WEBVTT - Ep112 "How is computer code like magic?" (with Sam Arbesman) 0:00:05.160 --> 0:00:08.960 What is software code and can it be thought of 0:00:09.119 --> 0:00:12.000 like a magic spell? Are we in the process of 0:00:12.039 --> 0:00:15.480 building a world so complex that we will lose the 0:00:15.520 --> 0:00:18.320 ability to understand it? Or has that already happened a 0:00:18.360 --> 0:00:20.319 long time ago? And what does any of this have 0:00:20.400 --> 0:00:24.279 to do with SimCity or knowledge that already exists but 0:00:24.400 --> 0:00:27.880 no one is thought to put together, or inventions that 0:00:28.000 --> 0:00:32.599 evolve beyond the grasp of their creators. And what coding 0:00:32.680 --> 0:00:38.840 looks like in the near future. Welcome to Inner Cosmos 0:00:38.840 --> 0:00:41.560 with Me and David Eagleman. I'm a neuroscientist and author 0:00:41.600 --> 0:00:44.680 at Stanford and in these episodes we look at the 0:00:44.720 --> 0:00:48.839 world inside us and around us to understand why and 0:00:48.920 --> 0:01:10.600 how our lives look the way they do. Today's episode 0:01:10.640 --> 0:01:15.520 is about computer code. We live in a world increasingly 0:01:15.560 --> 0:01:19.080 built out of symbols. We've got strings of code and 0:01:19.120 --> 0:01:24.000 lines of logic and invisible layers of computation stacked very 0:01:24.120 --> 0:01:26.959 deeply in our lives. This goes from traffic lights to 0:01:27.080 --> 0:01:32.520 financial markets, to weather predictions, to streaming recommendations, to all 0:01:32.560 --> 0:01:34.679 of our apps and our AI and on and on. 0:01:35.080 --> 0:01:40.240 We are surrounded by systems that hum quietly in the background, 0:01:40.520 --> 0:01:44.760 and that orchestrate everything about our modern lives. But it's 0:01:44.880 --> 0:01:50.160 very uncommonly that we stop to ask, what is code? Really? 0:01:50.520 --> 0:01:54.200 It's a tool, but it's also a massive force that 0:01:54.280 --> 0:01:57.800 has taken over the world. So where does it come from? 0:01:58.160 --> 0:02:00.920 Where is it taking us? And what does it mean 0:02:01.240 --> 0:02:05.800 to live inside systems that we ourselves have written but 0:02:05.880 --> 0:02:10.720 we can't possibly fully understand. So today's guest invites us 0:02:10.720 --> 0:02:15.079 to see code as having an aspect of magic. We're 0:02:15.080 --> 0:02:18.720 going to talk with Samuel Arbisman. He's a complexity scientist 0:02:18.760 --> 0:02:23.400 and a writer who thinks about the evolving relationship between 0:02:23.800 --> 0:02:26.840 humans and the tools that we build, and how our 0:02:26.880 --> 0:02:30.640 creations can outpace our comprehension. He's just written a new 0:02:30.680 --> 0:02:34.119 book called The Magic of Code, and here he takes 0:02:34.160 --> 0:02:38.359 a dive into the joy and the deeper nature of programming, 0:02:38.639 --> 0:02:42.520 in other words, beyond just an engineering discipline, but instead 0:02:42.800 --> 0:02:47.160 as a hopeful and enchanted practice. I think I would 0:02:47.200 --> 0:02:51.400 describe this book as a blurring of the boundaries between 0:02:51.560 --> 0:02:56.800 science and art, between logic and myth, between our intentions 0:02:56.840 --> 0:03:00.600 with writing code and what can actually emerge. Because the 0:03:00.720 --> 0:03:05.560 paradox is that code is built from rigid languages with 0:03:05.720 --> 0:03:09.800 strict syntax and rules, but it lets us create whole 0:03:09.960 --> 0:03:16.280 worlds from scratch. We can simulate galaxies and evolve virtual creatures, 0:03:16.320 --> 0:03:20.919 and test new economies and reimagine cities, and it gives 0:03:21.040 --> 0:03:24.560 rise to things that are complex and unpredictable and often 0:03:24.880 --> 0:03:34.040 quite beautiful. So here's my interview with Sam Arbusman. So, Sam, 0:03:34.080 --> 0:03:37.800 you're a scientist with very broad interests. What drew you 0:03:38.400 --> 0:03:41.840 to writing about your latest book on the subject of code? 0:03:42.400 --> 0:03:45.720 Yeah, So, one of the things I think about when 0:03:45.720 --> 0:03:48.360 I think about how people are talking about technology and 0:03:48.400 --> 0:03:51.880 computing and code is that right now, it feels like 0:03:51.880 --> 0:03:53.880 there's almost this kind of like a broken conversation in 0:03:53.960 --> 0:03:57.520 society where when we talk about code or computing the 0:03:57.520 --> 0:04:01.800 world of tech, there is this average real stance towards it, 0:04:02.040 --> 0:04:04.560 or or just worried about it. Sometimes people are just 0:04:04.880 --> 0:04:07.000 ignorant about it and unwilling to kind of learn more. 0:04:07.400 --> 0:04:10.280 And certainly some of the adversarial stuff is reasonable. But 0:04:11.160 --> 0:04:13.640 for me, like when I think about my own experience 0:04:13.680 --> 0:04:17.040 with computers growing up, it wasn't adversarial. It was kind 0:04:17.040 --> 0:04:19.360 of full of this kind of like wonder and delight. 0:04:19.800 --> 0:04:22.000 It also didn't feel that like computing was really just 0:04:22.000 --> 0:04:26.360 this branch of engineering. It also really connected to lots 0:04:26.400 --> 0:04:28.560 of different things. It was almost like humanistic liberal arts 0:04:28.760 --> 0:04:32.240 that drew in language and philosophy and biology and art 0:04:32.360 --> 0:04:34.839 and how we think in all these different areas. And 0:04:34.920 --> 0:04:38.560 so for me, I wanted to try to explain how 0:04:38.600 --> 0:04:41.640 to think about code and computing as this almost like 0:04:41.920 --> 0:04:44.039 liberal art that attracts all these different topics. 0:04:44.080 --> 0:04:47.240 And in the book, you compare this to philology, Right, 0:04:47.360 --> 0:04:50.000 what's this you tell us about philology? 0:04:50.120 --> 0:04:54.160 Yeah, so, so philology it's this branch of humanistic study 0:04:54.480 --> 0:04:59.600 that within the humanities, it was devoted to understanding the 0:04:59.600 --> 0:05:02.279 origin of words and kind of the nature of the 0:05:02.360 --> 0:05:06.040 history of language. But philology, in the process of doing philology, 0:05:06.480 --> 0:05:11.080 it required knowing about archaeology and anthropology and history and 0:05:11.120 --> 0:05:15.960 all these different topics. And then eventually philology sort of fractured, 0:05:16.000 --> 0:05:17.240 and that's kind of where we got a lot of 0:05:17.240 --> 0:05:20.200 the different domains within the humanities. And I kind of 0:05:20.279 --> 0:05:23.599 think that in some ways computing has at least some 0:05:23.960 --> 0:05:28.040 aspect of that kind of philology as unifier of lots 0:05:28.040 --> 0:05:31.240 of different topics. And so for me, my goal was 0:05:31.279 --> 0:05:35.120 to kind of try to show how computing and code 0:05:35.400 --> 0:05:39.920 can actually have that connective tissue between all these different domains. 0:05:40.120 --> 0:05:43.360 And so you described code as being magical. 0:05:43.680 --> 0:05:46.320 Why and so when I say magical, I'm not saying 0:05:47.160 --> 0:05:48.760 that it's like magic in the sense of like, oh, 0:05:48.800 --> 0:05:51.239 like this piece of software just works, it works like magic, 0:05:51.520 --> 0:05:53.640 although there is some of that. For me, it's actually 0:05:53.680 --> 0:05:56.640 this idea that when we think about the nature of 0:05:56.720 --> 0:06:01.159 code or magic, and we've had as a society this 0:06:01.279 --> 0:06:04.000 desire for millennia to kind of coerce the world around 0:06:04.080 --> 0:06:06.320 us through our language and our texts and our speech 0:06:06.600 --> 0:06:09.240 and make the world kind of do our bidding. And 0:06:09.279 --> 0:06:11.679 only in the past, I don't know, seventy five odd 0:06:11.760 --> 0:06:14.680 years since the advent of the modern digital computer, has 0:06:14.720 --> 0:06:16.680 this been a reality where we can actually write text, 0:06:16.720 --> 0:06:18.880 we can write code, and it can actually do things 0:06:18.880 --> 0:06:21.839 in the world. And so for me, there is this 0:06:22.000 --> 0:06:26.599 deep similarity between how we've thought about magic in the 0:06:26.600 --> 0:06:28.599 ancient or medieval days, or even in the stories that 0:06:28.640 --> 0:06:31.839 we tell ourselves and the reality of code. And so, 0:06:32.800 --> 0:06:35.760 of course this analogy and metaphor can only be taken 0:06:35.800 --> 0:06:38.360 so far before it kind of breaks down. I certainly 0:06:38.400 --> 0:06:41.200 take it to the bending point, if not the breaking point. 0:06:41.400 --> 0:06:44.240 But there are a lot of deep similarities between how 0:06:44.240 --> 0:06:47.720 to think about this. So, for example, magic often requires 0:06:47.760 --> 0:06:50.320 in our stories a certain amount of training and knowledge. 0:06:50.400 --> 0:06:52.600 And it's a craft. It's not just like a thing 0:06:52.640 --> 0:06:55.039 like that just works. It actually requires you to learn 0:06:55.080 --> 0:06:58.039 certain things. And so we have Hogwarts School of Witchcraft 0:06:58.120 --> 0:07:00.000 and Wizardry. You got to go there for seven years 0:07:00.080 --> 0:07:02.280 or whatever it is. And so too with code, like 0:07:02.320 --> 0:07:05.320 it doesn't necessarily just work. You actually have to understand 0:07:05.600 --> 0:07:07.919 how like the nature of syntax and the details of 0:07:07.920 --> 0:07:10.560 code and so and so that example, as well as 0:07:10.560 --> 0:07:12.720 other ones, I kind of us to show the ways 0:07:12.840 --> 0:07:15.280 in which this this idea of like the kind of 0:07:15.280 --> 0:07:17.920 the analogy of magic actually can be a productive and 0:07:18.000 --> 0:07:20.400 useful one to help us better understand how code works. 0:07:20.600 --> 0:07:22.760 So it's not just a set of instructions. It's like 0:07:22.920 --> 0:07:25.160 a spell in the sense that you puts a strange 0:07:25.160 --> 0:07:27.560 set of symbols and it does stuff in the world 0:07:27.640 --> 0:07:31.240 that moves electrons or launches rockets, or models pandemics or 0:07:31.280 --> 0:07:35.600 creates simulations. And that's the sense in which it's got 0:07:35.600 --> 0:07:39.480 this magic to it. But also you point to the 0:07:39.520 --> 0:07:44.400 fact that there's often unpredictability and emergence of things we 0:07:44.440 --> 0:07:47.280 didn't expect from code. So can you give us an 0:07:47.280 --> 0:07:50.240 example of this dual nature of code? 0:07:50.880 --> 0:07:53.000 Yeah, I mean so certainly. In the world of magic, 0:07:53.040 --> 0:07:55.240 we have a lot of these stories, and like there's 0:07:55.240 --> 0:07:56.840 like the story of like if the source is Apprentice, 0:07:56.840 --> 0:07:58.560 which I think there's like the old version and then 0:07:58.560 --> 0:08:02.160 the Disney makeme Out version, where some sort of magic 0:08:02.200 --> 0:08:05.960 has unanticipated consequences ands only you have you have brooms 0:08:06.040 --> 0:08:08.880 kind of walking around and flooding and flooding a basement, 0:08:09.040 --> 0:08:11.480 And the same kind of thing is true with code. 0:08:11.520 --> 0:08:14.120 Where in code, I think people who might not be 0:08:14.200 --> 0:08:16.480 familiar with with programming think of it as, oh, like 0:08:16.480 --> 0:08:18.000 I have this idea in my mind and I'm going 0:08:18.040 --> 0:08:20.200 to instantiate it into a computer program. And there is that, 0:08:20.480 --> 0:08:24.040 but there's also a huge amount of debugging and frustration 0:08:24.160 --> 0:08:26.840 because oftentimes when you write a program, there's a gap 0:08:26.880 --> 0:08:28.680 between how you think it will actually work and how 0:08:28.720 --> 0:08:32.440 it actually does work. And oftentimes the reality of the 0:08:32.440 --> 0:08:36.360 program and better understanding of it is only revealed through 0:08:36.480 --> 0:08:38.640 these bugs, through these glitches and edge cases and things 0:08:38.679 --> 0:08:42.559 like that, And so there are many situations where we 0:08:42.720 --> 0:08:45.760 only see these bizarre errors, and then based on those 0:08:45.880 --> 0:08:48.880 errors and these kind of unanticipated consequences, do we realize, oh, 0:08:48.920 --> 0:08:50.480 how this thing actually works, and so it can be 0:08:50.720 --> 0:08:53.600 So there's a there's a well known story of someone 0:08:53.640 --> 0:08:55.360 who I think was like a systems administrator for some 0:08:55.440 --> 0:08:58.520 university department. He was told by the chair of the 0:08:58.520 --> 0:09:01.640 department that their email was only able to be sent 0:09:01.679 --> 0:09:04.840 about five hundred miles away, and the systems like this 0:09:04.880 --> 0:09:06.920 is insane, like email, that's not how email works. And 0:09:06.960 --> 0:09:09.200 it turned out by delving into it, and he was 0:09:09.200 --> 0:09:11.480 able to find that I think that there was like 0:09:11.520 --> 0:09:14.920 an older piece of software that hadn't been upgraded, but 0:09:15.440 --> 0:09:18.160 the newer system didn't realize this and like would time out, 0:09:18.160 --> 0:09:20.160 but only after like some very small amount of time. 0:09:20.280 --> 0:09:22.440 And it turns out, based on like the speed of 0:09:22.480 --> 0:09:24.560 sound and that small amount of time, it ended up 0:09:24.559 --> 0:09:26.360 working out to about five hundred miles and it was 0:09:26.400 --> 0:09:29.120 this weird unanticipated consequence. But of course, and there's also 0:09:29.200 --> 0:09:32.079 other things that just the fact that when you stitch 0:09:32.480 --> 0:09:35.439 systems together and pieces of software together, they all interact 0:09:35.440 --> 0:09:38.120 in unexpected ways. And that's also the kind of unanticipated 0:09:38.120 --> 0:09:41.760 consequences we see. Whether it's like some weird little system 0:09:41.760 --> 0:09:44.880 fails to get upgraded and then suddenly all like the 0:09:44.880 --> 0:09:47.760 the airline systems go down for a first certain amount 0:09:47.760 --> 0:09:49.120 of time or whatever it is. And so there is 0:09:49.160 --> 0:09:52.319 that kind of unanticipated consequence in lots of different ways. 0:09:52.320 --> 0:09:54.679 And we're seeing this, of course even more so with AI. 0:09:56.440 --> 0:09:57.960 Okay, So this is the thing that you and I 0:09:58.040 --> 0:10:01.080 both love is the emergence of comp complexity in the world. 0:10:01.320 --> 0:10:04.880 And with code it's a very specific, detailed set of instructions, 0:10:04.920 --> 0:10:08.560 and yet it can break, it can decay, it can 0:10:08.600 --> 0:10:12.160 be opaque. All kinds of things can happen. And sometimes 0:10:12.200 --> 0:10:16.000 when we try to model complexity, we actually unleash complexity. 0:10:16.360 --> 0:10:19.440 So tell us your take on how code can do 0:10:19.520 --> 0:10:21.000 things that we didn't expect it to do. 0:10:21.400 --> 0:10:23.760 Yeah, and so when we think about and just engineered 0:10:23.800 --> 0:10:26.000 systems more broadly, and certainly when it comes to code, 0:10:26.360 --> 0:10:30.000 you think, oh, it's designed by people. It sounds very logical, 0:10:30.040 --> 0:10:32.440 it's kind of derived from mathematics. It should be very 0:10:32.480 --> 0:10:36.440 simple and straightforward, and very small bits of code are that. 0:10:36.920 --> 0:10:39.800 But the truth is it adds up and then through 0:10:39.840 --> 0:10:44.520 this combination of the sizes of programs, growing them beingcoming 0:10:44.600 --> 0:10:47.640 connected to various other bits of code that are out there, 0:10:47.880 --> 0:10:49.640 as well as also just engaging with kind of the 0:10:50.120 --> 0:10:52.720 messiness of the world around us, you end up getting 0:10:52.840 --> 0:10:54.920 sort of a certain amount of unexpectedness as well as 0:10:55.000 --> 0:10:57.840 kind of just a reduced understanding. And part of this 0:10:57.960 --> 0:11:00.200 is because there's and you mentioned this kind of like 0:11:00.440 --> 0:11:02.839 it's like breaking and things kind of growing over time. 0:11:02.960 --> 0:11:06.440 There is this whole phenomenon of legacy code where code 0:11:06.640 --> 0:11:08.760 has been around for a very very long time, and 0:11:09.360 --> 0:11:11.400 we have systems that are still being you like, that 0:11:11.440 --> 0:11:14.200 are still being used that were developed decades ago, but 0:11:14.240 --> 0:11:16.200 they might be involved in kind of the irs, but 0:11:16.200 --> 0:11:18.280 they were developed first developed or in the Kennedy administration. 0:11:18.320 --> 0:11:22.240 Like there's all these kind of crazy examples where things 0:11:22.400 --> 0:11:25.440 that have been developed the people who first made them 0:11:25.520 --> 0:11:27.600 they might be and they might be long retired, they 0:11:27.640 --> 0:11:30.520 might be dead. And we also just don't fully understand 0:11:30.520 --> 0:11:34.040 these systems. And so it's this weird situation where we 0:11:34.160 --> 0:11:36.680 have to recognize that even the systems of our own 0:11:36.840 --> 0:11:41.080 construction are actually like when they become big enough, they 0:11:41.080 --> 0:11:43.480 actually have this kind of qualitative difference where they almost 0:11:43.640 --> 0:11:47.480 they almost become almost biological or organic in their complexity, 0:11:47.559 --> 0:11:50.320 and as a result, we have a reduced understanding and 0:11:50.320 --> 0:11:53.240 we have to kind of take almost biological modes of 0:11:53.320 --> 0:11:56.080 studying these systems, whether it's kind of like the days 0:11:56.080 --> 0:11:57.680 of old with like the natural is kind of going 0:11:57.679 --> 0:12:00.480 out and collecting bugs in this case could be bugs 0:12:00.480 --> 0:12:03.040 and errors. It could be bugs like insects, as well 0:12:03.040 --> 0:12:05.000 as just kind of like trying to tinker at the 0:12:05.040 --> 0:12:08.320 edges and better understand a system, because the thing overall 0:12:08.400 --> 0:12:11.640 you fully don't understand. So there's this weird situation where 0:12:12.000 --> 0:12:16.880 these systems are engineered, but they also involve kind of 0:12:17.440 --> 0:12:20.120 a certain amount of humility in trying to understand these systems. 0:12:20.120 --> 0:12:22.000 And part of that is also because one of the 0:12:22.000 --> 0:12:25.240 other features of computing and software is this idea of 0:12:25.240 --> 0:12:27.800 abstraction that you can kind of build things on top 0:12:27.800 --> 0:12:31.800 of other pieces, and those pieces are then sophisticated kind 0:12:31.800 --> 0:12:33.640 of units, and you can kind of use them as 0:12:33.720 --> 0:12:35.800 standalone bits and then don't have to worry about the 0:12:35.840 --> 0:12:39.280 things underneath it. And so that modularity is very, very powerful, 0:12:39.320 --> 0:12:42.520 but as a result, there is a decreased amount of understanding, 0:12:42.559 --> 0:12:45.480 and so sometimes not understanding what's going on under the 0:12:45.480 --> 0:12:48.359 hood or kind of underneath these pieces, even when yourself 0:12:48.600 --> 0:12:51.040 are programming them or programming kind of the things that 0:12:51.080 --> 0:12:53.440 interact with them, means that you can kind of also 0:12:53.520 --> 0:13:10.760 have a certain amount of anticipated consequences. 0:13:12.840 --> 0:13:15.040 And so what are the key limitations that you see 0:13:15.040 --> 0:13:18.239 when we try to model very complex systems? 0:13:18.640 --> 0:13:19.880 So one of the things I think about when I 0:13:19.880 --> 0:13:23.720 think about modeling complex systems is is what is the 0:13:23.720 --> 0:13:27.360 goal of modeling the complex system? So there's some situations 0:13:27.400 --> 0:13:31.000 where we really want to have perfect fidelity to a 0:13:31.040 --> 0:13:33.720 real world system. And so for predictions, like with weather, 0:13:34.160 --> 0:13:35.880 as a weather prediction, like you want to and you 0:13:35.920 --> 0:13:38.760 want to not just understand kind of how kind of 0:13:38.840 --> 0:13:41.800 air moves around, You want to really understand whether or 0:13:41.840 --> 0:13:43.400 not it's going to rain tomorrow or in an hour 0:13:43.480 --> 0:13:46.120 or two hours or whatever it is. And and based 0:13:46.120 --> 0:13:48.360 on that you have to have a great deal of 0:13:48.440 --> 0:13:51.280 data and a great deal of complexity. And then oftentimes 0:13:51.480 --> 0:13:55.360 the resulting models might be very powerful, very sophisticated, but 0:13:55.400 --> 0:13:57.600 there might be a reduced amount of understanding and actually 0:13:57.720 --> 0:13:59.800 how these things are doing what they're doing. On the 0:13:59.800 --> 0:14:03.360 other hand, if you want to just understand the features 0:14:03.400 --> 0:14:06.400 of a system, you can sometimes get away with a 0:14:06.520 --> 0:14:09.880 much simpler model, which might not necessarily be exactly the 0:14:09.920 --> 0:14:12.320 way that the model works, but could at least capture 0:14:12.360 --> 0:14:14.040 some of the complexity and kind of the and the 0:14:14.080 --> 0:14:16.520 emergence of what you were talking about of that system. 0:14:16.559 --> 0:14:18.960 And so, for example, and this is a kind of 0:14:18.960 --> 0:14:22.080 trivial example, but the computer game SimCity. It is not 0:14:22.720 --> 0:14:26.440 modeling an actual city, but to give you an intuitive 0:14:26.520 --> 0:14:32.000 sense of how feedback operates, or unanticipated consequences work, or 0:14:32.080 --> 0:14:34.480 just the fact that, like complex systems can bite back 0:14:34.600 --> 0:14:37.800 and do weird things that you might not expect. SimCity 0:14:37.880 --> 0:14:40.080 is great for that kind of thing. And so, and 0:14:40.120 --> 0:14:41.640 it can also kind of give you a sense of, oh, 0:14:41.640 --> 0:14:43.240 when I do this kind of thing, according to this 0:14:43.320 --> 0:14:46.120 model of how Will Wright or whoever was programming it 0:14:46.280 --> 0:14:48.520 thought cities would work, this maybe would be the way 0:14:48.520 --> 0:14:50.040 it works. Whether or not that is actually how the 0:14:50.040 --> 0:14:52.680 city operates, that's entirely different, different thing. But so for me, 0:14:53.080 --> 0:14:55.680 I often think about, like, yeah, what is the ultimate 0:14:55.720 --> 0:14:57.520 goal with the model? Is the goal to kind of 0:14:57.600 --> 0:15:00.960 understand things? And we have to recognizing our human minds 0:15:00.960 --> 0:15:04.400 are really limited when it comes to understanding complex and 0:15:04.440 --> 0:15:07.680 nonlinear systems, and so we need these simplified models. If 0:15:07.720 --> 0:15:11.320 it is actually just prediction, then sometimes a really complex 0:15:11.360 --> 0:15:15.080 model can work, but at the cost of reduced understanding. 0:15:15.440 --> 0:15:19.640 So as AI generated code becomes more common, what does 0:15:19.640 --> 0:15:24.120 that do to our sense of authorship and even understanding? 0:15:24.120 --> 0:15:26.160 And could we end up in a situation where we're 0:15:26.160 --> 0:15:29.320 surrounded by systems that are running our world that we 0:15:29.360 --> 0:15:30.520 don't understand at all. 0:15:30.640 --> 0:15:33.040 I mean, to be honest, I think we're probably there already. 0:15:33.480 --> 0:15:35.800 It's just the situation where I think many people are 0:15:35.800 --> 0:15:38.000 not aware of that fact. This is also one of 0:15:38.040 --> 0:15:40.640 these situations where as we build more and more complex 0:15:40.680 --> 0:15:44.040 systems and everyday users are kind of more distant from them, 0:15:44.080 --> 0:15:46.880 we just don't realize the sheer complexity. When the Apple 0:15:46.920 --> 0:15:50.080 Watch first came out, this is years ago, there was 0:15:50.120 --> 0:15:52.600 an article in the Wall Street Journal, I think it 0:15:52.640 --> 0:15:54.680 was like the Style section about like, are people going 0:15:54.720 --> 0:15:56.960 to still use like biomechanical watch? As the answer is 0:15:57.040 --> 0:15:59.200 they still are that. They interviewed this one guy about it, 0:15:59.240 --> 0:16:01.360 like whether or not you want to buy mechanical watch 0:16:01.400 --> 0:16:03.720 or just smart watches, and this guy said something to 0:16:03.720 --> 0:16:06.400 the effect of, of course I want a mechanical watch. 0:16:06.400 --> 0:16:08.160 When I think about a mechanical watch. It's so complex 0:16:08.160 --> 0:16:10.360 as opposed to a smart watch, which is just a chip, 0:16:10.520 --> 0:16:12.440 and the thing is like a chip is orders of 0:16:12.520 --> 0:16:15.480 magnitude more complex than a mechanical watch. But we've been 0:16:15.480 --> 0:16:18.560 shielded from it, and I think as we have AI 0:16:18.640 --> 0:16:21.880 generated code, we're going to kind of have another level 0:16:21.920 --> 0:16:26.160 of shielding. I do think we need better mechanisms for 0:16:26.680 --> 0:16:28.720 so interrogating the system. So I actually so I think 0:16:28.760 --> 0:16:30.280 one of these situations we're On the one hand, it 0:16:30.360 --> 0:16:32.560 is very good that we can now generate code via 0:16:32.640 --> 0:16:36.760 AI and build simple tools and actually democratize the software development. 0:16:36.760 --> 0:16:39.200 I think there's lots of interesting things there. But I 0:16:39.560 --> 0:16:42.120 still also think understanding code to a certain degree and 0:16:42.160 --> 0:16:44.360 allowing you to kind of like dive into the code 0:16:44.360 --> 0:16:46.920 that is being generated and tweak it, not only does 0:16:46.920 --> 0:16:48.760 it give you a better understanding of what you're doing, 0:16:48.880 --> 0:16:50.920 but it's still actually really good to help make sure 0:16:50.960 --> 0:16:53.720 that it is doing at least partly what you hope for. 0:16:54.440 --> 0:16:57.200 That being said, our systems have always been imperfect, and 0:16:57.240 --> 0:16:59.000 I think right now, this is this moment is kind 0:16:59.000 --> 0:17:02.120 of just heightening that fact, and maybe we'll give people 0:17:02.200 --> 0:17:07.800 a better awareness that these systems have always been enormously complex, 0:17:08.520 --> 0:17:11.440 enormously imperfect, made by humans at least at some level, 0:17:12.560 --> 0:17:15.760 but maybe give us a greater appreciation for building systems 0:17:15.800 --> 0:17:18.680 on top of these, maybe also AI generated as well, 0:17:18.960 --> 0:17:23.040 that can allow us to make sure that the the 0:17:23.080 --> 0:17:25.560 unanticipated consequences are as minimal as possible. 0:17:27.359 --> 0:17:28.720 You know, I was just thinking about right after I 0:17:28.720 --> 0:17:31.320 asked my question about could we end up living inside 0:17:31.359 --> 0:17:34.520 systems that are too complex for us to understand. Obviously, 0:17:34.600 --> 0:17:37.480 we live inside our biology, and that is for sure 0:17:37.680 --> 0:17:41.120 that we don't understand. But a fraction of what's going 0:17:41.160 --> 0:17:43.480 on inside is biologically. But we try to eat the 0:17:43.560 --> 0:17:46.600 right foods and get exercise and just just right on 0:17:46.680 --> 0:17:49.320 top of this system. So we're actually quite used to 0:17:49.359 --> 0:17:52.920 living inside systems that are beyond our understanding. 0:17:53.200 --> 0:17:55.000 Yeah, And I think and that goes back to kind 0:17:55.000 --> 0:17:59.240 of like the biological nature of these massive, complex computational systems. Like, 0:17:59.280 --> 0:18:02.880 the more we recognize that these systems are really complex 0:18:02.920 --> 0:18:06.480 and almost have like this organic quality, we will have 0:18:06.520 --> 0:18:08.520 to realize, yeah, that we need different ways of approaching 0:18:08.560 --> 0:18:11.040 them and the way we approach our bodies. Right, It's not, 0:18:11.240 --> 0:18:13.720 oh like I have total ignorance about how the system 0:18:13.720 --> 0:18:16.440 works or have complete understanding. There's there's a lot in between, 0:18:16.840 --> 0:18:20.480 and using rules of thumbs and things are are very 0:18:20.520 --> 0:18:22.359 powerful and I but at the same time, though, we 0:18:22.359 --> 0:18:24.439 don't want to necessarily kind of like succumb to like 0:18:24.480 --> 0:18:26.600 the like the biohacking trend, which I feel like five 0:18:26.640 --> 0:18:29.360 ten years ago was a really big thing where it's like, oh, 0:18:29.440 --> 0:18:31.280 if I can just find this one chemical to ingest 0:18:31.320 --> 0:18:34.440 or this one cool trick to do, then I'll never 0:18:34.480 --> 0:18:37.240 need to sleep again or I'll be held and we 0:18:37.280 --> 0:18:40.399 realize that, I mean, yeah, our system, our bodies have 0:18:40.480 --> 0:18:44.840 evolved over millions of years and are optimizing a huge 0:18:44.920 --> 0:18:46.680 number of different things, and they're going to be imperfect 0:18:46.720 --> 0:18:49.000 and weird, and so it's going to be maybe we 0:18:49.040 --> 0:18:51.160 can find those things, but the odds that we are 0:18:51.400 --> 0:18:53.320 going to is very low. And I think the same 0:18:53.400 --> 0:18:56.200 kind of thing that that same kind of approach needs 0:18:56.240 --> 0:18:59.240 to be used when we think about these complex technologies 0:18:59.280 --> 0:19:01.000 that were building around us as well. 0:19:01.160 --> 0:19:03.200 So let me ask you this, if we fast forward 0:19:03.280 --> 0:19:07.280 one hundred years, are we still coding by using symbols 0:19:07.280 --> 0:19:09.919 and syntax or is it a completely different sort of 0:19:10.040 --> 0:19:15.320 thing where we are setting initial conditions and letting complexity evolve. 0:19:15.320 --> 0:19:17.480 I would say it's very different. I don't think we 0:19:17.520 --> 0:19:19.639 need to necessarily go even one hundred years into the future. 0:19:19.680 --> 0:19:22.159 It could be like five ten years until we're kind 0:19:22.160 --> 0:19:25.000 of managing these AI generated systems. I mean. But the 0:19:25.040 --> 0:19:28.840 truth is, when I think about what coding is, it's 0:19:28.880 --> 0:19:31.640 always been this moving target, like it's always been changing. 0:19:31.960 --> 0:19:34.439 So the way I learned how to program in some 0:19:34.480 --> 0:19:37.760 of the languages I learned, they're not totally extinct, but 0:19:37.760 --> 0:19:40.760 they're not things I would ever consider using nowadays. But 0:19:40.800 --> 0:19:43.760 also even when I think about how people before me 0:19:43.840 --> 0:19:46.959 learned how to program, that too was something very very different. 0:19:47.040 --> 0:19:50.199 It was like plugging in cables or flipping switches or 0:19:50.359 --> 0:19:53.760 writing things in binary or assembly code. Like I never 0:19:53.800 --> 0:19:55.960 did those kinds of things. Nor do I really have 0:19:56.040 --> 0:19:57.840 a strong desire too. And that's okay, And I think 0:19:57.840 --> 0:19:59.600 it's going to continue changing. And so one of the 0:19:59.600 --> 0:20:02.280 ways I think about this is actually I tell this 0:20:02.359 --> 0:20:05.800 story in the book, a story from the Tumud actually 0:20:05.920 --> 0:20:11.000 where it's this conversation that's describing a conversation between God 0:20:11.040 --> 0:20:14.240 and Moses and they're discussing like, oh, like who's going 0:20:14.280 --> 0:20:16.919 to be the greatest scholar in the future, and God says, oh, 0:20:16.920 --> 0:20:18.960 it's going to be some rabbi, like a thousand and 0:20:18.960 --> 0:20:20.879 two thousand years in the future, and Moses says, can 0:20:20.920 --> 0:20:22.600 you show him to me. So there's this weird time 0:20:22.640 --> 0:20:24.720 travel moment where he's transported to the hall of Study, 0:20:24.720 --> 0:20:26.280 like a thousand years in the future whatever it is, 0:20:26.600 --> 0:20:28.560 and Moses is sitting in the back listening to this 0:20:28.600 --> 0:20:32.680 illustrious scholar talking about things, and he realizes he doesn't 0:20:32.760 --> 0:20:35.560 understand anything this guy's talking about, and he's kind of overwhelmed. 0:20:35.560 --> 0:20:37.920 He's like, Oh, I'm the one who received the law 0:20:38.400 --> 0:20:40.480 from Heaven and I don't get it until at the 0:20:40.560 --> 0:20:42.640 very last moment, the rabbi says, oh, in the way 0:20:42.680 --> 0:20:46.320 we understand all this is because of the law received 0:20:46.359 --> 0:20:48.479 from Moses at Sinai, And at that point he's calmed 0:20:48.600 --> 0:20:51.239 and because he realizes that even if he doesn't understand it, 0:20:51.480 --> 0:20:55.080 there's this clear, continuous line and kind of continuous tradition. 0:20:55.119 --> 0:20:57.359 And I feel like when it comes to code, the 0:20:57.400 --> 0:20:59.520 same kind of thing is true. Coding has changed, it 0:20:59.520 --> 0:21:02.000 will continue to change. It'll be much more like managing 0:21:02.040 --> 0:21:05.080