WEBVTT

1
00:00:04.293 --> 00:00:07.855
Hey, Bankless Nation, we've got some bonus content for you here on the weekend.

2
00:00:08.075 --> 00:00:12.457
Every month, Bankless and Venice partner together to do a VVV community call.

3
00:00:12.858 --> 00:00:17.740
That's a one hour long episode that I host live on X with John, the head of strategy at Venice.

4
00:00:18.341 --> 00:00:21.863
And this call, this episode is really meant for VVV holders.

5
00:00:21.983 --> 00:00:30.287
We talk about VVV economics, Venice product updates, and cover the news in the AI industry as it relates to privacy and dignity in the world of AI.

6
00:00:30.827 --> 00:00:33.049
And as you might guess, there's always a lot to talk about.

7
00:00:33.709 --> 00:00:41.373
While typically we don't place content with our partners on the RSS feed, we've got enough requests from the community that we do so since that is where they consume their content.

8
00:00:41.693 --> 00:00:46.336
So we're adding the monthly VVV call to the RSS feed on the weekend after it's recorded.

9
00:00:46.876 --> 00:00:55.441
If you like the episode and you want to join the VVV Telegram where we talk, learn, and educate about VVV, you can hit me up in the Bankless Discord or in my Twitter account and I'll throw you the link.

10
00:00:56.001 --> 00:00:59.643
So here is the September VVV call with John from Venice.

11
00:01:00.003 --> 00:01:02.964
Welcome to the September VVV call.

12
00:01:03.264 --> 00:01:17.567
Every month, John, the head of strategy at Venice, and I go through the developments, the news, the events in the Venice and private AI space, keeping you up to speed with all things VVV and everything related to user privacy and sovereignty in the AI age.

13
00:01:17.767 --> 00:01:19.067
John, happy September.

14
00:01:19.548 --> 00:01:20.348
How are things over at Venice?

15
00:01:21.248 --> 00:01:22.848
Yeah, happy September, David.

16
00:01:23.088 --> 00:01:23.689
Things are good.

17
00:01:23.769 --> 00:01:29.530
It's been another, it's crazy that it's been a month already, but yeah, things happen fast.

18
00:01:30.656 --> 00:01:43.767
I think this week or this month specifically has been when the social zeitgeist around AI has kind of hit new highs with attention on Anthropic, attention on OpenAI.

19
00:01:43.807 --> 00:01:46.229
Dario's having a meeting with Donald Trump.

20
00:01:47.050 --> 00:01:55.978
The big AI labs keep on the market conversation around the big AI labs is looking at their IPO coming down the pike at the end of this year or early next year.

21
00:01:56.518 --> 00:02:20.545
And I think as like a, as a societal conversation, I think we're having the AI conversation kind of like more than ever, which kind of feels crazy to say, but because we've been having this AI conversation for years now, but it seems to be nonstop in terms of just like how much, how ready AI is just a conversation on like the daily podcast in the New York Times, like everyone, everyone is talking about AI everywhere.

22
00:02:21.740 --> 00:02:29.665
Yeah, I mean, I think as these labs push forward, as the models continue to get better, I just, I don't see how that doesn't continue to be the case.

23
00:02:29.805 --> 00:02:36.669
Like the thing just becomes more and more relevant as they become more and more capable and more people actually wake up to the capability.

24
00:02:36.929 --> 00:02:46.134
I'd actually argue that only a very small slice of the population has even woken up to what they can do today, much less what they're going to be able to do, you know, three months from now.

25
00:02:46.254 --> 00:02:49.076
So I don't see that train slowing down.

26
00:02:50.670 --> 00:03:08.813
In these VVV calls, we're going to go through just updates in the Venice ecosystem, updates to the VVV and DM tokenomics, updates to the Venice products, but we're also going to talk about some AI current events where I want to pick John's brain on some of the things happening around the Venice space in some adjacent companies, just to kind of keep us up with the news.

27
00:03:08.853 --> 00:03:10.854
But I want to start with this tweet from Eric here.

28
00:03:11.214 --> 00:03:12.894
This came in pretty early this month, I think.

29
00:03:12.914 --> 00:03:16.155
September 16th was when this was tweeted, so two weeks ago.

30
00:03:16.555 --> 00:03:16.895
Venice.

31
00:03:17.375 --> 00:03:22.078
Now at 250 billion daily tokens, this is Eric Fork, he's one of the founders of Venice.

32
00:03:22.098 --> 00:03:29.982
He's retweeting himself from a tweet a while ago when he tweeted out that Venice is at 100 billion daily tokens gone through the Venice platform.

33
00:03:30.022 --> 00:03:38.207
Now Venice was processing 100 billion daily tokens on June 25th, 250 billion daily tokens in September 16th.

34
00:03:38.227 --> 00:03:39.268
And if you look at this curve,

35
00:03:40.247 --> 00:03:42.008
I see a skateboard ramp.

36
00:03:42.448 --> 00:03:44.670
I see a curve going upward.

37
00:03:44.690 --> 00:03:48.352
So it's not just growth, but the growth itself is growing.

38
00:03:49.033 --> 00:03:54.756
John, talk to me about what this chart means for you as a strategy and for Venice as a company.

39
00:03:55.917 --> 00:03:57.378
Why is this chart so significant?

40
00:03:58.282 --> 00:04:03.144
Yeah, so I mean, at a high level, I mean, every time Eric posts this chart, it's always one of my favorites.

41
00:04:03.364 --> 00:04:10.067
And I always like to sort of, you know, half facetiously, but not really, like to say it's the best chart in crypto.

42
00:04:10.547 --> 00:04:15.029
Because this is just a chart of the fundamental business at the very base level.

43
00:04:15.049 --> 00:04:20.091
There's obviously a lot more that goes into Venice's business than just how many tokens are we doing a day.

44
00:04:20.691 --> 00:04:22.952
But at a high level, directionally, like...

45
00:04:23.583 --> 00:04:24.644
this is the growth.

46
00:04:24.704 --> 00:04:26.185
This is where everything is.

47
00:04:26.306 --> 00:04:32.291
And like, you know, and I think also when you look at comps, when you look at the open routers, et cetera, they have, you know, similar charts.

48
00:04:32.451 --> 00:04:36.074
AI growth is, AI usage is growing across the industry.

49
00:04:36.495 --> 00:04:38.316
We are certainly experiencing that at Venice.

50
00:04:38.557 --> 00:04:40.799
I think this chart is one of the best examples of just

51
00:04:41.439 --> 00:04:45.040
fundamentally we're just simply doing more tokens every day.

52
00:04:46.000 --> 00:04:49.561
And this chart, I believe, is specifically LLM tokens too.

53
00:04:49.661 --> 00:05:01.563
So like you're not even talking about image models or video models as part of this, but at a base level, it just means more users on the app, more users on the API, more overall usage.

54
00:05:02.043 --> 00:05:07.944
The models themselves get better or cheaper, and so that can also push a bit more token usage.

55
00:05:08.004 --> 00:05:10.405
But yeah, I mean, this story has been pretty consistent.

56
00:05:11.305 --> 00:05:12.705
certainly since the beginning of this year.

57
00:05:14.006 --> 00:05:24.167
When you say that this is the best chart in crypto, your favorite chart, is it accurate to say that if there was one chart to distill down the health of Venice, it would be this one?

58
00:05:24.568 --> 00:05:25.308
Would you say that's accurate?

59
00:05:25.888 --> 00:05:28.988
Yeah, I think at a high level, yeah, that is accurate.

60
00:05:29.408 --> 00:05:37.590
There's obviously things that go into this of how much revenue is Venice actually doing, what do the burns actually look like, and those derivatives, but...

61
00:05:38.171 --> 00:05:39.635
At a high level, this drives everything.

62
00:05:39.776 --> 00:05:44.209
You know, if this is growing, it's very likely everything else in the business is going to continue growing.

63
00:05:45.300 --> 00:05:58.090
I didn't realize that this is only LLM models and image and video models are actually not contained in these tokens because those have to be pretty consumptive in terms of inference, of a total inference.

64
00:05:58.410 --> 00:06:02.354
Do you have any sort of ballpark metric about what this chart encapsulates?

65
00:06:02.994 --> 00:06:08.178
Yeah, so I mean, what I can tell you is image and video have been growing in similar looking charts.

66
00:06:09.479 --> 00:06:12.422
In terms of total amount of tokens, just because of the way those...

67
00:06:13.426 --> 00:06:14.089
Models work.

68
00:06:14.210 --> 00:06:16.479
They're actually, it's not going to be as many tokens.

69
00:06:16.660 --> 00:06:17.644
LMs consume tokens.

70
00:06:18.388 --> 00:06:32.033
Basically the most tokens, because they're especially agentic LLMs tend to be doing multiple terms and they might go do a bunch of things and pull out a bunch of context and then they might even work themselves, you know, on an image model or a video like the LLM may actually interact with those things.

71
00:06:32.653 --> 00:06:44.237
So the actual amount of tokens on the video and the image are not going to be as much, but they're obviously they're much more spendy to, you know, you spend far more on, you know, a turn on a video model than you do on LLM generally.

72
00:06:44.257 --> 00:06:45.258
Yeah.

73
00:06:45.278 --> 00:06:45.378
Yeah.

74
00:06:46.297 --> 00:06:49.138
I want to focus on the fact that this chart is bending upwards.

75
00:06:49.698 --> 00:06:59.340
If you kind of linearly extrapolate the growth on daily token spend, you get to about 1 trillion tokens a day towards the end of January.

76
00:06:59.400 --> 00:07:07.621
But if you lean into the fact that this chart is curving upwards, so the growth is growing, you get to 1 trillion tokens by December.

77
00:07:08.001 --> 00:07:13.542
So I guess the question to you, John, that I have is, when will Venice hit 1 trillion tokens a day?

78
00:07:13.883 --> 00:07:15.143
Before or after 2027?

79
00:07:16.683 --> 00:07:17.584
Yeah, it's a good question.

80
00:07:17.644 --> 00:07:18.484
It's hard.

81
00:07:18.524 --> 00:07:22.405
That would still be a nearly 4x in that time from this point.

82
00:07:22.525 --> 00:07:38.132
So I don't want to be overly aggressive and I'll say probably this happens after December, but myself and many of us internally at Venice have been very wrong about the growth rates a number of times this year with it accelerating far past what we expected.

83
00:07:38.172 --> 00:07:38.412
Yeah.

84
00:07:38.932 --> 00:07:41.615
I don't think a trillion by the end of December would be at all crazy.

85
00:07:42.075 --> 00:07:50.124
But if I had to guess, I think it's probably more likely like something like January or February of next year, just based on the current growth rates.

86
00:07:50.324 --> 00:07:53.247
But yeah, I mean, the growth rate in general has just been nuts.

87
00:07:53.287 --> 00:07:57.512
And a lot of our models and forecasting, at the beginning of this year, we assumed something like,

88
00:07:58.200 --> 00:08:03.344
you know, if things were going amazing, that we'd be doing something like 15 to 20% month over month growth.

89
00:08:03.505 --> 00:08:07.768
And in the actuality of this year, I think it's been closer to like 35 or 40%.

90
00:08:07.808 --> 00:08:10.951
So we just tend to underestimate our own growth.

91
00:08:11.872 --> 00:08:13.253
Yeah, yeah.

92
00:08:13.353 --> 00:08:20.339
One thing that is nice that we have in the crypto ecosystem, I mean, I'm speaking as a crypto investor here, is that like,

93
00:08:20.559 --> 00:08:21.560
AI is very exciting.

94
00:08:21.660 --> 00:08:23.303
Anthropic, open AI is very exciting.

95
00:08:23.323 --> 00:08:24.784
People really want exposure to those things.

96
00:08:24.864 --> 00:08:29.110
It's nice to be able to see the growth of AI demand in a chart.

97
00:08:29.170 --> 00:08:36.419
This is like what that chart would look like, you know, slightly up and to the right with a slight crescendo of acceleration on growth.

98
00:08:36.719 --> 00:08:42.682
And it's also nice just to be able to have an instrument that has exposure to the economics of those tokens.

99
00:08:43.563 --> 00:08:49.306
Maybe there's other tokens out there in the crypto ecosystem that I'm not familiar with, like Tau comes to mind.

100
00:08:49.346 --> 00:08:58.832
But having something, having an instrument that has strong correlation to just AI inference demand, it's nice that we have that option in the cryptosphere.

101
00:08:59.598 --> 00:09:01.299
Yeah, yeah, I mean, I agree.

102
00:09:01.479 --> 00:09:06.143
That's definitely one of the reasons, you know, we have done the things we have done and we think it's important.

103
00:09:06.183 --> 00:09:20.154
And, you know, even as you started this conversation talking about these IPOs coming up from Anthropic and OpenAI and all these things, you know, we think it's important that people have access to something like this so that they can hopefully help.

104
00:09:20.734 --> 00:09:32.386
participate in all of this enormous wealth that AI is creating that, you know, many people, if they try to buy into those frontier lab IPOs, they're not going to get any exposure to until it's, you know, well north of a trillion dollars, most likely.

105
00:09:33.087 --> 00:09:33.708
Yeah, yeah.

106
00:09:34.591 --> 00:09:36.413
Let's talk about VVV emissions.

107
00:09:36.593 --> 00:09:44.461
We've talked about this for the last two VVV calls, but as we all knew, VVV emissions has gone down from 3 million a year down to 2.5 million a year.

108
00:09:44.481 --> 00:09:46.884
This is the first of two emission reductions.

109
00:09:46.904 --> 00:09:52.469
VVV issuance will be going down to 2 million a year again later in October, in a few days, actually.

110
00:09:52.509 --> 00:09:55.212
So if people, on Thursday, yeah.

111
00:09:55.232 --> 00:09:55.713
So it's Monday.

112
00:09:55.753 --> 00:09:57.394
On Thursday, it'll be down to 2 million a year.

113
00:09:58.295 --> 00:10:01.598
What do we know about the final equilibrium of VVV issuance?

114
00:10:01.638 --> 00:10:03.380
Do we like the 2 million number?

115
00:10:03.700 --> 00:10:05.561
How do we determine that that was correct?

116
00:10:05.762 --> 00:10:10.265
What would change your mind about that needing to either like increase or decrease further?

117
00:10:10.946 --> 00:10:11.687
What are the inputs here?

118
00:10:12.738 --> 00:10:16.780
Yeah, I mean, we look at a number of things as these emissions come down.

119
00:10:16.840 --> 00:10:23.603
So obviously that affects basically staking yield the most.

120
00:10:23.703 --> 00:10:31.007
What the yield is is both determinant of what is the emissions, but also how much VVV is actually staked.

121
00:10:31.627 --> 00:10:34.389
So those things kind of balance each other a little bit.

122
00:10:34.889 --> 00:10:40.573
So we'll be certainly watching closely what happens after this next emission reduction, how this all plays into things.

123
00:10:41.233 --> 00:10:48.138
I would say in terms of where we want to end up, we're still, my guess is even at two million a year, we're still going to probably be higher than...

124
00:10:48.798 --> 00:10:54.841
Then we need to be, you know, the goal is still a deflationary asset and that's likely going to require more emissions reduction.

125
00:10:54.881 --> 00:11:04.947
So this is our, you know, this is what we have announced and planned right now, but there will likely be more communication from that on us, maybe as soon as, you know, mid-October or something like that.

126
00:11:05.948 --> 00:11:11.715
Okay, so not official comms from Venice, but John's hunch is that 2 million might be on the high side.

127
00:11:12.075 --> 00:11:26.892
So the near-term plan for you and Venice is to get down to the 2 million yearly issuance number on Thursday and then see how the market reacts, see if people stake more, see if people unstake, see what people do, let the dust settle for a little bit, collect some data, and then proceed from there.

128
00:11:27.505 --> 00:11:29.887
Yep, I would say that's pretty much accurate.

129
00:11:30.207 --> 00:11:32.909
No decisions have been made past the current ones that have been announced.

130
00:11:32.989 --> 00:11:34.931
But yeah, we're always watching this closely.

131
00:11:35.031 --> 00:11:39.374
And, you know, we really mean it when we say we want this asset to become deflationary.

132
00:11:39.394 --> 00:11:43.758
And that's going to mean being very careful with emissions and the supply in general.

133
00:11:45.179 --> 00:11:48.260
Speaking of the supply, the DM target has been going up.

134
00:11:48.340 --> 00:11:51.761
It's now up to 40,000 is the new DM supply target.

135
00:11:52.282 --> 00:11:53.342
So same kind of question.

136
00:11:53.502 --> 00:11:54.542
What are the inputs?

137
00:11:54.603 --> 00:11:57.003
What's the data that you're looking at?

138
00:11:57.183 --> 00:12:01.185
Because the DM supply target has been going up for, I think, six weeks in a row now.

139
00:12:02.225 --> 00:12:03.326
What data have you got?

140
00:12:03.346 --> 00:12:09.028
What have you observed to kind of inform your decision making about whether this equilibrium of 40,000 is correct or not?

141
00:12:10.053 --> 00:12:11.614
Yeah, so same kind of thing.

142
00:12:11.654 --> 00:12:13.495
We observe the changes in real time.

143
00:12:13.715 --> 00:12:25.282
We see what impact this actually has on people staking, locking their stake to mint, whether they burn some of their diem to get their stake back and just where how much BBV is actually locked

144
00:12:26.583 --> 00:12:35.091
And what we had been seeing is that there was definitely an acceleration on the locking as this had happened, which was kind of what we hoped to and expected to see.

145
00:12:35.912 --> 00:12:46.382
But that has definitely come back down a little bit in the last week, mostly driven, I think, by price action and some people wanting to unstake their VVV, understandably, which required them to burn some DMs.

146
00:12:46.842 --> 00:12:50.103
So this one actually, it's going to be, it really, really need a little more time.

147
00:12:50.143 --> 00:12:56.084
So we only just had the last target increase take effect, what, like a week or two ago.

148
00:12:57.545 --> 00:13:01.806
And really this one is a more sensitive one in some ways.

149
00:13:02.346 --> 00:13:16.329
I want to see, you know, some months of data of how the market actually responds to this thing long-term where the VVV supply actually ends up now that we have this new target and have made it easier for people to mint some diem.

150
00:13:17.009 --> 00:13:19.191
Um, but that's not, you know, it's not in a vacuum.

151
00:13:19.291 --> 00:13:21.593
You have all these other market effects going on.

152
00:13:21.654 --> 00:13:25.397
So I think over time, this will be less noisy in the short term.

153
00:13:25.457 --> 00:13:26.278
It's a little bit noisy.

154
00:13:27.182 --> 00:13:27.922
Right, right, right.

155
00:13:27.942 --> 00:13:28.122
Yeah.

156
00:13:28.442 --> 00:13:33.964
It was something that you said, people will want to unstake their VVB when the VVB price goes up.

157
00:13:34.464 --> 00:13:37.865
September has been an incredible month for VVB price performance.

158
00:13:38.405 --> 00:13:49.128
And so does that imply that when VVB goes up in price, there is a DM demand because people who want to get liquidity at that higher VVB price have to go buy...

159
00:13:49.868 --> 00:13:53.010
back DM off the market so they can unlock their VVV.

160
00:13:53.070 --> 00:14:00.974
And so that can kind of like throw a little curveball, I think, in the stability of the VVV-DM relationship when there's a bunch of VVV volatility to the upside.

161
00:14:01.951 --> 00:14:08.735
Yeah, this is very much what I would call a feature, not a bug of the tokenomics.

162
00:14:09.336 --> 00:14:18.682
If anybody mints and sells VVV, what they're effectively doing is betting that the DM price, at least, will be lower at some point.

163
00:14:19.022 --> 00:14:23.825
That they'll be able to buy that DM back at a lower price and go from there.

164
00:14:24.166 --> 00:14:25.867
When you have things like this.

165
00:14:26.747 --> 00:14:41.673
When the price is going up, it basically creates an incentive for those to want to buy back their Diem, which naturally means it brings some more buy demand to Diem because generally, unless they're still holding that Diem, they're going to have to buy it off the market somewhere if they ever want to unlock their BBB.

166
00:14:41.693 --> 00:14:45.435
So yes, it creates kind of this natural pendulum swing.

167
00:14:46.035 --> 00:14:49.157
where when people want to unlock, they have to buy DM.

168
00:14:50.238 --> 00:14:57.824
When BVB price and DM are accelerating, it creates the opposite natural incentive for people to want to mint and potentially sell that into the market.

169
00:14:57.964 --> 00:15:03.067
And so the idea is that those kind of always, that pendulum swings back and forth and you get a relative equilibrium.

170
00:15:03.108 --> 00:15:03.528
And I think,

171
00:15:03.948 --> 00:15:16.760
It's interesting to point out, obviously, these things are correlated, but, you know, when we first announced that we were doing this DM supply, there was at least some feedback from a few community members who were concerned that we were, quote unquote, inflating DM.

172
00:15:17.561 --> 00:15:20.223
I still don't think that's really the right way to look at this.

173
00:15:20.764 --> 00:15:26.049
And I think the market kind of bared that out, like DM prices actually up and to the right, even as supply increased.

174
00:15:27.649 --> 00:15:27.949
Yeah.

175
00:15:28.089 --> 00:15:36.817
In the last like six months to a year or so since the VVV price went from like a dollar or two up to where it is now at like $30.

176
00:15:37.117 --> 00:15:45.886
This has been the first real substantive like price appreciation that has also the tail kind of has wagged the dog with impacting the VVV and DM economics.

177
00:15:46.386 --> 00:15:50.467
Now, with any future price appreciation, at least we have like some hindsight to look back on.

178
00:15:50.487 --> 00:15:51.807
We have some data to look back on.

179
00:15:51.827 --> 00:16:00.269
But nonetheless, there's always just like some inherent instability in the smoothness of everything when there is like rapid price appreciation.

180
00:16:00.649 --> 00:16:10.852
And also the DM economy so young as well that like this is really the first data set that you guys are really able to work with to make sort of like informed decisions.

181
00:16:10.892 --> 00:16:11.312
Is that right?

182
00:16:11.332 --> 00:16:11.752
Yeah.

183
00:16:12.223 --> 00:16:19.162
Yeah, I mean, yeah, the DM target supply in particular had stayed the same for, you know, a year since launch.

184
00:16:19.202 --> 00:16:19.944
And then this was our...

185
00:16:20.825 --> 00:16:33.873
our first actual change to the target and for us to see a chance of how that actually impacts the economics, how that actually impacts how many people are minting, what people are willing to pay for a DM, all these things that the market gives us good feedback on.

186
00:16:34.834 --> 00:16:43.560
But yeah, I do think it'll take a little time just like the initial DM tokenomics took, you know, I would say some months to kind of settle out into an equilibrium.

187
00:16:44.060 --> 00:16:46.922
I think we're still going to probably need a few months of time as people

188
00:16:47.442 --> 00:16:55.705
adjust to all these various changing variables to settle out what I think the long-term impact of this change actually is.

189
00:16:55.745 --> 00:16:58.486
So today, like I said earlier, I think it's a little bit noisy.

190
00:16:58.846 --> 00:17:05.329
There's so much going on that it's hard to draw too many conclusions from any of this right now, but I think within a couple of months, we'll have a better view.

191
00:17:06.881 --> 00:17:12.827
Now on previous VVV calls, we've talked plenty about the different mechanisms for how VVV gets burned.

192
00:17:13.087 --> 00:17:16.630
There's an automatic burn triggered by any time a new subscription comes.

193
00:17:16.650 --> 00:17:21.095
Somebody signs up for Venice, they buy a subscription, an amount of VVV is burned.

194
00:17:21.115 --> 00:17:25.359
There's also a 5% of API credits that people spend.

195
00:17:25.379 --> 00:17:29.523
So if somebody buys $100 of API credits, 5% of that goes to buy and burn VVV.

196
00:17:30.083 --> 00:17:34.467
One thing we haven't covered, one mechanism we haven't covered is the discretionary burn.

197
00:17:35.348 --> 00:17:37.270
Discretionary can mean a lot of things.

198
00:17:37.710 --> 00:17:39.632
It's a pretty big term.

199
00:17:40.032 --> 00:17:42.735
What does it mean for the burn to be discretionary?

200
00:17:43.335 --> 00:17:44.997
Can we discretionarily burn more?

201
00:17:45.477 --> 00:17:47.479
Can we discretionarily burn less?

202
00:17:47.499 --> 00:17:51.123
What are the variables that go into this discretionary word?

203
00:17:52.008 --> 00:18:03.961
Yeah, so the variables, so it's kind of right there in the word in that this component of the burn today is ultimately up to the team and the team's discretion.

204
00:18:05.282 --> 00:18:09.526
That generally, generally this has tracked more or less, you know, loosely

205
00:18:10.307 --> 00:18:12.589
you know, not exactly, but loosely revenue growth.

206
00:18:13.269 --> 00:18:20.816
And this was the first burn bucket that was introduced into VVV's tokenomics before we added any of the programmatic burns.

207
00:18:21.356 --> 00:18:32.105
So we have continued to use this lever as a way to effectively, you know, kind of capture some of the revenue appreciation and buy and burn that we

208
00:18:32.905 --> 00:18:35.566
We have not directly put into anything programmatic yet.

209
00:18:36.627 --> 00:18:48.213
What's interesting is I think this might have been, this was either the first or second month where the programmatic burns, I think, eclipsed the discretionary burn, which is great to see, even though the discretionary burn has continued to grow.

210
00:18:49.453 --> 00:19:16.902
so that's honestly a great sign exactly what goes into it is it's ultimately the team looking at what we think makes sense based on high level revenue numbers high level growth numbers what we think we can put into this and how the programmatic is scaling I think in the long run over the long term it might make sense right now we want to keep growing discretionary I think over the long term it might make sense for discretionary to kind of take a backseat to programmatic as the

211
00:19:17.850 --> 00:19:21.093
as the business really continues to grow and take off, hopefully.

212
00:19:21.513 --> 00:19:27.398
If that's the case, then I think we want more and more of that to be on the programmatic side and less up to us or anything like that.

213
00:19:27.478 --> 00:19:32.142
But for now, it's a nice lever for us to basically add on to the burn every month.

214
00:19:32.422 --> 00:19:34.504
And we're going to continue doing that at least for a bit.

215
00:19:35.671 --> 00:19:50.774
Is the idea here that on a month-to-month basis or on a quarterly basis, the team opens up the Venice P&L, the Venice team has an amount of burn associated with it, you guys have to pay salaries, you guys have costs, you have data center costs that you have to pay for.

216
00:19:51.375 --> 00:19:55.839
But then you also have like the revenue and you also have already the programmatic burns.

217
00:19:55.960 --> 00:19:58.842
And then when you net everything together, there's some money left over.

218
00:19:59.283 --> 00:20:08.252
And out of that money left over, you discretionarily decide to use some of that money, some arbitrary amount to buy and burn VVV as you guys so choose.

219
00:20:08.332 --> 00:20:09.053
Is that kind of how it works?

220
00:20:09.843 --> 00:20:26.326
Yeah, at a high level, I think that's a good way to put it, is it's some portion of our free cash flow after all those various things you mentioned, all the various expenses, all the imprints we need to buy, the salaries we need to pay of, in addition to our programmatic burns, what else can we put back here into the burn?

221
00:20:26.886 --> 00:20:32.487
And it's not like we decide to be clear, like on one number and then on one day this happens.

222
00:20:33.447 --> 00:20:37.428
These buys, these discretionary burns, just like programmatic burns, are actually happening all the time.

223
00:20:38.158 --> 00:20:50.417
effectively there's just a constant TWOP going in addition to the programmatic burns that's just buying all the time and that adds up over the course of a month and then you end up with a discretionary burn member based on that

224
00:20:51.692 --> 00:20:53.894
Right, so it is like the excess cash.

225
00:20:54.134 --> 00:20:57.555
It's not like a market smash the buy button on VVV.

226
00:20:57.596 --> 00:21:15.265
It's a TWOP, but the idea is like there's excess cash that Venice has rightly created a buffer with, and ultimately at the end of a month, or I don't know how long, a quarter, you guys have some like slush fund built up, and you're like, okay, let's TWOP that into VVV and burn it, and that's a discretionary burn.

227
00:21:15.845 --> 00:21:18.827
Yep, I think at a high level, that's the right way to think about it.

228
00:21:19.469 --> 00:21:22.390
I like my things being like 90% accurate.

229
00:21:22.410 --> 00:21:25.151
If I'm like 90% directionally correct, I think we're doing good.

230
00:21:25.551 --> 00:21:33.734
One other change that I saw that happened this week is the increase of the programmatic VVV burns from pro subscriptions now burns $2.

231
00:21:33.794 --> 00:21:39.396
If anyone signs up for a pro sub, it now burns $2 in VVV, which is doubling from a $1 VVV

232
00:21:40.896 --> 00:21:41.516
Same kind of question.

233
00:21:41.576 --> 00:21:50.782
What is the input into what gives you guys the room, the tolerance, the flexibility to feel that it's safe to like double the amount of VVV burn in the economics of Venice?

234
00:21:51.682 --> 00:21:54.664
Yeah, so this isn't, this is back actually in April.

235
00:21:54.944 --> 00:21:55.404
It looks like.

236
00:21:55.464 --> 00:21:56.425
Oh, this is in April.

237
00:21:56.745 --> 00:21:56.965
Yeah.

238
00:21:57.305 --> 00:21:57.906
Oh, this is old.

239
00:21:58.146 --> 00:22:05.390
Okay, well, similar kind of question though, because like this does kind of show that there's going to be changing VVV economics with a programmatic burns.

240
00:22:05.410 --> 00:22:07.191
Maybe you guys rate it up, rate it down.

241
00:22:07.511 --> 00:22:08.712
How do you guys make those decisions?

242
00:22:09.537 --> 00:22:15.919
Yeah, so I mean, we're still, you know, this is all kind of a work in progress, but we try to right-size these based on it.

243
00:22:15.939 --> 00:22:22.500
So when we first released these programmatic burns, it was just a basic $1 for whether it was Pro, Pro Plus, Max.

244
00:22:23.060 --> 00:22:36.084
And then shortly after that, like this was probably no more than a month after we originally released these programmatic burns, we made this change so that we wanted to stratify it so that there's a bit more burned per month

245
00:22:36.844 --> 00:22:42.252
per sub, and that it's actually more in accordance with the level of sub that people are buying.

246
00:22:42.512 --> 00:22:49.523
Part of this too was to just, this also allows the on-chain data to be more readable and knowable by something like AvenaStats.

247
00:22:49.603 --> 00:22:51.345
We know when it was all a dollar, you couldn't tell

248
00:22:52.066 --> 00:22:58.572
what was pro, what was pro plus, what was max, uh, with the different amounts, that's much easier for that on-chain data analysis to happen.

249
00:22:58.992 --> 00:23:06.698
And then obviously, you know, for us, it was just like, we wanted to have kind of a roughly similar proportion of the sub payment that goes to the burn.

250
00:23:06.779 --> 00:23:09.221
So the higher ones burn more, um,

251
00:23:09.801 --> 00:23:10.421
That was really it.

252
00:23:10.461 --> 00:23:15.064
This is something we will, you know, periodically reevaluate if something makes sense.

253
00:23:16.045 --> 00:23:20.127
And certainly if we did something like change the subscription pricing, this would probably change with it.

254
00:23:20.527 --> 00:23:25.450
But right now it's kind of just based, you know, based off the cost of those subs and what we think makes sense.

255
00:23:27.052 --> 00:23:42.938
When you guys make decisions like this, is one of the inputs into your guys' decision making, if we make it this number, that puts more data on chain so that the VVV community, mainly Gecko on Venice stats, but the VVV community gets more data to work with.

256
00:23:43.038 --> 00:23:47.280
And so it's like a selective disclosure of the internals of Venice.

257
00:23:47.660 --> 00:23:51.541
And you guys actually do think about when and when not to put data on chain.

258
00:23:51.581 --> 00:23:53.362
Is that something, a conversation that you guys have?

259
00:23:54.260 --> 00:24:01.367
Yeah, I would say that's definitely a conversation we have about what is the right data to put it on chain, what's the right way to do it and how to differentiate it.

260
00:24:01.487 --> 00:24:12.417
So like, yeah, actually, you know, like there was actually a problem when we when we added the the credit purchases as Burns because they had the same five dollar number as ProPlus.

261
00:24:12.977 --> 00:24:16.638
So you'll actually, if you look on chain, you'll notice there's actually a slight difference between those.

262
00:24:16.698 --> 00:24:24.259
It's like a five or six cent difference, but it's enough of a difference that those who are doing on chain analytics can tell the difference between those things.

263
00:24:24.299 --> 00:24:26.580
So it is definitely something that we think about.

264
00:24:27.460 --> 00:24:27.740
Nice.

265
00:24:28.140 --> 00:24:35.902
I like the idea of this event is intentionally leaving very explicit breadcrumbs for those who look for it can have the data that they want to find.

266
00:24:37.170 --> 00:24:40.652
Let's turn into some of the market commentary that happened this week.

267
00:24:41.092 --> 00:24:51.117
On September 8th, I believe, VVV broke through all-time highs on the same day as OpenAI had this Navier-Stokes Millennium Prize problem controversy.

268
00:24:51.537 --> 00:24:52.298
This was a big day.

269
00:24:52.398 --> 00:24:57.060
I think if anyone was holding VVV, they remember where they were when they were reading the news this morning because they—

270
00:24:57.320 --> 00:25:01.604
Maybe they did what I do and I looked at the charts and I'm like, oh, wow, wait a second.

271
00:25:01.624 --> 00:25:02.686
That's a very different price.

272
00:25:02.866 --> 00:25:06.389
Or they were looking at Crypto Twitter and Crypto Twitter was blowing up all about this anyways.

273
00:25:06.429 --> 00:25:07.891
But I'll kind of go through the story here.

274
00:25:08.792 --> 00:25:20.984
VVV started the day at $18 and peaked at $30 while a couple of mathematicians accused OpenAI of spying and stealing their work that they had prompted into OpenAI Codex.

275
00:25:21.824 --> 00:25:30.706
And so the TLDR of the controversy was that OpenAI claimed they found a solution to the $1 million millennium prize math problem.

276
00:25:30.907 --> 00:25:33.067
There's a handful of these math problems, and this is one of them.

277
00:25:33.747 --> 00:25:39.169
And this particular math problem involves like fluid dynamics that had like stumped mathematicians for something like 200 years.

278
00:25:40.700 --> 00:25:53.867
Roughly 12 hours before OpenAI claimed credit, two mathematicians made a public statement, said that they had been closing in and on solving the same problem using AI tools, including OpenAI's codex.

279
00:25:53.927 --> 00:26:06.673
Tristan, one of the mathematicians, insinuated that OpenAI may have secretly accessed his notes that he had deposited into codex because OpenAI's solution to the problem basically had the same logic processed

280
00:26:07.293 --> 00:26:08.916
that Tristan was using with.

281
00:26:08.956 --> 00:26:17.589
And so the claim is that OpenAI just used Tristan's work as like a jumping off problem after spying on his work, using his work that he had submitted into OpenAI.

282
00:26:17.669 --> 00:26:20.953
OpenAI didn't explicitly deny that this was a possibility and

283
00:26:22.215 --> 00:26:40.960
Overall, this, I think, awoken a lot of people outside of crypto Twitter, outside of people around Venice, to the idea that the big AI labs are kind of looking over your shoulder at your homework to see what value that they can get from whatever you submit into OpenAI, Anthropic, or whatever.

284
00:26:41.581 --> 00:26:45.382
John, give me a peek into the Venice Slack when this happened.

285
00:26:45.542 --> 00:26:46.622
What did you guys talk about?

286
00:26:46.642 --> 00:26:47.302
What did you think?

287
00:26:47.322 --> 00:26:48.362
What was your reactions?

288
00:26:48.843 --> 00:26:49.943
What was being said?

289
00:26:50.863 --> 00:26:53.505
Yeah, I mean, I think this was another one of those events.

290
00:26:53.565 --> 00:27:07.314
I mean, I feel like almost every month we end up talking about, like, this one was particularly big, obviously, and there was a big reaction online about it, you know, obviously outside of just crypto Twitter, but across, you know, all AI, people were very interested in this subject.

291
00:27:07.854 --> 00:27:10.736
And I think it's just another one of those reminders to people that, like,

292
00:27:11.448 --> 00:27:20.533
Yeah, actually, keeping your thoughts private might actually matter, especially if you're working on new novel, interesting things like, you know, novel mathematics solutions.

293
00:27:21.694 --> 00:27:27.397
Everything you put into a frontier model generally is kind of just getting hoovered up by default.

294
00:27:27.457 --> 00:27:37.003
And they don't always even know themselves what, like that was kind of part of the conversation here is OpenAI was like, well, we didn't specifically do anything here, but we can't rule out

295
00:27:37.423 --> 00:27:43.408
that that information basically had been taken and put into the training data of our newer models.

296
00:27:44.128 --> 00:27:45.729
And that's kind of always the case.

297
00:27:45.829 --> 00:27:52.014
Like they don't have like a whole list where they understand every single piece of information that goes into the training data.

298
00:27:52.394 --> 00:27:57.878
And these models, they want, you know, at a core level, they want novel information.

299
00:27:57.958 --> 00:28:03.082
Like that's kind of the best thing to put into the training of these things is any sort of novel information.

300
00:28:03.662 --> 00:28:07.405
novel ideas, novel theories, like novel mathematics solutions.

301
00:28:07.905 --> 00:28:20.094
So it's really hard to divorce the fact that these guys had been working in codex and likely supplying a lot of these and then OpenAI just happens to announce that they've solved this longstanding mathematics problem.

302
00:28:20.474 --> 00:28:25.137
I think it was another reminder to folks about the importance of privacy in AI.

303
00:28:25.618 --> 00:28:31.142
Obviously, we were very interested and excited about this from a Venice standpoint, but for us, it's kind of like

304
00:28:31.762 --> 00:28:50.218
another day, you know, another day at the office, another example of exactly why Venice exists, exactly to solve problems like this so that people don't have to worry that everything they're doing, the company's looking over their shoulder, that every thought they have is just getting taken and put into AI training data.

305
00:28:50.298 --> 00:28:52.340
And like, even when someone, you know,

306
00:28:53.224 --> 00:28:55.785
checks the thing on the Frontier Labs that says don't train on this.

307
00:28:55.906 --> 00:29:06.471
It's just like what level of verification, if any, do they have that that's actually what's happening and that something doesn't get scrubbed and anonymized and just brought into the training data because it's useful.

308
00:29:06.691 --> 00:29:07.672
Like we just don't know.

309
00:29:07.992 --> 00:29:10.053
And so you have to trust the labs at their word.

310
00:29:10.113 --> 00:29:13.235
And even at their word, they don't really know.

311
00:29:13.275 --> 00:29:20.099
So I think it's just a good reminder to everyone about the fact that when you talk to these things that they are likely taking everything you give it.

312
00:29:21.213 --> 00:29:28.597
Yeah, I think the conversation downstream of how actually the data got into hands of OpenAI was the most interesting.

313
00:29:29.097 --> 00:29:34.780
Because OpenAI distanced themselves from like, hey, we didn't quite know that this exactly was happening.

314
00:29:35.100 --> 00:29:38.722
But nonetheless, we can't rule it out that our models weren't trained on the data.

315
00:29:39.222 --> 00:29:43.305
And when they have to distance themselves from how they can't rule it out, to me...

316
00:29:44.045 --> 00:29:45.527
that shows how structural it is.

317
00:29:45.827 --> 00:29:49.811
As in, this is built into the way that the company works.

318
00:29:50.071 --> 00:29:58.500
As in, they must steal your data in order to have an edge over their competition, who, by the way, they are in an arms race with.

319
00:29:58.921 --> 00:30:03.666
And so it's kind of this toxic arms race that forces the Frontier AI labs

320
00:30:04.066 --> 00:30:06.948
to need to be able to access as much data as possible.

321
00:30:07.348 --> 00:30:14.772
And I think this is why I think societies, like heckles are being raised at some of the AI labs because they've seen this pattern before, right?

322
00:30:14.812 --> 00:30:21.295
Like Zuckerberg with Facebook, if the product is free, maybe the product is you actually.

323
00:30:21.335 --> 00:30:24.877
And so people are kind of pattern recognizing with the frontier AI labs.

324
00:30:25.257 --> 00:30:31.360
And I think the frontier AI labs are losing a lot of trust in society because of stories like this.

325
00:30:32.460 --> 00:30:33.860
Yeah, I think exactly.

326
00:30:33.920 --> 00:30:40.501
I think it's just, it reminds so many people use these products and don't think about these problems until they get stories like this.

327
00:30:40.662 --> 00:30:46.783
And they either are materially affected or in a similar field or something where they actually have to worry about this and then they think harder about it.

328
00:30:47.303 --> 00:30:48.443
Or something else comes along.

329
00:30:48.463 --> 00:30:59.505
You know, sometimes this story is a subpoena because people realize that the company has all their information and it comes up in like some various lawsuit where that information is discoverable.

330
00:31:00.125 --> 00:31:09.653
Or sometimes it's a story like this where, you know, some novel solution to a mathematics problem is seemingly taken from the people who had actually been working on it.

331
00:31:09.753 --> 00:31:12.896
And the big labs gain, you know, credit for it.

332
00:31:13.016 --> 00:31:15.098
And I think the labs themselves are even...

333
00:31:15.938 --> 00:31:24.283
reconsidering how they put out information about solutions like this because of this exact problem and worrying about how people respond to that.

334
00:31:24.584 --> 00:31:26.085
But that's not necessarily a good thing either.

335
00:31:26.145 --> 00:31:31.868
Like, you know, to the degree that we are getting things like novel solutions in mathematics, that's something we want widely shared.

336
00:31:31.928 --> 00:31:33.029
That's good for society.

337
00:31:33.089 --> 00:31:36.151
Those are who knows what things could be built on top of things like that.

338
00:31:37.512 --> 00:31:43.778
And we don't want the labs to be in an incentive where they don't even want to share that such a thing has occurred or not share the proof.

339
00:31:44.279 --> 00:31:45.640
Like we absolutely want that.

340
00:31:46.000 --> 00:31:53.308
But we also want them to, you know, respect those who are actually working on these things and have, you know, novel views in their field.

341
00:31:54.153 --> 00:32:01.698
and not just getting hoovered up and people just think it's so easy to forget that exactly like you said, like these are built into how these companies work.

342
00:32:01.778 --> 00:32:06.501
This is not like a one-off or they forgot to click the button or whatever it is.

343
00:32:06.861 --> 00:32:08.102
This wasn't oh, oopsies.

344
00:32:08.362 --> 00:32:19.529
No, these are products and companies that are built off of these massive, you know, or hugely expansive data collection because that's how you train the best model.

345
00:32:19.969 --> 00:32:24.331
And they have such competition with each other, with China, across the world.

346
00:32:24.791 --> 00:32:26.492
And these things are really not slowing down.

347
00:32:26.512 --> 00:32:34.715
Obviously, there was some various discussion that happened, you know, around potential slowdowns and among like Dario and Sam and some of the big labs, but...

348
00:32:35.195 --> 00:32:38.497
The reality is they just, they cannot slow down.

349
00:32:38.637 --> 00:32:48.683
They're just everybody, everyone in the world, in my opinion, is caught in this like prisoner's dilemma where if you slow down, you know, the result seemingly is worse.

350
00:32:48.723 --> 00:32:50.484
And so they're all trying to find a way to balance this.

351
00:32:50.524 --> 00:32:55.226
And it generally means they're going to use every edge they can, including every piece of data they can get their hands on.

352
00:32:55.967 --> 00:32:56.167
Yeah.

353
00:32:56.724 --> 00:33:05.306
You used the illustration of what if a lawsuit happens to one of these AI labs and you as a user, your data gets caught up in that lawsuit.

354
00:33:05.366 --> 00:33:13.269
Well, John, one of the news this week was that OpenAI must turn over 20 million chat GPT logs, according to a judge.

355
00:33:13.309 --> 00:33:16.210
So this is not just an illustration, not just an anecdote.

356
00:33:16.730 --> 00:33:23.492
This is real that OpenAI Inc. will have to turn over 20 million anonymized chat GPT logs.

357
00:33:23.512 --> 00:33:24.292
So there's something there.

358
00:33:24.732 --> 00:33:33.939
And a consolidated AI copyright case after it failed to convince a federal judge to throw out the judge's order about insufficient privacy concerns.

359
00:33:34.559 --> 00:33:37.001
And so this is happening in reality.

360
00:33:37.581 --> 00:33:44.026
And this was earlier this year in January, but I think the testament, the evidence is here nonetheless.

361
00:33:44.726 --> 00:33:51.714
Yeah, I mean, I'm sure you could do a search or ask AI about it and find 10 or 20 other such examples just like this.

362
00:33:51.814 --> 00:33:54.497
Like, I feel like I hear one every other week.

363
00:33:54.777 --> 00:33:57.340
Like, there's always something going on.

364
00:33:58.001 --> 00:34:04.068
Legal cases are kind of canaries in the coal mine because they will push for whatever data a company actually has.

365
00:34:04.900 --> 00:34:10.762
Um, you know, this is one of our barometers at Venice is like, how secure actually is the data?

366
00:34:11.062 --> 00:34:13.202
How effectively subpoena proof is it?

367
00:34:13.282 --> 00:34:15.063
Like we can only turn over what we have.

368
00:34:15.203 --> 00:34:16.964
You know, we get a legal request to Venice.

369
00:34:16.984 --> 00:34:18.024
We're going to turn over what we have.

370
00:34:18.064 --> 00:34:20.865
But the key is that we just simply don't have anything.

371
00:34:21.305 --> 00:34:22.945
There's nothing to turn over if you don't have it.

372
00:34:23.125 --> 00:34:29.007
But all these companies, they very much have it because it's their whole business model is built off having and using that data.

373
00:34:30.417 --> 00:34:39.187
I think two calls ago, John, we talked about the luxury that Venice has of not having to be in the arms race of model development.

374
00:34:39.507 --> 00:34:40.909
You guys don't develop models.

375
00:34:40.969 --> 00:34:42.230
That's not what you guys do.

376
00:34:42.450 --> 00:34:43.952
It's not the business that you guys have.

377
00:34:44.412 --> 00:34:49.258
You guys allow the rest of the world to develop models, and then you just plug it into the Venice platform.

378
00:34:49.698 --> 00:34:57.000
And I think that removes you from the AI arms race in a way that must feel kind of nice because A, you don't have to have the burden of training the models.

379
00:34:57.060 --> 00:34:58.601
That's not a cost that you guys have to bear.

380
00:34:58.621 --> 00:35:08.304
And as a result of not being in the arms race, you don't have to feel compelled to retain users' data or do anything you can to get your hands on more data.

381
00:35:08.704 --> 00:35:11.045
So I feel like it gives you guys a much more...

382
00:35:12.465 --> 00:35:17.575
relaxed position in the market where you don't have to worry about these sorts of things.

383
00:35:18.196 --> 00:35:18.957
That's my intuition.

384
00:35:18.997 --> 00:35:19.578
Does it feel that way?

385
00:35:20.654 --> 00:35:28.581
Yeah, I mean, this has very much been a strategic decision on Venice's part to not wade into this field of model training and everything that comes with it.

386
00:35:28.621 --> 00:35:33.786
And I think so far, it's very much been the right decision for Venice for the product we want to offer.

387
00:35:34.287 --> 00:35:37.369
This is not necessarily our expertise or our skill set.

388
00:35:37.870 --> 00:35:40.813
Doesn't mean we could never look at it or never lean into such a thing.

389
00:35:41.253 --> 00:35:46.518
But generally, you know, there's just so many, so many labs spending so much money on all these models.

390
00:35:47.058 --> 00:35:49.421
that it's better for us to just see them come out.

391
00:35:49.461 --> 00:35:58.192
And I, you know, I'm very happy personally that we don't have the stress of training these models because, you know, these companies, they spend so much time and effort on training these models.

392
00:35:58.232 --> 00:36:04.700
And at this point, it's like, if they're lucky, they get to basically, you know, their model gets to shine in the sunlight for years

393
00:36:05.000 --> 00:36:07.804
maybe two weeks before it's surpassed by something else.

394
00:36:08.264 --> 00:36:15.233
So it's just so much stress to get that thing out there as fast as possible so that you can maximize that time that you get in the sunlight.

395
00:36:15.533 --> 00:36:24.364
Where for us, it's just like, you know, Eric has this example that, or this analogy that I like to use where he basically says it's for us, it's like manna raining from heaven.

396
00:36:24.884 --> 00:36:29.950
of all of these models that just get dropped and plugged into Venice and useful for our users and great for us.

397
00:36:30.030 --> 00:36:32.213
And we don't care which one is the hot one right now.

398
00:36:32.654 --> 00:36:34.356
That's not it's just not really our game.

399
00:36:34.416 --> 00:36:43.747
So we the fact that we've kind of opted out of that, I think, is a big structural advantage for for us and how our company works compared to a lot of other projects.

400
00:36:45.066 --> 00:36:48.668
Mana reigning from heaven is very poetic from Eric, as always.

401
00:36:48.968 --> 00:36:53.111
There's one more in the Trad AI space, one more news event I want to talk about.

402
00:36:53.131 --> 00:36:57.414
This is actually from Anthropic's own privacy.cloud.com website.

403
00:36:57.794 --> 00:37:00.275
This was updated, I think, on September 8th.

404
00:37:00.355 --> 00:37:02.857
Anthropic released a new data retention policy.

405
00:37:03.277 --> 00:37:05.899
And so this is a quote from the page that I have pulled up.

406
00:37:06.259 --> 00:37:13.804
To ensure we are responsibly deploying covered models, we are requiring limited data retention and review as part of our safety work.

407
00:37:14.324 --> 00:37:23.248
Prompts submitted to and outputs generated by covered models are retained for 30 days to support our safety work on every platform where these models are offered.

408
00:37:23.588 --> 00:37:33.612
So if you are on a consumer plan from Anthropic, that's like Claude, Free, Pro, or Max, you don't have to worry because your data was already being collected and retained permanently by Anthropic.

409
00:37:33.652 --> 00:37:42.616
But according to Anthropic, this is for enterprises and organizations that have already set up workspaces with Anthropic with a zero data retention policy.

410
00:37:43.076 --> 00:37:50.300
So if you have that zero data retention agreement with Anthropic, it is now actually a 30-day data retention agreement.

411
00:37:50.740 --> 00:37:55.463
And Anthropic is citing safety concerns as the motivation behind this.

412
00:37:55.823 --> 00:38:04.588
Now, to put a negative spin on this, when I see Anthropic citing safety concerns, I kind of equate it to like the government citing like,

413
00:38:05.788 --> 00:38:14.432
state security, or this is because of terrorism, or you know, we got to protect the children, or something, something, something, but it's really something that the government wants.

414
00:38:14.812 --> 00:38:19.614
And I see that same sort of structure where, like, Anthropa gets to retain your data for 30 days because of safety concerns.

415
00:38:20.094 --> 00:38:24.216
But then if it gets to retain your data, it gets the value of that data.

416
00:38:24.236 --> 00:38:30.478
That's my, like, insipid way of looking at this, but I'm sure it's probably somewhat accurate, too.

417
00:38:30.798 --> 00:38:33.659
Yeah, unfortunately, I don't think that's an overly cynical view.

418
00:38:35.180 --> 00:38:37.262
I think it's relatively accurate.

419
00:38:37.302 --> 00:38:44.606
And yeah, you always have to wonder, you know, certainly with Anthropic and OpenAI of just like, what do they mean when they say these words safety?

420
00:38:45.007 --> 00:38:48.869
You know, it's everything they talk about lately, safety, alignment.

421
00:38:49.310 --> 00:38:50.210
What does it actually mean?

422
00:38:50.290 --> 00:38:51.231
Safety for who?

423
00:38:51.511 --> 00:38:52.692
Alignment for who?

424
00:38:54.393 --> 00:38:55.614
what does that mean?

425
00:38:56.054 --> 00:38:59.475
And so it's this very vague term that they use to justify something like this.

426
00:38:59.515 --> 00:39:06.857
Like we can't even, you know, they can't even ascribe to their own ZDR policies anymore because now they have to do 30 days of retention for safety.

427
00:39:07.257 --> 00:39:11.698
And, you know, and there may very well be like a few legitimate reasons why they have to do that, but they then,

428
00:39:12.498 --> 00:39:17.360
turn it into this broader catch-all that seemingly catches everybody in it and all of their data.

429
00:39:17.420 --> 00:39:20.421
And it just so happens that they also need that data to train on.

430
00:39:20.841 --> 00:39:28.404
So, you know, it's hard to not be slightly cynical in terms of their positioning and what they're actually doing and what they actually mean.

431
00:39:28.444 --> 00:39:34.546
And how concerned are they actually about the safety components of these things, especially when they are already concerned?

432
00:39:34.726 --> 00:39:43.669
building so many guardrails into their models to begin with about what they will actually respond to, how they will respond to things, even on things that are often innocuous.

433
00:39:43.729 --> 00:39:58.734
Like again, this remains one of the leading kind of funnels for Venice is people using OpenAI, using Anthropic and getting a refusal over something that they really don't think they should be refused on tends to lead them to search for alternatives and things like Venice.

434
00:39:59.494 --> 00:40:03.855
Um, so for us, yeah, it's hard to not be a little skeptical what they mean.

435
00:40:03.955 --> 00:40:08.217
Ultimately, they're their own private company, they're going to do what they what they want and what they feel they need to do.

436
00:40:08.737 --> 00:40:19.400
And I think that they are, I think that they just don't want to give up that that training data feels like more of the root of this than actually, you know, operating as safely as possible, in my opinion.

437
00:40:20.860 --> 00:40:25.263
So John, I think this is our fourth call together, the VVV call we do every single month.

438
00:40:25.303 --> 00:40:38.390
I think this month is definitely the hottest in terms of the rest of the AI world's conversations around privacy and concerns around user alignment, user alignment with the AI platform.

439
00:40:38.590 --> 00:40:43.053
Has that turned into any sort of change in signups or conversations with enterprises?

440
00:40:43.373 --> 00:40:46.275
How is that manifesting on the Venice side of the fence?

441
00:40:47.152 --> 00:40:54.341
Yeah, I would say it seemingly just keeps adding to the growth that we are seeing, which has just really not slowed down this year.

442
00:40:54.381 --> 00:41:01.249
I think even internally, we keep waiting to see if we're going to hit some sort of growth plateau or, you know, things are going to go sideways for a little bit.

443
00:41:01.870 --> 00:41:03.832
And it just really hasn't happened, to be honest.

444
00:41:03.932 --> 00:41:04.272
And I think

445
00:41:05.235 --> 00:41:10.696
I think part of that is these conversations and people waking up to the importance of privacy and AI.

446
00:41:10.737 --> 00:41:15.478
And it's still, you know, we live in a bubble, especially those of us who live on Twitter, etc.

447
00:41:15.818 --> 00:41:23.920
So we're seeing, you know, these loud conversations that are probably, you know, only representative of a very small percentage of the population even being aware of them or involved in them.

448
00:41:24.260 --> 00:41:25.481
But they're at least becoming louder.

449
00:41:25.721 --> 00:41:27.321
And I think that does wake up more people.

450
00:41:27.401 --> 00:41:30.482
And then you just, you get another story like this every week or two.

451
00:41:31.242 --> 00:41:32.442
And people realize that.

452
00:41:32.462 --> 00:41:41.285
So for us, we just kind of are seeing continually steady, you know, adjustments to the growth rate that seemingly keeps going up into the right.

453
00:41:41.305 --> 00:41:42.846
And so we're not complaining about that.

454
00:41:42.886 --> 00:41:47.127
But also AI itself just continues to grow and just become so much more capable.

455
00:41:47.187 --> 00:41:49.508
And some of the things you can do with these models.

456
00:41:51.068 --> 00:41:54.570
compared to what you could do even three months ago is just so wildly different.

457
00:41:54.590 --> 00:42:05.395
You know, some of my own tests that I run on some of these internal models, it's just wild how much better they are getting at things and the types of things they could do that you couldn't even imagine them doing.

458
00:42:05.895 --> 00:42:06.655
three or four months ago.

459
00:42:06.675 --> 00:42:11.017
And then a lot of those things tend to drive a lot more revenue and token usage itself.

460
00:42:11.238 --> 00:42:15.299
So it's like, the better they get, the better they also are at burning tokens.

461
00:42:16.460 --> 00:42:18.581
So you see just a lot more of that.

462
00:42:18.761 --> 00:42:22.563
It goes back to the chart we showed earlier, like all these things coalesce and then you just have

463
00:42:23.434 --> 00:42:24.215
growth of these things.

464
00:42:24.255 --> 00:42:34.325
And I just, it's, again, it's hard for me to imagine that train slowing down anytime soon because these things are just becoming so much more capable and so many more people are becoming aware of that capability.

465
00:42:34.705 --> 00:42:46.817
But at some point we're going to hit like an inflection point where like a large, a real large, you know, percentage of at least the American population, but really the world's population becomes more understanding of what these things could do.

466
00:42:47.377 --> 00:43:09.496
because you know we all feel like we understand it but we live in this bubble where you know we're using things like harnesses you know like the cloud codes and codexes and open clause and hermes agents and all these things that really you know are grokbots are now muse and like all these things that make these things far more accessible and useful but it's such a small percentage of people that are actually actually using any of those tools so like

467
00:43:10.236 --> 00:43:18.603
My best example is always that there's like, there's such a wide swath of the population that don't realize AI is anything more than souped up Google.

468
00:43:18.863 --> 00:43:20.445
Like that's how they use AI.

469
00:43:20.485 --> 00:43:22.506
Mostly they ask it questions like they would ask Google.

470
00:43:22.906 --> 00:43:24.748
They use ChatGPT's free product.

471
00:43:25.268 --> 00:43:26.209
They get an answer back.

472
00:43:26.229 --> 00:43:27.370
They're like, wow, this is really cool.

473
00:43:27.430 --> 00:43:28.551
I can have a conversation with it.

474
00:43:28.571 --> 00:43:30.893
But it's really just question, answer, question, answer.

475
00:43:32.065 --> 00:43:40.109
But the gap that's been forming is those who use AI in that way, which is basically really where AI was, you know, nearly 18 months ago.

476
00:43:40.729 --> 00:43:50.674
And then the people who, especially in December of last year, January of this year, realized that there was this leap where AI stopped answering questions and started doing things.

477
00:43:51.415 --> 00:43:54.536
And I just think so few people are actually turned on to that yet.

478
00:43:54.596 --> 00:43:57.958
It's still it's growing dramatically, but it's still such a small slice.

479
00:43:59.669 --> 00:44:07.776
I want to have a discussion with you actually about the way that the utility of all these models, where it gets expressed in like the tech stack.

480
00:44:08.396 --> 00:44:17.463
I think people will like the most normal thing for the unexposed to AI person, they'll go to OpenAI or Anthropic, just like one of the two AI apps.

481
00:44:17.544 --> 00:44:18.985
It's likely their first stop.

482
00:44:19.505 --> 00:44:26.814
And if they want to learn how useful an LLM is, how useful an AI Asian is, that's where they would go and they would tinker around there.

483
00:44:27.254 --> 00:44:32.621
But Venice has this product called Minds that allows for LLMs to get combined together.

484
00:44:33.061 --> 00:44:37.787
And that kind of opens up the conversation of like, is the most useful place for AI?

485
00:44:38.968 --> 00:44:42.249
at the layer of an AI agent with an AI lab?

486
00:44:42.349 --> 00:44:47.611
Or is there a layer, I kind of think Venice is an abstraction layer around above all the models.

487
00:44:48.051 --> 00:44:56.213
Is that, you guys are placing a bet on that's actually where a lot of the value of AI gets expressed because that's where agents get to talk to each other.

488
00:44:56.253 --> 00:44:59.234
That's where LLMs get to combine in weird ways.

489
00:44:59.574 --> 00:45:06.016
Do you think that's a more creative layer than just seemingly something on top of the AI labs that's super powerful?

490
00:45:06.036 --> 00:45:06.276
Yeah.

491
00:45:06.942 --> 00:45:10.645
Yeah, I think there's kind of a couple schools of thought on this question.

492
00:45:11.546 --> 00:45:20.814
So there's kind of the view of the way the Frontier Labs have viewed it, and they kind of have to view things this way because of the way their companies and their products work.

493
00:45:21.616 --> 00:45:39.660
Their general view is that these kind of specialized, you know, agents or outputs or even things that use multiple models, that they think that where everything's going is it's just like one big, fat, super smart model and that that's where everything will be.

494
00:45:40.140 --> 00:45:46.961
But that is not our view and it's not really the view that I think long term is going to be correct.

495
00:45:47.821 --> 00:45:57.043
So something like Mines, which is a product that we have in beta right now, we're going to be pushing that a lot more as we get some additional features shipped here over the coming weeks.

496
00:45:57.583 --> 00:46:14.427
But the idea of Mines is that a creator can assemble a model or various models together in a way with an agent that produces a far better output for a specific task than the kind of quote unquote naked model would by itself.

497
00:46:16.227 --> 00:46:32.341
I personally think that's more where things are going and that being able to direct models in the way, like effectively orchestrate or like give the agent instructions about how to orchestrate in specific ways for specific tasks will generally get the average kind of call it normie user

498
00:46:33.486 --> 00:46:35.508
a much better output for the thing that they're looking for.

499
00:46:35.548 --> 00:46:42.414
And that's kind of what the bet of something like mine is, is that there are really smart people out there who understand these models really well.

500
00:46:42.755 --> 00:46:46.538
And then there's a lot of people who just want the output and they have no idea how to get it.

501
00:46:46.578 --> 00:46:47.779
They don't know how to prompt well.

502
00:46:48.219 --> 00:46:50.421
They don't realize how much their prompt affects things.

503
00:46:51.322 --> 00:46:54.265
The models in general are getting better, even with very bad prompts.

504
00:46:54.805 --> 00:46:55.486
But the...

505
00:46:56.437 --> 00:47:09.832
the you can there can be a wide gap still in turn and I think this will persist of the outputs that are possible for those who know how to assemble things in a specific way and those who just are asking a naked model and really don't know what they're asking for.

506
00:47:10.312 --> 00:47:17.159
So the idea of minds is to make that far easier let creators kind of ship these AI experiences these types of AI products.

507
00:47:17.760 --> 00:47:21.621
that are assembled in specific ways to help you get a specific output.

508
00:47:22.341 --> 00:47:26.842
And I do think in many cases, that's going to be better for the average user.

509
00:47:27.302 --> 00:47:30.503
You know, they don't, and part of it is it demonstrates what these things can do.

510
00:47:30.523 --> 00:47:36.064
A lot of people, they just, they get to that chat GPT screen or whatever, and they don't, they just don't even know what to do.

511
00:47:36.584 --> 00:47:38.265
And that's why they end up using it like Google.

512
00:47:38.305 --> 00:47:40.445
They don't know what to do other than ask it a question.

513
00:47:40.845 --> 00:47:45.046
They haven't, they just aren't in that world where these models can do things.

514
00:47:45.486 --> 00:47:47.387
So something like Mines will more demonstrate

515
00:47:48.408 --> 00:47:51.370
hey, you want something that does X.

516
00:47:51.410 --> 00:48:02.077
You know, you want trading, a trading, an agent that's going to help you with your trading, or you want an agent that's going to help you with your accounting or spreadsheets or, you know, a PDF creator or whatever it is.

517
00:48:02.557 --> 00:48:12.184
And someone has assembled it in such a way to make that thing super, you know, dead easy for you where all you have to do is put the input in and get exactly the output you want.

518
00:48:12.464 --> 00:48:14.706
That's kind of the whole thrust and idea of minds.

519
00:48:14.986 --> 00:48:16.127
And I do think that that is,

520
00:48:16.887 --> 00:48:27.089
there is something very powerful there that is a very different place than the labs will go themselves because all of their resources are dedicated into just making their one model so smart.

521
00:48:27.109 --> 00:48:42.472
And I think we saw an example of this even in the industry over the last few weeks, because we had this whole thing with like this new class of models come out from Jev, a former co-creator of ChatGPT, who effectively created rather than a LLM text-based model, effectively a

522
00:48:43.112 --> 00:48:48.996
classifier or decision-based model, which is something that had existed before, but not exactly in this way.

523
00:48:49.476 --> 00:49:05.045
And now people are combining these things with the LLMs and realizing that these constituent parts, like an LLM, plus using JEV in the right places can be a lot smarter, a lot more efficient, a lot cheaper than just using the LLM on a bunch of things where it might get confused non-deterministically.

524
00:49:05.545 --> 00:49:11.989
So I do think you're going to see more and more of this type of thing, where the real power comes from

525
00:49:12.649 --> 00:49:18.713
combining various types of models and kind of system instructions in various ways to get a specific output.

526
00:49:20.334 --> 00:49:22.355
The word biodiversity comes to mind.

527
00:49:22.375 --> 00:49:27.078
I think when you started talking, you said there's two schools of thought.

528
00:49:27.459 --> 00:49:37.125
Maybe the way to define the schools of thought that I got from you speaking just now is you have the frontier AI labs, which they're trying to produce the one model to rule them all.

529
00:49:37.245 --> 00:49:39.186
They're trying to produce God, basically.

530
00:49:39.206 --> 00:49:40.247
Yeah, they want digital God.

531
00:49:40.407 --> 00:49:40.527
Yeah.

532
00:49:40.907 --> 00:49:43.148
They want they're trying to make digital God and it's just one model.

533
00:49:43.588 --> 00:49:47.210
And then there's the open router Venice side of things.

534
00:49:47.270 --> 00:49:49.411
Like I remember Alex Atala from open router.

535
00:49:49.952 --> 00:49:51.092
He was pretty early to this.

536
00:49:51.112 --> 00:49:51.913
He was like, no, no, no.

537
00:49:51.933 --> 00:49:53.633
There's actually going to be thousands of models.

538
00:49:53.653 --> 00:49:55.414
There's going to be more models than we can count.

539
00:49:55.794 --> 00:50:02.678
And what you're talking about with minds is like the combatorial value of unique specialized models.

540
00:50:03.318 --> 00:50:08.101
when you sum all the parts together, it's going to be greater than the whole, and that's going to be greater than God.

541
00:50:08.422 --> 00:50:10.663
And these are the two races going on.

542
00:50:10.923 --> 00:50:11.764
Both can be true.

543
00:50:11.944 --> 00:50:13.645
Like, they're not mutually exclusive.

544
00:50:13.945 --> 00:50:16.507
But, like, maybe one is more valuable than the other.

545
00:50:16.807 --> 00:50:23.092
And I think what you're saying is, like, with unique, specialized models, the combination is the valuable thing.

546
00:50:23.412 --> 00:50:27.275
And that's what Venice is trying to plant a flag and make a bet on with Minds.

547
00:50:28.135 --> 00:50:29.956
Yeah, I think that's a really good way to put it.

548
00:50:30.076 --> 00:50:38.720
It's the combination and the power of those who understand these and their ability to help deliver that output to someone who doesn't have any understanding of these things.

549
00:50:38.740 --> 00:50:41.722
You know, most people are not going to be good at prompting.

550
00:50:41.802 --> 00:50:54.288
Most people are not going to, you know, this has been a longstanding problem even at Venice and how we're evolving the product is like, you know, it's great that you can come to Venice and choose from all these models, but the average user gets in and they don't know the difference between

551
00:50:54.808 --> 00:50:57.130
you know, these hundreds of different models and what to choose.

552
00:50:57.170 --> 00:51:03.214
So we have to be smart about how we recommend them the right model, how we get them the right actual functionality they want.

553
00:51:03.494 --> 00:51:05.675
The average user doesn't care what model they're using.

554
00:51:05.935 --> 00:51:10.899
What they care about is, are they getting the intelligence that they want and are they getting the outputs that they want?

555
00:51:11.379 --> 00:51:20.265
So the idea is like, yes, the digital God thesis that a lot of these labs are going for, I think there's obviously something important and powerful about that.

556
00:51:20.745 --> 00:51:23.167
But, you know, for specific use cases,

557
00:51:24.027 --> 00:51:27.530
These combinations could be a lot better for a lot of specific use cases.

558
00:51:27.870 --> 00:51:29.392
They're not going to be generalized better.

559
00:51:29.512 --> 00:51:35.177
They're not going to be better at everything, but they can be better in very specific areas where the user really cares about.

560
00:51:35.617 --> 00:51:37.058
You know, the user really cares about

561
00:51:38.203 --> 00:51:40.644
do I get the output I want for the thing I'm looking for?

562
00:51:41.185 --> 00:51:44.286
Not is the model capable of doing everything under the sun?

563
00:51:44.507 --> 00:51:49.589
That's, you know, most users are going to use only such a small fraction of what a model is capable of anyway.

564
00:51:50.030 --> 00:51:55.152
What they really care about is can I get the thing I want and can I get it relatively fast and easy?

565
00:51:56.373 --> 00:51:59.755
That's where I think something like mine could be very powerful.

566
00:51:59.935 --> 00:52:01.376
That's what we are exploring.

567
00:52:02.323 --> 00:52:11.072
Is there any other company out there who kind of has a similar vision of combining models together to create combinatorial outputs?

568
00:52:11.152 --> 00:52:13.935
Or is Venice really the pioneer on this frontier?

569
00:52:15.020 --> 00:52:18.201
I definitely think we are in the phase of pioneering on this.

570
00:52:18.241 --> 00:52:19.741
I wouldn't say that there's nobody else.

571
00:52:19.821 --> 00:52:34.904
I mean, there are other companies that are doing things like, you know, that are effectively councils of models or ways to get models to route to each other, you know, in specific ways so that you use this model on this specific task and a smarter model on this specific task.

572
00:52:34.924 --> 00:52:36.924
So all the model routing is a combination of this.

573
00:52:37.304 --> 00:52:41.925
But I don't think anyone's combining these things in the way that we are in like a consumer-focused way.

574
00:52:42.706 --> 00:52:47.331
package even though there are an intentional direction towards the combination idea

575
00:52:48.018 --> 00:52:54.921
Yeah, I think that's definitely something hopefully we are leading on and that we are very excited about and that we think is really interesting.

576
00:52:54.941 --> 00:52:58.602
I would say the harnesses kind of steps into this world.

577
00:52:58.622 --> 00:53:00.803
You know, the various harnesses do this a little bit.

578
00:53:01.143 --> 00:53:06.366
But even those are more generally about, you know, very broad use cases, very broad generalized usage.

579
00:53:06.966 --> 00:53:14.609
We're interested in these very specific use cases that you otherwise cannot necessarily get as good of an output for unless you really know what you're doing.

580
00:53:15.822 --> 00:53:16.542
That's fascinating.

581
00:53:16.642 --> 00:53:21.903
I think that's for any VVB holder, I think that's something to absolutely keep an eye on to see the progress of this.

582
00:53:21.963 --> 00:53:41.367
If you accept the spectrum between the God models from the AI labs or the biodiversity realm of the spectrum, I think that is something that a VVB holder would be very interested in seeing is the biodiverse models coming up and coming together and making some sort of weird outputs, kind of like life.

583
00:53:41.727 --> 00:53:44.108
I always like the biodiversity metaphors.

584
00:53:44.748 --> 00:53:47.009
We need to get into the investor Q&A.

585
00:53:47.069 --> 00:53:48.489
John, we got a few questions from you.

586
00:53:48.549 --> 00:53:51.229
The first two from GW Jack in the VVV Telegram.

587
00:53:51.710 --> 00:53:55.670
How can Venice create additional demand to stake VVV?

588
00:53:55.750 --> 00:54:06.393
So as Venice improves the way, the way as holders can use staked VVV, how does the team think about designing these features so that they can create incremental demand to own and stake VVV?

589
00:54:08.120 --> 00:54:15.484
Yeah, thinking about this in a number of different ways, obviously over time, VVV staked has kind of been up and to the right.

590
00:54:15.504 --> 00:54:26.030
There's a, you know, sometimes with various price action, this will go up and down, but over any long time for anybody to zoom out enough, this has kind of been one directional since the genesis of Venice.

591
00:54:26.190 --> 00:54:28.251
It's definitely something I spend a lot of time thinking about

592
00:54:29.130 --> 00:54:33.031
What is additional utility that might make sense for the staking sphere?

593
00:54:33.451 --> 00:54:39.854
You know, obviously one of the big things even right now is just the fact that you can get effectively lifetime pro access if you have 100 staked EVV.

594
00:54:40.254 --> 00:54:42.795
That's something a lot of people like, that functionality.

595
00:54:42.815 --> 00:54:49.177
And then, of course, the primary ability to mint diem and get that functionality out is something that we expect to drive over time.

596
00:54:49.617 --> 00:54:51.858
I think that in some ways this will actually be...

597
00:54:53.680 --> 00:54:59.584
This question of like what drives VVV stake will be kind of downstream of DM utility.

598
00:54:59.704 --> 00:55:08.769
Like the more useful DM is, the more useful things you can do with it, the more places you can effectively resell that when you're not using it or get yield on it in various ways.

599
00:55:08.869 --> 00:55:18.675
The more demand there is to mint Venice or mint DM rather from VVV, the more likely you're going to see staking continue to rise and people locking up.

600
00:55:19.175 --> 00:55:19.856
to do those things.

601
00:55:20.397 --> 00:55:28.569
There's certainly a number of things, nothing I'm necessarily ready to talk about today, but a number of things that we're considering on that front that, you know, could get added over time.

602
00:55:28.630 --> 00:55:30.653
And it's something we pay a lot of attention to.

603
00:55:32.037 --> 00:55:35.518
This next question comes from Malcolm from the VVV Telegram.

604
00:55:36.238 --> 00:55:37.538
Excuse me, no, sorry.

605
00:55:37.558 --> 00:55:42.679
This one comes from Iguanarchist and also Gecko, gecko.eth, Gecko, the creator of Venetastats.

606
00:55:42.739 --> 00:55:47.140
Could reoccurring subscriptions become the next expansion of VVV Burn?

607
00:55:47.180 --> 00:55:54.642
So I think people are asking about if reoccurring subscriptions is where that is on the docket in the conversation of VVV Burn.

608
00:55:55.347 --> 00:56:00.229
Yeah, it's definitely something that, you know, obviously we have talked about, including even on these calls before.

609
00:56:00.249 --> 00:56:01.849
It's definitely something we're looking at.

610
00:56:02.509 --> 00:56:05.410
I don't know if it will be the next one necessarily.

611
00:56:05.911 --> 00:56:07.431
There's actually something interesting going on.

612
00:56:07.471 --> 00:56:10.472
One of our focuses that I think is going to, I think people are going to love.

613
00:56:10.752 --> 00:56:14.554
We're not necessarily ready to talk too much about these details yet.

614
00:56:14.974 --> 00:56:18.515
But Eric has hinted at this kind of incentive and reward program.

615
00:56:19.435 --> 00:56:20.256
that we've talked about.

616
00:56:20.316 --> 00:56:28.262
And I think that that will likely be one of the next big drivers of effectively additional VVV buys for that reason.

617
00:56:29.443 --> 00:56:32.365
So something to stay tuned for future calls.

618
00:56:32.505 --> 00:56:35.267
But yeah, some really cool things coming on that front.

619
00:56:35.307 --> 00:56:41.952
And then, you know, certainly things like renewals are something we're considering and could be on the docket sooner than later, but it may not be the next one.

620
00:56:43.627 --> 00:56:52.491
Another one from GW Jack, who wants to know about any timeline updates on the data center build out and any details that you can give about the Venice data center.

621
00:56:53.751 --> 00:56:55.752
Yeah, so I mean, this has been ongoing.

622
00:56:56.813 --> 00:57:08.518
You know, we've kind of, since we announced the raise back in July, this has been one of our, you know, big capital expenditures is building out this initial data center and then looking even at multiple data centers.

623
00:57:08.847 --> 00:57:11.249
potential for additional data centers after that.

624
00:57:12.750 --> 00:57:20.317
I'd have to follow up with my team on the exact timeline, but I believe that the first data center is supposed to come online, hopefully before the end of October.

625
00:57:20.797 --> 00:57:26.402
So maybe like late October, maybe if there's delays as there can be with a lot of these things, maybe that leaks into November.

626
00:57:26.962 --> 00:57:27.762
But it's pretty soon.

627
00:57:27.962 --> 00:57:30.504
It's coming up that we're going to be able to turn one of these online.

628
00:57:30.924 --> 00:57:36.886
We won't necessarily know how that affects things like margins and then how that affects things like VVV burns downstream right away.

629
00:57:36.927 --> 00:57:45.851
It'll take some time of actually getting these things online, optimizing them, seeing which models we can run most efficiently and experimentation on that front.

630
00:57:45.891 --> 00:57:48.912
But you'll definitely be hearing more from us as that gets built out.

631
00:57:48.932 --> 00:57:51.113
And I think, you know, over the next couple of months.

632
00:57:52.466 --> 00:58:01.571
For somebody who wants to pay attention to that topic of conversation with Venice, the increased margins from the data center, because you guys own your own data center, you're vertically integrating.

633
00:58:01.671 --> 00:58:02.371
Why are you doing that?

634
00:58:02.391 --> 00:58:03.672
Because you get better margins that way.

635
00:58:03.692 --> 00:58:06.273
You can absorb some of your own costs.

636
00:58:06.433 --> 00:58:12.196
Is that the part of the VVV burn where that would be expressed is the discretionary VVV burns?

637
00:58:12.417 --> 00:58:12.817
Is that right?

638
00:58:13.593 --> 00:58:25.457
It might be in the discretionary or it might, there might be a world where we decide that, you know, our margins are good enough that we can increase the amount for something like the credit purchase or the subs like we talked earlier.

639
00:58:25.537 --> 00:58:28.018
So it could, it could come in in that way.

640
00:58:28.278 --> 00:58:36.521
And then, yeah, probably initially it might come via discretionary if we feel it's creating enough additional free cash flow that we can route that back into BDD.

641
00:58:38.073 --> 00:58:42.255
Last question from Iguanarchist and also Cameron in the VVV chat.

642
00:58:42.355 --> 00:58:45.977
What role does Nier play in Venice's infrastructure?

643
00:58:46.858 --> 00:58:58.404
And basically, I think the last question I would love to know too is like, I know Venice uses Nier for accessing some very much of the top tier privacy AI products.

644
00:58:58.704 --> 00:58:59.885
How do you guys use Nier?

645
00:58:59.925 --> 00:59:01.206
How does it work into the Venice stack?

646
00:59:01.982 --> 00:59:11.709
Yeah, so NIR is used right now as kind of at our top privacy kind of levels of TEE and end-to-end encrypted inference.

647
00:59:11.769 --> 00:59:15.252
NIR is one of our providers for those types of models.

648
00:59:15.892 --> 00:59:19.055
So if you go into the Venice app and you try to look for things like

649
00:59:19.675 --> 00:59:26.557
TEE and end-to-end encrypted models, you'll see some of these models and some of those are offered by Nier.

650
00:59:26.977 --> 00:59:28.697
So that's kind of built into our stack.

651
00:59:28.757 --> 00:59:34.238
This is definitely a, what I would say is a smaller part of Venice's overall inference.

652
00:59:34.278 --> 00:59:46.621
These are not necessarily things that everybody cares about or wants, but it's really important and critical to our offering that it is available and that those who really care about kind of that cryptographic proof of

653
00:59:47.482 --> 00:59:49.664
the privacy that they're looking for, that they have those options.

654
00:59:49.904 --> 00:59:53.488
So Nier has been a great partner in allowing us to offer such a thing.

655
00:59:54.669 --> 00:59:55.029
That's true.

656
00:59:55.049 --> 01:00:00.094
Like for the parts of Venice that touches the Nier stack, that's for like the nuclear levels of privacy, right?

657
01:00:00.115 --> 01:00:06.001
Like the most hardcore, probably also the most expensive because now we have to start to talk to distributed systems.

658
01:00:06.802 --> 01:00:08.663
Yeah, it's the highest level of privacy.

659
01:00:09.543 --> 01:00:17.005
Some ways the functionality is a little more limited because of that, because there's certain things you can't get out of the context or that has to all be delivered at once.

660
01:00:17.205 --> 01:00:21.786
And yes, sometimes that means those are a little more expensive as a result, but it depends on the model.

661
01:00:22.246 --> 01:00:26.527
I think Nier's generally done a good job of being able to deliver those at relatively competitive pricing.

662
01:00:27.087 --> 01:00:32.509
So sometimes they're just good models to use if you care about privacy and you want that cryptographic proof.

663
01:00:33.897 --> 01:00:36.999
Well, John, this has been the September VVV call.

664
01:00:37.119 --> 01:00:38.460
I have learned a lot.

665
01:00:38.520 --> 01:00:42.602
There's just so much data to talk about with Venice and also just the wider world of AI.

666
01:00:42.622 --> 01:00:44.603
There's one last topic to talk about before we go.

667
01:00:44.984 --> 01:00:51.147
I think you and I are going to hang out here in about a month in New York because of the Lumara Film Festival.

668
01:00:52.148 --> 01:00:54.029
Talk to me about the Lumara Film Festival.

669
01:00:54.209 --> 01:00:54.689
What is it?

670
01:00:54.809 --> 01:00:55.850
Why are you excited about it?

671
01:00:55.890 --> 01:00:56.851
And what are you hoping to see there?

672
01:00:57.571 --> 01:01:02.918
Yeah, so this is a film festival that Venice has collaborated with Moonpay to put on.

673
01:01:03.198 --> 01:01:14.593
It was something internally that our team really cared about doing something like this because we talk to a lot of creators every day, creators especially since the advent of the better video models.

674
01:01:15.153 --> 01:01:22.162
They've really loved creating and creating via Venice because we don't block things that they get blocked on a lot of other platforms.

675
01:01:22.222 --> 01:01:27.688
And so their ability to effectively express themselves artistically is a lot less limited through Venice.

676
01:01:28.048 --> 01:01:30.612
And we wanted to give, you know, a demonstration of how

677
01:01:31.232 --> 01:01:52.156
you know, AI video has come along and how, how artistic some of this, some of these things are actually getting that this is real, you know, we're getting to the point where we can get real short films and soon enough, like real feature length films produced with AI that people are going to be really impressed of, obviously a far fraction of the cost of traditional film development and film production.

678
01:01:52.676 --> 01:01:58.517
So yeah, I think we've gotten something crazy, like over 700 submissions to this thing last I heard.

679
01:01:58.577 --> 01:01:58.637
And

680
01:01:59.237 --> 01:02:07.700
There's basically a process going down now of the judges of whittling those down and figuring out what's actually going to be aired at the film festival.

681
01:02:08.620 --> 01:02:10.300
But it's really cool to see this happening.

682
01:02:10.340 --> 01:02:12.721
It's really cool to see how excited people are about it.

683
01:02:12.781 --> 01:02:14.262
And it's going to be really fun to see that.

684
01:02:14.282 --> 01:02:25.265
And, you know, we've come a long way from, you know, Will Smith eating spaghetti in various terrible ways to things that actually look like they could have been produced by like, you know, a high tier film production.

685
01:02:26.720 --> 01:02:34.784
Would you say the creators creating on Venice are at the absolute frontier of what it means to be film creators with AI?

686
01:02:35.785 --> 01:02:36.726
I think, I think, yes.

687
01:02:36.826 --> 01:02:41.768
I mean, obviously there's varying levels of skill and professionality depending on who you're talking to.

688
01:02:41.808 --> 01:02:43.850
But some of these high tier creators, absolutely.

689
01:02:43.950 --> 01:02:49.193
You know, some of the initial previews of things I've seen, and I've only seen a small fraction of what's actually been submitted.

690
01:02:49.593 --> 01:02:53.415
But some of them are like incredibly impressive and hit, you know, aesthetics and design.

691
01:02:53.875 --> 01:03:01.538
you know, continuity in a way that you just would not necessarily expect or not necessarily call out that it was AI unless you already knew it was.

692
01:03:01.958 --> 01:03:16.883
And I just think it's going to kind of democratize the ability for artists, you know, singular artists to create full films that they're really proud of and that really speak to what they're trying to express where, you know, basically the production costs go from,

693
01:03:17.483 --> 01:03:24.388
you know, millions or tens or hundreds of millions of dollars to, you know, a few thousand dollars to produce something of similar quality.

694
01:03:24.408 --> 01:03:28.852
And I think, you know, like we always say in AI, this is the worst it will ever be.

695
01:03:28.912 --> 01:03:30.253
It's only going to get better from here.

696
01:03:31.367 --> 01:03:36.208
Well, if you are in New York in 28 days, it's the third week of October.

697
01:03:36.248 --> 01:03:38.889
This is when the Lamara Film Festival is, so come check it out.

698
01:03:39.249 --> 01:03:42.610
Get to see what the frontier of film development with AI is.

699
01:03:42.630 --> 01:03:46.051
And then one last CTA for the audience.

700
01:03:46.071 --> 01:03:48.172
You've heard me talk about the VVV Telegram.

701
01:03:48.212 --> 01:03:54.213
That is a telegram that we host to just talk about VVV, AI, inference, DM, tokenized economics.

702
01:03:54.233 --> 01:03:59.795
If you want an invite into that telegram, shoot me a DM on Twitter, and I'll throw you in there.

703
01:04:00.135 --> 01:04:02.576
John, it's been another month, another exciting month with Venice.

704
01:04:03.977 --> 01:04:05.258
Thank you for doing what you do.

705
01:04:05.298 --> 01:04:07.359
It's very valuable to have Venice here on the earth.

706
01:04:07.399 --> 01:04:11.461
And I can't wait to see everything coming down the pike with mines and the data center and everything else.

707
01:04:11.501 --> 01:04:13.442
And we will talk about all that stuff in a month.

708
01:04:14.282 --> 01:04:14.442
Yep.

709
01:04:14.742 --> 01:04:15.363
Thank you, David.

710
01:04:15.703 --> 01:04:16.283
Talk to you more soon.

711
01:04:16.303 --> 01:04:17.864
Bye, everyone.