WEBVTT 1 00:00:06.160 --> 00:00:09.039 Hello, and welcome to this week's episode of Alexa Mix. 2 00:00:09.119 --> 00:00:12.199 We are joined by, as ever, our fabulous panel of posts. 3 00:00:12.320 --> 00:00:16.760 We have Alex Kotman, Hey Alex. We've got Bruce t Hi. Everybody, 4 00:00:16.839 --> 00:00:18.960 Hello Bruce. We've got Steven Nunyez. 5 00:00:19.079 --> 00:00:19.480 Hello. 6 00:00:19.640 --> 00:00:22.280 I am SOPHIEA Benedetto and we are joined today by 7 00:00:22.440 --> 00:00:26.719 Randall Thomas, who has many potential introductions that I may 8 00:00:26.760 --> 00:00:28.719 choose from. So I'm going to go with my favorite two, 9 00:00:28.760 --> 00:00:31.320 which is the first. One's a little more informative CTO 10 00:00:31.480 --> 00:00:34.920 at a tech incubator called Geometer, as well as a 11 00:00:35.159 --> 00:00:37.479 thanks for Hacker and vonn Dufon, and I'm sure we'll 12 00:00:37.479 --> 00:00:40.840 be excited to dig into any of those topics. Welcome Randall, Oh, 13 00:00:40.920 --> 00:00:42.719 thank you so much for having me. Yeah, we're so 14 00:00:42.759 --> 00:00:44.719 thrilled you could be here. We've got so many great 15 00:00:44.719 --> 00:00:46.960 things to dig into today. But one of the things 16 00:00:46.960 --> 00:00:48.719 that we often ask us when they first come on 17 00:00:48.840 --> 00:00:50.320 is tell us a little bit about how you got 18 00:00:50.320 --> 00:00:52.039 into the Elixir community in the first place. 19 00:00:52.359 --> 00:00:54.799 Strangely enough, the short version is I've been around the 20 00:00:54.840 --> 00:00:57.600 Rube community for a long time and one day looked 21 00:00:57.600 --> 00:00:59.280 around in all the people that I knew from the 22 00:00:59.280 --> 00:01:02.159 old Rubey community, you weren't there anymore, and I figured 23 00:01:02.200 --> 00:01:04.319 out where they went, and it turns out that they 24 00:01:04.359 --> 00:01:06.560 all showed up inn Elixir. In fact, it was kind 25 00:01:06.599 --> 00:01:10.159 of accidental. I went to I think the first Elixir 26 00:01:10.200 --> 00:01:11.840 CONFI I had gone to, I think was in twenty 27 00:01:11.959 --> 00:01:14.560 eighteen or twenty seventeen, and I showed up there and 28 00:01:14.599 --> 00:01:17.280 I literally just walked around. I'm like, wait, so that's 29 00:01:17.280 --> 00:01:21.120 why you weren't at Ruby KMF. Oh wait, so you're here. 30 00:01:21.200 --> 00:01:23.239 What are you doing here? So I thought I would 31 00:01:23.280 --> 00:01:25.239 know no one, and it turned out that I knew everybody. 32 00:01:25.319 --> 00:01:27.439 So all the smart people that I sort of cocktailed 33 00:01:27.480 --> 00:01:29.640 on and Ruby ended up in Elixir. So I figured 34 00:01:29.680 --> 00:01:31.840 they're way smarter than me. I should go follow them 35 00:01:31.879 --> 00:01:33.799 see what they're doing. So it's kind of it. 36 00:01:34.000 --> 00:01:36.239 That's really as good a reason as any, and kind 37 00:01:36.280 --> 00:01:39.000 of similar to how I started learning Alixir. It's like, 38 00:01:39.359 --> 00:01:41.680 I don't know, maybe four or five years ago, one day, 39 00:01:41.760 --> 00:01:44.120 maybe five five thirty, and Steven says, hey, you want 40 00:01:44.120 --> 00:01:46.200 to see something cool, And that was kind of like 41 00:01:46.640 --> 00:01:49.640 theentre into the wide world of alexir. So following the 42 00:01:49.680 --> 00:01:50.840 smart people is the way to go. 43 00:01:51.280 --> 00:01:53.599 There was a funny thing. I'm pretty sure this is 44 00:01:53.680 --> 00:01:57.640 actually true, right, but like all apocryphal stories, because it 45 00:01:57.680 --> 00:02:00.959 happened in Barcelona. There was Baruko, Barcelon and Ruby KMF 46 00:02:01.120 --> 00:02:03.159 and at the time we were giving this talk on 47 00:02:03.519 --> 00:02:06.159 Ruby motion. If you remember, there was this big kerfuffle 48 00:02:06.200 --> 00:02:08.000 about hey man, you're gonna be able to write iOS 49 00:02:08.039 --> 00:02:10.439 apps in Ruby. And it was actually a really really 50 00:02:10.439 --> 00:02:12.000 great idea, I kind of think a little ahead of 51 00:02:12.039 --> 00:02:16.919 its time. And I think actually at that conference somebody 52 00:02:17.000 --> 00:02:19.520 was talking about, oh, yeah, you heard that somebody's trying 53 00:02:19.520 --> 00:02:22.919 to build a Ruby for Elixir or for Erlang, and 54 00:02:22.960 --> 00:02:26.680 I'm just like, yeah, good luck with that. I literally 55 00:02:26.719 --> 00:02:28.680 thought it was like because I was thinking about, like 56 00:02:28.680 --> 00:02:32.000 I tried learning Arlang once when the prag Prague released 57 00:02:32.000 --> 00:02:34.919 that first book from Joe on Arlang, and I remember 58 00:02:34.919 --> 00:02:37.199 getting into it and I remember thinking I'm like, okay, 59 00:02:37.439 --> 00:02:39.800 I'll put this somewhere right above pearl. I recant that 60 00:02:39.840 --> 00:02:42.520 statement now, but at the time I was thinking to myself, like, 61 00:02:42.599 --> 00:02:44.439 there's no way you're ever going to get Rubis to 62 00:02:44.479 --> 00:02:47.280 learn the syntax, because you know, the obsession with beauty 63 00:02:47.280 --> 00:02:49.439 and elegance and expressiveness and everything else. 64 00:02:49.599 --> 00:02:52.080 Yeah, and it's really funny that Jose kind of put 65 00:02:52.080 --> 00:02:55.280 it onto the Ruby syntax, not as this great statement 66 00:02:55.479 --> 00:02:59.439 of let's bring all the Ruby community over right. It 67 00:02:59.520 --> 00:03:03.240 was more like, okay, one last thing, right, no way yet, 68 00:03:03.360 --> 00:03:05.039 and then then he kind of layers on these these 69 00:03:05.080 --> 00:03:07.759 beautiful abstractions. But yeah, it seems like we were kind 70 00:03:07.759 --> 00:03:09.879 of circling to each other for a long time randal 71 00:03:10.000 --> 00:03:13.280 before for the elexur community that you know, the idea 72 00:03:13.400 --> 00:03:15.560 that that you were EngineYard and you know, us to 73 00:03:15.680 --> 00:03:18.800 point some stuff there, and you were kind of kind 74 00:03:18.800 --> 00:03:22.000 of involved from from the very beginning, and kind of 75 00:03:22.159 --> 00:03:24.919 we were. We were just just outside of each other's 76 00:03:25.199 --> 00:03:28.080 or bit and didn't really meet until the Garroxio course. 77 00:03:28.400 --> 00:03:30.960 Yeah, which is that's actually what I like to say is, 78 00:03:31.000 --> 00:03:35.000 you know, I do a lot to to Bruce for 79 00:03:35.000 --> 00:03:37.599 for really getting involved in the electric community. I don't 80 00:03:37.599 --> 00:03:40.199 actually did you just advertise that course online? Was that 81 00:03:40.240 --> 00:03:41.120 what it was? Yeah? 82 00:03:41.159 --> 00:03:44.159 I mean we literally just threw a couple of tweets 83 00:03:44.199 --> 00:03:47.120 up just to see if there was a test market, 84 00:03:47.400 --> 00:03:50.360 to test the market, and we said we would need 85 00:03:50.439 --> 00:03:52.560 more people to start the class, and then you were 86 00:03:52.599 --> 00:03:54.719 two signed up. But one was you right, and the 87 00:03:54.759 --> 00:03:57.439 other one was Patrick Thompson, who's already writing you know, 88 00:03:57.840 --> 00:04:01.960 articles and pull requests for or Phoenix or Phoenix Lide you. 89 00:04:02.199 --> 00:04:04.039 But that was kind of a cool moment because we 90 00:04:04.039 --> 00:04:06.319 had to decide whether to do the class or not, 91 00:04:06.520 --> 00:04:09.960 and at that point we decided to go against the 92 00:04:09.960 --> 00:04:14.199 grain of traditional courses to it's kind a you know, 93 00:04:14.439 --> 00:04:18.160 like a smaller Midge school right where everybody could and 94 00:04:18.199 --> 00:04:20.639 then that's that's where we kind of decided to flip 95 00:04:20.639 --> 00:04:22.600 the keyboard, and you know, the rest is history. 96 00:04:22.839 --> 00:04:24.759 Yeah, I was a pretty New York or program before 97 00:04:24.800 --> 00:04:26.439 I got to spend a week in pairing with Bruce, 98 00:04:26.480 --> 00:04:28.480 at least for a elixir, which is I think that's 99 00:04:28.639 --> 00:04:31.279 something that we take for granted. You know, if you 100 00:04:31.319 --> 00:04:33.639 come from procedural languages, it's really easy. If you know C, 101 00:04:34.040 --> 00:04:35.959 you can kind of learn C plus plus even though 102 00:04:35.959 --> 00:04:37.920 it might not be the right idioms, but it's the bridges. 103 00:04:37.959 --> 00:04:39.519 If you know job, you can kind of get into 104 00:04:39.560 --> 00:04:42.959 go and like. But learning how to actually think in 105 00:04:43.040 --> 00:04:45.519 elixir and break down the problem differently or thing functionally 106 00:04:45.639 --> 00:04:47.680 is actually one of the big barrier sentry I think 107 00:04:47.680 --> 00:04:51.120 people have and seeing it and doing it with other 108 00:04:51.120 --> 00:04:53.680 people and going through that evolution was really helped me 109 00:04:53.800 --> 00:04:55.360 but actually really leveled up my elixir. 110 00:04:55.839 --> 00:04:59.240 Yeah, so I don't I don't take any credit at 111 00:04:59.279 --> 00:05:02.160 all for any you're doing an elixir uh by the way, 112 00:05:02.199 --> 00:05:05.360 but I do think that there's a moment where your 113 00:05:05.360 --> 00:05:07.199 revolution is like a lot of a lot of other 114 00:05:07.199 --> 00:05:09.439 people's like mine as well. Right, it's why are the 115 00:05:09.480 --> 00:05:12.480 cool kids leaving the room? What do they see and 116 00:05:12.800 --> 00:05:15.879 how do I how do I make the mini that 117 00:05:16.319 --> 00:05:20.879 mental shift to do the things that that good programmers do. 118 00:05:21.120 --> 00:05:24.279 And you know that's what we were all able to go 119 00:05:24.360 --> 00:05:26.680 through together in that class, you and Patrick and I. 120 00:05:26.879 --> 00:05:28.759 Yeah, well I I still think it's one of the 121 00:05:28.800 --> 00:05:30.560 greatest things that where I recommend it. In fact, actually, 122 00:05:30.600 --> 00:05:32.360 if you remember, one of the first things I did 123 00:05:32.360 --> 00:05:35.079 at CTO was I got a subscription. I think I 124 00:05:35.120 --> 00:05:36.920 need you write a feature so I could get my 125 00:05:37.000 --> 00:05:40.279 team on Groxia. So shameless plug for Bruce's like learning 126 00:05:40.279 --> 00:05:43.240 platform Groxio. It's great to learn alixer in a browser. 127 00:05:43.319 --> 00:05:45.319 And I've actually used it for other people and they're like, well, 128 00:05:45.319 --> 00:05:46.839 what do you think about Elixer. I'm like, I don't know. 129 00:05:47.319 --> 00:05:49.279 Go try it out in a browser, like most people 130 00:05:49.319 --> 00:05:51.240 are just blown away that'll run in a browser anyway. 131 00:05:51.480 --> 00:05:55.279 So it really doesn't not in a browser, it doesn't. 132 00:05:55.360 --> 00:05:58.600 But like those types of abstractions, it's not important. It's 133 00:05:58.639 --> 00:06:01.360 like it's like everybody else. You know, first one's free, Amen, 134 00:06:01.720 --> 00:06:02.639 first one's free. 135 00:06:02.879 --> 00:06:07.519 Yeah. But so as your Elixir Generney journey has kind 136 00:06:07.519 --> 00:06:10.600 of started off, you didn't kind of peck around the 137 00:06:10.680 --> 00:06:12.279 edges like most people did. 138 00:06:12.439 --> 00:06:12.600 Right. 139 00:06:12.680 --> 00:06:15.480 It's you know, you went from kind of zero to sixty, right, 140 00:06:15.680 --> 00:06:18.439 so you went from Okay, now I'm in the community 141 00:06:18.560 --> 00:06:22.199 technic class to now I've got this incubator. Can you 142 00:06:22.240 --> 00:06:24.639 tell us a little bit about what you're doing at Geometer? 143 00:06:25.120 --> 00:06:28.839 Yeah, So some people I used to work with they 144 00:06:28.879 --> 00:06:31.800 started a tech incubator, like all good things, and then 145 00:06:31.879 --> 00:06:33.959 COVID hit. The most interesting part of actually what we 146 00:06:33.959 --> 00:06:36.160 were doing was we ended up pivoting towards doing COVID 147 00:06:36.160 --> 00:06:38.680 response work, and a lot of the work we ended 148 00:06:38.720 --> 00:06:42.360 up deploying for several states was actually built on Elixer interntline. 149 00:06:42.480 --> 00:06:46.639 So we actually have data transformation platforms and pipelines. A fact, 150 00:06:46.800 --> 00:06:49.480 we used Broadway for one of them. Two that are 151 00:06:49.639 --> 00:06:52.680 currently supporting contact tracing in two states on the East coast, 152 00:06:53.160 --> 00:06:56.720 so we learned a lot about both the scalability and functionality. 153 00:06:56.800 --> 00:06:59.120 So just to give you an idea, they were caught 154 00:06:59.120 --> 00:07:02.439 in batch world. Their best turnaround time started off at 155 00:07:02.480 --> 00:07:04.759 twenty four hours, and then when they were really trying 156 00:07:04.800 --> 00:07:06.759 to get up to speed, they could do it maybe 157 00:07:06.759 --> 00:07:08.720 in six but there were a lot of errors and bugs. 158 00:07:09.000 --> 00:07:12.079 We were actually able to in a week spike and 159 00:07:12.120 --> 00:07:15.279 prototype something and then in about another week actually get 160 00:07:15.399 --> 00:07:19.160 into deployment a solution that using Broadway and the inherent 161 00:07:19.199 --> 00:07:23.079 sort of scalability of OTP, that turned the Literally we 162 00:07:23.079 --> 00:07:25.560 were able to process the data faster than they could 163 00:07:25.600 --> 00:07:27.800 get it to us. So we went from processing and 164 00:07:27.800 --> 00:07:30.920 turning around all the COVID cases in about every couple 165 00:07:30.959 --> 00:07:34.079 hours to one minute and thirty seconds was our first test, 166 00:07:34.079 --> 00:07:36.160 and we got it down to sub one minute, and 167 00:07:36.199 --> 00:07:38.279 they were doing it incrementally. We were actually able to 168 00:07:38.279 --> 00:07:42.319 reprocess every single test, every single run, every single time, 169 00:07:42.399 --> 00:07:44.639 which allowed us to do a different type of quality control, 170 00:07:44.800 --> 00:07:47.160 which really matters for deduplication because a lot of the 171 00:07:47.639 --> 00:07:49.720 issues that you run into with COVID testing are that 172 00:07:49.800 --> 00:07:51.920 the lag between when people get tests and the test 173 00:07:51.920 --> 00:07:55.040 results are reported. I mean that duplicates and false positives 174 00:07:55.079 --> 00:07:58.480 are a real problem. So that was all done on 175 00:07:58.839 --> 00:08:02.240 Elixir and earline, right, And that was one of those things. 176 00:08:02.240 --> 00:08:04.480 By the way, we also wrote a telyphony shim that 177 00:08:04.560 --> 00:08:07.839 actually did calliphony control, that did call control for connecting 178 00:08:07.879 --> 00:08:11.600 calls to contact tracers, also done in arline, and both 179 00:08:11.639 --> 00:08:13.879 of them, if I remember correctly, I might be making 180 00:08:13.879 --> 00:08:15.439 this up a little bit, but both of them required 181 00:08:15.519 --> 00:08:18.600 so few resources that they ran in like a container 182 00:08:18.879 --> 00:08:21.480 inside of like an easy two medium, right. It was 183 00:08:21.560 --> 00:08:24.519 just so ridiculously efficient that it was kind of kind 184 00:08:24.519 --> 00:08:28.120 of sad, and we replaced a bunch of custom made 185 00:08:28.079 --> 00:08:30.199 over this Lambda functions that were doing this that had 186 00:08:30.240 --> 00:08:32.559 to do with somewhere aer prone. I think the Lamba 187 00:08:32.600 --> 00:08:35.759 function loads were a couple hours, like fourteen or fifteen 188 00:08:35.799 --> 00:08:38.399 hours to reload the entire universe of data just because 189 00:08:38.639 --> 00:08:41.000 of some architectural decisions. But the short version is Alexra 190 00:08:41.120 --> 00:08:42.720 was a big win because three people were able to 191 00:08:42.720 --> 00:08:46.279 rebuild that system in like a week. So long week, 192 00:08:46.360 --> 00:08:48.279 mind you, but it was really kind of impressive. That 193 00:08:48.279 --> 00:08:50.440 was one of those really aha moments. When you can 194 00:08:50.480 --> 00:08:54.159 think about Elixir and Arline and Beam are very different 195 00:08:54.159 --> 00:08:56.440 and they provide almost like what people or devs think 196 00:08:56.440 --> 00:08:57.360 of as superpowers. 197 00:08:57.679 --> 00:09:00.200 So I feel like I've lived the same exact story 198 00:09:00.240 --> 00:09:04.480 of Lamb does being painful and terribly slow, and then 199 00:09:04.559 --> 00:09:06.879 replacing it with an Elixir app and then it was, 200 00:09:07.120 --> 00:09:09.039 you know, like a whole different world entirely. 201 00:09:09.440 --> 00:09:09.679 Yeah. 202 00:09:10.000 --> 00:09:11.840 Actually, have you ever see that T shirt that says 203 00:09:11.840 --> 00:09:14.399 go away, I'll replace you with a small shell script. Like, 204 00:09:14.840 --> 00:09:17.080 you know, there's like this comic. I think it was 205 00:09:17.120 --> 00:09:18.960 floating around for a while, but it's I almost feel 206 00:09:19.000 --> 00:09:21.759 like the exact same thing. It's like almost any use 207 00:09:21.759 --> 00:09:23.480 case where you're using a Lamb to function, you can 208 00:09:23.639 --> 00:09:27.120 use Elixir and it just kind of works. It's kind 209 00:09:27.120 --> 00:09:27.519 of funny. 210 00:09:27.559 --> 00:09:29.759 I mean, a Lamb is effected with like one gen 211 00:09:29.879 --> 00:09:31.840 server and we can spawn like you know, thousands of 212 00:09:31.919 --> 00:09:33.279 thousands of them, no problem. 213 00:09:33.440 --> 00:09:36.200 So are you guys saying this? And the question I 214 00:09:36.240 --> 00:09:39.320 have is why isn't everyone doing this? Why is Elix 215 00:09:39.399 --> 00:09:42.360 or not like the language that everyone builds all the things. 216 00:09:42.600 --> 00:09:46.399 I mean, yeah, obviously it's the right soul for every 217 00:09:46.519 --> 00:09:49.440 job there. That's a great question. It has nothing to 218 00:09:49.480 --> 00:09:50.159 do with technology. 219 00:09:50.159 --> 00:09:51.639 I'd love to hear what you guys think, and then 220 00:09:51.720 --> 00:09:55.639 I'll opine with my own slightly derisive, sarcastic view. 221 00:09:55.480 --> 00:09:59.879 On this and and the darkness behind them, right, one 222 00:10:00.120 --> 00:10:04.039 language to rule them all, right, So it's it's interesting 223 00:10:04.480 --> 00:10:06.679 I'm starting to play with with a lot of different 224 00:10:06.720 --> 00:10:09.960 languages that that have different takes in the world. And 225 00:10:09.960 --> 00:10:12.120 so the one that really opened my eyes was Pony 226 00:10:12.799 --> 00:10:15.679 because it does so there are a lot of things 227 00:10:15.679 --> 00:10:18.559 that Elixer does really well that that Pony just really 228 00:10:18.559 --> 00:10:21.360 really sucks at, right, But one of the things that 229 00:10:21.399 --> 00:10:28.039 Pony does well is it enforces compiler constraints that enforce 230 00:10:28.200 --> 00:10:33.679 concurrency constraints, right, so incorrect programs won't compile. And so 231 00:10:34.200 --> 00:10:37.360 you know that that pointed me to to the to 232 00:10:37.480 --> 00:10:40.399 the idea that if you have to move a lot 233 00:10:40.440 --> 00:10:43.000 of data around, you know, if you have a lot 234 00:10:43.039 --> 00:10:47.120 of commutability, then then Elixor isn't it. And so I've 235 00:10:47.159 --> 00:10:49.159 been spending a little bit of time in Julia this 236 00:10:49.279 --> 00:10:52.759 year just kind of to flux the old math muscles. 237 00:10:52.799 --> 00:10:52.960 Right. 238 00:10:54.360 --> 00:10:57.279 So I was a math major in college or computer 239 00:10:57.360 --> 00:10:59.600 science and math, and so one of the things that 240 00:10:59.600 --> 00:11:04.000 that we were focused on was linear algebra I'm sorry, 241 00:11:04.480 --> 00:11:08.759 like matrix matrix algebra. And that's kind of what machine 242 00:11:08.840 --> 00:11:12.559 learning and machine modeling have have turned into, right. It's 243 00:11:12.600 --> 00:11:16.360 it's you have these these massive systems of equations and 244 00:11:16.360 --> 00:11:20.600 and algorithms to kind of slowly crank and provide more 245 00:11:20.600 --> 00:11:23.600 and more precision. And we're starting to get into the 246 00:11:23.639 --> 00:11:27.320 places where where you could take these these massive tables 247 00:11:27.799 --> 00:11:32.440 and apply these differential equations to them. And Elixir is 248 00:11:32.519 --> 00:11:34.519 just not good at that, and it is never going 249 00:11:34.559 --> 00:11:37.320 to be good at that. It's good at kind of 250 00:11:37.360 --> 00:11:40.200 the control. But yet, Yeah, I mean I think that 251 00:11:40.200 --> 00:11:42.240 that there are a lot of good things that Elixer 252 00:11:42.360 --> 00:11:45.720 is great at. It's funny, Randall, it's Elixer is great 253 00:11:45.799 --> 00:11:48.320 at almost all the things that we were doing. 254 00:11:49.240 --> 00:11:52.879 Where you're at these early Ruby communities, right, the deployment, 255 00:11:53.039 --> 00:11:57.039 the you know, the the babysit big fat relational database 256 00:11:57.120 --> 00:11:59.639 with a web UI, you know, on and on. 257 00:12:00.360 --> 00:12:02.759 Yeah. No, so I don't know, what do you get? 258 00:12:02.799 --> 00:12:04.480 Why do you guys think it's not winning? I mean, 259 00:12:04.679 --> 00:12:06.840 I guess the other question is what's the definition of winning. 260 00:12:07.200 --> 00:12:11.440 I don't think universal adoption of everything is the definition 261 00:12:11.480 --> 00:12:13.879 of winning. I don't think it's for everyone. Not all 262 00:12:13.960 --> 00:12:15.039 languages are for everyone. 263 00:12:15.440 --> 00:12:17.600 I have a question, how do we know it's not winning? 264 00:12:17.720 --> 00:12:17.919 Right? 265 00:12:18.320 --> 00:12:21.799 The language popularity? And so they make mistakes, right, Oh yeah, 266 00:12:21.840 --> 00:12:22.279 didn't they? 267 00:12:22.919 --> 00:12:27.519 Actually what was it? There's something that they included, like crystal, 268 00:12:27.519 --> 00:12:29.480 that's what it was, right, And they ended up playing 269 00:12:29.519 --> 00:12:29.879 like new. 270 00:12:29.840 --> 00:12:34.480 Age chygramming crystals, which is either a programming model or 271 00:12:35.159 --> 00:12:38.440 a great way to have your New Age crystals learn 272 00:12:39.039 --> 00:12:42.440 as you get more enlightened every day, I. 273 00:12:42.519 --> 00:12:45.639 Have to look into that. So no, So I think 274 00:12:45.639 --> 00:12:48.279 there are lots of things. It's really funny. So one 275 00:12:48.279 --> 00:12:50.639 of the things I learned from being a rubyist early on, 276 00:12:51.000 --> 00:12:52.679 and at the time I got into Ruby, I was 277 00:12:52.720 --> 00:12:55.799 actually an independent consultant and I was all in dot net, 278 00:12:56.080 --> 00:12:59.200 like I had actually started a long time ago. System 279 00:12:59.240 --> 00:13:02.320 five released three, that's how old I am. But I 280 00:13:02.360 --> 00:13:03.799 had gotten into that net because that's where all the 281 00:13:03.799 --> 00:13:06.639 consulting money was. So that was C plus plus microsup 282 00:13:06.679 --> 00:13:09.759 foundation classes all that crap, and the learning Aha moment 283 00:13:09.759 --> 00:13:11.480 for me is I had to take some visual basic 284 00:13:11.519 --> 00:13:13.440 programmers and try and teach them C sharp and they 285 00:13:13.440 --> 00:13:16.120 failed miserably, Like there was no way to teach them 286 00:13:16.159 --> 00:13:18.720 the basics of classes, and they just hadn't learned any 287 00:13:18.759 --> 00:13:21.480 of that stuff, and we had to deliver an ETL system. 288 00:13:21.679 --> 00:13:24.840 And Blake Miserini at the time he's like, hey, man, 289 00:13:24.879 --> 00:13:26.759 I know you like Pearl, but I've got something way 290 00:13:26.759 --> 00:13:29.320 better than Pearl. And I apologize for all the Pearl 291 00:13:29.360 --> 00:13:30.919 I ever wrote. I think we all need to do that. 292 00:13:31.399 --> 00:13:33.320 But he said, I got something way better than Pearl. 293 00:13:33.360 --> 00:13:35.799 It's this Ruby thing and like whatever. So I played 294 00:13:35.799 --> 00:13:37.240 around with it for a week and I'm like, huh, 295 00:13:37.240 --> 00:13:40.000 this is pretty neat, and I'm pretty sure at the 296 00:13:40.039 --> 00:13:42.960 time I actually had to go to like a bookstore 297 00:13:42.960 --> 00:13:45.639 and find the Ruby book, like an actual print version 298 00:13:45.679 --> 00:13:49.159 of the book. This was, you know, so like he 299 00:13:49.200 --> 00:13:51.879 had to like run around to borders and whatever and 300 00:13:52.039 --> 00:13:55.440 try and find a copy of Dave Thomas's book. And 301 00:13:55.799 --> 00:13:57.799 I did over the weekend, and then I tried it 302 00:13:57.799 --> 00:14:00.519 with these guys on one day and then I had 303 00:14:00.519 --> 00:14:03.279 them doing advanced string processing and classes. They just got 304 00:14:03.320 --> 00:14:05.399 it and rocked it. And that was the moment I 305 00:14:05.440 --> 00:14:08.519 knew that, like, oh yeah, that's where we need to go, 306 00:14:09.320 --> 00:14:11.759 and so I literally just stopped doing Dottin had stuff 307 00:14:11.759 --> 00:14:15.759 and started doing all in on Ruby and Erlin and Elixir. 308 00:14:16.200 --> 00:14:18.840 And I actually, when I say that earling Elixir Beam, 309 00:14:18.960 --> 00:14:21.759 I actually really like to separate those things because I 310 00:14:21.759 --> 00:14:24.480 think it's important that you don't mix them or blend them, 311 00:14:24.519 --> 00:14:26.840 like there are different things and that you're getting a 312 00:14:26.840 --> 00:14:29.120 lot there. I think it has that same feeling as 313 00:14:29.200 --> 00:14:32.720 early adoption for Ruby. Like literally when we started doing 314 00:14:32.799 --> 00:14:35.600 Ruby work, we had this other company called Quality Humans 315 00:14:35.600 --> 00:14:38.000 where we were all doing rails apps and everybody thought 316 00:14:38.000 --> 00:14:40.000 that we were crazy. You're like, I don't know this 317 00:14:40.120 --> 00:14:43.039 rails thing. I don't know. It seems pretty risky, Like 318 00:14:43.120 --> 00:14:45.000 only do guys in Chicago are doing it? 319 00:14:45.159 --> 00:14:45.360 Yeah. 320 00:14:45.360 --> 00:14:48.720 I mean I think that very often we as a 321 00:14:48.720 --> 00:14:51.679 programming community have to take a step back to take 322 00:14:51.679 --> 00:14:54.320 a step forward, right, And I think that that's one 323 00:14:54.360 --> 00:14:56.799 of the things that would be provided, right. So it 324 00:14:56.840 --> 00:14:58.960 was a huge step forward in a lot of ways, right, 325 00:14:59.039 --> 00:15:02.399 first in the idea that programming language should be intuitive 326 00:15:02.440 --> 00:15:07.000 and fun. Second in that that you didn't have to specify, 327 00:15:07.960 --> 00:15:13.200 it wasn't kind of this this really tedious artificial typing model. 328 00:15:13.559 --> 00:15:17.360 And third, we didn't have these really tedious and artificial 329 00:15:17.399 --> 00:15:21.879 configurations that were you know, XML that was explicit for everything, 330 00:15:22.360 --> 00:15:25.720 with no defaults. All those things were really important for Ruby, 331 00:15:26.240 --> 00:15:30.600 and it seems like by taking that step back to 332 00:15:30.639 --> 00:15:35.240 reset the kind of the bar on for application developers 333 00:15:35.360 --> 00:15:39.639 on the productivity of application developers. You know, we got 334 00:15:39.679 --> 00:15:43.320 a lot of the you know, the deployments, you know, immutability, 335 00:15:43.679 --> 00:15:46.440 a lot of a lot of things wrong, but there 336 00:15:46.519 --> 00:15:49.480 was a whole lot that was right about Ruby that 337 00:15:49.600 --> 00:15:50.600 liks Or learned from. 338 00:15:50.799 --> 00:15:51.480 Yeah. Good. 339 00:15:52.360 --> 00:15:54.559 I like that you mentioned productivity, Bruce, because there's something 340 00:15:54.559 --> 00:15:57.399 that I absolutely feel is true of Alexa and Phoenix. 341 00:15:57.399 --> 00:15:58.960 And then kind of back to what you were saying, Randall, 342 00:15:59.039 --> 00:16:02.519 when you were trying to, you know, move some folks 343 00:16:02.559 --> 00:16:04.759 into kind of this realm of like object oriented and 344 00:16:04.840 --> 00:16:06.919 when you found Ruby and kind of brought that to them, 345 00:16:06.960 --> 00:16:09.080 that people were able to grack it and learn it 346 00:16:09.080 --> 00:16:12.039 and be productive in it relatively quickly. And that's absolutely 347 00:16:12.039 --> 00:16:14.039 true of Alixir. It's absolutely true of Phoenix, and not 348 00:16:14.240 --> 00:16:16.759 just because of you know, Alixir's language features, in the 349 00:16:16.799 --> 00:16:19.399 fact that, similarly to Ruby, no, it is eloquent and 350 00:16:19.440 --> 00:16:21.879 easy to reason about. But I think there's so many 351 00:16:21.919 --> 00:16:26.000 sort of eco Elixir ecosystem features that allow developers and 352 00:16:26.039 --> 00:16:29.120 whole teams to be really highly productive. And I think 353 00:16:29.120 --> 00:16:31.279 that's why I'm not at all surprised to hear how 354 00:16:31.320 --> 00:16:34.000 Alexir and erlang has allowed you guys to solve these 355 00:16:34.039 --> 00:16:37.240 really big scale problems for like COVID contact tracing. It 356 00:16:37.320 --> 00:16:38.919 reminds me a bit of a talk that was one 357 00:16:38.919 --> 00:16:41.519 of the keynotes at Alexa kronk EEO this year, which 358 00:16:41.559 --> 00:16:44.159 will hopefully be out soon, about some folks in Kenya 359 00:16:44.200 --> 00:16:46.440 who are using Elixir to develop a new like nine 360 00:16:46.480 --> 00:16:50.279 to one one emergency contact system around the country. And 361 00:16:50.559 --> 00:16:52.799 it's just it's kind of the perfect combination of like 362 00:16:52.879 --> 00:16:56.039 scale and concurrency and fault tolerance, but also productivity for 363 00:16:56.279 --> 00:16:59.080 actual humans. And so I'd definitely be curious to hear, 364 00:16:59.320 --> 00:17:01.519 you know, from you and with your experiences lately with 365 00:17:01.559 --> 00:17:03.840 some of these COVID projects, like sounds of you guys 366 00:17:03.840 --> 00:17:06.480 spum this up really fast, like a week later you 367 00:17:06.559 --> 00:17:08.960 had these working prototypes. So like, what is it about 368 00:17:09.039 --> 00:17:12.839 alixer specifically or this ecosystem specifically that let you guys 369 00:17:12.880 --> 00:17:13.839 do that so quickly? 370 00:17:14.400 --> 00:17:16.039 So I'm going to give you the first one in 371 00:17:16.039 --> 00:17:17.799 no particular order, but the first one that pops to 372 00:17:17.799 --> 00:17:20.559 mind is people. So it's a really bizarre thing that 373 00:17:20.599 --> 00:17:23.440 I found is that people who are looking for those 374 00:17:23.440 --> 00:17:26.680 exact things that you mentioned, like people for whom tools matter, 375 00:17:26.839 --> 00:17:30.519 like productivity matters. Like this is not to poopoo or 376 00:17:30.559 --> 00:17:32.319 denegrate people who don't choose that, but there are some 377 00:17:32.359 --> 00:17:34.079 people who are like approach things that it's kind of 378 00:17:34.119 --> 00:17:35.720 like a craftsman, and some people who approach it it's 379 00:17:35.759 --> 00:17:37.519 just like this is my job. But there are literally 380 00:17:37.559 --> 00:17:39.400 people who go out and try and learn more because 381 00:17:39.400 --> 00:17:42.759 they want better tools, they want a better experience, like 382 00:17:42.839 --> 00:17:45.519 all those things holistically matter to them. And I find 383 00:17:45.519 --> 00:17:49.359 people who gravitate to the elixir ecosystem generally speaking, care 384 00:17:49.400 --> 00:17:51.480 about those things and as a result it ironically it 385 00:17:51.480 --> 00:17:54.279 makes them much better problem solvers, Like they literally care 386 00:17:54.359 --> 00:17:57.119 about the solution and the elegance solution on I think 387 00:17:57.160 --> 00:18:00.400 Buckminster Fuller had that statement. I might be misattributing the quote, 388 00:18:00.400 --> 00:18:02.680 but it's like, first I start with a solution, but 389 00:18:02.720 --> 00:18:04.319 if at the end it is an elegant, I know 390 00:18:04.359 --> 00:18:06.119 that I haven't solved it in the best way. It's 391 00:18:06.200 --> 00:18:08.359 kind of that's kind of the gist of it. So 392 00:18:08.640 --> 00:18:11.079 we actually because we had people who had experienced in elixir, 393 00:18:11.559 --> 00:18:14.000 they were really good at solving problems. Then the other 394 00:18:14.039 --> 00:18:17.400 thing that really works out is that beam Erling and 395 00:18:17.400 --> 00:18:21.799 Elixer provide great abstractions, Like I really love the fact 396 00:18:21.839 --> 00:18:23.799 that you get to pick to the right level of 397 00:18:23.839 --> 00:18:25.880 abstraction for your problem, and if you need to go 398 00:18:25.960 --> 00:18:28.680 more granular, you can't. If you don't have to go 399 00:18:28.759 --> 00:18:32.000 more granular, no need. Like I wrote C plus plus 400 00:18:32.039 --> 00:18:35.960 for a long time, probably professionally for the better part 401 00:18:36.000 --> 00:18:38.359