky. They may.

2460
01:37:44,740 --> 01:37:49,120
I mean, I could think that, and you may think that, but I don't know that the professionals

2461
01:37:49,120 --> 01:37:53,760
typically think that. But I think they're comparing it to something like a junk bond or high yield

2462
01:37:53,760 --> 01:38:07,798
bond And this still blows that out of the water And it looks much less risky because you don really need to worry about the business operations right And actually you saying that makes it more credible to say that you just said 11 might

2463
01:38:07,798 --> 01:38:08,678
be the sweet spot, right?

2464
01:38:08,838 --> 01:38:08,978
Right.

2465
01:38:09,438 --> 01:38:14,058
Well, so I was initially thinking, well, 9, 11, what's the difference compared to like

2466
01:38:14,058 --> 01:38:14,618
4 or whatever?

2467
01:38:14,718 --> 01:38:20,078
But if you're comparing it to junk bonds and the junk bonds are 7 to 9, well then 9% for

2468
01:38:20,078 --> 01:38:25,478
this whole new thing and this whole new risk that is new to you to assess, it's like, I'll

2469
01:38:25,478 --> 01:38:27,098
just go with junk bonds. I know that world.

2470
01:38:27,878 --> 01:38:29,578
But if you get 250 extra

2471
01:38:29,578 --> 01:38:30,378
basis points,

2472
01:38:30,758 --> 01:38:33,118
then it might

2473
01:38:33,118 --> 01:38:35,538
tip the scales. I also want to give

2474
01:38:35,538 --> 01:38:37,418
a shout out to Borja

2475
01:38:37,418 --> 01:38:39,538
and his company, Roxham. We've talked about them

2476
01:38:39,538 --> 01:38:41,478
before. I've mentioned many times, I've

2477
01:38:41,478 --> 01:38:43,418
become increasingly bullish on them over

2478
01:38:43,418 --> 01:38:45,498
the years. And I think it was yesterday or the

2479
01:38:45,498 --> 01:38:47,378
day before, they announced their new product

2480
01:38:47,378 --> 01:38:49,518
offering, which I think is genius. Absolute

2481
01:38:49,518 --> 01:38:51,118
genius. And it's the ability

2482
01:38:51,118 --> 01:38:53,698
to, they are now, through their

2483
01:38:53,698 --> 01:38:56,578
not available to US citizens right now, unfortunately.

2484
01:38:57,018 --> 01:38:58,498
But if you're outside of the United States,

2485
01:38:58,498 --> 01:39:02,598
you can buy Stretch on their exchange with Bitcoin

2486
01:39:02,598 --> 01:39:04,638
and get the yield paid in Bitcoin.

2487
01:39:05,558 --> 01:39:06,778
And that's pretty interesting.

2488
01:39:06,778 --> 01:39:09,618
I think we have at least one Canadian listening right now.

2489
01:39:09,738 --> 01:39:12,518
So maybe that person can test it out.

2490
01:39:12,518 --> 01:39:16,058
So just to make sure you deposit Bitcoin into Roxham

2491
01:39:16,058 --> 01:39:18,278
and they custody Bitcoin for you

2492
01:39:18,278 --> 01:39:22,078
and then they pay you 11% or 11.5%.

2493
01:39:22,078 --> 01:39:24,278
I don't know if they're taking any of the spread or not,

2494
01:39:24,458 --> 01:39:26,258
but yeah, maybe, maybe not.

2495
01:39:26,258 --> 01:39:26,778
In Bitcoin.

2496
01:39:27,318 --> 01:39:27,738
In Bitcoin.

2497
01:39:27,918 --> 01:39:29,318
You are giving up custody,

2498
01:39:29,558 --> 01:39:32,218
but you are getting a stretch yield on that.

2499
01:39:32,458 --> 01:39:33,218
That's my understanding.

2500
01:39:33,518 --> 01:39:34,818
Wow, that's kind of crazy.

2501
01:39:35,098 --> 01:39:35,878
It's kind of genius.

2502
01:39:36,458 --> 01:39:37,418
That seems, yeah.

2503
01:39:37,598 --> 01:39:38,858
Why is that not more common?

2504
01:39:39,318 --> 01:39:41,278
It's like, it's literally the smartest product I've ever heard.

2505
01:39:41,358 --> 01:39:43,318
And there's some net, some big, big things

2506
01:39:43,318 --> 01:39:44,658
I can imagine them doing next.

2507
01:39:44,838 --> 01:39:46,358
So, hats off.

2508
01:39:46,718 --> 01:39:48,158
Yeah, it's still, you know, it's still.

2509
01:39:48,578 --> 01:39:49,758
Yeah, not your keys, not your coin,

2510
01:39:49,918 --> 01:39:51,258
all the usual caveats,

2511
01:39:51,258 --> 01:39:54,718
but just from a what will actually work in the world, genius.

2512
01:39:54,918 --> 01:39:58,758
Yeah, from market's perspective, I think that seems savvy.

2513
01:39:59,198 --> 01:39:59,318
Yeah.

2514
01:39:59,838 --> 01:40:02,798
So we're way over, but just so we can check the box

2515
01:40:02,798 --> 01:40:04,258
that we covered things in the title,

2516
01:40:04,558 --> 01:40:07,418
quick newsflash, which we usually don't do here,

2517
01:40:07,538 --> 01:40:10,958
but there are two more product offerings

2518
01:40:10,958 --> 01:40:12,078
announced in the past week,

2519
01:40:12,238 --> 01:40:14,738
which support coins on Bitcoin,

2520
01:40:15,658 --> 01:40:18,558
or related to this.

2521
01:40:18,558 --> 01:40:21,038
One is Tether announced.

2522
01:40:22,238 --> 01:40:22,958
RGB thing.

2523
01:40:23,178 --> 01:40:25,758
U-E-T-X-O.

2524
01:40:25,998 --> 01:40:27,578
I might be butchering it, but something like that.

2525
01:40:27,638 --> 01:40:30,058
It's very close to U-Tree-X-O, but it's not U-Tree.

2526
01:40:30,238 --> 01:40:32,858
I think it's U-E-T-X-O, I think.

2527
01:40:33,398 --> 01:40:37,258
In any case, they've actually been working on this for like four years.

2528
01:40:37,778 --> 01:40:40,058
It uses LDK, so I've known about this for a long time.

2529
01:40:40,138 --> 01:40:53,317
But it an RGB LDK system So it using Lightning It better than the original Tether on Bitcoin because that was on chain Didn scale very well Now it using Lightning Scales better

2530
01:40:54,537 --> 01:40:55,197
And the second

2531
01:40:55,197 --> 01:40:57,577
news is Arc Labs raised

2532
01:40:57,577 --> 01:40:59,837
$5 million in change.

2533
01:41:00,637 --> 01:41:01,597
With Tether

2534
01:41:01,597 --> 01:41:03,177
I think was the lead, but there are

2535
01:41:03,177 --> 01:41:05,457
other notable investors

2536
01:41:05,457 --> 01:41:07,437
in it. But along

2537
01:41:07,437 --> 01:41:08,997
with that investment was

2538
01:41:08,997 --> 01:41:11,077
bringing stable coins and Tether

2539
01:41:11,077 --> 01:41:11,837
to

2540
01:41:11,837 --> 01:41:15,157
ARK Labs as well and ARK technology.

2541
01:41:15,677 --> 01:41:15,777
Yeah.

2542
01:41:16,097 --> 01:41:17,497
So the more the merrier.

2543
01:41:17,677 --> 01:41:19,197
It just adds to the...

2544
01:41:19,197 --> 01:41:20,657
Every day Bitcoin gets stronger.

2545
01:41:20,797 --> 01:41:22,117
What would need to happen then

2546
01:41:22,117 --> 01:41:24,577
for one or each of those to take off?

2547
01:41:24,657 --> 01:41:26,277
Do you need a major issuer

2548
01:41:26,277 --> 01:41:29,217
like a circle or a tether

2549
01:41:29,217 --> 01:41:30,737
to choose to do that system?

2550
01:41:30,737 --> 01:41:31,737
Well, you need a number of things,

2551
01:41:31,897 --> 01:41:33,417
but I mean, the first box they check

2552
01:41:33,417 --> 01:41:34,077
is that tether,

2553
01:41:34,617 --> 01:41:35,817
the one is from tether,

2554
01:41:36,037 --> 01:41:39,037
so obviously tether will exist on that

2555
01:41:39,037 --> 01:41:41,757
and then tether the token

2556
01:41:41,757 --> 01:41:44,317
and then because Tether

2557
01:41:44,317 --> 01:41:46,277
invested in Arc Labs, I mean they're bringing Tether

2558
01:41:46,277 --> 01:41:48,237
to that. So both of those solutions

2559
01:41:48,237 --> 01:41:50,317
will have Tether tokens.

2560
01:41:51,617 --> 01:41:52,457
Now you need

2561
01:41:52,457 --> 01:41:53,517
more than that.

2562
01:41:54,817 --> 01:41:55,937
You need

2563
01:41:55,937 --> 01:41:57,257
network effects,

2564
01:41:57,797 --> 01:42:00,017
adoption, usage, etc.

2565
01:42:01,137 --> 01:42:02,217
And that's a tough

2566
01:42:02,217 --> 01:42:02,777
nut to crack.

2567
01:42:03,197 --> 01:42:05,257
And Tether do things to

2568
01:42:05,257 --> 01:42:07,757
encourage that? Because presumably

2569
01:42:07,757 --> 01:42:09,957
Tether would like these other

2570
01:42:09,957 --> 01:42:11,857
solutions that are a little more open than maybe

2571
01:42:11,857 --> 01:42:13,937
Sol, Tron, ETH, stuff

2572
01:42:13,937 --> 01:42:15,737
that they may historically have been on?

2573
01:42:16,157 --> 01:42:18,157
So are they... Well, if I'm

2574
01:42:18,157 --> 01:42:19,237
Tether, I think

2575
01:42:19,237 --> 01:42:21,857
what I definitely don't want to see is

2576
01:42:21,857 --> 01:42:23,877
more of these, or like, corp chains

2577
01:42:23,877 --> 01:42:26,197
winning. So like a Tempo

2578
01:42:26,197 --> 01:42:27,377
base, etc.

2579
01:42:28,437 --> 01:42:30,017
So that would be like

2580
01:42:30,017 --> 01:42:32,017
enemy number one, and then

2581
01:42:32,017 --> 01:42:34,017
the next tier would be

2582
01:42:34,617 --> 01:42:36,337
Solana,

2583
01:42:36,517 --> 01:42:37,917
Ethereum, layer

2584
01:42:37,917 --> 01:42:39,477
one that are

2585
01:42:39,477 --> 01:42:41,717
relatively more decentralized

2586
01:42:41,717 --> 01:42:43,797
than the aforementioned ones.

2587
01:42:45,577 --> 01:42:45,937
But I think

2588
01:42:45,937 --> 01:42:48,777
the folks buying Tether

2589
01:42:48,777 --> 01:42:51,357
are pretty hardcore Bitcoiners.

2590
01:42:51,737 --> 01:42:52,877
And they get Bitcoin.

2591
01:42:53,097 --> 01:42:55,397
And they would rather see Bitcoin as a platform.

2592
01:42:56,497 --> 01:42:57,877
They could share more yield.

2593
01:42:58,077 --> 01:43:00,317
I mean, that wouldn't be able to do less.

2594
01:43:01,037 --> 01:43:01,757
Is that possible?

2595
01:43:01,897 --> 01:43:03,357
Can they take some of the yield

2596
01:43:03,357 --> 01:43:04,957
and share it back to drive

2597
01:43:04,957 --> 01:43:06,637
maybe wallet adoption?

2598
01:43:06,777 --> 01:43:08,357
You can't do that in the US,

2599
01:43:08,437 --> 01:43:09,177
but you could do that.

2600
01:43:09,477 --> 01:43:10,357
I don't know.

2601
01:43:10,437 --> 01:43:13,177
I mean, I don't know the rules of other jurisdictions,

2602
01:43:13,417 --> 01:43:15,097
but you could maybe do that outside the US.

2603
01:43:16,577 --> 01:43:19,537
That's why stretch tokens exist outside the US,

2604
01:43:19,597 --> 01:43:21,977
so it apparently is more regulatory friendly.

2605
01:43:22,057 --> 01:43:23,077
Well, this is what I was going to say,

2606
01:43:23,157 --> 01:43:24,637
is to come back to the whole stretch thing,

2607
01:43:24,697 --> 01:43:27,057
not to just like simp on microstrategy,

2608
01:43:27,217 --> 01:43:43,336
but like I don know what going to happen in the United States because of regulatory but like in my mind nothing can compete with wrap stretch for the stable coin Nothing I mean like you just think about I mean we had this one the buck token like maybe that one is in particular scammy I

2609
01:43:43,336 --> 01:43:47,616
don't know. But like if they're giving an 11 and a half percent, hell, they're giving 7% and you

2610
01:43:47,616 --> 01:43:53,356
can wrap that and keep 2% and give away the 5% to whatever, nine and a half percent to users.

2611
01:43:53,536 --> 01:43:59,336
Like that's just going to dominate. Like I think stretch, yeah, stretch is already one. It's like,

2612
01:43:59,336 --> 01:44:02,456
who's going to figure out how to like productize that for the stable coin offering maybe it'll be

2613
01:44:02,456 --> 01:44:06,016
tether maybe it'll be someone else but like i don't even see tether with their like measly

2614
01:44:06,016 --> 01:44:10,616
treasury yield able to compete in that world like i actually think tether is vulnerable right now

2615
01:44:10,616 --> 01:44:17,876
i'm i'm optimistic on you know what microstrategy is doing but i'll just ask the question um you know

2616
01:44:17,876 --> 01:44:25,276
the the yield the 11 and a half percent yield is coming from the presumed growth of the bitcoin

2617
01:44:25,276 --> 01:44:27,136
assets that underlie that.

2618
01:44:27,816 --> 01:44:29,516
And I think he generally models

2619
01:44:29,516 --> 01:44:31,536
it at 20 or 30% compounded annual

2620
01:44:31,536 --> 01:44:33,656
in the next 10 years or so.

2621
01:44:34,476 --> 01:44:34,976
And so

2622
01:44:34,976 --> 01:44:37,496
11.5 is maybe sharing

2623
01:44:37,496 --> 01:44:39,276
about half of that, maybe protecting

2624
01:44:39,276 --> 01:44:40,436
some buffer for volatility.

2625
01:44:41,056 --> 01:44:42,656
But what's to keep

2626
01:44:42,656 --> 01:44:45,376
an upstart from coming in and saying, we're offering

2627
01:44:45,376 --> 01:44:47,236
15% on a small stack

2628
01:44:47,236 --> 01:44:49,396
so we can't absorb as much liquidity, but we're

2629
01:44:49,396 --> 01:44:51,356
going to give 15% yield and

2630
01:44:51,356 --> 01:44:52,416
we're just going to hope that

2631
01:44:52,416 --> 01:44:55,216
the buffer and the

2632
01:44:55,276 --> 01:44:57,596
20-30% of the drawdown in sequence of returners

2633
01:44:57,596 --> 01:44:58,816
doesn't cause a problem.

2634
01:44:59,316 --> 01:45:00,556
If you have to ask, you are the yield.

2635
01:45:02,416 --> 01:45:03,316
Anybody who

2636
01:45:03,316 --> 01:45:04,696
accepts the yield is

2637
01:45:04,696 --> 01:45:06,116
that person.

2638
01:45:07,076 --> 01:45:08,896
To be clear, I'm not only for the yield, but

2639
01:45:08,896 --> 01:45:11,016
I think the short answer is, we've talked about this

2640
01:45:11,016 --> 01:45:13,736
before, but there are

2641
01:45:13,736 --> 01:45:15,296
moats

2642
01:45:15,296 --> 01:45:17,336
around who can raise capital

2643
01:45:17,336 --> 01:45:19,716
in the event of a downturn.

2644
01:45:19,716 --> 01:45:21,196
He has proven

2645
01:45:21,196 --> 01:45:23,536
at least through two bear markets so far

2646
01:45:23,536 --> 01:45:24,996
that he's not taking

2647
01:45:24,996 --> 01:45:28,316
sufficient leverage to sell down the stack,

2648
01:45:28,636 --> 01:45:30,436
I would be very skeptical of an upstart.

2649
01:45:30,936 --> 01:45:32,996
It would be not only difficult to build that trust,

2650
01:45:33,056 --> 01:45:34,536
but for them to actually manage the liquidity

2651
01:45:34,536 --> 01:45:35,636
to be able to raise the capital.

2652
01:45:35,936 --> 01:45:37,156
I'm not saying it's impossible.

2653
01:45:37,316 --> 01:45:39,316
There is clearly a buffer

2654
01:45:39,316 --> 01:45:41,976
between whatever the actual yield they're offering

2655
01:45:41,976 --> 01:45:42,716
and the CAGR is.

2656
01:45:43,936 --> 01:45:45,296
That's the value.

2657
01:45:46,376 --> 01:45:47,976
I just think they have so much

2658
01:45:47,976 --> 01:45:49,936
such deep capital relationships,

2659
01:45:50,396 --> 01:45:50,776
brand record.

2660
01:45:50,836 --> 01:45:51,836
I think it's going to be super hard.

2661
01:45:51,836 --> 01:45:53,876
So much volatility around the actual instruments

2662
01:45:53,876 --> 01:45:54,316
they're trading.

2663
01:45:54,316 --> 01:45:59,116
there's like markets on markets on top of micro strategy stock on top of stretch now so i mean

2664
01:45:59,116 --> 01:46:04,596
it's possible but like that that's not a game i would want to play to go after sailor yeah i agree

2665
01:46:04,596 --> 01:46:12,936
yeah all right is that a optimistic enough ending i think that's a wrap austin's happy

2666
01:46:12,936 --> 01:46:18,836
because we covered the topic so there we go have a good weekend everyone dk would say peace peace