WEBVTT - Making Electronics Better 0:00:15.316 --> 0:00:27.756 Pushkin. I'm Jacob Goldstein and this is What's Your Problem, 0:00:27.956 --> 0:00:30.716 the show where entrepreneurs and engineers talk about how they're 0:00:30.756 --> 0:00:33.556 going to change the world once they solve a few problems. 0:00:34.516 --> 0:00:38.196 My guest today is Anna Katrina d Letski. She's a 0:00:38.236 --> 0:00:41.716 mechanical engineer who started her career at Apple, where she 0:00:41.796 --> 0:00:46.836 worked on the Apple Watch that was responsible for we 0:00:46.956 --> 0:00:52.396 call it packing the suitcase. So the industrial design team 0:00:52.436 --> 0:00:55.236 makes beautiful surfaces of like what it's going to look 0:00:55.276 --> 0:00:58.716 like on the outside, and the product design engineers figure 0:00:58.756 --> 0:01:01.676 out how to get all the parts inside that suitcase. 0:01:01.836 --> 0:01:03.956 If you will, you got to get all the parts in, 0:01:04.276 --> 0:01:07.516 and then you know that product needs to be robust. 0:01:08.076 --> 0:01:10.556 When Anna was working on the watch, she saw a 0:01:10.596 --> 0:01:13.716 massive problem, not just with the watch or even just 0:01:13.756 --> 0:01:17.916 with Apple, but with all of electronics manufacturing. The problem 0:01:17.996 --> 0:01:23.876 is this electronics manufacturing is incredibly inefficient, incredibly wasteful. And 0:01:23.916 --> 0:01:27.556 I'm not talking about consumers throwing devices away. I'm talking 0:01:27.556 --> 0:01:31.676 about the manufacturing process itself, what happens before the product 0:01:31.716 --> 0:01:35.396 even leaves the factory. So Anna left Apple and started 0:01:35.436 --> 0:01:38.436 a company called Instrumental to try to solve that problem 0:01:39.516 --> 0:01:42.916 in our conversation. Anna laid out the problem really clearly 0:01:43.076 --> 0:01:46.276 when she explained what happens when a new device is 0:01:46.316 --> 0:01:49.196 about to go into production. It starts with a sort 0:01:49.236 --> 0:01:52.116 of practice build, when a few hundred or a few 0:01:52.156 --> 0:01:55.476 thousand devices are assembled as a test. If you're part 0:01:55.516 --> 0:01:58.516 of a team of engineers working in your cushy office 0:01:58.516 --> 0:02:01.516 in Kupertino, this is the moment when you and your 0:02:01.516 --> 0:02:04.356 team have to go and actually see this thing you've 0:02:04.396 --> 0:02:07.556 designed being built. You'd all get on a flight and 0:02:07.636 --> 0:02:10.716 you'd fly in earlatest China. It's where a lot of 0:02:10.716 --> 0:02:13.316 stuff is built, although today maybe it's somewhere else in 0:02:13.356 --> 0:02:17.876 Greater Asia. And then you land in a city, and 0:02:17.916 --> 0:02:20.436 then you'd get into a factory van and you might 0:02:20.516 --> 0:02:22.596 drive for an hour and a half or two hours 0:02:22.596 --> 0:02:25.876 out into a very rural part of China, and that's 0:02:25.876 --> 0:02:29.636 where you'd have a massive manufacturing center. And then we 0:02:29.676 --> 0:02:32.956 would be there for about ten or fourteen days. What 0:02:33.156 --> 0:02:35.396 is the point of you being there. I'm there trying 0:02:35.396 --> 0:02:40.596 to find issues, find problems, find parts that aren't fitting, 0:02:40.796 --> 0:02:44.756 find quality issues with parts. So like we expect there 0:02:44.756 --> 0:02:47.916 to be problems and quality issues with the parts, Anna said, 0:02:48.076 --> 0:02:51.356 she's not allowed to talk about specific preproduction problems they 0:02:51.396 --> 0:02:53.956 had with the Apple Watch, but she did give me 0:02:53.996 --> 0:02:58.876 a hypothetical source of problems the antenna. So, okay, picture 0:02:58.916 --> 0:03:01.956 of factory with hundreds of workers packing all these tiny 0:03:01.996 --> 0:03:04.676 little parts into the watch. One of those parts is 0:03:04.756 --> 0:03:07.876 the antenna, this tiny little thing the sizes of like 0:03:07.996 --> 0:03:11.036 a one inch piece of dried spaghetti, and it has 0:03:11.076 --> 0:03:13.036 to go in there just right so that the watch 0:03:13.076 --> 0:03:15.796 can talk to the world. If the antenna isn't just right, 0:03:16.036 --> 0:03:19.516 the watch doesn't work. Yeah, So let's say hypothetically, you 0:03:19.676 --> 0:03:21.156 at the end of the line, you have a pile 0:03:21.196 --> 0:03:24.916 of units that the antennas don't work. The antenna's really challenging. 0:03:24.956 --> 0:03:27.996 It's like near a lot of metal and antenna's a 0:03:28.036 --> 0:03:32.076 metal don't don't mix, and so the antennas don't work. 0:03:32.276 --> 0:03:36.636 What you have as an engineer is test station results 0:03:36.676 --> 0:03:39.476 that say this signal is low. It could be an 0:03:39.516 --> 0:03:42.956 eighty percent tissue, like eighty eight out of ten don't work. 0:03:43.276 --> 0:03:46.596 It could be a you know, point one percent issue. 0:03:46.756 --> 0:03:48.676 You only have one that doesn't work. It made a 0:03:48.716 --> 0:03:51.236 thousand and nine hundred and ninety nine of them work, 0:03:51.396 --> 0:03:53.356 But what's wrong with this one? Because if you make 0:03:53.396 --> 0:03:55.156 a million, what's wrong with that one? If you make 0:03:55.196 --> 0:03:58.596 a million, one in a thousand is a big problem, right, yes, 0:03:58.876 --> 0:04:04.516 exactly exactly, And so in development you care about every failure. So, okay, 0:04:04.516 --> 0:04:07.556 so you have this problem the antenna doesn't work whatever 0:04:07.836 --> 0:04:10.116 often or once in a while, what do you do? 0:04:10.276 --> 0:04:13.796 We would gather up those physical units, which is non 0:04:13.836 --> 0:04:16.116 trivial to get all those units in one place in 0:04:16.116 --> 0:04:20.316 this big factory, and then we would very carefully start 0:04:20.356 --> 0:04:22.996 to rip them apart to try to get more data 0:04:23.036 --> 0:04:26.076 about what's wrong with these units? Why don't they work? Now. 0:04:26.236 --> 0:04:30.636 One of the things that's interesting about antenna specifically and Watch, 0:04:30.716 --> 0:04:33.276 if you've ever seen a teardown of the watch, is 0:04:33.556 --> 0:04:36.556 when you're taking it apart, you could very easily be 0:04:36.676 --> 0:04:40.156 destroying the evidence of the thing you're trying to find. Oh, 0:04:40.196 --> 0:04:43.396 like you just budge the antenna a tiny little millimeter 0:04:43.516 --> 0:04:46.996 or whatever, and that matter, yes, and it matters. And 0:04:47.076 --> 0:04:49.956 so you might try to take some X ray images 0:04:49.996 --> 0:04:53.876 before you disassemble, so you have some idea literally there 0:04:53.956 --> 0:04:57.956 you have an X ray camera, yes, which is like 0:04:57.956 --> 0:05:02.236 a three dimensional X ray, And then you take them apart, 0:05:02.476 --> 0:05:05.116 and you'd very carefully try to figure out what was 0:05:05.156 --> 0:05:08.596 going wrong. And sometimes those things are obvious. You open 0:05:08.676 --> 0:05:11.036 it up, it has no antenna in it, So that's 0:05:11.036 --> 0:05:14.836 why the antenna doesn't where. Sometimes it's very obvious. Sometimes 0:05:14.836 --> 0:05:17.956 it's not obvious, and like, as a skilled engineer, you're 0:05:17.996 --> 0:05:19.996 looking at it, You're like, I have no idea, And 0:05:20.036 --> 0:05:23.076 then you start looking at the process. You go, you walk, 0:05:23.276 --> 0:05:25.716 you know, fifty meters up the line to look at 0:05:25.716 --> 0:05:28.396 where they're putting the antenna's in, and you're watching that 0:05:28.436 --> 0:05:31.556 process and trying to see is there something happening here? 0:05:31.676 --> 0:05:35.236 Are there like are there screwdrivers swinging across the line 0:05:35.236 --> 0:05:36.996 that are like going to hit the parts and knock 0:05:37.076 --> 0:05:39.196 the antenna? I mean, I feel like there's sort of 0:05:39.276 --> 0:05:44.836 two layers of problems here. Right. There's the fundamental problem of, oh, 0:05:44.916 --> 0:05:47.596 we manufactured a million of these things and we know 0:05:47.796 --> 0:05:49.796 ten thousand of them are going to break, right. That 0:05:49.956 --> 0:05:55.076 is just a straight up problem. There is also the 0:05:55.316 --> 0:05:59.556 like engineers like you have to go fly to China, 0:05:59.636 --> 0:06:01.876 you know, stand on an assembly line and just hope 0:06:01.876 --> 0:06:05.116 they catch things, hope they figure out where the problem is. 0:06:05.316 --> 0:06:07.676 Not even they're looking for problems, they don't even know 0:06:07.676 --> 0:06:10.516 their problems are yet, Like I would go to a 0:06:10.636 --> 0:06:13.596 build and it would be a successful build if we 0:06:13.716 --> 0:06:16.756 left the build with one hundred issues in a spreadsheet 0:06:16.836 --> 0:06:19.356 of different things that we needed to fix. So that 0:06:19.476 --> 0:06:22.756 was the success. Like, the more problems you find, the better. 0:06:22.916 --> 0:06:24.836 I mean, it's a success because you found it because 0:06:24.836 --> 0:06:26.716 you know the problems are there, and you're just worried 0:06:26.756 --> 0:06:28.556 that you're not going to find. Yes, that's like your 0:06:28.636 --> 0:06:31.996 real nightmare problem if you're like a design engineer, right, 0:06:32.036 --> 0:06:34.196 the one that like you don't know is a problem 0:06:34.276 --> 0:06:37.036 until you've shipped a million pairs of earbuds or something. 0:06:37.476 --> 0:06:40.076 The technical term for that is an escape. So that's 0:06:40.116 --> 0:06:44.716 a failure that has escaped the problem escaped the factory. Yes, yes, 0:06:44.796 --> 0:06:50.596 and so escapes are what essentially then cause returns and 0:06:50.676 --> 0:06:53.396 if we go back to that waste, yeah, which is 0:06:53.476 --> 0:06:55.836 hugely costly. Right, it's bad for the reputation of the 0:06:55.876 --> 0:06:59.716 manufacturer because it's like, oh, it's a crappy product, bad reviews, 0:06:59.796 --> 0:07:03.396 not a reliable problem Amazon, bad reviews, and you've got 0:07:03.396 --> 0:07:05.476 to like refund the money, send the person a new one. 0:07:05.516 --> 0:07:07.876 So it's costly on many levels. Yes, yes, and so 0:07:08.036 --> 0:07:13.396 returns should be avoided, and you avoid returns by reducing escapes, 0:07:13.796 --> 0:07:18.676 and you reduce escapes not only by figuring out all 0:07:18.676 --> 0:07:21.076 the things that you can possibly test for. And so 0:07:21.116 --> 0:07:23.596 that's why there are testations on the line, but the 0:07:23.636 --> 0:07:26.796 testation can't test for everything. Did you just live in 0:07:26.916 --> 0:07:34.396 fear all the time of missing some terrible problem? I 0:07:34.396 --> 0:07:38.116 don't even know how to answer that. I think that 0:07:38.316 --> 0:07:42.636 a good engineer is a paranoid engineer. So for me, 0:07:42.756 --> 0:07:45.316 as an engineer, I thought it was kind of silly 0:07:45.356 --> 0:07:47.676 that there is so much luck involved. You just have 0:07:47.716 --> 0:07:49.876 to get lucky and catch the problem before it goes 0:07:49.876 --> 0:07:53.196 into production. Luck like that, yes, and so you're trying 0:07:53.196 --> 0:07:56.076 to find you're relying on luck to find things that 0:07:56.156 --> 0:08:00.556 you can improve in development. And then there's also like 0:08:00.596 --> 0:08:03.516 the scramble and the heroism that happens when you do 0:08:03.636 --> 0:08:06.076 have a problem that everybody learns about. Oh, we have 0:08:06.156 --> 0:08:09.836 this problem. So then you have this fired drill where 0:08:09.836 --> 0:08:12.076 everybody flies in tries to figure out what's going on. 0:08:12.116 --> 0:08:14.996 There's tons of freneticism and activity, and this just seemed 0:08:15.036 --> 0:08:16.676 like a waste, Like why is it that we're so 0:08:16.756 --> 0:08:20.396 reactive versus proactive? Why is it that like I don't 0:08:20.436 --> 0:08:22.796 have access to the data that I need as an 0:08:22.796 --> 0:08:25.676 engineer to actually proactively solve these problems. I have to 0:08:25.756 --> 0:08:30.116 go and like hope versus No. After the break, Annah 0:08:30.196 --> 0:08:34.196 launches Instrumental, a company that's trying to help engineers solve 0:08:34.236 --> 0:08:38.316 the problem of finding problems. Help him go from hope 0:08:38.876 --> 0:08:49.556 to know. That's the end of the ads. Now we're 0:08:49.556 --> 0:08:51.916 going back to the show, So just a reset here. 0:08:52.276 --> 0:08:55.516 Anna has landed a job at Apple, the company where 0:08:55.556 --> 0:09:00.156 engineers dream of working, this amazing technology company, and yet 0:09:00.196 --> 0:09:02.796 she decides to leave. So I asked her in the end, 0:09:02.916 --> 0:09:05.356 what made you quit your job and go start this 0:09:05.396 --> 0:09:10.156 new company, Instrumental. I can share some thing that that 0:09:10.316 --> 0:09:12.436 you can use, but it's maybe it might be a 0:09:12.436 --> 0:09:14.636 little sideways to what you're talking about, but it's the 0:09:14.676 --> 0:09:18.436 true reason I started the company. So about a year 0:09:18.436 --> 0:09:21.436 before Apple Watch shipped and the world knew that it 0:09:21.516 --> 0:09:24.476 even existed, I actually had a personal tragedy happened in 0:09:24.476 --> 0:09:27.676 my life. My husband was killed by a drunk driver. 0:09:28.236 --> 0:09:31.356 Oh my god, and I was twenty seven, he was 0:09:31.396 --> 0:09:39.636 twenty seven. And I realized that life is precious and limited, 0:09:40.556 --> 0:09:42.276 and you don't know when your last day is going 0:09:42.316 --> 0:09:44.676 to be, and so you want to be proud of 0:09:44.716 --> 0:09:47.036 the days that you have and how you spent them 0:09:47.196 --> 0:09:49.636 and the impact you could have on others, and also 0:09:49.716 --> 0:09:54.836 reevaluate the justification for my existence on this planet. And 0:09:55.356 --> 0:09:58.676 I decided that Apple is great. I'm glad I was there, 0:09:59.036 --> 0:10:01.596 but it can't be my life's work for me. I 0:10:01.716 --> 0:10:05.876 felt like I needed to do something bigger. And the 0:10:06.036 --> 0:10:08.916 big problem that we all need to solve is how 0:10:09.156 --> 0:10:12.956 we build stuff is so wasteful. That is a more 0:10:13.076 --> 0:10:17.036 meaningful direction. And there is a big problem that I 0:10:17.076 --> 0:10:19.916 know that I have some intuition about how to go solve, 0:10:20.156 --> 0:10:23.116 And so I took what I understood as like, wow, 0:10:23.116 --> 0:10:24.876 it's like really hard to find and fix these issues. 0:10:24.876 --> 0:10:28.116 We're wasting so much money, so much physical stuff. We 0:10:28.196 --> 0:10:32.076 are pumping so much chemicals into our rivers, we are 0:10:32.196 --> 0:10:36.716 burning so much energy. We are wasting human lifetimes of 0:10:36.796 --> 0:10:44.596 time doing things that don't matter. It's just wasted, and 0:10:45.236 --> 0:10:47.276 we need to figure that out. We need to change 0:10:47.276 --> 0:10:50.156 how we build, We need to change how we think 0:10:50.196 --> 0:10:53.076 about building products. So this is like a little maybe 0:10:53.396 --> 0:10:56.556 sappy for your audience, but this is the true reason 0:10:57.476 --> 0:11:02.516 that like really made the change of thinking about away 0:11:02.596 --> 0:11:04.836 from thinking about, oh, my job as an engineer, I 0:11:04.996 --> 0:11:08.876 make stuff too. Oh, my job as a human is 0:11:08.876 --> 0:11:11.956 to figure out how to do the things we need 0:11:12.036 --> 0:11:15.396 to do better. So Anna decides she's going to try 0:11:15.436 --> 0:11:20.996 to make electronics manufacturing better, more efficient, less wasteful. And 0:11:21.076 --> 0:11:23.356 to do this, she and an engineer she worked with 0:11:23.396 --> 0:11:26.596 at Apple decide to start a company. They call it Instrumental, 0:11:27.196 --> 0:11:30.956 And when it started, they had one big idea. Instead 0:11:30.996 --> 0:11:33.996 of forcing engineers like Anna to frantically run up and 0:11:33.996 --> 0:11:37.676 down the assembly line hoping they'd spotted every potential problem, 0:11:38.196 --> 0:11:41.156 they would put dozens of cameras up and down the 0:11:41.196 --> 0:11:46.476 line to capture the entire assembly process for every single device. 0:11:46.876 --> 0:11:50.436 It's the idea that there is value in this data 0:11:50.676 --> 0:11:54.956 that today does not get collected, and so the actual 0:11:54.996 --> 0:11:58.476 core pieces, Let's go get all the data. That's the 0:11:58.916 --> 0:12:01.836 that's the core innovation here. Let's go get all the data. 0:12:02.036 --> 0:12:04.836 The data in this case is visual data. What does 0:12:04.876 --> 0:12:07.996 everything look like as this phone or this watch or 0:12:08.036 --> 0:12:10.356 whatever is getting put together? Is that right? That was 0:12:10.396 --> 0:12:13.076 the first data set that we went after was the 0:12:13.116 --> 0:12:17.196 image data set, because there isn't one. So we're creating 0:12:17.436 --> 0:12:20.156 a data set that didn't exist before. And as you know, 0:12:20.276 --> 0:12:22.516 as they say, a picture is worth a thousand words. 0:12:23.236 --> 0:12:24.956 When you take a picture, you don't need to know 0:12:24.996 --> 0:12:27.476 in advance what you're going to be looking for in 0:12:27.516 --> 0:12:31.196 the picture. It's a very high resolution data set, and 0:12:31.236 --> 0:12:34.196 then you can come back with twenty twenty hindsight and 0:12:34.236 --> 0:12:36.796 look at the picture and be like, oh, like, here's 0:12:36.796 --> 0:12:39.316 a problem in that picture. And so that was the 0:12:39.356 --> 0:12:41.956 initial concept of like, we need to find where these 0:12:41.956 --> 0:12:43.996 issues are, which means we need to see them. We 0:12:44.076 --> 0:12:46.116 work in the physical world, so we need to see them. 0:12:46.356 --> 0:12:48.556 Remember I was talking about the antennas, where you had 0:12:48.596 --> 0:12:51.356 to carefully take them apart, and as you took them apart, 0:12:51.436 --> 0:12:54.956 you'd sometimes damage the evidence you're looking for. Well, if 0:12:54.956 --> 0:12:58.036 you have images of them as they went together, then 0:12:58.196 --> 0:13:00.236 you don't maybe have to take them apart, or you 0:13:00.316 --> 0:13:03.876 maybe have additional data about what they looked like as 0:13:03.876 --> 0:13:05.796 they were going together. But you don't know which ones 0:13:05.836 --> 0:13:08.316 are going to be failures in advance, so you take 0:13:08.316 --> 0:13:10.596 pictures of all of them and then you have the 0:13:10.636 --> 0:13:13.076 ones after the fact that ended up being failures. You 0:13:13.116 --> 0:13:15.796 have those images. So that was like the first idea. 0:13:15.916 --> 0:13:18.036 So the basic ideas. You take pictures of the whole thing, 0:13:18.116 --> 0:13:20.756 and you identify, you know, whatever, the serial number on 0:13:20.836 --> 0:13:24.116 each say, watch each device that's going down the line, 0:13:24.276 --> 0:13:26.796 and then if there's one device that doesn't work, you say, 0:13:26.796 --> 0:13:28.196 show me all the pictures of that one, and then 0:13:28.196 --> 0:13:29.916 you say, how is that one different than all the 0:13:29.916 --> 0:13:32.756 ones that do work in the pictures? Yes, yes, But 0:13:33.036 --> 0:13:35.596 it's not just about figuring out what's wrong once you 0:13:35.636 --> 0:13:38.156 find a problem, right is It also about getting better 0:13:38.196 --> 0:13:41.236 at finding problems in the first place. So discovery, we're 0:13:41.236 --> 0:13:43.956 able to solve the discovery problem for our customer. We 0:13:44.076 --> 0:13:50.156 built algorithms that look for anomalies, and so what's an anomaly. 0:13:50.716 --> 0:13:53.796 An anomaly is something that's different from the other ones. 0:13:54.036 --> 0:13:56.916 So if you build a population of units, they should 0:13:56.916 --> 0:13:59.636 all look the same, Okay, a bunch of devices, right, 0:13:59.876 --> 0:14:03.796 The ones that look different are probably interesting, and so 0:14:03.836 --> 0:14:07.436 we build algorithms to find the ones that look different. 0:14:07.836 --> 0:14:11.196 And that was kind of the first offering that we provided, 0:14:11.396 --> 0:14:14.156 And so our customers were able to use that combination 0:14:14.716 --> 0:14:19.636 of the photo record with the algorithms that would highlight 0:14:19.676 --> 0:14:23.156 automatically the things that were different and interesting, and they 0:14:23.156 --> 0:14:26.436 were able to use that to find issues. Basically, one 0:14:26.476 --> 0:14:29.236 of these things is not like the other, yes, exactly, 0:14:29.356 --> 0:14:31.996 So like, oh, this one's bent a little bit, this 0:14:32.036 --> 0:14:34.396 one's the parts supposed to be black and it's white. 0:14:34.476 --> 0:14:37.916 Is that a problem? Maybe not? Maybe we highlight those 0:14:37.956 --> 0:14:41.716 things and then the next piece is okay, well what's 0:14:41.756 --> 0:14:45.436 causing like problems? Okay, so you've built a system that 0:14:45.556 --> 0:14:49.636 is better at identifying for finding problems. So now even 0:14:49.676 --> 0:14:53.436 if a device pass all the tests, your system will say, well, 0:14:53.436 --> 0:14:55.956 you want to check out this one missing screws because 0:14:55.956 --> 0:14:59.356 it just looks a little different missing screws. Okay, so yeah, 0:14:59.476 --> 0:15:02.396 like that missing screws, there's no screw tester. You're right, 0:15:02.436 --> 0:15:05.996 like missing screws. Devices will actually miss screws. I'm so 0:15:06.116 --> 0:15:07.916 naive that I'm like, well, surely a device wouldn't be 0:15:07.996 --> 0:15:10.996 missing a screw, but one hundred percent, it's like one 0:15:11.036 --> 0:15:13.796 of the top ten defects in production is missing or 0:15:13.996 --> 0:15:17.396 extra parts, extra screw just to make up for the 0:15:17.396 --> 0:15:19.396 other one that's missing a screw. On average, they have 0:15:19.436 --> 0:15:22.436 the right number of screws. Yes, on average they have 0:15:22.516 --> 0:15:25.716 the right number of scares. So we need to be 0:15:25.756 --> 0:15:28.156 able to find issues. We need to be able to 0:15:28.196 --> 0:15:31.716 help engineers fix them, and then we need to help 0:15:31.756 --> 0:15:33.756 them make sure they don't come back. Those are kind 0:15:33.756 --> 0:15:36.516 of the three things that we have to do. So okay, 0:15:36.596 --> 0:15:39.196 those are the three key things Anna's company does to 0:15:39.236 --> 0:15:42.596 solve this big problem and to tile those pieces together. 0:15:42.916 --> 0:15:45.436 She gave me a case study, a case study from 0:15:45.436 --> 0:15:48.396 a company that was trying to build a webcam and 0:15:48.476 --> 0:15:51.596 the company was having trouble with the antenna. The antenna, which, 0:15:51.756 --> 0:15:54.516 as Anna told me about earlier in the interview, is 0:15:54.556 --> 0:15:57.676 this crucial part that is surprisingly hard to get right. 0:15:57.876 --> 0:16:01.036 They always fail. There are very sensitive parts, and all 0:16:01.036 --> 0:16:04.036 of our connected devices really rely on their antenna performance. 0:16:04.036 --> 0:16:06.116 So you don't think about it, but you know it's 0:16:06.156 --> 0:16:08.716 an important part. And so we had a customer who 0:16:08.756 --> 0:16:13.116 had failing antennas and they were able to take that 0:16:13.236 --> 0:16:15.756 group of that population of units and they were able 0:16:15.796 --> 0:16:18.596 to actively, like root cause it to three different things. 0:16:18.636 --> 0:16:21.236 And so this is actually the next piece of the puzzle, 0:16:21.516 --> 0:16:24.476 which is we need to find root cause. And so 0:16:24.916 --> 0:16:27.436 in this case, the customer was able to see a 0:16:27.476 --> 0:16:29.956 couple of different things. The first thing they found is 0:16:29.996 --> 0:16:32.516 a group of those failures was actually had to do 0:16:32.676 --> 0:16:36.276 with the alignment of a connector that was connecting the antenna. 0:16:36.356 --> 0:16:38.556 So the connector, if it was like shifted a little 0:16:38.596 --> 0:16:40.796 bit this way, it was good, and it was shifted 0:16:40.836 --> 0:16:43.116 a little bit the other way, it was bad. And 0:16:43.156 --> 0:16:45.836 this was very subtle. This is not something they had 0:16:45.916 --> 0:16:50.916 like a specification. This is a discovery. Okay, so the 0:16:51.036 --> 0:16:53.916 angle mattered and they didn't know that in advance. Okay, 0:16:53.996 --> 0:16:57.556 Now this is also maybe too much in the weeds. 0:16:57.636 --> 0:17:00.236 But there are a lot of reasons, and antenna can 0:17:00.276 --> 0:17:02.636 fail in a lot of different ways. So it's not 0:17:02.716 --> 0:17:05.436 just like, oh, if they're failing, they're all the same 0:17:05.836 --> 0:17:08.916 root costs. Actually, there could be a group of different 0:17:09.236 --> 0:17:12.196 causes that cause failures. So you have two different phones 0:17:12.396 --> 0:17:15.276 from the same build, both of which have the antenna 0:17:15.316 --> 0:17:17.476 that doesn't work. It might be a different reason for 0:17:17.516 --> 0:17:19.756 each phone that the antenna doesn't work, which makes it 0:17:19.756 --> 0:17:21.516 harder to solve. So you solve one problem, but then 0:17:21.556 --> 0:17:22.916 some of the phones still don't work. Is that what 0:17:22.916 --> 0:17:26.276 you're saying, Yes, yes, that's very common. They also found 0:17:26.356 --> 0:17:29.796 that the software version that they were using on the 0:17:29.836 --> 0:17:33.396 tester was also correlated to the failures, so most of 0:17:33.436 --> 0:17:36.196 the failures were from one particular tester, So in fact, 0:17:36.236 --> 0:17:39.916 it wasn't a bad antenna. It was a test for 0:17:39.996 --> 0:17:41.836 some of those units, not all of them. Some of 0:17:41.836 --> 0:17:44.116 them still had the connector coming out at a wrong angle. 0:17:44.236 --> 0:17:47.636 Yeah yeah. And then the last one on the same 0:17:47.716 --> 0:17:50.316 group of failures, in the same group of failures, there 0:17:50.396 --> 0:17:54.756 was another group of units where what the image based 0:17:54.756 --> 0:17:57.716 algorithms found as a high correlation to the color of 0:17:57.796 --> 0:18:02.356 the circuit board. And so color doesn't affect performance, but 0:18:02.876 --> 0:18:05.756 it was a different vendor. One circuit board was like 0:18:05.796 --> 0:18:09.316 a slightly different shade of a color than another, and 0:18:09.436 --> 0:18:11.796 so that means like the vendor of that circuit board 0:18:12.996 --> 0:18:15.476 was for one of the vendors that made that circuit 0:18:15.476 --> 0:18:17.516 board was more likely to have failing in it. So 0:18:17.596 --> 0:18:20.236 that gives engineers then three different things. They can go 0:18:20.356 --> 0:18:22.556 chase down. They can go chase down a process is 0:18:22.636 --> 0:18:24.876 year around the angle of the connector. They can go 0:18:24.996 --> 0:18:28.476 chase down like the drifting test station and work on 0:18:29.436 --> 0:18:32.476 calibrating the testations. And then they can go and investigate 0:18:32.636 --> 0:18:35.636 if there's something meaningfully different between these two vendors of 0:18:35.836 --> 0:18:38.316 circuit boards. So they're able to continue building, but they 0:18:38.316 --> 0:18:41.236 have now three specific things they can go do, whereas 0:18:41.276 --> 0:18:43.876 without instrumental they would kind of be guessing. So, like, 0:18:44.316 --> 0:18:46.436 what's the sort of frontier for you? What's the thing 0:18:46.476 --> 0:18:48.916 you're trying to do that you haven't quite figured out yet, Like, 0:18:48.916 --> 0:18:51.956 what's the next problem to solve? Yeah, I mean so 0:18:52.036 --> 0:18:54.676 it goes back to the reason for being for the company, 0:18:54.996 --> 0:18:57.876 the reason for being being that twenty cents of every 0:18:57.876 --> 0:19:01.156 dollar that's wasted in manufacturing, we haven't solved that problem yet. 0:19:01.316 --> 0:19:04.516 What does that require? We need to change how we 0:19:04.556 --> 0:19:07.036 think about how we build things, how we design things, 0:19:07.116 --> 0:19:11.356 the process of iterating through that development process, the process 0:19:11.396 --> 0:19:14.156 of what happens when it goes into production, what happens 0:19:14.196 --> 0:19:16.316 when you return units? And what do we do with 0:19:16.356 --> 0:19:20.796 that information from your return as a consumer. And today 0:19:20.836 --> 0:19:23.836 we're really not doing that as an industry. Not gathering 0:19:23.996 --> 0:19:27.276 data in a meaningful way is that we're not gathering data, 0:19:27.436 --> 0:19:31.356 we're not thinking proactively. As an engineer. If I'm designing something, 0:19:31.396 --> 0:19:33.996 I'm thinking about, you know, the next build and production, 0:19:34.036 --> 0:19:38.276 but I'm not I'm not necessarily thinking about, like how 0:19:38.316 --> 0:19:40.996 can I develop this product's that I can capture the 0:19:41.036 --> 0:19:44.316 most data from it the fastest, the earlier in the 0:19:44.356 --> 0:19:47.636 process versus later in the product. Like everybody knows their 0:19:47.716 --> 0:19:50.156 data is valuable and there's stuff in there that could 0:19:50.156 --> 0:19:52.876 be valuable, but they don't know how to value it 0:19:53.116 --> 0:19:56.236 because the problems haven't happened yet. When they're on fire 0:19:56.316 --> 0:20:00.396 quote unquote and they are spending a million dollars a 0:20:00.436 --> 0:20:04.156 month on returns, they know exactly what the value is 0:20:04.196 --> 0:20:07.356 in the data that they don't have, Yes, but when 0:20:07.396 --> 0:20:10.116 they haven't started the program yet and they don't know 0:20:10.156 --> 0:20:12.836 what fires they're going to have, they don't know how 0:20:12.836 --> 0:20:16.356 to value having the data. And like preventing those problems 0:20:16.356 --> 0:20:18.516 in the first place, the vision is that that data 0:20:18.596 --> 0:20:21.956 is enough and then if you figured out how to 0:20:21.996 --> 0:20:24.476 harvest that data, you can actually build lines that improve 0:20:24.516 --> 0:20:27.756 themselves and then you eliminate the waste. So this new 0:20:27.796 --> 0:20:30.196 different way of doing things, like can you just tell 0:20:30.196 --> 0:20:32.876 me sort of specifically, like a few details, what would 0:20:32.876 --> 0:20:35.036 it be like, what would it look like, what would 0:20:35.076 --> 0:20:38.716 be happening in this world? Everything gets built cheaper, everything 0:20:38.756 --> 0:20:42.996 gets built with less waste. And when that happens, we 0:20:43.076 --> 0:20:45.716 have maybe we have cheaper products, we have maybe more 0:20:45.716 --> 0:20:49.716 profitable companies, we have less waste going into the world, 0:20:49.956 --> 0:20:53.756 less physical waste, chemicals and rivers, less energy use, less 0:20:53.796 --> 0:20:58.636 human lifetimes wasted, And it's just thinking completely differently about 0:20:58.756 --> 0:21:02.916 like that manufacturing actually is a machine itself, and that 0:21:02.956 --> 0:21:05.356 we need to optimize that whole process as a machine 0:21:05.356 --> 0:21:08.636 itself versus a means to justify the ends of like 0:21:08.836 --> 0:21:10.836 we need we need to get products out the other 0:21:10.916 --> 0:21:14.236 end of the line. In a minute, the Lightning Round, 0:21:15.116 --> 0:21:18.316 Anna tells us what mechanical engineers know about the world 0:21:18.876 --> 0:21:30.356 and the inefficiencies out in the wild that grind her tears. Now, 0:21:30.436 --> 0:21:34.116 let's get back to what's your problem? Great, let's do 0:21:34.156 --> 0:21:37.996 the Lightning Round. Are you ready? What's one piece of 0:21:38.036 --> 0:21:41.276 advice you'd give to someone trying to solve a hard problem. 0:21:41.356 --> 0:21:45.796 I love that question. I love it. I had a 0:21:45.836 --> 0:21:47.716 science teacher in high school taught me how to solve 0:21:47.756 --> 0:21:50.236 hard problems because I was working on science research. I'm 0:21:50.276 --> 0:21:53.916 a science fair kid, and he taught me to break 0:21:53.956 --> 0:21:56.996 those problems down into manageable pieces. So like, if you 0:21:56.996 --> 0:22:00.836 can take smaller steps and understanding a problem and the 0:22:00.916 --> 0:22:04.476 small steps you can do, then it becomes easier to solve. 0:22:04.956 --> 0:22:08.276 So as somebody whose job is sort of to find 0:22:08.516 --> 0:22:12.116 errors and inefficiencies in the world. Is there some some 0:22:12.276 --> 0:22:15.076 domain something you encounter in your daily life that you 0:22:15.196 --> 0:22:19.956