WEBVTT 00:00:00.020 --> 00:00:04.740 Hello and welcome to Python Bytes, where we deliver Python news and headlines directly to your earbuds. 00:00:05.180 --> 00:00:11.720 This is episode 465, recorded January 12, 2026, and I'm Brian Okken. 00:00:12.020 --> 00:00:12.920 And I am Michael Kennedy. 00:00:13.360 --> 00:00:18.260 And this episode is sponsored by us and you. 00:00:18.610 --> 00:00:24.359 So thanks to everybody that does support us through Patreon supporters 00:00:24.720 --> 00:00:27.900 and also through grabbing the pytest course 00:00:28.380 --> 00:00:29.960 or the Lean TDD book 00:00:30.200 --> 00:00:33.360 or so many of the massively cool courses 00:00:33.780 --> 00:00:35.020 at Talk Python Training 00:00:35.160 --> 00:00:39.380 and also Michael's fun book on deployment 00:00:39.740 --> 00:00:40.620 and stuff like that. 00:00:40.940 --> 00:00:42.920 So if you'd like to, 00:00:43.080 --> 00:00:45.180 we're recording this for the podcast, 00:00:45.520 --> 00:00:47.220 but also broadcasting it live. 00:00:47.420 --> 00:00:49.160 If you'd like to join the live, 00:00:49.220 --> 00:00:50.260 you can join us at YouTube, 00:00:50.560 --> 00:00:53.839 but you can just go to pythonbytes.fm/live 00:00:54.120 --> 00:00:56.600 and there's a link on how to get that 00:00:56.840 --> 00:00:58.280 and what the timing is and all that. 00:00:58.840 --> 00:01:01.060 Also, you don't need to take notes. 00:01:01.190 --> 00:01:03.200 You can just sign up for our newsletters. 00:01:03.480 --> 00:01:06.060 So go to pythonbytes.fm, grab the newsletter 00:01:06.520 --> 00:01:09.280 and we don't do very many spammy things. 00:01:09.410 --> 00:01:11.300 We just let you know what we talked about 00:01:11.580 --> 00:01:14.100 and with all the links and a little bit of extra information. 00:01:14.860 --> 00:01:17.340 Yeah, and we love growing that list. 00:01:17.390 --> 00:01:20.260 It grows by a little bit every week 00:01:20.640 --> 00:01:21.820 and it's so fun to see. 00:01:23.080 --> 00:01:26.200 And with that, Michael, what do you got for us first? 00:01:26.660 --> 00:01:28.460 Let's talk about PortKiller. 00:01:28.920 --> 00:01:31.980 So PortKiller is an interesting Mac app. 00:01:32.320 --> 00:01:33.600 And if you're not a Mac, I apologize. 00:01:33.840 --> 00:01:34.800 I know it's a Windows as well. 00:01:34.980 --> 00:01:35.680 It's a Windows as well. 00:01:36.220 --> 00:01:40.860 So if you're a Mac or Windows, which is a good chunk of us, here is a cool application. 00:01:41.500 --> 00:01:43.760 So let me just lay out the problem for you, Brian. 00:01:44.190 --> 00:01:52.040 Either you've got 20 terminal tabs open or in one of your four IDE windows, 00:01:52.600 --> 00:01:58.520 You've got your API running or you've got your web app running or something like that. 00:01:58.820 --> 00:02:01.420 And you go try to run it again, not knowing that fact. 00:02:01.760 --> 00:02:04.300 And it says, sorry, that port is already in use. 00:02:04.480 --> 00:02:06.320 You're like, oh man, where is it now? 00:02:06.820 --> 00:02:08.899 And that's on you to know where you ran it. 00:02:09.000 --> 00:02:12.660 You know, I mean, there's a button to say run in your IDE versus running the terminal. 00:02:12.900 --> 00:02:13.860 And so maybe just press the button. 00:02:13.910 --> 00:02:16.760 You forgot, oh yeah, that thing was like one of the running things. 00:02:17.180 --> 00:02:23.580 But since 2025, there's another reason this is a pain is some of these agentic AI things. 00:02:23.900 --> 00:02:27.400 They're like, oh, let me start the server for you so I can test it. 00:02:27.740 --> 00:02:33.660 And it'll like temporarily fire up the web server and make a request and like lose track of it and not kill it. 00:02:33.660 --> 00:02:35.060 And it's not in your terminal. 00:02:35.240 --> 00:02:35.900 It's not in your ID. 00:02:36.440 --> 00:02:39.140 It's just there somewhere in the background, hidden. 00:02:39.470 --> 00:02:40.900 You can't get to it to shut it down. 00:02:41.170 --> 00:02:46.680 You go to a task manager activity monitor and you say Python is like four things like, hmm. 00:02:47.140 --> 00:02:48.580 Well, three of those are supposed to be running. 00:02:49.100 --> 00:02:50.120 Which one do I kill, you know? 00:02:50.620 --> 00:02:51.460 So it's just a real hassle. 00:02:51.690 --> 00:02:55.720 And so this port killer solves that problem, not just for Python, but for applications 00:02:56.070 --> 00:02:56.680 in general, right? 00:02:56.920 --> 00:03:01.260 So it's a powerful cross-platform port management tool for developers. 00:03:01.790 --> 00:03:07.940 It monitors ports, manages Kubernetes port forwarding, and even does integrated Cloud 00:03:08.040 --> 00:03:08.480 Flare tunnels. 00:03:09.000 --> 00:03:11.320 So it's really, really slick, actually. 00:03:11.720 --> 00:03:25.500 So what you do is it's got a little notification section menu bar thing that drops down and you can have multiple CloudFlare tunnels, which we've talked about in Grok and I've talked about Rathole, like CloudFlare tunnels or a thing like that. 00:03:25.540 --> 00:03:27.680 Like put this on the internet so other people can get to it. 00:03:27.860 --> 00:03:30.880 So maybe I can help debug or show off something, right? 00:03:30.980 --> 00:03:36.300 Like if you're trying to debug a webhook, the only way to do that is online, right? 00:03:36.500 --> 00:03:38.820 Because some other server is trying to get to your dev machine. 00:03:39.300 --> 00:03:40.400 So these tunnels are super helpful. 00:03:40.460 --> 00:03:41.200 We talked about that. 00:03:41.380 --> 00:03:43.320 So that will actually manage these things. 00:03:44.020 --> 00:03:56.160 Like, for example, you can go to say your Flask app and I'm pretty sure you can, I don't see the UI exactly, but I think you can go to it and say, oh, expose this Flask application as a cloud flare tunnel. 00:03:56.570 --> 00:03:56.660 Boom. 00:03:57.120 --> 00:03:58.180 It's off to the races. 00:03:58.400 --> 00:03:58.920 That's pretty cool. 00:03:59.160 --> 00:03:59.400 Yeah. 00:03:59.660 --> 00:03:59.800 Yeah. 00:04:00.140 --> 00:04:04.880 But it also shows you all the local ports that are open and you can go kill them. 00:04:05.050 --> 00:04:09.580 So what you can do is you can actually go to either search for a process or a port. 00:04:09.720 --> 00:04:13.660 You're like, it says that port, you know, 5,000 or 8,000. 00:04:13.660 --> 00:04:14.340 Let's go with 8,000. 00:04:14.460 --> 00:04:14.980 It's a little safer. 00:04:15.400 --> 00:04:16.640 8,000 is taken. 00:04:17.079 --> 00:04:18.019 You can't run your app. 00:04:18.209 --> 00:04:18.340 Why? 00:04:18.560 --> 00:04:21.200 Just go in there and type 8,000, find the thing and press kill. 00:04:21.310 --> 00:04:25.540 Cause you're like stupid clod code left that thing running and I can't find it to kill it. 00:04:25.700 --> 00:04:26.660 So I'll do it here. 00:04:26.860 --> 00:04:27.300 You know what I mean? 00:04:27.530 --> 00:04:30.240 Or you can just see what ports are open, what's running, all sorts of stuff. 00:04:30.440 --> 00:04:31.540 And I think it's super neat. 00:04:31.890 --> 00:04:37.160 You install it on macOS via homebrew, or you can just download a DMG. 00:04:37.520 --> 00:04:38.860 So that's pretty cool. 00:04:39.260 --> 00:04:43.140 And yeah, I mean, that's pretty much what I got to say about it, but it is nice. 00:04:43.680 --> 00:04:44.060 That's cool. 00:04:44.480 --> 00:04:44.540 Yeah. 00:04:44.900 --> 00:04:49.640 I don't like this is this obviously I'm not even sure why I'm commenting on this, 00:04:50.820 --> 00:04:56.000 but I most have like one or two ports open when I'm working on a web app 00:04:56.240 --> 00:04:59.800 because I'm not heavy duty web developer, but yeah, it looks neat. 00:05:00.300 --> 00:05:01.900 Well, the thing is you might not have very many, 00:05:02.140 --> 00:05:04.440 but you can only have one app on that one. 00:05:04.860 --> 00:05:09.220 And if it gets lost, especially if the little AI thing goes off 00:05:09.240 --> 00:05:14.060 and then leaves it just dangling you're like oh man yeah you got a rogue agent you got to hunt down 00:05:14.460 --> 00:05:19.420 yeah exactly so it's even if you're just doing one it's like it's still valuable yeah yeah it's 00:05:19.440 --> 00:05:26.060 pretty cool neat uh well what am i talking about um fine question i'm going to talk about uh making 00:05:26.080 --> 00:05:33.479 things faster and profiling so henry schreiner sent this to us and um we're he's a fan of the 00:05:33.420 --> 00:05:39.940 friend of the show and we're fans of henry shiner's work so he's a core developer um the article is 00:05:40.100 --> 00:05:47.960 how we made python's packaging library three times faster and this is cool because um packaging is 00:05:48.280 --> 00:05:54.460 is one of the um is one of the heaviest hit uh packages on pypi even though a lot of people 00:05:54.980 --> 00:06:02.280 might not uh pip install it themselves it's used by a lot of other stuff um and um i uh let's see 00:06:02.300 --> 00:06:05.080 I often am doing it when I'll reach for packaging 00:06:05.220 --> 00:06:06.520 if I'm trying to compare versions. 00:06:07.100 --> 00:06:12.880 At runtime, I'll say if I'm testing with something at 3.15 00:06:13.220 --> 00:06:15.040 versus a different one or something like that, 00:06:15.080 --> 00:06:17.720 I'll often use packaging for grabbing versions. 00:06:17.980 --> 00:06:21.840 Anyway, there's a bunch of fun stuff in the packaging package. 00:06:23.060 --> 00:06:25.240 So how did they make it three times faster? 00:06:25.540 --> 00:06:28.000 The article talks about graphing and plotting 00:06:28.180 --> 00:06:30.540 and how he's figuring this out for a while. 00:06:30.920 --> 00:06:37.380 And I'm going to scroll past all of that and talk about the actual updates. 00:06:38.160 --> 00:06:47.300 One of the things I love about this is we often hear about premature optimization is bad. 00:06:47.600 --> 00:06:52.800 But there are times where you see stuff that's taking a while and you're like, I really wish this was faster. 00:06:52.960 --> 00:06:55.460 And so there is a time for optimization. 00:06:56.100 --> 00:06:58.860 And often it goes along something like this. 00:06:59.740 --> 00:07:09.960 You put some profiling together, you try to measure the different parts of it, and then measure it and try to reproduce those, make sure they're consistent times so that they're not random. 00:07:10.260 --> 00:07:12.360 But then you just do like 10% at a time. 00:07:12.620 --> 00:07:19.040 So this is an interesting one of like 10% speed up of just stripping zeros out of stuff. 00:07:20.260 --> 00:07:25.940 They had a reversed list, 00:07:26.280 --> 00:07:29.020 iter tools drop while away to reverse it, 00:07:29.100 --> 00:07:29.940 to strip the zeros. 00:07:30.520 --> 00:07:34.020 And they just made it a little different 00:07:34.440 --> 00:07:38.120 and more actually more readable and less elegant, 00:07:38.600 --> 00:07:39.380 but it's nice. 00:07:39.880 --> 00:07:41.820 So instead of one line, 00:07:41.920 --> 00:07:43.880 there's five lines of code with a for loop, 00:07:44.260 --> 00:07:47.340 but it sped up the whole thing by 10%. 00:07:48.160 --> 00:07:49.580 And that's awesome. 00:07:50.000 --> 00:07:52.620 And then, okay, so that's a little thing, 10% speed up. 00:07:52.840 --> 00:07:59.540 And then a faster regex, 10% to 17% faster on 3.11 and above only. 00:07:59.840 --> 00:08:05.120 And that's another thing is sometimes newer versions of Python have more tools at your disposal. 00:08:05.300 --> 00:08:11.060 So you might have to have two implementations, but why not make it faster for the newest? 00:08:12.000 --> 00:08:13.080 So that's pretty cool. 00:08:14.280 --> 00:08:16.900 Removing, I don't actually get this. 00:08:17.260 --> 00:08:18.400 Removing single disk. 00:08:18.660 --> 00:08:18.980 Oh, yeah. 00:08:19.160 --> 00:08:21.940 Removing single dispatch, 7% faster. 00:08:23.120 --> 00:08:26.480 Remove duplicate version creation, 37% faster. 00:08:26.820 --> 00:08:27.000 Awesome. 00:08:27.720 --> 00:08:30.980 Removing named tuples, 20% faster. 00:08:31.110 --> 00:08:33.159 Just all these different little tiny speedups, 00:08:33.229 --> 00:08:35.300 and they're like 10% to 20%. 00:08:35.300 --> 00:08:37.680 And in the end, it's a lot faster. 00:08:38.180 --> 00:08:44.660 So I really like write-ups on people just taking a little bit befores and afters. 00:08:45.040 --> 00:08:51.400 And that doesn't mean to say if you've got something that looks like the before code that you should speed it up. 00:08:51.550 --> 00:08:54.240 You should measure and make sure that it's readable. 00:08:54.620 --> 00:08:56.980 I mean, I think the first goal is readability. 00:08:57.500 --> 00:08:59.600 Next, measure to see if it's too slow or something. 00:09:00.420 --> 00:09:00.500 Yeah. 00:09:01.140 --> 00:09:06.120 Write it in the form of readable, understandable, and then figure out how to make it fast. 00:09:06.380 --> 00:09:06.480 Yeah. 00:09:06.700 --> 00:09:07.480 If needed. 00:09:07.800 --> 00:09:12.560 Yeah, but there are tools available if you are ready to try to make things faster. 00:09:13.660 --> 00:09:13.940 Indeed. 00:09:14.460 --> 00:09:14.820 All right. 00:09:15.760 --> 00:09:16.000 All right. 00:09:16.160 --> 00:09:18.360 I have a, I don't know how to phrase this. 00:09:18.370 --> 00:09:19.140 I think it's not good. 00:09:19.500 --> 00:09:22.180 Let's call this a not good story, but it's quite the story. 00:09:22.290 --> 00:09:27.040 So this is AI's impact on dev companies, open source. 00:09:27.720 --> 00:09:33.140 We start with a quote from Simon Wilson's sort of link blog, which itself is a quote. 00:09:33.540 --> 00:09:33.660 Okay. 00:09:33.900 --> 00:09:40.040 So Adam Watham is the CEO of Tailwind Labs, which is the maker of Tailwind CSS. 00:09:40.680 --> 00:09:45.460 And the story about what's happening right now with Tailwind CSS is bonkers. 00:09:46.100 --> 00:09:47.780 Like you would have never predicted it. 00:09:48.020 --> 00:09:50.800 But once you hear the story of like, ah, yes. 00:09:50.960 --> 00:09:51.080 Okay. 00:09:51.720 --> 00:09:58.420 So from Adam, he actually did this 30 minute morning walk where he like sort of walks and 00:09:58.580 --> 00:10:01.980 talks and it's kind of like a journal, but public or something like that, I guess. 00:10:02.340 --> 00:10:05.820 And he says, we had six months left. 00:10:06.160 --> 00:10:10.400 He says, I just laid off some of the most talented people I've ever worked with. 00:10:10.540 --> 00:10:11.440 and it blink and sucks. 00:10:11.800 --> 00:10:13.280 The direct written version is, 00:10:14.580 --> 00:10:18.240 this all actually came from this thing called LLMs.txt, 00:10:18.960 --> 00:10:19.820 which actually is pretty interesting. 00:10:20.170 --> 00:10:22.360 Like robots.txt, but for LLMs, 00:10:22.700 --> 00:10:24.480 I'm thinking about putting this on Python bytes and stuff 00:10:24.720 --> 00:10:26.820 to like help LLMs be more efficient 00:10:27.140 --> 00:10:28.660 working with what we produce. 00:10:29.140 --> 00:10:30.380 So if somebody wants to ask a question like, 00:10:30.410 --> 00:10:33.660 hey, what did Python bytes say about this? 00:10:33.800 --> 00:10:35.320 Like what did Michael and Brian cover here 00:10:35.460 --> 00:10:37.080 and what was their advice, right? 00:10:37.340 --> 00:10:39.360 If we put that as a, hey, LLM, 00:10:39.740 --> 00:10:42.780 you could technically scan the entire system 00:10:42.960 --> 00:10:43.720 and try and understand that, 00:10:43.840 --> 00:10:45.580 or you could use our search for these keywords 00:10:45.850 --> 00:10:47.380 and then here's how you access the transcript 00:10:47.640 --> 00:10:48.720 and then parse just, you know what I mean? 00:10:48.820 --> 00:10:49.780 You could give like a little advice. 00:10:50.200 --> 00:10:54.200 So someone had proposed as a PR to put that onto Tailwind 00:10:54.250 --> 00:10:56.140 and said, "Could you please put that 00:10:56.350 --> 00:10:57.580 onto the Tailwind documentation 00:10:58.160 --> 00:11:01.040 so that we'll be more efficient talking to Tailwind 00:11:01.480 --> 00:11:04.480 or asking questions of Tailwind from our AI agents?" 00:11:04.740 --> 00:11:06.720 And people lost it, people lost it. 00:11:07.170 --> 00:11:08.120 It wasn't good. 00:11:08.600 --> 00:11:11.980 And Adam said something to the effect of, 00:11:12.110 --> 00:11:13.740 like, this actually is a pretty decent request, 00:11:14.480 --> 00:11:16.640 but the reality is that 75% of the people 00:11:16.730 --> 00:11:18.820 in our engineering team lost their jobs yesterday 00:11:19.050 --> 00:11:20.580 because the brutal impact of AI. 00:11:21.180 --> 00:11:23.780 And if I add this feature, it's going to make it worse. 00:11:24.440 --> 00:11:27.380 All right, so here's the insane aspect. 00:11:28.040 --> 00:11:30.320 If you look at traffic to the, 00:11:30.630 --> 00:11:33.200 if you look at downloads, sorry, on NPM, 00:11:33.340 --> 00:11:34.380 I guess that's traffic of a form. 00:11:34.480 --> 00:11:37.460 If you look at the downloads of Tailwind, 00:11:37.680 --> 00:11:40.420 They're up six times year over year. 00:11:40.780 --> 00:11:40.960 Wow. 00:11:41.560 --> 00:11:42.840 Six times more usage. 00:11:43.420 --> 00:11:45.640 Don't you think that would make the product more viable? 00:11:46.200 --> 00:11:46.320 No. 00:11:46.840 --> 00:11:52.460 Because if you go to the Tailwinds documentation, it's all free, Tailwinds open source, but there's a thing that says get Tailwind Plus. 00:11:52.940 --> 00:11:53.840 You see that in the docs. 00:11:54.280 --> 00:12:06.839 However, even though there's a 6x increase in usage, there's a 40% drop in traffic to the documentation and revenue is down five times or down 80%. 00:12:06.860 --> 00:12:12.080 Why? Because instead of going to the docs to get help, you just say, hey, AI, install Tailwind and do this. 00:12:12.080 --> 00:12:18.140 Or you don't even say that. You say, make it pretty. And like Tailwind is the default styling mode of AI. 00:12:18.620 --> 00:12:19.200 That's pretty sad, huh? 00:12:20.160 --> 00:12:26.880 Yeah, it definitely is. I don't. Yeah, it sucks. I don't know what the solution is. 00:12:27.120 --> 00:12:32.980 I don't either. Avoiding LLMs.txt is probably not the answer, but it's also not a solution for sure. 00:12:33.400 --> 00:12:36.320 I mean, it might need a different business model. You know what I mean? 00:12:37.080 --> 00:13:00.100 Yeah. I mean, we've both seen this as well. I was faced with higher server costs on my blog. And then I put in, we talked about this earlier, I put in monitoring to see who it was, and it was mostly robots. And then I cut robots out, and we got zero traffic. 00:13:00.280 --> 00:13:02.300 So I'm like, not zero. 00:13:02.300 --> 00:13:03.860 That's not the answer either, though, is it? 00:13:04.500 --> 00:13:07.460 It's like saying, I do not want to appear in Google because I hate the traffic. 00:13:07.600 --> 00:13:09.860 I mean, I know it's not exactly the same, so don't email it. 00:13:10.180 --> 00:13:15.840 But with Google, they said, if you don't let our robots in to scrape your site for our LLMs, 00:13:15.900 --> 00:13:20.120 then we're not going to scrape your site for search traffic either. 00:13:20.500 --> 00:13:22.760 I think that's the thing. 00:13:23.000 --> 00:13:24.300 And that's lame. 00:13:24.900 --> 00:13:29.280 I think that you should be able to say, no, I don't want somebody to scrape my site for AI. 00:13:29.380 --> 00:13:32.100 but please do keep looking at it for search. 00:13:32.740 --> 00:13:34.220 We just like, just take a moment 00:13:34.740 --> 00:13:37.180 and just say that there are really good alternatives 00:13:37.900 --> 00:13:40.660 to Google, like Startpage, like Kaggy. 00:13:41.580 --> 00:13:43.480 Yeah, but a lot of them are all built off 00:13:43.590 --> 00:13:45.540 of the same data that Google's selling. 00:13:45.810 --> 00:13:47.300 And Startpage is built off the stuff. 00:13:47.470 --> 00:13:49.540 So it is in a sense, really, really the same. 00:13:50.400 --> 00:13:51.340 Yeah, that's a good point. 00:13:51.420 --> 00:13:52.900 Okay, I'm not even done with the story, Brian. 00:13:53.160 --> 00:13:54.880 Okay, let's hear the rest of the drama. 00:13:55.620 --> 00:13:56.680 Okay, so that's half the story. 00:13:56.820 --> 00:13:58.340 There's a really good video by this guy. 00:13:58.360 --> 00:14:00.580 I just came across Maximilian Schwarzmuller. 00:14:00.780 --> 00:14:03.020 He does a bunch of great developer videos. 00:14:03.180 --> 00:14:04.320 It's more of like a- 00:14:04.320 --> 00:14:05.740 - And an awesome name. 00:14:05.740 --> 00:14:06.980 - I know it is an awesome name. 00:14:06.980 --> 00:14:09.540 He's more like a JavaScripty guy, 00:14:09.540 --> 00:14:12.300 but he does a lot of just like thought pieces 00:14:12.300 --> 00:14:14.420 on programming in general, and he's good. 00:14:14.420 --> 00:14:18.240 So he did a 11 minute, 55 second video on this tailwind 00:14:18.240 --> 00:14:20.800 is in deep trouble, talks about it, 00:14:20.800 --> 00:14:22.720 actually shows a lot of graphs and data. 00:14:22.720 --> 00:14:23.960 So it's pretty interesting. 00:14:23.960 --> 00:14:25.900 Then the other is a video from the Primogen. 00:14:26.500 --> 00:14:27.820 And this is the other side of, 00:14:28.300 --> 00:14:30.780 this is another company in the same theme 00:14:32.200 --> 00:14:33.400 and Stack Overflow is cooked. 00:14:33.800 --> 00:14:35.780 It is insane. 00:14:36.640 --> 00:14:37.720 Let me see if I can find a spot here. 00:14:38.080 --> 00:14:39.600 Let me see if I can find a spot here with this graph. 00:14:39.860 --> 00:14:42.380 So there's an insane graph if you watch that video, 00:14:42.600 --> 00:14:43.540 but I can just tell you the words. 00:14:44.200 --> 00:14:45.300 I'll tell it to you more, Brian. 00:14:45.680 --> 00:14:47.600 So in, I don't have my show notes up, 00:14:47.740 --> 00:14:48.780 so I'm about to do this from memory, 00:14:48.960 --> 00:14:52.520 but I believe in the first month of Stack Overflow's life, 00:14:52.720 --> 00:14:55.320 there was 3,800 questions asked. 00:14:55.700 --> 00:14:56.400 Yeah, in 2009. 00:14:56.640 --> 00:14:57.260 I've got the notes up. 00:14:57.470 --> 00:14:57.720 Yeah, okay. 00:14:58.780 --> 00:14:59.500 Almost nothing. 00:14:59.580 --> 00:15:00.340 I mean, it's just getting going. 00:15:00.480 --> 00:15:00.940 It's like a forum. 00:15:01.010 --> 00:15:02.880 Like, no, of course, there's going to be almost no questions. 00:15:03.300 --> 00:15:11.040 It goes up and it stays 200,000 new questions asked per month almost through the pandemic. 00:15:11.480 --> 00:15:12.500 And it starts going down. 00:15:12.630 --> 00:15:14.940 And this is not purely from AI. 00:15:15.440 --> 00:15:21.760 It's going down like pretty precipitously for two years before ChatGPT comes out. 00:15:22.100 --> 00:15:26.180 When it does in like mid-2022, it drops even harder. 00:15:26.490 --> 00:15:28.640 So here's an insane stat. 00:15:28.840 --> 00:15:32.520 In 2016, there were 200,000 new questions asked at Stack Overflow. 00:15:32.880 --> 00:15:41.500 In 2026, well, technically last month in December 2025, there were 3,740-something questions asked. 00:15:41.800 --> 00:15:43.700 The same as the first month they opened. 00:15:44.040 --> 00:15:46.380 200,000 to three or whatever. 00:15:46.670 --> 00:15:47.160 3,000. 00:15:47.400 --> 00:15:48.000 It's insane. 00:15:48.360 --> 00:15:49.560 That is insane. 00:15:50.160 --> 00:15:54.940 So actually, I think the primary guide's got a premium to prepare your goodbyes. 00:15:55.400 --> 00:15:56.220 It's pretty much it. 00:15:56.540 --> 00:16:03.800 Now, it does seem to still be going strong, like selling this data, this pure data in a sense, to AI companies. 00:16:03.970 --> 00:16:05.280 But for how long is that going to last? 00:16:05.900 --> 00:16:09.460 That's like, we don't have any cows left and we can't get milk. 00:16:09.740 --> 00:16:11.820 But we got a lot of cheese that's really good in the freezer. 00:16:12.160 --> 00:16:14.500 So we can sell that for a while until it's gone. 00:16:15.320 --> 00:16:18.140 Without new questions, there's not new information there. 00:16:19.700 --> 00:16:22.720 And who's answering, even if there are new questions, who's answering them? 00:16:23.140 --> 00:16:26.280 Yeah, I bet you half of those 3,000 or 4,000 are like AI generated. 00:16:27.220 --> 00:16:28.260 Honestly, I can't believe this. 00:16:28.290 --> 00:16:34.040 I knew that they were going down, but 200,000 to a couple thousand is an insane level. 00:16:34.400 --> 00:16:40.020 Yeah, and they were actually one of the better ones as far as ads and sponsors and stuff. 00:16:41.279 --> 00:16:46.160 I mean, Stack Overflow, we make fun of it and stuff, but it's a part of software, 00:16:46.460 --> 00:16:48.139 and it wasn't really that spammy. 00:16:48.860 --> 00:16:52.340 It just was, it didn't have a bunch of pop-ups and stuff. 00:16:52.780 --> 00:16:55.420 So I actually kind of liked Stack Overflow. 00:16:55.880 --> 00:16:56.260 I know. 00:16:56.660 --> 00:16:57.640 I mean, I don't know. 00:16:57.810 --> 00:16:58.840 I don't know exactly how to feel, 00:16:58.960 --> 00:17:01.440 but it even went so far as we had the Stack Overflow keyboard 00:17:01.590 --> 00:17:02.600 that only had three keys. 00:17:03.340 --> 00:17:03.620 Yeah. 00:17:04.420 --> 00:17:04.560 Right? 00:17:04.670 --> 00:17:07.160 It had the, instead of the controller command, 00:17:07.430 --> 00:17:09.900 it had the Stack Overflow logo and then a C and a V. 00:17:10.660 --> 00:17:11.720 It's a beautiful joke. 00:17:12.319 --> 00:17:15.140 Just, you know, that's how important that used to be. 00:17:15.620 --> 00:17:16.939 Yeah, you take an error message 00:17:16.959 --> 00:17:21.600 and just throw it into Google search and you get a Stack Overflow question answering it. 00:17:21.780 --> 00:17:24.819 That's how you figure stuff out. 00:17:25.439 --> 00:17:25.620 Exactly. 00:17:26.209 --> 00:17:30.040 There was a period where we read books and then we copied and pasted from Stack Overflow. 00:17:30.220 --> 00:17:33.420 And now I guess we copy and we just take it from AI. 00:17:33.720 --> 00:17:39.500 But I think this is noteworthy in a big way in the developer world, right? 00:17:39.520 --> 00:17:40.500 This is big. 00:17:40.780 --> 00:17:41.040 Yeah. 00:17:41.320 --> 00:17:46.300 And I mean, I know that the people behind Stack Overflow, there's a lot of money there. 00:17:46.680 --> 00:17:50.560 And I think that the top of the chain, they're probably fine. 00:17:50.920 --> 00:17:54.600 But everybody that got laid off, that's lame. 00:17:55.740 --> 00:17:58.900 But Tailwind has always been a lightweight business model, 00:17:59.060 --> 00:18:01.720 even though they're doing some amazing things. 00:18:02.000 --> 00:18:04.020 And like you said, it's being used by almost everyone. 00:18:04.400 --> 00:18:05.600 So it's crazy. 00:18:05.940 --> 00:18:06.100 Yeah. 00:18:06.700 --> 00:18:07.940 Kiva out in the audience points out, 00:18:08.120 --> 00:18:10.100 it's insane how symmetric this graph is. 00:18:10.220 --> 00:18:10.640 I agree. 00:18:11.000 --> 00:18:11.920 It is insane. 00:18:12.180 --> 00:18:14.140 It's like almost a perfect trapezoid. 00:18:14.460 --> 00:18:15.200 Not quite, but almost. 00:18:15.680 --> 00:18:15.840 Yeah. 00:18:16.200 --> 00:18:17.780 While I'm pulling up stuff from the comment, 00:18:18.070 --> 00:18:21.320 Henry points out that he's just a PyPA member 00:18:21.700 --> 00:18:23.400 and builds many, many Python things, 00:18:23.640 --> 00:18:24.640 but not quite a core development. 00:18:25.000 --> 00:18:25.480 Oh, okay. 00:18:25.720 --> 00:18:27.320 Keep the record a little bit. 00:18:27.600 --> 00:18:27.820 Okay. 00:18:28.160 --> 00:18:32.940 Well, Henry Schreiner is a core asset in my learning journey. 00:18:33.290 --> 00:18:34.360 I guess I'll put it that way. 00:18:35.060 --> 00:18:35.620 Over to you. 00:18:35.940 --> 00:18:36.180 Okay. 00:18:37.220 --> 00:18:40.900 I wanted to talk about Henry Schreiner's art. 00:18:41.000 --> 00:18:41.060 No. 00:18:42.320 --> 00:18:42.740 Similar. 00:18:43.080 --> 00:18:48.060 So we talked about Henry Schreiner's article on Python packaging and making it faster. 00:18:48.520 --> 00:18:53.040 And it reminded me of something that I wanted to talk about for a while and I just forgot about. 00:18:53.340 --> 00:18:54.720 So thanks, Henry, for the reminder. 00:18:55.200 --> 00:18:58.920 And it was because I was and it's around profiling. 00:18:59.160 --> 00:19:03.220 So I was looking through not the packaging library. 00:19:03.220 --> 00:19:09.600 I was looking through FastAPI once looking at their I kind of do this. 00:19:09.620 --> 00:19:11.220 I kind of troll sometimes. 00:19:11.560 --> 00:19:16.740 go through GitHub workflows of different packages to see what sort of tricks they're using in their 00:19:16.860 --> 00:19:20.880 workflows. And I remember... It's like the view source of the modern day. 00:19:20.880 --> 00:19:25.760 Yeah, workflows. And it's one of the wonderful things about open source packages is you can do 00:19:25.860 --> 00:19:32.440 this. So FastAPI, plus the folks at FastAPI are really great about trying to stay cutting edge 00:19:32.660 --> 00:19:39.760 and helping other people. So looking through here, I saw COD speed benchmarks. And I'm like, 00:19:39.800 --> 00:19:47.480 what is this so um that's my topic right now is the cod speed tool um and um it's probably maybe 00:19:47.580 --> 00:19:53.420 it's code speed but it's spelled cod so i'm going to pronounce it cod speed um it could be measured 00:19:53.540 --> 00:20:00.780 in like how fast does a cod swim in the ocean and this is like a is it is it a laden cod or a uh 00:20:01.700 --> 00:20:06.100 a heavily laden cut bite your knee caps off get back here 00:20:07.480 --> 00:20:23.780 So I'll talk about the actual tool, but as we're looking through FastAPI, there's also in the merge requests for new features, there's a performance report that is linked and you can go and take a look at it. 00:20:23.840 --> 00:20:28.760 If you click on that, you go and see, like, you can go and poke around at that particular 00:20:28.980 --> 00:20:29.400 merge request. 00:20:29.540 --> 00:20:30.440 I'm looking at the overview. 00:20:30.920 --> 00:20:39.480 His FastAPI on codspeed.io has this graph of basically making is not very noisy of like 00:20:39.860 --> 00:20:44.840 the different times they've measured things and how the metrics, the performance metrics 00:20:45.100 --> 00:20:45.700 have gone up and down. 00:20:45.760 --> 00:20:47.080 And it's really kind of cool. 00:20:47.120 --> 00:20:48.300 So I was like, can I use this? 00:20:49.040 --> 00:20:49.260 Yes. 00:20:50.060 --> 00:20:51.240 So I'm going to try this out. 00:20:51.240 --> 00:20:51.940 I haven't tried it yet. 00:20:52.160 --> 00:20:53.780 but at codspeed.io. 00:20:54.280 --> 00:20:57.260 And there's a nice rabbit logo, which is nice. 00:20:57.600 --> 00:21:01.480 Anyway, it integrates into dev and CI workflows 00:21:01.690 --> 00:21:03.760 to measure performance, detect regressions. 00:21:04.160 --> 00:21:08.240 And if it was just a paid tool, 00:21:08.530 --> 00:21:10.440 I wouldn't probably be covering it. 00:21:10.620 --> 00:21:13.080 But it's one of those awesome tools 00:21:13.360 --> 00:21:16.240