Latest / The Innovation Show / The Magic of Code with Sam Arbesman | The Innovation Show (Thinkers50 Award-Winning Podcast)

Transcript

The Magic of Code with Sam Arbesman | The Innovation Show (Thinkers50 Award-Winning Podcast) · The Innovation Show

0:00
before we start, I want to Thank our sponsor Kyndryl, with a unique
0:02
blend of AI powered consulting built on unmatched service capability.
0:06
Kyndryl helps leaders harness the power of technology for smarter decisions, faster
0:10
innovation, and lasting competitive edge.
0:13
you can find Kyndryl at KY NDR YL ndl.com.
0:18
In the digital world, code is the essential primary building block,
0:22
the equivalent of the cell or DNA in the biological sphere, and
0:27
almost as mysterious code can create entire worlds real and virtual.
0:33
It allows us to connect instantly to people and places around the globe,
0:37
and it performs tasks that were once only possible in science fiction.
0:42
It is a superpower and not just in a technical sense.
0:46
It is also a gateway to ideas.
0:48
As vividly illustrated by today's guest code is the ultimate connector,
0:53
providing new insight and meaning into how everything from language
0:57
and mythology to biblical text even our patterns of thought connect with
1:03
the history and nature of computing.
1:06
Today's book is his attempt to reinvigorate the human in the world
1:10
of computing from understanding how our machines can help us to think
1:14
better, to seeing the similarities between the realm of language and text,
1:18
and the world of code and software.
1:20
But computing also extends its tendrils into simulating entire
1:24
societies, the nature of life itself, and the question of whether all of
1:29
reality is just a computer program.
1:33
It is a pleasure to welcome back for the finale of our trilogy, the
1:37
author of The Magic of Code, how Digital Language Created and Connects
1:43
Our World and Shapes our future.
1:45
Samuel Arbesman, welcome back to the show.
1:47
Thank you so much.
1:48
Yeah.
1:48
Good to be doing this again.
1:50
It's great to have you, man, and thank you.
1:52
Sam has just landed only a few hours ago at 5:00 AM and he still made our
1:57
call happen, so I appreciate it, Sam.
1:59
pleasure.
1:59
No, I'm, I think I'll be lucid enough.
2:04
Well, if you're in a computer program, it doesn't matter, but this is a brilliant
2:09
book for those who have ever heard of the idea of the similation hypothesis, if
2:15
we're living in a a similation, et cetera.
2:17
This book is the building blocks of how we got there, and that's towards the
2:21
part three of the book towards the end, where Sam talks about artificial life.
2:26
But let's, share your entire thesis behind this book because for me
2:31
it was such a privilege to see how you'd built everything to get to
2:35
here and still the common thread is the importance of humility and
2:41
curiosity as we progress in technology.
2:44
Oh, I love that.
2:45
Yeah, I definitely feel that the, yeah, that is definitely one of the themes that
2:48
kind of goes through all of my books.
2:51
But yeah, for this one, the way I think about it is, if you look around at the
2:55
current conversation around technology and computing, feels a little bit broken.
3:00
Like we're often like adversarial or worried about technologies, we're
3:03
concerned about certain things around AI or about social media or phones and kids.
3:09
And there's all these concerns and those concerns are all valid.
3:13
But when I remember my own experience with computers growing up, wasn't just worry.
3:19
It was also full of a huge amount of like wonder and delight.
3:22
And not only that, it didn't just feel like computation was
3:25
just this branch of engineering.
3:27
It was really this liberal art that connected to language and philosophy
3:32
and biology and art and how we think and all these different kinds of things.
3:34
And so.
3:35
For me, like thinking back to my own childhood of the Commodore Vic 20
3:40
and early programming and fractals in Sim City and all this kind of stuff.
3:42
Was like all, there was all these really exciting areas that touch
3:45
upon so many different aspects of our lives in this kind of like
3:49
humanistic liberal arts sort of thing.
3:51
And that's essentially the point of the book.
3:53
The book is designed to rekindle that sense of wonder and understanding
3:58
and hopefully provide a healthier way to approach computing.
4:02
One that kind of connects the human with the machine and can hopefully at
4:05
least somewhat repair a little bit of society's broken relationship with tech.
4:10
Reading it from the outside as somebody who wasn't a coder.
4:12
And I regret in some ways I missed that whole revolution that happened.
4:17
I remember we had actually Macs in the computer building
4:20
and we had a very old priest.
4:22
He didn't even know how to use them, but he was like this guardian of the computer.
4:25
So it wasn't very welcoming.
4:28
And actually you talk about the importance of serendipity as well here.
4:31
Like for example, you talk about SIM City and how Sim city
4:35
inspired a plethora of people to go into city planning as a result.
4:39
And I was thinking about how those important little serendipitous moments
4:43
can just redirect the course of your life.
4:46
But reading it, it was really clear to see that this reason you write
4:50
this, the magic of code is, it really is in a way, a history of how we got
4:55
there, but also an ode to code and your celebration of it in many ways.
5:00
Maybe you'll tell us a little bit more about that.
5:02
right, like it is this ode and yeah, this like this pan to this kind of stuff.
5:05
'cause I, and for me, and I think I, me maybe mentioned this in the end
5:09
notes, I can't remember exactly where.
5:11
But in many ways, this is the book that I wish I had when I
5:15
was first getting into computers.
5:17
Like I, I had to cobble together a lot of these different ideas on my own.
5:21
And so this is sort of my own way of like paying it forward to like
5:24
future young Sams who are beginning to think about these kinds of things.
5:28
Because right there is all of this wonder in these in these areas and
5:33
too often we've we've forgotten it.
5:35
And so I don't think I mentioned this story in the book, but so my grandfather
5:39
who lived, he lived to the age of 99.
5:41
He was a dentist but in addition to, being a dentist, he was a
5:45
lifelong fan of science fiction and he basically read it since like the
5:48
modern dawn of the genre he read.
5:50
I think he read Dune when it was like serialized in a magazine.
5:53
Like he knew all the different stories.
5:55
And so I remember when the iPhone first came out, back in, I
5:58
guess 2007, something like that.
5:59
I went with my grandfather as well as my father to the local Apple store.
6:03
To see the iPhones were, and we're playing with it.
6:06
And at one point grandfather says, he's like, this is it.
6:09
This is the object I've been reading about for all these years.
6:12
And like it really was something out of science fiction.
6:15
And yet we moved so quickly from that to, I dunno, complaining about battery
6:19
life or camera resolution or like weird buggy apps and like, we've forgotten that.
6:23
And for me, I think we need, I mean we certainly want to, I, I don't think you
6:27
can live with that like sense of like transcendental wonder at every moment
6:30
of your life, but we shouldn't take some of these kinds of things for granted.
6:34
And so for me, I about like, how can we have this sense
6:39
of wonder in our computing?
6:40
And I talk about the nature of like enchantment and disenchantment.
6:43
And so this is by analogy with the way people have talked about how like
6:46
we went from a world that was full of enchantment of like, I don't know,
6:50
magic and sorcery and gods and powers and all these different things to one
6:53
where that was in kind of the era of the enlightenment, no longer the default.
6:59
And using that kind of framing, I talk about digital enchantment, not
7:04
necessarily thinking of like this stuff is magical, but more in the sense of it
7:08
actually elicits the sense of wonder.
7:10
And there's a simple narrative that you can tell that says, oh, we back in
7:15
the day when computers were simpler and people were doing all these weird things
7:20
there was lots of enchantment and then things were disenchanted by virtue of
7:25
software and just building these things and software team, stuff like that.
7:29
The truth is that's not really that's not really the right narrative.
7:33
In fact, the digital enchantment and kinda like the wondering approach alongside
7:37
the, this more utilitarian kind of enterprise software kind of approach.
7:41
They've always coexisted.
7:42
So back in the day when you had really large computers, there were these
7:46
very large mainframes that were and they were the profits of companies
7:51
or universities or large labs.
7:54
And you were very distant from 'em.
7:55
But then alongside that, you had like mini computers.
7:57
Well, maybe a little bit later.
7:58
They were still very large and they're called mini.
8:00
They're actually like the size of a refrigerator.
8:02
And that was where some of the, like some of the earliest
8:04
computer games were developed.
8:05
So like Space War was developed during this time that was still very corporate.
8:08
And then once you moved to like personal computers, you both had the advent of
8:12
like the spreadsheet and computers and offices, but then also personal computers
8:17
that were geared more towards delight.
8:19
Like the Macintosh and all the early weird shareware and all this kind of stuff.
8:23
That was around then.
8:24
And then even now on the web.
8:26
both have things that are the province of big tech that feel not necessarily that
8:31
delightful, but then you have an entire realm of what people call the poetic web,
8:35
which is just the websites that are almost like works of art or fun and delight.
8:39
And so for me, only have these things always coexisted, but they
8:42
often play off of each other.
8:44
So oftentimes the more utilitarian frame, they actually provide the
8:50
foundation for the people who want to build the more wonderful stuff
8:53
to actually do it more easily.
8:54
And then conversely, the wondering stance, the really delightful things,
8:57
the poetic web, whatever it is, that can then actually help get people excited
9:01
about getting into software more broadly, and then maybe build more of these other
9:05
kinds of things that maybe are not as delightful, but still , very important.
9:09
And so for me, when I think about wonder, it's not just something that we
9:14
used to have or that we've forgotten.
9:15
It's always there, but we need to search for it and find it.
9:18
And then.
9:20
use it because it actually is very, very important for kind
9:22
of moving computing forward.
9:24
And what is that wondering what is that the source of wonder in computing?
9:27
It can be lots of different things and like going back to like this kind
9:30
of liberal, liberal art of computing, it is, for me it's building right
9:35
Building com computer programs that are works of art or thinking about how
9:38
to actually encapsulate biology in, in computers or simulations or Sims City
9:43
or all these different kinds of things.
9:44
And, and we can talk about all, all those, but that, that was the goal
9:48
was to show that this wondering stance not only is very powerful, it's always
9:52
been around and in fact some of the most interesting things around what
9:58
computers are doing, whether it's AI or weird things with artificial life.
10:01
They were actually present very nearly at the advent of the modern digital
10:05
computer like decades and decades ago.
10:07
But it took a long time for them to just, I guess, like percolate
10:10
and become, , a reality or at least something that people can use more.
10:13
So , we definitely need more of that sense of wonder in our computers.
10:19
That idea you share about shifting baselines, but also I think it was it John
10:24
McCarty you quote as well, like who's, it's not considered AI once it works.
10:27
Was it that ring a bell?
10:28
right.
10:28
Like, right, because ai, I mean, I certainly, I feel like now it's changed
10:31
a little bit because some of the things are just so impressive, but Right.
10:33
For a long time, the realm of artificial intelligence, it was at that
10:39
frontier of things that weren't quite able to be done that kind of mimic
10:42
certain features of human behavior.
10:44
And then once they were solved, you're like, oh, this is
10:46
actually just some weird trick.
10:47
It's not really ai, it's just some other branch of computing or whatever it was.
10:50
And we forget that, right?
10:51
That we have been making very, very powerful advances along the way, even
10:55
if, yeah, we're not willing to have that sense of wonder or consider it
10:59
to be truly AI or whatever it is.
11:02
I'm gonna mix it up now because I, there's so much in the book and I really
11:06
love part three and the end of part two as well, where you, you're a trained
11:10
biologist as well, which I didn't realize.
11:11
I knew you were a polymath, but I didn't realize that until this book,
11:15
and I can see now your curiosity around biology and also genetics
11:19
and all, everything in that realm.
11:22
But when I read about that, I was like.
11:24
AI essentially has evolved in a similar way, in a similar fashion, the same way
11:31
as biology does in DNA, for example.
11:34
. I'm grabbing pieces from the future and bringing them into part one
11:37
there, 'cause they mix well together.
11:38
. So in, in there's, yeah, two chapters near the end of the book where I
11:42
discuss various different aspects of biology and the realm of computing.
11:45
And one of them.
11:46
just simply to look at the ways in which biology and computation
11:51
are similar and different.
11:52
And so 'cause certainly analogies to be made to how computation
11:57
works, to how biology operates.
11:58
And so I discussed like two frames of how, how to think about biology one
12:03
and the first of which is the realm of the information where you look at the
12:06
informational aspect of biology and, and you can say, oh, by analogy with computers
12:12
that have binary zeros and ones DNA has four base pairs, the A, T, C and G.
12:17
And so like, oh yeah, also it's, it's, it's storing information somewhat
12:21
differently, but it's storing information.
12:22
Then that information is being processed in a certain way.
12:25
It goes from DNA to RNA and then eventually is turned into these
12:28
proteins, then act on, act in the system.
12:30
And so there's this very sort of computational aspect to it where
12:34
like machinery is taking information and actually making it do things.
12:39
So that's the, the very nice kind of information view.
12:41
But of course there's also the view of the mess, which is biology is
12:46
wildly different from computers.
12:48
It is deeply, deeply messy.
12:50
And so you have like at the organism level, like where these kind of
12:53
like weird wet squelching messes.
12:56
Very different from computers.
12:58
But then once you move down to the cell, the cellular level cells, they
13:02
might compute or do things and and, and manage information, but in a
13:07
very, very different sort of way.
13:09
So if you zoom down to a cell, it's not just, oh, here's some information.
13:12
It's being taken to this thing moving in very nice kind of flow chart.
13:15
It's a whole bunch of molecules just jittering around and
13:18
moving in this massive mess.
13:20
And regularities emerge from that mess.
13:25
But it really it, it's highly parallel.
13:28
It's very random and stochastic, it's a, it's a, it's a qualitatively
13:31
different sort of thing.
13:32
And for me, one of the inter and and that being said, you can still map on certain
13:36
kinds of things to the world of computing.
13:39
But for me, one of the most interesting things is recognizing that biology,
13:44
might still be doing something similar to what computers are doing in some
13:49
very broad way in this, in the case of like information processing.
13:51
But it's wildly different.
13:52
But it actually, what, what it then shows is that information processing or
13:56
computation is this much larger thing.
14:00
That, and biology is one example that's doing it in a very, very
14:03
different sort of way than we might traditionally think of computing.
14:06
And then you also have.
14:07
What happens on our laptops and happens on our computers.
14:09
That's another thing.
14:10
Which means though, that the space of computation is actually much larger
14:14
than May than people might realize.
14:15
And so there's an entire space of what's called unconventional computing, using
14:19
different sorts of substrates, not the kind of traditional transistors and things
14:23
like that to actually compute, to maybe optimize some sort of problem by using, I
14:29
dunno, a slime mold on, on a surface or, or DNA to help like factor large numbers.
14:35
I there, there's all these wild things.
14:36
And so for me, biology can help us expand our notion of what computation is.
14:43
Similarly, and then kind of the, the, the second chapter where I talk about biology
14:47
is this idea of artificial life where you can actually take biology, kind of
14:51
like strip it down to its, to its bones and say, okay, what are the features
14:55
necessary for biological complexity?
14:57
And can we actually them in com in computers?
15:01
And then from there result in.
15:05
Almost like a, like a digital kind of life, but then also
15:08
learn about biology itself.
15:09
Like find out what are the actual true feature, like the necessary features
15:13
of of artificial of, of biology.
15:16
so this, this field, this kind of artificial life, which is
15:18
often the, the description of it's like life as it could be.
15:22
It had its heyday, I guess back in the day in like the
15:25
eighties and maybe the nineties.
15:27
It went through a period where not many people were actually doing it.
15:30
Although my sense is one of the areas where it was still
15:34
actually being done was in Japan.
15:36
So Japan tended the fire while other parts of the world of artificial life
15:41
were not really working on this as much.
15:42
And now we've actually in the past couple years, we've actually seen a
15:45
pretty exciting resurgence of this field.
15:47
But one of the examples is like evolution.
15:49
Evolution is the fundamental feature of life.
15:52
And it turns out you can actually embody evolution and like the process
15:56
of natural selection in a computer because in many ways natural selection
15:59
and evolution, it's an algorithm.
16:01
you.
16:02
You have a population, there's variation, and then they're being selected for,
16:05
and then you have a next generation.
16:06
You do this kind of thing.
16:07
And of course, I mean there's many, many details that I'm skipping over
16:10
and the details really, really matter.
16:12
But the fact that you can actually embody evolution in a computer, this kind of like
16:18
evolutionary computation is unbelievable and actually has taught us a lot about
16:22
how to think about evolution itself.
16:25
I love that, and this is where I was getting to was the idea of the
16:29
Red Queen hypothesis, for example.
16:31
But there's loads in there, and I'd love you to give us a little whirlwind
16:34
tour of these, which was, for example, in the 1940s, there was, an Italian
16:40
researcher who created, , a DNA sequence.
16:42
Well, it was a, it was a computational sequence, but they, it was essentially
16:46
where they competed for resources.
16:49
And then what was remarkable was that parasites would be formed, which
16:54
I just was, that was mind blowing.
16:56
And then this was repeated again a few decades later with the Game of Life.
17:00
I'd love you to tell us about those, because these are just mind
17:03
blowing when you realize this.
17:04
so, so in going back to this idea that inherent in the history of computing
17:11
are often these ideas that may might, they may take a long time to germinate,
17:16
but a lot of these fundamental ideas of connecting to biology, connecting
17:21
it to certain aspects of simulation or artificial intelligence, like these
17:24
things were all inherent very, very early on in, in in the realm of the computer.
17:30
so there was this one man Neils Barricelli, or Barricelli who he worked
17:37
with the the Institute for Advanced Study.
17:40
So John von Neumann who is in and amazing, I think considered to be like one of
17:46
the smartest humans who ever lived.
17:48
He was involved in one of the one of the first computers, digital
17:52
computers that was developed at the Institute for Advanced Study.
17:56
In fact.
17:57
The architecture for all.
17:59
I wouldn't say all, but many modern computers derives from that
18:02
architecture, which we actually call the Von Neumann architecture.
18:05
And one of the first things that was done with this computer that was developed,
18:09
I guess like again, like the late 1940s, early 1950s in addition to large
18:14
scale simulation were these digital life, artificial life experiments.
18:18
And so, this researcher, yeah, he had these strings and they
18:21
all competed and yeah, were, were doing these kinds of things.
18:24
I don't know if the parasite thing was in there, and that might have actually
18:26
been in a later one called Tierra where
18:30
That's
18:31
right.
18:31
so there was a, there was a later experiment or program
18:34
I think maybe in the.
18:36
In the 1980s, I can't recall exactly where that would, that kind of, it took
18:41
some of these ideas, although I'm not necessarily sure he actually took those
18:43
ideas consciously or deliberately because a lot of that work from like the 1940s
18:47
was forgotten for, for a very long time.
18:49
And and it was only sort of, I dunno, re reinvented or or re not,
18:54
not quite rediscovered, but re popularized more, more recently.
18:59
But the the program Tierra is this really interesting thing where the idea is you
19:05
have, rather than having a like a like various different organisms or formulas
19:14
that are competing and you and you actually see which ones are the most fit.
19:17
The, the actual organism is itself a computer program.
19:20
There was very tiny little sim simple computer programs that
19:23
were competing for computational resources, if I recall correctly.
19:28
And one of the things that they found is like when you run this for long
19:31
enough, and actually fairly quickly, I think you would get parasites, you
19:34
would get snippets of code that cannot reproduce themselves, but are able to
19:40
reproduce due to the fact that they're working with like self-contained
19:45
organisms, like their own programs.
19:46
And and people have actually been building lots of other kinds of, of, of
19:50
programs as well to try to understand these, like the nation or the, the, the
19:54
nature of evolution within computation.
19:56
But another realm, and you mentioned John Horton Conway's Game of Life
20:00
is this idea of cellular automata.
20:02
So cellular automata are these very, very, and actually going back to John
20:07
von Neumann, he was actually one of the people who helped develop the
20:10
idea of cellular automata as well.
20:11
So there's a, these very deep histories, but so so as, so cellular automata are
20:17
imagine you have a grid and the grid has lots of squares and the squares
20:20
can have a number of different states.
20:22
And so, the simplest would be we can use the game of life as kind
20:26
of example, simple as imagine each square can be either on or off,
20:30
black or white, zero, one, and each generation, each, each time step.
20:37
The way in which you decide whether or not these these dots, the squares switch
20:42
on and off is you look to the neighbors.
20:44
And so and the, the way the rules are is that if if it's on, if, if a cell is
20:50
alive, that, like, if it's, if it's a black swear it looks around and if it's
20:55
like, I think you, it's not overcrowded.
20:58
So there's not too many neighbors and there's not too few neighbors.
21:01
So it's not dying of loneliness.
21:03
It continues to persist.
21:04
And if it has, and, and I think if a white square that so it's not alive
21:09
has exactly like three neighbors, then it is born and, and comes alive.
21:12
And then there's a, i i, it's a high level way of describing it.
21:15
And so it sounds very simple, you would assume, okay, you get maybe some like
21:18
little blinking things and nothing particularly exciting, it turns out.
21:23
Using these very, very simple rules.
21:26
This, this example of a, of a, of of cellular automata game of life, gives
21:31
you this unbelievably rich behavior.
21:34
So you have, you have these things called gliders, which are like a set of
21:37
like little little squares that, that they all move together and they waggle
21:42
across, diagonally across the screen.
21:44
You have a glider gun, which actually generates a, an
21:47
infinite amount of gliders.
21:49
People have actually made entire computers inside game of life.
21:53
So it actually is effectively equivalent to modern programming languages.
21:56
There's a huge amount, amount of variety.
21:59
And so this was developed by this mathematician, John
22:01
Horton Conway, around 1970.
22:04
And then soon after Martin Gardner, the the, the, the Science and Mathematics
22:10
journalist, he wrote about it in in a column in Scientific American.
22:14
And at the time I think after it, after this column was published, it was.
22:20
So popular and so many people basically just stopped whatever they were doing
22:23
and tried to find a computer to use to actually embody this in code.
22:27
That I think he estimated that like, I dunno, like millions of dollars
22:31
of computer, computer resources were, were, were dedicated to
22:34
the game of life, which is wild.
22:36
Also, when you think back, it's like 1970.
22:38
There weren't really, and there weren't personal computers.
22:40
These were all like large computers that were being, that were at
22:44
universities or companies or things.
22:45
And so people were actually using all this, these resources for that.
22:48
And of course since then.
22:50
have been continuing to develop more sophisticated cellular.
22:53
So there's another one that kind of uses a, a smoother version of this.
22:56
So instead of instead of the the discrete time steps and like the
23:00
squares of on and off you smooth it out to make it more like the real world.
23:04
And so it's called, there's this version that's called Linea.
23:06
it turns out there is a whole like bestiary of organisms that are
23:11
persistent in these digital worlds.
23:13
And it turns out you can, you can actually embody this all of millennia
23:16
and I think, I think in like 10 lines of code, which is, which is insane.
23:20
And so yeah, it's to, to see all the different ways that you can take
23:25
various aspects of biology, whether or not it's the sheer open-endedness
23:29
of biology which is still an open question of whether or not we can do
23:33
that fully within, within computation.
23:35
Certain aspects of evolution, certain aspects of persistence
23:38
and resilience and robustness.
23:39
All these things can actually be tested and embodied within computers.
23:43
It's incredible when you think of it that way.
23:45
And that's why I know you, you bring people on a journey in this book.
23:49
And for me it was quite a privilege to have read the other books because I kinda,
23:54
I'd seen where you were coming from in that early parts of the book, but for this
23:57
book, it was necessary to get people to part three, to bring them on that journey.
24:01
But there was a couple of things I just wanted to pull at.
24:04
One was where this all appears is in the chapter, ghosts in the Machine,
24:08
and there's a movie about that.
24:10
And it's an animé if it's an animé Japanese story, which is great.
24:14
I think Scarlet Johansson is the main protagonist.
24:17
But the idea of there will be emergent behavior in computation.
24:22
Therefore, if you have artificial intelligence, there will be.
24:25
Emergent behavior as well.
24:27
And I thought it was interesting.
24:28
There was a little line I just pulled out and I just would love
24:30
you to elaborate on this one.
24:31
You said all around us are clues that if the tape of life on earth where
24:36
we run it would look very different.
24:38
And I was like, that's an intriguing line.
24:41
Let's elaborate on that one.
24:42
Yeah.
24:42
And so an evolution ultimately is entirely contingent.
24:48
And so it is not a given that if intelligence arises, that intelligence
24:55
will come in the form of homo sapiens, of looking exactly like humans.
24:58
There's many, many different ways that evolution can solve problems
25:03
and, can solve problems of adapting to different types of environment.
25:06
And this is because like evolution by natural suction operates by
25:11
dealing with variation and exploiting random mutation and things like
25:14
that to then create organisms that are fit to their environment.
25:19
And of course there's coevolution, all these other kinds of things, but
25:21
if you look around, I mean, there are no dinosaurs anymore, but like.
25:26
Were dinosaurs not particularly fit for their environment.
25:29
They were actually very fit for their environment until there
25:32
was a massive cataclysm that kind of destroyed their environment.
25:36
And so it like the fact that mammals are now an intelligent,
25:41
mammals are the dominant form of life on the planet is not a given.
25:46
And, and you, if you look at the historical record, like the
25:49
fossil record, you see all these weird evolutionary dead ends.
25:53
Sometimes they were dead ends because maybe they were not fit.
25:55
Other times though, it's because of some sort of environmental
25:59
disaster, whatever it is.
26:00
And so you can see all the different paths that evolution.
26:04
Could have taken and and you see this, and sometimes you can even say, see it
26:07
in terms of even just the, you can see a little bit of that evolutionary history
26:11
or the history of the world in some of the things that have not adopted yet.
26:14
And so we talked in about when we were discussing our conversation with like
26:18
over complicated about like the honey locust tree that still has the giant
26:21
spikes to protect against megafauna of North America that don't actually exist.
26:25
And so there's just all, and, but then you also have I things
26:28
about like when you have these.
26:31
Environmental disasters, you can end up with a very small, non-random subset
26:36
of a population that then actually becomes the population for the future.
26:41
So there's actually so this is called like a population bottleneck, where
26:44
there's like some situation that causes a very small population to move
26:48
forward and then repopulate some area.
26:51
And of course if those, if the small number have some weird genetic trait
26:56
or whatever it is, then that is going to be the thing that is quite common in
27:00
the resulting population, even though it's not necessarily something that's
27:03
fit, it's just that's where it is.
27:04
And so I believe , there's some island that I think has a very
27:06
large amount of Colorblindness.
27:09
Pinglap
27:10
Yeah.
27:10
Right.
27:10
And I think, and I think the thinking is like, yeah, there was just some
27:13
sort of population bottleneck.
27:14
There was some sort of disaster and you can see, right, like there, like, like
27:17
certain populations might have relatively high rates of certain genetic diseases and
27:22
in certain situations it could be because population that had that relatively
27:26
high rate populated that new area.
27:29
And so that kind of led to certain things.
27:30
And so there's always a certain amount of weird contingency.
27:35
And and as a result, when you run these computation algorithms, you're
27:41
not always get going to get the exact same result, especially if
27:44
it's like some really complex problem that you're trying to solve or some
27:48
complex thing you're trying to do.
27:49
There might be many, many different ways of solving that kind of thing.
27:52
Equivalently.
27:53
I always think about that.
27:54
For example, with the great controversy of vaccination, do
27:58
you get vaccinated, do you not?
27:59
And you're going, well, that certainly plays it out differently
28:03
for different populations who did or didn't get vaccinated, et cetera.
28:06
We have.
28:06
Have no idea how that would die.
28:08
How will that will play out in the future?
28:10
And then I'm not gonna go there.
28:11
But then there's the whole quantum field is they're all happening.
28:15
All those different choice happened at the same time.
28:18
Mind blowing stuff.
28:19
But, but speaking of which we now get to the true end point of all this talk
28:24
about reality, physics and computing smashing together, as you say in the book.
28:28
And just as we can focus on the realness of computing, it's profound physicality.
28:34
What if we can posit that reality itself is a program found within a computer
28:39
as some kind of grand simulation?
28:41
So this is the core message you get to, well, not really the core
28:45
message you get to, but the, what you build towards in part three,
28:49
which is the simulation hypothesis.
28:51
And I thought we'd get there rather quickly because this is
28:54
something that people really find it hard to get their head around.
28:57
But actually reading your book was the easiest way I've got there.
29:01
And maybe we'll even rewind a little bit Samuel and go to stuff
29:05
like Sims and games you grew up with and played quite proficiently.
29:10
I mentioned, for example, SIM City earlier on, but how, how they work.
29:15
And then actually, well, if computing power is so powerful,
29:19
how would they work going forward?
29:21
Yeah, I mean, so yeah, taking a step back and like thinking about
29:24
like simulation in computers, right?
29:26
Simulation has also been one of these features of.
29:30
Digital computers since the very, like, the very early days of computing,
29:33
that was like often one of the things they were doing was like trying to,
29:36
I dunno, trajectories or nuclear simulations, things like that.
29:41
And the question becomes, right, how can we embody all the details
29:47
and complexity of reality within a machine and understand it in some
29:51
sort of reasonable and meaningful way.
29:54
'cause there's always going to be a distinction between
29:56
the map and the territory.
29:58
Like the model you make is always going to be some sort of simplification,
30:02
sometimes an like necessary simplification because you actually
30:05
wanna an abstract way details in order to understand some specific thing.
30:09
Sometimes it's also just because we don't have the, like a huge amount of
30:12
data or the necessary amount of data.
30:15
We don't necessarily, we might not necessarily understand all of
30:17
the forces that are relevant for the, the features of the world.
30:20
We're trying to understand being said.
30:24
And, and they're also just fundamental constraints of like
30:26
certain things around like.
30:27
Like chaos, a chaos theory, whatever, where like you run these things and it's
30:31
very hard to actually predict things beyond a certain time, time horizon.
30:35
That being said, like if you look at like weather for example, we've actually
30:37
gotten much, much better through the, the combination of like much better data
30:43
better algorithms where we can combine different models together, as well as just
30:46
massive advances in computational power.
30:48
And slowly but surely we have gotten better and better at understanding
30:52
the nature of the weather and actually being able to predict it.
30:55
of course, prediction is not the only reason.
30:58
We actually simulate and we make models.
31:00
And I alluded to that before when I said like, you, you can sometimes
31:03
make simplified models where you just wanna understand some
31:07
specific feature or with SimCity.
31:10
And SimCity certainly had an impact on a lot of people who
31:12
ended up going into city planning.
31:14
But it was a necessarily simplified version of the world
31:18
and probably so simplified that it was like wildly and accurate.
31:21
That being said, it was a lot of fun and it also gave certainly me and I imagine
31:26
many other people, a deep understanding of the massively non-linear nature of the,
31:31
of the world where like, if you are truly trying to understand a complex system, you
31:35
need to see how things have second order effects, unanticipated consequences, how
31:40
decisions bite back, how the system bites back, all these different kinds of things.
31:43
And you got to experience it in Sim City even though it was not making
31:46
any claim that it was the real world.
31:50
Now the question becomes right, like if we make these simulations
31:53
more and more sophisticated, what could the potential endpoint be?
31:57
And you also mentioned kinda like the physicality of computing.
32:00
I'll just kind of take a, take a step back and mention that with like we, we
32:04
think of code as this, like ephemeral thought stuff and in many cases it's
32:11
useful to think of that that way, Computation ultimately is deeply physical.
32:16
And like we're running things on on actual computers.
32:21
We have data centers that require huge amounts of energy.
32:23
And, and you can see like the deeply physical nature of computing, even
32:26
like in weird glitches and failures.
32:28
So there was this story I read, think it happened a couple years ago where
32:33
it was in some hospital and people noticed that their, like, that their
32:37
iPhones and their Apple watches stopped working around some MRI machine.
32:41
But Android devices were actually fine.
32:42
And it turned out it was that this MRI device had a helium
32:46
leak and it turned out helium.
32:47
The helium atoms were just the right size to like interrupt the switches in
32:51
Apple devices, but on Android devices.
32:53
And so we often like forget about the deeply physical nature
32:56
of computers at our apparel.
32:58
so when you think about what are these, what are the end points of, okay.
33:03
amounts of computational resources alongside kind of the, the end points of
33:07
computational, like how much computation you can actually make in like, I
33:09
dunno, a kilogram or whatever it is.
33:11
You end up doing these very weird thought experiments of thinking
33:14
about, okay, what are the true endpoint of computational resources?
33:17
And then that leads into can you actually simulate, what can you
33:20
model and could it be then that we are maybe inside a simulation.
33:26
So like the idea, this is the idea of the simulation hypothesis, that like
33:28
everything we see around us, like you, me, ourselves, the entire world,
33:34
we are all inside like a version of Sim City and just don't realize it.
33:39
And so, and for me, I mean, to be honest, I'm much less concerned with like
33:45
whether or not this is, this is true.
33:47
Like I, I don't view it as like this, this question of like great cosmic importance.
33:51
I feel like that's more kind of like the, like, like the kinds of, like the
33:54
kinds of discussions that is better for like undergraduates late at night talking
33:58
about versus for me, when I think about the simulation hypothesis it also, it,
34:04
it often feels like this kind of cry for mythos or like great deal of meaning.
34:09
And so if you look at like societies, we always have a need for mythological
34:14
tales or large overarching theories to understand the world.
34:18
And you see this even in certain kind of like theories within within
34:22
the realm of science and technology of like people talking about like
34:26
singularity or super intelligence or firm, these paradox and aliens.
34:30
And then also the simulation hypothesis.
34:32
I feel like it like slots into that place where people are
34:35
trying to find some bigger thing.
34:37
For me, I don't necessarily, I, I don't really find that compelling as
34:42
like, oh, like it's this big thing.
34:44
I view the simulation hypothesis as provocative in the sense of.
34:49
If we are in a computer program, let's think about what it would mean to be
34:53
inside a computer program and, I don't know, inject unexpected bits of code.
34:57
And so then that kind of leads you more towards, less towards the realm
35:00
of like cosmic importance and more towards thinking about glitches and
35:04
failures and like code injection and weird and weird, just like exceptions
35:08
and things in the realm of computation.
35:10
So for example there, there is so Super Mario World, which I think is for, it was
35:15
for the Super Nintendo like decades ago.
35:18
A group of people figured out that if you move around in this game
35:23
in a certain specific way and like jump in specific specific locations
35:28
and at certain timing you can actually reprogram the game itself.
35:32
And then eventually you can like, and so I, I, I watched this video where someone
35:35
did this in such a way and then turned Super Mario into a version of Flappy
35:40
Bird, like an entirely different game.
35:42
And which is not to say that like, if I just jump around in my office a certain
35:46
way, I'm gonna reprogram reality.
35:48
But these are fun thought experiments to think about.
35:51
And so, and, and it can also lead you to think about going back to like
35:54
the deep physicality of computing.
35:55
Like what are, like, what are, what are the end points or the the limits
36:01
of simulation, the limits of physical computing, what would that actually mean?
36:06
People have actually tried to, tried to calculate this and
36:09
explore these kinds of things.
36:10
And so for me, those are the kind of things that, that the
36:12
simulation hypothesis makes me find mantra much more interesting.
36:16
So for me it's like not like, oh, like this is the thing I need
36:19
to think about and worry about.
36:20
It's more simulation hypothesis is a wonderful provocation to get
36:24
me to think about more really big questions about computation itself.
36:28
I have a similar kind of approach to it.
36:29
Like, it's interesting and all those theories are interesting, but it's
36:34
more what do they mean to you with your beliefs and your belief system as well?
36:38
And also how do they help you?
36:39
Like for example injecting the code could be, well, I can actually have
36:44
a, a certain amount of control over my reality if if I, if I behave a certain
36:50
way, then the universe behaves back to me in a certain way and it creates this
36:54
feedback loop that actually can lead me on different pathways and stuff like that.
36:57
So they're the kind of things it, it forces me to think about as well.
37:02
But there's a great le there's a great saying that it's hard to
37:04
read the label when you're inside the jar and actually it's perfect
37:09
for the simulation hypothesis.
37:11
Well, how are you, how would you know?
37:12
And
37:13
. in the absence of like very clear evidence, right?
37:16
It's just a fun speculation.
37:18
And so for me it's like where, where can that fund speculation
37:21
lead in insightful ways about how we think about computing and reality?
37:25
That being said, there was a great novel that came out.
37:28
I think it's sometime in the past year it's called When We Were Real.
37:31
And the premise is like, I dunno, seven, eight years before the story starts,
37:35
world is revealed to be a simulation.
37:36
And then in the story itself, it is a group of people going on a bus tour
37:42
across the United States to visit all of the locations that are like basically
37:47
glitches in reality, that are these like roadside attractions that provide
37:50
evidence of the simulation hypothesis.
37:53
And it is wild and fascinating.
37:55
And you also just see and he explores like how people would respond to these
37:58
kinds of things and think about it.
37:59
And yeah.
38:00
I highly recommend that book.
38:01
It's fantastic.
38:02
It sounds like going to the certain places in Super Mario World
38:07
jumping over there in the corner.
38:08
And actually, just to mention Carl Sagan who noted, and you say this in
38:12
the book, "we have become a mechanism for the universe to study itself,
38:16
understanding physics and biology.
38:18
And even computation itself.'
38:19
I love that little quote, which brings us to the wisdom of
38:23
computation, the final chapter.
38:25
And that chapter I felt was less about machines, and it's a shift towards
38:29
mindset again and again while it's clear you're a lover of code and the history
38:35
of code and probably one of the most researched people I've ever read on this.
38:40
I left a review on Amazon and I said this, I was like, on Imagine a polymath
38:43
who was fascinated by code and language and a plethora of different subjects
38:49
driven by curiosity, writes a book.
38:51
Well, that's what this book is.
38:52
And,
38:53
you.
38:55
and I I just wanted to touch on this.
38:58
I, mentioned language.
38:59
Actually, let's, let's go there quickly because I, thought this was
39:03
a really interesting point you made.
39:05
You said that language is.
39:08
Really how we think and language that Wittgenstein quote, the limits
39:12
of my language is the limits of my mind, or something like that.
39:15
But also, I thought this was fascinating that you said our LLMs
39:19
can be biased because of our language.
39:22
Because at the bottom we train an LLM on text and it learns from our stories.
39:28
A narrative is based on what yields a good yarn, one that slices away
39:32
extraneous or complicating detail.
39:34
There is always a distinction between the map and the territory.
39:37
Since these models are ultimately built upon language, we might forever
39:42
be stuck in the world of stories, simplifying narratives and tales that
39:46
lack loose ends or confusing details.
39:48
I thought that was a really interesting insight that we might share.
39:52
thank you.
39:52
Yeah, yeah, right language, right?
39:54
Especially with these.
39:55
Large language models that we're building, they are built on
39:58
these like mountains of texts.
40:00
Now it's, it's changing a little bit more as we have like multimodal
40:02
ones, which are using images.
40:04
And maybe when we're beginning to train systems for robotics, they're actually
40:08
going to engage in the real world.
40:10
the large language models, they are trained on the amount of text
40:15
that we have generated as a society and as a species and text, right?
40:20
Like the text in the world and our stories, right?
40:22
That is not reality.
40:23
It is necessarily simplified and we forget this at our peril that the way
40:29
these systems work is they are trying to map things on to, and sometimes like
40:35
literary tropes, other kinds of things, or like certain, like in the same way
40:39
you see like a certain certain chatbots will have like certain personalities or
40:43
personas that they will take on because that's the thing that you see in, in text.
40:48
And so there a number of people have actually talked about this as I think
40:51
they describe it as like semiotic physics, where there's there's a
40:53
physics in the system, but it's not necessarily the physics of reality, it's
40:56
the physics of, of language and text.
40:58
And so of the examples someone did was like having these systems like
41:02
generate random numbers and it turns out these random numbers are not actually
41:06
statistically and mathematically random.
41:08
They're things that maybe feel random for humans.
41:10
And I actually re, I recently did this with Claude, where I said, okay, generate
41:14
like a hundred numbers like a hundred random numbers, like between one and 10.
41:18
And so it gave me this whole string of numbers.
41:21
And then I said, okay, now make a histogram.
41:22
And as soon as it made the histogram, it responded, oh, these
41:25
numbers are not random at all.
41:27
Because it was like, like the exact same number of, and I, and then I said, okay.
41:30
Then we had a little conversation about like like, I'm about this.
41:32
And I said, okay, now that you know this, now do it again.
41:34
And it did it again.
41:35
And it was like a little bit better, but it's still recognized.
41:38
It was not anywhere near random.
41:39
It's because right, like these systems, they are trained on, not on reality.
41:46
And you can, sometimes you can side it by having it actually write
41:49
code to do pseudo, like actual like math, like computational, pseudo
41:51
randomness as opposed to the weird things that feel random to humans
41:55
that are not actually random at all.
41:57
, But yeah, we're stuck in this kind of thing where it's almost like
42:00
the situation where if you know these narrative tropes, you will.
42:04
Better.
42:04
And like in the way stories are written or constructed, you'll better understand
42:07
how these systems make meaning.
42:11
And actually the the novelist Robin Sloan , his most recent novel Moon Bound, actually
42:16
discusses this kind of thing of like these systems can be, have huge amounts
42:20
of information, but they are ultimately bound by the realm of text and the kinds
42:26
of text that we generate as a society.
42:28
And that causes these kinds of things to be, to be biased away from reality
42:33
and towards certain aspects of language.
42:35
And so that's one of those things where we just have to be very
42:38
aware and returning to humility.
42:40
Like we have to have humility about the, the powers of
42:43
these kinds of, of of systems.
42:47
It's why I I write a weekly article.
42:49
I've included you in many of them, and actually in a few forthcoming as well.
42:53
But I, I'd never, I'd never use AI to come up with my ideas.
42:58
I'd feel like such a, I'm not saying you are.
43:01
If you do, if somebody's out there using it.
43:03
But I would feel like I'm a fake in some way and also that I'm not,
43:08
I'm not adding any value because I haven't drawn from my own reservoir
43:13
of thinking and ideas that marinate.
43:16
And it really leads me to this point that AI can help you maybe structure it.
43:22
And even when it does that, I don't like what it does because it,
43:25
it writes it a way that is too I dunno, it feels too sterile for me.
43:30
It's not human and broken enough when it, when it produces something, but I
43:35
don't feel I'm adding any value that's different and something that's truly human
43:40
and draws from very disparate fields.
43:44
Much like your writing as well, where you're drawing from different places.
43:46
I like to write an analogy and metaphor and I often find that by doing the show
43:51
and reading in these different areas, EC collectively you can come up with
43:55
great ideas that are just different.
43:56
And then you train it and you have a beautiful line where you say that.
44:01
AI or code is a model for thinking, not a replacement for it.
44:05
I think that's a really important aspect that we need to get across.
44:09
Yeah, no, I, I completely agree that like ultimately, right, these systems should
44:13
not be replacing the things we want to do.
44:16
They should be augmenting, there should be the kind of
44:18
this human machine partnership.
44:20
And so right, there's this whole realm of computational tools for thought,
44:24
like these tools that kind of help us to make connections or store notes or
44:29
better understand models of the world.
44:30
And these are very, very powerful.
44:32
And AI is certainly one of these kinds of things.
44:35
And and being said, though as AI is getting more and more powerful
44:42
it can cause worry of like, oh my God, like, can it do all the
44:45
different things that humans can do?
44:47
And so for me I view, I, I think about this kind of by analogy
44:51
with an idea from, from theology called like the God of the gaps.
44:55
And the idea behind this is.
44:56
You have, like what is like in this case, like how would you define God?
45:00
Oh God.
45:00
Is anything that science has not explained yet?
45:03
Now of course, that it, it, there's a fallacy there because as we develop more
45:08
and like astronomy and cosmology and, and evolution suddenly that conception of God
45:13
gets smaller and smaller and evaporates.
45:17
And, and I feel like we're doing the same kind of thing when we think about ai,
45:22
where we say, oh, like this AI system is impressive, but but it can't do the
45:27
following things that humans can do, like X and Y. And of course you wait
45:30
like 10 minutes and then suddenly AI can do those kinds of things as well.
45:33
And so for me it's less about saying, okay, what is essential to our humanity
45:37
of like, oh, we're two users, which then I know we've discovered alien or not, not
45:41
aliens, but we discovered animals can do these kinds of things or for language.
45:45
And the same kind of thing we're seeing in AI of like, oh, AI
45:48
can't do this, but now it can.
45:50
We need to focus less on like what is unique about humans and what is.
45:56
to our humanity.
45:57
And, and that is, that is something that each of us need to figure out.
46:00
Like what is quintessential to each of our humanity is like the
46:02
things that we truly want to do.
46:04
It can be thinking really well, it can be spending time with our children.
46:08
It can be gardening, it can be doing hobbies, things like that.
46:11
You wouldn't want a system to do those things for you, even if they could, if
46:16
they could do them, because these things are quintessential to your own humanity.
46:19
And so you have to find ways of using these tools to make sure you make
46:23
yourself the best version of that.
46:24
And so this kind of goes back related to the tools for thought idea of a
46:28
phrase from Steve Jobs called a, a computer as a bicycle for the mind.
46:33
And so.
46:34
idea is that it was, I think he, it was based on, he read an article in
46:38
Scientific American where there was this chart looking at like the, the energy
46:41
efficiency of different organisms.
46:43
And you have like, I dunno, certain like birds are maybe more
46:45
efficient or maybe less efficient.
46:46
I don't remember exactly what it was.
46:47
And like humans were okay, but they also include a human on a bicycle.
46:50
And suddenly everything changed.
46:51
Humans are much, much more efficient.
46:54
When it came to actually moving, like mobility on a bicycle and Jobs's
46:59
idea was like, computers are the same kind of thing, but for the mind.
47:02
Like they're like, they're a bicycle for the mind.
47:05
And we have to make sure that we constantly have that in hand when we're
47:10
thinking about these technologies.
47:11
They should augment us and make us more efficient and better versions of
47:15
ourselves, shouldn't necessarily guide our decisions entirely or remove the
47:23
actual things that we want to be doing.
47:24
And so they should be like, there should be this kind of human machine.
47:28
Symbiosis or partnership or whatever this, this kind of productive relationship
47:32
that allows us to actually retain what we view as our quintessential humanity.
47:36
Which brings us back to the Red Queen hypothesis.
47:39
When I thought about that, I thought again about the idea that we're seeing a bit
47:43
of a red queen, race, an arms race with ai, and most of it's in pursuit of power.
47:51
Or for example, I don't want to be the country left out here, so
47:54
I'm going to do it quicker, and, oh no, well, let's, let's cease,
47:59
let's cease the race for the moment.
48:00
And then you go, don't cease the race, keep going.
48:03
And that's a whole challenge that we're at the point in history as well
48:07
because we are opening up a Pandora's box and I'd love your thoughts on that.
48:13
Yeah, I mean, it's, it's one of these situations where, right as we
48:16
are developing more and more complex systems and complex technologies and
48:19
more advanced technologies that require huge amounts of resources, right?
48:22
To unilaterally disarm feels like, right?
48:27
Like just basically like emitting defeat.
48:31
And so how do you do that kind of thing?
48:35
And like how do you retain your humanity or kind of feel as if like you're choosing
48:39
to adopt certain things as opposed to just this like wave of technological
48:43
advancement kind of washing over us.
48:46
I do think, I mean, it's hard, but I do think we actually, we still do
48:50
retain a certain amount of agency and control and like choice in which
48:53
of the things we choose to adopt.
48:55
It is, it is very tough sometimes.
48:58
But I, but I wonder if like will be the ability to actually building.
49:06
To build things that are more at a human scale where we can make things more in
49:10
line with our humanity as opposed to just having big tech forced upon us.
49:16
And so and, and that kind of like that small human scale, I think that is
49:22
probably the key of like, how do we, like, how do you make things in line with
49:25
our humanity and what we actually want?
49:27
You probably have to make software and technology that feels not, like,
49:32
not so big and not so IM like not so much of an imposition upon us where
49:37
it's, we are not, are not bending ourselves to our technologies.
49:42
We are in fact using technologies to make us the best versions of ourselves.
49:45
And for me, I see hopes in that I guess maybe surprisingly, by
49:51
using some of these AI systems, which is around the idea of like
49:55
democratizing the creation of software.
49:56
So for a very long time, I mean, software has, like, you have to be.
50:00
Really, really well trained to under, like, to, you have to, it takes a lot
50:05
of effort to actually make software.
50:06
But there's been a dream for, for many, many years of actually allowing
50:11
a certain democratization of this, where each and every person can
50:14
build the software that, that they want for whatever needs they want.
50:18
And so actually to return to the novelist Robin Sloan, he has this, this wonderful
50:21
essay where he talks about an app as a home cooked meal where like, in the same
50:25
way that like when I'm cooking, I'm not cooking for like thousands or millions of
50:28
people I'm cooking for me or my family.
50:31
Like you can do the same kind of thing with software.
50:33
Like you should be able to build things just for yourself
50:35
at that kind of smaller scale.
50:37
And intriguingly ai, especially with like the advent of vibe coding and code
50:41
generation, it really is allowing us to do those kinds of things now, to
50:44
build much more sophisticated software that would require like, like, like it
50:48
being secure and maintaining privacy.
50:50
So like you need to really know how to actually.
50:53
Like build software, like you need to actually, how know how to code.
50:55
You can't just take the thing as given.
50:56
It could be not not so great, but for building something for yourself
50:59
or for your loved ones or whatever.
51:01
That's great.
51:02
And so I actually think this democratization of software creation
51:06
might actually allow us more easily build a world of technology that is more
51:11
in line with our own needs and desires and wants and on that human scale.
51:16
And to go back to something we talked about before, the limits of
51:20
human consciousness or limits of our cognitive abilities are overrun.
51:24
Then, for example, so you know one of the things, you're a polymath that's clear
51:28
from reading your work that all these different realms that you've drawn on.
51:32
That actually, that's not something many people can do.
51:36
And as we talked about in the last episode, we need more people with
51:39
that type of range or that type of breadth of knowledge because there's
51:43
often solutions for your challenge in a totally different realm over here.
51:47
And nobody's connected the dots.
51:48
And that's one of my hopes that AI can do is, is if it is fed these different
51:54
realms that you can actually solve big problems in different realms, different
51:58
arenas that are, have been hither to unseeable for, for many people as well.
52:03
But I mentioned the, the mindset piece, and I thought this was so
52:07
important that you talk about the importance of humility, restraint, and
52:12
curiosity comes up time and time again.
52:14
But alongside this humility, you say, we must think about code as something
52:19
that must be developed in line with our humanity rather than as something
52:23
that tries to subvert its nature.
52:26
To re to reiterate what was explored in the chapter on tools for thought,
52:31
we must avoid bending ourselves to the needs and whims of our machines.
52:35
Computation should not narrow our humanity.
52:38
The real world is messy and wonderful, and we must be deliberate
52:42
in how we balance all of this.
52:43
I, absolutely love how you wrote that there as well.
52:46
Oh, thank you so much.
52:47
I really appreciate that.
52:48
Yeah, no, I think, like that is really the goal and I feel like sometimes
52:52
when we get so obsessed with Yeah.
52:55
Technological advance for its own sake or kind of just like things that seem amazing
53:00
and powerful, like we forget, right?
53:01
Really what is ultimately the goal.
53:02
And so there was there was a TV show that came out like a decade ago
53:08
or so called halt and Catch Fire.
53:10
And it was about the early computing industry.
53:12
I think it went from like 1980 to like the mid nineties or so.
53:16
But early on in the show possibly even in the first episode, someone talks
53:20
about and they say computers aren't the.
53:23
They're the thing that gets us to the thing.
53:25
And it's all like, too often we, we just constantly want to advance our
53:29
technologies for their own sake without realizing no, there is an end point there,
53:33
which is like making us better versions of ourselves, making us in line with
53:37
our humanity, allowing us to, yeah, help us connect different ideas, acting as
53:42
tools for thought, making us making us better able to just live in the world.
53:47
Like these are all the things that we should be focusing on.
53:49
And so I think as long as we have that in mind of recognizing that these are
53:54
ultimately tools we have to choose how we actually use them in a way that
54:00
makes us the best versions of ourselves and in line with our humanity, then I
54:04
think we'll hopefully be in good shape.
54:06
I have a final quote that , I loved and this where you were
54:09
talking about your daughter, who is intrigued by watching you drive and
54:15
Looking at places and how you'd react to the traffic lights preparing for
54:19
her future of driving and you're like going, Hmm, , let's not talk about that.
54:23
Yeah.
54:23
'cause it might not happen because of autonomous vehicles, but here, there
54:27
was a little quote , that I'd love to share as my finale as I'm doing that,
54:30
I'd love you to think about your final message, but you wrote here beautifully.
54:34
"Change happens, but we live among the remnants.
54:37
Ruins is probably too strong a term of these forgotten technologies.
54:42
Cars have CD players for compact discs few people buy these days.
54:47
Libraries allow you to borrow DVDs even if you stream all your
54:50
movies and don't own a DVD player.
54:52
You view these less as the inertial burden of technology and more as layered mementos
54:57
of people's predictions of the future.
55:00
They are reminders of our inability to recognize that the future will be
55:04
different from what has come before.
55:06
Technological change is not a simple extrapolation of what we have
55:10
now, only a bit faster or sleeker.
55:13
It will be qualitatively different.
55:16
Now, there I share that for many reasons.
55:18
I love what you wrote there, but also, I always think about the poem of
55:22
Ozymandias and, you know, the, the great king that's has this statue to himself
55:28
and it's blowing in the wind and that.
55:30
Everything we do in life, we can fall into the trap of thinking It's so, so
55:35
important and forget that you gotta live and you have your family and
55:40
friends and all that kind of stuff.
55:41
And people fall in love with their job or, or don't, or just falls in
55:45
love with the dopamine hit and the stress of doing that job over and over.
55:49
And these kind of little quotes here remind me of that, that it's not really
55:54
that important and it's hard to read the label inside, you're inside the jar and
55:59
sometimes those kind of quotes where you read all those people who built all those
56:03
technologies, gave up so much to do so.
56:07
And then I go back to what you talked about earlier on about the idea of
56:11
asymptotically creating new technologies.
56:14
So you talked about the guy in the 1940s who, the Italian gentleman
56:19
who created the code sequences probably wasn't even remembered, but
56:23
that technology was built upon then later, decades later with the, the.
56:29
Oh, te terror, what was it called?
56:31
The game,
56:31
Yeah.
56:32
the Tierra Game or the Tierra sequence.
56:35
And surely that will be built upon in the future, and people forget about all
56:40
those people who have built a little bit towards where we're going in the future.
56:44
But it was so, so important for those people in the moment.
56:48
And I just loved what you wrote there as a reminder of that.
56:51
And that don't take it too seriously.
56:53
Be serious about it, but don't take it too seriously 'cause life is more important.
56:57
But I'd love you to close our finale of these series.
57:00
I've absolutely loved talking to you, learned loads taken, butt loads of notes
57:05
as I, I've shared with you as well.
57:07
But what's your final message and why you do the work that you do?
57:10
Yeah, no, thank you so much.
57:11
I mean yeah.
57:13
Final message.
57:14
I mean, and ultimately the way I right explore the world, right?
57:18
It's, and right there is a great deal that is, not known.
57:22
We don't necessarily know where these technologies are gonna lead
57:24
or how scientific advances are going to change and grow over time.
57:28
And so for me, certainly studying technological history
57:31
is incredibly valuable.
57:33
Understanding how ideas have changed over time is also like vital for
57:38
understanding these kinds of things.
57:40
But I would say, ultimately I'm driven by these two ideas of humility and curiosity.
57:45
Like humility in the face of right.
57:47
Our technologies are constantly changing.
57:48
It's very hard to predict what is next.
57:50
Like the sense that our work, right?
57:52
We might think it is of like great importance now, but it will be forgotten
57:55
and like lay laid bare by the sands of time in a short amount of time.
58:00
So having humility on so many different dimensions I think is vital.
58:06
But also it's a really good counterbalance towards so many of the forces
58:10
that we see in society right now.
58:12
But I also think it needs to be balanced with curiosity and like this
58:16
kind of like omnivorous curiosity.
58:17
Because as the world becomes more and more complex and we build giant
58:22
technologies that draw from many different areas or discoveries or
58:25
advances are connecting different things, being curious about everything
58:28
and then seeing ways of connecting these ideas together feels going back to
58:33
our discussion of our humanity, a very deeply human thing that we need to do.
58:37
And it also just, it provides a sense of context, a sense of wonder, and so
58:41
that like constant drive these, these constant drives of like keeping humility
58:46
in every aspect of what we're doing, of how we think about what we know,
58:49
what we understand, what we can build, what will have meaning, what we'll
58:52
have lasting significance alongside just this constant omniverse curiosity.
58:57
Those are the things that drive me and and yeah, hopefully I was
58:59
able to convey a little bit of those in the books that I wrote.
59:03
The common thread throughout, and it was an absolute pleasure.
59:06
Author of the Half-Life of Facts Overcomplicated on the shelves
59:10
there Behind me and his latest book, the Magic of Code Samuel Arbesman.
59:14
Thank you.
59:15
for joining us.
59:16
Thank you.
59:16
This was amazing.
59:17
And thank you as always to our sponsor, Kyndryl, with a unique blend
59:21
of AI powered consulting, built on unmatched managed service capability.
59:24
Kyndryl helps leaders harness the power of technology for smarter
59:28
decisions, faster innovation, and a lasting competitive edge.
59:30
You can find Kyndryl at Kyndryl.com.