WEBVTT 00:00:00.000 --> 00:00:01.700 Welcome to the Deep Dive, the show where we take 00:00:01.700 --> 00:00:03.859 a monumental stack of source material articles, 00:00:04.160 --> 00:00:07.139 biographies, technical papers, and really just 00:00:07.139 --> 00:00:08.880 drill down to give you the insights you need 00:00:08.880 --> 00:00:11.009 to understand the big picture. And today we are 00:00:11.009 --> 00:00:14.210 digging into the life and, I mean, the truly 00:00:14.210 --> 00:00:16.550 monumental intellectual architecture of Norbert 00:00:16.550 --> 00:00:18.890 Wiener. He's a fascinating figure, a man who, 00:00:18.929 --> 00:00:21.370 you know, long before Silicon Valley was even 00:00:21.370 --> 00:00:24.370 a concept, was building this theoretical bridge 00:00:24.370 --> 00:00:26.649 between the nervous system of a living creature 00:00:26.649 --> 00:00:29.410 and the electronic circuits of a machine. Absolutely. 00:00:29.449 --> 00:00:31.769 We're looking at the whole trajectory here. Norbert 00:00:31.769 --> 00:00:34.670 Wiener, who lived from 1894 to 1964, he was an 00:00:34.670 --> 00:00:36.409 American computer scientist, a mathematician 00:00:36.409 --> 00:00:39.340 of... really the highest order, and a philosopher. 00:00:39.640 --> 00:00:42.399 He's the figure who connects abstract math directly 00:00:42.399 --> 00:00:45.780 to engineering, to systems control, and maybe 00:00:45.780 --> 00:00:47.899 most importantly, to the ethics of it all. Right. 00:00:48.060 --> 00:00:50.960 And our sources for this deep dive cover a huge 00:00:50.960 --> 00:00:54.240 range, biographical reviews, his own writings, 00:00:54.420 --> 00:00:56.780 technical analyses. We're going to trace his 00:00:56.780 --> 00:01:00.799 story from this notorious child prodigy through 00:01:00.799 --> 00:01:03.820 his absolutely crucial and, as we'll see, ethically 00:01:03.820 --> 00:01:07.090 conflicted work during World War II. to his philosophical 00:01:07.090 --> 00:01:10.129 legacy. And our mission really is to unpack his 00:01:10.129 --> 00:01:12.769 core insight, this idea that intelligent behavior, 00:01:13.049 --> 00:01:15.090 whether you're looking at an animal or a machine, 00:01:15.250 --> 00:01:18.150 isn't some kind of magic. It's the predictable 00:01:18.150 --> 00:01:21.189 result of control achieved through feedback. 00:01:21.489 --> 00:01:23.939 Feedback mechanisms. Yeah. And we can't just 00:01:23.939 --> 00:01:25.819 talk about the philosophy. We have to get into 00:01:25.819 --> 00:01:27.959 the tools he built to make it real. Exactly. 00:01:27.959 --> 00:01:30.079 We have to look at the math, especially the Wiener 00:01:30.079 --> 00:01:32.680 filter, which is, you know, the bedrock for prediction 00:01:32.680 --> 00:01:34.900 in a world that's just full of noise and randomness. 00:01:35.040 --> 00:01:38.140 He didn't just have an idea. He gave us the equations 00:01:38.140 --> 00:01:41.500 to build it. So why should you, our listener, 00:01:41.680 --> 00:01:44.540 care about a mathematician who died six decades 00:01:44.540 --> 00:01:47.780 ago? Well, because Wiener's influence is, and 00:01:47.780 --> 00:01:50.519 I don't think this is an overstatement. It really 00:01:50.519 --> 00:01:54.239 is. His work literally defined the field of cybernetics. 00:01:54.420 --> 00:01:56.719 And that definition, the science of communication 00:01:56.719 --> 00:02:00.040 and control in the animal and the machine, that's 00:02:00.040 --> 00:02:01.939 the intellectual starting point for modern AI. 00:02:02.239 --> 00:02:05.060 In his shadow, it stretches so much wider than 00:02:05.060 --> 00:02:07.870 just computing. Oh, absolutely. His ideas were 00:02:07.870 --> 00:02:09.650 fundamental for Claude Shannon's information 00:02:09.650 --> 00:02:12.189 theory, for the computer architecture that John 00:02:12.189 --> 00:02:14.949 von Neumann was building. But they also resonated 00:02:14.949 --> 00:02:17.830 deep in the humanities, influencing anthropologists 00:02:17.830 --> 00:02:20.610 like Margaret Mead and Gregory Bateson. So he 00:02:20.610 --> 00:02:22.669 was showing that these same principles of control 00:02:22.669 --> 00:02:25.710 and communication applied just as much to, say, 00:02:25.849 --> 00:02:29.189 a social system as they did to a circuit board. 00:02:29.530 --> 00:02:31.990 Precisely. When we talk about autonomous cars 00:02:31.990 --> 00:02:34.370 or machine learning today, we are speaking the 00:02:34.370 --> 00:02:36.830 language of control and feedback that Norbert 00:02:36.830 --> 00:02:40.030 Wiener formalized. This deep dive is about understanding 00:02:40.030 --> 00:02:43.150 the origin story of our automated world. Okay, 00:02:43.210 --> 00:02:45.509 so let's get into act one of this really extraordinary 00:02:45.509 --> 00:02:49.030 life. It all starts with this intense and often 00:02:49.030 --> 00:02:52.150 very difficult. upbringing. He was born in Columbia, 00:02:52.210 --> 00:02:54.389 Missouri, but his intellectual roots were deep 00:02:54.389 --> 00:02:56.810 in European Jewish scholarship. And that background 00:02:56.810 --> 00:02:59.090 is so crucial to understanding the man he became. 00:02:59.389 --> 00:03:02.069 His parents were Jewish immigrants, his father, 00:03:02.150 --> 00:03:05.310 Leo Wiener from Lithuania, and his mother, Bertha 00:03:05.310 --> 00:03:08.169 Kahn from Germany. And there was a serious scholarly 00:03:08.169 --> 00:03:10.469 lineage there. Right. A very distinguished one 00:03:10.469 --> 00:03:13.469 on his father's side. The sources mention a possible, 00:03:13.550 --> 00:03:16.789 though less certain, link to Maimonides, the 00:03:16.789 --> 00:03:19.740 great medieval scholar. But Norbert himself was 00:03:19.740 --> 00:03:21.780 pretty sure he was descended from Rabbi Akiva 00:03:21.780 --> 00:03:24.419 Eger, a major rabbinical authority in the early 00:03:24.419 --> 00:03:27.919 19th century. So, you know, academic rigor wasn't 00:03:27.919 --> 00:03:29.639 just encouraged. It was basically the family 00:03:29.639 --> 00:03:33.039 business. And his father, Leo, was a force of 00:03:33.039 --> 00:03:35.740 nature in his own right. Absolutely central to 00:03:35.740 --> 00:03:38.020 the story. He was a language teacher at Harvard, 00:03:38.139 --> 00:03:41.159 a famous philologist in German and Slavic languages. 00:03:41.460 --> 00:03:43.560 But he was also highly skilled in mathematics. 00:03:44.039 --> 00:03:46.900 He had the perfect toolkit to shape Norbert's 00:03:46.900 --> 00:03:50.530 education. But maybe. Not the perfect temperament. 00:03:50.610 --> 00:03:53.050 Yeah, the sources describe this incredibly specific, 00:03:53.189 --> 00:03:55.990 intensive homeschooling that Leo set up, lasting 00:03:55.990 --> 00:03:59.069 until 1903. It clearly worked in terms of acceleration, 00:03:59.330 --> 00:04:01.729 but it sounds like it came at a huge cost. It 00:04:01.729 --> 00:04:04.349 was a massive paradox. Norbert was educated almost 00:04:04.349 --> 00:04:07.069 entirely at home using methods Leo just designed 00:04:07.069 --> 00:04:09.430 himself. And Wiener later said his father was 00:04:09.430 --> 00:04:11.949 usually calm, patient, you know, intellectually 00:04:11.949 --> 00:04:14.490 generous. Unless Norbert got an answer wrong. 00:04:15.180 --> 00:04:17.439 In those moments, Leo would just become explosively 00:04:17.439 --> 00:04:20.040 angry. Norbert saw this relentless pressure as 00:04:20.040 --> 00:04:22.639 a form of emotional abuse, believing his father 00:04:22.639 --> 00:04:25.360 wasn't just pushing him to succeed, but to be 00:04:25.360 --> 00:04:28.399 perfect. Wow. That kind of pressure on a child 00:04:28.399 --> 00:04:31.620 prodigy must instill this incredible work ethic, 00:04:31.660 --> 00:04:35.259 but also... Maybe a real fear of failure or intense 00:04:35.259 --> 00:04:37.779 self -criticism. Oh, for sure. And that intellectual 00:04:37.779 --> 00:04:39.879 intensity was paired with a moral and ideological 00:04:39.879 --> 00:04:42.860 one. Leo Wiener had developed this very strict 00:04:42.860 --> 00:04:45.500 Goldstein ethic, you know, nonviolence, simplicity, 00:04:45.800 --> 00:04:48.639 moral purity. And this just permeated the whole 00:04:48.639 --> 00:04:51.759 household. Which led to Norbert's childhood vegetarianism. 00:04:51.860 --> 00:04:54.579 A vegetarianism that was not a choice. It was 00:04:54.579 --> 00:04:57.279 enforced dogma and Norbert deeply resented it. 00:04:57.339 --> 00:04:59.560 He saw it as just another one of his father's 00:04:59.560 --> 00:05:01.759 rigid principles being forced on him. And this 00:05:01.759 --> 00:05:06.000 idea. It was. The sources even say Leo once had 00:05:06.000 --> 00:05:08.560 serious plans to go set up a commune in Central 00:05:08.560 --> 00:05:11.480 America based entirely on these Tolstoyan ideas. 00:05:11.860 --> 00:05:15.259 He never did. But it shows you the mindset, this 00:05:15.259 --> 00:05:18.420 tension, you know, between a genius kid who needs 00:05:18.420 --> 00:05:22.819 intellectual freedom and this rigid, almost puritanical, 00:05:22.819 --> 00:05:25.259 ethical framework from his father that really 00:05:25.259 --> 00:05:27.540 defined Wiener's entire life. And the result 00:05:27.540 --> 00:05:29.600 of this educational pressure cooker, whatever 00:05:29.600 --> 00:05:32.459 the emotional damage, was just pure academic 00:05:32.459 --> 00:05:34.939 velocity. The speed at which he went through 00:05:34.939 --> 00:05:37.259 the university system is almost hard to believe. 00:05:37.379 --> 00:05:39.319 It's dizzying. He graduated from high school 00:05:39.319 --> 00:05:43.100 at age 11 in 1906. Then, just three years later, 00:05:43.220 --> 00:05:46.060 in 1909, he has his Bachelor of Arts in Mathematics 00:05:46.060 --> 00:05:48.660 from Tufts College. He's 14 years old. And he 00:05:48.660 --> 00:05:50.839 doesn't even pause. He just jumps straight into 00:05:50.839 --> 00:05:52.680 grad school, moving from math to philosophy. 00:05:53.149 --> 00:05:55.050 And then back again. Right. He gets his master's 00:05:55.050 --> 00:05:57.569 in philosophy from Cornell at 17. Then he goes 00:05:57.569 --> 00:05:59.829 back to Harvard, earns his Ph .D. in June of 00:05:59.829 --> 00:06:02.410 1913. He was 18 when he finished the work, turned 00:06:02.410 --> 00:06:04.290 19 that fall. I mean, he's one of the youngest 00:06:04.290 --> 00:06:06.170 people ever to get a doctorate from Harvard. 00:06:06.329 --> 00:06:08.850 So let's talk about that doctoral work. It was 00:06:08.850 --> 00:06:11.529 in mathematical logic, comparing people like 00:06:11.529 --> 00:06:14.129 Ernst Schroeder, Whitehead and Bertrand Russell. 00:06:14.519 --> 00:06:16.920 It sounds very abstract, but there was a core 00:06:16.920 --> 00:06:19.019 insight there that really simplified things. 00:06:19.240 --> 00:06:22.160 It was profoundly foundational. At the time, 00:06:22.160 --> 00:06:24.319 mathematicians were trying to build all of math 00:06:24.319 --> 00:06:27.360 from the ground up using set theory. And a key 00:06:27.360 --> 00:06:30.120 piece of that puzzle is the idea of a relation, 00:06:30.560 --> 00:06:33.759 how things connect. The basic building block 00:06:33.759 --> 00:06:36.240 for that is an ordered pair, like a coordinate 00:06:36.240 --> 00:06:39.120 xy, where order obviously matters. Okay, so how 00:06:39.120 --> 00:06:41.639 do they handle that before Wiener? Before Wiener, 00:06:41.699 --> 00:06:44.480 you had to add extra rules or axioms into set 00:06:44.480 --> 00:06:47.100 theory just to make sure that the pair XY was 00:06:47.100 --> 00:06:50.379 different from YX. It felt a bit clunky. Wiener 00:06:50.379 --> 00:06:52.399 was the first to state publicly that you could 00:06:52.399 --> 00:06:55.000 define an ordered pair using only the most basic 00:06:55.000 --> 00:06:57.699 concepts of set theory itself. So he showed you 00:06:57.699 --> 00:06:59.920 could build the idea of order using the Lego 00:06:59.920 --> 00:07:02.339 blocks you already had without needing a special 00:07:02.339 --> 00:07:05.259 new piece. Exactly. It was a breakthrough in 00:07:05.259 --> 00:07:08.620 logical elegance. Now... The slightly cleaner 00:07:08.620 --> 00:07:10.860 definition we use today was proposed a few years 00:07:10.860 --> 00:07:13.879 later by Krutowski, but Wiener gets the credit 00:07:13.879 --> 00:07:16.060 for proving it could be done, for showing that 00:07:16.060 --> 00:07:19.139 ordered pairs didn't need new axioms. And after 00:07:19.139 --> 00:07:22.339 his PhD, he does the classic European tour in 00:07:22.339 --> 00:07:25.920 1914, studying with basically every big name 00:07:25.920 --> 00:07:28.439 in math and philosophy. And this exposure was 00:07:28.439 --> 00:07:30.259 just transformative. I mean, look at the list. 00:07:30.379 --> 00:07:32.399 Bertrand Russell and G .H. Hardy at Cambridge, 00:07:32.720 --> 00:07:35.120 David Hilbert and Edmund Landau in Göttingen, 00:07:35.199 --> 00:07:37.740 which was the absolute center of the mathematical 00:07:37.740 --> 00:07:40.000 universe. And it's so telling that he wasn't 00:07:40.000 --> 00:07:42.740 just studying pure math. No, he was a true polymath. 00:07:43.339 --> 00:07:45.300 While he was there, he also took courses with 00:07:45.300 --> 00:07:47.379 Edmund Husserl, the founder of phenomenology. 00:07:47.759 --> 00:07:50.100 So at the same time he's perfecting these tools 00:07:50.100 --> 00:07:53.120 of logic and analysis, he's also diving deep 00:07:53.120 --> 00:07:55.120 into the philosophical study of consciousness. 00:07:55.680 --> 00:07:58.060 That fascination with the mind would, of course, 00:07:58.079 --> 00:08:00.259 fuel everything he did later with cybernetics. 00:08:00.399 --> 00:08:02.920 But even with all this brilliance, his early 00:08:02.920 --> 00:08:05.600 career was pretty unstable. He really struggled 00:08:05.600 --> 00:08:07.939 to find his footing. He did. He taught philosophy 00:08:07.939 --> 00:08:10.459 for a bit at Harvard, worked as an engineer for 00:08:10.459 --> 00:08:12.680 General Electric, wrote for the Encyclopedia 00:08:12.680 --> 00:08:15.899 Americana. And that stint as a journalist is 00:08:15.899 --> 00:08:19.779 so revealing about his character, you know, echoing 00:08:19.779 --> 00:08:22.680 his father's uncompromising ethics. It really 00:08:22.680 --> 00:08:25.259 is. He was working for the Boston Herald. He 00:08:25.259 --> 00:08:27.519 writes this feature exposing the terrible labor 00:08:27.519 --> 00:08:29.839 conditions for mill workers in Lawrence, Massachusetts. 00:08:30.399 --> 00:08:33.139 And then shortly after, the newspaper owners 00:08:33.139 --> 00:08:35.860 tell him to write some puff pieces for a politician 00:08:35.860 --> 00:08:38.600 they're backing. And he just refused. Blood out 00:08:38.600 --> 00:08:41.000 refused. He cited his principles, said he had 00:08:41.000 --> 00:08:43.860 a moral duty to the truth, and he was fired on 00:08:43.860 --> 00:08:46.139 the spot. You can see the seeds of his later 00:08:46.139 --> 00:08:48.320 rebellion against the military right there. His 00:08:48.320 --> 00:08:51.080 actual military service in World War I was also 00:08:51.080 --> 00:08:53.440 pretty complicated. Yeah, he was eager to serve 00:08:53.440 --> 00:08:55.620 but had terrible eyesight, so he failed to get 00:08:55.620 --> 00:08:58.759 an officer's commission. But by 1918, he's working 00:08:58.759 --> 00:09:01.399 as a civilian on ballistics at the Aberdeen Proving 00:09:01.399 --> 00:09:04.700 Ground. This was his first real taste of applying 00:09:04.700 --> 00:09:07.480 advanced math to these complex, real -world dynamic 00:09:07.480 --> 00:09:10.399 systems. But he still felt he needed to serve 00:09:10.399 --> 00:09:13.720 in uniform. He did. And his moral code insisted 00:09:13.720 --> 00:09:16.470 on it. When he finally enlisted as a common soldier, 00:09:16.769 --> 00:09:19.110 he wrote home saying he'd consider myself a pretty 00:09:19.110 --> 00:09:21.250 cheap kind of a swine if I were willing to be 00:09:21.250 --> 00:09:24.070 an officer but unwilling to be a soldier. Wow. 00:09:24.210 --> 00:09:26.529 It's that Tolstoyian influence again, that belief 00:09:26.529 --> 00:09:29.889 in ethical equality. He was discharged in early 00:09:29.889 --> 00:09:33.389 1919 after the war ended. So this whole turbulent 00:09:33.389 --> 00:09:36.070 early period, it finally leads him to a place 00:09:36.070 --> 00:09:39.309 of professional stability at MIT. And that became 00:09:39.309 --> 00:09:41.809 the launch pad for, well, Everything that changed 00:09:41.809 --> 00:09:44.370 the world in the 1940s. But MIT wasn't his first 00:09:44.370 --> 00:09:46.850 choice. He really wanted to be at Harvard. That's 00:09:46.850 --> 00:09:49.669 a critical piece of the story. Wiener was convinced 00:09:49.669 --> 00:09:51.990 that his failure to get a permanent job at Harvard 00:09:51.990 --> 00:09:54.649 was at least partly due to anti -Semitism. And 00:09:54.649 --> 00:09:56.769 he specifically pointed the finger at a Harvard 00:09:56.769 --> 00:10:00.110 mathematician named G .D. Burkhoff. So a mentor 00:10:00.110 --> 00:10:02.830 suggested MIT and he took an instructor job there. 00:10:02.929 --> 00:10:04.649 He ended up staying for the rest of his career, 00:10:04.809 --> 00:10:07.110 eventually becoming a full professor. And while 00:10:07.110 --> 00:10:09.730 his mind was, you know, clearly on another level, 00:10:09.789 --> 00:10:12.309 his physical presence was often a bit chaotic. 00:10:12.549 --> 00:10:14.809 He became the legendary absent -minded professor. 00:10:15.210 --> 00:10:18.269 Oh, he was a total character, muttering equations 00:10:18.269 --> 00:10:20.269 as he walked down the hall, forgetting where 00:10:20.269 --> 00:10:23.100 he was going. The most famous story, which is 00:10:23.100 --> 00:10:24.620 probably a little one built. With his goddard. 00:10:24.720 --> 00:10:26.980 That's the one. He supposedly came home to an 00:10:26.980 --> 00:10:29.340 empty house after they'd moved, asked a little 00:10:29.340 --> 00:10:31.220 girl outside where his family went, and she replied, 00:10:31.519 --> 00:10:34.129 It's okay, Daddy. Mommy sent me to get you. And 00:10:34.129 --> 00:10:36.169 his daughter later said the moving part was true, 00:10:36.370 --> 00:10:38.590 but he hadn't actually forgotten who she was. 00:10:38.629 --> 00:10:41.929 He was just lost. Exactly. He was just so intensely 00:10:41.929 --> 00:10:44.750 focused on whatever problem was in his head that 00:10:44.750 --> 00:10:48.190 the real world kind of faded away. And in 1941, 00:10:48.289 --> 00:10:51.549 that intense focus was suddenly yanked back toward 00:10:51.549 --> 00:10:54.210 military problems with the start of World War 00:10:54.210 --> 00:10:57.350 II. Right. And despite his pacifist leanings 00:10:57.350 --> 00:10:59.990 from the first war, he accepts an appointment 00:10:59.990 --> 00:11:01.830 with the National Defense Research Committee, 00:11:01.990 --> 00:11:05.460 the NDRC. under Vannevar Bush. And his assignment 00:11:05.460 --> 00:11:08.399 was fire control. Which was this incredibly complex, 00:11:08.559 --> 00:11:12.100 dynamic, and deadly problem. How do you automatically 00:11:12.100 --> 00:11:15.080 aim an anti -aircraft gun to hit a fast -moving 00:11:15.080 --> 00:11:17.379 bomber? You can't just aim where it is. You have 00:11:17.379 --> 00:11:19.440 to predict where it's going to be. And you have 00:11:19.440 --> 00:11:22.120 to do that based on this noisy, intermittent 00:11:22.120 --> 00:11:25.379 data from very early radar systems. And that 00:11:25.379 --> 00:11:28.639 necessity, the need for prediction in the face 00:11:28.639 --> 00:11:32.059 of noise, led directly to his greatest invention. 00:11:32.840 --> 00:11:34.940 the wiener filter okay let's spend some real 00:11:34.940 --> 00:11:36.779 time on the wiener filter this is the gateway 00:11:36.779 --> 00:11:39.100 to cybernetics it was classified published in 00:11:39.100 --> 00:11:41.840 1942. and it was an immediate hit even if it 00:11:41.840 --> 00:11:43.960 was so dense that it got the nickname the yellow 00:11:43.960 --> 00:11:46.019 peril because of the color of its classified 00:11:46.019 --> 00:11:49.460 cover and how hard the math was so functionally 00:11:49.460 --> 00:11:52.480 what is it doing It's a statistical tool. It 00:11:52.480 --> 00:11:55.340 lets you take a messy, noisy, signal -like radar 00:11:55.340 --> 00:11:57.759 static mixed with the actual reflection of a 00:11:57.759 --> 00:12:00.759 plane and produce the best possible guess of 00:12:00.759 --> 00:12:03.600 what the clean, desired signal actually is. And 00:12:03.600 --> 00:12:05.539 how does it get that best possible guess? It's 00:12:05.539 --> 00:12:07.299 not just averaging, right? Oh, no, it's way smarter. 00:12:07.759 --> 00:12:10.100 The filter minimizes the error between its output 00:12:10.100 --> 00:12:12.840 and the real signal you wish you had. And the 00:12:12.840 --> 00:12:14.980 genius of it is that it does this by understanding 00:12:14.980 --> 00:12:17.820 the statistical nature of the noise itself. You're 00:12:17.820 --> 00:12:19.960 not just smoothing things out. You're mathematically 00:12:19.960 --> 00:12:22.120 describing the randomness so you can predict 00:12:22.120 --> 00:12:24.320 what the signal should be. So it's a proactive 00:12:24.320 --> 00:12:26.879 predictive engine. And it was developed at MIT's 00:12:26.879 --> 00:12:29.820 Rad Lab specifically for that German bomber problem. 00:12:30.039 --> 00:12:32.710 Right. But the minute he solved it, he saw the 00:12:32.710 --> 00:12:35.769 profound generality of the math. This wasn't 00:12:35.769 --> 00:12:38.509 just a gadget for a gun. This was a theory for 00:12:38.509 --> 00:12:41.669 predicting the future as best you can based on 00:12:41.669 --> 00:12:44.210 incomplete, noisy information about the past. 00:12:44.970 --> 00:12:47.690 prediction, he realized, is just an optimal form 00:12:47.690 --> 00:12:49.710 of filtering. And this research put him on a 00:12:49.710 --> 00:12:52.210 path parallel to Claude Shannon, who was basically 00:12:52.210 --> 00:12:54.429 inventing information theory at the same time. 00:12:54.549 --> 00:12:57.090 Yes, their approaches were different, but equally 00:12:57.090 --> 00:13:00.450 foundational. Wiener pioneered this idea of modeling 00:13:00.450 --> 00:13:02.570 an information source like the bombers movement 00:13:02.570 --> 00:13:05.769 as a specific kind of random process, a type 00:13:05.769 --> 00:13:08.250 of noise. If you can define the randomness, you 00:13:08.250 --> 00:13:10.629 can build a control system to handle it. That 00:13:10.629 --> 00:13:12.929 was the leap. And this leap took him from guns 00:13:12.929 --> 00:13:15.750 to biology. He started to see that the same feedback 00:13:15.750 --> 00:13:17.990 that aimed the gun was at work when, you know, 00:13:17.990 --> 00:13:20.029 we reach for a cup of coffee. That's the birth 00:13:20.029 --> 00:13:22.610 of cybernetics right there. It led to this pivotal 00:13:22.610 --> 00:13:26.409 1943 paper, Behavior, Purpose, and Teleology, 00:13:26.629 --> 00:13:29.830 that he wrote with the physiologist Arturo Rosenbluth 00:13:29.830 --> 00:13:33.049 and his assistant, Julian Bigelow. And this paper 00:13:33.049 --> 00:13:35.649 was the first to explicitly say that goal -seeking 00:13:35.649 --> 00:13:38.289 behavior purpose in both living things and machines 00:13:38.289 --> 00:13:41.590 comes from self -regulating feedback. The analogy 00:13:41.590 --> 00:13:44.009 is perfect. When you reach for a cup, your nervous 00:13:44.009 --> 00:13:46.009 system is constantly getting feedback on your 00:13:46.009 --> 00:13:48.690 hand's position and sending control signals to 00:13:48.690 --> 00:13:50.769 adjust the muscles until you get there. And that's 00:13:50.769 --> 00:13:53.690 the exact same logic as the anti -aircraft gun 00:13:53.690 --> 00:13:56.190 correcting its aim based on radar feedback. It's 00:13:56.190 --> 00:13:59.230 the same fundamental process. And that all culminated 00:13:59.230 --> 00:14:02.570 in his book Cybernetics in 1948. He coined the 00:14:02.570 --> 00:14:05.049 term from the Greek word for steersman or governor. 00:14:05.289 --> 00:14:07.269 And the definition he gave is still the one we 00:14:07.269 --> 00:14:09.570 use. The science of communication as it relates 00:14:09.570 --> 00:14:13.309 to living things. And he argued really forcefully 00:14:13.309 --> 00:14:15.789 that all intelligent behavior is just the result 00:14:15.789 --> 00:14:18.129 of these feedback mechanisms, which could be 00:14:18.129 --> 00:14:21.129 quantified and, crucially, simulated by machines. 00:14:21.590 --> 00:14:24.169 This idea was an atom bomb in the intellectual 00:14:24.169 --> 00:14:26.830 world. It just exploded, influencing engineering, 00:14:27.190 --> 00:14:30.049 computer science, biology, even how we think 00:14:30.049 --> 00:14:32.090 about societies. It's so interesting that he 00:14:32.090 --> 00:14:34.509 wrote a lot of this foundational book while he 00:14:34.509 --> 00:14:37.289 was in Mexico. He did. He started writing the 00:14:37.289 --> 00:14:39.350 bulk of cybernetics while he was in residence. 00:14:39.720 --> 00:14:42.360 Mexico's National Institute of Cardiology, starting 00:14:42.360 --> 00:14:45.559 in 1945. And that's important because it shows 00:14:45.559 --> 00:14:48.059 his focus was already shifting away from military 00:14:48.059 --> 00:14:50.379 applications and toward biology and medicine. 00:14:50.639 --> 00:14:53.440 He was captivated by the universal nature of 00:14:53.440 --> 00:14:55.539 the math. Okay, so to really get cybernetics, 00:14:55.639 --> 00:14:58.100 we have to appreciate the mathematical tools 00:14:58.100 --> 00:15:01.139 Wiener built. He was a pioneer in studying stochastic 00:15:01.139 --> 00:15:03.559 processes, basically the math of systems that 00:15:03.559 --> 00:15:06.139 change randomly over time. And that field is 00:15:06.139 --> 00:15:08.639 so critical for the 20th century, which was all 00:15:08.639 --> 00:15:11.480 about... complex systems, from electronics to 00:15:11.480 --> 00:15:14.360 finance. Wiener's big contribution was that he 00:15:14.360 --> 00:15:16.980 didn't treat randomness as the enemy to be eliminated. 00:15:17.299 --> 00:15:20.159 He treated it as a quantifiable part of the system 00:15:20.159 --> 00:15:22.240 itself. Which brings us to that famous quote 00:15:22.240 --> 00:15:24.220 from cybernetics that you still see everywhere 00:15:24.220 --> 00:15:26.399 in data science. Information is information, 00:15:26.740 --> 00:15:29.929 not matter or energy. I mean, what a revolutionary 00:15:29.929 --> 00:15:32.769 statement. It says that information is an abstract 00:15:32.769 --> 00:15:35.169 thing separate from whatever is carrying it, 00:15:35.210 --> 00:15:38.169 the wire, the radio waves. It's a quantifiable 00:15:38.169 --> 00:15:41.149 entity defined by its reduction of uncertainty, 00:15:41.330 --> 00:15:44.850 and its flow is the key to control. And the most 00:15:44.850 --> 00:15:47.370 famous mathematical outcome of this work is the 00:15:47.370 --> 00:15:49.830 Wiener process. which is the mathematical model 00:15:49.830 --> 00:15:51.929 for Brownian motion. Right, Brownian motion, 00:15:52.029 --> 00:15:54.470 that jittery random dance of a particle in a 00:15:54.470 --> 00:15:57.169 fluid getting bumped around by molecules. Wiener 00:15:57.169 --> 00:15:59.870 was the first to give this physical thing a truly 00:15:59.870 --> 00:16:02.370 rigorous mathematical description. And what was 00:16:02.370 --> 00:16:05.220 his key insight about that random path? He proved 00:16:05.220 --> 00:16:08.539 that the paths of Brownian motion are non -differentiable. 00:16:08.759 --> 00:16:11.720 Now that sounds super technical, but the idea 00:16:11.720 --> 00:16:15.220 is huge. If a path is differentiable, it's smooth. 00:16:15.480 --> 00:16:17.840 You can define a clear velocity at any point. 00:16:18.240 --> 00:16:20.440 Wiener showed that the path of a Brownian particle 00:16:20.440 --> 00:16:23.799 is so chaotic, so jittery at every scale, that 00:16:23.799 --> 00:16:26.480 its velocity is changing infinitely often. It's 00:16:26.480 --> 00:16:29.289 mathematically rough. So he formalized the idea 00:16:29.289 --> 00:16:32.529 that true randomness isn't smooth. It's fundamentally 00:16:32.529 --> 00:16:35.710 erratic. Exactly. And that model, the Wiener 00:16:35.710 --> 00:16:38.610 process, is essential everywhere. In physics, 00:16:38.789 --> 00:16:41.929 it explains diffusion. In finance, it's the basis 00:16:41.929 --> 00:16:44.529 for modeling the unpredictable fluctuations of 00:16:44.529 --> 00:16:46.389 the stock market. He even came up with a great 00:16:46.389 --> 00:16:48.809 name for a piece of this concept, the Wiener 00:16:48.809 --> 00:16:51.129 sausage. You gotta love the name. It's a brilliant 00:16:51.129 --> 00:16:53.529 visualization. Imagine a particle moving randomly 00:16:53.529 --> 00:16:55.950 and you wrap a little cylinder of a fixed radius 00:16:55.950 --> 00:16:58.679 around the path it traces out. That tube is the 00:16:58.679 --> 00:17:01.480 Wiener sausage. It's a mathematical tool to help 00:17:01.480 --> 00:17:03.700 calculate the volume of space a random particle 00:17:03.700 --> 00:17:06.400 explores over time. Okay, moving beyond Brownian 00:17:06.400 --> 00:17:09.220 motion, he had these core theorems that are still 00:17:09.220 --> 00:17:11.779 used everywhere in signal analysis. Let's start 00:17:11.779 --> 00:17:14.099 with the Wiener -Kinchen theorem. This one is 00:17:14.099 --> 00:17:16.519 absolutely foundational for any electrical engineer. 00:17:17.049 --> 00:17:18.890 It connects two different ways of looking at 00:17:18.890 --> 00:17:21.769 a random signal. First, you have the power spectral 00:17:21.769 --> 00:17:24.390 density. That tells you how the signal's energy 00:17:24.390 --> 00:17:26.630 is spread out across different frequencies. Like 00:17:26.630 --> 00:17:30.170 the EQ on a stereo. Sort of, yeah. And the second 00:17:30.170 --> 00:17:33.190 part is the autocorrelation function. That measures 00:17:33.190 --> 00:17:35.809 how much a signal at one point in time is related 00:17:35.809 --> 00:17:37.730 to itself a little bit later. It's like the signal's 00:17:37.730 --> 00:17:40.670 memory. The theorem says that for a certain class 00:17:40.670 --> 00:17:43.269 of stable signals, these two things, the power 00:17:43.269 --> 00:17:45.809 spectrum and the autocorrelation, are just Fourier 00:17:45.809 --> 00:17:48.309 transforms. forms of each other. And why is that 00:17:48.309 --> 00:17:50.529 relationship so important for an engineer? Because 00:17:50.529 --> 00:17:52.589 it means you can completely describe the nature 00:17:52.589 --> 00:17:55.250 of a signal, including its noise, whether you 00:17:55.250 --> 00:17:57.509 analyze it in the time domain or the frequency 00:17:57.509 --> 00:18:00.009 domain. It gives you the mathematical certainty 00:18:00.009 --> 00:18:02.329 you need to characterize complex noise and then 00:18:02.329 --> 00:18:05.170 crucially build the perfect Wiener filter to 00:18:05.170 --> 00:18:08.029 get rid of it. Okay, next up, a big one from 00:18:08.029 --> 00:18:11.710 pure math. Wiener's Tolberian Theorem from 1932. 00:18:12.450 --> 00:18:14.930 The sources say this was a real paradigm shift. 00:18:15.319 --> 00:18:18.500 It really was. Before Wiener, Tauberian theorems 00:18:18.500 --> 00:18:21.880 were this collection of really specific, difficult 00:18:21.880 --> 00:18:25.000 results about infinite series. It was very piecemeal 00:18:25.000 --> 00:18:27.599 work. Wiener's genius was realizing that all 00:18:27.599 --> 00:18:29.619 these separate results were just special cases 00:18:29.619 --> 00:18:34.140 of one single huge principle from harmonic analysis, 00:18:34.519 --> 00:18:37.380 the study of waves and frequencies. So instead 00:18:37.380 --> 00:18:40.279 of solving a dozen little puzzles, he basically 00:18:40.279 --> 00:18:43.640 created one master key that opened all the locks. 00:18:44.170 --> 00:18:46.150 That's a perfect way to put it. He lifted the 00:18:46.150 --> 00:18:48.609 problem out of the narrow world of series and 00:18:48.609 --> 00:18:51.130 put it into the much more powerful world of functional 00:18:51.130 --> 00:18:54.549 analysis and Banach algebras. It was a major 00:18:54.549 --> 00:18:56.650 shift in mathematical thinking, from computation 00:18:56.650 --> 00:18:59.089 to abstraction and structure. And that leads 00:18:59.089 --> 00:19:01.069 to the Paley -Wiener theorem. Right, which he 00:19:01.069 --> 00:19:03.410 co -authored with Raymond Paley. This is another 00:19:03.410 --> 00:19:05.990 cornerstone of harmonic analysis. It creates 00:19:05.990 --> 00:19:08.190 this beautiful link between how fast a function 00:19:08.190 --> 00:19:10.250 can grow and what its frequency content looks 00:19:10.250 --> 00:19:12.670 like after a Fourier transform. What's the takeaway 00:19:12.670 --> 00:19:15.119 for a non -specialist? The core idea is about 00:19:15.119 --> 00:19:17.859 limits. It tells you that if a signal is confined 00:19:17.859 --> 00:19:21.519 in time like it only exists for one second, then 00:19:21.519 --> 00:19:23.700 its frequency content can't also be confined. 00:19:23.960 --> 00:19:26.019 It has to spread out. It's a fundamental tradeoff 00:19:26.019 --> 00:19:28.240 in physics and signal processing. And it seems 00:19:28.240 --> 00:19:30.200 like he was developing some of these ideas at 00:19:30.200 --> 00:19:32.500 the same time as other great minds. Absolutely. 00:19:32.759 --> 00:19:35.819 The notion of a Banach space, this key structure 00:19:35.819 --> 00:19:38.619 in modern analysis, was independently discovered 00:19:38.619 --> 00:19:41.660 by both Wiener and the Polish mathematician Stefan 00:19:41.660 --> 00:19:43.769 Banach at almost the exact same time. same time 00:19:43.769 --> 00:19:46.890 around 1923 finally let's connect all this abstract 00:19:46.890 --> 00:19:49.430 math back to the machines through his work on 00:19:49.430 --> 00:19:52.470 nonlinear control theory right because most simple 00:19:52.470 --> 00:19:54.930 control theory assumes systems are linear you 00:19:54.930 --> 00:19:56.829 put twice the input in you get twice the output 00:19:56.829 --> 00:20:00.660 but the real world is messy and nonlinear So 00:20:00.660 --> 00:20:03.140 how did he tackle that chaos? He used something 00:20:03.140 --> 00:20:05.900 called polynomial chaos based on the Hermite 00:20:05.900 --> 00:20:08.420 -Laguerre expansion. The method is actually quite 00:20:08.420 --> 00:20:10.960 elegant. You take an unknown nonlinear system 00:20:10.960 --> 00:20:13.779 and you input the simplest possible random signal 00:20:13.779 --> 00:20:16.500 -like noise. Then, by analyzing the system's 00:20:16.500 --> 00:20:19.000 output using this expansion, you can create a 00:20:19.000 --> 00:20:22.119 mathematical map of its nonlinear behavior. And 00:20:22.119 --> 00:20:23.880 once you have the map... Once you have the map, 00:20:24.000 --> 00:20:26.539 you can control it. It was a massive step toward 00:20:26.539 --> 00:20:29.319 creating automation that could work robustly 00:20:29.319 --> 00:20:32.720 in the real unpredictable world. So we've established 00:20:32.720 --> 00:20:34.759 that Wiener had this incredible mathematical 00:20:34.759 --> 00:20:38.059 toolkit for controlling systems. But as the machine 00:20:38.059 --> 00:20:41.920 age really began, he became maybe the first major 00:20:41.920 --> 00:20:44.799 scientist to publicly and dramatically wrestle 00:20:44.799 --> 00:20:47.240 with the moral consequences of his own inventions. 00:20:47.559 --> 00:20:49.420 This is what you could call the scientist rebels 00:20:49.420 --> 00:20:52.039 phase of his career. He had helped build weapons 00:20:52.039 --> 00:20:54.279 during. the war. And afterwards, he had this 00:20:54.279 --> 00:20:57.759 profound ethical crisis. He became deeply, deeply 00:20:57.759 --> 00:20:59.740 concerned with the militarization of science 00:20:59.740 --> 00:21:01.859 and the political interference that came with 00:21:01.859 --> 00:21:04.000 all that defense money. And this wasn't just 00:21:04.000 --> 00:21:06.119 a private concern. He took this radical public 00:21:06.119 --> 00:21:09.240 stance. He refused all government funding and 00:21:09.240 --> 00:21:11.559 refused to work on any military projects after 00:21:11.559 --> 00:21:13.759 the war. That must have made him a pariah in 00:21:13.759 --> 00:21:16.769 some circles. It really did. He became an iconoclast. 00:21:17.029 --> 00:21:20.190 He laid out his case in this famous scathing 00:21:20.190 --> 00:21:23.309 article in the Atlantic Monthly in 1947 called 00:21:23.309 --> 00:21:27.029 A Scientist's Rebels. It was a call to arms telling 00:21:27.029 --> 00:21:29.730 scientists they had a moral agency, that they 00:21:29.730 --> 00:21:31.630 had to think about how their work could be abused 00:21:31.630 --> 00:21:33.930 before they took the money. And this ethical 00:21:33.930 --> 00:21:36.630 position puts him in direct philosophical conflict 00:21:36.630 --> 00:21:39.609 with other giants of the time, especially John 00:21:39.609 --> 00:21:42.170 von Neumann. This contrast is so fundamental 00:21:42.170 --> 00:21:45.470 to AI ethics today. For Wiener, technology was 00:21:45.470 --> 00:21:48.210 inherently moral. The creator had a responsibility 00:21:48.210 --> 00:21:51.269 to guide its use and even withhold it if necessary. 00:21:51.609 --> 00:21:53.589 Whereas von Neumann's view was totally different. 00:21:53.789 --> 00:21:56.029 Completely. Von Neumann believed technology was 00:21:56.029 --> 00:21:58.569 ethically neutral. For him, the scientist's job 00:21:58.569 --> 00:22:01.029 was just to see what was possible. If the military 00:22:01.029 --> 00:22:03.150 wanted to use it, whether it was a computer or 00:22:03.150 --> 00:22:05.250 an atomic bomb, that was a political problem, 00:22:05.329 --> 00:22:08.329 not a scientific one. Wiener saw that as dangerously 00:22:08.329 --> 00:22:11.670 irresponsible. saw it as professional objectivity. 00:22:11.829 --> 00:22:14.289 While he was fighting this big external battle, 00:22:14.569 --> 00:22:16.890 he was also dealing with a really devastating 00:22:16.890 --> 00:22:19.690 internal conflict that basically destroyed his 00:22:19.690 --> 00:22:22.170 own cognitive science team at MIT. This is one 00:22:22.170 --> 00:22:24.490 of the really heartbreaking stories in the history 00:22:24.490 --> 00:22:27.930 of AI. After the war, Wiener helps recruit this 00:22:27.930 --> 00:22:31.269 brilliant team, Warren McCulloch, the young prodigy 00:22:31.269 --> 00:22:33.990 Walter Pitts, to work on the connection between 00:22:33.990 --> 00:22:36.250 the brain and computers. But then the collaboration 00:22:36.250 --> 00:22:40.539 just shatters. Almost overnight. Wiener suddenly 00:22:40.539 --> 00:22:43.380 and completely cut off all contact with them. 00:22:43.460 --> 00:22:45.880 It was emotionally devastating, especially for 00:22:45.880 --> 00:22:48.619 Walter Pitts, and it basically derailed his career. 00:22:48.819 --> 00:22:50.599 And the sources have a couple of different theories 00:22:50.599 --> 00:22:52.359 for why this happened, right? Two main ones. 00:22:52.500 --> 00:22:55.500 The first is domestic. Apparently, Wiener's wife, 00:22:55.680 --> 00:22:58.559 Margaret, really disliked McCulloch's bohemian, 00:22:58.559 --> 00:23:01.039 hard -drinking lifestyle and might have pressured 00:23:01.039 --> 00:23:03.339 Wiener to break ties. And the other theory is 00:23:03.339 --> 00:23:06.059 more professional. Yeah. The second theory is 00:23:06.059 --> 00:23:08.680 that Wiener had based a big new theory of the 00:23:08.680 --> 00:23:11.200 nervous system on some physiological facts that 00:23:11.200 --> 00:23:13.359 McCulloch had given him. He presents the theory 00:23:13.359 --> 00:23:15.339 at a conference and gets, you know, publicly 00:23:15.339 --> 00:23:17.740 humiliated by experts who say the facts are wrong. 00:23:17.880 --> 00:23:19.940 He became convinced McCulloch had set him up. 00:23:20.059 --> 00:23:22.940 Given his sensitivity to intellectual criticism, 00:23:23.279 --> 00:23:24.960 which goes all the way back to his father, that's 00:23:24.960 --> 00:23:27.519 also very plausible. Either way, the result was 00:23:27.519 --> 00:23:29.819 that this pioneering research just got stopped 00:23:29.819 --> 00:23:32.440 in its tracks. A huge loss for the field. It 00:23:32.440 --> 00:23:34.480 just shows you how the progress of science is 00:23:34.480 --> 00:23:38.119 often tangled up in these very messy, very human 00:23:38.119 --> 00:23:40.299 relationships. Now, moving to his international 00:23:40.299 --> 00:23:43.640 influence, he was really a global thinker. He 00:23:43.640 --> 00:23:45.759 advised the government of India in the 1950s. 00:23:45.759 --> 00:23:48.500 He was a genuine believer in technology for global 00:23:48.500 --> 00:23:52.460 progress. He strongly pushed India to adopt automation 00:23:52.460 --> 00:23:55.380 as a way to leapfrog economic underdevelopment. 00:23:55.559 --> 00:23:59.140 He saw cybernetics as a tool for equality. And 00:23:59.140 --> 00:24:01.259 his commitment to open science got him into some 00:24:01.259 --> 00:24:03.259 trouble during the Cold War. It absolutely did. 00:24:03.359 --> 00:24:05.200 He believed in sharing knowledge freely, and 00:24:05.200 --> 00:24:07.319 that included collaborating with Soviet researchers. 00:24:07.740 --> 00:24:10.000 In the American political climate of the time, 00:24:10.019 --> 00:24:12.160 that made him an object of serious suspicion. 00:24:12.440 --> 00:24:15.269 But he never wavered. On a more personal note, 00:24:15.390 --> 00:24:17.269 his relationship with his own Jewish heritage 00:24:17.269 --> 00:24:21.069 was complex. Very. Neither he nor his father 00:24:21.069 --> 00:24:23.769 were religious, but he always credited the Jewish 00:24:23.769 --> 00:24:26.190 social and cultural structure with shaping his 00:24:26.190 --> 00:24:29.069 character and his intellectual discipline. He 00:24:29.069 --> 00:24:31.509 saw it as a resilient community, not just a religion. 00:24:31.809 --> 00:24:33.869 And he had this interesting theory, which he 00:24:33.869 --> 00:24:36.690 shared with the geneticist JBS Haldane, about 00:24:36.690 --> 00:24:39.490 the role of scholarship in Jewish history. Yeah, 00:24:39.569 --> 00:24:42.130 they speculated that because Jewish culture historically 00:24:42.130 --> 00:24:45.390 valued scholars so highly, those scholars were 00:24:45.390 --> 00:24:48.150 more likely to have larger families. And over 00:24:48.150 --> 00:24:50.289 generations, this might have led to an amplification 00:24:50.289 --> 00:24:53.670 of certain cultural and, they suggested, even 00:24:53.670 --> 00:24:56.950 biological traits related to intense intellectual 00:24:56.950 --> 00:24:59.769 work. And his view on Zionism was consistent 00:24:59.769 --> 00:25:02.730 with his broader philosophy. It was. He was ambivalent 00:25:02.730 --> 00:25:05.809 about nationalism. For him, Jewish identity was 00:25:05.809 --> 00:25:08.769 this ancient, continuous cultural thing. that 00:25:08.769 --> 00:25:11.309 existed completely independently of any political 00:25:11.309 --> 00:25:13.230 state. We should also mention the health issues 00:25:13.230 --> 00:25:15.710 he talked about in his autobiography, I Am a 00:25:15.710 --> 00:25:18.289 Mathematician. Yes. In a moment of real candor, 00:25:18.309 --> 00:25:21.150 he admitted to regularly using Benzedrine, which 00:25:21.150 --> 00:25:23.730 is an amphetamine, throughout his life, often 00:25:23.730 --> 00:25:26.009 without being fully aware of the dangers. It 00:25:26.009 --> 00:25:27.769 might help explain some of that relentless energy 00:25:27.769 --> 00:25:30.130 and maybe some of the volatility, too. And finally, 00:25:30.190 --> 00:25:33.130 his official legacy. The awards and memorials 00:25:33.130 --> 00:25:35.559 really tell the story of his impact. they do 00:25:35.559 --> 00:25:38.200 you have the national medal of science the bosher 00:25:38.200 --> 00:25:41.579 prize for pure math posthumously his book god 00:25:41.579 --> 00:25:44.099 and golem inc won the national book award the 00:25:44.099 --> 00:25:46.420 titles of the awards themselves show the breadth 00:25:46.420 --> 00:25:48.740 of his thought. Exactly. There's Crater Wiener 00:25:48.740 --> 00:25:50.980 on the Moon and the Norbert Wiener Prize in Applied 00:25:50.980 --> 00:25:53.519 Mathematics. But I think the most telling one 00:25:53.519 --> 00:25:55.839 is the Norbert Wiener Award for Social and Professional 00:25:55.839 --> 00:25:58.619 Responsibility. That award basically enshrined 00:25:58.619 --> 00:26:01.319 his central ethical conflict as a permanent concern 00:26:01.319 --> 00:26:05.099 for all technologists. His key books, Cybernetics, 00:26:05.200 --> 00:26:08.420 The Human Use of Human Beings, and God and Gollum, 00:26:08.440 --> 00:26:11.799 Inc., are still essential reading. He died in 00:26:11.799 --> 00:26:14.359 Stockholm in 1964 from a heart attack. He was 00:26:14.359 --> 00:26:17.309 69. So we've spent this deep dive tracing the 00:26:17.309 --> 00:26:19.569 path of a mind that really did bridge the gap 00:26:19.569 --> 00:26:22.890 between pure math and autonomous machines. As 00:26:22.890 --> 00:26:24.930 we wrap up, what are the core takeaways you, 00:26:25.029 --> 00:26:27.109 the learner, should carry forward about Norbert 00:26:27.109 --> 00:26:29.190 Wiener? I think you have to remember him as the 00:26:29.190 --> 00:26:31.549 architect who made feedback the universal language 00:26:31.549 --> 00:26:34.430 connecting biology and engineering. He gave us 00:26:34.430 --> 00:26:36.450 the framework that shows a nervous system controlling 00:26:36.450 --> 00:26:39.269 a muscle is solving the same basic problem as 00:26:39.269 --> 00:26:41.430 a machine tracking a target. And on the mathematical 00:26:41.430 --> 00:26:45.160 side, the Wiener filter is his monument. It proved 00:26:45.160 --> 00:26:47.519 that prediction is fundamentally an act of optimally 00:26:47.519 --> 00:26:50.380 filtering noise. That's a concept that is still 00:26:50.380 --> 00:26:52.500 central to, I mean, everything from your phone's 00:26:52.500 --> 00:26:55.259 GPS to financial models. But for me, the most 00:26:55.259 --> 00:26:58.240 powerful legacy is his ethical one. He was worried, 00:26:58.380 --> 00:27:01.079 deeply worried, that by creating these intelligent 00:27:01.079 --> 00:27:03.920 machines, these golems, we might lose control 00:27:03.920 --> 00:27:06.539 of them. Or even worse, just hand over our moral 00:27:06.539 --> 00:27:08.680 judgment to them. That's what his final book, 00:27:08.819 --> 00:27:11.579 God and Golem Inc., is all about. We've seen 00:27:11.579 --> 00:27:13.680 that Wiener's brilliance was always driven by 00:27:13.680 --> 00:27:16.640 this deep desire to understand systems, whether 00:27:16.640 --> 00:27:19.200 they were chaotic, like Brownian motion, or predictive, 00:27:19.420 --> 00:27:21.839 like an anti -aircraft gun. His life was this 00:27:21.839 --> 00:27:23.660 constant struggle between technical possibility 00:27:23.660 --> 00:27:27.099 and ethical responsibility. As you look at the 00:27:27.099 --> 00:27:29.500 growth of AI today, Wiener's legacy forces you 00:27:29.500 --> 00:27:31.859 to ask that essential question. If intelligence 00:27:31.859 --> 00:27:34.460 is just feedback loops, what is our responsibility 00:27:34.460 --> 00:27:36.920 in designing those loops? Who decides the goal? 00:27:37.289 --> 00:27:39.529 We are building new golems at a speed he couldn't 00:27:39.529 --> 00:27:41.509 have imagined, but we are still faced with this 00:27:41.509 --> 00:27:43.950 most important question. What is the human use 00:27:43.950 --> 00:27:44.730 of human beings?