TRANSCRIPT 🔗
1
00:00:00,000 --> 00:00:04,799
The irony is I think for better or for worse,
when you're in a capital dense geographic
2
00:00:04,799 --> 00:00:09,039
location like San Francisco or New York, you
get signals about what's happening in in the
3
00:00:09,039 --> 00:00:10,880
fundraising markets whether you want them or
not.
4
00:00:10,880 --> 00:00:14,884
And if you're not there, you're reading about
these rounds six to nine months later, and
5
00:00:14,884 --> 00:00:18,164
you're probably scratching your head saying,
how did that company get all of that money?
6
00:00:18,164 --> 00:00:21,364
I feel bad because a lot of times those people
show up and they're like, here's my report
7
00:00:21,364 --> 00:00:21,765
card.
8
00:00:21,765 --> 00:00:22,884
I finished all the problems.
9
00:00:22,884 --> 00:00:24,724
I was like, but you were in the wrong chapter.
10
00:00:24,884 --> 00:00:27,125
It's the week of October 26.
11
00:00:27,125 --> 00:00:30,219
Welcome to the learning corner a Precursor
Ventures podcast.
12
00:00:32,619 --> 00:00:34,379
Welcome to the Learning Corner by Precursor.
13
00:00:34,379 --> 00:00:38,939
I'm Mia Foranum, and I'm joined by our GP and
managing partner, Charles Hudson, where we will
14
00:00:38,939 --> 00:00:42,859
be talking through the articles and Goodreads
that caught our team's attention over the past
15
00:00:42,859 --> 00:00:43,179
week.
16
00:00:43,575 --> 00:00:48,854
Our first piece this week is, of course, the
news that Anthropoc has, in fact, filed its s
17
00:00:48,854 --> 00:00:52,614
one with SEC, setting up an Nasdaq listing
under the ticker ANT.
18
00:00:52,614 --> 00:00:55,574
The numbers in the prospectus are staggering.
19
00:00:55,574 --> 00:01:00,320
They've got 16,300,000,000 in revenue in the
first half of twenty twenty six, up from about
20
00:01:00,320 --> 00:01:06,560
1,160,000,000 in the same period a year earlier
and well past the 4,600,000,000 it did in all
21
00:01:06,560 --> 00:01:07,680
of 2025.
22
00:01:07,680 --> 00:01:14,025
The headline net loss of $80,700,000,000 looks
alarming, but everyone knows at this point most
23
00:01:14,025 --> 00:01:20,504
of that is a $72,000,000,000 noncash charge
tied to convertible notes, derivative, safes,
24
00:01:20,504 --> 00:01:26,530
and the company actually posted 974,000,000 in
operating income in the second quarter.
25
00:01:26,609 --> 00:01:32,769
Also discloses that more than $430,000,000,000
are locked in cloud compute commitments, and
26
00:01:32,769 --> 00:01:36,689
that roughly one fifth of its revenue comes
from just two customers.
27
00:01:36,769 --> 00:01:42,525
On governance, the seven cofounders keep around
50.1% of voting power through a single founder
28
00:01:42,525 --> 00:01:47,325
share, while the long term benefit trust they
set up, elects a majority of the board, last
29
00:01:47,325 --> 00:01:51,325
valued at 965,000,000,000 in a private round in
May.
30
00:01:51,325 --> 00:01:57,209
This listing will, of course, be the final test
of how public markets price a frontier AI lab.
31
00:01:57,209 --> 00:01:57,450
Yeah.
32
00:01:57,450 --> 00:02:01,930
Charles, wanted to ask you what the general
pulse is feeling like in San Francisco with the
33
00:02:01,930 --> 00:02:04,250
impending IPO on the horizon.
34
00:02:04,534 --> 00:02:09,014
So I had a meeting yesterday at an office with
one of the people who was doing the book
35
00:02:09,014 --> 00:02:12,455
underwriting on the IPO, and it was it was busy
in that office.
36
00:02:12,455 --> 00:02:15,574
I think everyone's trying to figure out what
will happen.
37
00:02:15,574 --> 00:02:17,254
Of course, with the headlines like, oh my god.
38
00:02:17,254 --> 00:02:19,259
All of this wealth infusion into San Francisco.
I'm
39
00:02:19,099 --> 00:02:22,300
I'm like, well, someone who lives here, there's
already a lot of wealth in San Francisco.
40
00:02:22,300 --> 00:02:27,419
And two, unlike the Google Facebook era, a lot
of that wealth has already leaked into the
41
00:02:27,419 --> 00:02:29,500
market through secondaries and tenders.
42
00:02:29,500 --> 00:02:33,264
And also just like the general appreciation in
mag savvy stocks, and a lot of those people
43
00:02:33,264 --> 00:02:33,745
live here.
44
00:02:33,745 --> 00:02:37,665
So I think it's gonna be fascinating to see
what happens.
45
00:02:37,824 --> 00:02:42,224
A few of my friends were like, oh, well, once
this goes public and people start to leave, I'm
46
00:02:42,224 --> 00:02:47,099
like, I've seen some crazy stats about the
number of centimillionaire and decimillionaires
47
00:02:47,099 --> 00:02:48,219
that are gonna be made.
48
00:02:48,219 --> 00:02:49,180
They're gonna be minted.
49
00:02:49,180 --> 00:02:52,139
I'm like, look, I wouldn't assume that all
these people are just gonna go start tech
50
00:02:52,139 --> 00:02:52,780
companies.
51
00:02:52,860 --> 00:02:57,655
For a couple reasons, one, they might have a
really robust view of what Anthropics near term
52
00:02:57,655 --> 00:02:58,134
potential is.
53
00:02:58,134 --> 00:03:00,215
And so, like, I'm not jumping in front of that
buzzsaw.
54
00:03:00,215 --> 00:03:01,814
They might just say, I have enough money.
55
00:03:01,814 --> 00:03:03,495
I wanna go become a pastry chef.
56
00:03:03,495 --> 00:03:06,215
And I bring these things up because, like, this
happened in the .com era.
57
00:03:06,215 --> 00:03:07,414
There are people who are just like, wow.
58
00:03:07,414 --> 00:03:09,735
I suddenly have all the money I need.
59
00:03:09,735 --> 00:03:13,340
Now that I have freedom, I wanna go pursue
things outside of tech.
60
00:03:13,340 --> 00:03:14,620
I think it'll be fascinating.
61
00:03:14,620 --> 00:03:15,659
And for me, the...
62
00:03:15,659 --> 00:03:20,539
One of the more interesting things is we'll
finally have public eyes on these companies.
63
00:03:20,539 --> 00:03:26,574
I think everything else has been more innuendo
or leak reports or company status stuff.
64
00:03:26,574 -->
Load full transcript
00:03:30,735
And I think we'll finally have, you know, 10 q
level reporting on these companies.
65
00:03:30,735 --> 00:03:35,294
And I think the interesting thing is it'll give
people a chance to inspect these companies
66
00:03:35,294 --> 00:03:41,009
maybe who are slightly more skeptical or who
think differently about the long term value of
67
00:03:41,009 --> 00:03:41,489
these companies.
68
00:03:41,489 --> 00:03:44,370
And I think it's a real bellwether moment for
Venture.
69
00:03:44,370 --> 00:03:48,449
I think it's a bellwether moment for a handful
of firms like Spark and Menlo.
70
00:03:48,610 --> 00:03:49,250
And I don't know.
71
00:03:49,250 --> 00:03:50,930
I think it's gonna be fascinating.
72
00:03:50,930 --> 00:03:53,810
I keep looking at what's happening in the macro
environment.
73
00:03:54,235 --> 00:03:58,955
And the one thing that I think is fascinating
is if I were Anthropic, I'd be like, compute is
74
00:03:58,955 --> 00:04:01,354
this, like, existential risk for us.
75
00:04:01,354 --> 00:04:06,634
We're able to cobble it together by calling our
friends the hyperscalers and SpaceX.
76
00:04:06,634 --> 00:04:09,194
And we have these neo clouds, and we find
together...
77
00:04:09,194 --> 00:04:13,280
At some point, I'm like, at $500,000,000,000,
it's clearly existential.
78
00:04:13,280 --> 00:04:17,040
And I'll be very curious to see how they think
about financing compute going forward.
79
00:04:17,040 --> 00:04:23,120
Well, there is the undercurrent conversation I
felt like on X that this is as good as what the
80
00:04:23,120 --> 00:04:26,625
business is gonna look like ever, and so that's
why they won IPO today.
81
00:04:26,625 --> 00:04:27,904
Do you have any comments on that?
82
00:04:27,904 --> 00:04:29,504
I guess it's all skeptical at this point.
83
00:04:29,504 --> 00:04:30,785
I think it might've...
84
00:04:30,785 --> 00:04:31,904
I'm not sure who said it.
85
00:04:31,904 --> 00:04:33,425
It might've been Benedict Evans.
86
00:04:33,665 --> 00:04:36,384
Someone said that Anthropic s one is a
Rorschach test.
87
00:04:36,384 --> 00:04:38,384
And on the one hand, you're like, wow.
88
00:04:38,464 --> 00:04:43,930
This business is highly dependent on a very
small number of customers spending a whole lot
89
00:04:43,930 --> 00:04:44,569
of money.
90
00:04:44,569 --> 00:04:49,129
That's a bad signal for the health and wellness
of the AI business.
91
00:04:49,129 --> 00:04:54,329
The flip side is this is a gigantic business
where only a very small number of people have
92
00:04:54,329 --> 00:04:58,464
figured out how to take advantage of this
technology, imagine how much greater it can
93
00:04:58,464 --> 00:05:01,024
become when others get on the same train.
94
00:05:01,024 --> 00:05:04,865
And in some ways, I feel like that's the
existential question about all of this AI
95
00:05:04,865 --> 00:05:05,185
stuff.
96
00:05:05,185 --> 00:05:10,529
It's like, are the San Francisco frontier
billboard people, are they living
in the future
97
00:05:10,529 --> 00:05:12,209
that we will all inhabit?
98
00:05:12,209 --> 00:05:18,289
And years from now, many people will be AI
maxing and using AI as aggressively as the
99
00:05:18,289 --> 00:05:20,129
people in my zip code?
100
00:05:20,129 --> 00:05:24,125
Or are we on some weird feature branch of the
future where we're running an experiment where
101
00:05:24,125 --> 00:05:27,644
a very small number of people are using this
technology very intensely, and that feature
102
00:05:27,644 --> 00:05:29,564
branch will, like, never end up on main.
103
00:05:29,564 --> 00:05:32,524
Because most people will be like, I'm not gonna
do what you're doing.
104
00:05:32,524 --> 00:05:33,644
I'm not gonna live that way.
105
00:05:33,644 --> 00:05:37,240
AI is not gonna become that integral into my
business or life.
106
00:05:37,240 --> 00:05:40,439
And I think if we're on a future branch, we're
probably overdoing it.
107
00:05:40,439 --> 00:05:44,759
And if that future branch ends up back on main,
then, like, we're just living in the future.
108
00:05:44,839 --> 00:05:47,079
We're just, like, a few years ahead of what
everybody else will do.
109
00:05:47,084 --> 00:05:50,204
And I think it's funny if you live in San
Francisco, you're like, well, of course, we're
110
00:05:50,204 --> 00:05:51,725
just living in the near future.
111
00:05:51,725 --> 00:05:56,125
Everybody else is going to be using this stuff
in the same way with the same intensity as we
112
00:05:56,125 --> 00:05:56,365
are.
113
00:05:56,365 --> 00:06:00,889
And my experience has been, that's not how
technology diffusion It always takes longer
114
00:06:00,889 --> 00:06:01,850
than people think.
115
00:06:01,850 --> 00:06:06,170
And I think we're gonna be in FDE land for a
lot longer than people think.
116
00:06:06,170 --> 00:06:07,930
I think it's gonna take a lot longer for...
117
00:06:07,930 --> 00:06:09,930
Because I've seen it even in our little firm.
118
00:06:09,930 --> 00:06:14,490
The changes you have to make to the way that
you work to really get value out of AI are not
119
00:06:14,490 --> 00:06:14,810
trivial.
120
00:06:15,115 --> 00:06:16,954
And it's not just like laying off people.
121
00:06:17,034 --> 00:06:21,435
It's also reimagining the way that work gets
done, ripping stuff up, and, you know, getting
122
00:06:21,435 --> 00:06:22,954
people to change habits and habits.
123
00:06:22,954 --> 00:06:25,115
Habits are like a safety blanket in corporate
America.
124
00:06:25,115 --> 00:06:29,160
You're like, we do this thing this way, and if
I follow this rule and do it this way, I'm
125
00:06:29,160 --> 00:06:29,479
safe.
126
00:06:29,479 --> 00:06:34,199
It's a good tie into our second piece this
week, which comes from a post by Nick Rubin
127
00:06:34,199 --> 00:06:37,319
over at First Round titled The Gap Between GPUs
and Compute.
128
00:06:37,319 --> 00:06:41,879
It was a recap of a dinner he recently hosted
with around six people who buy, they deploy,
129
00:06:41,879 --> 00:06:45,964
and operate compute for some of the biggest AI
labs and infra companies.
130
00:06:45,964 --> 00:06:51,324
The headline is that demand is being treated as
infinite for now, so people don't necessarily
131
00:06:51,324 --> 00:06:52,125
care about that.
132
00:06:52,125 --> 00:06:54,444
The real constraint isn't customers.
133
00:06:54,444 --> 00:06:55,644
It's capital and power.
134
00:06:55,980 --> 00:06:57,580
It creates a catch 22.
135
00:06:57,580 --> 00:07:02,540
Buyers will pay a premium for capacity today,
but hesitate to commit to a specific generation
136
00:07:02,540 --> 00:07:04,540
of GPUs a year or two out.
137
00:07:04,540 --> 00:07:08,379
And without those commitments, providers can't
really get the financing to build what they
138
00:07:08,379 --> 00:07:09,020
need to.
139
00:07:09,100 --> 00:07:14,185
Even after a contract is signed, getting a
cluster running is so hard and labor intensive
140
00:07:14,185 --> 00:07:19,064
that one operator offered to send their own
people into a provider's data center, actually,
141
00:07:19,064 --> 00:07:20,345
from that dinner conversation.
142
00:07:20,425 --> 00:07:25,990
And another attendee watched a fully approved
project nearly get blocked by locals at a city
143
00:07:25,990 --> 00:07:28,870
hearing, which are dynamics we've talked about
a lot on the podcast.
144
00:07:28,870 --> 00:07:30,870
Nick's takeaway sums it up pretty well.
145
00:07:31,029 --> 00:07:36,230
Quote, AI moves fast, but the infrastructure
underneath it takes years to build.
146
00:07:36,629 --> 00:07:39,845
There's been more and more chatter about how
the financing should work for the actual
147
00:07:39,845 --> 00:07:41,525
infrastructure that enables compute.
148
00:07:41,525 --> 00:07:43,685
This post is obviously a really good reminder
of it.
149
00:07:43,685 --> 00:07:46,085
We haven't even talked about the neo cloud
drama
150
00:07:46,165 --> 00:07:46,564
Yes.
151
00:07:46,564 --> 00:07:49,365
That swirled last week, which is a bit adjacent
to this.
152
00:07:49,365 --> 00:07:54,009
But the concept that there is a disconnect
between owning GPUs and having usable compute
153
00:07:54,009 --> 00:07:56,490
was the main takeaway in this post.
154
00:07:56,569 --> 00:08:00,329
Any additional thoughts you've been having on
this entire subject right now?
155
00:08:00,329 --> 00:08:03,689
I know you talked about kind of like different
errors that you've lived through before.
156
00:08:03,689 --> 00:08:08,884
We've talked about fiber and laying all of that
into the ground more than what we needed to.
157
00:08:08,884 --> 00:08:09,925
How does this compare?
158
00:08:09,925 --> 00:08:11,284
I think it's fascinating.
159
00:08:11,284 --> 00:08:15,444
And maybe I'm being a little bit of a nervous
Nelly, but I feel like people aren't paying
160
00:08:15,444 --> 00:08:19,490
enough attention to what's happening in the
credit markets when it comes to finance. If
161
00:08:19,170 --> 00:08:24,290
If I look at a lot of this AI data center
spend, it's not all speculative in the sense
162
00:08:24,290 --> 00:08:28,050
that, like, if we had this capacity, I think a
lot of it would get used today.
163
00:08:28,050 --> 00:08:31,730
It's speculative in the sense that, like, the
cash flows for this stuff are gonna come many,
164
00:08:31,730 --> 00:08:33,090
many years in the future.
165
00:08:33,090 --> 00:08:37,304
And I don't know how you finance the level of
compute expansion that's required without
166
00:08:37,304 --> 00:08:39,225
capital markets being supportive.
167
00:08:39,304 --> 00:08:43,784
And I think with the Neo Cloud thing, it's the
first wobble we've seen where people are like,
168
00:08:43,784 --> 00:08:45,065
well, maybe these...
169
00:08:45,065 --> 00:08:46,904
All these Neo Clouds won't work out.
170
00:08:46,904 --> 00:08:50,399
And, also, like, I've talked to a handful of
our portfolio companies, including one one
171
00:08:50,399 --> 00:08:54,559
founder who's done a real deep dive on sub 10
megawatt data center construction.
172
00:08:54,559 --> 00:08:59,519
What he describes to me sounds kind of like
buying any illegal substance on the black
173
00:08:59,519 --> 00:08:59,919
market.
174
00:08:59,919 --> 00:09:01,120
He's like, I found this guy.
175
00:09:01,120 --> 00:09:02,639
He's like, I got GPUs.
176
00:09:02,720 --> 00:09:04,044
And he's like, you got GPUs.
177
00:09:04,044 --> 00:09:05,085
Let me see the GPUs.
178
00:09:05,085 --> 00:09:05,485
No.
179
00:09:05,485 --> 00:09:06,125
Well, okay.
180
00:09:06,125 --> 00:09:07,004
I actually don't have GPUs.
181
00:09:07,004 --> 00:09:08,365
I got a guy who's got g p...
182
00:09:08,365 --> 00:09:12,365
Everything is murky and fast and loose because
the demand is so high.
183
00:09:12,365 --> 00:09:17,290
And so I don't know what happens if the credit
markets begin to tighten for some of this AI
184
00:09:17,290 --> 00:09:17,929
demand.
185
00:09:17,929 --> 00:09:19,370
I don't know what happens if some...
186
00:09:19,370 --> 00:09:20,970
Know, everything gets contracted compute.
187
00:09:20,970 --> 00:09:25,129
I'm like, well, the risk in the contracted
compute scenario is that the contracted entity
188
00:09:25,129 --> 00:09:26,090
cannot pay.
189
00:09:26,090 --> 00:09:26,490
Yes.
190
00:09:26,490 --> 00:09:30,090
They have signed and they have committed, but
you've gotta go build the thing and they've
191
00:09:30,090 --> 00:09:32,085
gotta be around to pay you in the future.
192
00:09:32,085 --> 00:09:33,684
And I think that is risky.
193
00:09:33,684 --> 00:09:37,125
And if you treat that as, a 100% certainty,
that's probably not good business.
194
00:09:37,125 --> 00:09:40,325
I really wonder what happens if the credit
markets either...
195
00:09:40,325 --> 00:09:42,565
I don't think they'll shut off AI spend.
196
00:09:42,565 --> 00:09:43,924
I think that would be wild.
197
00:09:43,924 --> 00:09:46,730
But the cost of capital could go up pretty
materially.
198
00:09:46,730 --> 00:09:50,409
I think that's what you're seeing with, like,
some of the the Corweave stuff around pricing
199
00:09:50,409 --> 00:09:52,090
and the credit default swaps.
200
00:09:52,090 --> 00:09:56,009
And even with Oracle, these things to pencil,
they're harder to pencil if, basically, the
201
00:09:56,009 --> 00:09:59,115
effective cost of capital or interest rates you
have to pay and that money goes And if you
202
00:09:59,115 --> 00:10:03,115
think compute is tight and hard to get now, I
think it was AWS said, hey, we're basically
203
00:10:03,115 --> 00:10:04,475
sold out through next year.
204
00:10:04,475 --> 00:10:08,955
And again, a lot of people are contracting for
compute that they expect to need, but you can
205
00:10:08,955 --> 00:10:15,759
envision a a cycle where capital for data
center build gets expensive, and so people
206
00:10:15,759 --> 00:10:21,759
build less and also get tighter on ROI, maybe
charge more, which means people say, oh, that
207
00:10:21,759 --> 00:10:24,924
thing I was doing that was fun and free, maybe
I can't do that. You can
208
00:10:24,524 --> 00:10:27,644
You can envision, like, a cycle, a feedback
loop that gets negative.
209
00:10:27,644 --> 00:10:31,164
I don't think we're on the cusp of that, but
I'm certainly keeping an eye on this because by
210
00:10:31,164 --> 00:10:35,565
all accounts, we need a lot more compute, and
the only way I can see to finance that is with
211
00:10:35,565 --> 00:10:35,884
credit.
212
00:10:36,410 --> 00:10:40,649
And even if I were to bring it to the equity
side of conversations, especially for some of
213
00:10:40,649 --> 00:10:44,490
the folks who are building in that space, I
mean, rounds that have been announced from
214
00:10:44,490 --> 00:10:47,370
folks like Etch, they need hundreds of millions
of dollars.
215
00:10:47,370 --> 00:10:52,575
The question becomes, at what point will those
funds realize if it's working or not working?
216
00:10:52,575 --> 00:10:55,375
Because I think Nick said it well, capital is
becoming the moat.
217
00:10:55,375 --> 00:10:57,934
You absolutely need it in order to execute.
218
00:10:57,934 --> 00:11:02,175
You could dump in hundreds and hundreds of
millions of dollars into something for a large
219
00:11:02,175 --> 00:11:03,934
period of time to give yourself some runway.
220
00:11:04,389 --> 00:11:07,509
At what point do you realize it's not going to
come to fruition?
221
00:11:07,509 --> 00:11:11,750
Again, credit markets, when they fall apart,
are far more detrimental to ripple effects on
222
00:11:11,750 --> 00:11:12,470
the economy.
223
00:11:12,470 --> 00:11:17,669
One thing that they talked about, which we have
a bit, is on the public opposition side.
224
00:11:17,669 --> 00:11:22,884
The clearest hope, I think he said, at the
table was that AI starts delivering tangible
225
00:11:22,884 --> 00:11:28,245
benefits that the average person can feel, and
and that being maybe a reason why people could
226
00:11:28,245 --> 00:11:31,365
become more supportive around data centers
being built in their backyards.
227
00:11:31,365 --> 00:11:35,759
But I think the reality is the infrastructure
has to be built before those benefits arrive.
228
00:11:35,759 --> 00:11:36,240
Yeah.
229
00:11:36,240 --> 00:11:40,559
How much runway do you think the industry has
before local and political pushback
230
00:11:40,559 --> 00:11:42,240
meaningfully slows down?
231
00:11:42,240 --> 00:11:46,975
Do you think that maybe the anthropic IPO will
make this easier or further for the public to
232
00:11:46,975 --> 00:11:47,375
stomach?
233
00:11:47,375 --> 00:11:51,615
I'm speaking more specifically about people who
have been opposed to larger data centers, the
234
00:11:51,615 --> 00:11:53,855
infrastructure of it all, be built so closely
to them.
235
00:11:53,855 --> 00:11:59,294
I mean, if you look at really the story of the
Anthropic IPO, there's a very small handful of
236
00:11:59,294 --> 00:12:02,014
people who live in San Francisco who are going
to make a lot of money.
237
00:12:02,289 --> 00:12:03,169
And and if you say, okay.
238
00:12:03,169 --> 00:12:04,450
Maybe not San Francisco proper.
239
00:12:04,450 --> 00:12:05,330
San Francisco Bay...
240
00:12:05,330 --> 00:12:09,649
And that includes early investors, employees,
real estate agents, commercial lead...
241
00:12:09,649 --> 00:12:11,649
Like, a lot of people in a very localized...
242
00:12:11,649 --> 00:12:13,730
That tends not to build goodwill.
243
00:12:13,730 --> 00:12:18,295
If the data center's going up in your backyard
and somebody in my zip code is making a lot of
244
00:12:18,295 --> 00:12:22,934
money, you're like, how does how does getting
rich help me deal with this big monstrosity in
245
00:12:22,934 --> 00:12:23,735
my backyard?
246
00:12:23,735 --> 00:12:29,335
I think also we have this weird duality even in
the s one where it's like this is the greatest
247
00:12:28,889 --> 00:12:30,970
technology and wealth creation thing ever.
248
00:12:30,970 --> 00:12:32,170
Also, it might kill us all.
249
00:12:32,170 --> 00:12:35,850
It might kill us all and we'll probably take
your job, but otherwise, everything is amazing.
250
00:12:35,850 --> 00:12:40,009
And it's almost as if nobody put any thought
into two things.
251
00:12:40,490 --> 00:12:44,725
One, is there any validity whatsoever to
people's concerns and fears?
252
00:12:44,725 --> 00:12:47,125
And some of this is fear, and you should never
dismiss it.
253
00:12:47,125 --> 00:12:52,644
And and two, to the extent that their fears are
valid, how do we wanna respond to them other
254
00:12:52,644 --> 00:12:57,240
than being dismissive or pretending that those
fears are people trying to slow down the rate
255
00:12:57,240 --> 00:12:57,960
of progress?
256
00:12:58,039 --> 00:12:59,399
It turns out they can.
257
00:12:59,399 --> 00:13:01,879
People are just like, oh, they're Luddites and,
like, they don't get it.
258
00:13:01,879 -->
00:13:03,559
They're trying to slow down the the wheels
of...
259
00:13:03,559 --> 00:13:06,840
I'm like, well, they seem to be succeeding by
blocking these data center projects.
260
00:13:06,840 --> 00:13:11,345
So whether out of necessity or out of care, I
think at some point, the AI collective
261
00:13:11,345 --> 00:13:14,625
marketing machine is gonna have to figure out
how do we really wanna talk to people about
262
00:13:14,625 --> 00:13:15,504
this technology.
263
00:13:15,504 --> 00:13:16,705
Also, people aren't dumb.
264
00:13:16,705 --> 00:13:21,350
They see their bosses at work telling them to
use AI, and they're getting AI slop from their
265
00:13:21,350 --> 00:13:21,829
colleagues.
266
00:13:21,829 --> 00:13:25,029
And they're seeing people get laid off so that
the company can do more with AI.
267
00:13:25,029 --> 00:13:30,070
And people are just like, other than maybe Muse
and Instinct being free and useful, how does
268
00:13:30,070 --> 00:13:30,309
that...
269
00:13:30,309 --> 00:13:32,629
Why does that matter if I don't have a job and
can't pay my mortgage?
270
00:13:32,964 --> 00:13:35,684
Was talking to somebody else yesterday, and
they're like, I don't think the average person
271
00:13:35,684 --> 00:13:39,605
is, like, reading through an s one for an
impending IP owned.
272
00:13:39,605 --> 00:13:42,964
I'm like, like, I'm sure everyone has more
impactful things to do with their time who
273
00:13:42,964 --> 00:13:44,725
don't sit with an tech NVC.
274
00:13:44,725 --> 00:13:48,579
But to your point, it's not like they haven't
heard this down the pipeline, and it's only
275
00:13:48,579 --> 00:13:49,779
been in a negative way.
276
00:13:49,779 --> 00:13:53,940
And they still haven't quite figured out how to
position this in a way that benefits everyone.
277
00:13:53,940 --> 00:13:58,179
And when I read this piece, I think that one
that one part really stood out to me.
278
00:13:58,179 --> 00:14:02,179
In order to get everyone excited about it and
to feel the benefits of it, you have to build
279
00:14:02,179 --> 00:14:03,220
the infrastructure infrastructure first.
280
00:14:02,914 --> 00:14:03,394
First.
281
00:14:03,394 --> 00:14:05,475
You can't do it before.
282
00:14:05,794 --> 00:14:07,714
So there has to be some other give.
283
00:14:07,714 --> 00:14:10,355
And I don't know what that'll look like, and I
don't know if we we've I
284
00:14:10,355 --> 00:14:15,394
think if most people were like, you mean with
ChatGPT, I can get paid for five days of work
285
00:14:15,394 --> 00:14:17,990
and work for four days, and my boss will never
know.
286
00:14:17,990 --> 00:14:21,910
AI would be like the most most popular
technology like ever created.
287
00:14:22,070 --> 00:14:23,830
But that's not most people's experience.
288
00:14:23,830 --> 00:14:24,389
No.
289
00:14:24,389 --> 00:14:28,149
If it was that of great, I can win back three
hours of my day and sign off.
290
00:14:28,309 --> 00:14:31,764
Most people aren't like, oh, I'd love to win
three back three hours back of my day so I can
291
00:14:31,764 --> 00:14:32,644
become more productive.
292
00:14:32,644 --> 00:14:34,245
The average person isn't thinking like that.
293
00:14:34,245 --> 00:14:35,205
Like, they're just not.
294
00:14:35,205 --> 00:14:39,125
The average person is like, I can win back
three hours of my day, spend time with my
295
00:14:39,125 --> 00:14:42,165
family, cut back costs on childcare because I
can do it at home.
296
00:14:42,165 --> 00:14:44,565
No one's thinking like, so I can become a
superhuman.
297
00:14:44,929 --> 00:14:47,730
That's the average person, and I just don't
think we've bridged that gap yet.
298
00:14:47,730 --> 00:14:50,529
And and I think the employer view is like, my
gosh.
299
00:14:50,529 --> 00:14:56,129
For the same eight hours, I can now get 1.3,
1.4 x the productivity of this person.
300
00:14:56,129 --> 00:14:59,004
And I think this is like the lived experience
of a lot of people.
301
00:14:59,004 --> 00:15:02,284
It's like, oh, AI allows me to do more work per
unit of time.
302
00:15:02,284 --> 00:15:05,644
It's not to, like, do the same fixed amount of
work faster.
303
00:15:05,644 --> 00:15:06,924
It's a little bit of John Henry.
304
00:15:06,924 --> 00:15:08,605
It's like, now I have to keep up with the line.
305
00:15:08,605 --> 00:15:09,485
Keep up the machine.
306
00:15:09,485 --> 00:15:10,845
More is being asked of me.
307
00:15:10,845 --> 00:15:12,764
This is probably great for the person who owns
the
308
00:15:12,819 --> 00:15:13,139
Yes.
309
00:15:13,139 --> 00:15:16,100
The five people who run each company are are
all...
310
00:15:16,100 --> 00:15:16,419
Yeah.
311
00:15:16,419 --> 00:15:18,100
All get to see the benefits first.
312
00:15:18,100 --> 00:15:23,059
Well, our last piece comes from Jay Gibson over
at Better Tomorrow Ventures titled On Risk, and
313
00:15:23,059 --> 00:15:26,500
his argument is that venture is starting to
look a lot like public markets.
314
00:15:26,500 --> 00:15:28,100
Everybody owns the same handful of names.
315
00:15:28,295 --> 00:15:30,134
Everyone is benchmarked against each other.
316
00:15:30,134 --> 00:15:30,375
Mhmm.
317
00:15:30,375 --> 00:15:32,695
And closet indexing is on the rise.
318
00:15:32,934 --> 00:15:34,134
All stats we've talked about.
319
00:15:34,134 --> 00:15:38,695
Established firms took 88.2% of all US venture
capital raised so far this year.
320
00:15:38,695 --> 00:15:45,110
Megadeals of a $100,000,000 or more captured
around almost 88% of the $412,000,000,000
321
00:15:45,110 --> 00:15:46,389
deployed in the first half.
322
00:15:46,389 --> 00:15:51,269
Jake says this is the opposite of 2021 because
now even high quality, capital efficient
323
00:15:51,269 --> 00:15:55,634
companies get told they're not venture scale
unless they check exactly the right boxes on
324
00:15:55,634 --> 00:15:58,274
category, pedigree, cap table, and AI
narrative.
325
00:15:58,274 --> 00:16:04,514
He also calls
out the rise of Vibe revenue from
two week pilots annualized into ARR to
326
00:16:04,514 --> 00:16:08,090
contracts with ninety day opt outs, counted as
reoccurring.
327
00:16:08,090 --> 00:16:12,090
Jake says that the market is trading career
risk or financial risk, saying for most of the
328
00:16:12,090 --> 00:16:17,049
ecosystem, it's becoming well understood for
your reputation to fail conventionally than to
329
00:16:17,049 --> 00:16:18,330
succeed unconventionally.
330
00:16:18,330 --> 00:16:21,929
So invest in the safe, expensive options, the
more legible things to everyone.
331
00:16:22,235 --> 00:16:27,595
His prescription is patience at a time when
patience is the world's most scarce resources.
332
00:16:27,595 --> 00:16:29,115
Wanted your thoughts on this one.
333
00:16:29,274 --> 00:16:35,274
Obviously, again, with the anthropic IPO on the
horizon and likely folks going to use that as
334
00:16:34,429 --> 00:16:37,949
ammunition to continue piling into consensus
market, if it goes well.
335
00:16:37,949 --> 00:16:39,789
I read Jake's post and I texted him.
336
00:16:39,789 --> 00:16:40,829
I thought it was very good.
337
00:16:40,829 --> 00:16:43,309
And Also, comments on it were funny.
338
00:16:43,309 --> 00:16:44,750
People were like, who hurt you?
339
00:16:44,750 --> 00:16:45,069
I know.
340
00:16:46,684 --> 00:16:50,125
But we talked in an earlier episode about the
Nifty Fifty and that whole era.
341
00:16:50,125 --> 00:16:51,084
I read Jake's post.
342
00:16:51,084 --> 00:16:51,485
I don't know.
343
00:16:51,485 --> 00:16:51,964
The one...
344
00:16:51,964 --> 00:16:56,924
There's one book I come back to all the time
because I think it explains a lot of the world.
345
00:16:56,924 --> 00:17:01,610
It's The Unaccountability Machine, which shout
out to James Cham for pointing out that book to
346
00:17:01,610 --> 00:17:02,330
me a long time ago.
347
00:17:02,330 --> 00:17:06,409
One of the premises of the book is, like, every
system works exactly as designed.
348
00:17:06,490 --> 00:17:10,250
And I'm like, well, what if we're in a moment
where, like, really what you're seeing is the
349
00:17:10,250 --> 00:17:12,809
reflected preferences of the people who
allocate capital?
350
00:17:13,105 --> 00:17:18,625
And that whether they came to this view on
their own or they were influenced by VCs, the
351
00:17:18,625 --> 00:17:22,944
view is, look, there's a very small number of
companies that are consequential.
352
00:17:22,944 --> 00:17:26,625
And it's hard to argue that SpaceX and
Anthropic aren't two of those five.
353
00:17:27,070 --> 00:17:31,789
Those companies are so consequential that very
little else can measure up and there's a
354
00:17:31,789 --> 00:17:37,470
relatively small number of people who we
believe will have consistent access to those
355
00:17:37,470 --> 00:17:38,109
companies.
356
00:17:38,430 --> 00:17:40,349
Why don't we just give them all of the money?
357
00:17:40,654 --> 00:17:43,134
And I'm like, but you'll never know the
counterfactual.
358
00:17:43,214 --> 00:17:47,134
You'll never know like what other people would
have found amazing companies if you'd done
359
00:17:47,134 --> 00:17:47,934
something different.
360
00:17:47,934 --> 00:17:52,894
You'll never know whether those people caught a
heater and been like right for some extended
361
00:17:52,894 --> 00:17:54,894
period of time but that that's not
reproducible.
362
00:17:54,990 --> 00:17:57,470
There are all of these things you simply cannot
know.
363
00:17:57,470 --> 00:17:59,549
And I think a lot of people were just like,
it's fine.
364
00:17:59,549 --> 00:18:00,429
This is easy.
365
00:18:00,750 --> 00:18:05,150
Give all of the money to the small number of
people who we think can find exceptional
366
00:18:05,150 --> 00:18:05,789
companies.
367
00:18:05,869 --> 00:18:07,789
Hope that they find them and just be done.
368
00:18:08,125 --> 00:18:10,044
If you think about it, everything else is
harder.
369
00:18:10,044 --> 00:18:15,005
Meeting hundreds of fund managers to find a few
people that you think have a special talent for
370
00:18:15,005 --> 00:18:19,565
company selection, and then waiting ten years
to see if those people are right, that's
371
00:18:19,565 --> 00:18:20,284
harder.
372
00:18:20,284 --> 00:18:24,519
And so sometimes I read things like what Jake
wrote, I'm like, maybe the system is designed
373
00:18:24,519 --> 00:18:27,480
to reflect the preferences of the people at
Capital who are just...
374
00:18:27,480 --> 00:18:28,839
I want this to be easy.
375
00:18:29,159 --> 00:18:33,079
I wanna make fewer decisions, and I wanna
increase my probability that I'm in the
376
00:18:33,079 --> 00:18:36,200
relatively small number of names that matter,
and that's it.
377
00:18:36,315 --> 00:18:42,315
I was getting coffee with a founder yesterday
morning, and it struck me that it's never been
378
00:18:42,315 --> 00:18:48,074
so important or inherent to the job of a
founder to understand the business of venture
379
00:18:48,075 --> 00:18:49,115
now more than ever.
380
00:18:49,115 --> 00:18:53,079
Whereas before, like, you could really just
focus on executing as a founder and being
381
00:18:53,079 --> 00:18:55,799
obsessed with an idea in a market and going in
that direction.
382
00:18:55,799 --> 00:19:00,759
I would say, you need to know what's going on
in venture land from the growing fund
383
00:19:00,759 --> 00:19:05,720
perspective, and then need to be a 5 to
$20,000,000,000 outcome in order to be remotely
384
00:19:05,720 --> 00:19:06,039
interesting.
385
00:19:06,345 --> 00:19:09,304
Previously, I really do think you could just be
a founder.
386
00:19:09,304 --> 00:19:11,625
Now I'm like, you really need to start thinking
like a VC.
387
00:19:11,625 --> 00:19:12,984
I don't think that used to be the case.
388
00:19:13,224 --> 00:19:18,184
One thing I was gonna ask in 2021 when we were
in the last cycle, obviously different.
389
00:19:18,184 --> 00:19:18,984
He said it
well.
390
00:19:18,984 --> 00:19:23,929
Like, every good, bad company was getting
overly funded at overly valued prices.
391
00:19:23,930 --> 00:19:28,009
And we did have funds at larger sizes back
then, obviously not at the magnitude that they
392
00:19:28,009 --> 00:19:28,889
are today.
393
00:19:29,130 --> 00:19:35,049
Why didn't we see, like, an immediate drop off
of funding for, like, sub $5,000,000,000 market
394
00:19:34,484 --> 00:19:35,365
companies back then?
395
00:19:35,365 --> 00:19:38,565
Is it really just because we didn't know those
were that possible?
396
00:19:38,724 --> 00:19:40,804
Why didn't we see as much of a drop off?
397
00:19:40,804 --> 00:19:43,044
I think this cycle is gonna last.
398
00:19:43,045 --> 00:19:46,644
If you think about it, there's still a lot of
zombie corn, whatever you wanna call it.
399
00:19:46,644 --> 00:19:51,099
A lot of companies that raised a bunch of money
in 2021 cut expenses, and they're still around.
400
00:19:51,099 --> 00:19:51,819
They're still around.
401
00:19:51,819 --> 00:19:52,460
They're still alive.
402
00:19:52,460 --> 00:19:57,259
They're still operating because they cut costs
and held onto that cash because they knew there
403
00:19:57,259 --> 00:19:58,380
was, like, no more coming.
404
00:19:58,380 --> 00:20:00,859
And for all those companies, we still don't
have a final reckoning.
405
00:20:00,859 --> 00:20:01,819
Like, they haven't shut down.
406
00:20:01,819 --> 00:20:02,539
They haven't sold.
407
00:20:02,539 --> 00:20:06,555
They're on people's books at some price,
probably pretty close to the top of market.
408
00:20:06,555 --> 00:20:10,394
And I think we're kinda doing the same thing
here, except we're doing it two ticks earlier.
409
00:20:10,394 --> 00:20:15,115
Instead of giving companies with a few million
in ARR, a bunch of money on a high future
410
00:20:15,529 --> 00:20:16,409
revenue multiple.
411
00:20:16,409 --> 00:20:19,369
We're giving it to companies at inception,
which means there'll be companies that raise
412
00:20:19,369 --> 00:20:22,970
$2,030.40, $50,000,000 pre product market fit.
413
00:20:22,970 --> 00:20:27,609
If things tighten, those companies will still
have 20 or $30,000,000 in the bank and Yeah.
414
00:20:27,769 --> 00:20:30,809
Will abandon whatever original scaling and
business model they had.
415
00:20:30,825 --> 00:20:34,264
And those companies could take seven to ten
years to play out.
416
00:20:34,264 --> 00:20:39,224
But the people who invested in them will
probably be able to raise new funds while those
417
00:20:39,224 --> 00:20:41,544
founders are stuck sorting all that stuff out.
418
00:20:41,544 --> 00:20:43,384
And so this is to me, like, the weird irony of
419
00:20:43,384 --> 00:20:43,865
these
420
00:20:43,865 --> 00:20:49,150
very large inception stage or growth stage
financings is it puts the day of reckoning for
421
00:20:49,150 --> 00:20:53,549
these companies way into the future because
they can afford to just operate and live for a
422
00:20:53,549 --> 00:20:54,190
really long time.
423
00:20:54,190 --> 00:21:00,109
That, I think, is the biggest irony is, I would
say, the larger your market or opportunity is,
424
00:21:00,474 --> 00:21:04,474
the less validation you need in the earlier
stages to lock in.
425
00:21:04,554 --> 00:21:09,274
And Jake talks about Vibe Revenue saying if the
bar for a series a is a specific growth curve,
426
00:21:09,274 --> 00:21:10,474
founders will figure it out.
427
00:21:10,474 --> 00:21:13,274
And I think that was my question, and I'd love
your perspective.
428
00:21:13,274 --> 00:21:18,429
You think founders building at this, like,
preconception stage are catching on as well.
429
00:21:18,429 --> 00:21:21,309
There are actually no benchmarks and things
that I need.
430
00:21:21,309 --> 00:21:26,029
There's people I meet who deeply understand the
way that VCs are making decisions.
431
00:21:26,424 --> 00:21:30,345
I'd argue a lot of those people are in San
Francisco or in New York.
432
00:21:30,345 --> 00:21:34,105
They're close enough, and they're watching
people raise, and they're like, oh, this is
433
00:21:34,105 --> 00:21:38,869
some mix of vibes, having a timely problem,
having a team that people understand.
434
00:21:38,869 --> 00:21:40,070
These are the things that matter.
435
00:21:40,070 --> 00:21:43,509
And I have other people who are still chasing
ARR benchmarks from 2020.
436
00:21:43,509 --> 00:21:47,750
And I feel really bad for the people who are
chasing ARR benchmarks from 2021.
437
00:21:47,750 --> 00:21:50,710
Was like, you're playing the wrong game and you
don't even you don't even know it.
438
00:21:51,025 --> 00:21:51,424
No.
439
00:21:51,424 --> 00:21:52,304
It's response...
440
00:21:52,304 --> 00:21:55,184
I get it, like, why you think that it is what
you should be shooting for.
441
00:21:55,184 --> 00:21:59,505
And I almost feel gross saying, like, you need
to abandon that if we're going for something
442
00:21:59,505 --> 00:21:59,984
bigger.
443
00:21:59,984 --> 00:22:03,025
And I feel bad because a lot of times those
people show up, and they're like, here's my
444
00:22:03,025 --> 00:22:03,744
report card.
445
00:22:03,744 --> 00:22:05,180
I, like, have all of the...
446
00:22:05,180 --> 00:22:06,220
I finished all the problems.
447
00:22:06,220 --> 00:22:07,900
I was like, but you were in the wrong chapter.
448
00:22:07,900 --> 00:22:12,619
The irony is I think for better or for worse,
when you're in a capital dense geographic
449
00:22:12,619 --> 00:22:16,859
location like San Francisco or New York, you
get signals about what's happening in in the
450
00:22:16,859 --> 00:22:18,539
fundraising markets whether you want them or
not.
451
00:22:18,805 --> 00:22:20,565
So you don't have to be actively fundraising.
452
00:22:20,565 --> 00:22:25,365
You've got friends, neighbors, people you know,
friends of friends who are raising rounds, and
453
00:22:25,365 --> 00:22:27,125
you're like, how do they get that crazy round?
454
00:22:27,125 --> 00:22:27,684
Oh, dude.
455
00:22:27,684 --> 00:22:31,509
Like, this person said in their DMs and heard
that they like poker game.
456
00:22:31,509 --> 00:22:33,430
You you get a sense of how things are
happening.
457
00:22:33,430 --> 00:22:37,349
And if you're not there, you're reading about
these rounds six to nine months later, and
458
00:22:37,349 --> 00:22:40,630
you're probably scratching your head saying,
how did that company get all of that money?
459
00:22:40,630 --> 00:22:42,789
And all along, you've been playing a different
game.
460
00:22:42,789 --> 00:22:45,750
Would you say one of the most important things
that you need to do right now if you're a
461
00:22:45,750 --> 00:22:48,785
founder and you're wanting to build
consequential comedy, you absolutely have to be
462
00:22:48,785 --> 00:22:49,904
in those two geographies.
463
00:22:49,904 --> 00:22:50,945
Ideally, probably SF.
464
00:22:50,945 --> 00:22:54,225
I think if you're not there, you need
incredible connectivity.
465
00:22:54,225 --> 00:22:55,984
You need incredible connectivity.
466
00:22:56,225 --> 00:23:00,065
So much so that I'm just like, it's probably
easier to just be in one of those places.
467
00:23:00,359 --> 00:23:01,960
Well, we are all out of time today.
468
00:23:01,960 --> 00:23:03,799
Thank you, Charles, for jumping in.
469
00:23:03,799 --> 00:23:05,480
Thank you to all of our listeners as well.
470
00:23:05,480 --> 00:23:08,599
We had a bunch of new folks join us last week,
so stay tuned.
471
00:23:08,599 --> 00:23:10,119
We'll drop a new episode next week.
472
00:23:10,119 --> 00:23:10,519
Thanks.
473
00:23:10,519 --> 00:23:11,079
Bye.
474
00:23:11,079 --> 00:23:11,240
Bye.