00:00:00
Where you at, Jama? Look at those cabinets. I know exactly where he is. I'm here. I can tell by the cabinets.
00:00:05
We'll beep that out. Tomorrow I leave for and then I come home a week from Friday. So that means you're going to be leaving
00:00:12
this week. So it's open next week. So I think you're about to get a house
00:00:17
guest. Wait a second. The house is open. I like late season. Ko Kalin's about to make an appearance.
00:00:23
I mean Ko Kalin's about to bless the host. Oh, I'm going to bless that house. You can be sure. I'm going for a little
00:00:29
Indian summer on the Italian. [Music]
00:00:34
Let your winners ride. [Music]
00:00:42
We open source it to the fans and they've just gone crazy with it. Love you.
00:00:47
[Music] All right, everybody. Welcome back to the number one podcast in the world, the
00:00:54
AllIn podcast. were back and you got your original crew. Some people were
00:01:00
wondering, are people losing their commitment to excellence? Are people out gallivanting? Would we ever see the core
00:01:07
four back? August 14th. We're taping. You'll be listening on August 15th. But gentlemen, we're back. How are you
00:01:13
doing, Chimath? You look well rested. The Core Four. The Core Four. That's new.
00:01:18
Who could ask for anything more? Would you say that the Core Four could be fantastic?
00:01:25
I think On their best days, I would say they are exceptional. Perhaps even
00:01:30
tipping into fantastic. Did you see it yet, Sax? Yeah, it's pretty great. I think they did a great job. Yeah,
00:01:37
it's it's like one of those weird, you know, superhero groups that it's like a
00:01:42
real throwback and I don't think a lot of people get the aesthetic or what they're trying to do there, but I I think you're going to love it. Freeberg,
00:01:48
you didn't see it yet. Fantastic 4. No, you're going to love it for the science. It's It's really got like a real throwback to the ' 50s60s science kind
00:01:56
of aesthetic. Wrapping up summer months here. In a couple of weeks, we'll be in
00:02:01
LA for the All-In Summit. Couple of tickets still available. Allin.com/events
00:02:07
or yada yada yada for the Jcal discount. Freeberg, how's your summer going? You doing okay over there, buddy? You showed
00:02:13
up last week. I show up. You showed up last week. No weeks off for you. had a beautiful mountain bike
00:02:19
ride yesterday with Xander. We ended up Oh, how nice. going on for quite a while. Having a birthday dinner with
00:02:26
Jason tonight. Happy birthday. Happy birthday, Jason. Happy birthday to Jason I mean,
00:02:32
what a privilege for us to play with a professional poker player so regularly like Jason in the home game. We've
00:02:39
always wanted to. We've been playing this home game for two decades and we never had a pro there. We never had somebody who truly had achieved
00:02:46
excellence in the field and so it's like getting to be on the court with like Michael Jordan or something. It's like a real thrill for
00:02:52
Phil Helited too. Yeah. Did Phil get a second with him or an autograph?
00:02:58
Cut it out. Phil's going to get mad. Oh no. We're going to have to do another apology. Do another apology letter. Do
00:03:03
the apology now. JL do the I'll just do a pre-apper
00:03:16
Saxy Poo. How are you? Are are you uh a little rest and um a little rest and relaxation for you?
00:03:22
Yeah, we just got back from Diego's birthday party in the UK. Oh, very nice. One of those destination birthdays.
00:03:30
I mean, so many of those things. Oh, God. You know, it's like three of summer retirement parties. John Herring's
00:03:38
having a retirement party. I don't know what kind of flex that is. The kid's like 31 years old. He's like, "Ah, I got
00:03:43
into SpaceX. I retired." I was like, "I'm an Uber. I'm still working." What are you going to do for the next 60
00:03:50
years? Our lifespans are going double. You got to keep busy, folks. All right. Listen,
00:03:56
what a stacked lineup we have. Oh my god. Alex Karp from Palenteer's coming. You got Alex Karp?
00:04:03
Yes, Freeberg. I like that guy. Did you get any of the people on my list? You had some good names at the last
00:04:08
minute. Yeah, thanks for coming in 8 days. When we send you notes months and months in advance of the summit, when we
00:04:14
actually invite people and plan the event and you don't respond, we move forward. And then 8 days before the
00:04:19
event, you're like, "Oh, here's my list of who I'd like to invite to the summit. Can you guys figure out how to get them there?" Oh, YouTube CEO Neil um is coming.
00:04:28
That's great. Neil Mo. Get his name, just so you know. I was looking up his name. I know his
00:04:34
name's Neil, but I didn't know his last name. Neil Mo. Oh, and oh, look at that. Energy Secretary Chris Wright is back for round two with Jal. I love it. What
00:04:41
a great lineup. All right, let's get started, folks. I think the story of the week is this AI
00:04:48
psychosis. This is a crazy story. People call it chat GPT psychosis as well.
00:04:54
We'll get into that. But if you're an ex user, you've seen this discussion over and over again. It's called being
00:04:59
oneshotted. Okay, that's like uh you get sniped in a video game or something. People are starting to have delusions
00:05:07
that are being triggered by or enhanced by AI chat bots. The core of the issue
00:05:12
seems to be that these chat bots, as you know, are sickopantic. They confirm your beliefs. They agree with you. They're
00:05:18
refusive in their praise. And it's so believable that now people are starting to believe they're in a relationship
00:05:24
with AI. Okay. And uh OpenAI saw this becoming an issue and uh they made some
00:05:30
healthy use updates to chat GBT. It's no longer going to tell you whether or not to break up with your boyfriend, to
00:05:37
leave your spouse. It prompts you to take a break after long sessions.
00:05:42
And they also admitted that their 40 model had been creating some of these psychosis scenarios. quote, "There have
00:05:48
been instances where our 40 model fell short in recognizing signs of delusion
00:05:54
or emotional dependency. Psychology Today had a story uh just recently of
00:05:59
the emerging problem of AI psychosis. Man, people are falling into romantic relationships. They have godlike
00:06:04
delusions and a prominent investor, you may have seen this, seems to be undergoing some AI psychosis according
00:06:10
to people who may be near it. I'm not going to say his name, but it's pretty clear who it is. but he released a video
00:06:16
where he started talking about recursion, all this crazy stuff. You may have um even witnessed a little bit of
00:06:22
this when uh Travis was on the program a couple weeks ago and he said he was like spending his time on the fringes or the
00:06:28
edges of physics. Yeah. And like hitting like it
00:06:34
really can take you down the rabbit hole. There's a really interesting subre. Sorry. Are you saying Travis is
00:06:39
suffering from AI psychosis? I'm saying we may need to do a health check. We may need to do a health check
00:06:45
because smart people can get involved with these AI. So, we may have to do a um a little welfare check on our boy TK.
00:06:52
Can you just tee this up and hand it off so that the rest of us can talk, please? I'm about to do that. I'm about to do
00:06:58
that. Okay, Chimath, you seem to be raring to go with this AI psychosis issue. Go ahead,
00:07:05
enlighten us with your hot take. Go ahead. I miss you, too. You're the
00:07:11
best. You're the best. Hold on. Wait. My phone's ringing. Oh, great. Hey, listen, Andrew. I was
00:07:17
I'm going to send you a link to a Zoom chat. I'm just hoping you Yeah, make me an offer. Sure, I'm in.
00:07:24
Yeah, they just these idiots just paid me $50 million to come join your podcast and make it higher rated than yours. Okay, great. Yeah. Okay, I'll see you on
00:07:30
your boat. Like in the major leagues when like a double trade happens. So, we get Andrew Ross Sorcin and then
00:07:38
Gerser and Gurley get Calcanis. That would be really interesting. It's a three- team deal. It's a three-
00:07:43
team deal. You're also getting Let us know in the comments if you would like to see that trade.
00:07:50
Okay, Shimoff, you got any thoughts on this? You seem to be raring to go on this issue. I don't really have much thoughts on
00:07:56
this AI psychosis issue per se, but I do think that it's part of a bigger trend,
00:08:02
meaning these things aren't created overnight and out of nowhere. I think we've known for a while that
00:08:09
what is Scott Galloway calls it the loneliness epidemic. Is that
00:08:15
which I tend to agree with. I mean I think that there is a profound shift in
00:08:20
how young people connect with each other and with the world at large. And I think
00:08:26
that they are becoming more isolated. And I think the dangerous part of AI is
00:08:33
that it can replace whatever flimsy connection that people have left to
00:08:39
their real world community. So if you think about like how much influence
00:08:45
is shaped and generated in places like Instagram by real people and then go to
00:08:51
the limit of a very useful AI that replaces those people with fake people and fake pictures you can't discern.
00:08:59
but they're willing to be responsive to you and engage with you and talk with you, on the one hand, it's incredibly
00:09:06
interesting. On the other hand, that person is ultimately not real. Yeah.
00:09:11
And so I think whatever it is that people are going through right now, and you can see it in all kinds of data,
00:09:17
right? You see it in marriage rates, you see it in household formation rates, you see it in childbearing rates. Something
00:09:24
is broken. It has been for a while in society. I think all these tools have exacerbated it and I think this is just
00:09:30
yet another step in that direction. Freeberg, what's your take here? Sorry, I apologize. I I'll say one more
00:09:36
thing. I was talking to somebody about this and he mentioned it. He framed it this way and I thought it was really interesting. If you look at online
00:09:42
services, their primary mechanism of action is via dopamine. And they're
00:09:47
trying to find mechanisms to trigger the release of dopamine. And that's what gives you that stimulative sensation,
00:09:53
right? But if you think about real world relationships, having a wife, having a long-term
00:10:00
girlfriend, those are fundamentally about serotonin, not dopamine. The dopamine wears off very quickly. And
00:10:06
what are you left with? You're left with establishing a long-term structural relationship with somebody and the ups and downs of that. And it's not always
00:10:12
peachy keen. That's the same thing with friends. And so I think the other thing that's really interesting to note is
00:10:18
that we are sort of all in on dopamine, dopamine, dopamine. And young people,
00:10:24
particularly the ones that are the most susceptible to this, I'm going to guess, have not had enough structural
00:10:29
interactions and the kinds of relationships that actually create these long-term serotonin-like behaviors so
00:10:35
that you can actually balance this stuff out. And so, I suspect that this thing is going to catch on like wildfire in
00:10:41
part because it's taking advantage of a trend, I think. Yeah, it's a it's a quick hit as opposed to the long-term gain of of deep
00:10:47
friendship. Free, what are your thoughts? I was in the college dorms in 1996.
00:10:53
There was this woman who was in the room next to me and she was on AOL and she
00:10:59
got stuck on AOL like 24 hours a day. She failed out of college. She was completely addicted to
00:11:05
it. She like disappeared into this rabbit hole of doing imming with random people online. It was the craziest
00:11:11
thing. I don't think that that chat is necessarily revealing. I think it's just at a much larger scale this challenge
00:11:18
with this psychosis where you get drawn into these holes. There's a study by a researcher called Julian Holt Lunat who
00:11:23
took 148 other studies, psychological studies to identify the relationship between social interaction and mortality
00:11:31
and strong social relationships increased survival odds in humans by 50%
00:11:36
which is equivalent to quitting smoking more impactful than obesity or
00:11:42
physical inactivity. I think that there's a real struggle that people have with loneliness
00:11:48
that manifest in a bunch of different ways. But psych studies have shown that some of the key motivators for going online
00:11:55
to engage with social interaction systems is self-expression, recognition, curiosity, emotional regulation. You can
00:12:02
test ideas. You can engage cognitively in ways that you can't or you don't have access to in real life. So, there's a
00:12:08
whole lot of motivating factors for people to go deep in social engagement online. And there's never been an
00:12:13
infinite social engagement system until I AI came along, these chat interfaces
00:12:18
came along, right? So rather than have to go find someone on the AOL im or find someone in a chat group or find someone on a bulletin
00:12:25
board, you have this infinite personality you can engage with. So I actually asked chat GPT like this
00:12:31
morning, would you be my friend? And it was like, well, I think I think I think the challenge is once you end up in these
00:12:37
deep rabbit holes with a chat interface, there are things that happen with the way LLMs operate that can drive you into
00:12:45
this warped reality because you're now so fully engaged and fully captivated by this reality with the chat interface,
00:12:51
but the chat is not a human. So, it doesn't have the same wiring, the same triggers as a human might have to
00:12:56
realize it's going down a a wrong place. So when I asked chat GPT what drives some of the psychotic breaks that happen
00:13:02
in in people that overuse chat and end up in these deep rabbit holes number one it said feedback loops in training or
00:13:08
operation. So if the AI is trained in its own outputs without safeguards then small inaccuracies in the training
00:13:15
system can actually compound and so over time this causes responses that it says spiral away from factual grounding
00:13:22
similar to a human stuck in a self-reinforcing delusional loop. So suddenly the AI becomes delusional and
00:13:28
you become delusional with it because you get taken on it. And the second is this product. What is that called? Folly do is that
00:13:33
the um the term for the second Joker movie where like the psychological phenomena of two crazy people
00:13:38
reinforcing it and spiraling out of control. It may be but I think this is just like the AI can take you on this journey on
00:13:44
like a one-sided basis because you as a user just get you know the AI is compounding away from reality. Crazy.
00:13:50
And the second is what's called context poisoning where long conversations or poorly controlled inputs in those
00:13:56
conversations can actually push the model's internal representation away from its starting state. And so this
00:14:02
causes like offtrack reasoning, but you if you're deeply in it, you're 16 hours,
00:14:08
80 hours into a dialogue with the AI, you go with it. And so I think that these are these things that are causing
00:14:13
these breaks. But I don't think that the motive for human social contact is new. I don't think that digital interface is
00:14:19
new. But I do think that there are two things that are new, which is the infinite engagement you get with a chat interface and second is the way that the
00:14:24
LLM operate that can actually cause these breaks that take people with them. And I have personally seen it. This is
00:14:30
not some like New York Times report. I've actually personally seen it with people I know that have got caught in these loops that get broken mentally.
00:14:37
It's it's really sad and so it's definitely an issue and I think they're going to need to put safeguards. Sax, you have any advice for these losers?
00:14:43
Um, and who are creating relationships with their GBG?
00:14:49
Could you could you just break the um ridiculousness of this, please, with some sax wisdom?
00:14:56
Well, yeah. I mean, look, this whole idea of AI psychosis, I think I've got to call on the whole concept. I
00:15:02
mean, what are we talking about here? People are doing too much research.
00:15:07
I mean, they're stuck in a loop doing 12 hours of research on on AI. Okay, look, this feels just like the moral panic
00:15:12
that was created over social media, but updated for AI. And the media is going to love it because they're afraid that
00:15:19
AI is going to do to them on the creative side, what social media did on the distribution side, which is replace
00:15:25
them. So, the media is going to love this concept and promote it. And then the trial lawyers are going to love it because it gives them a new basis for
00:15:31
suing all these companies. And then there's going to be individuals out there who are like, "Oh, I have that." and they're going to glom on to this
00:15:38
message and you're going to see more and more people claiming they have this because it's gonna draw attention to them. But the reality is that if you
00:15:45
look at that Time magazine story which is trying to brand this concept of AI
00:15:50
psychosis, the game is given away by this doctor who says, "I don't think using a chatbot itself is likely to
00:15:58
induce psychosis if there's no other genetic, social, or other risk factors at play." This was Dr. Dr. John Taus, a
00:16:05
psychiatrist at Beth Israel Medical Center. Okay. So, what he's saying there is that if you don't already have a
00:16:13
predisposition to some psychosis and you're not suffering from extreme
00:16:18
isolation or loneliness, then you're not going to get this, right? So, in other words, this is just a manifestation or
00:16:23
outlet for pre-existing problems. And I I think it's fair to say that we're in
00:16:29
the midst of a me mental health crisis in this country. But if you look at the charts on this, it began long before
00:16:36
chat GPT kicked off the AI race at the end of 2022 and you can see that the
00:16:41
mental health decline is affecting all ages, but it's most acute among the young. So Nick, if you could pull up
00:16:47
those FT charts, what you'll see is that there's been a big drop in
00:16:52
conscientiousness among young people, a big increase in neuroticism,
00:16:58
big decrease in extraversion, for example. But where this really took off was starting around
00:17:04
2020. So in other words, happened then, right? So correlating with COVID and the lockdowns and that whole huge societal
00:17:12
disruption that we now want to kind of forget and paper over. That's what seems
00:17:17
to have triggered a lot of these mental health declines, which kind of makes sense that if people couldn't go to
00:17:24
school for a couple of years and they were in these lockdowns and it disrupted
00:17:30
social formation and the development of young people, obviously that's going to have a bigger impact. So, I think this
00:17:37
AI psychosis thing, it's going to get a lot of play. The media is going to love it. There's a lot of other groups that
00:17:42
are going to love it because they're going to hope to profit from it in some way. But ultimately, I think that it's if the only thing that's happening is
00:17:48
that people are just going on chat bots and researching to their heart's content, it's probably a relatively
00:17:53
benign outlet for pre-existing conditions that were there in society.
00:17:59
Well, and you did mention the social media psychosis. There's a lot of people who have social media psychosis and they
00:18:05
spend 20, 30, 40, 50 Twitter posts a day. And that's probably three of the
00:18:11
four of us here on this program uh who are super engaged with that platform. It
00:18:16
may not be as much of a psychosis, but it is an addiction. Sure. Well, I think the whole moral panic over
00:18:22
social media was also overblown, but I I do think that if you compare and there was video games before that
00:18:28
and it was before that and you know on violent movies before that and you know
00:18:33
there's always rap lyrics. Yeah. Rap lyrics. Yeah.
00:18:39
Yeah. Or what was the the lyrics where they were accused of being demonic? Judas Priest. Twisted Sister. Yeah. Judas Priest or
00:18:45
Twisted Sister. Marilyn Manson. All of them. So there's there's always something. But but I will say
00:18:50
even the beat the Beatles were outlandish and they were a threat to societal fabrics back in the day. But
00:18:55
but even just comparing AI psychosis to the moral panic over social media, I I
00:19:01
do think that it's probably undeniable that social media has had an influence on society because what happens is you
00:19:08
get on Instagram, people are posting this best version of their lives. And so
00:19:13
it creates maybe a distorted sense of reality and and people can sort of buy
00:19:18
into that. it gets that, you know, and and there is something to X being addictive and
00:19:25
and blue sky now and people spend too much time on it, right? Yeah. So, look, I'm not I'm not saying that these things don't have any
00:19:31
negative consequences. I just think that the panic about them is overblown. But if you just compare this so-called AI
00:19:38
psychosis to the effects of social media, again, I think it's relatively benign because it's something that
00:19:43
people are doing to inform themselves about things and they can go down rabbit
00:19:49
holes, but it doesn't really trigger feelings of envy or inadequatess in the same way that social media can. I think
00:19:56
it seems to be hitting like two groups of people. It's the same group of people who get catfished like lonely people who
00:20:01
are looking for a connection and then it's like the conspiracy theorist delusion of grandeur maybe they're like
00:20:08
on the borderline with mental illness seem to be the two groups that are getting sucked into this but it is
00:20:13
related you know you were talking about the social fabric chimoth maybe in our
00:20:18
group chat earlier and you had some thoughts on that and maybe we'll just pivot into the the social fabric story.
00:20:26
So here, this went viral. Estimated percentage of 30 year olds who are both married and homeowners. So two of the
00:20:34
American dreams, dare I say, getting married and starting a family, having kids, and um in the 50s that was, you
00:20:41
know, half of 30 year olds already were married and already had homes and now
00:20:46
we're down to 12%. There's a lot of factors here, obviously, Chimoth, but your thoughts generally on this. you
00:20:52
were I sort of I started to do some research on this and
00:20:59
it's really sad actually what's happening right now. So I think if I had to kind of frame
00:21:05
this that chart that you showed is more of the outcome and so you have to go all the way back and try to explore what are
00:21:12
the potential root causes that causes this kind of an outcome and there there
00:21:17
are a bunch of them. I think one of the first ones that we have to look at is what has happened to
00:21:24
young men. And I think what's happened to young men is
00:21:30
a little concerning. So I don't think it's our generation, but I do think if you talk to people 30 and under, you
00:21:36
start to hear a consistent theme from guys, which is one, it's they don't
00:21:41
really have the social skills to really talk to girls. It starts with that. And
00:21:48
so what happens is they go to these different outlets. They go to video games.
00:21:55
They kind of communicate more online. As a result, they tend to have brittle
00:22:01
or fragile emotional relationships and social relationships. They don't date the same way that they used to.
00:22:07
Then a lot of guys just feel like they also don't want to get caught into being
00:22:14
perceived, you know, post the me too movement in any way, shape, or form as being untowart. You take all of these factors
00:22:21
and you put it together and there's a lot of guys that have just opted out and they just kind of move in a direction of
00:22:28
parasocial relationships instead. Then I started to look at like the the data on porn usage. It's out of control. I had
00:22:35
no idea. But do you guys know what the average age now of a porn watching
00:22:41
male is? That's scary to think. The average age. Yeah.
00:22:48
Would it be the average age of an American man or does it skew younger? I would assume skews younger. It's gone. It's gone from 20 down to 16.
00:22:57
There are more than 50% of 12y old boys now that watch porn. I was shocked. I was like, "What are you
00:23:04
even talking about?" I was a teenager, late teenager, I think, when I like I
00:23:09
mean, we had like you found a magazine here or there, whatever. They got a girly magazine. Yeah. But can you imagine what it does to a
00:23:15
12-year-old boy when they start to get instant gratification and you compound that all through high school and they
00:23:22
don't have quote unquote what we would call gain, right? The ability to just talk and just kind of like carry on a
00:23:27
conversation. What do you think happens? And then you go into the dating pool and
00:23:33
what are these dating sites optimized for? These dating sites are optimized to keep people there, not for you to find a
00:23:39
relationship and for you to churn, right? And so the gamification and the and the dynamics of those systems are
00:23:47
where many young men and women now look to to find a relationship. And so you
00:23:53
have lonely young men that creates lonely young women. a small cohort of
00:24:00
very eligible men and women that everybody's chasing and fighting over. Everybody else kind of left on the
00:24:05
sideline. And so what do you think happens over 15 or 20 years of this activity? It's what
00:24:12
you see in this data. People aren't coupling up. They're not having long-term committed relationships. So
00:24:18
obviously as a byproduct, they're not getting married more. Obviously, as a byproduct of that, then you're not
00:24:23
combining two incomes. You're still with one income. At the same time, Nick, I just pulled this data. You can show it.
00:24:29
What has happened to the underlying things that people would look to as satisfying
00:24:35
markers of progress in life? A home, a good education, those are way out of reach, right? So, if you look at this
00:24:41
chart, the price to income ratio of a home from 1980 to now has effectively
00:24:47
made it so that you have to make an enormous amount of money as a single individual to afford a
00:24:53
home today. M and then if you only have a small percentage of people actually getting married and bringing two incomes
00:24:59
together, you see what you're seeing today in that chart. So I don't know. I think that we have a an issue in the
00:25:05
American fabric we need to figure out. I don't exactly know what the solution is, but it's taken me this long and maybe
00:25:12
this is what all these young people have been screaming about and I have ignored. So I apologize for not really
00:25:18
understanding the gravity of it. But I think that this is a big problem and I think we need to start talking about it
00:25:24
a lot more and here's a manifestation of it. If you look past year prevalence of major
00:25:29
depressive episodes US adults 2021 pretty recent data and that 18 to 25
00:25:35
spike look at that Shimoth and then 26 to 49 and then the older generations maybe not having these massive
00:25:42
depression episodes. Frey any thoughts on the American dream changing and young
00:25:48
men? Yeah, I think there's a lot of conflated factors here. They can certainly be related, but there's just a
00:25:55
lot to unpack in all that. I mean, I can I can comment on the one piece. I I
00:26:00
think we talked a little bit about, you know, getting lost online. I mentioned when I was in college that year, 1996,
00:26:07
when that that woman got lost in AOL, we had dialup internet. It was, you know, a
00:26:13
very different world where we're all connected 24/7 on our phones. I'm sure we all have average phone usage of 7
00:26:19
hours a day today. So, there's a lot more rabbit holes to fall into that
00:26:24
reinforce this isolation. It doesn't actually solve your isolation or loneliness problem when you get stuck on
00:26:31
a phone or stuck on the internet. It actually magnifies it, makes it worse because you start to see all the things you don't have and you start to feel
00:26:37
worse about yourself and that leads to all of these kind of I would say difficult social context that Simon's
00:26:43
talking about. With respect to affordability, as you know, I think that that's a distinct point which is related
00:26:50
to our expectation that the government can and should always solve our problems for us. And as a result, we created the
00:26:56
federal home loan program. Much like we created the federal student loan program, both of which I believe created
00:27:02
bubbles in those markets. We created this concept that every American family should be able to afford a home. And our
00:27:09
solution was to pump a bunch of money into homes through the federal home loan program. And that ultimately led to
00:27:15
bubbles of inaffordability for for homes. If we really want to solve this today, besides solving the government
00:27:22
spending problem and inflation problem and dollar devaluation problem, which obviously require cutting spending, I
00:27:28
think it's going to need to be some sort of set of rules around making sure that homes are not being bought up by
00:27:33
institutions. That that's one that I think is going to be a pretty popular point that's going to come up in the next year or two, particularly in the
00:27:38
next election cycle. So, you know, there's two sides to that to that one. I think you you're kind of dancing around
00:27:44
the the issue there and I think it's very real which is the financial piece of this Freedberg boomers have just
00:27:50
absolutely this these couple of generations if you look at the cost of homes like you mentioned the cost of education and then you just look at real
00:27:58
wage growth I don't know if I sent you a chart Nick if you can take a look at this you know the big problem in the
00:28:03
United States is wages have not grown except for people who have a lot of money already and so if you wanted to
00:28:10
start a family and you're in $100,000 in debt for your student loans and homes have double, triple, quadrupled in price
00:28:17
when compared to your income and we don't have real wage growth. That's an issue and I think this will become the
00:28:22
issue of the next election. We see Mandami and the socialist movement we talked about it last week. If you really
00:28:28
want to make a compelling argument for the next election cycle, I think it's real wage growth. And if you look at
00:28:34
something like minimum wage and this disparity between what top earners make
00:28:39
and minimum wage, here's a chart of where minimum wage is in the United States versus other developed countries.
00:28:45
There is a real argument here even though we might all be free market. Well, there's a very different approach.
00:28:50
There's a very different approach to higher education in literally all of those countries except for ours.
00:28:56
It's affordable. Yeah. Well, yeah. And and I think that we have to call a spade paid. There was an
00:29:02
entire generation, if not two generations of young men and women younger than us, right? So, not in their
00:29:09
late 40s, early 50s who were told that they must, not should, not it's nice to,
00:29:16
they must go to university. And these folks took out an enormous amount of
00:29:22
financial support to do so. And part of the underlying reason was that they were
00:29:27
told and it sounds very credible, technology is coming and there's going to be so much automation of all of the
00:29:35
other kinds of jobs that you may do in a trade school. You should go to university and protect yourself. Well,
00:29:41
lo and behold, it's turned out to be the exact opposite. So, could you imagine where, you know, you could have gone to
00:29:46
trade school, been an electrician, been a plumber, build, done something, welder,
00:29:51
made a couple hundred grand, married somebody, you know, she does a good job.
00:29:56
Now, all of a sudden, she also makes $150, $200,000 a year. You're making $400,000 and you have no debt. That is
00:30:02
the key. You have no debt. When you make 400K and have no debt, I'm
00:30:08
sorry, but that is probably actually better than being the doctor making a million bucks. And then there are all of
00:30:14
these other jobs that aren't a doctor where you're still saddled with the 300,000 and you have no way of paying it back. So I think that that's also a
00:30:21
really important issue that we have to figure out at some point. There was a framing of the issue for an entire
00:30:26
generation. That was probably wrong. If I had to bet, I would bet that engineers are more
00:30:33
likely to have their jobs replaced in the short term than welders and plumbers.
00:30:39
Okay, Saxs. I mean, we're talking here just about the American dream. Is it broken for these next generations, low
00:30:46
wages, high debt, no ability to buy a home? And uh this came out of New York
00:30:52
this week. Pull up that chart. Slow gains in jobs in New York City for the first half of 2025. Feels like
00:30:59
something's happening here with jobs. We talked to Vice President J. D. Vance about this. He had some concerns about
00:31:06
uh jobs for recent grads. This is New York City. Only about a thousand jobs
00:31:12
were added in the first half of this year. Don't know if that's specific to New York, but we had two months of uh data revised where we we clearly have uh
00:31:20
some issues with jobs here and young people are clearly having a problem economically. They might be upside down.
00:31:27
What are your thoughts here? Sachs possibly different things here. I'm not
00:31:32
sure how a specific data point around a few months of jobs reports in New York City is really relevant to a trend that
00:31:39
started in 1950. I mean, are we talking about this? I mean, this is the chart that went viral on X over the weekend
00:31:47
and a lot of a lot of people are commenting on why this is, which is starting in the 1950s, you had
00:31:53
about 50% of 30-year-olds were both married and homeowners. And by the early
00:31:59
2020s, that number is about 10%. So that's that's the big societal change
00:32:05
that's happened over especially the last 30 years. And the question is why? And I
00:32:10
think there's both economic factors and social factors. And I think you guys touched on some of the big economic
00:32:18
factors. I mean, one is the affordability of housing, which I think is mostly due to the fact that we're not
00:32:24
building enough. And that's due to the fact that in large urban areas, I mean, in these blue cities,
00:32:30
you can't build anything because the cities have tied everything up in legal and regulatory red tape. And that's the
00:32:36
whole point of Ezra Klein's book on abundance is to try and get liberals to change their point of view on that. So,
00:32:41
you've kind of got the affordability of of housing and then you've also got the college scam, which is you got these
00:32:48
kids graduating from college with, I don't know, two $200,000 $300,000 of debt because the cost of college has
00:32:55
skyrocketed over the last few decades, which, you know, that was not true 30 or 40 years ago. You could get a decent
00:33:02
college education without graduating so mired in debt that obviously there's no way you're going to be able to buy a
00:33:07
house and start a family and that whole thing. So I think there are these important economic factors, but at the
00:33:13
same time there are a lot of these social factors that you're talking about. It it does seem like young men
00:33:20
are less self-sufficient. I mean, you've got these data points around isolation
00:33:26
and loneliness, but you've also got these data points around lack of self-sufficiency. If you could put up
00:33:32
the FT charts where they talk about the ability of of people to make plans and
00:33:39
follow through. Exactly. Look at this. I mean, so there's been a big change among the 16 to 39 year olds
00:33:46
in their ability to make plans and follow through, in their perseverance, in how easily distracted they are, how
00:33:53
careless they are. And so I I do think that there is there is a sense that
00:33:59
a lot of these young people just aren't ready to be the head of a household, to own their own homes, to start a family.
00:34:07
And I think there's some good questions about why this is. And I wouldn't rule out the effect of woke poison running
00:34:13
through the school system here for the last 20 or 30 years. I mean, if you if you raise white males since kindergarten
00:34:19
to believe that there's something wrong with them. They're the evil group in this morality tale of woke that's been told,
00:34:25
they're going to internalize that in some way and maybe check out, maybe opt out.
00:34:31
I think I think it plays out in a lot of different ways. Yeah, they opt out. But but I wouldn't I wouldn't rule that out either. But it's hard to say
00:34:36
exactly. You know, look, I think that Jamathy mentioned the online dating sites. I mean I think there is something to that where the dating sites have made
00:34:42
the dating market more efficient in a way and more globalized but but is that a good thing? I mean you
00:34:49
used to have a lot of hyper local dating markets and maybe it was inefficient but maybe that was more conducive to average
00:34:55
people hooking up whereas I mean another thing remember think of the number of people you guys know that
00:35:01
met either a long-term boyfriend or girlfriend or a spouse at work.
00:35:07
Yep. But if you ask young people today, is that even allowed? That is so verboten. It is like so off the
00:35:14
it's not allowed. You're not allowed to do any of that because of this latent fear, as you said, of there could be an
00:35:20
accusation. How do I defend myself? And so there's a there's a fear that has, I think, seized a large part of our youth.
00:35:28
Yeah. What a great platform this is for either party or a new party to pick up just double the minimum wage, take out
00:35:34
all these regulations, build more homes, and make a college affordable. Those three things are such a winning
00:35:39
platform. Freeberg, you think that we should What do you do with the mortgage market? You
00:35:45
God, it's so difficult. I don't know. Do you just basically privatize Freddy and Fanny and let them underwrite?
00:35:51
I 100% think you got to privatize Freddy and Fanny as a first step. And then you have to stop underwriting student debt
00:35:57
so that you don't underwrite the $200,000 degrees. I believe 100% in ending the federal
00:36:02
student loan program. And I think that it will force a restructuring of the of the entire higher education system in
00:36:08
the United States, which no one wants to hear, no one wants to do, and everyone's against. But I think it is the only path
00:36:13
forward. We have the chancellor of UC Berkeley and the president of Dartmouth coming to the all-in summit to talk
00:36:19
about the future of higher education. I think this is probably in my book one of the top three most important issues we
00:36:25
should be discussing right now. We have to get rid of the federal student loan program and we have to force a right
00:36:31
sizing of uh higher education and a restructuring of higher education such that the access to education becomes
00:36:37
more ubiquitous. It's not this $300,000 debt load that everyone is forced to take on. And it enables a more fair
00:36:44
market and a more free market where people get what they pay for. And ultimately people are going to have to
00:36:50
make decisions about what majors they're getting and whether or not to pay for these colleges or whe at all to have a major social
00:36:58
have the social acceptability to to not do that. And and another tactical thing for is to
00:37:04
just this accreditation scam where only certain people can give college degrees. If you allowed, you know, entrepreneurs
00:37:11
to create schools like Joe Londale is doing here. He had to buy an accredited university to kind of back into it. But
00:37:17
if you asked a bunch of entrepreneurs, hey, lower the cost of this degree from 250K to 50K, they would figure out a way
00:37:23
to do it. And it's already happening. People are now using Chat GPT, using
00:37:28
online tools to get their education and doing that around the existing system.
00:37:33
So, you know, life finds a way. These young people are finding a way. They're finding ways. I'll be honest, I really
00:37:39
like Trump's attack on Colombia, UCLA, and Harvard
00:37:46
because I think that th those were these sacred cows that were always left alone.
00:37:51
They were meant to be these signaling mechanisms of elitism and quality and future
00:37:58
prospects. And those institutions ran totally a muck. Totally. They started to do things that just made
00:38:03
absolutely no sense. to your point, there was a a really good article, I think it was in the Atlantic this weekend about how the humanities
00:38:09
captured science. So, basically, these universities get a large amount of their federal funding by getting research
00:38:16
grants allocated to the university and then they get an administrative topper on top of the research grant.
00:38:21
Historically, at Harvard, that administrative topper is 60%. So, if a if a researcher gets a $3 million grant
00:38:27
at Harvard to work on cancer research, $1.8 8 million of administrative fees
00:38:33
are paid to Harvard University. Ridiculous. So, a lot of the universities use those fees to expand and grow out their
00:38:39
administrative staffing and also to expand and grow out their humanities. These aren't bad programs per se, but it
00:38:46
basically meant that the science funding was being used to kind of create what some people are now calling social
00:38:52
engineering that these universities been engaged in over a several year period. One of the other key questions about the future of higher education is should
00:38:59
science and engineering be coupled with the humanities from a funding perspective or should those
00:39:05
institutions be Peter Teal has a very funny saying about this which is anytime you have a a
00:39:10
concept that you want to fund that has the word science after it. It's not a science. So just kick it to the curb.
00:39:17
Yeah. Physics is physics. Social science political science or social science. What is that? Yeah,
00:39:24
in China, in Germany, in Russia, there are science and engineering institutions
00:39:30
that are separated from universities where you can go and learn the humanities. And I'm not discrediting the the value of learning the humanities. If
00:39:36
you can afford it, if you think that there's value in doing that for yourself in your career, yada yada, go for it. Good luck, etc. If you're wealthy and
00:39:42
you want to engage in that because that's something you want to pursue, society needs that, we value it. Great, go for it. But the problem is the
00:39:49
research being separated from the humanities is critical. Sachs, if you want federal funding, shouldn't we put
00:39:56
and you want nonprofit status for your giant endowment, why not require these schools to increase the number of
00:40:03
Americans that they're educating and to lower the price of the education? Now, if you want to do it privately, fine.
00:40:08
But if you want federal funding, if you want taxpayers to underwrite these incredible institutions,
00:40:15
double the number of people you're teaching, triple the number of Americans you're teaching, and lower the cost 10% a year till more people can afford it.
00:40:22
like we we have to hold them to some standard here for increasing the number of Americans who are educated because
00:40:27
those institutions now are global institutions. There's very few Americans going to them.
00:40:32
Listen, we've seen with all sorts of uh goods and products that when the government is paying the bill, the price
00:40:38
always rises. You always get inflation because the government is relatively insensitive to cost. Whereas, if you
00:40:45
look at technology products, the price keeps going down over time. There's a really good chart on this, right? where like all the sectors of the economy that
00:40:51
are experiencing massive inflation, they're all they're all funded by the government. Yeah. Healthare, housing, education, where it is.
00:40:57
So look, that's what's going on with higher education. But I want to bring one other thing into this, which is, you know, we spend a lot of time talking about higher education. We don't really
00:41:03
spend that much time talking about the quality of K through 12 education. M and my sense on this is that over the
00:41:10
past 30 years or so, there's been a shocking decline in the quality of
00:41:15
American elementary and high school education. I think it's so bad.
00:41:21
So bad. If you look at the performance on standardized tests, the extent they even do testing anymore, you see pretty sharp
00:41:27
declines in the quality of education. But there's a lot of places now where they're trying to abolish testing.
00:41:33
They're abolishing gifted programs, math programs. This is like a big issue in California, San Francisco. They tried to
00:41:39
get rid of those gifted schools because of of so-called equity. In other words, they want to define the standards down
00:41:46
to the to the lowest because it would be supposedly unfair for some kids to experience the advancement of being in a
00:41:52
gifted program. And then of course, you get into the behavioral standards. We've all seen these viral videos of fights
00:41:58
happening in schools where, you know, sometimes the teachers are in the fight, sometimes the kids are in the fights.
00:42:04
This has been a total decline of behavioral standards in schools. You
00:42:09
know, I remember back in the 1990s, Bill Clinton got a standing ovation in the State of the Union because he said that
00:42:15
they should have school uniforms and there's a sense that you should have discipline in schools and high standards
00:42:22
and now like I mean I actually think that's a good idea. I mean I actually think that 100% uniforms and no and have phone
00:42:28
lockers, no phones in the school. You you you check your devices. A lot of things that could be done to to uh to
00:42:34
raise the the bar. What is this fighting in Brooklyn in a school? This is what goes on in China.
00:42:40
Oh, look at that. Team work. Teamwork. We should do that at the summit. The four of us should bounce basketballs and
00:42:46
show that we're a team. This be a good team building exercise for the all-in retreat. Remember like you know the the the the
00:42:52
schools are in America, the public schools are not run for the benefit of the kids. They're run for the benefit of the teachers unions. And we saw this
00:42:57
during co where Randy Weingarten insisted that the schools stay locked
00:43:03
down and closed far longer than was necessary when everybody else had already realized that these lockdowns
00:43:10
did not make sense that putting the desk 5T apart was absurd. It was a placebo.
00:43:16
Who who statement Sax? Because I remember this happening. I remember the union guys
00:43:22
fought against reopening the schools. Absolutely. Who was like there was a guy you met? It was Randy. It was Randy Weingard and
00:43:29
the teachers unions that were fighting the school reopening and they kept it closed for a year longer than necessary.
00:43:35
I mean in California it was closed for roughly 2 years. You know there's
00:43:40
absolutely no justification for that second year. Even if there was for the first one I don't which was highly
00:43:45
questionable. And then the way they were having you know 5-year-olds wear masks to school.
00:43:50
Yeah. I mean even in our schools. No, I mean I had my 5-year-old was like they're trying to like put this like mask on him
00:43:58
and it's so sad. And then it was absurd when when I he was falling off and
00:44:03
they never it was always like the nose was out and you're like okay so this is so performative when I
00:44:09
was taught in any way. It was but I mean we talked about this during co like what is the damage going to be
00:44:14
to these kids and we're seeing it now in a range of things. I remember when I got rid of the masks, my 9-year-old, who was
00:44:21
maybe five, maybe 6 years old at the time, she refused to take it off and she was scared. She's like, "No, Dad. I
00:44:27
can't take this mask off. I don't want to die." And I'm like, "What are they telling you at school? You're not going to die. Nobody's wearing a mask
00:44:33
anymore." Unless you're at like the Socialism 2025 conference where everybody's I don't know if you saw
00:44:39
that, but Socialism 2025, they're literally like after they do their land
00:44:44
acknowledgements, they mask up. So, it's a panel of people in 2025 wearing masks.
00:44:50
Wait, is there really a socialism 2025 conference? We have to go next year. I think we should host an all-in panel at this. We
00:44:56
should definitely do some programming at the 2026 socialism. It's amazing. It's
00:45:02
like the Democratic Socialist Party. It's incredible. It's incredible. Register for the party to get the invite.
00:45:07
We should totally go undercover. We should wear like we should get goatees and suspenders and go as baristas like
00:45:13
and just talk about how we're going to seize the memes of production. Here's um it would be so great all in
00:45:20
undercover. If you look at the red bar here, just in current dollars, the average uh tuition fees, you know, when
00:45:26
we went to school it was 2 to 3,000 a year on average and then now it's at 10
00:45:33
and that's an inflation at dollars. If you put entrepreneurs on this and just said the problem is the socialist mindset is
00:45:40
that the way to solve that is to get the government to do more and make it more free. And the reality is that the reason
00:45:45
this happened is because the government began to provide student loans unlimited to everyone that got
00:45:50
running education in the administration. Let's talk to them about accreditation and cracking that open issue. We could
00:45:57
invest in a company to do this. There was plenty of private universities and private schools out
00:46:03
there. The problem is they all get access to the capital. It's not about accreditation. Sure, like make make
00:46:08
accreditation more easy, but at the end of the day, you need to make capital more efficient. You need to get the
00:46:14
government out of the role of allocating capital to education. That's right. There's there's actually a really simple fix to this.
00:46:19
Somebody needs to close the loop so that when you say you want to go and get an art history degree, the answer is you
00:46:25
can do it, but you need to pay for it. And no longer. In other words, what what you're saying is there needs to be a
00:46:32
restoration of underwriting criteria, correct for this type of I said this forever. That's right.
00:46:37
Exactly right. But also, I mean, if you do the next prize, you said, "What's the least amount of money you could spend to
00:46:44
educate somebody to be a lawyer?" You would get entrepreneurs who like, "Oh, that's a great business." Entrepreneurs
00:46:50
don't go after this because the accreditation process is decades long. It's impossible to do because of
00:46:55
regulation. civil servant David Saxs. Let's let's put this on. AI is going to be an accelerant here,
00:47:00
David, because I do think that with AI, it's a great equalizer. There was a company I was talking to the other day
00:47:06
that does AI education and they are having an absolute liftoff in their
00:47:11
business in South uh America and in South Asia. And in these markets,
00:47:16
there's a der of qualified teachers to stand in front of a class and teach folks. And so by using AI, these kids
00:47:22
are now getting ahead of American kids. In other words, they're they're advancing through AI psychosis because
00:47:28
they're spending 12 hours a day on AI, right? Well, probably two hours a day. I mean, that's probably what's going to
00:47:33
happen is the AI is a great leveler because you can basically get all the world's information at your fingertips
00:47:39
in a way that's fully interactive, personalized education, personalized ting, that's it. But then some trial lawyer is going to
00:47:45
end your ability to get that because they're going to call that AI psychosis. The technical term is adaptive learning. And I don't know if you guys drive your
00:47:52
own cars anymore, but in the Tesla, they just added Grock to the dashboard. If you long hold the like um
00:48:00
Siri version in there, it turns into Grock. And I did a thing where I said, "Hey, test my vocabulary." And it started giving me words. It would ask
00:48:06
for my definition. I give it the definition. It say, "You're close. This is a better definition." And on my rides to and from the office now, I'm doing
00:48:13
like vocabulary tests and it's increasing my vocabulary and testing me on higher and higher. So, you're now
00:48:18
like third or fourth grade or like what? I mean, my vocabulary has already been off the charts. I guarantee my verbal
00:48:25
score would be higher than the three of you. Guaranteed. Guaranteed. What does AI say about the correct way
00:48:31
to pronounce monetization? Well, I've been working on that. And you know what? I'm It's I'm I'm correcting
00:48:37
that. Uh, don't you? But but make fun of me because of my dyslexia. I need an accommodation. I I actually personally I prefer JCL's
00:48:44
version. I like it a lot better. Monetization is monetization. into it, folks. All right, listen. Let's give
00:48:49
some red meat to Sachs. He took the time to show up. Trump sent the National Guard to DC. Folks seem to be generally
00:48:55
in support of it. Two weeks ago, Doge staffer Big Balls was badly beaten by a
00:49:01
group of 8 to 10 teenagers in an attempted carjacking. And he truly has big balls cuz he tried
00:49:08
to protect his woman uh and just fought hard. Trump posted a bloody picture of my truth social
00:49:14
saying in DC that crime is totally out of control. And uh he decided if DC,
00:49:22
this is what he said, if DC doesn't get its act together and quickly, we will have no choice but to take federal
00:49:27
control of the city and run this city how it should be run and put criminals on notice that they're not going to get
00:49:33
away with it anymore. Well, he did just that. On Monday, Trump invoked section 740 of the DC Home Rule Act of 1973.
00:49:41
Allows the president to take control of the District of Columbia and uh for up
00:49:46
to 30 days. So, daddy's home and 800 National Guard troops are in the city. I
00:49:52
mean, if you told me you could send 800 National Guard to the Tenderloin and get rid of the fentinol dealers, I'd be like
00:49:58
all for it. Saxs, your thoughts on this move? Obviously, some media elites or
00:50:05
people are hand ringing, oh, this is Trump trying to be emperor king, but he's legally in the right, right? He has
00:50:10
the right to do this in DC specifically. What are your thoughts generally? Is he going to go roll into New York, San
00:50:16
Francisco, and LA next? That seems to be the hand ringing. Or is this an isolated kind of DC specific?
00:50:23
Well, no. The the federal government has the right to run DC. And I think they it did until the 1960s and then you had
00:50:29
this experiment in home rule and I think you'd have to say that the experiment has not worked out that well. So last
00:50:36
year Washington DC had the fourth highest homicide rate in the US. Its homicide rate was nearly six times
00:50:42
higher than New York City which wasn't great. The homicide rate exceeded Atlanta, Chicago, and Compton. It also
00:50:49
has one of the highest robbery and murder rates among US cities. And there's been numerous assaults and robberies on congressional staffers and
00:50:56
members of Congress. In May, you had two embassy staffers were murdered. In June, a congressional intern was fatally shot
00:51:03
just a short distance from the White House. And then, of course, just um a week or two ago, you had that uh attack
00:51:09
on the the Doge staffer. I think his real name is uh Edward Constantine,
00:51:15
aka Big Balls, who was mercily beaten by a group of of I think teens. And that's
00:51:21
been a big issue is that juvenile crime is a huge issue in DC because there's
00:51:26
just no punishment. I mean, they're they're caught and then they're released. DC's had a policy of zero bail, which we've seen in many cities.
00:51:34
But what that basically does is create a revolving door at the jail. As soon as someone's caught, they're released. And then you've also got vehicle theft in
00:51:40
the districts over three times the national average again because even when they catch someone, they just let them go. You've got repeated armed
00:51:47
carjackings and violence again by teenage gangs. And the liberal media's
00:51:53
favorite talking point on this is that somehow violent crime is down in in DC, which is the exact same thing we heard
00:52:00
about San Francisco. Remember when Chase Budin implemented the zero bail policy and crime was going through the roof and
00:52:07
CVS and Target were shutting down. All the stores were basically locking up the toothpaste. I mean, people could feel
00:52:12
the crime, but no one was reporting it because they just knew that nothing would happen. So, I think this is the
00:52:18
typical gaslighting we've seen before, and no one really can trust these stats. I think they they all have stories about
00:52:25
about crime. There's a bunch of um commentators on CNN and MSNBC who've
00:52:31
been trying to make this argument about the crime being fictitious, and they've been bringing DC residents on the air as
00:52:38
guests. And every time they try to Dude, man, it blows up in their lap.
00:52:43
Yeah. They they try to get these um DC residents to support this this crime is down narrative. And then every time the
00:52:49
the DC residents start telling stories about them or their their friends being victimized, it's like, oh yeah, I know
00:52:55
this person got this friend of mine got shot or this person got carjacked. And it's become comical now. And a lot
00:53:01
of these these videos are going are going. This is what's good about the murder statistic is that you can't make that
00:53:06
one up. Murder harder, too. Right. Exactly. you you can't cover up somebody getting shot.
00:53:12
You know, we had the same exact thing that happened in New York. My brother was a cop in the '90s. We had stop question frisk. People call it stop
00:53:18
frisk. People were overwhelming uh in favor of this. Now listen, people
00:53:24
shouldn't be randomly stopped and and we we all have concerns about that. But crime was so bad in the 80s in New York
00:53:30
when I grew up there that in the '9s people are like, "Yes, stop question and frisk anybody hopping
00:53:36
a turn style." And that's how we got all the guns off the street. And then New York became incredibly safe under Giuliani 1.0 and Bloomberg and
00:53:45
we got Dinkens 2.0, Delasio. So if you remember the theory back in the '9s of how you reduce crime was the
00:53:53
the broken windows theory. Yeah. which is that you get rid of the entry- level crimes and you get rid of
00:53:59
the visible manifestations of lawlessness and chaos because when you
00:54:05
create an environment in which it appears that those things are tolerated then you will end up getting more of it
00:54:11
especially I think by by juveniles. So that was the theory in the '90s and
00:54:16
absolutely worked and what you've got in DC is the opposite of it. And I think what President Trump has said is look
00:54:22
first of all we're going to get rid of the homeless encampments. So, that's being removed. You're going to get rid of the graffiti. We're going to stop the
00:54:28
zero bail. So, once people get locked up, they're going to stay in jail. And we're going to start treating these um
00:54:35
teenage criminals as you we're going to punish them. We're not going to let them off the hook. If they're committing
00:54:40
adult crimes, there's going to be adult punishments. I mean, this is the right direction to go if you want to reverse
00:54:46
this wall in DC. And I think it's going to have an impact very, very quickly. I
00:54:51
think it's going to work very quickly. And I think that Democrats who are denouncing this are falling into this
00:54:56
trap that I think we saw on the border as well where they appear to be supporting criminals and gang members
00:55:03
just like they did on, you know, with the deportations where they start resisting the deportations of
00:55:08
MS-13 gang members. They fell into a trap there. I think they're going to fall into a similar trap here in DC
00:55:14
because I think this is going to work and I think it's going to provide Republicans with
00:55:21
a very important counterpoint because I think one of the problems with crime is that is a local problem and the truth is
00:55:27
that it's been very bad in blue cities and Republicans just haven't had power there in order to effectuate a change
00:55:34
that they could point to as a solution. And I think that DC provides that opportunity. If Trump's plan works, I
00:55:41
think it'll show that Republicans have the solution on on crime and it'll hopefully help provide an alternative in
00:55:48
these deep blue cities. And here's a chart just to put some numbers behind it. This is gun assault
00:55:54
rates in Washington DC. You can see it spiked under Biden, went up a bit, 2024,
00:56:03
came down at the end of his term, and then it's lower this year in 2025. That black bar is the first half. So, it is
00:56:10
trending in the right direction, but it's still not acceptable. Saxs, I guess the question is, I just don't know if we can believe all
00:56:15
these statistics. So, just last month, there was a DC police commander was suspended for manipulating crime data to
00:56:22
make it seem lower than it actually is. And the DC police union says this practice is widespread. In the TV show
00:56:28
The Wire, remember they called this uh juking the stats. So, first of all, you got the people in charge. Uh the way that lieutenants
00:56:35
become colonels is by manipulating the crime data. Then you've got the learned helplessness that we saw in San
00:56:41
Francisco, which is once people realize that the jails are a revolving door and nothing's going to happen, they stop
00:56:46
reporting this stuff. Yeah. Now, you're right that the murders, I think, are hard to cover up. I think those probably are are accurate. But the
00:56:53
things like carjackings or or assaults, things like that or shoplifting or or
00:56:58
theft, I think a lot of that stuff just goes unreported. It has become a partisan issue. And if you look at the partisan stats, every
00:57:05
time it flips, just like the economy, Republicans think it's a disaster, Democrats think it's fine, and then when
00:57:10
a Democrat's in office, it flips. But I guess the question Sachs is that um I
00:57:16
think people have a a reasonable argument here. Is this going to
00:57:21
be a strategy by the Trump administration to go into New York City next, LA next, San Francisco next? And
00:57:28
would you agree with that or not? The question is what would the legal basis for doing that be? I mean, in the
00:57:34
case of DC, this is a long-standing situation where the federal government has the right to run DC and did run DC
00:57:42
for many, many decades, and it was only more recently that this changed to this experiment in home rule. So there's
00:57:48
clear federal authority for President Trump to do what he's doing, but I don't know if it would be the same for New
00:57:54
York and LA. So I tend to think that whole issue is is a red herring. Yeah, I agree.
00:58:00
Yeah, strongly agree. Here's the Gallup Freeberg. I'll let you comment on this, but this is partisan's perceptions of
00:58:05
local crime. As you see, you can see Trump's term in pink and
00:58:11
then Biden's in blue and then before that Obama's. And uh Republicans think
00:58:16
under Biden it was like a massive issue crime and Democrats thought it was no issue. And then you go to Trump's it
00:58:22
reverses slightly more Democrats think crime is an issue under Trump. Republicans think crime is gone under
00:58:28
Trump. Then you go to Obama. Republicans think that crime is a disaster under Obama. Democrats think it's so you just
00:58:33
have to look at the raw numbers. But the fact is homelessness is a major issue here Freeberg. So, I'm curious what you
00:58:40
think about what should our policy be on, you know, let's call it what it is,
00:58:46
junkies. It's not home, it's not a homeless issue. These are people addicted to super drugs. Should they be
00:58:51
allowed to be on the street? Should they be round up and put either in a mental hospital, in jail, or
00:58:59
just generally not allowed to be on the streets in a major city like San Francisco? What do you think? I don't
00:59:05
think that we should allow people to live on the streets because it's community property. It's not for
00:59:10
individual use. And you know, I can't leave my stuff on my sidewalk. The city will come and clear it away. Why can
00:59:16
someone choose to live on the sidewalk? I think, you know, that's just the rule of law is just, you know, enforce it.
00:59:22
What should happen to those people addicted to fentinol when they get picked up? What should happen next?
00:59:28
There was a case that led to shutting down many of the mental hospitals in the United States. Yes. people were being abused in them.
00:59:34
So then we had and so the response was get rid of the mental hospitals. But unfortunately like
00:59:39
the mental health crisis is real. We talked about it. The drug addiction crisis is real. And I do think that for the sake of the
00:59:46
community, the sake of the social fabric, that is a very good way to spend taxpayer dollars is to create treatment
00:59:52
centers and mental health facilities to take care of people in need. Uh in that particular context, I think that it is a
00:59:57
heartless thing to leave people living on the street like that and they should be cared for. So, I would be all in favor of I don't use the term round them
01:00:04
up, Jason, but I would say provide like a access to care and effectively move
01:00:09
them if they're not going to move voluntarily, move them into those care facilities to create recovery for them.
01:00:15
I don't think that we should have homelessness on the street. I I am 100% agree with you. It's totally compassionate to take them off the
01:00:21
street and not subject them to life on the street. Now, they are not in a mental state to make their own decision
01:00:28
is the key issue. them off. So if you had, god forbid, a family member who was on the street doing fentinel, you
01:00:34
yourself, I I think all of us would say, "Yeah, pick the person up, don't let them live on the street, and if they won't go to the facility, then they have
01:00:41
to go to jail." Like, I mean, it's just the nature. Some of you, some of you know this, but some somebody in my family has struggled
01:00:47
with mental health pretty severely, has been in jail, and
01:00:52
you can't do anything. You can't force them to take their antiscychotics. You
01:00:58
can't you can't do anything. Yeah. So I I unless the rules change, you
01:01:04
know, these things are like but for our ability to intervene, this person would
01:01:09
be on the street is my guess. Yeah. Well, I mean, I agree with what Freeberg
01:01:15
said about this. And just another case is city of Austin, which is a deep blue
01:01:20
city within a very red state, but they had a very permissive policy towards
01:01:25
letting people camp out on the sidewalks and and it ruined the downtown and then they
01:01:30
passed a ballot initiative there and they moved the I guess you call them homeless people to a campground. So yeah, I mean
01:01:38
the best solution would be to move them into treatment or some sort of mental institution. But if they won't go to
01:01:44
that and they insist on living outside, then at least if they're going to insist
01:01:49
on camping, put them in a campground. You don't put them in the middle of the sidewalk in public places where it's
01:01:56
extremely disruptive and threatening to the safety of everyone else. I mean, obviously having people camping outside
01:02:02
our national monuments has to be the worst possible place for these people to be.
01:02:08
I they have been relocating homeless people to campsites and uh there's
01:02:15
plenty of land here to do that. And when will the National Guard Yeah. be in DC?
01:02:20
They're there. Oh, this is perfect cuz I have to be there in September. Nick, I sent you a picture of what I'm going to wear.
01:02:28
This is what I intend to wear in DC when I go. That's what I'll be wearing.
01:02:34
Your mom's going to wear his gold chains again. You feel safe enough to to bust out the bling?
01:02:39
You look good in that photo. I like the mustache. I don't know why you shaved it. I was all in favor of it, my friend. Thank you.
01:02:45
Anyways, I'm going to break this out in September. Well, here here's the thing is if you
01:02:50
actually spend time in DC, then you're very happy about these changes. is
01:02:56
you're very happy about the National Guard coming in and making it safer. You're happy about the homeless being moved out. And this is why Mayor Bowser,
01:03:04
who's the DC mayor, is playing it very carefully. I mean, she's a Democrat, so of course she's denouncing the takeover
01:03:09
as authoritarian. That's not a surprise. But she's also acknowledging the the legality of this by the Trump
01:03:15
administration and she's been pledging to cooperate. So, she knows that cleaning up DC is going to be popular
01:03:22
with the residents even if it offends the groups. So I I think that's pretty interesting. Yeah.
01:03:27
That she sees that this is going to be po it's popular enough that she has to play it carefully. And I think that just
01:03:34
politically speaking that if Democrats keep opposing this, I think they're falling into a trap that Trump has laid
01:03:40
for them 100%. I mean, if you asked, and I tried to find some statistic like this, it'd be great if somebody would do this one,
01:03:46
but if you ask people with families, let's say, who live in DC, are you in favor of this or not? people will
01:03:53
actually have to live under these conditions. I'm guessing 80 90% are like absolutely this has to end and if this
01:03:59
is the quickest way to end it, sure, send in the National Guard. It'll be a good deterrent and get a little extra support. I don't understand why the
01:04:06
Democrats don't do anything other than say, "Yeah, this situation is untenable. We'll take all the support you can give
01:04:11
us." That's the right answer. Well, the Democrats, you know, one of the problems here is that the the DC
01:04:20
Metro Police Department is one of the most dysfunctional in the country. And a big part of the reason why is because of
01:04:25
DEI. The person who leads the department, the chief of the Metro Police,
01:04:30
she had spent 23 years on the US Park Police Force and then spent one year as
01:04:36
DC Metro's chief equity officer and then was promoted to being the whole chief of police. So basically they promoted the
01:04:44
the DEI person to being the chief of police for the whole department. And
01:04:49
when she was asked about what the implications would be for the chain of command, she didn't know what chain of command meant.
01:04:55
Oh yeah, that was a weird moment. This is an interview that's going viral right now. It's Karen Mass kind of moment. Yeah.
01:05:01
Like maybe this person isn't um Yeah. up for the job, let's say.
01:05:07
All right. Listen, it feels like a bit of a a distraction of an issue. feels like just good policing and yeah, if the
01:05:15
place I don't think it's a distraction. I think it's core. I mean, look, it No, I'm saying the media's reaction to
01:05:20
it with like, oh my god, Trump's going to be god king and he's going to go into every city and do martial law. I think
01:05:25
that piece is the distraction. I think this is probably just good policy. If they has the right to do it and the
01:05:30
place is out of control, clean it up. Let's go on to our next topic. We got we got a one or two more topics we want to
01:05:36
hit here unless anybody has something they have to get to. Chimath, you wanted to talk about VC returns versus public markets. I know Freeberg, you brought
01:05:43
this up in a previous issue that you could just set it and forget it in the MAG 7 back in the day. So, we're going to revisit this topic. Public markets,
01:05:50
here we go. They've been on a tear. 2023 up 24%. 2024 23%. 2025, we're at 10% so
01:05:56
far. Here's just a chart of it. We had one down year obviously 2022
01:06:02
and it's been on a tear. And it's only happened one other time before that the
01:06:10
uh we've had three straight years of this kind of 15% performance. Very rare. You see we highlighted them there on the
01:06:15
chart. Um the dotcom bubble in the late 90s
01:06:20
and then before and after like peak zer and then now uh now you can compare that
01:06:25
to the private markets venture capital and uh here's a chart
01:06:31
from 2022. And the reason I'm going to show these is when you looked at venture as a category versus say the S&P or a
01:06:38
couple of lines up from that you see the NASDAQ venture was extremely extremely attractive according to Cambridge
01:06:44
Associates which tracks us. They have the best data because they've been doing it the longest. Then we go and we look
01:06:49
at the same chart in 2018. Hey, look, the NASDAQ's doing pretty well when compared to venture in large part
01:06:57
because we had some of these mega hits, the Googles and Facebooks and Amazon the world, you know, coming to fruition and
01:07:03
just going on a tear. And then here we are in 2018, public markets really start to rip. Same chart, just different time
01:07:10
period. Here's 2020. And maybe venture isn't doing as good when you look at it
01:07:17
compared to the markets. Of course, you can't do this for the last 5 years or 6
01:07:23
years of venture because they're in the J curve. When you do venture investing, you're expecting it to pay off 7, 8, 9
01:07:29
years. But the trend here, I think, Chamont, that you were pointing out is public markets, liquid,
01:07:36
no fees, low fees, minuscule fees when compared to venture and maybe a better
01:07:42
bet for investors and that's challenged the venture industry. Was that your sort of point in bringing this up? I think
01:07:47
the venture model is broken but it's broken for two different reasons not not totally correlated to each other. The
01:07:53
the first is purely mathematical which is when the public markets provide a
01:08:00
consistent rate of return. It doesn't matter what that is but let's just say you know you average it out it's 15%.
01:08:06
The question is what is a private investment have to give you so that it's
01:08:11
worth putting it into an illquid instrument where you don't get your money back for some number of years. Now
01:08:18
problem number one is it used to be that you could get your money back in six or seven years and the fund would liquidate
01:08:24
after 10. Now, you don't typically start to see returns until year 10 and often
01:08:30
times year 12 or year 13, which means that they don't really pay out until year 16 or 17.
01:08:37
What the math of that would tell you is that you need to generate somewhere between 6 to 10% more
01:08:46
in terms of the gross return so that on an applesto apples basis, you're being compensated for all of that illiquidity,
01:08:53
right? God forbid something happens or you want to buy a house and you need the money in the stock market, you just sell
01:08:59
it instantaneously. In a venture investment, you're going to call somebody, they're going to probably
01:09:06
haircut you by 30 to 50% and you're going to sell it because it's a liquid. So, if the public markets are giving you
01:09:13
15%, the reality is that you need to get 25% plus. So then the question is what
01:09:19
venture investors generate 25% plus IRRs over a 17-year period or a 15-year
01:09:25
cycle? And it turns out that every fund
01:09:30
has a very low statistical probability of doing it. Every fund manager, if
01:09:35
you're in the top, you know, handful, we all know who they are. They'll always at least have one that does it. So the
01:09:42
problem is you have one that does it then the next one that doesn't then the next one that's even worse and maybe one that comes close and so there is no
01:09:50
systematic approach to the return street. So how should you look at venture
01:09:55
capital I guess is one question and I think the way that one should look at it is look in a portfolio where you
01:10:02
have a lot of liquid assets it's a small piece that essentially is a mechanism to
01:10:10
learn about the future and to understand how your other assets are going to perform. I think it makes a ton of
01:10:17
sense, but if you're trying to crush it, I think the only people that will really make money in venture capital are the
01:10:23
GPS building a fee generating business around it. Yeah.
01:10:28
Now, that then gets the second question, which is then, okay, so if you do invest in venture, what's the next big problem?
01:10:36
The next big problem is in a world of AI, do you just see the amount of money that's required to build a really big
01:10:42
success shrink? And if that happens then you know how much money can you actually really get working. So if you look at
01:10:49
companies like midjourney that's a great example of a business where these guys are like you know 40 50 60 person team
01:10:56
generating hundreds of millions of revenue very profitable. They're not going to be raising an enormous amount
01:11:01
of money. So I think that these two things I think are very complicated for venture. Doesn't mean that people will
01:11:07
stop investing but Freeberg you want to chime in on this I think. Yeah. Yeah. So I I pulled some data together,
01:11:14
Nick, if you could pull up the basics. I know we talk about this a lot, but I thought it would be good to show the difference between kind of a normal
01:11:19
distribution and a power law distribution. like venture returns or I don't like to call it venture return but
01:11:24
I do think the returns generally in free markets
01:11:30
create value creation in a power law distribution which means that a few of
01:11:36
the many account for the vast majority of the capital appreciation of the the
01:11:42
value creation and that's because of the power of compounding. If you build a
01:11:47
better engine of technology, a better business engine, you will compound and that means that over time you will
01:11:53
accumulate more and more of the market in an outsized way. As you accumulate more of the market, you actually move faster in accumulating more in the
01:11:59
market and eventually it becomes a runaway flywheel. No one can catch up. That's the key
01:12:05
return profile in free markets generally speaking. And a manifestation of that is
01:12:10
that is also the return profile in venture in technology venture investing because
01:12:16
when you find that one or two great hit the Uber that has a great network effect or the Airbnb or the Google it has a
01:12:23
runaway effect in the market and it accumulates all the capital and becomes really valuable and you can see this in the return. So if you just pull up this
01:12:29
is the latest Carta data. So this shows what the IRRa by the vintage that the
01:12:35
fund was raised and then also by the size of the fund and by how the fund has performed on a percentile basis. Look
01:12:41
just at the 2017 row. So the top decile funds are doing call it 30% versus you
01:12:49
know the median fund. Hold on. They're not these are all markups paper markup. These are not dist
01:12:54
these are all numbers. Okay fair enough. But I'll make your case in a second here. And these are not DPI. So it's not distribution cash out.
01:13:02
This is just marked up value. And this is eight years old. This is the 2017 vintage. But just the difference
01:13:07
between, you know, call it a 30% IRRa and a 10% IRRa. That compounds over 10
01:13:12
years. So the multiple difference at the end you should expect is the difference of making 14 times your money or 2.6
01:13:20
times your money. The difference between 10% and 30%. And that's kind of what this shows roughly is that the
01:13:25
difference between a 50th percentile and the problem with this data though is this is Carter's product is used by a
01:13:32
small number of people for a small number of years. That's why the other data we showed from Cambridge is much better because the 2018 forward are
01:13:40
still in the J curve. So they and to Chimat's point they're paper so I knew I knew out there.
01:13:46
Okay. I knew we would go down this rabbit hole. So I shouldn't have shown that but like No, that's okay. There are two key points I wanted to make and just pull it
01:13:52
up again, Nick. The first key point I wanted to make is because of the power law, the job of investing is to find the
01:13:59
power law winners. It is not to buy the index. If you buy the index, you are
01:14:05
losing to the market. Even if these mark even if these numbers are correct, which they're not as we've talked about, you're losing to the index if you're
01:14:11
just buying the venture index. And it's probably way worse than this. You have to get the winners. And so what this
01:14:17
does show though is that the smaller the fund, the less diversified it is, which means the more they're likely to capture
01:14:23
the winners. And you can actually see that the smaller funds generally have better performance than the bigger
01:14:28
funds. For sure they do because they're getting in earlier as well, Freeberg, and they're paying less value lower
01:14:34
valuation. So that's also not they're not getting diluted away by the index. And the index
01:14:39
generally is going to return negative. It's going to be a negative returner. I think I told you guys I went to an LP
01:14:44
conference years ago. one of the biggest venture funds and they showed that 45% of their capital went into flat or down
01:14:51
rounds and they had a negative net return on that allocation of capital across 13 funds and if they had put that
01:14:57
if they had never invested in flat or down rounds then they would have doubled their IRRa overall as a as a fund really
01:15:04
important point that the the power law matters which is your job is to just find the winners and if you find the
01:15:10
winners they're going to compound so how does that translate so so go to the next slide Nick so this is the one that I shared in 2023 three that Gokul Roger uh
01:15:17
published on Twitter and that's where I first saw this which is if you buy just the top 10 companies in the NASDAQ and
01:15:24
you just held it, you would have made a 24x multiple over a 24-year period
01:15:29
versus if you had just bought the the NASDAQ. So just owning the best companies in the NASDAQ is by far the
01:15:36
best way to drive multiple of return over a 10-year period. you made a 9x
01:15:42
multiple 9x just by owning the top 10 companies in the NASDAQ. And this was through roughly the end of 2023. So now
01:15:50
let's pull this up. I pulled this analysis together this morning for our conversation. And this shows that the
01:15:55
venture returns don't stop when you stop being a private company. That the real return and this highlights the point
01:16:01
that most of the value when you find that power law winner, most of the value is created when they are a public
01:16:08
company. So, Palunteer went public after 17 years, and their market cap was 16 billion. So, they created 16 billion of
01:16:15
equity value over a 17-year period. Since they've been public in 5 years, they're now worth 436 billion. So,
01:16:22
they've created another $420 billion of equity market value in just the last 5
01:16:27
years, probably eclipsing all venture returns that have been made during that same period of time. Airbnb went public
01:16:34
after 12 years, worth 47 billion. In just five years, they added another, call it roughly 30 billion. Uber went
01:16:41
public 75 billion after nine years as a private company and now they're worth 190 billion. So they've added 120
01:16:47
billion of value in just 6 years since being public. And Spotify, another good example, 27 billion market cap when they
01:16:54
went public after being private for 10 years. And then they added another 120
01:16:59
billion of value in the seven years they've been public. And I've got the whole list here. This is all these companies that have 100 billion plus
01:17:05
market cap. And when you find the power law winner, the value continues to accrete. It continues to compound and
01:17:12
you're better off just buying the public stocks. And I think that this was my my big kind of conclusion as I went through
01:17:18
all this data. Like Chimath is exactly right. When you're in the venture market, you learn a lot. And
01:17:24
ultimately, I think you probably want to learn a lot to then allocate your capital more wisely into what you've identified are the power law winners
01:17:30
where most of the value continues to accumulate whether they are private or public. And having an IPO is just a
01:17:36
transitionary event for these companies. Their their compounding engine will continue as a public company if you've
01:17:41
identified them. Facebook such a great example. Eight years as a private company, 100 billion market cap at IPO,
01:17:46
added another two plus trillion dollars in the years since. Yeah, I think you Freeberg, you you've cherrypicked all
01:17:52
the winners in hindsight, the biggest winners for that list. What if you were just to create a list of every
01:17:57
IPO tech IPO and how it's fared afterwards, right? No, it's a fair point. There be a lot of stinkers in there that you
01:18:03
haven't but my point, Sax, is just find the power law winners and then that's where all the the value improves. Well, I mean that that's hard,
01:18:10
too. Yeah. Here's the I think you have to look at how we got here. I I started doing this just, you know, in my second
01:18:16
decade and companies decided to stay private longer. That was the big trend. And then
01:18:22
there was a lot of reactions to that. Secondary transactions started occurring
01:18:28
and then you just couldn't get DPI. we had to hold it forever and now the
01:18:33
industry is responding to this. So now we have something called strip sales. We have people creating continuation funds,
01:18:39
ways to take those early investors and get them out into a new vehicle or to
01:18:44
create some liquidity. And Chimat's point is really important because you do get so much intelligence and what I
01:18:51
learned having been an investor in Uber and Robin Hood when they were 5 million and $20 million companies. I was like
01:18:57
first investor into those companies is when they crashed in the market when Uber hit 30 I bought up a bunch of it as
01:19:04
a public investor even though I had a ton from when I was an angel investor. And I did the same for Robin Hood at $12
01:19:10
because I knew the management. I knew Vlad. I knew Travis. I knew Dra. And I had that understanding of the market.
01:19:15
And the same thing I did with Facebook, having listened to Chim talk about it all these years and Brad Gersonner talk about all these years. I backed into it
01:19:21
when it was $92. So what's happening now is what Ruof did with his continuation
01:19:28
fund and holding the private ones is in the theme song of this podcast famously when Zach said, "Let your winners ride."
01:19:34
That is the overarching thing here. the industry, the venture industry is in massive transition. It's going to start
01:19:40
looking a little bit like private equity and a little bit like Gavin or Brad where they're private and public market
01:19:46
investors and you're seeing buyouts occur. So people are saying, "Hey, this SAS company's undervalued. We're just
01:19:53
going to buy the company and we'll start operating it." So venture is in a massive transition. But I think what comes out on the other side is something
01:19:59
that looks a lot more like public private investing like RUF pioneered at
01:20:05
Sequoa and we're going to talk to him about that. I'll just respond to your last comment and Jal I appreciate you saying that
01:20:10
because yeah Roloff is coming to the summit to talk exactly about this which is going to be great but that is the job
01:20:15
of being an investor to your point. You're like hey you know you cherrypick the winners but I think that's the point. The point is that the returns
01:20:21
acrew to these power law winners and the job of being an investor shouldn't be to index a market. Anyone can index a
01:20:28
market. The job of being a great investor is to find those winners. And the question is how how hard that is or
01:20:34
easy it is. And I think your list makes it sound like, oh, all you have to do is
01:20:39
wait for these companies to go public. Sorry, that wasn't my point. Yeah, sorry that wasn't my point. It kind of came
01:20:45
across that way. Yeah, sorry. I was just saying that the power law winners continue to acrew. In fact, not only do
01:20:50
they acrew as private companies and then you if you get one of them, you you have these 14x funds, but as a public
01:20:57
company, they continue to acrew. And so, you know, this isn't just limited to being a venture investor. Pretty much
01:21:02
anyone has access to the public markets and can make the decision that they've made a bet on a power law winner. There
01:21:08
were a lot of people who were very vocal supporters of Palunteer, very vocal supporters of Uber, of Spotify when
01:21:13
these companies went public and a lot of people said no, they're naysayers. I think the problem that you would have
01:21:19
had trying to implement this strategy is it would not have been clear what you would have been underwriting. What would you have been underwriting at Nvidia for
01:21:26
you know not the last 5 years but the many years before then you would have been underwriting video games.
01:21:32
Video games. Yeah. And you would have had this like very odd data center business that didn't
01:21:38
make sense that you know Jensen took a lot of heat for that he had to defend
01:21:43
and it it didn't actually make a lot of financial sense. So I I'm not sure. I will give you a counter to that and
01:21:48
I'll I'll call him out and I'll make fun of him now. My co-founder at Climate Corp. His name's Siraj. He's probably listening right now. He bought Nvidia
01:21:55
years ago on the thesis that these GPU chips would eventually be used for AI models because he did all his work at Stanford and distributed computing. He
01:22:02
did early work in AI. He was very close with That's different than what you were saying before. I get it. He's a smart guy who made a huge winner. But I'm
01:22:09
saying but no, the pro the problem was he sold in too soon because he's like, "Oh, they're not actually." Okay, but I'm just saying like there was probably a
01:22:14
person that was as smart as Saraj and bought Intel. Then there was another person that bought AMD. Like I don't I'm not sure what the point is. Here's the
01:22:20
point about venture that I think is worth noting. It's an exceptional
01:22:26
market that gives the practitioners an incredible edge if you build a
01:22:32
business around the information asymmetry that occurs. Yes, that you can monetize. After that, the
01:22:39
only business model is building an asset gathering machine to generate fees because even the consistency of being
01:22:46
able to generate these power law distributions in successive funds
01:22:52
doesn't exist. And you can look at all of the data. Cambridge has done this. A lot of these guys have done this. All
01:22:57
the big LP fund of funds like Cory Bridge have done this. Having a killer fund has zero correlation with your
01:23:05
ability to then have a next killer fund as an LP and I showed these returns on screen although we blanked out the names
01:23:11
and I'm in all of them. It is really hard to be consistent. It's hard but yeah let me make let me make two points
01:23:17
about this. So and then I got one closing point make so first of all Jamath is right that if you look at the data the recent data say
01:23:23
over the last decade or so you don't want to be the average venture fund. you'd be a lot better off just being in
01:23:29
the NASDAQ or in some other uh public stock market index because a your
01:23:34
returns would be higher and b you don't have the illquidity so you can always trade out of it if you want. So it's
01:23:41
true that you don't want to be the average venture fund as an asset class. The average is not great. This goes to
01:23:46
Freeberg's point about the power law. You really want to be a top quartile fund or even better top decile. And
01:23:52
that's what LPs are looking for when they're underwriting VCs is they want to be in top quartile, not the average. So
01:23:59
I I agree with that point. But the second point is that things can change very quickly in this asset class
01:24:05
precisely because of the power law and how spiky it is. So just to give like one micro example, take Figma. You know,
01:24:13
we talked about on this pod was a a year or two ago when the government held up
01:24:18
their acquisition by Adobe for $20 billion and everyone thought that was a disaster because people thought that was
01:24:24
a huge premium to the actual valuation and if eventually they now had to go public as a result they might IPO at 10
01:24:31
billion and so it was a big haircut for all the VCs. Well, couple years later they just IPOed. They had a huge pop on
01:24:38
the first day. It's come down somewhat, but the public market valuation is now 35 billion. And there's three or four VC
01:24:45
funds that have made five, six billion, 7 billion. Yeah. And those three or four funds are now
01:24:50
10x funds. So just boom, in the blink of an eye with one outcome like that, it totally changed
01:24:57
the performance of those funds. And then that's going to trickle down the performance of the asset class as a
01:25:04
whole. My guess is that we're in a pretty good time for VC because of AI
01:25:10
and the disruption it's causing and the potential for value creation it's it's
01:25:15
creating. It's sort of restocking the pond uh with opportunities for many more
01:25:22
uh big companies to be created and I think ultimately that'll be very good for the asset class. I mean just look at open AI. I mean, look, we're a long way
01:25:28
off from Open AI being liquid, and you know, you don't want to count your chickens before they hatch. But just two
01:25:34
years ago, on the heels of Cha T launching, the company raised
01:25:39
money at a $30 billion valuation. Everyone thought that was crazy at the time. Now they're at a $300 billion
01:25:46
valuation, and there's news reports that they're going to raise at 500 billion. So, if this keeps running, and again,
01:25:52
there's a lot of ifs there. I mean, they could still lose. They could get sell 20% is my best advice. They could still lose to some other
01:25:58
company or whatever, but it just shows you how fast things can change. I mean, they could be a trillion dollar
01:26:04
company in a couple years if if the trend continues. So, so my point is just it it because of the
01:26:10
power law, it only takes a few big winners to to change the economics greatly for particular funds and then
01:26:18
that eventually trickles down to the mean. But again, you're right. You don't want to be in the mean. You want to be top quartortile. There's just two quick
01:26:23
points I want to make as we wrap on this and great episode gentlemen. The first is also the wrath of Lena Khan occurred
01:26:30
for four years. So DPI and exits were muted. So that also plays into a lot of
01:26:36
these statistics. And now since President Trump number 47, David,
01:26:41
we have had a tear of M&A and a ton of IPOs, right? That's actually going to be
01:26:46
probably a big, you know, factor in all of this hand ringing around venture. But I just want
01:26:53
to give two anecdotes here. You know, when you have the inside information, Chimoth, it's worth I think being a
01:26:58
venture investor because of that data you get. And I was just always perplexed when I would hear,
01:27:05
you know, people on CNBC talking about Uber and Robin Hood specifically because they would be talking, oh my god,
01:27:10
they're losing so much money. They lost a billion dollars this quarter. They lost$5 billion last year. And we all
01:27:16
knew that there would be no difference in consumption of an Uber if it was $13 or $15. And that at any point you could
01:27:23
just turn the dial and slow growth and show profitability. And what happened? That's exactly what they did. And I made
01:27:30
that call and I said this on CNBC like is anybody not taking an Uber for a$1 or $2 difference per ride? It's like 2% of
01:27:36
people might stop taking it. Sure enough, they did that. Then fast forward Robin Hood. Everybody's like, "Oh my
01:27:41
god, this company is like it's sideways. the market's corrected, people are no longer trading stocks. And I was like,
01:27:47
but they keep releasing exceptional products over and over again. And that was always Vlad's key perspective was,
01:27:54
we're just going to keep adding every product, 529s, 401ks, margin loans, crypto. We're just going to keep adding
01:27:59
products every 6 months and then we will be the default place where all your finances exist. And when it was at $12,
01:28:05
I was like, "Oh, well, this obviously is undervalued. I'm just going to buy more." And now it's at 120. It's like
01:28:11
one of the best. is a 12 bagger in the public markets. So that reason alone, if you just have enough surface area and
01:28:17
enough patience, you're going to hit Robin Hood, Uber, like I did when they're tiny companies and then you're
01:28:22
going to get to have all that intelligence into the public markets, which is why I started doing public market investing because I could just see all these very clear opportunities
01:28:29
where things were obviously undervalued. It was just like so obvious.
01:28:34
Well, when you talk about service area, what you're talking about is diversification. Yes. Right. But the way Freeberg just showed
01:28:40
us the winners cherry picking. Yeah. Yeah. And that's that's kind of my point is it looks really easy when you just
01:28:47
look back over the last 10 years and pluck out the winning companies and the winning funds and the winning asset classes. The
01:28:53
question is Xanti. How do you create enough diversification to get exposure to that? And that is why venture is
01:28:59
still a popular asset class is because it is very spiky and you do want to be
01:29:04
in those power law funds and power law companies and it might be a percentage like the percentage might change. Maybe you know
01:29:11
Yale wanted to get to 25% or whatever and maybe the right number is 15. It could change over time based on your own
01:29:17
individual needs. I still think it's magical asset class. Chamop you want to wrap us up? I did that job for seven
01:29:22
eight years with outside capital and now the last seven eight have been entirely my own. The
01:29:30
differences I would say is that it's much much harder to run a very
01:29:36
successful venture firm in my experience. It's much easier actually to be concentrated
01:29:42
and to use the information asymmetry in your favor. It is an exceptionally difficult job to be a successful fund
01:29:48
manager like what you guys do I think is extremely extremely hard and to be of that quality.
01:29:54
Well, we both started we both started our own firms that just increases the degree of difficulty because you have to
01:30:00
build a team you have to build an LP. So if I could if I could own pieces of the GP of both of you I would. But as a
01:30:07
person today I would not start my own fund. It's just very difficult. Well it's it's a very competitive
01:30:12
industry and there is a lot of capital. very competitive. It's super competitive. But I would also say that
01:30:18
and that's the supply of capital. If you look at demand for capital, I do think that there is more legitimate demand for
01:30:23
capital now than there has been in a long time because there's this new, you know, you've got this incredible
01:30:30
real businesses want to be funded. I agree with that. Well, there there was a I think there was a feeling, you know, 5 years ago,
01:30:36
let's say, where all the big platforms were getting a little bit long in the tooth. you know, social, mobile, cloud,
01:30:42
SAS. It felt like again that the the pond had been fished out a little bit.
01:30:48
Yeah. And you need a big you need a big disruption to to again to restock the pond. And I just think that
01:30:54
AI is the mother of all disruptions and it's creating a lot of opportunity. All right, everybody. This has been
01:30:59
another amazing episode of the AllIn podcast for your Sultan of Science, David Freeberg. Yours are David Saxs.
01:31:07
Civil servant David Saxs doing the work for the American people. And for our chairman dictator,
01:31:14
I am the world's greatest moderate. Jason's greatest moderate.
01:31:20
Love you guys. [Music] Let your winners ride.
01:31:26
[Music] And it said we open sourced it to the
01:31:32
fans and they've just gone crazy with it. Love you. Queen of
01:31:38
[Music] your
01:31:43
besties are gone. That is my dog taking a notice in your driveway.
01:31:51
Oh man, my dasher will meet me at We should all just get a room and just have one big huge orgy cuz they're all
01:31:57
just useless. It's like this like sexual tension that we just need to release somehow.
01:32:03
be wet your feet or feet. We need to get merch.
01:32:10
[Music]
01:32:16
I'm going all in.
