WEBVTT - Spotlight on Gordon Moore 0:00:04.360 --> 0:00:12.200 Welcome to Tech Stuff, a production from iHeartRadio. Hey there, 0:00:12.240 --> 0:00:15.760 and welcome to tech Stuff. I'm your host Jonathan Strickland. 0:00:15.800 --> 0:00:19.200 I'm an executive producer with iHeartRadio and how the tech 0:00:19.280 --> 0:00:23.720 are you. In yesterday's news episode, I opened the show 0:00:23.720 --> 0:00:28.400 by talking about the passing of Gordon Moore, an important 0:00:28.480 --> 0:00:32.240 figure in the world of tech today. I thought we'd 0:00:32.560 --> 0:00:35.280 look a little bit into his history and talk about 0:00:35.320 --> 0:00:38.279 the man and where he came from and some of 0:00:38.320 --> 0:00:42.600 his contributions to technology. Along the way, we'll learn why 0:00:42.720 --> 0:00:47.159 he was dubbed traitorous by a former employer, along with 0:00:47.280 --> 0:00:50.200 seven of his co workers. We'll talk a bit about 0:00:50.240 --> 0:00:54.000 Moore's law and what that originally meant versus how we 0:00:54.120 --> 0:00:57.720 typically interpret it today, and of course we'll chat a 0:00:57.760 --> 0:01:01.440 bit about a tiny little company he co found called Intel. 0:01:02.760 --> 0:01:08.080 Gordon Moore was born on January third, nineteen twenty nine. 0:01:09.000 --> 0:01:11.720 That would turn out to be a pretty darn eventful year, 0:01:12.080 --> 0:01:16.480 because in August the world would enter the Great Depression, 0:01:16.840 --> 0:01:23.640 which would affect the economy for a full decade. Fortunately 0:01:23.680 --> 0:01:26.680 for the Moore family, they were not you know, Wall 0:01:26.720 --> 0:01:31.600 Street investors. They didn't lose their shirts in the Great Depression. Now, 0:01:31.640 --> 0:01:35.399 his father was a deputy sheriff of a little town 0:01:35.480 --> 0:01:39.880 called Pescadero, which was on the coast of California, so 0:01:39.959 --> 0:01:43.720 just a little bit south of San Francisco, and his 0:01:43.840 --> 0:01:48.760 mother ran a general store in town, so his family 0:01:48.920 --> 0:01:53.040 was not as directly hit by the depression as many 0:01:53.120 --> 0:01:58.000 others were. Now. Early on, Gordon Moore said that his 0:01:58.160 --> 0:02:02.760 ambition was to become a teacher. He was very interested 0:02:02.800 --> 0:02:07.720 in science, specifically in chemistry, and he enrolled in San 0:02:07.840 --> 0:02:11.280 Jose State College, and that's where he would meet his 0:02:11.320 --> 0:02:14.760 future wife, Betty Whittaker. The two would be married for 0:02:14.919 --> 0:02:19.840 more than seventy years. They actually got married in nineteen fifty. 0:02:20.800 --> 0:02:23.800 That was the same year that Moore had graduated with 0:02:23.880 --> 0:02:28.160 his undergrad degree. He had transferred to the University of California, 0:02:28.200 --> 0:02:32.560 Berkeley and completed his undergrad degree there, and he earned 0:02:32.600 --> 0:02:37.239 his degree in chemistry. He then continued his studies at 0:02:37.280 --> 0:02:42.000 cal Tech. He had postgraduate work and Moore would earn 0:02:42.120 --> 0:02:46.520 a doctorate, also in chemistry. Rather than pursue a career 0:02:46.560 --> 0:02:50.880 in teaching, he applied for a position at Dow Chemical. However, 0:02:50.919 --> 0:02:55.880 accloid Moore during a psychological evaluation for the part of 0:02:55.880 --> 0:03:00.480 the hiring process, a Dow representative determined that More didn't 0:03:00.480 --> 0:03:03.760 have what it takes to be a manager. They said 0:03:03.800 --> 0:03:07.919 that his technical skills were suitable, that he had strong 0:03:08.320 --> 0:03:11.520 skills in the field of science, but that he would 0:03:11.600 --> 0:03:15.440 not make a good manager. Considering what would happen later 0:03:15.520 --> 0:03:17.760 in Moore's life, this assessment should go down as one 0:03:17.800 --> 0:03:21.160 of the dumbest ones in tech history. But hey, once 0:03:21.240 --> 0:03:24.040 upon a time I worked for consultants who would do 0:03:24.120 --> 0:03:28.280 things like leadership assessments, and based upon the stuff that 0:03:28.320 --> 0:03:32.359 I was reading, my own personal opinion is a lot 0:03:32.400 --> 0:03:35.880 of those assessments are based on who he Again, that's 0:03:35.920 --> 0:03:38.200 my own personal opinion, and I could be totally wrong 0:03:38.240 --> 0:03:41.120 about that. I just got very jaded about it. Anyway, 0:03:41.200 --> 0:03:45.320 back to the story. More needed a gig, So then 0:03:45.320 --> 0:03:48.640 he took a position at Johns Hopkins University. But this 0:03:48.680 --> 0:03:53.560 is in Baltimore. He joined their applied physics Laboratory. But 0:03:53.920 --> 0:03:57.080 you know, Baltimore is in Maryland. That's on the opposite 0:03:57.160 --> 0:04:00.160 side of the United States from California. So he was 0:04:01.120 --> 0:04:04.520 about as far away as he could get from his 0:04:04.560 --> 0:04:09.960 home while still being in the United States. And Moore's 0:04:09.960 --> 0:04:12.960 extended family still lived out in California, so he didn't 0:04:12.960 --> 0:04:16.680 really like working on the East Coast. So even though 0:04:16.680 --> 0:04:19.240 he had steady work, he kept looking for opportunities that 0:04:19.279 --> 0:04:21.880 would let him move back across the United States to 0:04:21.920 --> 0:04:25.880 the West Coast, and one such opportunity arose at the 0:04:26.080 --> 0:04:32.080 Lawrence Livermore Laboratory. Now this lab had Ernest Lawrence and 0:04:32.320 --> 0:04:37.000 Edward Teller as co founders. Now those names might not 0:04:37.120 --> 0:04:40.760 mean anything to you, but Edward Teller in particular is 0:04:41.000 --> 0:04:44.920 also sometimes referred to as the father of the hydrogen bomb. 0:04:45.560 --> 0:04:49.279 Though he didn't He didn't like that. He was a 0:04:49.320 --> 0:04:53.919 theoretical physicist and was notoriously difficult to work with. He 0:04:53.960 --> 0:04:59.520 also earned a bad and arguably unjustified reputation after testifying 0:04:59.560 --> 0:05:03.280 in a hear to determine whether or not Robert Oppenheimer 0:05:03.360 --> 0:05:06.440 should have his security clearance revoked. Teller was of the 0:05:06.480 --> 0:05:11.960 opinion that it should, and ultimately it was. Anyway, episode 0:05:12.040 --> 0:05:14.160 is not about Edward Teller. I will have to do 0:05:14.200 --> 0:05:17.320 another one about Teller in the future. I think I've 0:05:17.320 --> 0:05:19.800 talked about him in past episodes. I'm certain I have, 0:05:19.880 --> 0:05:23.279 and things like episodes about the Manhattan Project. But the 0:05:23.400 --> 0:05:27.520 Livermore Lab was known primarily for work involving the research 0:05:27.560 --> 0:05:31.760 and development of nuclear weapons. So Gordon Moore had this 0:05:31.800 --> 0:05:34.760 opportunity to go and work for the Livermore Lab, and 0:05:35.680 --> 0:05:39.000 he thought about this and ultimately decided that he was 0:05:39.160 --> 0:05:43.400 not keen on the idea of working in the field 0:05:43.520 --> 0:05:48.640 of nuclear weapons research and development. And you know, even 0:05:48.680 --> 0:05:52.320 though he had the skills in chemistry and he had 0:05:52.360 --> 0:05:55.039 a good understanding of physics, he just didn't feel right 0:05:55.080 --> 0:05:59.640 about dedicating his expertise toward making things go boom. But 0:06:00.040 --> 0:06:03.680 nineteen fifty six he got a different opportunity with a 0:06:03.720 --> 0:06:09.360 brand new startup division a company that was called Shockley Semiconductor. 0:06:10.080 --> 0:06:14.320 This was a division of another company called Beckman Instruments Incorporated. 0:06:14.839 --> 0:06:19.240 So the founder of the parent company, Arnold O. Beckman, 0:06:20.040 --> 0:06:23.080 was a chemist and inventor who saw a potential in 0:06:23.120 --> 0:06:26.359 a brand new technology, one that was developed under the 0:06:26.400 --> 0:06:31.320 management of William Shockley when he was working at Bell Labs. 0:06:32.040 --> 0:06:35.160 This would be in the very late nineteen forties. Now 0:06:35.200 --> 0:06:38.200 I've talked about William Shockley in old episodes of tech Stuff. 0:06:38.520 --> 0:06:41.679 He led a solid state physics group in Bell Labs, 0:06:41.720 --> 0:06:47.080 and under his watch that lab produced the world's first transistor. 0:06:48.480 --> 0:06:54.360 It is impossible for me to overstate how important the 0:06:54.400 --> 0:06:58.960 transistor is, how important it was, and how important it 0:06:59.000 --> 0:07:04.239 continues to be. It really is the basis for all 0:07:04.400 --> 0:07:07.920 of electronics moving forward from that point. So to call 0:07:07.960 --> 0:07:12.200 it a pivotal moment is being, you know, kind of 0:07:13.120 --> 0:07:17.040 is kind of downplaying it. Really, Transistors can do all 0:07:17.080 --> 0:07:20.280 sorts of stuff that's useful in circuits, but like the 0:07:20.320 --> 0:07:24.360 basic ones, or things like being a way to direct 0:07:25.720 --> 0:07:28.760 electricity so that it flows along a specific path, or 0:07:29.000 --> 0:07:32.440 to act as a switch to stop electricity from passing 0:07:32.480 --> 0:07:35.360 through certain places, to being able to amplify a weak 0:07:35.440 --> 0:07:40.880 signal into a stronger signal. Transistors would replace things that 0:07:40.920 --> 0:07:45.080 were much larger components like vacuum tubes. So this meant 0:07:45.120 --> 0:07:50.240 that the transistor would make it possible to miniaturize components 0:07:50.320 --> 0:07:55.600 and create integrated electronics. So this is what allowed for 0:07:55.640 --> 0:07:58.000 the era of miniaturization. When you hear things like a 0:07:58.040 --> 0:08:02.920 transistor radio that was based upon this technology, and it 0:08:03.000 --> 0:08:04.960 meant that a radio no longer had to be like 0:08:05.000 --> 0:08:08.600 a piece of furniture that was large and had to 0:08:08.600 --> 0:08:11.000 sit either on top of a table or heck, it 0:08:11.080 --> 0:08:13.920 might have been its own cabinet in the old radio days, 0:08:14.120 --> 0:08:17.160 to something that you could fit in your pocket. That 0:08:17.200 --> 0:08:20.840 would be what transistors would make possible, and it would 0:08:20.880 --> 0:08:26.720 transform the technological world. So William Shockley his team developed 0:08:26.880 --> 0:08:31.400 the very first transistor, and his research served as sort 0:08:31.400 --> 0:08:36.000 of the foundation for this work. But it was actually 0:08:36.040 --> 0:08:40.040 two of his lab text two of his employees who 0:08:40.440 --> 0:08:44.680 did the actual building. It was John Bardeen and Walter Bretagne, 0:08:45.240 --> 0:08:48.480 two of Shockley's team who created the first working transistor. 0:08:49.559 --> 0:08:52.800 There was a lot of animosity between Shockley and his team, 0:08:53.400 --> 0:08:57.360 with disagreements regarding who should get credit for the invention 0:08:57.360 --> 0:08:59.800 and that kind of thing. Shockley was always upset whenever 0:09:00.600 --> 0:09:03.720 there were patents being filed for technologies that came out 0:09:03.720 --> 0:09:05.839 of his lab that didn't have his name on it, 0:09:06.120 --> 0:09:08.560 and then his team members would often bristle at that 0:09:08.760 --> 0:09:12.560 saying like, yeah, but you didn't do it, so your 0:09:12.640 --> 0:09:15.600 name shouldn't be on it. And it just turned out 0:09:15.640 --> 0:09:21.000 that Shockley, like Edward Teller, was also notoriously difficult to 0:09:21.000 --> 0:09:27.800 work with anyway. Putting aside his managerial limitations, Shockley had 0:09:27.840 --> 0:09:34.000 received some support from Beckman to create Shockley Semiconductor Beckman 0:09:34.160 --> 0:09:38.240 recognized that the semiconductor industry was going to be the 0:09:38.280 --> 0:09:40.480 next really big thing, and he wanted to make sure 0:09:41.160 --> 0:09:44.079 that he had a handle on it. So he felt 0:09:44.120 --> 0:09:47.840 that it was important to have Shockley be able to 0:09:47.920 --> 0:09:53.000 establish his own company. Although he had some reservations about 0:09:53.040 --> 0:09:56.680 Shockley's ability to lead people, it turns out those reservations 0:09:56.679 --> 0:09:59.679 were well founded. One thing Shockley did have, though, was 0:09:59.720 --> 0:10:02.400 a really good eye for talent. He started to recruit 0:10:02.679 --> 0:10:06.760 very bright prospects. Folks fresh out of postgraduate studies who 0:10:06.760 --> 0:10:09.640 were on the cutting edge of scientific research were among 0:10:10.320 --> 0:10:15.000 the top candidates. One person he had his zion was 0:10:15.120 --> 0:10:19.439 Gordon Moore. Gordon Moore was someone that Shockley approached during 0:10:19.440 --> 0:10:25.040 a meeting of the American Physical Society, so he actually 0:10:25.080 --> 0:10:29.520 recruited more in person. In other cases, he reached out 0:10:29.559 --> 0:10:34.720 to various West Coast post graduates, and in a few 0:10:34.800 --> 0:10:38.960 cases he actually got attention by running advertising in the 0:10:39.000 --> 0:10:45.800 Northeast because at the time, the semiconductor transistor world was 0:10:45.880 --> 0:10:49.360 sort of coalescing in the New York area. That's where 0:10:49.360 --> 0:10:54.520 Bell Labs was located. So that was a challenge because 0:10:54.520 --> 0:10:58.319 Shockley was locating his lab on the West coast. So 0:10:58.360 --> 0:11:02.719 to attract more talent, he actually ran essentially classified ads 0:11:02.760 --> 0:11:06.560 in newspapers in the Northeast to attract people. And so 0:11:07.000 --> 0:11:11.120 this initial group came together to become the first team 0:11:11.320 --> 0:11:15.240 at Shockley Semiconductor. All right, I've got more to say 0:11:15.280 --> 0:11:19.400 about what happened over at Shockley and how Moore's career 0:11:19.440 --> 0:11:21.880 would go from there. But before we get to that, 0:11:21.960 --> 0:11:34.920 let's take a quick break. Okay, So Gordon Moore joins 0:11:34.960 --> 0:11:38.520 a team along with a bunch of other notable people 0:11:38.600 --> 0:11:42.760 in technology over at Shockley Semiconductor, and it would not 0:11:42.880 --> 0:11:45.400 take long for things to go sour. More and his 0:11:45.520 --> 0:11:50.640 co workers faced very tough conditions. They didn't always agree 0:11:50.640 --> 0:11:53.760 with each other. For one thing, they had different ideas 0:11:53.800 --> 0:11:58.839 of what the best approach was toward developing semiconductors and transistors, 0:11:59.280 --> 0:12:05.319 so there was some disagreement even within their community of coworkers. 0:12:05.360 --> 0:12:10.400 But more than that, William Shockley was an aggressive, autocratic, 0:12:10.440 --> 0:12:14.959 and paranoid employer. I get the sense that he felt 0:12:14.960 --> 0:12:16.880 the need to be the smartest person in the room. 0:12:17.440 --> 0:12:21.840 But then he also gathered together talent for his pledgling 0:12:21.880 --> 0:12:24.800 company that consisted of some of the brightest minds and 0:12:24.960 --> 0:12:29.760 science and engineering of the day. He reportedly distrusted the 0:12:29.800 --> 0:12:34.400 work that his team created and would send the output 0:12:34.559 --> 0:12:37.080 to contexts that he still had at Bell Labs to 0:12:37.280 --> 0:12:41.480 double check their work. Which if you are, you know, 0:12:41.679 --> 0:12:45.040 like a pre eminent researcher in your field, you are 0:12:45.120 --> 0:12:48.320 going to be insulted if your employer says, well, let 0:12:48.400 --> 0:12:51.679 me just send this off to a competitor to double 0:12:51.760 --> 0:12:55.960 check your numbers. Shockley also insisted on testing each person 0:12:56.000 --> 0:12:59.440 with psychological analysis before hiring them on which you know, 0:13:00.000 --> 0:13:02.760 and Moore had already gone through when he had applied 0:13:02.800 --> 0:13:06.440 for dal Chemical. William Shockley also allegedly planned to force 0:13:06.480 --> 0:13:09.240 his entire team to submit to a lie detector test 0:13:09.280 --> 0:13:12.719 at one point, but because everyone objected to it, he 0:13:13.240 --> 0:13:16.640 let it drop. And I know that I talk a 0:13:16.679 --> 0:13:19.720 lot about demanding bosses in tech, you know, folks like 0:13:19.880 --> 0:13:23.920 Tim Cook or Elon Musk, who would demand that employees 0:13:23.960 --> 0:13:27.840 returned to the office. But honestly, when you look at 0:13:27.880 --> 0:13:30.560 what William Shockley did, they ain't got nothing on him. 0:13:31.040 --> 0:13:34.760 So anyway, Gordon Moore had joined Shockley Semiconductor in nineteen 0:13:34.840 --> 0:13:39.360 fifty six. By mid nineteen fifty seven, things had approached 0:13:39.440 --> 0:13:43.360 a point where they were intolerable. So more, with the 0:13:43.400 --> 0:13:47.199 support of several of his compatriots, met with Arnold Beckman, 0:13:47.640 --> 0:13:52.040 the founder of the parent company that Shockley Semiconductor was under, 0:13:52.600 --> 0:13:56.960 and he said, Hey, either Shockley goes or we go. 0:13:57.840 --> 0:14:01.960 We cannot continue this. You need to remove Shockley from 0:14:02.040 --> 0:14:06.080 being in a management position, put someone who has actual 0:14:06.120 --> 0:14:10.120 management experience in his role, and I don't know, give 0:14:10.160 --> 0:14:15.840 Shockley like a really cushy professor position at some university. 0:14:15.880 --> 0:14:19.960 But Beckman declined to acquiesce to this request, and so 0:14:20.040 --> 0:14:25.360 in September nineteen fifty seven, Gordon Moore and seven others 0:14:25.400 --> 0:14:30.760 who were important people with Shockley Semiconductor left and founded 0:14:30.800 --> 0:14:34.120 a new company with the help of a wealthy entrepreneur 0:14:34.520 --> 0:14:38.720 named Sherman Fairchild. Fairchild already had a couple of companies 0:14:38.720 --> 0:14:42.880 to his name, including a camera company that worked closely 0:14:42.880 --> 0:14:46.480 with the military, and so they created a division called 0:14:46.720 --> 0:14:53.040 Fairchild Semiconductor. These eight defectors from Shockley Semiconductor became known 0:14:53.080 --> 0:14:57.400 as the Traitorous Eight, though I don't know if Shockley 0:14:57.480 --> 0:15:02.760 himself ever referred to them as the Traitors Eight. Shockley 0:15:02.800 --> 0:15:05.240 would go on to alienate a lot more people and 0:15:05.360 --> 0:15:10.120 sadly adopted increasingly extremist views as he got older, including 0:15:10.160 --> 0:15:14.240 an interest in eugenics. But we'll leave him behind because 0:15:14.320 --> 0:15:19.760 Gord Moorscher did. Now, if everyone within the eight were 0:15:20.320 --> 0:15:23.080 on good terms with each other and were in agreement 0:15:23.160 --> 0:15:27.480 with what the best approach to semiconductors and integrated circuits was, 0:15:28.000 --> 0:15:32.240 our story would likely pretty much stay with Fairchild Semiconductor 0:15:32.280 --> 0:15:36.120 from here on out. But they didn't. There were disagreements. 0:15:36.560 --> 0:15:41.280 There were different teams that were focused on different pathways 0:15:41.760 --> 0:15:46.960 toward a goal, pathways that were not compatible with one another, 0:15:47.200 --> 0:15:50.320 and so there were disagreements between the different teams about 0:15:50.360 --> 0:15:54.720 which design or methodology was the best one to pursue, 0:15:54.800 --> 0:15:58.000 which one was going to create the superior product, and 0:15:58.400 --> 0:16:03.200 things got pretty heated. They did have some unifying beliefs. 0:16:03.240 --> 0:16:05.680 There were things that the whole team agreed upon, and 0:16:05.760 --> 0:16:08.360 one of those was that silicon was going to be 0:16:08.400 --> 0:16:11.120 the semiconductor material of the future. So at the time, 0:16:11.920 --> 0:16:16.800 most research labs were working with germanium as the material 0:16:16.840 --> 0:16:23.080 for semiconductors. But silicon is really very cheap. It is 0:16:23.160 --> 0:16:28.680 like essentially sand, It's plentiful, it is not expensive, and 0:16:28.800 --> 0:16:32.240 the group argued that it would allow for cheaper electronics 0:16:32.840 --> 0:16:36.600 once you reached a point where the integrated circus you 0:16:36.640 --> 0:16:40.280 were building were reliable and efficient, which would require a 0:16:40.280 --> 0:16:42.840 lot of R and D to get there. But they 0:16:42.880 --> 0:16:46.920 saw silicon as being the most promising of the materials 0:16:46.960 --> 0:16:49.600 and something that would ultimately be scalable, that you would 0:16:49.600 --> 0:16:55.440 be able to build much more complicated electronics using silicon 0:16:56.000 --> 0:16:58.800 because of, you know, the fact that it's cheap and 0:16:58.880 --> 0:17:02.960 plentiful and it wouldn't be holding you back. So they 0:17:02.960 --> 0:17:06.680 were talking about a future where electronics would become so 0:17:06.720 --> 0:17:10.359 inexpensive that it would actually be less costly to just 0:17:10.400 --> 0:17:14.040 go out and buy a new whatever it was, rather 0:17:14.080 --> 0:17:18.400 than to have your broken one repaired. World's turn on 0:17:18.520 --> 0:17:24.119 such thoughts. Fairchild Semiconductor began producing transistors and found customers 0:17:24.119 --> 0:17:28.240 in the form of other big companies like IBM, so 0:17:28.440 --> 0:17:31.480 Fairchild was not selling stuff straight to consumers. You wouldn't 0:17:31.520 --> 0:17:35.600 go into your store and buy a fair Child Semiconductor product. 0:17:35.920 --> 0:17:39.440 Instead other companies that would use these components to build 0:17:39.440 --> 0:17:44.800 out their products, which you might possibly buy, although IBM 0:17:44.880 --> 0:17:47.920 wasn't selling directly to consumers at that time either, so 0:17:48.000 --> 0:17:51.960 the computer systems they were designing were for big companies 0:17:52.119 --> 0:17:55.320 or the military. In fact, one of the earliest uses 0:17:55.400 --> 0:17:59.280 for Fairchild semiconductors was in a computer system that was 0:17:59.320 --> 0:18:04.199 for a military bomber. One of Gordon Moore's co workers 0:18:04.320 --> 0:18:10.400 and also someone with whom Moore would clash, was Jean Hoernie. Now, 0:18:10.480 --> 0:18:15.320 Jean developed a new process for transistor manufacturing that improved 0:18:15.400 --> 0:18:19.919 reliability and performance, and all Gordy and Jean were leading 0:18:20.200 --> 0:18:23.520 to different groups that we're working on different approaches to transistors, 0:18:24.040 --> 0:18:26.200 and this would cause a bit of a divide, especially 0:18:26.200 --> 0:18:30.080 since Gordon Moore had the ear of Robert Noyce, who 0:18:30.200 --> 0:18:32.240 was kind of seen as a leader within the group. 0:18:32.720 --> 0:18:36.600 So Moore was able to get his projects moved forward, 0:18:37.320 --> 0:18:39.919 sometimes at the expense or at least perceived to be 0:18:40.400 --> 0:18:43.159 at the expense of others like Jean hoo Ernie, and 0:18:43.240 --> 0:18:47.960 so there was tension within the group. By nineteen sixty, 0:18:48.000 --> 0:18:51.680 the Fairchild team oversaw the development of a silicon wafer 0:18:51.720 --> 0:18:56.000 containing a four transistor circuit on it. This was a 0:18:56.119 --> 0:18:59.359 huge deal. I mean, four transistors is nothing today, all right, 0:18:59.400 --> 0:19:02.679 We're talking about we're in an era where billions of 0:19:02.760 --> 0:19:07.680 components are on a microprocessor we have far far outstripped 0:19:07.800 --> 0:19:11.879 the capabilities of the nineteen sixty four transistor circuit, but 0:19:12.040 --> 0:19:16.800 at the time it was truly a remarkable achievement. It 0:19:16.880 --> 0:19:21.080 was the first integrated circuit that used silicon as the 0:19:21.119 --> 0:19:25.440 semiconductor material. So fair Child was not the first company 0:19:25.520 --> 0:19:29.000 to develop an integrated circuit, because Texas Instruments had already 0:19:29.000 --> 0:19:32.359 done it. Jack Kilby had created an integrated circuit, but 0:19:32.440 --> 0:19:36.159 that one used germanium as the semiconductor material, and he 0:19:36.200 --> 0:19:39.240 had done that back in nineteen fifty eight. The fair 0:19:39.359 --> 0:19:42.600 Child integrated circuit was the first to use silicon, and 0:19:43.080 --> 0:19:46.399 unlike germanium, silicon was much more cheap and plentiful and 0:19:46.480 --> 0:19:52.440 thus scalable. It made integrated circuits a viable technology for 0:19:52.480 --> 0:19:57.720 all sorts of applications, including cheap electronics. So again, this 0:19:57.800 --> 0:20:01.800 was one of those developed ments that would really shape 0:20:02.119 --> 0:20:06.480 the way technology would evolve from that point forward. By 0:20:06.520 --> 0:20:11.320 early nineteen sixty one, the Traitorous eight had split into 0:20:11.359 --> 0:20:14.800 two groups of four because four of the eight would 0:20:14.880 --> 0:20:19.080 leave Fairchild Semiconductor and join a different company called Amelco, 0:20:19.560 --> 0:20:22.760 but Gordon Moore stayed with Fairchild along with Robert Noise, 0:20:23.200 --> 0:20:28.000 and they continued to work on integrated circuits at Fairchild Semiconductor. 0:20:28.560 --> 0:20:31.600 In nineteen sixty five, Gordon More submitted an article to 0:20:31.680 --> 0:20:37.840 the journal Electronics that would forever be associated with him. 0:20:37.960 --> 0:20:43.200 That article has the title Cramming More Components onto Integrated Circuits, 0:20:43.680 --> 0:20:47.000 but the observation More made in that article would later 0:20:47.280 --> 0:20:51.440 be dubbed by Carver Mead of cal Tech as Moore's Law. 0:20:51.920 --> 0:20:55.920 I'll explain more about Moore's law when we come back 0:20:56.080 --> 0:21:09.560 from this quick break. Okay, So if you read Cramming 0:21:09.600 --> 0:21:12.320 More Components onto Integrated Circuits, which you can do. The 0:21:12.600 --> 0:21:15.359 article is freely available on the internet. You can just 0:21:15.400 --> 0:21:19.359 search for that title or search Gordon More article. They'll 0:21:19.400 --> 0:21:22.600 find lots of links to it. Moore has a couple 0:21:22.600 --> 0:21:25.240 of bangers right out of the gate. I mean, you 0:21:25.359 --> 0:21:27.439 might not think of them as that exciting, but I 0:21:27.480 --> 0:21:31.840 think they were incredibly prescient. So the opening statement says, quote, 0:21:32.240 --> 0:21:35.520 with unit cost falling as the number of components per 0:21:35.560 --> 0:21:40.080 circuit rises by nineteen seventy five, economics may dictate squeezing 0:21:40.119 --> 0:21:43.720 as many as sixty five thousand components on a single 0:21:43.800 --> 0:21:48.000 silicon chip end quote. He also says at the beginning 0:21:48.000 --> 0:21:52.239 of the article quote the future of integrated electronics is 0:21:52.560 --> 0:21:58.120 the future of electronics itself end quote. And yeah, these 0:21:58.160 --> 0:22:01.280 were really good observations. So what the heck is this 0:22:01.359 --> 0:22:07.440 article all about. It's mostly covering how economics push companies 0:22:07.440 --> 0:22:11.960 to develop increasingly smaller components for integrated circuits and how 0:22:12.000 --> 0:22:17.520 those circuits power new kinds of electronics, and this perpetuates 0:22:17.520 --> 0:22:20.119 the need to pour more money into developing ways to 0:22:20.119 --> 0:22:25.760 cram even more components onto an inch of silicon wafer 0:22:26.080 --> 0:22:29.679 like a square inch of silicon wafer. So, in other words, 0:22:30.200 --> 0:22:35.880 it's the cycle of being able to create more powerful chips, 0:22:36.400 --> 0:22:41.800 which then drives down the cost per component on the chip. 0:22:41.960 --> 0:22:45.040 The chips themselves might get more expensive, but the components 0:22:45.600 --> 0:22:49.080 get less expensive the actual individual components on the chip 0:22:49.119 --> 0:22:52.000 do and that this in turn creates the ability for 0:22:52.040 --> 0:22:57.639 companies to make cheaper electronics, which goes onto fuel the 0:22:57.960 --> 0:23:04.719 development cycle. So again, this cyclical pattern that allows for 0:23:04.800 --> 0:23:11.240 the progression of of chip development. So balancing this out 0:23:11.440 --> 0:23:14.760 is the scale of production and the ability to integrate 0:23:14.760 --> 0:23:17.879 these circuits into different products. So it almost becomes like 0:23:17.920 --> 0:23:21.119 a you can make it up in volume, because even 0:23:21.160 --> 0:23:25.919 though the component price might go down. The fact that 0:23:25.960 --> 0:23:28.480 you're putting more components on a chip means the individual 0:23:28.560 --> 0:23:31.199 chip might get more expensive to make. So yeah, like 0:23:31.800 --> 0:23:36.000 per component, you're spending less, but then you're cramming way 0:23:36.080 --> 0:23:39.199 more components on a chip, so you spend more. And 0:23:39.400 --> 0:23:41.560 like I said, you have to find that sweet spot, 0:23:41.800 --> 0:23:46.920 that spot where the cost of manufacturing makes sense and 0:23:46.960 --> 0:23:50.360 that you're able to achieve profit by integrating that circuit 0:23:50.400 --> 0:23:56.040 into whatever electronic product you're pushing out there. Then more 0:23:56.160 --> 0:24:01.879 makes another observation that quote the complexity for minimum component 0:24:02.000 --> 0:24:05.199 costs has increased at a rate of roughly a factor 0:24:05.280 --> 0:24:09.160 of two per year end quote. So here he's talking 0:24:09.240 --> 0:24:12.160 about how using the most cost effective means of producing 0:24:12.160 --> 0:24:16.720 integrated circuits, you would expect to see circuit complexity double 0:24:17.440 --> 0:24:20.760 every year, at least over the short term. And he 0:24:20.800 --> 0:24:23.840 does acknowledge that this should change a little bit over time, 0:24:23.960 --> 0:24:27.280 but that within ten years time from the publication of 0:24:27.280 --> 0:24:30.960 the article quote, the number of components per integrated circuit 0:24:31.000 --> 0:24:34.479 for minimum costs will be sixty five thousand end quote. 0:24:35.119 --> 0:24:40.000 So again he's saying that based upon what we're seeing 0:24:40.080 --> 0:24:42.960 right now, we should see the number of components on 0:24:43.000 --> 0:24:48.800 a square inch of silicon chip double every year. Eventually 0:24:48.880 --> 0:24:52.080 it would be every two years. So the entire article, 0:24:52.119 --> 0:24:54.320 like I said's available online. You can read the whole thing, 0:24:54.320 --> 0:24:57.000 and it does touch on some other elements besides the 0:24:57.080 --> 0:24:59.960 economic factors that drive the development in the first place, 0:25:00.520 --> 0:25:04.600 including stuff like dealing with the heat that's generated by 0:25:04.640 --> 0:25:08.280 an integrated circuit that has tens of thousands of components 0:25:08.320 --> 0:25:11.680 that are all crammed next to each other. But later 0:25:11.720 --> 0:25:16.800 on people would reframe More's observation a bit and they 0:25:16.840 --> 0:25:19.040 started to look at it from a slightly different angle. 0:25:19.119 --> 0:25:22.560 So instead of talking about the economic factors that would 0:25:22.560 --> 0:25:25.439 allow for this to happen and ultimately would result in 0:25:25.560 --> 0:25:31.080 cheaper electronics, they saw it as every two years, companies 0:25:31.119 --> 0:25:35.520 double the number of components. Typically we're talking about transistors 0:25:35.520 --> 0:25:39.240 that they can fit on a silicon wafer. Later on, 0:25:39.320 --> 0:25:42.800 people would reframe it even more and say that essentially 0:25:42.880 --> 0:25:47.439 this means that computer companies double the processing power or 0:25:47.480 --> 0:25:52.040 in some cases, processing speed every two years. So you 0:25:52.040 --> 0:25:53.960 could say that if you order to buy a brand 0:25:53.960 --> 0:25:58.040 new computer today, that brand new computer would be twice 0:25:58.040 --> 0:26:01.440 as powerful as a similar model that you had purchased 0:26:01.560 --> 0:26:05.760 two years earlier. Now this also depends upon how you 0:26:05.800 --> 0:26:09.080 define powerful, So there's there's a lot of wiggle room 0:26:09.480 --> 0:26:12.720 for a law. Most laws don't allow for that much 0:26:13.160 --> 0:26:16.040 wiggle room, but Moore's law does. But again, it's because 0:26:16.560 --> 0:26:22.160 this all stems from an observation about the economics of 0:26:22.320 --> 0:26:27.119 producing semiconductor chips. It has since been seen as a 0:26:27.160 --> 0:26:33.680 benchmark that semiconductor companies aim to maintain. There's this pressure 0:26:33.960 --> 0:26:36.840 to keep up with the demands or expectations that have 0:26:36.920 --> 0:26:41.320 been set by Moore's law. Now everyone pretty much acknowledges 0:26:41.359 --> 0:26:44.560 that this trend is not infinitely sustainable, and that we're 0:26:44.560 --> 0:26:47.919 already bumping up against the physical limits of what we 0:26:47.960 --> 0:26:52.000 can do with traditional microchip architecture. But there is still 0:26:52.000 --> 0:26:55.760 this compulsion to keep the spirit of Moore's law going. 0:26:56.359 --> 0:26:58.280 No one wants to be the one to say, all right, well, 0:26:58.320 --> 0:27:01.400 we had a good run. From here on out, we're 0:27:01.400 --> 0:27:04.520 not going to see progress at that same rate. It 0:27:04.600 --> 0:27:07.480 will still see progress, but it's going to slow down. 0:27:07.520 --> 0:27:10.080 Maybe it'll be every four or five years instead of 0:27:10.080 --> 0:27:13.160 every two. Now I'm fairly sure that at the time 0:27:13.240 --> 0:27:16.520 Gordon Moore had no idea how widely known his observation 0:27:16.560 --> 0:27:20.000 would become or the incredible impact it would have on 0:27:20.040 --> 0:27:23.720 the tech sector. But we're not finished with Gordon Moore yet. 0:27:24.240 --> 0:27:27.280 So nineteen sixty five, he publishes this article. At that time, 0:27:27.320 --> 0:27:30.240 he was still hard at work at Fairchild semi Conductor, 0:27:30.640 --> 0:27:34.399 but things were a bit rocky over there. For a 0:27:34.440 --> 0:27:37.680 few years, Fairchild was performing pretty well in the tech 0:27:37.760 --> 0:27:41.400