0:00
Curiosity is a shit starter.
Search this video
0:02
And he goes, man, creativity is such a fucking curse.
0:05
Like you, your biggest strength is usually your biggest weakness.
0:08
Feynman says, no matter how beautiful your idea, if it disagrees with experiment, it's wrong.
0:16
So much of society today is so specialized because that's the dominant strategy to get somewhere.
0:24
The most important thing is to not be fooled and you're the easiest person to fool.
0:28
I wish we'd have more effort going into less prestigious areas.
0:31
We don't always get it right.
0:33
We often get it wrong, and having this information about what doesn't work is still useful.
0:37
You learn from other peoples mistake.
0:39
Hopefully avoiding some of yours.
0:41
I think networks are gonna become vastly more important than they were in the past.
0:45
There seems to be a new kind of meritocracy.
0:47
Now people are immigrating to the online world, right?
0:51
We left the stone Age because we got better ideas, we got better explanations, and as these better explanations flourished, We flourished.
1:00
Progress doesn't depend on the calendar.
1:03
People have to make it happen, and if for a while people don't make it happen or block it, it's not gonna happen.
1:08
What we are on the cusp of here is this explosion of new learning and we can't know what's gonna come out of it. Just figure it out.
1:19
Just find what's interesting to you Jim O'Shaughnessy: Hi, I'm Jim O'Shaughnessy and welcome to Infinite Loops.
1:22
Sometimes we get caught up in what feel like infinite loops when trying to figure things out.
1:24
Markets go up and down, research is presented and then refuted, and we find ourselves right back where we started.
1:26
The goal of this podcast is to learn how we can reset our thinking on issues that hopefully leaves us with a better understanding as to why we think the way we think and how we might be able to change that to avoid going in infinite loops of thought.
1:29
We hope to offer our listeners a fresh perspective on a variety of issues and look at them through a multifaceted lens — including history, philosophy, art, science, linguistics, and yes, also through quantitative analysis.
1:32
And through these discussions help you not only become a better investor, but also become a more nuanced thinker.
1:34
With each episode we hope to bring you along with us as we learn together.
1:35
Thanks for joining us, now please enjoy this episode of Infinite Loops.
1:36
Disclaimer: Jim O'Shaughnessy is chairman and Co-Chief Investment Officer of O'Shaughnessy Asset Management, where Jamie Catherwood is an associate.
1:38
All opinions expressed by Jim, Jamie and podcast guests are solely their own opinions and do not reflect the opinions of O'Shaughnessy Asset Management.
1:43
This podcast is for informational purposes only, and should not be relied upon as a basis for investment decisions.
1:48
Clients of O'Shaughnessy Asset Management may maintain positions in the securities discussed in this podcast.
1:53
Jim O'Shaughnessy: Well, hello, everyone.
1:57
It's Jim O'Shaughnessy with yet another Infinite Loops.
2:00
I was talking to somebody the other day and they said, well, clearly you shouldn't even number.
2:05
You should just put an infinite sign instead of a podcast number because you're going to have Infinite Loops.
2:10
And I said, well, that requires Sim Jim to be fully online and the AI to really work well.
2:17
Many of these topics I am going to discuss with my good friend Liberty Richard P. Feynman.
2:23
He uses only the initials on Twitter and on his Substack.
2:28
He and I have gotten to know each other really, really well over the last couple of years.
2:32
He is one of the most interesting people that I get to talk to.
2:37
He has all sorts of insights on not only markets but AI, presentation, style, curation, creativity. So welcome.
2:48
I'm going to really grill you today.
2:49
Liberty RPF: Thank you for inviting me again, Jim.
2:52
I guess I didn't make the blacklist for the first time.
2:55
Jim O'Shaughnessy: But here you are yet again.
2:58
So my first question, and I got to make sure that I don't use your actual Christian name here, so I've just got to think Liberty, Liberty, Liberty, Liberty. Okay.
3:07
So my first question is around curation.
3:09
One of the things that drew me initially to you was I found myself, I had yet not yet subscribed to your Substack, but I found myself, I almost got this weird feeling if I hadn't read it.
3:30
And so I would actually go search my email and, ah, I missed one.
3:35
And one of the reasons why I think that my question is going to be to you, is this something that can be learned, like practicing your scales on the piano?
3:46
Because I think of curation as a creative endeavor. What are your thoughts?
3:53
Liberty RPF: That's a really good question.
3:57
I think there's something to the repetition, the iteration of it.
4:00
I think it's part of it may be just having taste like this intangible kind of Steve Jobs thing of some people have more taste than others and you can develop taste in certain fields if you spend a lot of time in them and learning about them.
4:14
And after a while I guess you've seen enough that you can recognize what's really interesting in it and what's just kind of like the day's BS that's going to be a fish wrapper tomorrow.
4:22
I feel like, I don't know, that may be part of it.
4:25
But the other part is that so much of society today is so specialized, because that's the dominant strategy to get somewhere, to advance in your career to, you get a degree in fruit flies, genetics, or something.
4:38
And you have to spend years and years to get at the top of that field, and it doesn't leave you a lot of time to explore other, even things right next to it, but things very far away like the arts and I don't know, some parts of engineering about building bridges or something.
4:55
And my own wiring is very opposite to that.
4:59
It's very of a generalist wiring.
5:02
I can't spend all my time in one field. I'd get bored.
5:05
There's always something shiny over there, over elsewhere.
5:09
And I feel like the only advantage I can have over there is to do the 80 20 rule and try to expand my 20% over here and there and there, and build up just enough knowledge in a bunch of fields that I can try to connect them, find something at the intersection that other people may not have seen.
5:24
I think in my case that's the kind of curation I do.
5:26
I'm sure there's specialists that do curation in a very specific field and that's different.
5:31
But in my case it's just purely powered by curiosity and just being bored quickly with just the same thing over and over again. Jim O'Shaughnessy: Yeah. So I agree with that.
5:40
What is your read to publish ratio?
5:44
So how many articles or things do you read versus end up putting, getting put into your Substack?
5:52
Liberty RPF: That's a good question.
5:54
It feels like some weeks...
5:57
It feels like I'm down a particular rabbit hole, that's a very rich vein.
6:01
Like lately I've been into generative AI, which I'm sure we'll talk about.
6:05
To me that's a very rich vein.
6:06
I throw a tennis ball and I find something interesting in there.
6:09
I keep finding more and more interesting stuff.
6:12
So during that time, the read to publish ratio is pretty high because I keep finding good stuff.
6:18
Other times I'm not sure what I'm looking for.
6:20
I find stuff once in a while.
6:21
It's like all of the old film is on the cutting room floor and I keep searching for the good parts and it's harder. So I think it varies. I think it's cyclical.
6:30
I feel like this, when you get into a real flow state on something, some idea interests you enough that you can spend weeks just immersed in it, during that time, you can learn more than into the next six months where you don't get that same vein.
6:45
So on average is probably kind of low, but during certain weeks it's super high.
6:50
Jim O'Shaughnessy: Yeah, I am a fellow rabbit hole jumper as you know.
6:55
And it's interesting because we mentioned before we started recording that my new company O'Shaughnessy Ventures is the worst kept secret of any new company that was meant to be in stealth mode ever.
7:08
And you made a funny joke about Steve Jobs telling everyone about things that he had previously said, don't tell anyone.
7:16
And then he did himself and you said it was leaking from the top.
7:19
And I'm like, well, yes, the good ship O'Shaughnessy Ventures is definitely leaking from the top.
7:26
And one of the things that we are trying to do and endeavoring to do is, essentially the so what of it is I truly believe we are going into a pretty incredible period in human history.
7:42
And I think that the tools available to people continue to stun me.
7:48
And now that I'm involved with Stability AI and I've seen what's coming down the road over the next couple of years, it is literally and legitimately mind blowing.
8:01
And so I get caught up in those rabbit holes, and luckily I have the luxury to do that.
8:10
But a big thing that we're going to be doing at O'Shaughnessy Ventures is, you know the Thiel fellowships.
8:14
And there's a bunch of these, right?
8:16
We're actually basing ours on the mythical Atlantis.
8:21
And one of the things that I was reading, because by the way, if you really want to go down a rabbit hole, type in Atlantis mysteries into Google or your preferred search engine.
8:32
And there was a great book that talked about how they sent people out, scouts, if you will, to all corners of the earth.
8:40
And they have these very antiquated names.
8:43
So my folks at O'Shaughnessy Ventures have updated what we're going to be doing, but we're going to be, this idea is going to have 12 Atlantian scholars and we're going to fund them.
8:53
And it's still in the thought process phase, but I think your future given your skill set is just going to be so amazing because of your versatility.
9:11
I think that I completely agree on the whole specialization in the sciences, for example, in medicine, in engineering, of course, yes, of course I want people who really know how to build a bridge.
9:29
But I think that we are going to see also an awakening of creativity.
9:34
And when I say creativity, I mean in all fields.
9:38
I'm not just talking about art.
9:39
I'm not talking about just generative AI.
9:43
But entrepreneurs are creative, musicians are creative.
9:48
I view creatives as a much larger group of people.
9:52
And so our basic thing is we're going to find them.
9:54
We're going to fund them.
9:56
We're going to hopefully do business with them, plug them into our network.
10:00
But it leads me now to the big topic of AI, because you and I have been having a really great chat.
10:05
And I felt really bad and I actually DMed you, because I used what you said when we were talking about the various use cases for AI specifically in medicine.
10:15
So now I want you to be able to [inaudible] for them and outline the three that you gave me, because I think they're brilliant ideas.
10:22
The people that I've spoken to in my group in AI think they're brilliant ideas.
10:28
But I'd love it if you would outline them for our audience because they really are insightful.
10:33
Liberty RPF: Oh, thank you.
10:35
I'm just happy that the ideas are there because as you say, I feel like for so much of humanities history, execution and having the idea were so tied together.
10:48
The person needed to have both.
10:49
And now with many of our so powerful tools, it's kind of becoming a bit disconnected and you can have the ideas and then have them executed by software, by a machine, or somewhere else, or OSV may be the execution machine for these people that have the ideas but don't have the capital or the tools or access or whatever to execute them.
11:07
So bringing those things together is amazing.
11:10
About the ideas I had, I think I remember two out of the three, so you may have to jog my memory for the last one.
11:15
But the first one I was thinking about is there's been a big scandal basically about fraudulent research about Alzheimer's recently.
11:22
And it made me think about how there's so much, thousands and thousands and probably hundreds of thousands of studies in every field that will never have enough people to go back and go through them and go with a fine tooth comb trying to figure out if there's maybe some somewhere if there's some good faith errors or some outright fraud.
11:41
But with machine learning, I think we could data mine these things and try to find all kinds of stuff that we could never have found before.
11:49
And we may figure out that some branches of sciences have been going in the wrong direction for a long time because they're basing their current research on some bad foundation somewhere, that the house is built in a foundation on sand or something and they're wasting so much time and money and effort and that has real consequences.
12:06
If the Alzheimer's research spend years and billions of dollars going after emulate plagues, plagues or something because of some fraud, that's terrible.
12:15
People suffering from this disease should have research going in the right direction for them to find a cure.
12:22
So yeah, I feel I, that's one of the top uses I can think of for this kind of AI.
12:26
Not that it's an easy things and maybe it's just probabilistic.
12:30
You flag some stuff as potentially to need human regulate, let's say.
12:34
The other one I wish I could see, and that's probably a bit farther down the line, when you can model things in silico better, I've computational models and in biology, we're already starting to do that.
12:42
Google has kind of cracked a lot of computational protein folding stuff and eventually we can have more complex models where you can take past experiments and rerun them in silico to try to see if they replicate.
12:56
If you had to do it kind of in the real world, it would take forever and cost billions of dollars and you could maybe share, pick a few studies that you could try to replicate and a few needles in the haystacks.
13:07
If you can do it at scale in AI models basically, that's another area where you could figure out if there's a replication crisis in that part of it.
13:19
But also you can rerun the same studies with slight variations to try to maybe optimize them if you had a good result on some study.
13:28
But the people doing the study have only this much funding and they can only try, I don't know, 500 variations on that compound or on those animals or whatever.
13:36
Well, maybe if you rerun it much more cheaply and quickly, you can run 500,000 variations and find a much better, one more optimized where you can improve procedure for drugs or whatever.
13:48
And I don't know if that's the third one I mentioned to you, but another one that I'd love to see is there's, science is, it's very about prestige and you want citations, you want advanced career.
13:58
So everybody wants to work on the most prestigious and the sexiest stuff.
14:02
There's a line I love by a scientist who said, it seems tragic to me that all of the top scientists and engineers want to work in the fields where they make the least difference.
14:11
Where if they were hit by a truck five minutes later, someone else would come up with the same thing.
14:16
I wish we'd have more effort going into less prestigious areas, areas where there's less competition from the top minds where you can make more discoveries, but also where you can find a bunch of new hypothesis, right?
14:27
Where you can find a bunch of stuff that doesn't work, and having this information about what doesn't work is still useful.
14:33
Well, the next person working the field, if they have huge database of a million failed experiments, they can much better target what they want to do in the future or maybe just avoid doing something expensive and it takes a long time to solve resources, avoid wasting resources is just as good as having more resources.
14:50
So I wish AI could help us with those kind of less sexy parts of science.
14:55
That's another one where in some fields it's going to be easier 'cause they're more based on information and data and can be done in software.
15:00
Some other fields going to be harder.
15:01
But I feel like over time, because of our good friend Claude Shannon, basically anything in the world can be represented by information and you can act on it in that information realm.
15:13
It may not be easy, but at the rate at which things are improving, it's definitely going to be possible at some point. Jim O'Shaughnessy: Yeah. You nailed all three.
15:20
And when I was discussing it with these...
15:23
I love my new circumstances because I've assured that I'm always, almost always the stupidest person in the room when I'm having these conversations.
15:31
Liberty RPF: That's the way to be.
15:32
Jim O'Shaughnessy: And, oh, I love it, because it's just like in awe of watching these minds work.
15:37
The forensic accounting example of previous studies, I think is brilliant.
15:45
Because we know about, for example, I always use Ancel Keys at the University of Minnesota.
15:52
He was an, he's now, he's dead, but now we know that he faked all his numbers.
15:57
Liberty RPF: But how long did it take and how long did his influence have in the mainstream view of nutrition?
16:02
Jim O'Shaughnessy: Exactly.
16:03
Liberty RPF: That's the thing that there's so much inertia. Jim O'Shaughnessy: Yeah.
16:06
But also from my point of view, I'm willing to go further.
16:09
How much destruction did that one man cause by lying because he had this pet theory that failed our hero Feynman says no matter how beautiful your idea, if the experiment fails, you are wrong. Liberty RPF: Yeah.
16:30
Nature of the final arbiter. Right?
16:32
Jim O'Shaughnessy: Exactly.
16:33
So this one really I'm passionate about, all credit to you.
16:38
But as I think about it, just there's got to be, actually I love to bet on things.
16:43
So we're setting the bar and I set the bar at, first at 25% and after some conversations it's inching its way up to 30% that I believe are going to, when AI does the forensic accounting, are going to be either outright fraud where data that was inconvenient was removed.
17:03
We know about those things too.
17:05
And or legitimate, or sorry, honest minded data mining going, not attempting to fool anyone but fooling themselves.
17:14
Again, Feynman, the most important thing is to not be fooled and you're the easiest person to fool.
17:21
And because if you're passionate, again all the things that make we humans so cool also can be real problems.
17:29
If I'm passionate about, I don't know, bone broth, I'm doing a bone broth experiment right now and the research comes back that, yeah, go ahead and do it but it doesn't do any good.
17:41
It is just our human nature to seek out the information that says, no.
17:45
Bone broth, Kobe Bryant says bone broth was the fountain of youth.
17:51
So the other thing that I understand now better about AI that I didn't before is the, we, humans, another thing we have trouble with is grokking exponential anything, right?
18:06
So there's a great quote, I can't remember whose it is, but it's something like, the problem with human beings, and Shannon of course knew all about this as well, is one, two, three, four, a lot.
18:18
The phone numbers are seven digits.
18:20
We are not configured to be able to comprehend literally millions, hundreds of millions of data items, whereas AI has no problem with that.
18:34
Liberty RPF: You just gave me an idea for new things that I wish AI could do.
18:39
What if, because one of the hard things for scientists is that they have to be an expert in, I don't know, the livers of mice, and then they also have to be great at metabology and statistics and understand all the probabilities to power the study correctly, have the right sample rate, the right control group.
18:57
This part of it is also very difficult and it's hard to be an expert at both.
18:57
So I think a lot of mistakes come because the scientists may not understand properly the right methodology to have the study say what they wanted to say.
19:09
And I know that they can get help with that and they have other experts that can consult.
19:13
But what if you could have an AI that you run the study through first and it tells you if the methodology is correct, 'cause it happens.
19:20
Peter Attia wrote, writes about this all the time about how, okay, the headline in the New York Times or something is like, oh, vitamin D does nothing.
19:28
And then okay, you go look at the study and they gave like 400 units to [inaudible] and then they, self-reported, they never took blood samples. And that's the study.
19:35
It's like, this is a terrible study.
19:36
It doesn't say anything, right?
19:37
Jim O'Shaughnessy: It's awful.
19:38
Liberty RPF: It doesn't say anything, but people read the headline.
19:41
They stop taking their vitamin D and they probably are worse off because of that.
19:44
Because vitamin D is known to be extremely, toxicity is very, very difficult to get with vitamin D. It's very inexpensive.
19:52
So the downside is basically nothing and the potential upside is pretty high.
19:55
So it's a good asymmetric bet to make, even if you don't have super great studies about it.
20:00
But if the studies are bad too, that's a huge disservice to everybody. So that's another thing.
20:05
If you could have this AI service for scientists where it's like, okay, I've designed my study.
20:10
Either you write it down in natural language like LLM, right?
20:15
Large language model style like GPT-3 and the service can understand it.
20:18
Or you have some other software where you have to enter it more formally.
20:21
And then the AI can spit back out, okay, for what you're trying to do, you would need a bigger, I don't know, sample size or this or that or your probabilities over there are you've made a math mistake basically.
20:32
A bunch of scientists would probably love to have that kind of stuff.
20:34
The other thing I want to say as an aside, because we're talking about Claude Shannon and information is about around the same time this podcast should come out, I'm doing a podcast about Claude Shannon with Jimmy Soni who wrote a biography.
20:46
Jim O'Shaughnessy: Oh, Jimmy, I know him well. Yeah. Great guy.
20:50
Liberty RPF: And David Senra, both of them on the same podcast with me.
20:52
So we're trying to bring... Oh, wow.
20:54
Jim O'Shaughnessy: Claude Shannon a bit in the spotlight.
20:55
So I think you'll enjoy this one. Oh, I love that. I'm a bit jealous.
20:58
Those two guys are two of my favorite guys.
21:01
So you're going to have a blast.
21:02
Liberty RPF: Likewise, great friends.
21:03
Jim O'Shaughnessy: So staying on the AI, I think that's a brilliant idea.
21:08
And one of the things that I do is I just generate use.
21:13
I can't help it because of what I did for 40 years essentially was I generate algorithms in my head all the time trying to figure things out.
21:24
And so now I've seen it change over to generating use cases for AI.
21:30
And people like you are just, I envy because you're so good at just there.
21:38
That's a fantastic use case that you just came up with.
21:40
Liberty RPF: Well, you gave me the idea. You gave me the idea. I'll credit you.
21:44
Jim O'Shaughnessy: But the other one that I really like very much is the null hypothesis.
21:49
We at OSAM a keep what we call a research graveyard, which is essentially the same thing.
21:55
We don't dispense with the studies that are not, or proved to have no efficacy.
22:01
We keep them because we want to see how they were structured, what was involved, everything else, so that our researchers go to that library which is growing and growing and growing.
22:14
And I mean I had one guy come into my office a few years ago and say, I love this data set because you just saved, just because of that, I'd saved a month of my time.
22:27
He said I was bummed out when I came in because I thought I had this really great idea.
22:32
And he goes, and then I go to the research graveyard and I'm like, oh, shit. Liberty RPF: Yeah.
22:38
And it's a different application of the principle.
22:42
Someone like Munger would say about life in general, you learn from other people's mistake, hopefully avoiding some of yours.
22:48
You make different mistakes, but at least you're more hopefully do the same, right?
22:51
Oh, this person drinks too much, maybe that's not a good idea.
22:54
This person has too much debt.
22:56
Well, that's a good way to go, bro. Right?
22:58
In science, it should be, or quant or whatever.
22:59
It should be the same thing. How can we go wrong?
23:01
Let's learn about that stuff before we attempt something else.
23:05
And too many people try to shove that all under the rug and pretend that they're batting average is a thousand.
23:10
And if nobody sees it, it's like it doesn't exist.
23:12
Well, I don't know, there's good information you're throwing in the trash over there.
23:17
Jim O'Shaughnessy: Totally agree.
23:18
And I love this idea of good information because one of the things that excites me the most about AI is that I see literally, and I got to be careful because I get really excited about this.
23:31
And in the Deutschian sense, I'm very optimistic, but in the David Deutsch sense.
23:39
In other words, I have no illusions that there aren't going to be big problems that are going to be secondary or tertiary effects from the primary effect.
23:50
Look, I can't remember whose quote it is, but it's like we discovered fire.
23:55
And after we discovered fire, we started cooking our food.
23:56
And our brains got so much better and bigger, but fire was also a little bit dangerous.
23:59
And he goes, "So we invented fire alarms, fire departments, fire warnings," all of these things.
24:04
And so one of the things that I want to kind of shift gears to now, still on AI, is...
24:11
So I've collected contemporary art almost my entire adult life.
24:16
We have this big thing where we like to, if we can, buy living artists because my wife is a expert at art history, and she was fascinated by the idea of patronage and by being able to affect someone's actual life, and we've had some wonderful experiences where that is the case.
24:36
But one of the things that...
24:38
So I've been talking to a lot of my artist friends, people we collect and things, and I'm not surprised that some of them are really unhappy about generative AI.
24:53
And I always try to reposition it with them.
24:57
I have up actually today, on my two thoughts from, a quote from Degas, which is, "Art is not what you see, but what you make others see." What do you think?
25:08
Liberty RPF: I saw that one in my inbox this morning. It's good.
25:11
I wrote a bunch about this, and I feel like there's kind of the political solution and there's what I think is the more, I don't know, my own solution.
25:22
So the political situation is probably, "Well, lots of people are going to be angry, so put a bunch of money in the pool over there and give money to people, and they're less angry.
25:31
And then you keep doing what you want to do." I guess that works.
25:35
But it kind of sets the precedent that people are right to look at it this way.
25:41
And I don't think they're right.
25:42
I'll back up a bit because it's abstract, but a bunch of artists are saying, "Well, this model that's creating this art has been trained on my work and on people's work, and so we should be compensated for that."
25:51
And I feel like if I generate a picture of a copyrighted Disney character with the model and I try to sell it, I should be stopped on that side of it.
26:00
Disney has the right to this, we have laws.
26:02
But on the learning side, I feel very, very uncomfortable with trying to restrict learning, because we wouldn't do it for humans.
26:09
You wouldn't tell humans like, "Oh you're a painter, but you can't look at the paintings of anyone that's copyrighted.
26:13
You have to learn only from un-copyrighted works."
26:16
Or, "You musicians, never play a cover of the Beatles or of Nirvana."
26:23
Every artist learns from what comes before them and then they kind of remix it all in their heads.
26:27
And for a while, they probably just copy. That's imitation.
26:30
That's a great way to learn.
26:32
And when you imitate great stuff, at first, it's the only great stuff you're going to do, right?
26:37
It's not going to come from you right away unless you're kind of a freak of nature.
26:41
So that's the way to learn.
26:43
I feel like the AI is learning the same way.
26:45
And one thing that I think people misunderstand is the AI is going to learn on billions of images, but the inference model at the end that generates the images, doesn't contain those images.
26:57
It has been trained on those images, but what it keeps is the learning is under the form of parameters.
27:04
It learned that Mickey Mouse has these 5,000 parameters of this shape and these edges and these colors and these textures and all that together is Mickey Mouse, right?
27:13
Otherwise, if it was just learning from single images instead of from like, okay, I'll back up again.
27:21
A human can recognize Mickey Mouse from any angle.
27:24
We don't just learn to recognize them from the front and then you turn five degrees to the left and then what is that?
27:29
No, what we've learned is not the static image.
27:31
What we've learned are the patterns. Well the AI is the same.
27:34
It learns a bunch of patterns and when you ask it to generate something, it'll use those patterns and that's why you can mix the style of three different painters, then it'll mix the patterns from those three and output you something, right?
27:46
So I feel like this learning on copyrighted material, it's not even in the model that's used to generate stuff.
27:53
So it's hard to say, well you took my image and you put it out.
27:56
It's not even in there, your image.
27:57
What I've learned is the patterns of your style of painting.
28:00
So I don't know, there may be something unethical there, but I have trouble seeing it because I don't think we apply it to humans.
28:07
I also feel like over time the AI is going to learn more and more from images generated by AI.
28:14
And I don't know if that's good or bad.
28:15
I don't know if it's going to dilute the pool or increase the variety or what direction it's going to take humanity and art.
28:21
But I feel like there's kind of like a AC and BC line where before this year almost everything online was generated by humans and you train the models on it and then in 10 years you're going to train the models on a pool of data that's going to be like 80% AI generated or something. And is that good, bad?
28:38
That's another thing to worry about in some ways, right?
28:41
Because there's not going to be good only good actors in there.
28:45
There's going to be all kinds of spammers and scammers and propagandists and all that kind of stuff.
28:49
So yeah, I'll leave it there for now.
28:52
Jim O'Shaughnessy: Yeah, I think that's very insightful.
28:55
And one of the things that I try to do is just help them reframe.
29:00
Because I also have a lot of artists that I collect who are the most enthusiastic about what's going on with generative art.
29:10
I was talking to one and she basically said, "Oh, I am in love.
29:15
I put my own images in, the AI learns on them and then it presents myself back to me."
29:24
And she was basically saying, "I have come up with some of what I think is my best work by working with this iteration."
29:34
And she said, "I don't have any feeling that the art that I put out at the end, which I interacted with the AI on, is anything but my art."
29:46
And she goes, "So I look at this as just really and incredible tool for me as an artist."
29:56
And so I'm hoping like we both know, because we're both students of history, whenever something new comes along, be it like a new form of entertainment, like movies or radio, people freak out, this is going to ruin the youth.
30:12
And you know, can go all the way back to Aristophanes and his plays in ancient Greece and same things were being said.
30:21
And when the camera came out, when the camera came out, portrait artists were wearing black arm bands.
30:28
But that over time they get tuned in, oh wait a minute, this might actually be a really interesting new tool.
30:36
Or maybe this frees me to not try to have to do an exact likeness of this person I'm painting and you get Francis Bacon and you get people like that.
30:48
And so I look at this as just like we are continually evolving and that's a good thing.
30:55
And I think, yeah, we don't always get it right, we often get it wrong.
31:00
And I think you always have to acknowledge that because then you bring error correction into the equation.
31:06
But I think that directionally to steal another line from a friend, the error of progress is up.
31:12
Liberty RPF: I think, yeah, and I think we shouldn't only look at it from the artist's point of view.
31:19
It's a very important point of view, but there's also a lot of people who consume art, who enjoy art, whose lives are enriched by art.
31:29
And so if there's more good art out there because of these tools, that's a very good thing for society, I feel like.
31:38
I have a folder where I save my favorite images created by these Ais, and I have hundreds and hundreds of images there and many of them are some of my favorite images ever, like digital paintings and all that.
31:50
And they wouldn't exist without it.
31:51
And I don't know, I feel like it's kind of the long tail thing, right?
31:55
Oh, because of the internet, there are so much crap out there, but the absolute number of brilliant things that increased and that's kind of all that matters if you have good filters, good curation and so on.
32:05
So I feel like with these tools, you still need to have pretty good taste to produce good images.
32:11
People will just jump in and it's easy to make crap too.
32:15
Some of the best images I've seen, like when I save them in a folder, I get the username of the person and over time I realize that some usernames come up again and again and again.
32:24
And so that person probably has really good taste, really good imagination.
32:28
As I was saying earlier, the execution and having the idea has been kind of disconnected.
32:34
You don't need to be a great painter to paint the idea, but some people have tons of great ideas that they would never have executed otherwise.
32:41
And so these ideas exist now because of these tools.
32:45
Even I feel like, because I've had all kinds of crazy projects with generative AI lately.
32:49
I spent two weeks doing series of images that are digital paintings of orcs.
32:54
And so I have orcs working in offices on cruise ships, on vacation.
32:58
I have orcs and fine dining restaurants, I have cyberpunk orcs with exoskeletons working in spaceship space ports.
33:05
All that stuff wouldn't exist, right?
33:07
Because I don't have the means to produce that stuff otherwise.
33:09
So I feel there's value there.
33:11
I don't want to forget the artists because it's good for the consumers of art.
33:15
I don't like the word consumer, but the appreciators of art.
33:18
But I think these people need to be taken into consideration too, right?
33:24
Jim O'Shaughnessy: I agree.
33:26
And I think that obviously people know where I stand.
33:30
I'm a huge believer in both open AI because I think these powerful programs should not be controlled by a small group of people.
33:41
Because that, in my opinion, Lord Acton, absolute power corrupts absolutely.
33:48
And if we have a system where AI is closed and controlled by a few plutocrats, that's not a good world, it becomes a panopticon.
33:58
And I have lots of grandchildren and I don't want them growing up in that kind of world.
34:03
And also, you make the point quite eloquently, that what we are on the cusp of here is this explosion of new learning and we can't know what's going to come out of it.
34:17
That's another Deutschian thing that I am always quoting.
34:20
We cannot prophesize the future.
34:23
We cannot say what new inventions and innovations are going to happen.
34:28
And so this idea that we should close the books, right?
34:33
Anyone who says the science is settled has never read anything by Shannon, by Von Neumann, Feynman, any of my favorite scientists.
34:46
That's scientism, it's not scientific method.
34:51
So I believe that we're really literally at the beginning of what could be a golden age for human development and any attempts to throttle that or thwart it...
35:05
Now listen, I am fully aware of the dangers that some of these technologies, so for example, one of the initiatives that we are announcing at Stability AI is we are doing a deep fake detection software contest Stability AI is putting $100,000 US up as a prize and O'Shaughnessy Ventures is matching that with another $100,000.
35:31
So there's going to be a $200,000 prize for the winning algorithm or a program.
35:37
And if it succeeds over our hurdle rates, we're going to give it away for free to everyone.
35:43
And so we are aware of the problem, some are more intractable than others, but AI ethics for example, we want to get an open dialogue.
35:56
And I would even love a wiki where people could come and present their ideas.
36:01
It would be moderated of course, because there's going to be a lot of crackpots who come in and say everything, luddites, et cetera, et cetera.
36:09
Liberty RPF: Or even non people, right?
36:12
[inaudible] writing into the Wiki. Don't turn me off.
36:16
Jim O'Shaughnessy: Exactly.
36:18
And so I just think, I share with you the general optimism that is tempered by the acknowledgement that we have to deal with problems that, undealt with, could be very dangerous.
36:33
Liberty RPF: Yeah, that's the thing.
36:35
I think it's probably easy to hear that I'm very optimistic and I'm generally pretty optimistic about that stuff.
36:41
So because of that, I have to remind myself of the dangers and the bad side.
36:46
And I try to take that very seriously.
36:48
I've been interested in AI for maybe, I don't know, 17 years or something like that.
36:53
And a lot of people, it's funny because a bunch of what I used to read about back then is kind of happening now.
36:58
Oh, we can do this and that in 25 years, 50 years.
37:01
And now it's all more quickly than we expected.
37:04
The things I'm worried about are not the small problems that always come with new powerful technology.
37:11
The way I try to separate it in my mind is there's a bunch of recoverable problems where you make a mistake and you figure it out and you fix it.
37:18
And that's always been like that with every technology.
37:21
And people complain about this problem.
37:24
It's like, okay, but the problem we used to have that was fixed by this was bigger.
37:29
There's no good old days for humanity.
37:31
It used to be pretty terrible in many ways.
37:33
We take it for granted now. But that's on one side.
37:36
What I try to keep in mind, and I try to keep it in the conversation as much as I can, is there's also the potential for nonrecoverable problems with AI.
37:46
Because if you think about it, all of humanity's most powerful tools and technologies and weapons, they're all basically IP.
37:53
A nuclear bomb, that's an idea and then we made it.
37:56
But that's the idea that created it.
37:59
AI, as it becomes better and better and you get AGI at some point probably, or even without AGI, you can make all kinds of very, very scary stuff with it too.
38:08
Bio weapons that are synthetic biology that our immune system cannot recognize and you leave it dormant in someone for years before activating it.
38:15
And so everybody has it by the time, I can imagine terrifying scenarios with that.
38:20
So I want to make sure that the people making the AI make humanity better with all kinds of cool tools.
38:26
And then we fix the recoverable problems that as we get them. And that's fine.
38:30
But we always keep our eye on the big nonrecoverable things because as Buffet would say, you can have this lines of great [inaudible] and then you multiply once by zero and even if you're cured cancer and 99 great things, if you have one big nonrecoverable thing that it all didn't matter.
38:50
So that's the thing, I always want to make sure that the people working on this, and I'm sure they do because they're much smarter than I am, but that's a little part that every time I'm super optimistic, I'm like, yeah, but I hope we really don't screw it up too much.
39:02
Jim O'Shaughnessy: Same with me.
39:03
I have exactly the same attitude and I'm lucky to be able to, from being the chairman of Stability AI, I can have some input that I wouldn't otherwise been able to have.
39:14
Liberty RPF: I'm very happy that you're there, yeah.
39:16
Jim O'Shaughnessy: And what I have found, if it's worth anything, I went in a little skeptical of the motivations of what I had seen big tech develop into, say over the last 20 years.
39:34
And I must tell you that I was so pleasantly surprised by the amount of people who didn't just say things, they acted.
39:46
We have some major new people who work for Stability that, I mean people like David [inaudible] for example, who was Google Brain in Tokyo, he now works for Stability AI.
40:00
And he's a great example of a brilliant person who got the opportunity to...
40:10
What we're seeing is we're seeing so many people who are really tech geniuses, coders, thinkers, everything.
40:18
I'm seeing that they want this to, in other words, they don't want to optimize the objective function for revenue.
40:27
They want to optimize it for creativity, for learning, for all of these things.
40:32
So I have been very, maybe I'm just naive, but I'm actually watching this happen with them coming of their own steam and own accord.
40:43
We're not even recruiting them. They're coming to us.
40:45
Liberty RPF: That's a good sign.
40:47
Jim O'Shaughnessy: I think it's a great sign, but I'm totally with you on the existential risk front.
40:55
Liberty RPF: The way I look at it is AI is so powerful, right?
41:01
Because humanity, our only special thing is really intelligence.
41:05
Everything else comes from that.
41:06
All of our tools, all of our systems of organization, that's all created there.
41:11
If we can augment our intelligence and we can have this augmentation on the curve.
41:17
Because human intelligence, it's building on accumulated knowledge like stuff we learn in books, but the basic hardware in our heads doesn't get much smarter over time.
41:27
Maybe we have better nutrition and vitamins and this and that and we've gotten a bit smarter than people from a thousand years ago or something. But it's not advancing.
41:35
But if we have this other intelligence on a curve that's going to get smarter than us and then keep going and keep going, I feel like we've got it made.
41:43
If we just avoid [inaudible].
41:45
It's like Berkshire that's going to keep compounding.
41:48
If it's in a position of not going to zero at some point, all I want for humanity, I don't want even to go as fast as possible.
41:55
I want us to go as safe as possible because if we're safe, the rest is all going to take care of itself eventually.
42:01
We're going to find the cures for all the diseases we're going to cure aging.
42:04
We're eventually, if we have this augmented intelligence that's going to create all these other tools that everybody's going to use.
42:10
And that's all going to build in itself as long as we don't multiply by zero at some point.
42:14
So that's kind of like, if we can optimize for that, the rest is really going to take care of itself, I feel like.
42:20
Jim O'Shaughnessy: Yeah, I agree.
42:21
Liberty RPF: And if we optimize also for not having these tools at the hands of some dictator or somewhere that's going to create this panopticon that you talk about or this totalitarian society like turbocharged by all these tools, that's a scary thing.
42:37
And it's a good thing that free societies create so much more innovation and tech than closed societies.
42:45
And so it's probably going to come out first in these societies than somewhere like North Korea.
42:50
That's a very, very good thing.
42:51
So this part I'm a little bit less worried about though.
42:54
I don't know that it's scary too in places like China.
42:57
Jim O'Shaughnessy: And that's an argument that I have long made.
43:01
If open versus closed, open is always going to win, open is going to win.
43:07
Because most things are complex adaptive systems that require a variety of cognitive styles and- Liberty RPF: Error correction.
43:16
Jim O'Shaughnessy: Error correction and all of these things and where there isn't a massive forbidden list.
43:23
Or in the case of these, North Korea is a great example, that is a country built on lies.
43:31
And so I can't imagine them, the AI that they could develop would just suck because it would be wrong about everything.
43:39
Liberty RPF: They have nuclear weapon but don't have food.
43:45
Where are the scientists there?
43:48
All these regimes, if they can't steal their technology from someone else, it's not going to be homegrown Jim O'Shaughnessy: Exactly.
43:55
Because of their closed nature and complex adaptive systems, emergence comes from the bottom, not the top.
44:02
And if you want to have a top down, our dear leader type society, it's just not crap.
44:09
It's going to be horrible, horrible stuff.
44:13
So I, like you, am of a similar disposition in terms of we must always be vigilant and the more minds working on it, the better as far as I'm concerned.
44:25
And of course it will be moderated.
44:26
And of course there will be people who know more than other people.
44:30
I'm surrounded by all those people every time I get on a Stability AI call. And so that's great.
44:36
But the open nature, cognitive diversity, I think everyone, diversity, diversity, diversity.
44:43
The key diversity that we need is cognitive diversity.
44:47
Liberty RPF: Yes, yes, so much yes.
44:50
Jim O'Shaughnessy: Please.
44:51
Liberty RPF: Oh no, I was just agreeing loudly with you.
44:56
People try to measure diversity all kinds of ways, but at the end of the day, the real thing you're trying to measure really is cognitive diversity.
45:03
And if you have a bunch of people will look different, dress different, come from different places, but they all think the same, that's not real diversity.
45:09
Jim O'Shaughnessy: Not at all. Not at all.
45:12
And that's been a big deal of mine.
45:13
It's like I think networks are going to become vastly more important than they were in the past.
45:19
Because in the past, networks were basically bound by your geography or your school tides or your religion or some external association.
45:32
So you went to an Ivy League because of the networking. Made a lot of sense. It was smart to do.
45:37
Very religious, you get involved with all the things that your church does and those become the members of your network.
45:47
In this new world, the world is my network and I can find talent anywhere and talent can find me.
45:58
And so just my own experience in building out Infinite Loops and now O'Shaughnessy Ventures, we have people all around the world and we don't care what you look like.
46:09
We don't care what sex you are, we don't care who you choose to love.
46:13
We don't care what your religion is.
46:15
All we care about is can you do this great work?
46:19
And then we want to give you that platform to be able to do it.
46:22
And so the network effect which everyone knows about, so we don't even have to talk about, becomes more profound in my opinion, because it allows nodes to come on to the network that in the past had no way of joining that network. Here's an example.
46:42
One of my main guys, Vasel at Infinite Loops, he lives in India.
46:46
He is a very bright young man, pre-internet, pre explosion of these tools, zero chance that I would meet him in any capacity. Zero.
46:58
Liberty RPF: Oh, [inaudible] right? Where would we be?
47:01
Jim O'Shaughnessy: And you, exactly.
47:03
Well, there would've been a slight probability because I have so much business in Canada, we could have- Liberty RPF: I've never worked in finance.
47:10
I've never worked in tech.
47:11
Jim O'Shaughnessy: That's true.
47:14
But the only reason I would give it a slightly higher probability is we- Liberty RPF: Yeah, [inaudible] is a bit more compatible, yeah.
47:20
Jim O'Shaughnessy: We would be in situations where we were physically proximate to each other because I go to Quebec or went to Quebec all the time.
47:28
I go to Montreal all the time.
47:31
So I mean, I could have literally bumped into you on the street in Montreal, but that's farfetched.
47:39
But it could theoretically have happened.
47:42
The chances of that are almost nil.
47:46
So the variety of just my experience with my network, of which you are a member, is just so the synthesis that is happening when I talk to different people in my network is just incredible.
48:03
And because I'm so lucky to have access to guys like you, to have access to guys like Jimmy, who wrote the book about Claude.
48:13
And David with his, I love the guy.
48:17
Liberty RPF: David's the best.
48:18
Jim O'Shaughnessy: He really is. Liberty RPF: N of one.
48:20
Jim O'Shaughnessy: Yeah, he definitely is.
48:22
But my point is, all of these people can now meet and interact and create, right?
48:28
I'm going to come away from this podcast with four new ideas that you just gave.
48:33
Liberty RPF: And I think I may have a new one listening to you, it's half baked, right?
48:37
I think I'm going to have to write about it at some point.
48:39
But in our previous episode we talked a lot about the interest graph, which was kind of this new thing.
48:44
There used to be the social graph and you tried to replicate online your social ties with coworkers, whatever.
48:49
And now there's the interest graph on places like Twitter where people with similar interests and skills can find each other and instead of, as I was saying, instead of gossiping and having fun, they're creating and learning.
49:00
And that's a different kind of network.
49:02
The new idea is that on top of that, there seems to be a new kind of meritocracy.
49:08
And I feel like in the old world, meritocracy was so limited by geography and by these kind of networks that you could have meritocracy inside of a company or in a town with the civic people or maybe in the region or in a longer network of, I don't know, pediatricians or something.
49:26
But it was kind of limited.
49:27
And now I feel like there's two things that overlap, where you have this global network of people with similar interests where the really good people can kind of rise to the top.
49:36
But you also have this infinity of niches where you have all these little mountains where you can get to the top of that niche, right?
49:44
Because if you had this worldwide thing, but it was still the three channels and the middle of the bell curve, there's no space there to rise up to the top.
49:53
So meritocracy would be limited, but if you have the combination of infinite niches and subcategories and subfields with a way of finding the good people in each of those fields, I feel like that's a new thing.
50:04
I don't know what it is, what it's called, but I think that's a new thing.
50:07
So I'm going to have to write about that.
50:08
Jim O'Shaughnessy: I definitely agree that it's a new thing and it's like one of the things that spurred me on to create O'Shaughnessy Ventures, honestly.
50:17
I believe that all the old models are collapsing and that people who understand that and are tuned into the new models are going to be able to do both really well, but they're also going to be able to contribute to the common good.
50:30
And one of the things that you just outlined is a great example of the ability of the common good to grow exponentially.
50:40
And it's this whole idea of the human colossus.
50:43
The human colossus gets smarter and smarter, and then you add AI, and it just keeps getting bigger and bigger.
50:50
And then you destroy all of the bounds of geography, space, time.
50:56
And those are all gone now too and when you conceive of this and you see this network, I tend to see it in my head.
51:05
It's just the flashing lights man are, it's incredible.
51:10
And, so I think everyone loves to joke about Twitter, but I still think Twitter, whoever owns it by the way has the chance to emerge as a true real time worldwide global intelligence network.
51:28
And, does that mean that 90% of it is still not going to be shit posting, "No, it doesn't."
51:34
It leaves room for that, right?
51:35
Liberty RPF: As long as the absolute number of really good ideas and people is higher than without, that's still a real net good.
51:42
Because, all the shit posting, you can ignore it if you want.
51:45
I feel like what immigration to the US used to do, now people are immigrating to the online world.
51:54
Elon Musk, if he had stayed in South Africa maybe a story would be different.
51:59
So many people in the US came from elsewhere and where they were, maybe, there was not this opportunity to build companies or run the lab that did something great.
52:10
I feel like our people now are taking that ambition, that intelligence, that work ethic, that perseverance, and putting it kind of the online world and applying it there.
52:23
Instead of, when you talk about geographic kind of collapsing and you know can work with someone anywhere.
52:29
Only so many people can immigrate to the US and come to somewhere where they're freer to do what, attempt to do what they want to do.
52:35
But, if the online world can stay open enough and full of opportunities, it could be that kind of place too.
52:42
Jim O'Shaughnessy: Totally.
52:43
Liberty RPF: It's pretty big in the real world, but it could be additive. Jim O'Shaughnessy: Yeah.
52:46
And, it's much easier to change your digital zip code than to change your actual one.
52:52
And, so this freedom of movement and of ideas and of concepts is unprecedented in human history, it is unprecedented.
53:02
And, I'm just giddy over it in terms of what can come out of it.
53:09
Again, being very mindful that we always have to keep an eye on that existential risk over there.
53:16
But, yeah, I'm extremely bullish on it.
53:17
I want to switch gears here, because you are also, and I know this because of all of our private chats and DMs and whatnot.
53:26
You're very, very creative.
53:27
There's a great quote, I can't remember who said it, I just jotted it down for coming on with you.
53:32
But, "Curiosity is a shit starter."
53:35
And, one of the things that we're developing as a campaign for O'Shaughnessy Ventures.
53:39
Is that, this first part of it is going to be aimed at creatives.
53:43
Now, I think of creatives in a much broader category than many people do, I think somebody who's a founder of a company that's a creative.
53:52
I think that a scientist understanding new disease and or whatever, that person, he or she is a creative.
54:01
But, in the more conventional sense, one of our branding guys, during one of our Zooms said, "Man he's a creative and he's a creative director."
54:13
And, he goes, "Man creativity is such a fucking curse."
54:18
But, everybody lit up, everyone was like...
54:22
We have another guy who's doing the analytical side, he goes, "But, no, it's a magnificent obsession."
54:28
So, we're actually running with this idea, because I talk to a lot of my friends who I consider like the most creative.
54:38
And, do you think that creativity is a curse? "Oh, my God, yes."
54:42
And, then, when we talk about it, not a curse in the, I curse you sense.
54:46
But, in the sense that it just compels you, that you just can't help yourself.
54:53
Liberty RPF: You can't turn it off, right?
54:54
Jim O'Shaughnessy: Right.
54:55
Liberty RPF: Your biggest strength is usually your biggest weakness.
54:58
So, sometimes you wish you could focus on one thing or just be linear about something, but then you're all over the place.
55:05
And, yeah, I think I understand what you meant. Jim O'Shaughnessy: Yeah.
55:08
But, as a creative tell me about a time where you had the aha moment, thinking "Oh, my God, why am I," whatever you happen to be doing.
55:19
"Why do I have 500 images of orcs in my folder?"
55:24
And, how did you overcome it or take advantage of it, is what I'm interested in?
55:32
Liberty RPF: That's a really good question, I've never really thought about it. I guess it's a...
55:38
Introspection is a skill that has to be also used and sometimes you use it on certain parts of your life and not on others.
55:45
I think it's funny, because I used to play music, I know you're also a huge music fan.
55:52
I used to do all kinds of things, and then for a while I was just investing and all of my notes were kept private.
55:58
And, I also used to write in public a lot, like 10 years ago, and then for a while everything was private.
56:04
And, I was talking to a friend of mine who's a novelist and he kept like, "Oh, don't you miss it, don't you miss creating?"
56:10
And, I'm like, "Oh, no, I'm creating a ton, I'm writing tons of words in my notes."
56:15
And, it wasn't until I started the Substack on a whim, as a hobby, it wasn't supposed to be anything.
56:22
And, then, I realized, oh I really did miss it.
56:24
I really missed an outlet that's public where you can put stuff out there.
56:28
And, it's not the same when it's public as when it's private.
56:31
And, it all feeds on itself, because you do one thing and you get feedback and it gives you an idea for the next thing.
56:37
And, someone says, "Oh, this reminds me of that," and they point you to someone else who did something similar.
56:42
And, once the machine is started you don't know where it's going to go, it's kind of a self-directed thing with 15 hands on the wheel.
56:50
You are trying to hold on, but other people are like, "Come this way, come that way."
56:54
And, I feel like the newsletter has been that realization, that I kind of missed it in my life.
56:59
I missed even stupid ‘Deadwood’ jokes or whatever, that stuff has to come out of me somewhere and I needed the outlet.
57:05
Jim O'Shaughnessy: So, what I hear you saying is, dare to take the step, dare to put your stuff out there in public.
57:15
Because, cumulatively what's going to happen is what you've just described, I think.
57:19
If you have good ideas like you do, and you put them out in public, what's going to happen is your ideas are going to get better.
57:26
Because- Liberty RPF: That's the thing, you start with bad ideas, but then you repeat and repeat.
57:30
And, I wrote for 10 years and did different job every day and you have to get those words out of your system.
57:36
It's going to be shit for a long time and then it starts to get a little bit better.
57:40
And, then, even in this project, if I count everything I've done, maybe 400 editions in two and a half years, that's a lot of iteration.
57:48
And, if I go back and look at the early stuff it's like it burns my retinas, but you have to go through that stuff.
57:54
And, now I'm sure that in two years I'm going to look back at what I'm doing today and it's going to be like, "Ah, that really sucked."
58:00
But, that's the good thing, if you keep looking back and finding that you sucked it means you've probably improved.
58:06
Maybe, you still suck, but at least you've improved.
58:08
If you look back and you're like, "Oh, that was brilliant, that was a genius."
58:11
Either you really are a genius or you're just fooling yourself, and you're just stagnating and you're not improving.
58:16
Jim O'Shaughnessy: Yeah, I agree.
58:18
And, that seems to be one of it.
58:21
It does not surprise me, your answer at all, because that was one of the things.
58:24
As we were batting it around, it was my same creative director guy by the name of John Foster.
58:32
He was like, "Yeah, but look at what we're doing right now."
58:35
We are iterating back and forth, people who think very differently than I do, John, who's our creative director.
58:42
And, we're coming up with a better version of this.
58:46
And, so that gets us back to well if I can collaborate with people everywhere in the world I'm going to be able to have access to ideas that would never even occur to me.
58:59
And, so I'm very bullish on that, but it also becomes that sort of, first you got to take the first step.
59:06
Because, like- Liberty RPF: I feel there's huge leverage to encouraging others to do it.
59:12
Part of what I like about my newsletter is I'm mostly directing some attention to what other people are doing.
59:18
Like, "Oh, I found this great stuff over there, I found this great study over here.
59:21
I found this great Substack, this great writer."
59:23
And, I try to highlight that stuff, that's why the title is Liberty’s Highlights.
59:28
And, I can only produce so much myself, but by trying to highlight the stuff of others, well some of them would do great regardless.
59:37
But, a few of them they come to me in the early stages and like, "Oh, I have nothing to say."
59:42
It's like, well bring people along with you, a lot of people are at the same stage as you are.
59:48
They're just learning, so you can learn along.
59:49
There's always a version that you can put out there that's going to benefit someone.
59:53
That's going to be useful to someone, or that's going to be useful to you.
59:55
Because, you're going to grow as you do it, so just encouraging people to do it.
59:59
And, OSV, that's another way to lever, where one person can do a bunch of stuff themselves.
1:00:05
But, if you can provide the tools and the capital or the distribution or the attention or whatever to others, that's real leverage that creates a ton of value in the system.
1:00:14
And, having these people with the taste and the kind of distribution.
1:00:18
Even Ben Thompson has started doing more of that, where it's not only Ben Thompson now.
1:00:22
It's Ben Thompson interviewing people, it's Ben Thompson having podcasts with other people that hosted by others, that he's not even on, but distributed to a channel.
1:00:30
These people can generate tons of value as kind of nodes in the network.
1:00:35
Jim O'Shaughnessy: Totally agree, and you've just outlined the business model of O'Shaughnessy Ventures.
1:00:40
It is, we now live in an era where I can meet other people who we can have a win-win relationship.
1:00:50
And, that's another big, I mean part of our investment thesis, because we have different verticals so we're branding them all.
1:00:58
Because, of Infinite Loops we're branding them all, Infinite Films, Infinite Media, Infinite Adventures.
1:01:05
Infinite Adventures is investments, because venture capital was originally called adventure capital.
1:01:11
Which, I like so much more.
1:01:12
Liberty RPF: That's good.
1:01:14
Jim O'Shaughnessy: But, our litmus test is, is this an asymmetric opportunity that B presents a win-win opportunity for everyone involved.
1:01:25
And, if it fails either one of those tests we won't do it.
1:01:30
And, so it's really interesting, because you find yourself.
1:01:34
I'd never really kind of thought about it that way, you know what I mean.
1:01:39
And, so I went back and I went through, because we've been doing private investment for a long time.
1:01:46
So, since the early 2000s, because we had, all of our assets were in long world equities.
1:01:53
Which, worked out fine, but you wanted to diversify.
1:01:56
And, so I went back and I started looking and sure enough at least 80% of them meet both of those criteria.
1:02:04
And, so for example, just off the top of my head, permanent equity that's Brent Beshore.
1:02:10
My son Patrick raised Brent's first fund for him, we have been big investors with Brent.
1:02:17
And, how is that a win-win opportunity?
1:02:20
He's reinventing private equity, which, in many respects is not the greatest business model.
1:02:27
And, he is turning it into a vastly better business model.
1:02:31
And, all of those people who would've lost right now, if they're working with Brent they're winning.
1:02:39
Because, his model's completely different, his model isn't to get rid of assets that are non-performing.
1:02:45
And, put the lipstick on the pig and lever up five to six times and then sell it to someone else.
1:02:53
I don't see a lot of value add there, but his model is tons of value added.
1:02:57
Another one that we're invested in is Matt Clifford's Entrepreneur First.
1:03:02
And, talking about a win-win, it's like he looked at the venture model and said, "This is broken."
1:03:10
How can we ever get the sense for somebody if it's one and done, or if we go through, we comb through the data room and check all the boxes.
1:03:21
We don't really know, we need to spend actual time with that person, seeing how they think, all of this.
1:03:28
And, he came up with this idea that he co-founds, he and his partner at Entrepreneur First are co-founders with other founders.
1:03:38
They put them on a clock and they give them time and it's a really fascinating improvement to that model.
1:03:52
So, I'm lucky that a lot of the things that we already had invested in followed this rubric of win-win.
1:04:00
Because, I also believe that we are going into an era of different names, scarcity is leaving.
1:04:11
And, because of that plenitude, human beings have never experienced plenitude other than rich plutocrats, most humans haven't.
1:04:24
And, you looked like you were going to say something really profound, so I'm going to let you say it.
1:04:29
Liberty RPF: Oh, I doubt it's profound.
1:04:30
But, I like, you're saying a lot of my favorite words, win-win and abundance.
1:04:35
And, there's a few mindsets, I wish we could try to inject into the zeitgeist a bit more.
1:04:42
And, these are some of them, just being a rational optimist about things.
1:04:45
Not being a blind optimist, but rational optimism, generally leads to better places.
1:04:49
Because, you're trying to picture a better place and then you can build it, if you don't even think it's possible you're never going to get there, it's self-fulfilling.
1:04:57
And, a mindset of abundance is like the difference between, like okay we have to I don't know, do these environmental things.
1:05:05
Like, let's try to maintain a little bit of what we have, let's try to replace a little bit of the old thing with the new things.
1:05:10
And, it's going to be okay, it's like, "No, I want abundance."
1:05:13
Build a thousand nuclear reactors and make tons of electricity.
1:05:16
So, we can desalinate waters and have jungles growing in the deserts.
1:05:21
And, I want a grand vision for a better world, because if you don't even have the vision you're not going to have the world.
1:05:27
You have to start with the ideas.
1:05:28
When I said that every dangerous weapon is IP first, well every great thing that we build is IP first.
1:05:35
And, people don't even dare to dream that IP right now.
1:05:39
People like Elon Musk, yeah he's really smart, but mostly he has grand vision.
1:05:44
I want a bunch of other people with different grand visions than Musk that can execute on them, there's not enough of that.
1:05:52
There's tons of stuff that we discovered in Year X that could have been discovered 50 years before or a hundred years before.
1:05:58
Some stuff depends on some technology, you need the microscope before you can do X, but other things are basically just someone that think of it.
1:06:05
Well, I think there's a bunch of these ideas laying on the floor right now.
1:06:08
We have all the pieces required we just need the people to come up with them, to dare to dream them.
1:06:13
I don't know, I wish we had a bit more of that stuff in the zeitgeist.
1:06:17
People are depressed about the future and I get it, there's a bunch of scary stuff.
1:06:21
But, the way to turn it around is not by giving up.
1:06:25
Jim O'Shaughnessy: Totally.
1:06:27
Again, I think one of the reasons we get along so well and enjoy talking to each other so much, is we share a lot of the same ideas.
1:06:34
And, we're both in our own way trying to put a little dent in the zeitgeist.
1:06:41
And, I think as long as we're persistent, yeah we'll have losses for sure, but being persistent in this idea of vision.
1:06:52
Because, that is something else I am really obsessed with.
1:06:56
Everything we see, everything, even when I was on safari in Africa, everything you see came out of the mind of a human being.
1:07:06
The only thing that didn't come out of a mind of a human being is if you could find some piece of land that a human being had never set foot on or changed in any way.
1:07:21
It's like we didn't leave the stone age because we ran out of stones, we left the stone age because we got better ideas, we got better explanations.
1:07:31
And, as these better explanations flourished, we flourished and so on.
1:07:38
And, so the idea that, "Oh, that I want to be resoling and return to nature."
1:07:45
I don't, I tell you what, if somebody said to me, "Jim, there's this time machine you're the only guy who can go through it.
1:07:54
And, you can only go to the past, where are you going to go?"
1:07:57
And, I'm like, "I'm staying right here brother," because like...
1:08:01
I was reading, there's a great book, Bill Bryson's ‘America One Summer, 1927’.
1:08:06
And, he's talking about how deadly even that seemingly modern world was.
1:08:13
Like, Coolidge, when he was president his son died from getting a splinter when he was playing tennis.
1:08:21
Liberty RPF: Infections and anything could kill you, basically.
1:08:23
Jim O'Shaughnessy: Anything could kill you.
1:08:24
And, so this romanticizing that world, I think is really, really wrong.
1:08:31
Liberty RPF: We cherry pick what we look back on, we don't look back on the average life from those days.
1:08:36
We look back at stuff that was cool enough, that someone wrote about it and still reads about it today.
1:08:41
It survived all that time, because it was special.
1:08:43
It's like, there's a good example for survivorship bias with log cabins, and people are like, "Oh, wow log cabins they used to be so well built, they last forever, the design is great.
1:08:53
And, look at all these log cabins they're all great."
1:08:56
It's like, No, no, no, those that survived are the great ones.
1:08:58
The shitty ones all fell down or people tore them down or didn't maintain them.
1:09:02
So, that's the same thing with the past, we only look at those things.
1:09:06
There's a thought experiment, I'm sure you've run it because how can you not think about it.
1:09:11
It feels like humanity, the arrow of progress is mostly going in one direction, but not all the time.
1:09:17
So, if you think about ancient Greece and Alexandria and all that thing, and then you went back a lot in the Middle Ages.
1:09:24
What if the Middle Ages didn't happen for a lot of progress would we be in the stars right now having defeated aging?
1:09:35
And, progress doesn't depend on the calendar, people have to make it happen.
1:09:40
And, if for a while people don't make it happen or block it, it's not going to happen.
1:09:45
People assume that Moore's law is like this thing that's written in the stars and chips are going to get better over time.
1:09:49
It's like, no, there's the bunch of really smart engineer in Holland working really hard on the machine that costs 150 million.
1:09:55
And, has 120,000 parts that atomizes tin with lasers and has a mirror so flat that if it was the size of Germany you wouldn't find a millimeter bump on it.
1:10:04
That's the kind of people that make it happen and I can't help but wonder where would we be right now if humanity had been just 1% better at it, 1% over 500 years. Where would we be?
1:10:17
So, we have to be- Jim O'Shaughnessy: Compounding.
1:10:19
Liberty RPF: The best right now for the future, for our kids.
1:10:21
Jim O'Shaughnessy: No question.
1:10:22
And, it's the idea that's near and dear to both of our hearts, because we're both big on the various investment schools and everything.
1:10:31
It's misattributed to Einstein, I know for all of the- Liberty RPF: Everything is Einstein, Oscar Wilde, George- Jim O'Shaughnessy: George Bradshaw or Mark Twain.
1:10:40
Liberty RPF: Mark Twain has like 50% of them.
1:10:42
Jim O'Shaughnessy: Oh, totally.
1:10:43
Liberty RPF: He said everything.
1:10:44
Jim O'Shaughnessy: He did, everything that was ever uttered was uttered by either Oscar Wilde- Liberty RPF: That's something- Jim O'Shaughnessy: Or Mark Twain.
1:10:50
Liberty RPF: You could do with an AI, right?
1:10:51
Just have a 50 petabyte book of every possible sentence, copyright it, and then in the future everything can be attributed to you.
1:10:58
Jim O'Shaughnessy: Yes, I love it.
1:10:59
One of the things that, as I was listening to you about progress, not depending on the calendar people, have to have the vision to create it.
1:11:07
I love that line, I'm going to put it up on Twitter, I'm going to credit you with it of course.
1:11:11
But, is this idea, one of the things, and this is a good example of cross pollenization.
1:11:18
So, I'm a big Joseph Campbell fan, I know you are as well.
1:11:21
And, I was reading Campbell the other day and it just hit me right between the eyes.
1:11:27
Because, one of the reasons why Stability AI has a pretty good advantage is our compression algorithms are better than everyone else's, basically.
1:11:39
Now, will that last forever?
1:11:41
No, not unless we continue to make them better and better and better.
1:11:45
But, this idea of compression has been around forever, and I was reading Campbell and there it was.
1:11:53
He was like, what people should understand is that stories that have been around for a very, very long time.
1:12:01
And, have been told and retold and retold and retold.
1:12:06
And, he uses the term, he said have been compressed to just the best parts of that story.
1:12:14
And, it's like your thing about the log cabins, yes we are sitting in a world where the log cabins that we see were the best log cabins.
1:12:24
We are also in a world though where many of the stories that most of us are aware of, despite where we grew up or whatever, are also the result of getting compressed, compressed, compressed and containing real information.
1:12:39
Back to Claude Shannon right.
1:12:41
Liberty RPF: Claude Shannon.
1:12:42
Jim O'Shaughnessy: And, I sometimes think all roads lead to either him or Feynman, or one of those guys.
1:12:48
But, I'm fascinated, because that's what's going to happen here.
1:12:54
And, one of our projects when we can do it is we want, A we want AI to be programmed in every language in the world.
1:13:03
Because, right now it has a very western bias.
1:13:07
Liberty RPF: Language includes culture too.
1:13:10
You can not- Jim O'Shaughnessy: Culture, too.
1:13:12
Liberty RPF: Have culture without the language.
1:13:13
Jim O'Shaughnessy: Well, for example, our chief operating officer is from Japan and I was spending time with him in San Francisco last week.
1:13:19
And, he had this sudden insight when we were talking about manners of communicating.
1:13:24
And, he's like, he looked at me, he stood up and he goes, " Jim, I am you, but only in Japanese."
1:13:34
And, I'm like, "Ren, what are you talking about?"
1:13:37
He goes, "I just had this satori," he goes, "Like you say exactly what you mean, you don't pussy foot around, you're very straightforward."
1:13:45
He goes, "It's one of the things that I love about you."
1:13:47
And he goes, and I was talking to my friend and I'm like, "God, I wish I could be like that."
1:13:54
And, he goes, "Well, it's because it's English, English is your second language."
1:13:57
And, Ren was like, "I am that person in Japanese."
1:14:02
And, ultimately we ended up having this great conversation about it, about Wittgenstein, the limits of my language are the limits of my world.
1:14:14
And, when we apply it to why should AI be familiar with all languages and all cultures, because we want it to learn more and more and more, and better and better and better.
1:14:24
And, having just one point of view might be not so great.
1:14:29
Liberty RPF: And, I think it's also a tool, as the listener may be able to tell English is my second language.
1:14:35
And, I feel like being bilingual or trilingual, or whatever, allows you to step outside of a language to look at it from the vantage of another language.
1:14:44
It's kind of like a introspection tool in a way, it's harder to examine something with the thing itself.
1:14:52
So, I wonder if AI doesn't benefit in some way from looking at certain dataset from the point of view of other data sets, or of other model, of other inferences.
1:15:02
Maybe, there's a way to create a, you know how they have these tournaments where, I don't know the name of the model.
1:15:10
But, this against this, and you could have the English model looking at the Japanese model.
1:15:16
You keep them separate, but you have one analyze the other.
1:15:17
And, then, have the Chinese model look at the, I don't know, the Brazilian model.
1:15:21
Instead of meshing it all together, maybe there's a way of better looking at certain models.
1:15:27
Or extracting certain insights that can't come from inside the model, from the vantage of under the model.
1:15:32
I don't know, maybe it makes no sense at all, but it- Jim O'Shaughnessy: It actually makes sense.
1:15:35
You keep going on here and we're going to hire you over at Stability AI to design these things.
1:15:41
And, what's interesting is you intuit, your intuition was correct.
1:15:45
Because, where the conversation ultimately came to was me saying that one of my greatest regrets in life was not being able to speak a second language, because of what you just outlined.
1:15:57
The ability to step outside of your dominant way of thinking and speaking, and examine it from another way that you're fluent in, I think is a huge advantage.
1:16:10
And, of course I've got to tell the old joke.
1:16:11
Which is there's a woman and her little daughter walking down the street here in Manhattan.
1:16:17
And, the little girl hears people up in front of them, they're speaking English first, but then they switched to French.
1:16:24
And, the little girl looks up at her mom and says, "Mommy what language is that?"
1:16:30
And, her mom says, "Well, it's French."
1:16:32
And, she goes, "So, they speak both languages?" And, mom says, "Yeah."
1:16:37
And, little girl goes, "Well, what's that called?"
1:16:39
And, the mother says, "Well, it's called being bilingual."
1:16:42
Little girl goes, "Oh, that makes sense, and mommy what do you call somebody who speaks three languages?"
1:16:49
And, the mom said, "Well, they're called trilingual."
1:16:53
Little girl says, "Well, that makes sense too, mommy what do you call a person who only speaks one language?"
1:16:59
And, the mother looked at the girl and said, "An American."
1:17:04
Liberty RPF: Well, America has lots of immigrants who- Jim O'Shaughnessy: No, I know. Liberty RPF: Language...
1:17:11
I get the joke, no, that's the first thing I thought about on languages.
1:17:16
One thing you made me think of is how there's a kind of a quantum leap, where when you're learning a different language at first you're translating it and that's like okay.
1:17:26
But, at a certain point you start being able to think in the other language.
1:17:30
And, that's where I catch myself, sometimes I'm in the kitchen at night just thinking about something.
1:17:36
Like, "Hey, have been thinking in English for an hour?"
1:17:38
And, that stuff is weirder, because I'm not sure I'm the same person in my head, in English, as I am in French.
1:17:46
And, I've seen videos on TikTok or something, they're like, I think they're Japanese women or something, or maybe, I don't know.
1:17:53
And, they're like, "When I speak in my native tongue like in Chinese or Japanese I sound like this."
1:17:59
And, then, they speak and they have a very distinct voice.
1:18:01
And, when they speak English and they speak, it's totally different.
1:18:05
It's like an octave lower and it's very different.
1:18:07
And, it's like, is it only the voice that changes or is the personality influenced, because the language encode so much of the culture.
1:18:14
Jim O'Shaughnessy: I think so, because that's where Ren and I got to.
1:18:19
Like his insight, he was really excited, he's like, "I am you in Japanese."
1:18:24
And, I'm like, "I'm sorry," but then when he explained it to me it was this, what you just outlined.
1:18:32
He feels like a different person when he is communicating simply in English.
1:18:38
And, now you got to remember this is, he worked in the government.
1:18:42
Jim O'Shaughnessy: He worked in the government services.
1:18:43
He was an aide to Prime Minister Abe, who unfortunately was assassinated.
1:18:48
He learned English at a very young age.
1:18:51
He lives in London, so he thinks in English all the time.
1:18:55
But then he was just like, "I was talking to a good friend who's Japanese, and obviously we're talking in Japanese."
1:19:02
And he goes, "That exchange just lit up in my brain.
1:19:06
Oh, I behaved very differently when I was talking to my friend in Japanese."
1:19:11
Liberty RPF: So the next logical question would be, is there a diminishing return, right?
1:19:16
If you know three languages, four languages, at some point, do you stop having these kind of split personalities or does it keep going on? Right?
1:19:23
Does someone who knows six languages, very different languages, are they switching between partial personalities inside of their head when they switch languages?
1:19:33
I wish there was a study on this.
1:19:34
Jim O'Shaughnessy: Yeah, I do too because I think it's really exciting.
1:19:38
And again, one of my great laments is that I was too lazy to learn how to speak in another language.
1:19:46
So for example, for many, many years, we went to Italy every year.
1:19:49
That was interrupted by the pandemic, but I had been in Italy long enough so that I could understand, not a rapid fire Italian conversation, but when they came up and just asked simple questions of me in Italian, I knew what they were asking.
1:20:10
And it got to the point where I didn't even have to translate it in my head.
1:20:14
I just knew, oh, they wanted this or that, or anything.
1:20:18
But try responding in Italian, forget about it.
1:20:22
And we were with some friends where the guy spoke fluent English, he had been trained in English from his youth, and his wife, who was a lovely person, said she didn't speak English.
1:20:37
And so, we were having conversation and he was being very polite and speaking in English to us.
1:20:42
And anyway, I made a joke, or he made a joke, I can't remember who, in English.
1:20:47
And his wife started laughing and I looked at him and I said, "Hey."
1:20:51
And he goes, "No, no, no.
1:20:54
She understands everything in English."
1:20:57
Liberty RPF: Sounds like a little bit shy, right, because it's much harder to speak it than- Jim O'Shaughnessy: Exactly, Exactly.
1:21:02
Liberty RPF: I was just telling my cousin's daughter about this recently, about when I learned English quite late.
1:21:09
I was like 15, 16 when I started.
1:21:11
So I never had the chance to have two native tongues.
1:21:15
My kids are going to be like that because they're already bilingual.
1:21:18
But to me, English is always going to be kind of a different part of the brain.
1:21:22
And I was telling her about what made the biggest difference to me was when I stopped giving a crap.
1:21:29
For a long time, you try to speak your second language and you're like, you're just good enough at it that you can understand how much you suck at it. Right?
1:21:36
You're like, oh I'm not pronouncing that, and what's the word for this?
1:21:38
And you're always tripping over yourself.
1:21:39
And so you get self-conscious and you kind of stop and you just don't practice because you're kind of afraid.
1:21:46
But then I was part of a startup and we had a weekend in New York brainstorming like 72 hours from morning to the night.
1:21:53
And all day, I spoke English for hours and hours and hours.
1:21:56
At this point, I was so tired I just stopped giving a crap and I just went for it.
1:22:00
And forever after that I was better.
1:22:02
I just had to have this kind of leap forward.
1:22:05
And then it was like something unlocked in my brain and it was like, okay, now I can do this.
1:22:10
And I feel like a lot of people, that's what's stopping them. It's not their capacity.
1:22:13
It's like the driver in the car has their foot on the brake.
1:22:17
That's what you have to get over. Jim O'Shaughnessy: Yeah. Yeah. Yeah.
1:22:22
And I think that it's pretty cool in terms of just the ability to, again, cognitive diversity.
1:22:28
There's language diversity too.
1:22:30
And being able to take advantage of them, I think, are going to be cumulatively much better for us.
1:22:37
Let's switch gears for a second here to learning.
1:22:40
This is something you and I talk a lot about too and we, big surprise, agree on this idea that- Liberty RPF: You should find something we disagree on.
1:22:50
Jim O'Shaughnessy: I know. I will.
1:22:53
For our next chat that we're going to publish.
1:22:55
I am going to scour all of your Substacks and I'm going to find something.
1:22:59
I'm going to say- Liberty RPF: Looking forward to it.
1:23:01
Jim O'Shaughnessy: Liberty, I disagree with this.
1:23:03
Liberty RPF: Controversy sells.
1:23:04
That's going to be my biggest ever.
1:23:06
Jim O'Shaughnessy: Yeah, it'll be huge.
1:23:07
But this whole idea that you have said, why does everybody think that if something is serious, it has to be important?
1:23:16
And another thing you've said that I totally agree with is learning is so much easier if it's fun.
1:23:25
Liberty RPF: Oh, 10000%, right? That's the thing.
1:23:31
You can never make someone learn something.
1:23:33
They can learn it, you can put them in the right position, you can surround them with the right tools and the right incentives.
1:23:39
But if someone doesn't want to learn, they can be sitting in MIT all day long and they could learn nothing at all, or they could learn just enough to pass the test and then forget it the next day.
1:23:49
That's something I think we agree too.
1:23:51
I think the current education system has been created for a certain thing.
1:23:55
And that thing was mostly when you're young, get out of the way of the parents so they can go work in the factory or something, and then you kids, you're going to learn to be a good worker, be on time.
1:24:06
Everybody's the same, everybody's doing the same things.
1:24:08
And it's kind of a big daycare slash prison system with some learning on top of it.
1:24:15
And I feel like my kid, we spend an hour discussing.
1:24:18
He's going to ask me questions he's curious about and he's going to learn more in an hour of intense curiosity in conversation than he may spend in a week with a teacher at the blackboard or whatever they have now, the digital board.
1:24:30
I feel like schools should be about lighting the flame of curiosity and giving you basic tools, like reading and arithmetic, and there's stuff that they teach you that's super important for all of the rest.
1:24:43
But the rest should more be about...
1:24:47
Children and are naturally curious.
1:24:49
They're naturally learning machines and what schools should do is not kill that.
1:24:54
It should be about avoiding errors rather than trying to do something special because it's already there.
1:25:00
And I wish, I don't know, because it's a hard problem.
1:25:04
If you put me in charge of redesigning school systems, it's not that easy.
1:25:09
But I think we can know what not to do.
1:25:13
A lot of what we're doing right now is what not to do and that's part of the internet's superpower.
1:25:18
Everything I've done since I was a teen was based on stuff I've learned by myself in books or online.
1:25:25
And most of the books I've read, I've found online.
1:25:28
If I didn't have that, I'd be a completely different person.
1:25:30
All of my life's trajectory would be different.
1:25:33
So just figuring out how to have more people.
1:25:37
That's one of my big things, right, is the exploration versus exploitation kind of framework.
1:25:42
I feel more people should be more on the exploration side of it and they should...
1:25:47
That's the thing we should encourage, right?
1:25:50
That's what I tell my cousin's daughter when I see her.
1:25:52
I try to be kind of a mentor to her and be like, "At your age, don't be locked on just one thing that you think is going to be the thing for your life because you haven't seen enough things yet to know what you really like, what's really interesting, what's out there."
1:26:04
You can't know what's out there just by passively absorbing it because TV shows and this and that, they all show the same things.
1:26:10
Everybody's a doctor or lawyer or a firefighter or a soldier.
1:26:14
There's way more stuff out there.
1:26:16
There's way more stuff that didn't exist when your parents grew up that they can't tell you about.
1:26:20
New jobs are being created all the time.
1:26:22
Being a software writer is not exactly like an ancient job. So I don't know.
1:26:28
If school could somehow be reorganized to foster this curiosity and this exploration and have the flexibility of adapting itself to what kids want, rather than trying to bring kids to what the school wants, to form kids in the shape of the peg of the school, that would be wonderful.
1:26:46
But I don't quite know how.
1:26:47
Jim O'Shaughnessy: So, I will admit to taking advantage of you by incepting that idea in you because I think that some of the biggest new and incredible innovations coming are going to be on the educational side.
1:27:06
And so, I'm talking to a lot of my super smart friends like you and getting them to start thinking about it. And I know you.
1:27:15
You're not going to be able to stop thinking about it.
1:27:18
Liberty RPF: Probably not.
1:27:19
Jim O'Shaughnessy: And then I'm going to get a great, and who knows, we could go into business together on one of your great ideas, because I could not agree with you more.
1:27:29
The old way was essentially pounding creativity out of people because the needs of society in an industrialized world were very different than the needs of society today.
1:27:45
And what we need today is teach the children how to learn, not what to learn.
1:27:54
And so I'm convinced that there's going to be a couple of new models that work pretty well that could really change lives of children.
1:28:08
I'm a Montessori kid, and I don't have any real memory of that, other than all of the teachers of Montessori when I was in there in the Paleolithic era came from India.
1:28:21
And so my dad set up the first Montessori in Minnesota right around when I was three, in 1963.
1:28:28
And so the only thing that I can actually remember was being really upset about having to go to the Montessori classroom on a Saturday when all my friends were going for the double feature at the movie theater.
1:28:41
And the woman went in her beautiful saree and everything.
1:28:45
It was a Parent's Day and they needed kids.
1:28:48
And no one had heard of what Montessori was, its methodology.
1:28:54
Now, you've given me an idea.
1:28:56
I probably should have my sister on here. She probably knows more.
1:28:58
She's the former trainer for North America in the Montessori method.
1:29:02
But I didn't even think about all the advantages that gave me until way later in life.
1:29:11
And my sister did this thing.
1:29:12
I was just going through a box.
1:29:14
I'm moving out of my office because I'm retiring from OSAM and going with the new company.
1:29:19
And there were these baseball cards that my sister had made up and they called Montessori All Stars.
1:29:25
And I'm looking at it, guess who else was in there? The founder of Google. Oh, Jeff Bezos.
1:29:33
And the list just keeps going on and on and on.
1:29:37
And so, Montessori is an older version of being able to get kids involved, and kids can do so much more than we think that they can do.
1:29:47
And that's why Montessori has you work with knives, real knives, and things like that, and dexterity is very important.
1:29:55
But my point is, I think that there are other models, combination of other models, et cetera that are going to work beautifully.
1:30:06
And to me, that's super exciting because talk about win-win.
1:30:10
Talk about asymmetric opportunity.
1:30:12
Liberty RPF: Literally, you know how you have a curve and the sooner you can have an effect on the curve, the more you can bend it. Right?
1:30:20
If you can affect someone's childhood, you can change the whole rest of their life, and by a huge amount.
1:30:25
There's such simple stuff that should be like...
1:30:28
I've been reading Jocko Willink's books to my kid and some of the principles in there are so powerful, and now my kid is doing pull- ups and trying to...
1:30:39
The right idea at the right time in someone's mind can change everything, right?
1:30:43
I'm sure you've seen Carol Dweck stuff about growth mindset versus fixed mindset.
1:30:47
Well, just putting these ideas out there, and I'm sure there are studies that say, "Oh, the effect is not this."
1:30:54
There's still stuff that we know can work.
1:30:55
And even if it works on 30% of kids, that's much better than zero.
1:30:59
And then you talk to them about stoics, or with the power of technology, well the best teacher in the world, Khan Academy or something, can teach millions of people and the teachers can be more about one on one time with the kids rather than having to reteach the same things over and over again.
1:31:14
We can record things now.
1:31:16
We can spend our times on better stuff.
1:31:18
My version of that was I went to regular school, but at the time, I didn't know it was a good thing, but my childhood was extremely unstructured.
1:31:26
I basically had all the time in the world to do whatever I wanted just by myself.
1:31:31
And I think that's what kind of preserved my curiosity.
1:31:34
If I had been super structured like, oh, at this time you do this and then you have this class and then you have a piano lesson, I'm sure some good would've come of that, and sometimes I wish I had that, but now I realize that some of my peers maybe never had this. Just figure it out.
1:31:51
Just find what's interesting to you.
1:31:52
Just go down the rabbit hole looking at the...
1:31:56
I had this encyclopedia with pictures, and I spent days and days just looking through them and volcanoes and submarines and all kinds of stuff, I still kind of do today a version of. Jim O'Shaughnessy: Yeah.
1:32:09
And then back to our earlier topic, wait until AI gets involved, and we can tailor make certain things to certain dispositions and encourage...
1:32:21
The prompt would be, oh, student A is showing a strong interest in subjects A, B, and C.
1:32:27
What are the adjacent subjects that that student might be interested in?
1:32:32
Serve some of those up to the student.
1:32:35
They might not want them, but if you have the ability to immediately serve that up for the student, they're going to find like, oh my goodness.
1:32:44
Like you with the encyclopedia, my father, I was punished for doing reading assignments too much in my traditional grade school, and it was a Catholic grade school.
1:32:55
And so, they made me sit at my desk with the little...
1:33:00
For people listening, I'm showing him the Christmas tree you make with your hands.
1:33:04
And so, I felt like I was being punished for getting the work done too soon.
1:33:08
They should have sent me to the library, but they made me sit like this, which again, I have to thank them because it really stoked my rebellious nature. Liberty RPF: Yeah.
1:33:19
You get the reverse of what you punish most of the time.
1:33:21
Jim O'Shaughnessy: Always.
1:33:23
And so I went back, and I'm this little kid, and we always had dinner as a family, which I did in my family as well.
1:33:30
It's just a great- Liberty RPF: Yeah. I do too.
1:33:34
Jim O'Shaughnessy: Way to communicate.
1:33:35
Anyway, so my dad is saying, "Well, how was your day today?" And I'm like, "Sucked." "Why did it suck?"
1:33:43
"Well, because I'd read through all of the SRA thing through the end and they told me that I was doing it wrong and that I just had to sit at my desk while all the other kids did their thing."
1:33:57
And my dad goes, "Oh that's horrible."
1:33:59
And he goes, "Come with me."
1:34:02
And he walked me into our more formal living room where we had the Encyclopedia Britannica, and he took the first volume out and he goes, "Just look through this."
1:34:17
And I wasn't yet reading well enough to comprehend a lot of it.
1:34:21
So like you, my first trip through the Encyclopedia Britannica was, I was drawn to the pictures and to, "Oh, I don't even... What animal is this?" Right?
1:34:34
But it's just that simple thing alone, what a great thing to have happened to me, and maybe it's my Daoist nature, but we'll see.
1:34:49
Other people might tell that story and say, "Look at how horrible those nuns were and punished him for doing the reading."
1:34:59
But okay, what happened next? What happened next?
1:35:01
Liberty RPF: Yeah, it's what you do with it. Right?
1:35:04
Someone could have been discouraged or someone can be like, "This is fire, this is putting a fire under me and I'm going to do even more of that." Right?
1:35:13
Jim O'Shaughnessy: Exactly.
1:35:14
Liberty RPF: And if I look at my kids, if I had to design a way for them to learn stuff with the power of AI, and maybe this is another idea for Stability, but what I would do is, they love games.
1:35:27
Jim O'Shaughnessy: Oh yeah. Liberty RPF: So okay.
1:35:30
You want to teach my oldest about some math concept. Here I go.
1:35:35
AI, generate me some Minecraft world with some math puzzles that you have to solve to get to the netherite sword or something. Right?
1:35:43
I want my youngest to learn about ancient Greece.
1:35:46
Generate me a Zelda style Breath of the Wild world where you're walking through ancient Greece and then you meet people and they talk to you about, I don't know, stoicism and stuff.
1:35:56
And that's just part of the story and the puzzles.
1:35:58
If you can generate something that's tailored to the interest of the person, but that's also kind of related to the curriculum you want, I think you're going to learn much faster.
1:36:06
If you gave this to my kid the next day, they'd be like, "What's next?"
1:36:09
It wouldn't take them a month to learn their time tables or about stoicism or whatever.
1:36:17
There's a great essay by Derek Silvers called There's no Speed Limit.
1:36:19
Jim O'Shaughnessy: Yeah, I love his work.
1:36:20
Liberty RPF: Yeah, there's no speed limit.
1:36:22
I don't know if you've seen this one, but it's how you learn, I don't know, a few semesters of music theory in a couple weeks or something, because you don't have to learn at the pace that others are learning.
1:36:32
This pace is mostly designed for people that are not super self-motivated, that are not...
1:36:37
Some people need that pace, that's fine.
1:36:39
But if you are more motivated, you can go much, much faster.
1:36:42
And that's the terrible thing, is that school is for people who have problem learning or something, they're going to be left behind.
1:36:50
And people who learn very, very fast, they're going to get so frustrated and bored that they're also going to do badly.
1:36:55
So the people at the center of the bell curve, they kind of do okay, but you kind of chop off both ends and that's terrible because both ends need more support.
1:37:03
And when you listen to interviews that Patrick does, how many times a year...
1:37:08
I was a terrible student.
1:37:09
I didn't do well at school.
1:37:11
A bunch of people did great things and created great products and technologies and hired employees, thousands of people now, the system failed them.
1:37:24
How many people that have been failed never recovered and we're missing out on their greatest life's work?
1:37:30
Because some people can recover, but some people like, to no fault of their own, they just won't recover from that. That's tragic.
1:37:37
Jim O'Shaughnessy: Again, we are just so eerily simpatico on these things because the game idea- Liberty RPF: What was the joke?
1:37:44
Do you have a long lost son in Canada or something.
1:37:48
Jim O'Shaughnessy: Right. Yeah. Yes.
1:37:52
Well, we won't talk about that.
1:37:53
Liberty RPF: It's only a joke.
1:37:54
Jim O'Shaughnessy: Yeah, it's only a joke. We're kidding.
1:37:56
But yeah, the game idea is one that I think could be massive.
1:38:00
Liberty RPF: Oh, I think it's going to replace everything if it can be done well. Jim O'Shaughnessy: Yes.
1:38:05
Liberty RPF: Because why would you sit down and read at a blank page with a bunch of boring stuff on it if you could play a game, right?
1:38:10
Jim O'Shaughnessy: Totally agree. Totally agree.
1:38:12
Well, I didn't get to our agreement on that we should build a thousand nuclear plants around the world. We agree on that too.
1:38:20
And maybe we'll lead in with that as our next conversation, and maybe for people who were expecting a bunch of stock market talk, we'll talk a little bit more about investing next time too.
1:38:32
But so, this is even better than I anticipated, and I had very high expectations for you.
1:38:39
So you have exceeded them.
1:38:41
Always love talking to you, love your ideas.
1:38:44
I love seeing you succeeding with the Substack.
1:38:47
And I know it's been a slow go, and it's like that, as you point out, sometimes people get a...
1:38:55
Tim Urban does these graphs so well.
1:38:58
It's like they give up right here, and then the next thing is a vertical.
1:39:03
And I love the fact that you're not giving up, but more importantly that you're having a good time doing it.
1:39:08
Liberty RPF: I always try to keep that number one.
1:39:12
My own thing is number one priority is to keep it fun for myself, because if I don't do that, I'm going to stop it and I'm not going to achieve any other goals.
1:39:21
So it's pointless to have any other goals if it's not fun.
1:39:23
Jim O'Shaughnessy: Totally aligned with you.
1:39:25
All right, so we always have the same question that we've modified a little, I used to go in much more elaboration on it, but I am bestowing on you a magical microphone in which you get to utter two phrases, and these phrases will be incepted into every human being on the planet.
1:39:49
And whenever they wake up for whenever their next day is, they're going to sit up in bed and think that both of these ideas were their own, that they're the best ideas they've ever had, and that they're going to act on them.
1:40:02
What two things are you incepting?
1:40:04
Liberty RPF: Can I keep my crown and my royal outfit from the last time?
1:40:09
Jim O'Shaughnessy: Yes, you may. Liberty RPF: Awesome.
1:40:11
That stuff, I think I'm going to reference a few things we talked about.
1:40:13
I think the first one would be on the explore versus exploit, and I don't like that name a lot.
1:40:20
I think I prefer explore versus create, right?
1:40:23
I feel like specialists are creating in one field.
1:40:25
I'm going to try to incept people to move the dial a little bit more in the exploration direction.
1:40:33
That's kind of what I try to do with the Substack.
1:40:35
Maybe people don't have time to do it themselves, but if you read what I'm writing, it's like you're going to learn about nuclear submarines and about all kinds of random stuff that you may not have found otherwise.
1:40:45
Maybe half of it, tutors of it, is not interesting.
1:40:47
But if a couple times a week I give you something interesting that you wouldn't have known otherwise, maybe you'll discover your new favorite album or TV series or a new book or whatever.
1:40:57
To me, if I can do that for even one person, right, perfect. I've done my job.
1:41:02
So explore a little bit more, inject a little bit more chaos into your life, stuff that doesn't fit usually.
1:41:08
If you only listen to country, well go listen to some gangster rap for a while, that kind of stuff.
1:41:13
The second one is more serious.
1:41:15
I wish everybody would take very, very seriously what we discussed about when it comes to nonrecoverable mistakes about ruin, right?
1:41:23
We want to really avoid multiplying by zero because if we avoid that, I think the rest is going to take care for itself because our technology, our science, is getting powerful enough that a bunch of problems that have been unsolvable, we've put everything against them and they were not solvable.
1:41:40
I think now, we'll have the tools to solve them.
1:41:43
So if we just stick around, billions of people could have better lives.
1:41:49
We just have to avoid that nonrecoverable mistake. It just takes one.
1:41:54
And the danger is that there's multiple ways this could go.
1:41:58
Nick Bostrom is a good book about this called ‘Existential Risk’, right? Jim O'Shaughnessy: Yeah. I love his work. Liberty RPF: Yeah.
1:42:04
Super volcanoes and that, there's some stuff we can do more about, some stuff we can do less about, but whatever we can influence, nudging it by 0. 1% for all future lives.
1:42:15
If you avoid one existential event, you're not only saving people alive to date, you're saving the trillions of future humans that are not born yet.
1:42:26
So the utility of that, nothing compares.
1:42:28
It's the same reason why I'm so in favor of working much, much harder on the diseases of aging, curing all these, because it's an evolutionary blind spot, right?
1:42:38
Evolution doesn't care what happens to you after you reproduce because your genes are already passed.
1:42:42
So all of the sex pressure is after reproducing.
1:42:45
That's why everything goes wrong in our bodies, but we don't have to have the same priorities as nature.
1:42:50
We can fix that stuff, right?
1:42:52
Anyway, I'm getting on a tangent as I always do, but that would be what I would try to incept.
1:42:57
Jim O'Shaughnessy: I love both of them.
1:42:58
I think they're both great ideas. Again, here we are.
1:43:02
I'm going to find things we disagree on for our next chat, but immediately my thinking goes to ‘Deadwood’.
1:43:08
For listeners, Liberty and I are both massive fans of ‘Deadwood,’ the Western series, and we'll talk about it at any chance that we get.
1:43:18
And the other thing, we share a love of many of the same types of movies.
1:43:23
For example, I saw you watched, ‘Amadeus’ again, one of my favorite movies of all time. But listen to me.
1:43:31
I haven't watched Amadeus in 25 years probably, Liberty RPF: And I bet it's that memorable that you could almost see it in your mind scene by scene. Yeah?
1:43:43
Jim O'Shaughnessy: I can. I absolutely can. I absolutely can.
1:43:46
Anyway, thank you so much.
1:43:47
As always, it has been my pleasure and privilege to talk to you and keep inventing these cool new ideas.
1:43:55
Who knows, we might do something together with them.
1:43:58
Liberty RPF: Thank you so much, Jim.
1:44:00
I really appreciate you and appreciate you inviting me, and it's always a great pleasure talking with you.