WEBVTT 00:00:00.020 --> 00:00:04.220 OpenAI just acquired Astral, the company behind uv, Ruff, and ty. 00:00:04.900 --> 00:00:08.400 And if your first thought was, wait, is uv toast? 00:00:08.870 --> 00:00:09.500 You're not alone. 00:00:10.220 --> 00:00:12.500 But here's the twist Charlie Marsh shared with me. 00:00:13.080 --> 00:00:17.740 He thinks they may ship more open source at OpenAI than they ever did at Astral. 00:00:18.180 --> 00:00:22.440 On this episode, we get into the acquisition, the mixed feelings, and the future of your 00:00:22.650 --> 00:00:27.180 favorite Python tools, and what it's like to build right at the center of the AI universe. 00:00:27.960 --> 00:00:33.340 This is Talk Python To Me, episode 552, recorded June 2nd, 2026. 00:00:51.399 --> 00:00:56.180 Welcome to Talk Python To Me, the number one Python podcast for developers and data scientists. 00:00:56.700 --> 00:00:58.080 This is your host, Michael Kennedy. 00:00:58.430 --> 00:01:02.020 I'm a PSF fellow who's been coding for over 25 years. 00:01:02.660 --> 00:01:03.760 Let's connect on social media. 00:01:04.100 --> 00:01:07.240 You'll find me and Talk Python on Mastodon, BlueSky, and X. 00:01:07.520 --> 00:01:09.360 The social links are all in your show notes. 00:01:10.120 --> 00:01:13.640 You can find over 10 years of past episodes at talkpython.fm. 00:01:13.820 --> 00:01:17.020 And if you want to be part of the show, you can join our recording live streams. 00:01:17.360 --> 00:01:17.840 That's right. 00:01:18.030 --> 00:01:21.300 We live stream the raw uncut version of each episode on YouTube. 00:01:21.880 --> 00:01:26.320 Just visit talkpython.fm/youtube to see the schedule of upcoming events. 00:01:26.570 --> 00:01:30.200 Be sure to subscribe there and press the bell so you'll get notified anytime we're recording. 00:01:31.100 --> 00:01:32.760 This episode is brought to you by Sentry. 00:01:33.080 --> 00:01:34.320 Don't let those errors go unnoticed. 00:01:34.600 --> 00:01:36.120 Use Sentry like we do here at Talk Python. 00:01:36.620 --> 00:01:39.480 Sign up at talkpython.fm/sentry. 00:01:40.400 --> 00:01:42.320 Charlie Marsh, welcome back to Talk Python and me. 00:01:42.640 --> 00:01:43.840 Lovely as always to have you on the show. 00:01:44.040 --> 00:01:44.260 Thank you. 00:01:44.360 --> 00:01:45.960 Yeah, it's great to be on the show again. 00:01:46.180 --> 00:01:46.720 It's always a pleasure. 00:01:47.080 --> 00:01:49.060 We talk about so many interesting things. 00:01:49.520 --> 00:01:50.720 We've kind of shared the journey. 00:01:51.140 --> 00:01:55.420 your journey together a little bit in the sense that I had you on originally to talk about Ruff 00:01:55.520 --> 00:01:59.560 and how cool and how fast it was. And when uv came out, I'm like, oh, you have to, we have to have 00:01:59.560 --> 00:02:05.040 you on the show to talk about uv and all these different things. pyx as well. And now something, 00:02:05.420 --> 00:02:10.820 something a little bit different, but awesome. I mean, congratulations that you're now part of 00:02:10.960 --> 00:02:16.539 OpenAI. I know that there's mixed feelings out in the universe about this, and we're going to talk 00:02:16.560 --> 00:02:22.120 about that but personally just happy for you this must be a dream come to true in some ways and 00:02:22.390 --> 00:02:27.660 yeah just congrats thank you yeah thanks so much it's funny it's like uh yeah a lot of the maybe 00:02:27.780 --> 00:02:33.000 all of the biggest moments for the for the company have been punctuated by appearances on talk python 00:02:33.300 --> 00:02:39.360 so uh we have really been we have we have been on this journey together certainly all of our big 00:02:39.540 --> 00:02:43.659 you know like product and tool releases i'm excited to come on here and yeah and talk more about the 00:02:43.680 --> 00:02:47.580 acquisition, joining OpenAI, all that kind of stuff. So yeah, thanks. Thanks for having me on 00:02:47.740 --> 00:02:51.240 again. And yeah, it's been, it's been fun. Hopefully we can do some more episodes too. 00:02:51.480 --> 00:02:57.560 You're welcome. And you know, I think having you on the show, especially now, certainly for uv, 00:02:57.820 --> 00:03:02.900 but even more so now, kind of like a public service to the community type of thing. Cause 00:03:04.160 --> 00:03:08.760 there's a lot of people out there just wondering what tools should they use? How does this change 00:03:08.980 --> 00:03:13.640 this thing that I've become so dependent upon and love so much and things like that. So I think just 00:03:13.660 --> 00:03:17.300 to share thoughts and future and all that is really good. 00:03:17.550 --> 00:03:17.760 Yeah, definitely. 00:03:18.260 --> 00:03:20.020 Now, before we do any of that business, 00:03:20.570 --> 00:03:21.120 believe it or not, 00:03:21.320 --> 00:03:23.120 not everyone has listened to every single episode 00:03:23.230 --> 00:03:24.100 of Talk Python To Me. 00:03:24.620 --> 00:03:26.620 Y'all out there, if that's you, you got some homework, 00:03:26.860 --> 00:03:28.040 but it may be a quick introduction 00:03:28.340 --> 00:03:29.180 for folks who don't know you. 00:03:29.420 --> 00:03:29.840 Yeah, of course. 00:03:30.020 --> 00:03:31.100 So I'm Charlie. 00:03:31.500 --> 00:03:34.940 I am the founder and was the CEO of Astral. 00:03:36.040 --> 00:03:37.440 We were a company. 00:03:37.490 --> 00:03:39.620 We build high-performance developer tools 00:03:39.840 --> 00:03:40.860 for the Python ecosystem. 00:03:41.380 --> 00:03:45.180 So we're best known for Ruff, which is our linter and formatter. 00:03:46.040 --> 00:03:48.960 And then we also build uv, which is a package manager and a manager. 00:03:49.640 --> 00:03:52.240 And then we're also working on a type checker called ty. 00:03:52.680 --> 00:03:56.280 We joined OpenAI about a month ago, I think at the time of recording. 00:03:56.590 --> 00:04:00.040 I continue to lead the Astral team here at OpenAI. 00:04:00.460 --> 00:04:02.120 We continue to work on all those tools. 00:04:03.180 --> 00:04:08.120 And hopefully in the future, we'll continue to work on all sorts of new cool tools too, 00:04:08.250 --> 00:04:09.480 in addition to all the things we've already done. 00:04:09.940 --> 00:04:10.820 But that's my background. 00:04:11.160 --> 00:04:13.540 We build developer tools, things that we hope, 00:04:13.630 --> 00:04:15.580 building software, effective, more fun. 00:04:15.760 --> 00:04:17.060 Well, you certainly nailed that. 00:04:17.420 --> 00:04:19.420 Developer tools, especially around packaging 00:04:19.820 --> 00:04:22.360 and other things like Ruff, but especially with uv, 00:04:22.600 --> 00:04:24.100 you really, really nailed it. 00:04:24.100 --> 00:04:25.860 And people certainly connected to it. 00:04:26.100 --> 00:04:28.240 Before we talk about this current and future, 00:04:28.540 --> 00:04:30.780 let's just do a little bit of looking back. 00:04:31.170 --> 00:04:33.180 So I remember having you on the show 00:04:33.680 --> 00:04:36.860 when Ruff was still just a project that you were doing. 00:04:37.040 --> 00:04:39.080 You're like, what if I learned this weird language, Rust, 00:04:39.540 --> 00:04:41.040 that was not that popular at the time? 00:04:41.400 --> 00:04:46.460 And I built this thing that was like black, but it was really fast and then released that. 00:04:46.780 --> 00:04:51.160 And my first experience was worth it was, oh, this is probably broken, or I just don't understand 00:04:51.230 --> 00:04:56.580 how to use it because I ask it to analyze a hundred thousand lines of Python code in a project. 00:04:56.900 --> 00:05:00.280 And it just went done. I'm like, huh, that probably didn't work. What did I miss? You know, 00:05:00.350 --> 00:05:04.000 maybe it's supposed to be like, look, look at the sub directories or something. I don't know 00:05:04.140 --> 00:05:10.139 what happened here. So take us on this journey of sort of from starting your open source journey 00:05:10.160 --> 00:05:13.440 until you pre-acquisition, just building up Astral? 00:05:13.780 --> 00:05:14.240 Yeah, definitely. 00:05:14.700 --> 00:05:17.680 So, I mean, I started the company, 00:05:17.990 --> 00:05:19.360 I was at a moment in my life 00:05:19.560 --> 00:05:21.460 when I thought I wanted to start a company, 00:05:21.700 --> 00:05:23.800 but I wasn't really sure what I wanted to do. 00:05:24.740 --> 00:05:27.780 And Ruff at the time was really like the side project 00:05:28.120 --> 00:05:30.920 that I wasn't really supposed to be spending time on, 00:05:31.700 --> 00:05:33.860 but I found really fun and interesting and cool. 00:05:34.180 --> 00:05:36.580 And I sort of published a blog post 00:05:36.650 --> 00:05:38.220 when I got it to what I'd considered 00:05:38.270 --> 00:05:40.000 to be sort of a proof of concept stage. 00:05:40.100 --> 00:06:09.820 And, you know, it turns out people were very interested in like the ideas behind it. And, and, and also the, I guess the instantiation of those ideas, like the fact that they ran this thing and it was fast and it seemed useful. And so I started working on it full time pretty soon after that. Started the company pretty soon after that. That was like late 2022. My first kid was born right around the same time. So I was kind of like, I had like a baby and, you know, this company that was at the time, just me. 00:06:10.200 --> 00:06:14.260 um then eventually started going to team like five that's a decent amount of work to have a baby and 00:06:14.300 --> 00:06:18.720 a new company i remember i think the baby made a bit of an appearance on yeah i think the baby made 00:06:18.720 --> 00:06:23.000 a bit of an appearance on the podcast one time which was all good yeah it was an amazing um 00:06:23.450 --> 00:06:30.680 yeah it was an amazing time in my life and um uh you know my uh my son was also born like seven 00:06:30.840 --> 00:06:35.280 weeks early and so i thought i had kind of like seven weeks to like figure out what i was doing 00:06:35.330 --> 00:06:39.039 with this company and then my son was just like born and and thankfully he's doing great but 00:06:39.180 --> 00:06:44.320 it was like a time in my life when uh everything was was kind of like things were going crazy like 00:06:44.700 --> 00:06:51.600 in all arenas um uh but isn't that emblematic of of being a dad just like yeah yeah i thought i had 00:06:51.640 --> 00:06:55.860 this time or i thought this was lined up and it completely just yeah and i was like fixing issues 00:06:56.120 --> 00:07:01.260 like from the hospital and like anyway i look back on it all like very fondly though you know yeah 00:07:02.900 --> 00:07:07.660 crazy to balance it all and then after a while i started growing the team um like we raised some 00:07:07.540 --> 00:07:12.820 money i was able to grow the team and um you know we started with rough which was our linter 00:07:13.160 --> 00:07:18.180 that kind of expanded um after that we did our formatter that was all that's also part of rough 00:07:18.660 --> 00:07:23.280 um and then you know a little bit after that we basically decided to start working on packaging 00:07:23.580 --> 00:07:28.380 which was um uh i mean at the time for us it was actually really different because we were working 00:07:28.380 --> 00:07:33.599 on python tooling but everything we'd built was like static analysis like linter formatter and i 00:07:33.620 --> 00:07:38.680 I kind of felt like if we wanted to be, you know, the Python tooling company, like aspirationally, 00:07:39.040 --> 00:07:43.340 that we sort of had to find a way to make packaging feel really different. 00:07:44.840 --> 00:07:47.820 And so we were like, let's take on this whole new set of challenges around. 00:07:47.940 --> 00:07:53.260 That was uv, did a couple of big uv releases, and the tools just kept like growing, growing, 00:07:53.460 --> 00:07:55.640 growing and put together. 00:07:56.780 --> 00:08:00.519 I mean, I just feel really lucky to be working with the team that we were able to put together 00:08:00.540 --> 00:08:06.940 because we were able to attract, I think, some really amazing people by building this company 00:08:07.220 --> 00:08:12.460 around build open source, write Rust, programming language tooling. And over time, you build momentum 00:08:12.760 --> 00:08:17.180 because people liked what we were doing. And then we had just huge amounts of usage. And so the work 00:08:17.320 --> 00:08:22.780 becomes more and more impactful. Last year, we started thinking seriously about commercialization 00:08:23.100 --> 00:08:28.259 because up to that point, we were, I mean, up until the point we were acquired, we were obviously 00:08:28.280 --> 00:08:32.860 venture funded. I mean, it was a pretty efficient company. If you think about it, we weren't spending 00:08:33.020 --> 00:08:37.260 crazy amounts of money. We always kept the team pretty small. But obviously the goal was to build 00:08:37.419 --> 00:08:43.020 a successful independent business. And so the question becomes, well, how do we want to monetize 00:08:43.180 --> 00:08:49.740 it? And I think our strategy around monetization was always quite consistent, which was we wanted 00:08:49.820 --> 00:08:58.600 the tools, what we considered the tools, like Ruff, uv, ty, to be free, open source, very 00:08:58.760 --> 00:09:05.580 permissively licensed, and ideally kept pristine from the monetization portion of the business. 00:09:06.240 --> 00:09:10.840 And what we wanted to do was basically build cloud services, like hosted services that people 00:09:10.860 --> 00:09:14.520 would pay money for that were kind of like the natural next thing you need if you were 00:09:14.800 --> 00:09:15.680 using our developer tools. 00:09:16.660 --> 00:09:35.000 And that was basically, that was actually like the vision that I like wrote out from the start. Not that it was like so, so like ingenious or whatever, but it was like, I was like, okay, we'll like build the tools and then we'll have like a kind of like a Python cloud that people will use our tools and the tools will create distribution, like companies that use tools will buy our products. 00:09:35.340 --> 00:09:41.800 And then maybe we can have like technical advantages because we work on both the tools and the hosted services and all that kind. 00:09:42.020 --> 00:09:48.680 So pyx was like the first piece of that platform. 00:09:49.260 --> 00:09:54.920 And pyx was our or is our hosted Python registry. 00:09:55.140 --> 00:09:57.000 It's sort of like the server to the uv client. 00:09:57.940 --> 00:10:07.640 So we had lots of issues in uv from people who, you know, companies basically that have needs around how they do Python packaging and distribution that we couldn't really solve in the client. 00:10:07.730 --> 00:10:17.400 Like maybe they need to distribute private packages or maybe they need to have various kinds of security scanning built into their registry instead of using PyPI. 00:10:18.120 --> 00:10:20.540 And we basically said, well, let's go build a really good registry. 00:10:20.960 --> 00:10:46.900 And we thought we could use that to solve not only those kinds of enterprise problems around security and private package distribution, but also a bunch of what we considered user experience problems, like trying to make packaging even faster, trying to make a big dent in PyTorch and GPUs and CUDA, a lot of the user friction that people run into when they're trying to build with GPUs. 00:10:47.260 --> 00:10:53.080 And so we were like, we could solve that if we have our own server, you know, and people are using our unique client. 00:10:53.620 --> 00:10:57.580 Yeah, and people want to know what you're alluding to with the PyTorch and all the issues. 00:10:57.890 --> 00:11:08.000 We did a whole episode both on pyx and on Wheel Next recently, especially the Wheel Next one, which was, I don't know, months ago, less than a year ago, something like that. 00:11:08.220 --> 00:11:11.220 Tons of hard problems that I've spent lots of time thinking about. 00:11:11.680 --> 00:11:12.400 It's very hard, yeah. 00:11:12.800 --> 00:11:15.860 Yeah, we're kind of like working on that in standards and we're still continuing to do that. 00:11:16.160 --> 00:11:20.220 And at the same time, we were like, let's see if we can solve this for paying customers. 00:11:20.500 --> 00:11:24.360 Because machine learning researchers, that's very high value users. 00:11:24.760 --> 00:11:27.020 Those people are very expensive for companies. 00:11:27.260 --> 00:11:38.920 So if you can make them more productive, and part of the goal of pyx is basically make those people more productive by trying to solve problems that require expertise, both in GPUs and CUDA and Python packaging, which is sort of a rare combination. 00:11:39.260 --> 00:11:43.440 So it's a rare combination, but it's one you basically need if you want to be productive in that space, unfortunately. 00:11:43.560 --> 00:11:47.100 Yeah, yeah. And the old Python packaging standards are kind of misaligned. 00:11:47.520 --> 00:11:48.740 So that's why it means... 00:11:48.930 --> 00:11:49.320 Yeah, exactly. 00:11:51.180 --> 00:11:53.660 This portion of Talk Python is brought to you by Sentry. 00:11:54.140 --> 00:12:01.820 I've been using Sentry personally on almost every application and API that I've built for Talk Python and beyond over the last few years. 00:12:02.140 --> 00:12:05.000 They're a core building block for keeping my infrastructure solid. 00:12:05.620 --> 00:12:07.560 They should be for yours as well. Here's why. 00:12:08.220 --> 00:12:12.440 Sentry doesn't just catch errors. It catches all the stuff that makes your app feel broken. 00:12:12.680 --> 00:12:18.160 The random slowdown, the freeze you can't reproduce, that bug that only shows up once real users hit it. 00:12:18.500 --> 00:12:22.040 And when something goes wrong, Sentry gives you the whole chain of events in one place. 00:12:22.280 --> 00:12:25.260 Errors, traces, replays, logs, dots connected. 00:12:25.640 --> 00:12:29.380 You can see what's led to the issue without digging through five different dashboards. 00:12:30.240 --> 00:12:34.740 Seer, Sentry's AI debugging agent, builds on this data, taking the full context, 00:12:35.400 --> 00:12:40.320 explaining why the issue happened, pointing to the code responsible, drafts a fix, 00:12:40.440 --> 00:12:43.400 and even flags if your PR is about to introduce a new problem. 00:12:44.240 --> 00:12:45.260 The workflow stays simple. 00:12:45.700 --> 00:12:47.420 Something breaks, Sentry alerts you, 00:12:47.740 --> 00:12:49.420 the dashboard shows you the full context, 00:12:49.760 --> 00:12:52.900 Seer helps you fix it and catch new issues before they ship. 00:12:53.620 --> 00:12:56.380 It's totally reasonable to go from an error occurred 00:12:56.720 --> 00:12:58.700 to fixed in production in just 10 minutes. 00:12:59.760 --> 00:13:02.480 I truly appreciate the support that Sentry has given me 00:13:02.720 --> 00:13:05.080 to help solve my bugs and issues in my apps, 00:13:05.700 --> 00:13:08.280 especially those tricky ones that only appear in production. 00:13:08.660 --> 00:13:10.120 I know you will too if you try them out. 00:13:10.440 --> 00:13:12.080 So get started today with Sentry. 00:13:12.360 --> 00:13:17.260 Just visit talkpython.fm/sentry and get $100 in Sentry credits. 00:13:17.780 --> 00:13:18.520 Please use that link. 00:13:18.530 --> 00:13:19.900 It's in your podcast player show notes. 00:13:20.080 --> 00:13:25.320 If you're signing up some other way, you can use our code talkpython26, all one word, 00:13:25.920 --> 00:13:28.440 talkpython26, to get $100 in credits. 00:13:29.240 --> 00:13:30.900 Thank you to Sentry for supporting the show. 00:13:32.020 --> 00:13:36.280 So yeah, we started, I think we did the, we launched the closed beta for that in like 00:13:36.470 --> 00:13:37.700 August of last year, maybe. 00:13:38.780 --> 00:13:44.680 And so from there, we basically had like, I would say three quarters of the team was working purely on the open source. 00:13:44.980 --> 00:13:53.680 And then maybe a quarter of the team, me included, was working on the commercial platform and starting a real kind of like sales process around that and like going and selling it to companies. 00:13:54.180 --> 00:14:00.620 And probably not widely known, but we had like real revenue and actually grew pretty well. 00:14:00.740 --> 00:14:07.120 I mean, it was no, you know, you see all these blog posts about companies going from like zero to a hundred million dollars in like three months or something and breaking all sorts of records. 00:14:07.420 --> 00:14:11.440 It wasn't like that, but it certainly grew pretty well. 00:14:11.560 --> 00:14:15.480 And we had, I would say, a small number of very high quality customers. 00:14:16.700 --> 00:14:20.420 And a lot of that basically came from the open source, people who used our open source 00:14:20.520 --> 00:14:21.960 tooling and saw we were building this thing. 00:14:22.000 --> 00:14:26.840 And that led to a lot of companies coming to us and wanting to see this product that 00:14:26.880 --> 00:14:29.020 we had talked about and written about. 00:14:29.720 --> 00:14:31.280 So we spent a while building that. 00:14:32.140 --> 00:14:36.940 And then, you know, eventually we ended up in conversations with OpenAI. 00:14:38.000 --> 00:14:41.920 That is a, it turns out to be a long and interesting process. 00:14:43.060 --> 00:14:50.020 But we announced that deal, I think in April, I think in April. 00:14:50.360 --> 00:14:51.760 Sorry, you can fact check me on that. 00:14:51.860 --> 00:14:52.840 I'm just scrolling off my memory. 00:14:53.290 --> 00:14:54.420 But we ended up announcing 00:14:54.420 --> 00:14:57.040 Let's see, March 19th, 2026. 00:14:57.040 --> 00:14:58.040 How about that? 00:14:59.240 --> 00:15:04.020 We announced that deal in March and then we joined OpenAI about a month ago. 00:15:04.340 --> 00:15:04.600 Okay. 00:15:05.060 --> 00:15:09.620 So a couple of months in between, figure out all the details and so on. 00:15:09.920 --> 00:15:16.300 So I want to take you through kind of reliving that experience and goals. 00:15:16.460 --> 00:15:21.680 But first, what was OpenAI primarily interested in your tooling? 00:15:21.840 --> 00:15:22.660 Was it UV? 00:15:23.260 --> 00:15:24.860 Was it rough? 00:15:25.100 --> 00:15:27.260 I mean, to me, let me just speculate out loud. 00:15:28.320 --> 00:15:31.260 then maybe I can channel some of the thoughts of the people out in the audience. 00:15:31.490 --> 00:15:37.400 Like uv, obviously for packaging and managing Python apps, tooling, dependencies, all of that, 00:15:37.600 --> 00:15:37.700 right? 00:15:37.860 --> 00:15:42.160 We've seen Claude acquire Bun, which is kind of a JavaScript equivalent. 00:15:42.900 --> 00:15:46.040 Maybe equivalent's not quite, but in the same ballpark type thing as uv, 00:15:46.980 --> 00:15:49.680 sort of like npm a little bit, but also running stuff. 00:15:50.300 --> 00:15:52.160 Yeah, it's a package major and a runtime. 00:15:52.520 --> 00:15:54.100 So it's like Node and npm. 00:15:54.520 --> 00:15:54.860 Exactly. 00:15:55.080 --> 00:15:56.140 It's like, yeah, Node plus npm. 00:15:56.280 --> 00:15:56.920 That's a good way to put it. 00:15:57.160 --> 00:16:04.540 But then also rough, like if you've got Codex, the team that you're on, you know, I see some of the AIs out there just going grep, grep, grep, grep. 00:16:04.620 --> 00:16:07.080 And I'm like, that seems like a really shallow way to understand it. 00:16:07.160 --> 00:16:13.480 So what if you could parse a concrete syntax tree and truly understand the code, not just fragments here and there? 00:16:13.580 --> 00:16:14.880 Like that might be super interesting. 00:16:15.260 --> 00:16:15.960 So these are my thoughts. 00:16:16.080 --> 00:16:19.260 And I'm like, well, why did OpenAI come and show an interest in you all? 00:16:19.600 --> 00:16:19.780 Yeah. 00:16:19.920 --> 00:16:20.820 What can you say about it? 00:16:21.000 --> 00:16:21.380 Yeah, definitely. 00:16:21.620 --> 00:16:23.160 I mean, there's a few different pieces to it. 00:16:23.240 --> 00:16:36.480 And, you know, I think one is, you know, OpenAI, there's a lot of, this has been written about publicly, like there's a lot of Python at this company, also a lot of Rust at this company. 00:16:37.170 --> 00:16:38.460 And we work at the intersection. 00:16:38.710 --> 00:16:40.080 We work on Python tooling. 00:16:40.190 --> 00:16:40.980 We write a lot of Rust. 00:16:42.100 --> 00:16:44.200 And they're also big users of our tools. 00:16:45.420 --> 00:16:51.180 So part of it is, you know, how can we help accelerate OpenAI as like one user of our tools? 00:16:52.160 --> 00:17:05.680 I think another is we, as an independent company for the acquisition, had been thinking a lot about how do our tools need to change as the way we build software changes. 00:17:06.120 --> 00:17:11.660 So we have to think a lot now about a lot of our users are building software with agents. 00:17:12.069 --> 00:17:18.160 And so we're thinking a lot about how do we make tools that are great for humans and for agents and what changes and what stays the same. 00:17:18.540 --> 00:17:28.540 And there are lots of questions that we want to explore there that I believe, and I think OpenAI believe this too, we can more effectively explore together. 00:17:28.650 --> 00:17:39.760 If you basically think about kind of co-designing like the models and the harness and tools, like what are interesting questions and where can you like innovate in order to try and build like a really different developer experience? 00:17:41.020 --> 00:17:54.480 And part of it too is I think just that we're lucky because I think OpenAI had a lot of respect for our team and the impact that we've been able to have with the resources that we had. 00:17:54.500 --> 00:17:58.880 And that if we were put in part of a bigger platform, we could have an even bigger impact. 00:18:00.320 --> 00:18:01.680 So it's a mix of things. 00:18:01.880 --> 00:18:06.520 Part of it is about accelerating OpenAI. 00:18:07.160 --> 00:18:13.640 Part of it is thinking about like how tools change and like the future of software as we, the Codex team. 00:18:14.600 --> 00:18:22.060 So, you know, this, this also speaks, I think, to a bit to, you know, sort of like what changes and what stays, what stays the same. 00:18:23.020 --> 00:18:30.700 Because for us, we, like we tend to view OpenAI as like a user of our tools. 00:18:31.520 --> 00:18:40.280 And what we want to do is we want to keep building what we hope is like great software that benefits everyone, you know, including OpenAI. 00:18:41.120 --> 00:18:48.580 So, you know, for us, we kind of use this platform in part as, oh, here's a big, you know, like a huge user of our tooling. 00:18:49.100 --> 00:18:56.460 How can we find areas where our tooling is like insufficient that help accelerate internally, but then also tools better for everybody? 00:18:57.460 --> 00:19:03.140 And so like still very much working on like all of our tools, like very much still a priority 00:19:03.220 --> 00:19:04.460 to keep, keep building those. 00:19:05.740 --> 00:19:10.260 And then similarly, if you think about that second piece of like, how can we bigger about 00:19:10.420 --> 00:19:11.400 like the future of software? 00:19:11.780 --> 00:19:14.800 We now have lots of opportunities that we, we sort of didn't have before. 00:19:14.960 --> 00:19:19.280 Like, I think before we were a bit more constrained because one, you know, the things that we 00:19:19.420 --> 00:19:24.260 worked on, they did have to be coupled in a realistic way to kind of like the commercial 00:19:24.500 --> 00:19:25.340 vision of the company. 00:19:26.220 --> 00:19:32.600 And two, we were pretty focused on Python and just Python. 00:19:32.800 --> 00:19:36.300 And the idea of expanding scope in different ways was, I think, more challenging. 00:19:36.600 --> 00:19:42.760 And so when I look at what we're doing here, part of it is I think we can actually be a 00:19:42.800 --> 00:19:49.260 little bit more ambitious and maybe a little bit crazy in terms of the experiments that 00:19:49.260 --> 00:19:51.540 we run or the things that we try to build. 00:19:52.480 --> 00:19:58.600 And even think beyond Python if we want to in terms of how can we build tools that basically make programming more productive? 00:19:58.790 --> 00:20:00.060 Like that was always our goal. 00:20:01.000 --> 00:20:06.220 And so here at OpenAI, it's like I want to keep and we're going to keep building our open source tools. 00:20:06.310 --> 00:20:09.880 We want to use those to build the best possible programming experience we can for Python. 00:20:10.900 --> 00:20:17.600 And we also want to be thinking about like what's the future of how we build software and how can we go and sort of like test those waters? 00:20:17.860 --> 00:20:26.340 And if we sort of like think a few steps ahead, what is like the next generation of tooling look like, whether it's for Python or for Rust or for, you know, other parts of the software development? 00:20:26.840 --> 00:20:35.500 That's pretty interesting. I guess one thing I left out in my enumeration of reasons they might care about you all is just Python intersect Rust super team, right? 00:20:35.680 --> 00:20:37.420 Like just the folks that work there. 00:20:37.800 --> 00:20:43.620 Yeah, that must be really interesting to just have such a bigger space to explore in, right? 00:20:43.630 --> 00:20:47.020 Because you were known as Astral, the team that makes awesome Python tools. 00:20:47.560 --> 00:20:50.920 But if you're like, oh, we got this really great idea about C++ and packaging, 00:20:51.200 --> 00:20:54.040 that wouldn't be congruent with how you're trying to grow 00:20:54.600 --> 00:20:58.180 into your Python cloud offerings and pyx and those types of things. 00:20:58.280 --> 00:21:00.120 But now the constraints are off. 00:21:00.580 --> 00:21:02.280 Constraints are just very different, right? 00:21:02.440 --> 00:21:06.540 I mean, I think we're still extremely committed to our tools. 00:21:07.660 --> 00:21:10.760 And we actually just finished, the first thing I did, 00:21:11.140 --> 00:21:13.640 maybe not the first day, but starting the second week, 00:21:13.820 --> 00:21:16.180 is I wrote up a charter for our team 00:21:16.640 --> 00:21:18.160 that we all kind of like aligned around 00:21:18.780 --> 00:21:19.740 to basically try to figure out like, 00:21:19.770 --> 00:21:20.440 what are we doing here? 00:21:21.180 --> 00:21:22.980 Like, what is our mission here at OpenAI? 00:21:23.090 --> 00:21:24.940 And like, how is it different than what we did before? 00:21:25.600 --> 00:21:26.820 How do we prioritize? 00:21:27.110 --> 00:21:28.720 How do we figure out whether something's worth doing, 00:21:28.920 --> 00:21:30.440 whether it fits into like our mandate 00:21:30.780 --> 00:21:32.000 and everything to do? 00:21:32.680 --> 00:21:34.820 And we just, we sort of like went through 00:21:34.960 --> 00:21:35.660 that mission exercise. 00:21:36.380 --> 00:21:37.720 And then we did a bunch of planning 00:21:37.910 --> 00:21:39.580 over the past two weeks to kind of set our goals 00:21:39.690 --> 00:21:41.780 for the next, we're doing like two month cycles. 00:21:41.930 --> 00:21:43.280 So we set our goals for the next two months. 00:21:43.560 --> 00:21:50.480 and I will say like of like most of what's in there is like continuing to build our open source 00:21:50.700 --> 00:21:55.040 tools and like trying to solve the problems that we see there and then there's obviously some new 00:21:55.320 --> 00:21:59.600 experimental different stuff that's like we wouldn't have done before you know especially 00:21:59.860 --> 00:22:03.820 around you know we're thinking about experimenting with like Rust tooling we think Rust is really 00:22:04.000 --> 00:22:07.240 interesting we spend all our time writing Rust we think Rust is going to be like an increasingly 00:22:07.500 --> 00:22:11.679 important language are there things that we could do to make like Rust more effective you know are 00:22:11.700 --> 00:22:16.340 there other ways that like agents are going to fit into the broader software lifecycle of like 00:22:16.700 --> 00:22:21.900 filing issues and and code review and all those kinds of things so you know it's a mix of keep 00:22:22.080 --> 00:22:27.520 building the tools that like so many people depend on like including open ai um like keep but also 00:22:27.680 --> 00:22:30.820 solving problems that honestly some of them probably don't matter that much to open ai but 00:22:30.820 --> 00:22:36.080 they matter a lot to our users um and go was better and part of it is like thinking big about 00:22:36.280 --> 00:22:40.819 new areas where we could be having you know hopefully like a 10x impact over what things 00:22:40.840 --> 00:22:45.740 are today in the same way that hopefully we did with our existing tools. So, you know, I think 00:22:46.160 --> 00:22:50.900 like we will, you know, it's easy for me to say this. I think like ultimately time will tell whether 00:22:51.020 --> 00:22:55.840 this is true, but I, at least right now, I genuinely think it's possible that we end up writing more 00:22:55.990 --> 00:23:01.520 open source here than we did at Astral. Because if you look at the arc of what was happening at 00:23:01.640 --> 00:23:04.980 Astral, I mean, it was going, like the company was going well, but also we were increasingly having 00:23:05.010 --> 00:23:09.839 to think like how do we commercialize? Yes. I was just thinking that actually. How do we build a 00:23:09.860 --> 00:23:15.460 commercial product and yeah yeah here we have like no constraints or like you know no rules or 00:23:15.470 --> 00:23:21.360 whatever but like we certainly um we certainly do have the ability right now to like keep building 00:23:21.640 --> 00:23:25.120 in open source and it's definitely uh you know there's a lot of alignment in that throughout 00:23:25.120 --> 00:23:30.340 the company so um i don't know i'm just like i'm very to be honest with you i'm just like super 00:23:30.640 --> 00:23:37.540 happy right now because um i think it was obviously a very stressful process uh in a lot of different 00:23:37.560 --> 00:23:43.240 ways. Just a lot of responsibility to weigh, you know, a lot of high stakes moments, a lot of 00:23:43.560 --> 00:23:49.660 different groups to consider, whether it's like the people on the team, our users, our customers, 00:23:50.260 --> 00:23:55.380 right? There's just a lot, a lot, a lot to think about. But like sitting here now a month in, 00:23:55.580 --> 00:24:00.320 I'm like, I'm super happy with, with like where the, where we're situated on as a team, 00:24:00.920 --> 00:24:05.320 like the things that we're committed to doing over the next few months, the resources we have to do 00:24:05.340 --> 00:24:10.920 them, it's like, I don't know, it feels like a really special like moment in time. And I think 00:24:11.210 --> 00:24:15.600 I really think the amount of like open source that we will ship over the next year will be like, 00:24:15.920 --> 00:24:19.780 yeah, again, possibly more than we would have done otherwise. I mean, obviously hard to hard to prove 00:24:19.830 --> 00:24:22.960 and who knows how that will play out. But that's how I feel like sitting here today, at least. 00:24:23.140 --> 00:24:27.520 That's great. You know, one thing I was thinking about while you're saying this is I run my own 00:24:27.650 --> 00:24:33.759 company here at Talk Python, right? And there's a remarkably high percentage of my time. It just 00:24:33.780 --> 00:24:36.660 feels like administration support, et cetera. 00:24:37.130 --> 00:24:40.900 You know, I'm doing accounting and I'm trying to figure out like, chasing 00:24:41.080 --> 00:24:44.840 down sponsors for the show, following up on invoices that don't get paid, 00:24:45.280 --> 00:24:48.280 keeping the website running, you know, just like all these things, none of 00:24:48.440 --> 00:24:50.140 which delivered direct value. 00:24:50.640 --> 00:24:54.780 And so you've, you probably live this like even to a higher degree than I did, 00:24:55.200 --> 00:24:58.900 but, but then now you've been like plucked from that environment, leaving 00:24:59.080 --> 00:25:02.300 all that infrastructure, like accounting and stuff aside. 00:25:02.740 --> 00:25:07.200 And now you're more like pure product, which is in a sense could actually be better, as you were saying. 00:25:07.480 --> 00:25:15.400 I mean, you know, to be honest, like it's like that's a lot of what I was hoping for is I thought that this would put us in a position. 00:25:16.440 --> 00:25:22.680 And again, like I should be held to account, you know, at some point in the future to see whether this works out or not. 00:25:22.880 --> 00:25:29.080 But like I I thought I would be putting us in a position to keep doing the things that we love to do and that we're really good at. 00:25:30.020 --> 00:25:39.780 And, you know, in a place that was that didn't require us to compromise on a lot of our values in terms of how we think about users and how we think about building. 00:25:41.340 --> 00:25:44.120 And I'm I don't know, I'm excited about like what we have planned. 00:25:44.510 --> 00:25:49.660 So, yes, I I certainly, you know, my my job has changed quite a bit, as you can imagine. 00:25:50.060 --> 00:25:56.100 Like I it's funny because like I'm not I'm probably working more or the same amount. 00:25:56.460 --> 00:25:59.620 But I like I kind of like always I just like work a lot. 00:26:00.039 --> 00:26:02.640 but I do feel less stressed to be honest. 00:26:04.220 --> 00:26:06.560 Like I feel less stress about, 00:26:07.480 --> 00:26:09.100 sort of like existential stress about 00:26:09.600 --> 00:26:11.680 how do I make sure the company doesn't like completely fail, 00:26:12.820 --> 00:26:13.860 you know, especially for 00:26:13.860 --> 00:26:14.800 It's probably more deep. 00:26:14.800 --> 00:26:17.220 Even all these people who came and joined, right? 00:26:17.220 --> 00:26:19.200 But now I feel less stress. 00:26:19.200 --> 00:26:21.200 I feel way more, I feel just like, I don't know. 00:26:21.200 --> 00:26:22.160 It's very cool. 00:26:22.440 --> 00:26:22.600 Amazing. 00:26:22.620 --> 00:26:24.520 All right, before we move off this topic of, 00:26:24.580 --> 00:26:26.520 you know, you joining OpenAI and all that, 00:26:26.520 --> 00:26:29.019 I kind of just want to have you relive the experience 00:26:29.040 --> 00:26:34.360 a little bit for people like what what was it like to hear from open ai like oh my gosh they want to 00:26:34.520 --> 00:26:38.560 acquire they want to have a discussion about maybe acquiring us and then maybe telling the team 00:26:39.500 --> 00:26:43.400 telling your wife like just and your family like just to the extent you can just give people a 00:26:43.540 --> 00:26:47.540 sense of like what those first couple yeah yeah of course i mean there's like you know there are 00:26:47.780 --> 00:26:54.859 there's obviously a lot i can't say um you know the things i will say is like one um i i think 00:26:54.880 --> 00:27:02.080 think I'm pretty naive, but it was like, it was a lot of work and takes a lot of time. And, you 00:27:02.080 --> 00:27:05.640 know, there's just a bunch of different phases to the process where sometimes things move very fast 00:27:05.760 --> 00:27:10.200 and sometimes they move slow and then they move very fast. And it just has all sorts of ups and 00:27:10.360 --> 00:27:17.860 downs and twists. And, you know, I think, I think people had, you know, lots of different feelings 00:27:18.240 --> 00:27:24.180 about the acquisition. And yeah, that's right. There was a lot of positive, but also a lot of 00:27:24.180 --> 00:27:28.100 negative stuff that you probably got thrown at you when this came out, right? So that must have been 00:27:28.150 --> 00:27:32.620 a little bit challenging to deal with. Yeah. I mean, I think for me, like it wasn't, 00:27:33.100 --> 00:27:40.240 to be totally honest, it wasn't anything that I didn't expect. And it reminded me a bit of when 00:27:40.310 --> 00:27:43.540 we announced that we were, you know, to be honest, it reminded me a bit of when we announced that we 00:27:43.660 --> 00:27:49.520 were forming a company and sort of like, you know, the range of reactions that you get around that. 00:27:49.720 --> 00:27:53.619 Some people are very excited for you. Some people are like, this should be an open source project 00:27:53.640 --> 00:27:54.740 that isn't part of a company. 00:27:55.000 --> 00:27:56.700 Some people are worried about the future of the project 00:27:56.860 --> 00:27:58.480 or think that you're not happy with your choices. 00:28:00.100 --> 00:28:02.500 And I kind of tell myself a lot of the same things 00:28:02.680 --> 00:28:03.940 that I told myself then, 00:28:04.980 --> 00:28:09.120 which are, no matter what you say 00:28:09.200 --> 00:28:10.160 and how genuine you are 00:28:10.160 --> 00:28:11.180 and how you feel about those things, 00:28:11.320 --> 00:28:14.260 you won't convince everybody just with words. 00:28:15.200 --> 00:28:16.620 And what you really have to do 00:28:16.620 --> 00:28:18.560