Latest / Financial Survival Network / AI's Influence on Financial Markets and Productivity - Max Osbon #6087

Transcript

AI's Influence on Financial Markets and Productivity - Max Osbon #6087 · Financial Survival Network

0:00
I think like it's tempting to say like AI is going to automate everything.
0:04
I think you really have to still have a view and have a driving bisis
0:08
and then understand basically how that AI is going to hold your hand through those
0:12
ups and downs and wherever it's going, so you know first first kind of
0:17
concept I would say is like for markets, the biggest risk I find is
0:21
when all markets are being driven by a single factor. You are listening to
0:24
Carrie Letz's Financial Survival Network, where you get valuable information you just can't find
0:30
anywhere else to thrive in today's trying times. You need the Financial Survival Network
0:36
now more than ever. Go to Financial Survivalnetwork dot com and get your free
0:41
newsletter and gift. Financial Survival Network now more than ever. And welcome you
0:52
are listening to and watching the Financial Survival Network. I'm your host, and
0:57
right now we've got Max Osmond with us. And Max you are a eventure
1:02
capitalist, fintech guy, crypto guy, and big tech trend follower. We
1:07
love tech on this show. I couldn't be doing this show without tech.
1:14
That right now, we just take it for granted, Max, But you
1:17
know, ten years ago Zoom was a real pain to use. Oh yeah,
1:23
you know ten years ago to do a podcast, really you had to
1:26
really jump through hoops. Now it's all commonplace and we just take it for
1:32
granted. But one day, at one time, this was somebody's vision.
1:37
So you specialize in working with people who have vision, right, a vision
1:42
of creating a product or service based on tech that's going to have an effect
1:49
in the future once their vision is realized. How how do you separate people's
1:55
visions from pie in the sky to hey, this guy really is onto something.
2:00
Well, so my primary framework on anything is that there are no companies
2:05
or tradings to strategies or anything that are run autonomously. It's all run by
2:09
people. So at the end of the day, the companies any company,
2:14
but it's Apple, Google, Amazon, It's run by people. So the
2:17
primary filter, beginning, middle, and end of a zero to one kind
2:22
of decision making process, it comes back to the person. So do you
2:24
believe in this person, their vision, their capability, and really, more
2:28
importantly than that, the context. So how do they arrive at this?
2:32
Why? Now did they fail at a whole bunch of other stuff? And
2:36
this is what they're like kind of last shot is it desperation or is it
2:39
something that lined up perfectly for them because of their career history, their opportunity.
2:45
You can look at public markets, you can look at private markets.
2:46
I think all in the same way. Why is this person doing what they're
2:51
doing? Is this something that I'm willing to get on board with right now?
2:54
I think it's a great framework for evaluating public stocks too, and that's
2:57
why people like founder led companies in public markets. Okay, so as far
3:01
as when you you know, because mostly these people are considered somewhat eccentric,
3:07
They certainly don't fit the moldy because they wouldn't be doing what they're doing.
3:14
If they did right, they'd be working, you know, coding and makeing
3:19
nice living. But that's not enough. So what is it about a particular
3:23
founder where you will look at him and say, Wow, this guy is
3:28
really onto something. It's who they're connected to and what they'd done previously.
3:35
So you know, we do we do some direct investing at an early stage,
3:40
and you know, I won't go in like to too many specifics because
3:45
they're not like publicly trade not available wouldn't be helpful in this case. But
3:47
you know, in the last investment we made was a guy who'd been in
3:52
the hedge fund industry for twenty years in the utilities sector and retired and put
3:57
you know, a bunch of money into utility startup that decentralizes ammonia and and
4:02
then started the company and has all the connections and is a unique CEO,
4:08
founder and h and it's the kind of person that you'd want to follow into
4:11
battle, right this is this is a is a good fight, and these
4:15
are talented people. As a talented person, so the previous company before that,
4:18
you know, it's like in the regulation fintech regulation space, same same
4:24
story. Someone who worked at levels of government, connected affected well within government
4:30
and has a technology background. And then the problem with starting companies that have
4:34
to interact with the government is that so many people come in too hot with
4:40
the government. You know, they'll be kind of insulting, and you got
4:42
to realize there are people on the other side too. They don't like to
4:44
be insulted, even though they know that their regulations are frustrating. And so
4:47
if you come in from both angles. I think the reason why they've been
4:50
successful is because too many people came in and insulted the government and they work
4:55
collaboratively. You know. This is when they started. It was around the
4:58
time when Uber was still getting and kind of shut down, you know,
5:01
locally because in favor of the taxi drivers because they're being too aggressive. Same
5:04
with airbnbing. Yeah, like that was a fascinating story, and they came
5:12
to the conclusion it's better to ask forgiveness, better to seek fitness that it
5:17
is ask permission. Uh, you know, and that works in some cases,
5:21
maybe not. This one that really annoyed me about that was the scooter
5:27
companies that just dumped their scooters and like the Middle Texas, right, Yeah,
5:33
not a good model in my opinion. No, that was an interesting,
5:38
like terrible, terrible execution. Maybe would never have worked, but like
5:42
the financial model was always interesting online scooters because you know, you're basically buying
5:46
them cheap and then you're trying to rent them out for like a dollar a
5:49
day or a dollar an hour or whatever it is, and so the financial
5:53
model may be made sense, but ultimately, yeah, you'd ended up just
5:57
creating a ton of trash. I don't I don't focus a ton on the
6:00
on the line story or that, but but I am I am focused you
6:05
know, a lot right now on like you know AI, AI's impact,
6:12
whether it's as impact on a portfolio or AI is impact on how you spend
6:15
your time efficiently to be productive. So let's talk about that. So AI,
6:21
you know, I don't know certainly it can identify trends. It kind
6:28
of is a a progression from neural networks and quantum computing right well, the
6:38
way I look at it is like zeros and ones for classical computering. That's
6:42
just like it's binary, it's if then statements. It's like if I go
6:46
over here, if it's Tuesday, or if it's raining outside, if it's
6:48
raining outside, bringing umbrella. If it's not, you don't need it,
6:53
And it's just statements. What a AI has done in machine learning is has
6:57
just made it probabilistic. So if you add up enough of the zero ones,
7:00
you can do an insane amount of probability calculations and instead of getting you
7:06
know, should I bring the umbrella or not if it's raining, it's like,
7:09
well maybe sixty percent chance you should bring it, you know, And
7:12
then you add those things up enough, and then you end up replicating language.
7:15
You end up bropricating analysis where it becomes really interesting for financial markets is
7:20
I would say I now basically have a team of about twelve different analysts AI
7:25
analysts who are helping me, you know, kind of continuously evaluate markets.
7:30
Markets are probabilistic, which you know fits really well with the AI theme.
7:34
You know, so there are really no binary I mean, there are binary
7:38
outcomes in markets. You know you're either successful or not, but the analysis
7:44
is super probabilistic. Okay, So so where do you think AI fits in
7:50
to finance and all that. I mean, hey, yeah, thinks me
7:55
like fraud detection, money laundering, and maybe not so much as predicting the
8:03
future, but it's saving you from things that you might otherwise be better off
8:11
without. Yeah, well, well, I think like it's tempting to say
8:16
like AI is going to automate everything. I think you really have to still
8:20
have a view and have a driving thesis and then understand basically how that AI
8:24
is going to hold your hand through those ups and downs and wherever it's going.
8:28
So you know, first first kind of concept I would say is like
8:31
for markets, the biggest risk I find is when all markets are being driven
8:35
by a single factor. In twenty twenty two, all markets were being driven
8:39
by rate hikes. Is like did not matter what else was going on,
8:43
every company is falling be through the rate hikes and the fear around that,
8:46
and of course COVID crash, all markets being being driven by onset of COVID.
8:52
So right now, I would say like the most useful kind of AI
8:56
tool is sort of like analyze are being markets by a variety of factors or
9:01
by a single factor, And I think what explains market's doing well right now
9:05
is markets being driven by a variety of factors. It's maybe rate cuts,
9:09
maybe increases and decreases in labor availability, increases or decreases in inflation. You
9:16
know, it's like I could list off ten or fifteen different factors that are
9:20
influencing each sector, and to me, that's a healthy market. That's a
9:24
healthy market that can continue. Okay, So yeah, so kind of talking
9:28
about markets that are just being dominated by emotions rather than fact But you know,
9:37
some would argue there's only two emotions, fear and greed, and then
9:41
you have fear and greed's fomo and kind of a combination of the two.
9:46
But it's a little more complicated than that, isn't it. Yeah, it
9:50
definitely is. I mean, we're definitely in a fomo stage right now.
9:52
I'm getting more calls of like, why do we hold any cash at all?
9:56
You know, we should, you know, And that's just like rearview
9:58
mirror, and you have to make decisions based in the future and not the
10:01
past, and buying the S and P now or whatever, it doesn't give
10:05
you the eighteen or twenty percent that you missed out on in the last year
10:09
or so. And of course that breeds risk right like call markets make make
10:15
for risky your markets. Some of the things I'm concerned about right now.
10:20
One is just like massive institutionalization of call selling. You know, there's been
10:24
a long, long time, a long project by JP Morgan other banks to
10:31
sell products that just have automated call selling behind them. And anytime you make
10:37
a one sided market where they're just like, oh, we're just going to
10:39
harvest yield on call selling carves siealed, it's eventually going to crack, right.
10:43
It's it's too much imbalance on one side of an options market. And
10:48
so like I wouldn't touch any of those products that are that are that are
10:52
call selling. I also don't love the concept. You know, if I
10:56
own equity, I want. I want unlimited upside. I don't want to
11:00
my upside by selling a call trying to eke out one or two percent on
11:03
a position. I know people do it opportunistically, but what I've seen,
11:07
you know you don't want to sell calls on your apples so I can get
11:11
a called away missed? You know whatever next generation of you know you think
11:18
tech can't jump twenty percent, and then you see you do it in a
11:20
day and then kick yourself about why you sold calls on it. Kicking yourself.
11:26
Yeah, well, you know, I understand a covered call strategy.
11:33
I think it can be effective. But the problem is, like who thought
11:37
the market was taken off in April? Everybody was predicting the recession, you
11:43
know, the old saying economists has successfully called five out of the last two
11:48
recessions. Yeah, they're never right. And anytime you see experts any headlight
11:54
of any article or it says experts say you can just put in parentheses.
12:01
People who never get it right. Yeah, well the experts are get it
12:05
right. The problem with calling him an experts is like you're giving them a
12:07
veil of an anonymity. Like I want to hear which expert like is it
12:13
Paul Krugman, because like I don't want to hear what he has to say
12:16
about where tech is going or stand Drucklin Mall or some of that. Who
12:20
is like incredibly influential for financial markets? Yeah yeah, yeah, Like okay,
12:26
one man's expert is another man's uber driver. You know, yeah,
12:31
yeah, exactly, your drivers are experts and getting it from point A to
12:35
point bat watching the map on the screen, you know, But does that
12:43
mean they're an expert? And who determines boom is an expert? Right right?
12:50
Well? Definitely a personal decision, or at least it's it's been a
12:54
super personal decision for me, and I like, I like being aggressive in
12:58
my own mind of like I will absolutely never listen to a thing this person
13:01
says. I mean, maybe they'll be right on some topics for me.
13:05
I don't know. Maybe we'll we'll we'll have some fun to take some shots
13:09
of people. But like Fessor Galloway has been like a perennial like issue for
13:15
me, Like I don't want to I don't want to follow his ideas or
13:18
his trends, even though he's got a lot of press. I just think
13:22
I just don't like his uh, his framing and his angle. You know,
13:26
ultimately you're framing is a culmination of your experience, and whatever experience he
13:31
has has brought him to this level. But meanwhile, I love listening to
13:35
Dan Dan Miller. Yeah, you know, kind of jury's out when it
13:39
comes to who's the bond? Who's the bond? Guy Double Linear, Jeff
13:46
Gunlot. I'm kind of like Jurry's you have to you're on expert? Yeah,
13:50
no, I got you, And you know, it's kind of you
13:54
got to choose your expert. But eventually experts are like employees. Eventually they
14:01
all disappoint you. Yeah, it's like I never meet your hero, right,
14:07
and you held them in higher status in your mind than they could ever
14:09
drew up to. You know, I'm going to remember that one because I
14:13
think it's totally true. And yet you can't know everything. You got to
14:18
trust somebody in this world, and you know, with you, trust is
14:24
a big issue because you're putting in, you know, millions of dollars into
14:28
a startup and you got to know and you know, Leeyakoka once said there's
14:33
two things you could never know about a person. Number one is if he
14:37
or she is honest, and number two is if they're hard working. Now,
14:43
I think number two you can figure that out based on prior accomplishments.
14:48
You can get some idea, but the honesty test is really difficult. You
14:52
could give them a battery of of psychological profiles. I would say what he
14:58
was really saying there. You can never tell if somebody is a sociopath until
15:03
you see them in action, you know. Yeah, with one of the
15:09
some of the really seasoned veteran venture capitalists that they've factored in that like ten
15:13
percent of their you know, founders or some number that will be like flat
15:16
out lying to him, not out to them, right, and they just
15:20
have to model. Hey, and you know, like nobody wants to be
15:26
the bearer of bad tidings. Easier to be a yes man than it is
15:31
to be a no man, right right, Yeah, yeah, I mean
15:37
I I I tend to think optimism is the is the better strategy, especially
15:39
being a young person. It's like lots of time to let things not go
15:43
well and then and then get kittyer wins. You know. The best thing
15:48
is I think all about asymmetric upside. So you have to take a risk,
15:52
you have to let some things fail I think even Ken Ken Griffin.
16:00
You know, if Citadel says like we've lost more money than any other edge
16:03
fund, he's also made more money than any other edge fund. I think
16:07
it just goes hand in hand. You have to have you have to be
16:08
willing to take losses. Yeah. Yeah, yeah, there's a lot of
16:15
wisdom out there, but it's choosing between it. Now, with the Internet
16:18
and all that, you have unlimited access to it, all the good,
16:22
the bad, and the ugly, and most of it is ugly. Hey,
16:26
So trends looking ahead? Fin tech is your your baireiwick there? What
16:33
what can we look forward to here? You know? Well, I would
16:37
say, you know, within within fintech. Just one of the frustrations with
16:42
the field is if you watch the Apple Development conference from this week and you
16:48
see the things that they're able to do with the devices, right, They're
16:52
they're leveraging machine learning there, you know, optical character recognition, you know,
17:00
an enormous number of automations to roll them out to Apple products, and
17:03
it's like pushing the technology to the absolute limit of what it can do.
17:07
And then you get back to fintech and you get like we struggle with like
17:11
basic math or ledgers or like you see all the different like failures of the
17:15
different kind of what was the most. I haven't been following it down to
17:18
the detail, but there was a couple. There was a regional bank that
17:22
failed that was maybe commingling accounts, and like it's such it's from a technology
17:27
perspective and from a math perspective because my background is in mathematics. It's just
17:33
like regular algebra, right, and it's so it's not hard from a math
17:37
perspective, it's not hard from a concept perspective. For some reason, it's
17:40
like the technology lags and fintech in that area by just I don't know,
17:45
decade two decades. Like if we take what Apple is doing, we would
17:49
dramatically improve, you know, what's going on in the fintech world. One
17:56
other trend, though, what the trend? The trend that I would say
17:59
is capturing my attention the most right now, I have a call later on
18:03
this today I'm trying to piece it together is quantum and it's not necessarily to
18:07
fintech, but it is a really important kind of emerging merging concept and if
18:15
I get a couple of minutes, I can frame it for you. Please,
18:18
I think is interesting. So we're hitting the physical limits of what you
18:22
know in video GPUs and computing can do. The physical limits are defined right
18:27
now by energy. So it's like, you can build a bunch of data
18:30
center. It's actually not hard. It's not as hard to build the data
18:33
center as it is to build more electricity. And how do you get the
18:36
electricity? Is it going to be sole It can't be solar because you need
18:41
it can't be renewables because you need consistency for the data centers. You can't
18:44
afford to go down. I mean, I guess you could put a battery,
18:48
a megapack battery or two or three or some of them are so big
18:53
you probably need twenty of them ton and then they consume just an unbelievable amount
18:59
of ER and then you have, you know, the five nines of reliability.
19:03
You can only have four minutes a year of downtime, not like anything
19:06
more than that. No one's going to use your data sare So it's like,
19:08
so do you do nuclear? Well, it's going to take I think
19:11
Bill Gates said today they're building a nuclear power planet. It'll be alive by
19:15
twenty thirty. So it's like, what are we going to do for the
19:18
next six years. You know, it's like you need you need things now.
19:21
So anyways, I think about quantum because if we're going to hit the
19:25
limits of what we can do with traditional computing, you know, quantum.
19:30
The word quantum is really into etymologies. It helps you, like understand the
19:33
framing. Quantum is a reference to the pieces that are inside the atoms.
19:38
And so when we build viewers now the smallest we can get the transistors to
19:44
sort of like twenty atoms wide. But now I know that's not a perfect
19:48
description for actually how it works, but that's kind of the that's the limit
19:52
we've had. So if we want to get even smaller, we have to
19:55
use pieces of atoms. Those are quantum, you know, quantum or quantum,
20:00
and then if you want to use the computer based on those, you
20:03
can basically dramatically scale up its processing power. I don't really understand completely how
20:07
it works on every level, and I don't think anyone really does. There
20:11
are some really weird things that happen in quantum computing, quantum mechanics. All
20:15
of it is experimental right now. So IBM has the most you know,
20:18
kind of advanced current quantum structure. But my understanding is again like they make
20:25
more money selling quantum preparedness software and stuff then they do actually selling quantum assets.
20:32
So what's going to happen? So my view is, if you could
20:37
if it takes six months to train a model an a model on Nvidia GPUs
20:41
and it costs three hundred million dollars, you know, you develop in machine
20:45
learning algorithm on quantum when it actually functions that would do the same amount of
20:51
work in like a minute, right, and that would be you know,
20:53
we talked about orders of magnitude of a lot and AI taking six months of
20:59
work and compressing town to a minute or something like that. That is like,
21:03
you know, it reminds me of what it used to be, like
21:06
a computer in an entire room. Now, you know, with the Apple
21:11
devices that we're probably talking on, like dramatically more efficient, more powerful,
21:17
and they fit you know, in tiny bags. So it's like, you
21:19
don't think these order of magnitudes improvements can happen, but they certainly can happen.
21:23
So so when is the why an't we going to get breakthroughs with quantum?
21:26
I don't know, like it's but to me, it's the only thing
21:30
that's going to answer the efficiency issue with with machine learning, Like, we
21:34
can't be spending hundreds of billions of dollars to create giant warehouses to train these
21:40
models when there is eventually going to be a breakthrough that simplifies that. It's
21:45
just that that is a breakthrough. So so where do we stand with it
21:48
now? I'm working on this now assuming that at some point over the next
21:52
five years there's going to be a really interesting quantum related investment. I don't
21:57
know what it's going to look like. I don't know where it's going to
21:59
be. If it's going to be IP related, I don't know if it's
22:03
going to be momentum because they solved it and they're going to be running away
22:06
with the technology into the future. But I think now is the time to
22:11
start thinking. Just like if you were thinking about AI ten years ago,
22:15
you eventually arrived on GPUs and Nvidia, I think you have to think about
22:18
quantum now to see what's going to happen. All right, So you know
22:22
during the gold Rush, you know, only a very few miners got wealthy,
22:26
but the companies that sold the pickaxes and the dynamite and the clothes the
22:34
Leli Strausser's of the world they cleaned up what So you outline the problem here
22:41
of the energy shortages electricity particular, because these still are highly regulated industries,
22:51
not like what they used to be because got pseudo competition and distribution. You
22:57
know, in the favorable economic climates and climates too where people are going to
23:06
want to build these multi billion dollar data centers. You know, these utilities
23:11
are going to kill it. And I would also say the utilities around the
23:15
nation's capital because I don't think they aren't the biggest investor in AI out there
23:21
in every aspect of what the government does. So you know, those utilities
23:29
are in a bind because, like you said, nukes take ten years with
23:33
licensing and everything else. And don't believe modularity because that really doesn't speed up
23:40
the process regulation, bog and everything else. So we're going to need more
23:45
energy. To think it's going to come from renewables is naive and so you're
23:53
you're you're walking into a really interesting topic that's that's fun to explore and not
23:59
not an area that I've done enough work on yet. But but you know,
24:03
natural gas is the is the sort of fossil fuel solution it's cleaner,
24:08
more efficient, you know it it is. It's funny because you know,
24:15
we have to have trade offs. We have to have cheap energy, so
24:17
we can't like throw out fossil fuel, but you can use natural gas and
24:21
get some really compelling results. And so that's still not popular at all among
24:27
you know, a Democrat and not the politics you know, conversation. But
24:32
it's just not the democratic kind of narrative. You know, it doesn't even
24:36
want to talk about natural gas. So there's a lot of energy assets that
24:40
I think are really compelling, probably compelling because of what you're talking about,
24:45
Like you need to have the energy of the electricity to support either this transformation
24:49
or just like power the world. Like I like these stresses, Like there's
24:53
no such thing as wealthy com economy without cheap energy. It's like there's not
24:57
a still developed economy that didn't have acces to cheap energy. It's like the
25:00
mental space. Yeah, so so yeah, I wonder because the last decade
25:07
prior to COVID, the last decade of energy Exxon or anything, there's been
25:11
like it was horrible, Like from a from an index perspective, if you
25:17
look at xl E, that's sort of the benchmark I'm referring to or exon
25:21
it was horrible and then all of a sudden, out of COVID. You
25:23
know, it's been an unbelievable asset class to own. It's been the only
25:26
other kind of forming asset class outside of tech. And so how do we
25:30
look at energy going forward? I think it's a really important question to answer
25:33
for investors. You know, do you get deep down in the weeds and
25:37
you pick you know stocks so you get you know, really well integrated into
25:41
the different assets out there, or do you do you benchmarket with an index?
25:45
I don't know. Well, you know the good thing if you're getting
25:49
into that gas, let's show you're speaking my language now, been a moon
25:55
and of it. It's been trading at a discount per BTU to oil for
26:00
decades now, and when I was a child growing up through two energy crises,
26:04
it traded pretty close to parody. And nice thing is you actually get
26:11
paid to weight because they pay out above average dividends almost the couple points behind
26:21
the actual inflation rate. But the technology to manage the grid all right,
26:26
because the problem with the grid is and most people don't even know it,
26:32
like the you have transmission loss fifteen twenty percent, and they kept saying,
26:37
well, we're going to come up with the superconductor power transmission lines and we'll
26:42
have one or two percent and they'll be nitrogen cooled. Well, evidently that
26:48
wasn't too practical, so grid management, and you know, for the rest
26:53
of US residential users and commercial users out there in these states where they start
27:00
building these data centers, it's going to be a fight for electricity. Now,
27:06
in theory, they could bring in their own generator and do it themselves,
27:10
and maybe if they own some that gas wells, they could just feed
27:14
that generator right off the wells. I mean, cryptomiding, it's like insane,
27:21
you know, it's basically the same thing as AI. So you know,
27:25
just interesting things. Well, I think you've like helped us look at
27:29
trends in a different way here, Max. If we want to connect with
27:33
you on the web and take a look at what you're doing, how do
27:37
we do that? Yep? So I write a pretty widely read article on
27:42
my website, osboncapital dot com. It comes out weekly, so I'll be
27:47
writing it today and it'll be out tomorrow morning, Thursday, Thursday morning.
27:51
To pay them in this area. It comes out every Thursday morning. That
27:56
that article is where I both write down my own thoughts for my own to
28:02
codify my own thinking, and to share it with other people, to put
28:06
it out publicly, either to you know, get people to throw throw stones
28:11
at it and tell me why I'm wrong, or to reansworth it to reamport
28:15
at it. Yes, so that's that's my primary things. I write that
28:18
article I've been doing every week for many, many, many years. It's
28:22
my commitment, all right. Hey, well, the link to your site
28:26
is in the show notes to this interview on Financial Survival Network dot com.
28:30
When you go there, please sign up for your free newsletter. I got
28:33
a question for Max or myself. We'd love to hear from you. Shoot
28:37
me an email to kl at Carrie LUTs dot com. Max. Been a
28:41
pleasure. We'll definitely talk to you again. And hey, we're gonna sign
28:45
up to your blog your newsletter there. Thank you, Gerry, thanks for
28:48
listening to carry lets this Financial Survival Network your solution to today's trying times.
28:56
For the latest, go to Financial Survivalnetwork dot Cominancial Survival Network now more than ever