WEBVTT 00:00:00.000 --> 00:00:02.359 I was looking back at the calendar for October 00:00:02.359 --> 00:00:05.339 2011 recently, and it really struck me. It was 00:00:05.339 --> 00:00:09.580 this bizarre, pivotal moment in tech history. 00:00:09.839 --> 00:00:11.679 Oh, absolutely. It was one of those weeks where 00:00:11.679 --> 00:00:14.820 the world really shifted, but half of the shift 00:00:14.820 --> 00:00:17.839 was screamingly loud and the other half was almost 00:00:17.839 --> 00:00:20.359 completely silent. It's a perfect study in contrast, 00:00:20.579 --> 00:00:21.820 isn't it? Yeah. You're talking about the week 00:00:21.820 --> 00:00:24.359 that digital age lost two of its founding fathers, 00:00:24.480 --> 00:00:27.809 really. Exactly. But the way we handled it, the 00:00:27.809 --> 00:00:30.870 way the world reacted was just so different. 00:00:31.050 --> 00:00:33.649 On one side, you had Steve Jobs. When he died 00:00:33.649 --> 00:00:36.670 on October 5th, it was a global event. A phenomenon. 00:00:36.890 --> 00:00:38.850 A total phenomenon. I remember walking past the 00:00:38.850 --> 00:00:41.350 Apple store downtown and the windows were just, 00:00:41.390 --> 00:00:43.270 you know, plastered with sticky notes. People 00:00:43.270 --> 00:00:45.030 were crying in the streets. It was front page 00:00:45.030 --> 00:00:48.009 news in every single language. It felt like a 00:00:48.009 --> 00:00:50.969 head of state had died. It really did. And in 00:00:50.969 --> 00:00:54.560 a way, it made sense. Jobs was the face. Of the 00:00:54.560 --> 00:00:56.719 technology we touch, the technology we hold, 00:00:56.920 --> 00:01:00.079 he was the showman. The ultimate showman. He'd 00:01:00.079 --> 00:01:01.719 hold the device up in his hand and tell us it 00:01:01.719 --> 00:01:04.359 was magic. And we believed him. We felt this 00:01:04.359 --> 00:01:06.980 personal connection to the guy in the black turtleneck 00:01:06.980 --> 00:01:09.140 because he was the one selling us the future. 00:01:09.400 --> 00:01:12.359 Then, just seven days later, literally a week 00:01:12.359 --> 00:01:15.620 around October 12th, another man died. He was 00:01:15.620 --> 00:01:18.760 70 years old, living alone in a town called Berkeley 00:01:18.760 --> 00:01:21.700 Heights, New Jersey. His name was Dennis Ritchie. 00:01:22.680 --> 00:01:24.920 crickets. Almost total silence. Unless you were 00:01:24.920 --> 00:01:27.500 a serious computer nerd, a programmer, you probably 00:01:27.500 --> 00:01:28.980 didn't even see the headline. It was buried. 00:01:29.239 --> 00:01:31.640 It was there at all. And that silence, I think, 00:01:31.640 --> 00:01:33.480 is actually quite profound. It speaks volumes 00:01:33.480 --> 00:01:36.579 about the nature of his work. While Steve Jobs 00:01:36.579 --> 00:01:39.159 was designing the, let's call them the fashion 00:01:39.159 --> 00:01:41.700 items of the tech world, the beautiful, sleek 00:01:41.700 --> 00:01:44.959 glass rectangles, Dennis Ritchie was doing something 00:01:44.959 --> 00:01:48.060 much, much deeper. He wasn't designing the case. 00:01:48.140 --> 00:01:50.719 He was designing the physics of the entire digital 00:01:50.719 --> 00:01:52.879 universe that lives inside that case. That's 00:01:52.879 --> 00:01:55.040 the core of what we want to get into today. This 00:01:55.040 --> 00:01:57.879 claim that this quiet guy who nobody stopped 00:01:57.879 --> 00:02:00.519 to mourn in the street actually built the, quote, 00:02:00.680 --> 00:02:03.500 numeric soul of the device in your pocket right 00:02:03.500 --> 00:02:06.560 now. It's not just a claim, you know, it's the 00:02:06.560 --> 00:02:09.430 reality of the situation. If you could take a 00:02:09.430 --> 00:02:12.250 powerful enough microscope and look inside the 00:02:12.250 --> 00:02:14.930 logic of your iPhone or your smart fridge or 00:02:14.930 --> 00:02:17.610 the massive servers running this very conversation. 00:02:17.949 --> 00:02:20.409 You wouldn't see Steve Jobs. You would not. You 00:02:20.409 --> 00:02:22.169 would see the work of Dennis Ritchie everywhere. 00:02:22.490 --> 00:02:25.129 His fingerprints are on every single chip. So 00:02:25.129 --> 00:02:28.169 if he's that fundamental, why is he the unsung 00:02:28.169 --> 00:02:31.030 hero? Why don't we have movies about him written 00:02:31.030 --> 00:02:33.770 by Aaron Sorkin? Why isn't he a household name? 00:02:33.909 --> 00:02:36.030 Because infrastructure is invisible when it works 00:02:36.030 --> 00:02:38.580 perfectly. You only notice the plumbing when 00:02:38.580 --> 00:02:41.520 a pipe bursts. Right. You only think about the 00:02:41.520 --> 00:02:43.879 power grid when the lights go out. Dennis Ritchie 00:02:43.879 --> 00:02:45.699 built the plumbing for the entire information 00:02:45.699 --> 00:02:48.139 age. So for this deep dive, we're going to unpack 00:02:48.139 --> 00:02:50.199 that. We're going to figure out how one man, 00:02:50.300 --> 00:02:53.300 working in a probably messy office at Bell Labs, 00:02:53.500 --> 00:02:56.460 created the two most important, most foundational 00:02:56.460 --> 00:02:59.199 tools in computing history. The Unix operating 00:02:59.199 --> 00:03:01.860 system and the C programming language. The twin 00:03:01.860 --> 00:03:04.639 pillars. Yeah, and I hear those names Unix and 00:03:04.639 --> 00:03:06.539 CNOT, like I vaguely know they're important. 00:03:06.800 --> 00:03:09.159 But I want to get past the Wikipedia summary 00:03:09.159 --> 00:03:12.280 today. I really want to understand why they matter. 00:03:12.580 --> 00:03:16.139 Because I've heard people, serious people, say 00:03:16.139 --> 00:03:18.400 that without Richie, the internet as we know 00:03:18.400 --> 00:03:21.379 it simply wouldn't exist. And that's not hyperbole. 00:03:21.460 --> 00:03:24.400 It's a statement of fact. Brian Kernaghan, who 00:03:24.400 --> 00:03:26.840 worked right alongside him, famously said that 00:03:26.840 --> 00:03:29.180 the tools Ritchie built and their direct descendants 00:03:29.180 --> 00:03:32.259 run pretty much everything today. Everything. 00:03:32.439 --> 00:03:34.860 Everything. So our mission here is to find out 00:03:34.860 --> 00:03:36.719 how that happened. And it really has to start 00:03:36.719 --> 00:03:39.139 with understanding who Dennis Ritchie actually 00:03:39.139 --> 00:03:41.819 was as a person. Right, because he wasn't a Silicon 00:03:41.819 --> 00:03:43.979 Valley dropout, was he? He didn't start some 00:03:43.979 --> 00:03:46.000 company in a garage with the goal of getting 00:03:46.000 --> 00:03:48.159 rich. Oh, not at all. He was from a completely 00:03:48.159 --> 00:03:50.620 different world, a different mindset. Dennis 00:03:50.620 --> 00:03:54.340 McAllister Ritchie was born in 1941 in Bronxville, 00:03:54.400 --> 00:03:56.780 New York. And he was basically born into the 00:03:56.780 --> 00:03:59.400 family business of science and engineering. His 00:03:59.400 --> 00:04:01.759 father was also in the field. His father, Alistair 00:04:01.759 --> 00:04:04.139 Ritchie, was a legend in his own right. A longtime 00:04:04.139 --> 00:04:07.139 scientist at Bell Labs. The same Bell Labs. The 00:04:07.139 --> 00:04:09.879 very same. Alistair co -authored a really important 00:04:09.879 --> 00:04:12.659 book called The Design of Switching Circuits. 00:04:13.289 --> 00:04:16.189 Now, that sounds incredibly dry. It does sound 00:04:16.189 --> 00:04:18.769 a little dry, yeah. But switching circuits are 00:04:18.769 --> 00:04:21.509 the fundamental logic of how machines make decisions. 00:04:21.730 --> 00:04:25.310 It's the physical manifestation of zero and one. 00:04:25.649 --> 00:04:29.430 On and off. True and false. Dennis literally 00:04:29.430 --> 00:04:33.089 grew up in Summit, New Jersey, with that kind 00:04:33.089 --> 00:04:35.490 of deep structural thinking just floating around 00:04:35.490 --> 00:04:38.160 the dinner table. So he was steeped in this idea 00:04:38.160 --> 00:04:40.220 of logical structures from the time he was a 00:04:40.220 --> 00:04:42.399 kid. It wasn't something he discovered in college. 00:04:42.660 --> 00:04:44.800 Completely. It was his native language. And of 00:04:44.800 --> 00:04:46.519 course, he was a brilliant student. He went to 00:04:46.519 --> 00:04:49.860 Harvard, graduated in 1963. But here's the key 00:04:49.860 --> 00:04:52.040 distinction. And this is so important for understanding 00:04:52.040 --> 00:04:55.579 his later work. He didn't study computer science. 00:04:55.759 --> 00:04:57.920 Because the major didn't exist yet. It barely 00:04:57.920 --> 00:05:00.470 existed as a concept. He studied physics and 00:05:00.470 --> 00:05:02.629 applied mathematics. And I think this is absolutely 00:05:02.629 --> 00:05:04.930 crucial for understanding how his mind worked. 00:05:05.089 --> 00:05:07.370 Well, computer science today can sometimes feel 00:05:07.370 --> 00:05:09.389 more like engineering. You're gluing different 00:05:09.389 --> 00:05:11.329 libraries and frameworks together to build a 00:05:11.329 --> 00:05:16.170 product. But physics and pure math. They're about 00:05:16.170 --> 00:05:18.490 discovering fundamental truths. They're about 00:05:18.490 --> 00:05:21.129 finding the cleanest, the most elegant, the simplest 00:05:21.129 --> 00:05:24.170 possible equation to describe a complex universe. 00:05:24.509 --> 00:05:26.730 And you're saying you can see that elegance in 00:05:26.730 --> 00:05:29.029 his code later on? You absolutely can. Yeah. 00:05:29.110 --> 00:05:31.290 There is a minimalism and a power to his work 00:05:31.290 --> 00:05:34.110 that is just beautiful. It's not messy. It's 00:05:34.110 --> 00:05:36.610 incredibly disciplined. He approached software 00:05:36.610 --> 00:05:38.370 not as a hacker trying to make something work 00:05:38.370 --> 00:05:40.870 for a demo, but as a mathematician trying to 00:05:40.870 --> 00:05:43.470 find the correct and most fundamental solution 00:05:43.470 --> 00:05:45.920 to a problem. OK, so speaking of his time at 00:05:45.920 --> 00:05:47.939 Harvard, there's this really weird story about 00:05:47.939 --> 00:05:49.980 his Ph .D. that I was reading about. The notes 00:05:49.980 --> 00:05:52.879 say he sort of lost his degree. How on earth 00:05:52.879 --> 00:05:55.019 do you lose a Ph .D.? This is one of those anecdotes 00:05:55.019 --> 00:05:57.420 that perfectly, perfectly encapsulates his personality. 00:05:57.860 --> 00:06:00.720 So in 1968, he's at Harvard. He writes his entire 00:06:00.720 --> 00:06:03.660 thesis. It's got this very academic title, Computational 00:06:03.660 --> 00:06:05.980 Complexity and Program Structure. Sounds impressive. 00:06:06.399 --> 00:06:09.399 It was. It was supervised by a big name. Patrick 00:06:09.399 --> 00:06:12.079 Fisher. He does all the research. He writes the 00:06:12.079 --> 00:06:14.259 draft. He submits it. He's done all the hard 00:06:14.259 --> 00:06:18.279 work. But he never officially received the degree 00:06:18.279 --> 00:06:21.660 at the time. So what happened? Did he fail his 00:06:21.660 --> 00:06:23.519 defense or something? No, no, nothing like that. 00:06:23.560 --> 00:06:25.939 He just he just didn't bother with the final 00:06:25.939 --> 00:06:28.740 administrative wrap up. The story goes that he 00:06:28.740 --> 00:06:30.339 didn't want to pay the binding fee to have a 00:06:30.339 --> 00:06:32.899 fancy copy put in the library, or he just didn't 00:06:32.899 --> 00:06:34.899 file the final piece of paperwork. He already 00:06:34.899 --> 00:06:37.439 started working at Bell Labs, and he was so completely 00:06:37.439 --> 00:06:39.819 engrossed in the actual real -world problems 00:06:39.819 --> 00:06:41.819 there that the piece of paper, the credential 00:06:41.819 --> 00:06:44.389 itself, it just didn't matter to him. that is 00:06:44.389 --> 00:06:47.029 the ultimate nerd flex i'm literally too busy 00:06:47.029 --> 00:06:49.189 inventing the future to go pick up my diploma 00:06:49.189 --> 00:06:52.230 from the registrar's office it is and for decades 00:06:52.230 --> 00:06:55.069 people thought the thesis was actually lost it 00:06:55.069 --> 00:06:57.529 became this kind of biographical mystery a holy 00:06:57.529 --> 00:07:01.850 grail then just a few years ago in 2020 the computer 00:07:01.850 --> 00:07:04.230 history museum worked with his family and finally 00:07:04.230 --> 00:07:07.209 dug it up from his home but it just shows you 00:07:07.209 --> 00:07:09.170 where his head was at he wasn't chasing titles 00:07:09.170 --> 00:07:11.529 or accolades he was chasing interesting problems 00:07:12.110 --> 00:07:14.790 And he went to chase those problems at Bell Labs. 00:07:15.250 --> 00:07:17.350 I feel like we need to pause on Bell Labs for 00:07:17.350 --> 00:07:20.250 a second. We all hear the name. But in the 1960s, 00:07:20.250 --> 00:07:22.870 what was this place really? Was it just a corporate 00:07:22.870 --> 00:07:26.310 lab? It was Mecca. It was the absolute engine 00:07:26.310 --> 00:07:29.610 room of American and really global innovation 00:07:29.610 --> 00:07:33.730 for half a century. You have to imagine a place 00:07:33.730 --> 00:07:37.490 that had the nearly unlimited funding of a massive 00:07:37.490 --> 00:07:40.470 corporation, AT &T, but the intellectual freedom. 00:07:41.209 --> 00:07:43.689 of a top -tier university. That's a rare combination. 00:07:44.550 --> 00:07:46.589 Incredibly rare. You could walk down the hallway 00:07:46.589 --> 00:07:48.170 and bump into a guy who had just won the Nobel 00:07:48.170 --> 00:07:50.430 Prize for discovering the cosmic microwave background 00:07:50.430 --> 00:07:52.649 radiation and then go have lunch with the team 00:07:52.649 --> 00:07:54.689 that was inventing the laser. The transistor 00:07:54.689 --> 00:07:57.110 was invented there. Information theory was born 00:07:57.110 --> 00:07:59.689 there. It sounds like a playground for geniuses. 00:07:59.750 --> 00:08:02.230 It was. And the management style was completely 00:08:02.230 --> 00:08:04.189 unique. They basically hired the most brilliant 00:08:04.189 --> 00:08:05.990 people they could find and said, go do something 00:08:05.990 --> 00:08:08.430 interesting. We trust you. They didn't have these 00:08:08.430 --> 00:08:10.829 quarterly product targets or marketing driven 00:08:10.829 --> 00:08:13.310 deadlines in the same way companies do today. 00:08:13.670 --> 00:08:16.449 And that freedom is the only reason Richie and 00:08:16.449 --> 00:08:17.949 his colleagues were able to do what they did. 00:08:18.089 --> 00:08:21.170 So he lands at Bell Labs and he meets Ken Thompson. 00:08:21.430 --> 00:08:23.730 And this is the partnership that changes everything. 00:08:23.949 --> 00:08:26.250 Lennon -McCartney of computing. Right. And they 00:08:26.250 --> 00:08:28.149 start working on this project called Multics. 00:08:28.370 --> 00:08:31.509 What was that? OK, so Multics. The name stands 00:08:31.509 --> 00:08:34.649 for Multiplexed Information and Computing Service. 00:08:35.009 --> 00:08:37.809 This was the Apollo program of computing in the 00:08:37.809 --> 00:08:42.409 60s. A moonshot. It was this massive joint project 00:08:42.409 --> 00:08:45.649 between Bell Labs, MIT, and General Electric. 00:08:45.870 --> 00:08:48.029 What was the goal? They wanted to build a computing 00:08:48.029 --> 00:08:50.879 utility. The idea was that computing would be 00:08:50.879 --> 00:08:52.899 like electricity or water. You just plug your 00:08:52.899 --> 00:08:55.679 terminal into the wall, and this massive mainframe 00:08:55.679 --> 00:08:57.820 somewhere else would serve up computing power 00:08:57.820 --> 00:09:00.600 to hundreds of users at once. That sounds suspiciously 00:09:00.600 --> 00:09:03.659 like the cloud or timesharing. It was a proto 00:09:03.659 --> 00:09:05.659 -cloud. It was an incredibly visionary idea. 00:09:06.240 --> 00:09:08.100 But the problem was they tried to do everything 00:09:08.100 --> 00:09:10.799 at once. The system became monstrously, unbelievably 00:09:10.799 --> 00:09:14.059 complex. It was slow. It cost a fortune to run. 00:09:14.240 --> 00:09:16.500 It was a classic case of design by committee 00:09:16.500 --> 00:09:19.299 on a grand scale. And Bell Labs, despite having 00:09:19.299 --> 00:09:21.360 deep pockets, eventually got sick of burning 00:09:21.360 --> 00:09:24.720 money on it. They pulled the plug. In 1969, they 00:09:24.720 --> 00:09:26.620 officially withdrew from the Multics project. 00:09:27.139 --> 00:09:30.480 They just realized it was a bottomless pit. So 00:09:30.480 --> 00:09:32.740 suddenly, Richie and Ken Thompson, who had been 00:09:32.740 --> 00:09:35.649 living in this world for years, were left. With 00:09:35.649 --> 00:09:39.289 nothing, no project, no massive mainframe to 00:09:39.289 --> 00:09:41.509 play with. This seems like the moment where most 00:09:41.509 --> 00:09:43.889 people in a big company would just, you know, 00:09:43.889 --> 00:09:45.870 dust themselves off and go find another corporate 00:09:45.870 --> 00:09:48.309 assignment. OK, boss, tell me what to work on 00:09:48.309 --> 00:09:51.190 next. But they didn't. They were frustrated. 00:09:51.570 --> 00:09:53.850 They had tasted what an operating system could 00:09:53.850 --> 00:09:55.529 be, something interactive, something that let 00:09:55.529 --> 00:09:58.350 multiple people work at once. And they did not 00:09:58.350 --> 00:10:00.309 want to go back to the old way of doing things. 00:10:00.429 --> 00:10:02.450 And what was that old way? Remind us what they 00:10:02.450 --> 00:10:04.230 were trying to avoid. The alternative was called 00:10:04.230 --> 00:10:08.129 batch processing. And it was soul crushing. Imagine 00:10:08.129 --> 00:10:10.289 you want to run a program. You can't just type 00:10:10.289 --> 00:10:12.690 it. You have to punch your code onto a stack 00:10:12.690 --> 00:10:15.470 of physical cards. Like with holes in them? Exactly. 00:10:15.649 --> 00:10:18.230 A stack of punch cards. You take your stack of 00:10:18.230 --> 00:10:21.009 cards to an operator, a guy in a white lab coat 00:10:21.009 --> 00:10:23.970 behind a big glass wall. You hand them over. 00:10:24.460 --> 00:10:26.000 He puts them in a queue with everyone else's 00:10:26.000 --> 00:10:28.340 jobs. You come back six or eight hours later, 00:10:28.460 --> 00:10:30.899 and he hands you a printout that says, error 00:10:30.899 --> 00:10:33.940 on line three. That sounds like actual hell. 00:10:34.080 --> 00:10:37.100 The feedback loop is hours long. It was an incredible 00:10:37.100 --> 00:10:39.779 amount of friction. Multics, for all its flaws, 00:10:40.059 --> 00:10:42.580 was supposed to be interactive. You type a command, 00:10:42.820 --> 00:10:45.919 the computer responds instantly. So when Multics 00:10:45.919 --> 00:10:48.379 was canceled, Thompson and Ritchie were desperate 00:10:48.379 --> 00:10:51.299 to keep that interactive environment alive. And 00:10:51.299 --> 00:10:53.500 Ken Thompson found an old, discarded computer 00:10:53.500 --> 00:10:57.139 in a corner of the lab, a PDP -7. A discarded 00:10:57.139 --> 00:10:59.659 computer, not the latest and greatest. Far from 00:10:59.659 --> 00:11:01.299 it. It was already obsolete, sitting unused. 00:11:01.740 --> 00:11:04.500 It was a mini -computer, which in 1969 meant 00:11:04.500 --> 00:11:06.299 it was only the size of a few filing cabinets, 00:11:06.480 --> 00:11:09.440 not an entire room. But it had terrible memory, 00:11:09.580 --> 00:11:12.240 very little processing power. The key thing, 00:11:12.240 --> 00:11:14.659 though, was that it was theirs. Nobody was watching 00:11:14.659 --> 00:11:16.820 them. No management oversight. So they start 00:11:16.820 --> 00:11:18.639 building their own operating system. Was it just 00:11:18.639 --> 00:11:20.820 for fun? Was there a business case? Not at all. 00:11:20.879 --> 00:11:24.500 It was for utility, born of necessity. Constance 00:11:24.500 --> 00:11:26.639 had written this game he liked called Space Travel, 00:11:26.820 --> 00:11:29.399 which simulated the physics of flying through 00:11:29.399 --> 00:11:31.120 the solar system. Our video game. Basically, 00:11:31.220 --> 00:11:33.919 yeah. And running it on the other available systems 00:11:33.919 --> 00:11:36.799 was expensive and clunky. So he decided to port 00:11:36.799 --> 00:11:39.669 the game to the little PDP -7. But to do that 00:11:39.669 --> 00:11:42.809 effectively, the machine needed a basic operating 00:11:42.809 --> 00:11:46.049 system, a file system. So Thompson starts writing 00:11:46.049 --> 00:11:48.590 one and Ritchie joins in. They started stripping 00:11:48.590 --> 00:11:51.090 away all the insane complexity of multics and 00:11:51.090 --> 00:11:53.850 keeping only the clean, simple ideas. And this 00:11:53.850 --> 00:11:55.389 is where the name comes from, right? This is 00:11:55.389 --> 00:11:57.950 a great story. It is. Their colleague, Brian 00:11:57.950 --> 00:11:59.889 Kernighan, who we mentioned earlier, joked that 00:11:59.889 --> 00:12:02.820 since the bloated multics was multiplexed. meaning 00:12:02.820 --> 00:12:05.620 many things at once. This new, scrappy, minimalist 00:12:05.620 --> 00:12:07.840 system was the opposite. It was Uniplex. It did 00:12:07.840 --> 00:12:10.759 one thing. Single tasking system. Right. So he 00:12:10.759 --> 00:12:13.840 called it Unix, U -N -I -C -S, as a pun. And 00:12:13.840 --> 00:12:16.519 that eventually got stylized as U -N -I -X. The 00:12:16.519 --> 00:12:18.740 most important piece of software in the world 00:12:18.740 --> 00:12:22.639 started as a dad joke. A nerd pun, yes. But the 00:12:22.639 --> 00:12:25.700 name perfectly reflected the philosophy. And 00:12:25.700 --> 00:12:27.639 this is where we really need to dig in. Because 00:12:27.639 --> 00:12:31.340 the Unix philosophy is the idea that truly changed 00:12:31.340 --> 00:12:34.350 the world. Multics tried to be a Swiss army knife 00:12:34.350 --> 00:12:36.629 with a thousand blades that was so heavy you 00:12:36.629 --> 00:12:39.710 couldn't lift it. Unix was a kitchen drawer full 00:12:39.710 --> 00:12:42.629 of individual, really sharp knives. Okay, break 00:12:42.629 --> 00:12:44.490 that analogy down for me. What does that mean 00:12:44.490 --> 00:12:47.110 in practice? The Unix philosophy, in a nutshell, 00:12:47.269 --> 00:12:50.509 is this. First, write programs that do one thing 00:12:50.509 --> 00:12:52.929 and do it well. Second, write programs that can 00:12:52.929 --> 00:12:56.039 work together. Third... write programs to handle 00:12:56.039 --> 00:12:58.840 text streams because that is a universal interface. 00:12:59.179 --> 00:13:00.899 Can you give me a real -world example of what 00:13:00.899 --> 00:13:02.799 that looks like? Absolutely. Let's say you have 00:13:02.799 --> 00:13:05.440 a huge text file, maybe a whole book, and you 00:13:05.440 --> 00:13:07.240 want to find every unique word in that book that 00:13:07.240 --> 00:13:09.899 contains the letter Z, sort them alphabetically, 00:13:09.899 --> 00:13:12.159 and then count how many there are. Okay, a pretty 00:13:12.159 --> 00:13:14.799 specific task. In the old way of thinking, you'd 00:13:14.799 --> 00:13:17.519 have to write a single, massive, complex program 00:13:17.519 --> 00:13:20.139 from scratch to do specifically that one task. 00:13:20.750 --> 00:13:23.769 In the Unix way, you use small tools and connect 00:13:23.769 --> 00:13:26.269 them with a concept called a pipe. You take a 00:13:26.269 --> 00:13:28.909 small program that just reads text. You pipe 00:13:28.909 --> 00:13:31.250 its output to another small program that just 00:13:31.250 --> 00:13:33.990 finds words matching a pattern. You pipe that 00:13:33.990 --> 00:13:36.149 output to a program that just sorts lines of 00:13:36.149 --> 00:13:38.570 text. You pipe that to a program that removes 00:13:38.570 --> 00:13:41.870 duplicates. And finally, you pipe that to a program 00:13:41.870 --> 00:13:44.649 that just counts lines. It's like a digital assembly 00:13:44.649 --> 00:13:47.830 line. Each station does one simple job. Precisely. 00:13:48.039 --> 00:13:50.360 You chain these small, simple, single -purpose 00:13:50.360 --> 00:13:52.840 tools together to accomplish a complex task. 00:13:53.500 --> 00:13:56.259 This concept of the pipe, represented by a vertical 00:13:56.259 --> 00:13:58.960 bar on the command line, was utterly revolutionary. 00:13:59.299 --> 00:14:01.379 It meant you didn't have to reinvent the wheel 00:14:01.379 --> 00:14:03.899 every time you had a new problem. You just rearranged 00:14:03.899 --> 00:14:05.779 the Lego blocks you already had. And Richie's 00:14:05.779 --> 00:14:09.600 role in this was what? He was crucial. While 00:14:09.600 --> 00:14:11.360 Thompson wrote a lot of the initial kernel and 00:14:11.360 --> 00:14:13.639 commands, Ritchie was deeply involved in the 00:14:13.639 --> 00:14:16.480 design philosophy. And very importantly, he was 00:14:16.480 --> 00:14:18.580 the primary designer of the Unix file system. 00:14:18.820 --> 00:14:21.179 What was special about the file system? The big 00:14:21.179 --> 00:14:24.419 idea was everything is a file. Your document 00:14:24.419 --> 00:14:27.500 is a file, obviously. But in Unix, your keyboard 00:14:27.500 --> 00:14:30.019 is also treated like a file. The screen you're 00:14:30.019 --> 00:14:32.840 typing on is a file. The printer is a file. Even 00:14:32.840 --> 00:14:35.200 the processes running on the system are represented 00:14:35.200 --> 00:14:38.019 as files. This incredible abstraction made the 00:14:38.019 --> 00:14:41.860 system unbelievably simple and flexible. You 00:14:41.860 --> 00:14:44.000 could use the same tools to read from a document 00:14:44.000 --> 00:14:46.200 as you did to read from the keyboard. Okay, so 00:14:46.200 --> 00:14:48.820 they have this cool, elegant OS running on their 00:14:48.820 --> 00:14:52.539 junker PDP -7 in a closet in New Jersey. But 00:14:52.539 --> 00:14:54.480 that's a long way from running the entire world. 00:14:54.639 --> 00:14:57.059 How does it escape the lab? That is the next 00:14:57.059 --> 00:15:00.039 great revolution, porting. See, back then, if 00:15:00.039 --> 00:15:02.139 you wrote software for a specific computer, like 00:15:02.139 --> 00:15:05.200 a PDP -7, it only ran on a PDP -7. It was completely 00:15:05.200 --> 00:15:07.139 locked to the hardware. The code was written 00:15:07.139 --> 00:15:09.039 in what's called assembly language, which is 00:15:09.039 --> 00:15:11.139 tied directly to the wiring and instruction set 00:15:11.139 --> 00:15:13.600 of that specific machine. So if Bell Labs bought 00:15:13.600 --> 00:15:16.139 a new, better computer a year later, all their 00:15:16.139 --> 00:15:18.980 Unix work would be trash. Basically. You'd have 00:15:18.980 --> 00:15:20.980 to start over and rewrite the whole thing from 00:15:20.980 --> 00:15:23.700 scratch for the new machine's architecture. It 00:15:23.700 --> 00:15:27.159 was a huge problem. Richie and Thompson did something 00:15:27.159 --> 00:15:30.580 that was considered heretical at the time. They 00:15:30.580 --> 00:15:32.399 decided they wanted to move Unix to a newer, 00:15:32.440 --> 00:15:35.399 more powerful machine, the PDP -11. But to do 00:15:35.399 --> 00:15:37.700 that effectively, they knew they needed a way 00:15:37.700 --> 00:15:40.360 to write the operating system in a language that 00:15:40.360 --> 00:15:42.899 wasn't tied to any one piece of hardware. And 00:15:42.899 --> 00:15:45.120 this is the bridge to the next huge topic. They 00:15:45.120 --> 00:15:47.840 needed a new language. Yes. This is the moment. 00:15:47.899 --> 00:15:50.879 To make Unix universal. To make it so it could 00:15:50.879 --> 00:15:53.000 live on any computer, past, present, or future. 00:15:53.789 --> 00:15:55.909 Dennis Ritchie had to invent the C programming 00:15:55.909 --> 00:15:58.090 language. All right. This is the part I've been 00:15:58.090 --> 00:16:00.149 dreading a little bit because I'm not a coder. 00:16:00.230 --> 00:16:03.789 I hear C and my brain just sees semicolons and 00:16:03.789 --> 00:16:06.309 error messages. How do we explain the importance 00:16:06.309 --> 00:16:08.350 of C to someone who has never written a line 00:16:08.350 --> 00:16:10.529 of code? Let's stick with analogies. Yeah. We 00:16:10.529 --> 00:16:12.289 talked about assembly language before. Imagine 00:16:12.289 --> 00:16:14.330 you want to build a simple brick wall. Assembly 00:16:14.330 --> 00:16:16.289 language is like giving instructions to a robot 00:16:16.289 --> 00:16:18.990 where you have to specify the movement of every 00:16:18.990 --> 00:16:22.139 single muscle fiber in its body. Contract bicep 00:16:22.139 --> 00:16:25.639 by 20%. Rotate wrist exactly four degrees. Grip 00:16:25.639 --> 00:16:28.919 brick with 10 .2 newtons of force. Extend arm. 00:16:29.200 --> 00:16:31.879 That sounds absolutely exhausting. And if you 00:16:31.879 --> 00:16:34.039 get one muscle twitch wrong, you drop the brick 00:16:34.039 --> 00:16:36.840 and the whole thing fails. Precisely. And even 00:16:36.840 --> 00:16:39.019 worse, if you get a new different model of robot, 00:16:39.259 --> 00:16:41.360 none of those muscle instructions make any sense 00:16:41.360 --> 00:16:43.700 anymore because the new robot has different arms 00:16:43.700 --> 00:16:46.779 and motors. That's assembly. It's incredibly 00:16:46.779 --> 00:16:49.299 fast and powerful if you get it right, but it's 00:16:49.299 --> 00:16:52.139 tedious, error -prone, and completely tied to 00:16:52.139 --> 00:16:54.580 one specific machine. So what's the opposite 00:16:54.580 --> 00:16:57.259 of that? What's a high -level language? A high 00:16:57.259 --> 00:16:59.460 -level language like Python or JavaScript today 00:16:59.460 --> 00:17:02.259 is like telling a human contractor, build me 00:17:02.259 --> 00:17:04.619 a four -foot -high brick wall right there. The 00:17:04.619 --> 00:17:06.599 contractor handles all the details of muscles 00:17:06.599 --> 00:17:09.400 and mortar. It's easy for you, the boss, to give 00:17:09.400 --> 00:17:11.859 the instruction, but you lose fine -grained control. 00:17:12.329 --> 00:17:14.269 Maybe the contractor builds it a bit slower than 00:17:14.269 --> 00:17:16.809 you'd like or uses a bit too much mortar. It's 00:17:16.809 --> 00:17:19.309 abstract and easy, but less efficient. Okay, 00:17:19.369 --> 00:17:21.809 so we have the micromanaging robot and the independent 00:17:21.809 --> 00:17:25.569 contractor. Where does C fit in this picture? 00:17:25.869 --> 00:17:28.769 C is the perfect middle ground. It's like you, 00:17:28.970 --> 00:17:31.730 the human, are wearing a powered exoskeleton. 00:17:32.470 --> 00:17:34.809 You are still moving the bricks yourself. You 00:17:34.809 --> 00:17:36.990 have all that direct control and speed and power, 00:17:37.089 --> 00:17:39.710 but the suit handles the tedious micromuscle 00:17:39.710 --> 00:17:41.730 movements for you. It lets you think in terms 00:17:41.730 --> 00:17:44.230 of pick up brick and place brick, and it translates 00:17:44.230 --> 00:17:46.490 those concepts into the machine's native instructions 00:17:46.490 --> 00:17:49.789 with incredible efficiency. So it's powerful 00:17:49.789 --> 00:17:52.769 and fast like the robot, but readable and logical 00:17:52.769 --> 00:17:55.190 like a human instruction. That's the magic of 00:17:55.190 --> 00:17:58.119 it. Richie created a language that was close 00:17:58.119 --> 00:18:00.400 to the metal, as programmers say. It allowed 00:18:00.400 --> 00:18:02.779 you to do things that were previously only possible 00:18:02.779 --> 00:18:05.500 in assembly, like managing memory addresses directly, 00:18:05.740 --> 00:18:08.019 which is risky but very powerful, while still 00:18:08.019 --> 00:18:10.200 using logical words and structures that humans 00:18:10.200 --> 00:18:13.180 can understand, like if, then, while, and for. 00:18:13.480 --> 00:18:15.680 You mentioned memory addresses. Why is that so 00:18:15.680 --> 00:18:18.599 important? I mean, today, I just save a file. 00:18:18.660 --> 00:18:20.039 I don't have to think about where on the hard 00:18:20.039 --> 00:18:23.000 drive it goes. In the 1970s, you absolutely had 00:18:23.000 --> 00:18:26.200 to think about it. Memory was scarce and incredibly 00:18:26.200 --> 00:18:30.079 expensive. A core innovation in C was the concept 00:18:30.079 --> 00:18:33.319 of pointers. A pointer is basically just a variable 00:18:33.319 --> 00:18:36.220 that holds a memory address. It allowed a programmer 00:18:36.220 --> 00:18:39.700 to say, hey, don't copy this entire massive encyclopedia 00:18:39.700 --> 00:18:41.940 from one part of memory to another new spot. 00:18:42.180 --> 00:18:44.460 Instead, just create a little note that points 00:18:44.460 --> 00:18:46.279 to the encyclopedia that's already on the shelf 00:18:46.279 --> 00:18:48.579 over there. So it saved a ton of space and time. 00:18:48.700 --> 00:18:51.339 A massive amount. It made programs incredibly 00:18:51.339 --> 00:18:54.220 fast and efficient. But it also made them dangerous. 00:18:54.539 --> 00:18:57.200 How so? Because if you're not careful, your pointer 00:18:57.200 --> 00:19:00.019 could point to the wrong shelf. Or you could 00:19:00.019 --> 00:19:02.079 change the data on the shelf when another part 00:19:02.079 --> 00:19:03.880 of the program was still expecting the old data 00:19:03.880 --> 00:19:06.279 to be there. Or you could forget to clean up 00:19:06.279 --> 00:19:08.059 the shelf when you're done, leading to memory 00:19:08.059 --> 00:19:11.059 leaks. This is the great trade -off of C5 -3. 00:19:12.160 --> 00:19:16.099 Richie gave programmers a chainsaw. Yes. It's 00:19:16.099 --> 00:19:18.559 an unbelievably powerful and efficient tool. 00:19:18.759 --> 00:19:21.690 You can build a log cabin in record time. But 00:19:21.690 --> 00:19:23.869 if you're careless for a split second, you can 00:19:23.869 --> 00:19:27.089 cut your own leg off. Many modern languages have 00:19:27.089 --> 00:19:30.029 safety guards on the chainsaw. C does not. It 00:19:30.029 --> 00:19:32.029 assumes you're an expert and trusts you with 00:19:32.029 --> 00:19:34.349 the full power of the machine. And this language, 00:19:34.609 --> 00:19:37.630 C, was what allowed them to finally move Unix 00:19:37.630 --> 00:19:40.789 off that one machine. This was the absolute game 00:19:40.789 --> 00:19:43.490 changer. They rewrote almost the entire Unix 00:19:43.490 --> 00:19:46.750 kernel and all its utilities in C. This was unheard 00:19:46.750 --> 00:19:48.869 of. Operating systems were supposed to be written 00:19:48.869 --> 00:19:52.420 in assembly. But by rewriting it in C, the operating 00:19:52.420 --> 00:19:55.099 system was suddenly no longer a PDK -11 thing. 00:19:55.240 --> 00:19:57.319 It was a software thing. It was abstract. So 00:19:57.319 --> 00:19:59.660 if you wanted to run Unix on a brand new supercomputer 00:19:59.660 --> 00:20:01.619 from a different company. You just had to do 00:20:01.619 --> 00:20:04.420 one relatively small job. Write a C compiler 00:20:04.420 --> 00:20:07.500 for that new machine. A C compiler is a program 00:20:07.500 --> 00:20:10.240 that translates C code into that specific machine's 00:20:10.240 --> 00:20:12.500 native language. Once you had that compiler, 00:20:12.839 --> 00:20:15.079 the entire Unix operating system, millions of 00:20:15.079 --> 00:20:17.589 lines of code, would just compile and run. This 00:20:17.589 --> 00:20:19.630 broke the monopoly that hardware manufacturers 00:20:19.630 --> 00:20:22.430 had on software. It completely destroyed it. 00:20:22.450 --> 00:20:24.549 It decoupled the software from the hardware. 00:20:24.710 --> 00:20:27.430 It meant the software was now portable, independent 00:20:27.430 --> 00:20:30.869 of the machine it ran on. In a sense, it commoditized 00:20:30.869 --> 00:20:33.190 the hardware. The value shifted from the box 00:20:33.190 --> 00:20:35.750 to the code running inside it. This brings us 00:20:35.750 --> 00:20:38.450 to the book. When I was preparing for this, I 00:20:38.450 --> 00:20:40.990 was told I absolutely have to ask about K &R. 00:20:41.170 --> 00:20:45.339 Ah, yes. The white book. The C Programming Language, 00:20:45.640 --> 00:20:48.440 published in 1978, written by Brian Kernaghan 00:20:48.440 --> 00:20:51.579 and Dennis Ritchie. In the industry to this day, 00:20:51.680 --> 00:20:54.259 it is known simply as K &R. What is it, just 00:20:54.259 --> 00:20:56.180 a textbook, a manual? It's so much more than 00:20:56.180 --> 00:20:58.440 that. It is a masterpiece of technical writing. 00:20:59.400 --> 00:21:01.680 Computer manuals then and now are like reading 00:21:01.680 --> 00:21:04.279 a tax code. They are thick, dense, and impossible 00:21:04.279 --> 00:21:07.319 to get through. K &R was thin. It was concise. 00:21:07.539 --> 00:21:10.299 It was beautifully, clearly written. It didn't 00:21:10.299 --> 00:21:12.039 just teach you the syntax of the language. It 00:21:12.039 --> 00:21:13.940 taught you the style and the philosophy. It set 00:21:13.940 --> 00:21:16.440 the vibe for the entire C programming community. 00:21:16.900 --> 00:21:19.519 It absolutely did. It preached clarity over cleverness. 00:21:20.720 --> 00:21:23.210 Simplicity. Don't write tricky code. Write clear 00:21:23.210 --> 00:21:26.049 code. That book is still on the shelves of professional 00:21:26.049 --> 00:21:28.829 software engineers today, almost 50 years later. 00:21:29.210 --> 00:21:31.529 It's a classic for a reason. So now we have the 00:21:31.529 --> 00:21:34.029 two pillars in place. We have the OS, Unix, and 00:21:34.029 --> 00:21:37.490 the language C. Now let's fast forward. We need 00:21:37.490 --> 00:21:40.190 to connect the dots. Why does the person listening 00:21:40.190 --> 00:21:43.210 to this on an iPhone 16 or an Android phone owe 00:21:43.210 --> 00:21:45.549 Dennis Ritchie a thank you? Okay, let's trace 00:21:45.549 --> 00:21:48.200 the family tree. And it's actually shocking how 00:21:48.200 --> 00:21:50.160 direct the lineage is. Let's start with Unix. 00:21:50.500 --> 00:21:52.519 Richie himself said in an interview late in his 00:21:52.519 --> 00:21:55.240 life that he views Linux as a direct and healthy 00:21:55.240 --> 00:21:58.559 continuation of Unix ideas. And Linux today is 00:21:58.559 --> 00:22:01.220 everywhere. Linux runs the Internet. The vast 00:22:01.220 --> 00:22:03.259 majority of servers that make up the cloud, Amazon, 00:22:03.480 --> 00:22:06.619 AWS, Google Cloud, Microsoft, Azure, they all 00:22:06.619 --> 00:22:08.039 run on Linux. So everything you stream a movie 00:22:08.039 --> 00:22:10.599 on Netflix, do a Google search, use an AI chatbot, 00:22:10.759 --> 00:22:13.579 you're using a system that is, for all intents 00:22:13.579 --> 00:22:15.920 and purposes, a clone of Unix. What about the 00:22:15.920 --> 00:22:17.319 phone in my pocket? Let's start with Android. 00:22:17.519 --> 00:22:20.140 Android uses the Linux kernel. Hold on. Define 00:22:20.140 --> 00:22:22.200 kernel for us. I hear that word all the time, 00:22:22.220 --> 00:22:24.279 but what is it really? The kernel is the core. 00:22:24.759 --> 00:22:27.019 It's the absolute heart of the operating system. 00:22:27.160 --> 00:22:30.099 It's the first piece of software that loads when 00:22:30.099 --> 00:22:32.579 the computer boots up. And it's the piece that 00:22:32.579 --> 00:22:34.900 talks directly to the hardware. It manages the 00:22:34.900 --> 00:22:37.740 CPU, the memory, the battery, the radios, the 00:22:37.740 --> 00:22:40.059 interface you see, the buttons and icons you 00:22:40.059 --> 00:22:42.920 click. That's like the body of the car. The kernel 00:22:42.920 --> 00:22:45.579 is the engine. So Android's engine is Linux. 00:22:45.779 --> 00:22:47.900 Which is a Unix clone. So every Android phone 00:22:47.900 --> 00:22:50.440 is a Unix machine. Okay, what about Apple? Apple 00:22:50.440 --> 00:22:52.640 is famous for being the odd one out. They always 00:22:52.640 --> 00:22:54.140 like to do their own thing, the think different 00:22:54.140 --> 00:22:56.680 mantra. Not here they don't. This is the ultimate 00:22:56.680 --> 00:22:59.839 irony. Steve Jobs' greatest software achievement, 00:23:00.079 --> 00:23:03.700 macOS and by extension iOS, which runs the iPhone 00:23:03.700 --> 00:23:06.519 and iPad, is built directly on top of a system 00:23:06.519 --> 00:23:10.220 called BSD. And BSD is? Berkeley Software Distribution. 00:23:10.319 --> 00:23:13.200 And what is BSD? It's a version of Unix that 00:23:13.200 --> 00:23:14.579 was developed at the University of California, 00:23:14.799 --> 00:23:17.160 Berkeley in the 1970s, based on the original 00:23:17.160 --> 00:23:19.380 source code from Bell Labs. So you're telling 00:23:19.380 --> 00:23:22.119 me when I slide to unlock my iPhone deep, deep 00:23:22.119 --> 00:23:24.859 down in the basement of that software? It's Dennis 00:23:24.859 --> 00:23:28.160 Ritchie's logic. In fact, if you open the Terminal 00:23:28.160 --> 00:23:30.779 app on any Mac today, you are dropped into a 00:23:30.779 --> 00:23:33.880 Unix command line. You can type commands like 00:23:33.880 --> 00:23:36.839 Ls to list files or grep to search for text, 00:23:37.000 --> 00:23:39.779 the very same tools with the very same names 00:23:39.779 --> 00:23:42.180 that Ritchie and Thompson designed on a PDP -11 00:23:42.180 --> 00:23:45.460 in 1971. That is just wild. It's like finding 00:23:45.460 --> 00:23:47.680 out that the engine in a brand new Tesla is actually 00:23:47.680 --> 00:23:50.299 a steam engine from the 1800s, but, you know, 00:23:50.339 --> 00:23:52.640 in a good way. It's more like finding out that 00:23:52.640 --> 00:23:54.759 the laws of physics that govern the Tesla's electric 00:23:54.759 --> 00:23:57.480 motor are the same ones Newton discovered 300 00:23:57.480 --> 00:24:00.160 years ago. Ritchie didn't just invent a product. 00:24:00.279 --> 00:24:02.519 He discovered and defined the fundamental physics 00:24:02.519 --> 00:24:04.420 of how modern software should be structured. 00:24:04.940 --> 00:24:06.849 And then there's the other side of it. The C 00:24:06.849 --> 00:24:09.769 family tree. Right. Because C itself isn't really 00:24:09.769 --> 00:24:12.150 used for writing most iPhone apps anymore, is 00:24:12.150 --> 00:24:14.529 it? Programmers use newer languages like Swift. 00:24:14.809 --> 00:24:17.509 You write the app in Swift, yes. But the Swift 00:24:17.509 --> 00:24:20.950 compiler, the tool that takes your human -readable 00:24:20.950 --> 00:24:23.109 Swift code and turns it into an app the iPhone 00:24:23.109 --> 00:24:27.049 can actually run, is written in C++ Mark. And 00:24:27.049 --> 00:24:29.470 what is C++? It's literally just C with a bunch 00:24:29.470 --> 00:24:31.490 of extra features bolted on. It was originally 00:24:31.490 --> 00:24:35.240 called C with glasses. So C is building the tools 00:24:35.240 --> 00:24:37.160 that build the apps. It's the foundation of the 00:24:37.160 --> 00:24:39.900 whole tool chain. What about Java, the language 00:24:39.900 --> 00:24:42.319 that runs huge corporate systems and was the 00:24:42.319 --> 00:24:45.059 original language for Android apps? Its syntax 00:24:45.059 --> 00:24:47.680 is based directly on C JavaScript, the language 00:24:47.680 --> 00:24:49.980 that makes the entire web interactive. C style 00:24:49.980 --> 00:24:52.740 syntax. Python, the language of data science 00:24:52.740 --> 00:24:55.440 and AI. The main Python interpreter, the program 00:24:55.440 --> 00:24:57.779 that actually runs Python code, is written in 00:24:57.779 --> 00:25:00.599 C. So it really is the DNA. It is the absolute 00:25:00.599 --> 00:25:03.339 genetic ancestor. Every single major programming 00:25:03.339 --> 00:25:05.920 language in use today is either a direct descendant 00:25:05.920 --> 00:25:08.779 of C, is heavily influenced by C's syntax, or 00:25:08.779 --> 00:25:18.079 is built using tools written in C. I remember 00:25:18.079 --> 00:25:21.279 reading one commentator who said, Richie co -invented 00:25:21.279 --> 00:25:24.380 the DNA of effectively every single computer 00:25:24.380 --> 00:25:27.859 software product we use. I honestly thought that 00:25:27.859 --> 00:25:29.859 was an exaggeration when I first read it. It's 00:25:29.859 --> 00:25:31.890 not. It's probably an understatement. Yeah. Doug 00:25:31.890 --> 00:25:33.769 McElroy, one of their brilliant colleagues at 00:25:33.769 --> 00:25:36.390 Bell Labs, put it best. He said the names of 00:25:36.390 --> 00:25:38.410 Ritchie and Thompson may safely be assumed be 00:25:38.410 --> 00:25:41.369 attached to almost everything not otherwise attributed. 00:25:41.750 --> 00:25:44.029 Meaning if you look at a piece of foundational 00:25:44.029 --> 00:25:46.630 software and you don't know who built it, it 00:25:46.630 --> 00:25:48.670 was probably them. That's the gist of it. Their 00:25:48.670 --> 00:25:51.130 influence is the default. I want to pivot a bit 00:25:51.130 --> 00:25:53.569 to the man himself, because usually when you 00:25:53.569 --> 00:25:56.440 have someone this influential. Steve Jobs, Bill 00:25:56.440 --> 00:25:59.140 Gates, and Elon Musk, they are these massive, 00:25:59.220 --> 00:26:01.740 complicated, often difficult personalities. They 00:26:01.740 --> 00:26:03.259 seek out the spotlight. They want the credit. 00:26:03.599 --> 00:26:07.220 Dennis Ritchie seems different. He was the absolute 00:26:07.220 --> 00:26:10.640 antithesis of the modern tech bro CEO. He was 00:26:10.640 --> 00:26:13.480 quiet. He was famously humble. He lived alone. 00:26:13.740 --> 00:26:16.240 He wasn't trying to disrupt industries or make 00:26:16.240 --> 00:26:18.480 billions of dollars or become a celebrity. He 00:26:18.480 --> 00:26:20.759 was a scientist. He just wanted to build better 00:26:20.759 --> 00:26:23.289 tools for himself and his friends. There's this 00:26:23.289 --> 00:26:25.589 story about the NSA that I think is just incredible. 00:26:25.849 --> 00:26:27.589 It feels like it's pulled from a spy thriller 00:26:27.589 --> 00:26:30.869 for a second. It really does. So in the 1970s, 00:26:30.869 --> 00:26:33.730 Ritchie was also dabbling in cryptography, because 00:26:33.730 --> 00:26:36.250 of course he was. He was working with two colleagues, 00:26:36.490 --> 00:26:39.329 James Reeds and Robert Morris, on what's called 00:26:39.329 --> 00:26:42.329 a ciphertext -only attack. Which means cracking 00:26:42.329 --> 00:26:44.390 a secret code without having the decryption key, 00:26:44.529 --> 00:26:46.849 right? Exactly. You just have the scramble messages. 00:26:47.250 --> 00:26:51.130 And they were targeting the US M209 cipher machine. 00:26:51.740 --> 00:26:54.079 This was a mechanical encryption device used 00:26:54.079 --> 00:26:57.000 by the U .S. military during World War II and 00:26:57.000 --> 00:26:59.119 the Korean War. It was supposed to be secure. 00:26:59.720 --> 00:27:02.140 Well, Ritchie and his team wrote a program that 00:27:02.140 --> 00:27:04.140 could crack it. They figured out that if they 00:27:04.140 --> 00:27:07.880 had about 2 ,000, 2 ,500 letters of coded text, 00:27:08.059 --> 00:27:10.660 they could solve it. That's a huge deal. They 00:27:10.660 --> 00:27:12.940 basically proved a piece of U .S. military encryption 00:27:12.940 --> 00:27:15.519 hardware was fundamentally broken. It was a massive 00:27:15.519 --> 00:27:17.799 discovery. So they wrote a formal paper on it. 00:27:17.839 --> 00:27:19.299 They were getting ready to publish it in an academic 00:27:19.299 --> 00:27:22.450 journal. The NSA, the National Security Agency, 00:27:22.569 --> 00:27:24.750 came knocking on Bell Labs' door. The actual 00:27:24.750 --> 00:27:27.230 men in black. Pretty much. They came and they 00:27:27.230 --> 00:27:29.450 had a quiet conversation with Ritchie and Bell 00:27:29.450 --> 00:27:32.390 Labs management. And they said, look, we know 00:27:32.390 --> 00:27:35.410 you cracked it. We're not surprised. But you 00:27:35.410 --> 00:27:39.210 cannot publish this paper. And their reason was? 00:27:39.470 --> 00:27:41.869 Their reason was that several foreign governments, 00:27:42.109 --> 00:27:44.190 some of them adversaries of the United States, 00:27:44.369 --> 00:27:47.950 were still using that same M209 machine because 00:27:47.950 --> 00:27:50.200 they thought it was secure. The NSA was happily 00:27:50.200 --> 00:27:52.420 reading their mail. If Ritchie published his 00:27:52.420 --> 00:27:54.839 paper, these governments would find out their 00:27:54.839 --> 00:27:57.279 codes were broken, they'd switch to a new system, 00:27:57.400 --> 00:27:59.700 and a valuable source of intelligence would go 00:27:59.700 --> 00:28:02.539 dark. Wow. Okay, so this is a real ethical dilemma. 00:28:02.680 --> 00:28:04.779 This is the moment where the modern information 00:28:04.779 --> 00:28:07.680 wants to be free. Hacktivists would just dump 00:28:07.680 --> 00:28:09.839 the paper on the Internet anyway. Consequences 00:28:09.839 --> 00:28:11.980 be damned. Right. You can imagine that happening 00:28:11.980 --> 00:28:14.099 today. But what did Dennis Ritchie do? He just 00:28:14.099 --> 00:28:16.220 said, okay, and he put the paper in a drawer. 00:28:17.319 --> 00:28:20.119 And it stayed there, unpublished, for decades. 00:28:20.460 --> 00:28:22.740 He didn't fight it? He didn't argue for academic 00:28:22.740 --> 00:28:26.740 freedom? No. He just complied. He wasn't an anarchist. 00:28:26.759 --> 00:28:29.279 He was a scientist. But he was also a pragmatist. 00:28:29.319 --> 00:28:31.859 He didn't have this massive ego that needed the 00:28:31.859 --> 00:28:35.539 fame of being the guy who beat the NSA. The intellectual 00:28:35.539 --> 00:28:38.059 challenge was solved. That was enough for him. 00:28:38.490 --> 00:28:41.250 So he just went back to work on Unix. That humility 00:28:41.250 --> 00:28:43.690 seems to be the central theme of his life. He 00:28:43.690 --> 00:28:46.470 wins the Turing Award in 1983, which is... The 00:28:46.470 --> 00:28:49.390 Nobel Prize of Computing. The highest honor. 00:28:49.630 --> 00:28:51.789 He gets the National Medal of Technology from 00:28:51.789 --> 00:28:54.609 President Bill Clinton in 1998. The citation 00:28:54.609 --> 00:28:57.890 says he stimulated the growth of an entire industry. 00:28:58.599 --> 00:29:00.839 But he never really cashed in, did he? Never. 00:29:01.039 --> 00:29:03.920 He stayed at Bell Labs, which later became Lucent 00:29:03.920 --> 00:29:06.900 his entire career. He was managing a small research 00:29:06.900 --> 00:29:08.519 group towards the end, but he was still writing 00:29:08.519 --> 00:29:10.900 code. He wasn't on the board of directors. He 00:29:10.900 --> 00:29:12.700 wasn't doing a book tour or giving TED Talks. 00:29:12.740 --> 00:29:14.640 There's a visual description of him that I love 00:29:14.640 --> 00:29:17.460 from his colleagues. He was often just seen walking 00:29:17.460 --> 00:29:19.619 the halls in a simple striped sweater, sometimes 00:29:19.619 --> 00:29:21.920 with a big beard, looking like a slightly distracted 00:29:21.920 --> 00:29:24.920 but very calm professor. The calm genius archetype. 00:29:24.980 --> 00:29:27.400 Exactly. And he was incredibly generous with 00:29:27.400 --> 00:29:29.890 credit. Whenever people interviewed him about 00:29:29.890 --> 00:29:32.829 Unix or C, he would constantly deflect the praise. 00:29:33.170 --> 00:29:35.329 He'd insist that Unix was just a continuation 00:29:35.329 --> 00:29:38.789 of ideas from Multics and other systems. He refused 00:29:38.789 --> 00:29:42.509 to take sole credit. He always, always pointed 00:29:42.509 --> 00:29:44.490 Ken Thompson to Brian Kernighan to the whole 00:29:44.490 --> 00:29:46.710 collaborative environment of Bell Labs. Which 00:29:46.710 --> 00:29:48.569 makes the end of his life and the lack of public 00:29:48.569 --> 00:29:51.369 recognition even more poignant. We talked about 00:29:51.369 --> 00:29:54.089 the timing, October 2011. Yeah, it was a rough 00:29:54.089 --> 00:29:56.490 end. He had been in frail health for a while. 00:29:56.529 --> 00:29:58.730 He was... battling prostate cancer and heart 00:29:58.730 --> 00:30:01.769 disease. He died alone in his home. The news 00:30:01.769 --> 00:30:04.410 actually broke because Rob Pike, another former 00:30:04.410 --> 00:30:07.069 colleague and a giant in the field himself, posted 00:30:07.069 --> 00:30:09.789 about it on Google Plus Skip. Google Plus Skip. 00:30:09.849 --> 00:30:11.829 Wow, that dates it right there. It really does. 00:30:12.049 --> 00:30:14.910 And because the entire global media machine was 00:30:14.910 --> 00:30:16.789 in the middle of this week -long deification 00:30:16.789 --> 00:30:19.630 of Steve Jobs, the obituaries for Richie were 00:30:19.630 --> 00:30:21.710 small. They were footnotes. But the tech community, 00:30:21.950 --> 00:30:24.009 the people who actually build things, they knew. 00:30:24.309 --> 00:30:26.980 Oh, they knew. The tributes from that community 00:30:26.980 --> 00:30:30.779 were quiet, but so heartfelt. And they were geeky 00:30:30.779 --> 00:30:34.119 and beautiful. The Fedora 16 Linux distribution, 00:30:34.500 --> 00:30:36.740 which was released a month after he died, was 00:30:36.740 --> 00:30:39.779 officially dedicated to his memory. FreeBSD 9 00:30:39.779 --> 00:30:42.279 .0, another major open source operating system, 00:30:42.440 --> 00:30:44.960 was also dedicated to him. And my personal favorite 00:30:44.960 --> 00:30:48.680 tribute. The asteroid, yes. In 2012, astronomers 00:30:48.680 --> 00:30:51.380 Tom Glenos and David H. Levy officially named 00:30:51.380 --> 00:30:55.329 asteroid 294727 Dennis Ritchie. So he's literally 00:30:55.329 --> 00:30:58.109 floating out there in the cosmos. His legacy 00:30:58.109 --> 00:31:00.490 is in the sky and in the cloud that runs on his 00:31:00.490 --> 00:31:03.049 OS and in the phone in your pocket. He's everywhere. 00:31:03.269 --> 00:31:05.309 I want to circle all the way back to that comparison 00:31:05.309 --> 00:31:07.609 we started with, Jobs versus Richie. Now that 00:31:07.609 --> 00:31:09.730 we've gone through this whole history, how do 00:31:09.730 --> 00:31:11.549 you see it? It feels like we need both types 00:31:11.549 --> 00:31:14.609 of people, but our culture over -indexes so heavily 00:31:14.609 --> 00:31:16.750 on one of them. That is the perfect way to put 00:31:16.750 --> 00:31:19.789 it. We over -index on the visionary marketer. 00:31:19.970 --> 00:31:22.849 And Steve Jobs was a genius at that. Let's not 00:31:22.849 --> 00:31:25.230 take anything away from him. He understood human 00:31:25.230 --> 00:31:28.549 psychology. He understood desire. He knew how 00:31:28.549 --> 00:31:31.950 to package technology into a beautiful, simple 00:31:31.950 --> 00:31:33.970 object that we would fall in love with. He was 00:31:33.970 --> 00:31:36.910 a master of the user experience. But Ritchie 00:31:36.910 --> 00:31:39.509 was the visionary architect. He understood the 00:31:39.509 --> 00:31:42.150 structural integrity of the entire digital world. 00:31:42.329 --> 00:31:44.970 The Economist had that great line in his obituary. 00:31:45.269 --> 00:31:48.250 The gizmos of the digital age owe a part of their 00:31:48.250 --> 00:31:51.220 numeric souls to Dennis Ritchie. And the Deccan 00:31:51.220 --> 00:31:54.019 Herald, a newspaper in India, went even further. 00:31:54.200 --> 00:31:56.500 They just flat out said Ritchie was far more 00:31:56.500 --> 00:31:59.119 influential than Jobs. And you can make a strong 00:31:59.119 --> 00:32:01.259 case for that. What's the case? Here's the litmus 00:32:01.259 --> 00:32:04.240 test I always use. If Steve Jobs had never existed, 00:32:04.420 --> 00:32:06.259 your smartphone today would almost certainly 00:32:06.259 --> 00:32:08.559 look different. Maybe you'd have a physical keyboard. 00:32:08.619 --> 00:32:11.259 Maybe the interface wouldn't be as slick. But 00:32:11.259 --> 00:32:13.319 it would very likely still exist in some form. 00:32:13.440 --> 00:32:15.900 Another company would have figured it out. But 00:32:15.900 --> 00:32:18.079 if Dennis Ritchie had never existed... We might 00:32:18.079 --> 00:32:20.319 still be punching cards. It's a real possibility. 00:32:20.680 --> 00:32:23.900 Or, more likely, we would be stuck in a horribly 00:32:23.900 --> 00:32:27.180 fragmented world of proprietary, incompatible 00:32:27.180 --> 00:32:30.000 systems where nothing talks to anything else. 00:32:30.180 --> 00:32:33.740 The internet as we know it, which relies on open 00:32:33.740 --> 00:32:35.720 standards and portable software that can run 00:32:35.720 --> 00:32:39.140 on any machine, might have been impossible. Richie 00:32:39.140 --> 00:32:41.720 gave us the universal language and the universal 00:32:41.720 --> 00:32:43.880 blueprint. It really makes me think about where 00:32:43.880 --> 00:32:46.500 we are today in the tech world. Richie and his 00:32:46.500 --> 00:32:48.799 colleagues worked in this open, collaborative 00:32:48.799 --> 00:32:51.900 environment. Bell Labs shared the Unix source 00:32:51.900 --> 00:32:54.420 code with universities. For free, basically. 00:32:54.559 --> 00:32:56.240 They wanted people to play with it, to learn 00:32:56.240 --> 00:32:58.240 from it, to improve it. It was an academic project, 00:32:58.420 --> 00:33:00.759 not a commercial one. Exactly. And now you look 00:33:00.759 --> 00:33:03.200 at our phones. They are the definition of walled 00:33:03.200 --> 00:33:05.420 gardens. They have locked bootloaders. They have 00:33:05.420 --> 00:33:07.579 app stores that you cannot bypass. The hardware 00:33:07.579 --> 00:33:11.240 is sealed shut. That is the big, final, provocative 00:33:11.240 --> 00:33:13.680 question for you, the listener, to think about. 00:33:14.220 --> 00:33:17.279 We built this entire modern world on the back 00:33:17.279 --> 00:33:20.740 of Richie's open, modular, tool -based philosophy. 00:33:21.549 --> 00:33:24.670 But the dominant companies of today are increasingly 00:33:24.670 --> 00:33:27.490 locking that world down into closed proprietary 00:33:27.490 --> 00:33:31.970 ecosystems. Are we still funding the kind of 00:33:31.970 --> 00:33:35.190 fundamental curiosity -driven research that allows 00:33:35.190 --> 00:33:38.349 a Dennis Ritchie to play with a junker PDP -7 00:33:38.349 --> 00:33:41.089 for a year with no immediate commercial goal? 00:33:41.309 --> 00:33:43.769 Or are we just funding the next iteration of 00:33:43.769 --> 00:33:46.380 the shiny fashion item? Exactly. Are we building 00:33:46.380 --> 00:33:48.539 foundational tools for the future, or are we 00:33:48.539 --> 00:33:50.380 just building disposable products for today? 00:33:50.759 --> 00:33:53.440 Richie built tools, and tools are what empower 00:33:53.440 --> 00:33:55.140 the next generation to build something that we 00:33:55.140 --> 00:33:57.619 can't even imagine yet. That's a heavy, important 00:33:57.619 --> 00:33:59.759 thought to end on. I can guarantee that the next 00:33:59.759 --> 00:34:01.380 time I pick up my phone, I'm not going to see 00:34:01.380 --> 00:34:03.220 the screen and the apps. I'm going to be thinking 00:34:03.220 --> 00:34:05.400 about the engine underneath it all. Think about 00:34:05.400 --> 00:34:08.039 the Unix philosophy. Think about the C code running 00:34:08.039 --> 00:34:10.099 the kernel. Think about the quiet guy in the 00:34:10.099 --> 00:34:12.619 striped sweater. Think about Dennis. Thanks so 00:34:12.619 --> 00:34:14.400 much for taking us through this deep dive. It's 00:34:14.400 --> 00:34:17.199 always a pleasure. His is a story that needs 00:34:17.199 --> 00:34:17.780 to be told.