WEBVTT 00:00:00.020 --> 00:00:00.980 Every company has one. 00:00:01.330 --> 00:00:03.980 The little internal tool that Jane built back in 2021, 00:00:04.970 --> 00:00:05.820 and then Jane left. 00:00:06.220 --> 00:00:07.120 Nobody understands it. 00:00:07.480 --> 00:00:08.220 Nobody touches it. 00:00:08.620 --> 00:00:11.960 There are two unwritten rules around this internal software. 00:00:12.640 --> 00:00:15.740 First, if it's working, don't change it. 00:00:16.440 --> 00:00:19.280 Second, if you break it, you bought it. 00:00:19.740 --> 00:00:21.560 That's Dark Matter Enterprise software. 00:00:21.760 --> 00:00:23.680 For every app that you can actually see, 00:00:23.920 --> 00:00:26.680 there are tens of these sitting in the shadows, frozen. 00:00:27.300 --> 00:00:29.140 Michael Booth thinks that just changed. 00:00:29.400 --> 00:00:34.620 He read my article on hyper-personal software and ran with it, writing about hyper-team software. 00:00:34.940 --> 00:00:39.500 Small teams inside big companies finally building the tools that were never going to get built, 00:00:39.880 --> 00:00:42.380 or modernizing the ones trapped in the 90s. 00:00:42.760 --> 00:00:47.940 We cover where this works, where it quietly goes wrong, and the guardrails that keep it from turning into a mess. 00:00:48.120 --> 00:00:53.540 Let's get into it. This is Talk Python To Me, episode 558, recorded July 15, 2026. 00:01:11.600 --> 00:01:15.920 Welcome to Talk Python To Me, the number one Python podcast for developers and data scientists. 00:01:16.440 --> 00:01:21.760 This is your host, Michael Kennedy. I'm a PSF fellow who's been coding for over 25 years. 00:01:22.400 --> 00:01:23.460 Let's connect on social media. 00:01:23.580 --> 00:01:27.000 You'll find me and Talk Python on Mastodon, BlueSky, and X. 00:01:27.180 --> 00:01:29.100 The social links are all in your show notes. 00:01:29.880 --> 00:01:33.400 You can find over 10 years of past episodes at talkpython.fm. 00:01:33.660 --> 00:01:36.760 And if you want to be part of the show, you can join our recording live streams. 00:01:37.100 --> 00:01:37.600 That's right. 00:01:37.800 --> 00:01:41.060 We live stream the raw uncut version of each episode on YouTube. 00:01:41.640 --> 00:01:46.060 Just visit talkpython.fm/youtube to see the schedule of upcoming events. 00:01:46.240 --> 00:01:49.940 Be sure to subscribe there and press the bell so you'll get notified anytime we're recording. 00:01:50.740 --> 00:01:53.280 This episode is sponsored by Sentry's Seer. 00:01:53.580 --> 00:01:55.920 If you're tired of debugging in the dark, give Seer a try. 00:01:56.580 --> 00:01:58.620 There are plenty of AI tools that help you write code, 00:01:58.880 --> 00:02:01.780 but Sentry's Seer is built to help you fix it when it breaks. 00:02:02.380 --> 00:02:06.640 Visit talkpython.fm/sentry and use the code talkpython26, 00:02:07.100 --> 00:02:10.240 all one word, no spaces, for $100 in Sentry credits. 00:02:11.000 --> 00:02:14.280 Talk Python and Python Bytes both now have MCP servers. 00:02:14.820 --> 00:02:19.200 Point your AI at 10 plus years of Python episodes, transcripts, and show notes. 00:02:19.640 --> 00:02:19.780 Free. 00:02:20.340 --> 00:02:24.400 click mcp in the nav at talkpython.fm and at pythonbytes.fm 00:02:25.320 --> 00:02:29.580 Michael welcome to Talk Python To Me great to have you here it's a great name I think this is the 00:02:29.600 --> 00:02:34.320 the Michael square there's a lot of things we can do with this is going to be really fun I think 00:02:34.540 --> 00:02:38.200 Michael squared is the most I don't think I've ever had a Michael cubed on the show with three 00:02:38.400 --> 00:02:42.780 Michaels but yeah it's Michaels across the board or it could be Michaels all the way down like 00:02:42.920 --> 00:02:46.700 turtles yeah it's going to be fun whatever it is and thanks for having me you know what I saw your 00:02:46.720 --> 00:02:52.960 article. And I thought, oh, we got to talk. So you wrote an article called Hyper Team Software. 00:02:53.460 --> 00:02:58.080 And right before that, I wrote a blog post, an article called Hyper Personal Software. 00:02:58.660 --> 00:03:02.340 And we were both, I could tell we were both thinking kind of in the same ways and 00:03:02.580 --> 00:03:07.240 your article riffed off of mine. And so I thought, well, let me try to do a little bit of like 00:03:07.480 --> 00:03:13.000 technical jujitsu here. You wrote an article off of mine and I'm going to have you on the show to 00:03:12.980 --> 00:03:17.940 talk about your article and like I said Michael's all the way down. Your article really resonated 00:03:18.340 --> 00:03:23.180 when I had a look at it the first time. Yeah well thanks. We're going to talk about we've got a 00:03:23.310 --> 00:03:28.120 couple of names here so mine was hyper personal software because I wrote this for me. Yours is 00:03:28.200 --> 00:03:33.860 hyper team software because the same concept applied to the enterprise honestly is way more 00:03:34.080 --> 00:03:37.200 powerful than anything that I'm doing although I'll talk about some of the things I'm doing and people 00:03:37.320 --> 00:03:42.940 can decide that but the potential for what you can do to make companies better and teams work 00:03:42.960 --> 00:03:49.200 better is tremendous. And I honestly feel like it's an opportunity squandered for a lot of folks. So 00:03:49.620 --> 00:03:54.280 hopefully you can inspire them to, you know, crack the door open a little bit and see what they can 00:03:54.290 --> 00:03:59.040 do. Coding assistants are changing the world pretty quickly, right? So in good ways and bad. 00:03:59.180 --> 00:04:05.860 In good ways and bad and over and over again, right? For example, six months ago to today could 00:04:06.000 --> 00:04:10.100 be a totally different experience of what's, how things work or how well they work or what's 00:04:10.120 --> 00:04:15.060 possible and so on. So yeah, it's just, I don't know. It's a weird time, Michael. It's a weird 00:04:15.220 --> 00:04:18.739 time. It is a weird time, especially when you've been doing it for a while. It's very weird. 00:04:19.120 --> 00:04:24.000 Exactly. I mean, I've been programmed for over 25 years in tech and I feel like you must be in a 00:04:24.240 --> 00:04:30.860 similar time journey there. And I don't know, weird. Just so many times I've seen technology 00:04:31.080 --> 00:04:35.440 come up that, oh, this is the thing that's going to kill programming. This is going to be the end 00:04:35.460 --> 00:04:42.940 of programmers, UML, drag and drop, visual basic, low code, no code, et cetera. And all those things 00:04:42.960 --> 00:04:47.700 are fine, but they're absolutely not the end of software. I don't think coding agents are either, 00:04:48.060 --> 00:04:52.760 but they're certainly the biggest dent that they can put into software is anything that's come along 00:04:52.960 --> 00:04:57.480 before it. Yeah. The speed and scale is sort of interesting, but sometimes you can just do a lot 00:04:57.480 --> 00:05:02.120 of bad things much faster. Yeah. I mean, that's a big debate, right? Is it, is, is our agentic 00:05:02.140 --> 00:05:07.640 coding tools. Are these amplifiers or are they skilled little separate things that do their own 00:05:07.840 --> 00:05:11.840 work? I think there's a case to be made for amplification to a large degree. Before we get 00:05:11.900 --> 00:05:15.320 into all that, let's just hear a little bit of background about you. Tell us about yourself. 00:05:15.760 --> 00:05:23.080 I started off sort of academic life as doing a PhD in computational chemistry. I've worked in 00:05:23.280 --> 00:05:28.940 quantitative finance for a long, long time. More recently, I started the data booth consultancy 00:05:28.960 --> 00:05:34.380 things. So just the brand for me working as a single consultant and I've been doing that sort 00:05:34.420 --> 00:05:39.020 of coming out of COVID, I decided to do something different. So yeah, about the last five years, 00:05:39.140 --> 00:05:43.380 I've been doing that. Computational chemistry, huh? Are you familiar with a program called 00:05:43.620 --> 00:05:49.000 HyperChem? I think I am actually, that's pretty old. We're talking 90s. I mean, I think we're 00:05:49.100 --> 00:05:54.860 talking 90s. Yes, definitely 90s. So I finished my PhD in late 90s. I vaguely remember the tool, 00:05:55.140 --> 00:05:57.600 But I vaguely remember chemistry now as well. 00:05:57.780 --> 00:06:00.840 I studied a lot of chemistry and I loved organic chemistry. 00:06:00.960 --> 00:06:05.560 And I found this program called HyperChem and it was just so incredible. 00:06:05.620 --> 00:06:09.100 It was like the Google Earth equivalent of complex chemistry. 00:06:09.440 --> 00:06:11.660 And wow, was it a neat thing to play with. 00:06:11.660 --> 00:06:14.300 And I figured maybe it's like right around a similar time frame. 00:06:14.540 --> 00:06:16.840 Yeah, I do remember it now that you mentioned it. 00:06:17.060 --> 00:06:20.680 Organic chemistry was my most unfavorite element of chemistry. 00:06:21.100 --> 00:06:25.920 That's why I did theoretical because experiments are much easier to control on a computer than real life. 00:06:26.260 --> 00:06:28.380 And organic chemistry smells bad. 00:06:28.460 --> 00:06:28.960 It does. 00:06:29.520 --> 00:06:36.940 Honestly, I would tell you that probably is the reason that I didn't work in chemistry is like I'd get headaches all the time from like the super volatile chemicals and stuff. 00:06:37.120 --> 00:06:40.300 I'm like, if I did this every day, I don't know what this would do to me, but it wouldn't be good. 00:06:40.460 --> 00:06:41.320 So let's go do math. 00:06:42.160 --> 00:06:43.460 Math doesn't hurt physically. 00:06:44.820 --> 00:06:47.100 Yeah, I did math and computing in a chemistry school. 00:06:47.320 --> 00:06:50.100 Give us a bit of information about Data Booth. 00:06:50.160 --> 00:06:55.420 You've got some nice articles there. You've got some consulting. Yeah, just tell us about DataBooth before we move on. 00:06:55.620 --> 00:07:00.860 As I said, coming out of COVID, I decided to do something different. I've been working for others for a long time. 00:07:01.060 --> 00:07:08.480 So I thought, and I'd also directly before this been reviewing a lot of models and things that are used in finance. 00:07:08.860 --> 00:07:14.360 So I thought it'd be nice to actually build some things again. I sort of identified data science as a place to hang out. 00:07:14.580 --> 00:07:21.320 I did a boot camp in Oakland, actually, in 2017, and just decided this was sort of a growth area. 00:07:21.530 --> 00:07:24.500 I think that conclusion was correct, as it turns out. 00:07:24.870 --> 00:07:27.460 Working as a single consultant can not always be easy. 00:07:27.690 --> 00:07:29.160 Being seen are tricky at times. 00:07:29.580 --> 00:07:31.360 I still wanted to just have a go. 00:07:31.590 --> 00:07:33.740 It's the first time I've done something on my own. 00:07:34.010 --> 00:07:36.160 And I've had a chance to do some interesting things. 00:07:36.380 --> 00:07:40.200 So I worked with a lady who does experiments with horses, for example. 00:07:40.700 --> 00:07:45.760 And she trying to understand when a horse is stressed, for example. 00:07:46.260 --> 00:07:48.840 So she had a Raspberry Pi based experiment. 00:07:49.260 --> 00:07:53.560 I wrote code to control the experiment and then I wrote the code to analyze the results. 00:07:53.860 --> 00:07:57.940 So I guess being small, you can make choices to do small projects. 00:07:58.360 --> 00:08:02.020 And for me, that was helping out somebody doing fundamental research. 00:08:02.520 --> 00:08:04.680 And I guess it got me back to my roots a bit. 00:08:05.020 --> 00:08:06.460 So I've been able to do something like that. 00:08:06.940 --> 00:08:09.580 And then I've been able to do stuff in finance. 00:08:09.980 --> 00:08:15.400 Again, I worked for a small media company at one point as well, helping them do some analysis. 00:08:15.700 --> 00:08:20.140 So it's been fun, but it's also, as I say, it can be hard to be seen when you're small. 00:08:20.700 --> 00:08:26.260 I'm fortunate I have a wife who is happy to give me some flexibility around those things. 00:08:26.680 --> 00:08:29.580 So yeah, shout out to her as being a great support. 00:08:29.980 --> 00:08:32.400 I think the connection's a little bit up and down. 00:08:32.440 --> 00:08:37.360 I think it's super important to have a spouse who is understanding of those kinds of things, 00:08:37.479 --> 00:08:39.140 because it lets you do really, really interesting things. 00:08:39.280 --> 00:08:42.460 Like I can say the same thing for my family as well in a slightly different way. 00:08:42.880 --> 00:08:44.460 What a cool story about the horse. 00:08:45.000 --> 00:08:49.240 You know, it's mixing biology and IoT and tech. 00:08:49.460 --> 00:08:53.640 And that's the kind of stuff that you learn a whole bunch in a real short period of time. 00:08:53.920 --> 00:08:55.660 But also you can make a big difference, you know. 00:08:55.740 --> 00:08:56.300 I learned a heap. 00:08:56.460 --> 00:08:59.360 I'd never worked with a Raspberry Pi, for example. 00:08:59.620 --> 00:09:04.120 So it's quite interesting when you don't play in a space, you don't realize just how 00:09:04.640 --> 00:09:05.840 far certain things have advanced. 00:09:06.120 --> 00:09:09.320 So it's pretty cool what you can do on very small devices these days. 00:09:09.520 --> 00:09:10.180 That's Data Booth. 00:09:10.250 --> 00:09:10.500 Very cool. 00:09:10.660 --> 00:09:11.400 People can check that out. 00:09:11.740 --> 00:09:16.420 Let's take it out a bit and talk about sort of internal company software. 00:09:16.920 --> 00:09:23.120 There's a lot of people listen to this show who are not what you would call enterprise developers, right? 00:09:23.260 --> 00:09:24.240 They are students. 00:09:24.960 --> 00:09:31.420 They're maybe data scientists kind of on their own or they work at a research or they work at a startup. 00:09:32.000 --> 00:09:36.660 So give us a sense of what internal software, like these bigger companies, looks like today 00:09:36.900 --> 00:09:39.880 and like maybe before the AI world and so on. 00:09:39.940 --> 00:09:41.360 Every company is different. 00:09:41.560 --> 00:09:43.740 And I'm sure you've experienced that as well. 00:09:44.140 --> 00:09:46.120 Lots of different operating models. 00:09:46.640 --> 00:09:51.580 But I guess enterprises always need to make sure that things are orderly to some degree. 00:09:52.100 --> 00:09:58.600 So when I work on my MacBook as a solo practitioner, I can make decisions about what software I 00:09:58.660 --> 00:10:00.540 want to use, how I want to use it. 00:10:00.860 --> 00:10:05.040 Ideally, when I'm working with data, you've got to make sure that you've got certain controls in place. 00:10:05.210 --> 00:10:07.400 But within companies, those things really matter, right? 00:10:07.660 --> 00:10:14.980 So you do need to make sure that things are secured appropriately, that people only have access to the things that you want them to. 00:10:15.340 --> 00:10:21.680 So I think in that context, there's always going to be sort of overheads and friction on developing any sort of software. 00:10:22.390 --> 00:10:25.900 So you've just got to be aware that you need to allow for those things. 00:10:26.320 --> 00:10:39.280 So I think the idea that we'll talk about coming up is the sense that when you're actually working within a team, within an enterprise, maybe there's certain things that you can do that don't need to impact too many other people. 00:10:39.720 --> 00:10:43.240 And therefore, you can achieve a greater degree of autonomy. 00:10:43.660 --> 00:10:47.160 There's certainly the things that you need to be thinking about within the enterprise. 00:10:47.380 --> 00:10:48.760 So it's not choose your own adventure. 00:10:49.080 --> 00:10:52.180 It's a work within the constraints which are appropriately there. 00:10:52.580 --> 00:10:57.900 And particularly between industries, you can have some industries which have little regulation. 00:10:58.230 --> 00:11:02.340 And then in other industries, they're highly regulated for very good reasons. 00:11:02.860 --> 00:11:10.240 But again, that can provide an additional layer of complexity as you start to think about how you develop and what you develop. 00:11:10.340 --> 00:11:12.740 I've not spent a ton of time working at large companies. 00:11:12.890 --> 00:11:16.980 I worked at a company for maybe six months that had 1,200 people. 00:11:17.520 --> 00:11:21.900 And I got there by being acquired from a company that had 10 full-time employees. 00:11:21.940 --> 00:11:27.260 So it was a little bit of a shift, but I did do professional software training for a long 00:11:27.440 --> 00:11:27.580 time. 00:11:28.050 --> 00:11:31.360 So I taught week-long courses over a hundred different companies. 00:11:31.660 --> 00:11:34.660 And, you know, you just, you get a look inside these places. 00:11:34.780 --> 00:11:36.920 You're like, wow, this is, this one is really different. 00:11:37.160 --> 00:11:40.060 This is like really unique in this way or that. 00:11:40.140 --> 00:11:45.100 But the, one of the, I mean, maybe not with the tech companies, but maybe even so, but 00:11:45.160 --> 00:11:49.680 like for many of the companies, I would say there's, there's two golden rules of a lot 00:11:49.680 --> 00:11:50.400 of this kind of stuff. 00:11:50.800 --> 00:11:52.700 One of them is don't change it. 00:11:53.120 --> 00:11:53.480 It's working. 00:11:53.960 --> 00:11:54.640 Just don't change it. 00:11:54.940 --> 00:11:55.520 You know what I mean? 00:11:55.780 --> 00:11:56.840 And there's something to that, right? 00:11:57.140 --> 00:11:59.080 If it is working, why would you fix it? 00:11:59.460 --> 00:12:01.640 It can be very expensive to change things, right? 00:12:01.720 --> 00:12:01.840 Yeah. 00:12:01.940 --> 00:12:04.060 So maybe we'll call that the principle of least touching. 00:12:04.300 --> 00:12:05.080 Like, just don't touch it. 00:12:05.400 --> 00:12:06.260 It's just leave it. 00:12:06.460 --> 00:12:06.940 It's over there. 00:12:07.060 --> 00:12:07.720 Just don't touch it. 00:12:08.020 --> 00:12:09.820 And that leads to like you're talking about on your MacBook. 00:12:10.120 --> 00:12:14.200 Like, oh, I could use like Python 3.14 and we could pick, oh, let's try FastAPI. 00:12:14.300 --> 00:12:15.140 And they're like, that's cute. 00:12:15.240 --> 00:12:15.960 We use Python 2. 00:12:16.380 --> 00:12:17.320 You're like, huh, really? 00:12:17.720 --> 00:12:17.860 Okay. 00:12:18.080 --> 00:12:18.680 You don't want to change it. 00:12:18.760 --> 00:12:18.820 No. 00:12:19.339 --> 00:12:20.080 Principle of least touching. 00:12:20.240 --> 00:12:20.860 We don't touch this. 00:12:21.070 --> 00:12:21.380 It works. 00:12:21.620 --> 00:12:25.860 When I was in a previous company, they were using 3.7, for example. 00:12:26.060 --> 00:12:27.840 I'm like, yeah, but it's out of support. 00:12:28.860 --> 00:12:29.880 It's end of life. 00:12:30.420 --> 00:12:31.360 Aren't you going to upgrade things? 00:12:31.560 --> 00:12:33.440 And people just stared at me. 00:12:33.440 --> 00:12:34.660 It was like, why would we? 00:12:35.140 --> 00:12:39.400 So, yeah, I guess it's the what you should do on paper versus what you do in reality. 00:12:39.520 --> 00:12:41.180 Both of those ways are right, I think. 00:12:41.300 --> 00:12:43.780 Like, you maybe don't want to mess it up because it is working. 00:12:44.460 --> 00:12:49.520 But at the same time, we had things like Log4J and other stuff where, like, there are really big problems. 00:12:49.820 --> 00:12:51.940 And it's just because people didn't want to update their software. 00:12:52.420 --> 00:12:57.460 So the other core principle, I think, is you break it, you bought it. 00:12:57.710 --> 00:13:02.680 And that's like, if you do touch it and it was working and now it's not working, this 00:13:02.710 --> 00:13:03.520 is now your problem. 00:13:03.630 --> 00:13:06.740 And maybe the person who knows how it works doesn't even work here anymore. 00:13:07.100 --> 00:13:11.020 The person that did it almost certainly does not work there anymore, right? 00:13:11.160 --> 00:13:15.900 If you look at the stats, I think sort of typical tenure, at least these days, is less 00:13:15.950 --> 00:13:16.540 than two years. 00:13:16.640 --> 00:13:24.060 So the likelihood that you're going to find anybody in the team that knows anything is, can be just very, very small, right? 00:13:24.160 --> 00:13:24.820 It could be very small. 00:13:24.940 --> 00:13:31.600 And so that leads to just some interesting gaps or in a more half glass, half full way, like opportunities. 00:13:31.940 --> 00:13:39.720 There's so much of these little tools and these little connectors and these things or just software that was never built because you don't want to take it on if you build it. 00:13:39.820 --> 00:13:45.320 and they're just sitting there inside these huge companies that have so much internal software, 00:13:45.490 --> 00:13:50.560 but they just, they're either really outdated or they're just really ugly and they work clunky, 00:13:50.740 --> 00:13:54.460 but people just like the person who built it left and no one else can even build it. 00:13:54.470 --> 00:13:57.620 So like, be thankful it works, not that it's hard to use, you know? 00:13:57.940 --> 00:14:02.100 So there's just so much opportunity for this concept of hyper team software, I think. 00:14:02.280 --> 00:14:06.420 Yeah, no, I think there's lots of opportunities out there, but I think like any problem, 00:14:06.760 --> 00:14:10.320 you want to think through a solution maybe before you start changing anything. 00:14:10.530 --> 00:14:12.480 I think it's very tempting just to jump in. 00:14:12.980 --> 00:14:15.240 And I think, yeah, taking a few steps back. 00:14:15.530 --> 00:14:19.240 And again, just understanding the scope of the problem that you're trying to solve. 00:14:19.620 --> 00:14:23.860 So like most of life, if you choose a smaller problem to start with, 00:14:24.440 --> 00:14:28.020 maybe you learn something from it and it helps you as you start to tackle bigger problems. 00:14:28.220 --> 00:14:33.140 I think it's also just, you know, as an aside for people who are looking to advance their career, 00:14:33.760 --> 00:14:38.000 If you can improve these tools and make life better for everybody at the company, that's 00:14:38.000 --> 00:14:39.720 a really high visibility thing to do. 00:14:39.870 --> 00:14:44.600 It used to be all the data scientists had this super annoying thing that they had to 00:14:44.700 --> 00:14:44.840 do. 00:14:45.270 --> 00:14:46.880 And then you spent two weeks on it. 00:14:46.960 --> 00:14:49.220 And all of a sudden, everybody's happier at work. 00:14:49.400 --> 00:14:50.660 And they know that it was because of you. 00:14:50.760 --> 00:14:53.360 Like that shines a pretty good halo on you. 00:14:53.500 --> 00:14:57.840 No, like there's great opportunities to like make a real impact, I think. 00:14:58.180 --> 00:15:03.760 And for me, coming out of a sort of science background, and I try and see everything as 00:15:03.760 --> 00:15:04.760 an experiment in life. 00:15:05.240 --> 00:15:09.200 And I try and put all of my experiments orderly in GitHub, for example. 00:15:09.800 --> 00:15:15.800 You can move quickly and you can experiment and you can identify the paths not to go down 00:15:16.540 --> 00:15:19.340 so that you can actually work out the paths that are worth pursuing. 00:15:19.840 --> 00:15:23.160 And to your point, yeah, there are opportunities to be seen. 00:15:23.580 --> 00:15:24.520 And you can move from that. 00:15:24.720 --> 00:15:33.080 I just wanted to draw a diagram on a piece of paper and convince you that this was a good idea to, well, here's the working prototype or proof of concept. 00:15:33.460 --> 00:15:38.280 And we actually now can talk about, have a real discussion rather than a hypothetical one. 00:15:38.380 --> 00:15:39.760 That's so interesting that you bring that up. 00:15:39.910 --> 00:15:51.300 Because one of the things that I found really challenging working at bigger companies, even not huge companies, so just in general, is you can describe something and people go, oh, maybe that's nice. 00:15:51.920 --> 00:15:53.100 You can tell them you can do it. 00:15:53.100 --> 00:15:54.540 And they're like, yeah, probably you can. 00:15:54.860 --> 00:16:00.460 But then you just take a little bit of time and you show them like, no, look, it's not I could do this. 00:16:00.630 --> 00:16:02.340 It took me two hours to do this. 00:16:02.760 --> 00:16:04.500 If we keep going on it, it'll be really nice. 00:16:04.610 --> 00:16:07.940 Look, wait, you already kind of did this? 00:16:08.120 --> 00:16:09.740 Like, this is actually really amazing. 00:16:09.980 --> 00:16:16.020 I really think that makes a huge difference in how much trust and sort of adoption you get for like your ideas. 00:16:16.300 --> 00:16:16.740 Absolutely. 00:16:17.060 --> 00:16:24.460 Like in a previous employer, you sort of had the business analysts go, okay, let's draw up wireframes and whatnot. 00:16:24.860 --> 00:16:27.140 And then we can prove out this and do that. 00:16:27.700 --> 00:16:29.340 Like, no, we're not using wireframes. 00:16:29.560 --> 00:16:31.520 We're just building it because it doesn't take very long. 00:16:31.880 --> 00:16:34.340 And then people can see it, see the real thing. 00:16:34.640 --> 00:16:37.360 That's not a critique of the sort of BAs at all. 00:16:37.540 --> 00:16:41.100 It's just the world's changed, I think, from when you used to work like that. 00:16:42.620 --> 00:16:46.200 This portion of Talk Python To Me is brought to you by Sentry and Sear AI. 00:16:46.860 --> 00:16:49.320 There are plenty of AI tools that help you write code, 00:16:49.530 --> 00:16:52.560 but Sentry Seer is built to help you fix it when it breaks. 00:16:53.100 --> 00:16:54.160 The difference is context. 00:16:54.860 --> 00:16:57.040 Seer isn't just guessing based on syntax. 00:16:57.360 --> 00:16:59.400 It's analyzing your actual Sentry data, 00:16:59.760 --> 00:17:01.720 your stack traces, logs, and failure patterns. 00:17:02.280 --> 00:17:03.740 Because it has the full context, 00:17:03.830 --> 00:17:06.860 it can A, spot buggy code in review 00:17:06.930 --> 00:17:08.520 and help prevent issues before they happen, 00:17:09.140 --> 00:17:11.939 and B, identify the root cause of production errors. 00:17:12.880 --> 00:17:15.079 It can even draft a fix and hand the work off 00:17:15.020 --> 00:17:17.439 to an agent-like cursor to open a PR for you. 00:17:18.140 --> 00:17:20.040 Seer turns Sentry into a complete loop. 00:17:20.560 --> 00:17:22.720 You have your traces, errors, logs, and replays 00:17:22.720 --> 00:17:25.180 to see the problem, and now AI to help solve it. 00:17:25.459 --> 00:17:27.920 Join millions of devs at companies like Claude, 00:17:28.380 --> 00:17:29.800 Disney+, and even Talk Python, 00:17:30.220 --> 00:17:31.460 who use Sentry to move faster. 00:17:31.960 --> 00:17:34.700 Check them out at talkpython.fm/sentry 00:17:34.780 --> 00:17:38.160 and use code talkpython26, all one word, 00:17:38.620 --> 00:17:40.700 for $100 in Sentry credits. 00:17:41.220 --> 00:17:43.060 Thank you to Sentry for supporting Talk Python. 00:17:44.140 --> 00:17:49.520 This like, I think leads into where we're going. There's, we're going to build software in a manner 00:17:49.680 --> 00:17:56.200 that a bank could run on it because that's our company philosophy and our culture way of working. 00:17:56.380 --> 00:18:02.520 We build solid software. That's no frills that works versus, yeah, but do you need to do that 00:18:02.620 --> 00:18:06.280 for the log parsing thing? That's just going to help us see if the server's up. Like, no, 00:18:06.320 --> 00:18:09.799 you probably don't, but don't, you know what I mean? Like it's, it's, I think it's hard for some 00:18:09.820 --> 00:18:14.220 of these groups to kind of break away from treating everything as like mission critical 00:18:14.650 --> 00:18:18.240 and i think that's pretty germane to this conversation you know yeah like i think you 00:18:18.380 --> 00:18:23.680 need to work out what mission critical actually means right and whether certain things particularly 00:18:23.850 --> 00:18:29.540 within teams actually are mission critical because that will sort of help guide maybe how you develop 00:18:29.700 --> 00:18:35.140 things and who should be developing them maybe just picking up on the logging comment there if i 00:18:35.100 --> 00:18:39.880 understood it correctly the argument used to be that putting in sort of baseline good practices 00:18:40.180 --> 00:18:45.540 was expensive and i think that's also changed so the effort to put logging into a piece of software 00:18:45.760 --> 00:18:50.420 is like almost trivial these days right so i think there's a whole bunch of things where people go oh 00:18:50.700 --> 00:18:56.200 we couldn't possibly do that or maybe that isn't warranted but if it doesn't really cost you 00:18:56.340 --> 00:19:00.659 anything why wouldn't you put it in like if it's a point pointless piece of functionality then fair 00:19:00.680 --> 00:19:06.420 enough right we don't want we don't want animated gifs in everything but um yeah the there's just 00:19:06.580 --> 00:19:11.140 basics of good software engineering that i think you should put into sort of any tool that you're 00:19:11.360 --> 00:19:16.440 developing that's a good point it used to be you'd build software and you might have some kind of 00:19:16.860 --> 00:19:21.920 team lead review it for security or run bandit against it or something like that and hope that 00:19:21.960 --> 00:19:27.539 it's kind of okay i mean you don't call this function wrong like um print s print f or something 00:19:27.540 --> 00:19:32.220 or, you know, the unsafe version of YAML load, whatever, right? 00:19:32.580 --> 00:19:34.060 Those are the kinds of things that you should get checked. 00:19:34.120 --> 00:19:38.680 But now, I mean, we basically have Mythos as a co-worker. 00:19:39.320 --> 00:19:43.260 You know, if we subscribe to Claude Code or the new ChatGPT stuff or Codex, 00:19:43.640 --> 00:19:45.880 and you can just say, look, do a security review on this. 00:19:46.340 --> 00:19:50.640 And what used to take like two weeks and a $20,000 pen testing engagement 00:19:50.700 --> 00:19:52.540 is now like an hour or less. 00:19:52.680 --> 00:19:53.340 It's pretty crazy. 00:19:53.400 --> 00:19:57.800 I think you, of course, still need qualified experts to review the results, right? 00:19:58.220 --> 00:20:05.980 So you don't want to outsource your thinking, I would say, to a machine that's very good at multiplying and adding matrices together very quickly. 00:20:06.460 --> 00:20:09.020 The machines bring a lot to the table, right? 00:20:09.180 --> 00:20:16.240 But yeah, even if you're doing those sort of reviews, you want to leverage the machine's capability and not just completely rely on it. 00:20:16.250 --> 00:20:19.800 And I think that's one of the things that we talked about good and bad at the start. 00:20:19.920 --> 00:20:24.100 I think that's one of the things where I get slightly concerned when people seem to be 00:20:24.440 --> 00:20:28.420 outsourcing their thinking that's not across the board. Turn off their critical thinking. 00:20:29.040 --> 00:20:33.400 Yeah, 100%. It is interesting. I noticed the University of Washington teaches a course now 00:20:33.520 --> 00:20:37.620 on critical thinking, which is somewhat interesting, right? Why would you need to 00:20:37.860 --> 00:20:42.540 teach people critical thought? Like clearly, clearly you get taught critical thought through 00:20:42.540 --> 00:20:47.460 a whole bunch of experiences and courses in life. But I think it speaks somewhat to the age that 00:20:47.480 --> 00:20:51.860 you'd actually have an official course on that. I guess that's what experience brings. 00:20:52.200 --> 00:20:58.140 Right, right. And the more you can sort of shortcut your way through literature and philosophy and 00:20:58.420 --> 00:21:01.620 math and so on, then you kind of, I guess, need to make it up somewhere else. 00:21:01.790 --> 00:21:06.480 I think experience plays into this a lot, right? So the longer you've been around, 00:21:06.880 --> 00:21:11.340 just the more things that you've seen them gone wrong, right? And that's not unexpected in life, 00:21:11.350 --> 00:21:16.240 I don't think, because we all like to take shortcuts. But I think, as I often say to people, 00:21:16.520 --> 00:21:17.580 Just choose the right shortcuts. 00:21:17.880 --> 00:21:20.560 This one, I had to come in the middle of the night because the server went down. 00:21:20.920 --> 00:21:21.920 We don't want to do this anymore. 00:21:22.460 --> 00:21:24.400 Or the car broke down on the side of the road because of this. 00:21:24.600 --> 00:21:25.360 We're not doing that anymore. 00:21:26.120 --> 00:21:28.040 You learn the lessons over time, I suppose. 00:21:28.360 --> 00:21:30.140 Hopefully you learn the lessons over time. 00:21:31.020 --> 00:21:31.900 That's the goal, I think. 00:21:32.200 --> 00:21:33.540 As I say, everything's an experiment. 00:21:34.040 --> 00:21:35.740 The question is, will you learn from the experiment? 00:21:36.020 --> 00:21:36.480 Maybe you'll learn. 00:21:36.680 --> 00:21:37.140 Not 100%. 00:21:37.220 --> 00:21:37.720 There's no guarantees. 00:21:38.160 --> 00:21:43.159 Before we jump into the details of your article, I do want to put one more concept out there 00:21:43.180 --> 00:21:48.260 for people that I think can sometimes get lost in the fray as well. And that's the concept of 00:21:48.580 --> 00:21:54.100 competing against non-consumption. So for, put it in context for my example of like, oh, we have a 00:21:54.200 --> 00:21:59.280 pen tester now that can work on this thing. You would hire a pen tester for the most important 00:21:59.480 --> 00:22:07.120 part of your software, but your little tiny, like expense tracking tool that Jane from 2021 wrote 00:22:07.200 --> 00:22:11.479 and left the company, you're probably not going to hire a pen tester for that. But now these days 00:22:11.500 --> 00:22:17.280 with these tools, you can, quote, hire a pen tester for it, and it takes half an hour. 00:22:17.840 --> 00:22:21.320 And it's not that, ah, you probably should have hired, you probably should really use 00:22:21.360 --> 00:22:22.120 like a real person. 00:22:22.180 --> 00:22:25.160 Like there was never going to be a pen test of that at all. 00:22:25.580 --> 00:22:26.420 And now there can. 00:22:26.540 --> 00:22:30.500 Like there will be for the real software, maybe, but not for like these little internal tool 00:22:30.680 --> 00:22:32.260 things that like, I think we're going to talk about. 00:22:32.520 --> 00:22:36.760 So I think this concept of competing against non-consumption is worth keeping in mind. 00:22:36.960 --> 00:22:37.320 Good point. 00:22:37.400 --> 00:22:42.300 I guess the question still is, would you, is there a case to actually use the machine for that? If 00:22:42.380 --> 00:22:46.700 it's sort of free, then I guess, why not? But there's still lots of other things that you might 00:22:47.080 --> 00:22:51.640 investigate, right? So do you actually want to spend some of your token budget on that even? 00:22:52.120 --> 00:22:56.540 So I think, again, it's thinking well about problems and going, what are the actual pros 00:22:56.540 --> 00:23:01.839 and cons here? Or what are the risks that we're trying to mitigate? Okay, let's jump into Hyper 00:23:01.860 --> 00:23:09.360 team software and we're going to talk through your excellent essay from hyper personal software to 00:23:09.460 --> 00:23:14.020 hyper team software small team built ai assisted tools inside the enterprise because i like i said 00:23:14.020 --> 00:23:20.700 i think there's just a gold mine of opportunity here to solve some of these principle of least 00:23:20.840 --> 00:23:25.780 touching and you break it as yours sort of type like things that are a hassle for everyone and 00:23:25.780 --> 00:23:29.219 just no one's going to go near it and no one understands it i think there's a huge possibility 00:23:29.240 --> 00:23:34.260 there. So, but since you said, you know, this is, this is based on my concept of hyper-personal, 00:23:34.300 --> 00:23:38.360 let me just give people a really quick, definite, like a starting point from there. So I said like, 00:23:38.440 --> 00:23:43.720 look, I think there's a lot of people that say, if this AI stuff is so good, why don't we see an 00:23:43.800 --> 00:23:48.880 explosion of software? And my article said, I think there is an explosion of the software, 00:23:49.000 --> 00:23:53.800 but so much of it is like dark matter in a sense. Like you can't see it because it's not worth 00:23:54.340 --> 00:23:58.639 bringing out into a great big flash. It's just a bunch of these little things that solve or polish 00:23:58.780 --> 00:24:04.280 a bunch of stuff. So for my example was I was using the start page search engine and they'd 00:24:04.440 --> 00:24:08.880 started putting ads that would fill like two pages before you even got to the first search result. 00:24:09.010 --> 00:24:13.300 Like I don't mind having an ad. I want to support you, but two pages, this is ridiculous, right? 00:24:13.680 --> 00:24:16.960 Yeah. So I said, I just had Claude build me an extension that would just remove, 00:24:17.840 --> 00:24:22.460 remove the ads just for me. I have no intention of ever shipping it. I just thought, here's a 00:24:22.530 --> 00:24:27.900 problem I have. I bet in 20 minutes I can make it go away, you know? So that sets the stage. 00:24:28.020 --> 00:24:33.280 tell us about this concept here like tell us about your idea apologies for the name first of all 00:24:33.430 --> 00:24:38.300 because i don't think it's particularly the right name but i just wanted to riff off your name so 00:24:38.440 --> 00:24:45.600 that was the initial thinking behind it and also apology for the length of my articles i try not 00:24:45.650 --> 00:24:50.940 to write so much but sometimes that's just what happens so yeah i guess as i reflected on your 00:24:51.180 --> 00:24:56.679 article and thought well if michael can build stuff that solves his problem surely people in 00:24:56.920 --> 00:25:03.080 smaller teams within companies actually have plenty of problems to solve and you don't want to be 00:25:03.440 --> 00:25:09.640 necessarily starting up a capital p project and having lots and lots of people involved which is 00:25:09.760 --> 00:25:15.460 sort of a typical model because of the way companies are structured but if you can just yeah solve a 00:25:15.640 --> 00:25:21.800 small problem and it doesn't involve any sort of other people outside your team or any great 00:25:21.820 --> 00:25:26.860 infrastructure so you're not trying to set up a client server database necessarily to solve it 00:25:27.200 --> 00:25:31.920 you're not trying to include sort of authentication because the security of it doesn't really really 00:25:32.080 --> 00:25:37.460 matter in this case you're just solving a small problem within your team then why not take a 00:25:37.500 --> 00:25:41.020 similar approach like i've already beat to death i think there's tons of opportunity here but you 00:25:41.040 --> 00:25:46.440 know give me your sense of this at my limited experience i would say for everything that you 00:25:46.620 --> 00:25:50.400