WEBVTT 00:00:12.340 --> 00:00:15.900 Welcome to the bikeshed podcast where we dive into all things software engineering 00:00:16.670 --> 00:00:20.220 argue over which terminal is truly worthy of your deity-like devotion 00:00:21.160 --> 00:00:26.599 and remind you weekly why neovim is the one true ide while everything else is just an electron app 00:00:26.680 --> 00:00:33.940 cosplaying as productivity. I'm your co-host, Tempest TypeScript Troll, Gargoyle of Golang, 00:00:34.150 --> 00:00:41.720 Scott Kaye. And alongside me, my co-hosts. His blog post is your blueprint to success. His PR comments 00:00:41.960 --> 00:00:46.920 read like a sacred text. And every meeting he steps into becomes an architectural whiteboard 00:00:47.220 --> 00:00:54.180 exercise. Matt Hamlin. And my other co-host, the man whose opinions are so scorching hot, 00:00:54.580 --> 00:00:56.260 They don't just melt your performance metrics. 00:00:56.740 --> 00:00:58.440 They optimize and ship them. 00:00:58.960 --> 00:01:01.580 Dillon, spicy, take, curry. 00:01:02.800 --> 00:01:05.300 Fellas, what are we cooking up today? 00:01:08.740 --> 00:01:09.960 We got a short episode today. 00:01:10.340 --> 00:01:11.180 Well, maybe it'll be short. 00:01:11.240 --> 00:01:11.460 Who knows? 00:01:11.660 --> 00:01:13.020 Maybe we'll end up babbling on for an hour. 00:01:13.200 --> 00:01:15.640 We had a few, maybe sort of like a newsflash episode. 00:01:15.860 --> 00:01:17.400 We got to work on the branding and the naming of it. 00:01:17.760 --> 00:01:18.800 But a few topics we wanted to cover. 00:01:19.100 --> 00:01:22.520 Bun getting acquired by Anthropic, which is pretty interesting, exciting. 00:01:22.660 --> 00:01:23.540 Maybe we talk about that. 00:01:23.740 --> 00:01:30.880 And then right before we started recording, React just announced that they had a severity 10 vulnerability with React server components. 00:01:31.120 --> 00:01:35.620 So we're kind of scrambling to figure out what that is and how we fix it. 00:01:36.260 --> 00:01:38.980 Dillon, let's maybe bounce it to you. 00:01:39.040 --> 00:01:39.800 What do you want to start with first? 00:01:39.880 --> 00:01:40.900 Do we talk about React first? 00:01:41.340 --> 00:01:41.740 Sure. 00:01:42.600 --> 00:01:52.960 It turns out if you're using Cloudflare and you use the WAF, which stands for Web Application Firewall, they just like patched it for you. 00:01:53.560 --> 00:01:54.820 which is actually really cool. 00:01:56.000 --> 00:01:59.160 It sounds like even if you're using the vulnerable version 00:01:59.880 --> 00:02:01.680 and you have this firewall, 00:02:02.640 --> 00:02:06.260 you're protected already, which sounds great. 00:02:06.540 --> 00:02:08.080 Hopefully it's actually doing what it's supposed to. 00:02:08.539 --> 00:02:12.640 We were able to apply it in minutes in reaction to this blog post, 00:02:12.680 --> 00:02:13.700 which is pretty sick. 00:02:14.860 --> 00:02:17.780 I don't know anything about what this can actually cause 00:02:18.340 --> 00:02:19.240 to people's systems. 00:02:19.380 --> 00:02:21.960 It sounds like you can just run whatever code you want on somebody's server. 00:02:22.820 --> 00:02:23.180 Potentially. 00:02:25.340 --> 00:02:29.600 Also, in addition to Cloudflare patching at the firewall, I think 00:02:29.620 --> 00:02:30.780 Vercel did something similar. 00:02:30.860 --> 00:02:32.800 I saw railway announced something as well. 00:02:33.460 --> 00:02:35.400 And I think Netlify as well and Deno deploy. 00:02:35.560 --> 00:02:40.240 So like a number of different sort of deploy infrastructure companies, maybe 00:02:40.340 --> 00:02:43.980 we're keyed in early also on the report, which is nice. 00:02:44.480 --> 00:02:47.820 One of the things that I'm finding interesting is like most of the server 00:02:47.900 --> 00:02:50.880 component frameworks built on top of react weren't keyed in. 00:02:51.440 --> 00:02:56.780 So Waku, which I use, had no idea about it until the public announcement of it. 00:02:56.860 --> 00:02:57.500 The disclosure of it. 00:02:58.040 --> 00:03:00.820 It seemed like next JS knew about it though, ahead of time, obviously, 00:03:00.980 --> 00:03:03.540 because they're basically the react team at this point. 00:03:04.260 --> 00:03:08.400 And there's like a number of other like meta frameworks and whatnot that also 00:03:08.640 --> 00:03:12.560 weren't keyed in early, which is to me is like kind of, I don't know, just yet 00:03:12.600 --> 00:03:16.840 again, yet another sort of bad taste in my mouth about the react team and a 00:03:17.060 --> 00:03:20.200 framework of like not keeping the community up, up to date with things 00:03:20.400 --> 00:03:20.820 that are happening. 00:03:22.180 --> 00:03:26.540 is your hot take that the meta investors were told ahead of time 00:03:28.820 --> 00:03:34.600 yeah yeah this was priced in at the market um yeah everyone that invested in the meta 00:03:35.120 --> 00:03:37.660 knew about it they put in some puts this morning yeah 00:03:42.440 --> 00:03:49.739 scott looks pissed he didn't know 00:03:42.440 --> 00:03:49.739 i'm not pissed at all it doesn't actually affect me 00:03:49.760 --> 00:03:55.860 I mean, someone could take down my blog probably, but that's about it. 00:03:55.860 --> 00:03:57.480 I mean, it doesn't really affect me. 00:03:57.490 --> 00:03:59.360 We don't use these at work currently. 00:03:59.760 --> 00:04:00.920 So it's kind of nice. 00:04:01.020 --> 00:04:06.220 We had the whole episode about React Server Components where I was like somebody pushing 00:04:06.420 --> 00:04:08.780 for them at work and it's not something we're using yet. 00:04:08.940 --> 00:04:10.880 So like, it's just less work. 00:04:11.100 --> 00:04:12.820 So I'm actually excited to hear that. 00:04:13.240 --> 00:04:19.720 That's probably the biggest pro for me is that I don't have to now sell why React Server 00:04:19.720 --> 00:04:24.160 components are still great, even though they have this epic vulnerability in them right now. 00:04:24.860 --> 00:04:28.240 I took the day off of work because we're going on a trip, which I'll share on my stand-up update. 00:04:28.420 --> 00:04:32.900 But the first thing that came to mind was like, one, yeah, at work, we're probably not even, 00:04:33.340 --> 00:04:36.460 maybe no one's even talking about it. I mean, I assume someone shared it in a discussion channel, 00:04:36.680 --> 00:04:40.520 but like we're not impacted, fortunately, or maybe unfortunately, depending on how you look at it. 00:04:41.030 --> 00:04:47.020 But then I also realized that I have countless side projects that are using maybe really old 00:04:47.040 --> 00:04:53.860 versions of Next or using waku or using the vite rsc plugin and like all those are going to be i 00:04:53.940 --> 00:04:58.440 need to go through and update all those so that's going to be a good activity on the flight 00:05:00.340 --> 00:05:05.640 ehh shouldn't be that bad there's an upgrade path for at least i didn't read the through the whole 00:05:05.840 --> 00:05:09.960 article i was on the move i think next had an upgrade path but was there upgrade paths for 00:05:10.320 --> 00:05:15.340 for everything that you're using i know you use maintainer waku like didn't know about this so 00:05:15.360 --> 00:05:18.400 So they just started looking at it after I shared it in the Discord channel for it. 00:05:20.180 --> 00:05:23.740 Yeah, Next.js was basically the only one that was keyed in on the update. 00:05:24.000 --> 00:05:29.000 I guess the parcel setup also was keyed in just because they maintain their integration within the React repo. 00:05:29.970 --> 00:05:31.140 So it was patched at the same time. 00:05:33.380 --> 00:05:34.460 But yeah, I don't know. 00:05:34.560 --> 00:05:39.220 There's a number of different things that will sort of have a trailing, 00:05:39.400 --> 00:05:41.440 sort of long tail of updates that might need to happen. 00:05:42.740 --> 00:05:44.760 Maybe it's time to just stop using React. 00:05:47.060 --> 00:05:49.960 maybe yeah maybe we should point people back to our episode about uh remix, The Downfall of React episode 00:05:51.680 --> 00:05:58.540 we're getting there remix needs to actually be in a state where we can use it right 00:05:58.540 --> 00:06:06.360 that's yeah that's true yeah march uh or or april i'm sure it'll slip from there though 00:06:06.980 --> 00:06:10.700 yeah march 2030 00:06:06.980 --> 00:06:10.700 yeah exactly 00:06:06.980 --> 00:06:10.700 remix 12. 00:06:12.960 --> 00:06:14.820 Any further thoughts on the React vulnerability? 00:06:15.360 --> 00:06:17.400 I think it's maybe not too surprising for most folks 00:06:17.440 --> 00:06:19.300 of like React is going into the server. 00:06:19.400 --> 00:06:20.600 So it obviously opens the door 00:06:20.760 --> 00:06:24.300 to server vulnerabilities as well. 00:06:25.980 --> 00:06:27.880 Yeah, I don't know if I have any further hot takes. 00:06:31.600 --> 00:06:33.000 Just that it's, I guess, unfortunate. 00:06:33.160 --> 00:06:35.300 I feel like they worked on server components for a while. 00:06:35.480 --> 00:06:37.900 I don't like normally I would say, oh, rushed, rushed. 00:06:38.280 --> 00:06:40.720 But I don't think that was the case here. 00:06:41.620 --> 00:06:43.320 I just think more so maybe unfortunate. 00:06:44.160 --> 00:06:47.560 It isn't widely adopted yet, I think, everywhere. 00:06:47.880 --> 00:06:50.880 So it may be on smaller apps and maybe in some places. 00:06:51.020 --> 00:06:55.320 I know we adopted it at Wayfair, but I don't fully believe that it's widely adopted throughout 00:06:55.340 --> 00:06:57.940 the industry just yet, which I find interesting. 00:06:58.720 --> 00:06:59.580 It is unfortunate. 00:06:59.860 --> 00:07:01.040 It's a bad look for React. 00:07:01.240 --> 00:07:06.360 Perfect time for Remix to get deployed and ready to be used in an environment. 00:07:07.600 --> 00:07:10.820 Because I still think there's space for a de facto framework. 00:07:10.920 --> 00:07:13.420 written on top of React to get Remix isn't. 00:07:13.820 --> 00:07:14.140 I don't know. 00:07:14.280 --> 00:07:15.060 Maybe it's how I feel, 00:07:15.220 --> 00:07:17.980 but I feel like Next.js was going to be this dominant player. 00:07:18.120 --> 00:07:20.380 It still kind of is, but I do, I don't feel like, 00:07:21.480 --> 00:07:22.980 I feel like I've seen so many developers 00:07:23.880 --> 00:07:25.140 who have had issues with it, 00:07:25.200 --> 00:07:25.780 that there's just like, 00:07:25.940 --> 00:07:28.640 there's no one true de facto solution still. 00:07:28.980 --> 00:07:31.640 There's use case by use case related solutions 00:07:32.500 --> 00:07:34.200 that are better based on what you're building. 00:07:34.800 --> 00:07:38.579 I do think that there could be a dark horse player here 00:07:38.960 --> 00:07:44.340 that could start to be like a new, um, a new framework around react or just a new framework 00:07:44.430 --> 00:07:44.740 in general. 00:07:47.600 --> 00:07:47.680 Yeah. 00:07:47.810 --> 00:07:52.900 Maybe we save the, yeah, maybe we have a deeper dive on react and if there's something better 00:07:52.900 --> 00:07:56.260 than next, maybe we do some like, maybe TanStack Start is the way to go. 00:07:56.530 --> 00:07:56.860 I don't know. 00:07:57.520 --> 00:08:01.780 Could be what one, I guess like one other thought as you were talking that came to mind is like, 00:08:02.420 --> 00:08:04.060 you know, next had, had released some patches. 00:08:04.180 --> 00:08:11.880 they only patched back to the v15 major so v16 was patched v15 was patched but not v14 00:08:12.600 --> 00:08:16.340 you know it could be any number of things could be that they don't support v14 anymore because 00:08:16.480 --> 00:08:21.900 it's three summer majors behind or whatever but uh the one thing that also that's not clear from 00:08:21.900 --> 00:08:28.019 the react announcement of like is there they announced like the 19.0 and and above releases 00:08:28.380 --> 00:08:35.479 were impacted, but RSC existed in the Canary releases and experimental releases prior to 00:08:35.860 --> 00:08:41.940 the 19 stable cut. So I'm wondering if there's vulnerable code that might not get updated just 00:08:41.940 --> 00:08:46.820 because it hasn't been flagged or has been patched. That'll be interesting to see. 00:08:47.380 --> 00:08:52.800 Yeah, I feel bad for all the large apps that are on like maybe 17 or 18 and they just have a really 00:08:52.740 --> 00:08:59.460 hard time upgrading and now they're going to be forced to make a upgrade decision 00:08:59.460 --> 00:09:06.240 i think if they're not using next or like another rc framework then and they're on v18 that then it should be 00:09:06.350 --> 00:09:10.480 safe or like you know they don't need to worry about upgrading but right i'm specifically talking 00:09:10.610 --> 00:09:17.239 about using react server components on those versions yeah yeah 00:09:17.260 --> 00:09:23.420 this this made me realize how annoying it is that you you think you're using like react 18 in your app when you're using next 00:09:24.340 --> 00:09:30.020 because you are within your app using react 18 but the next js is just using some other random 00:09:30.520 --> 00:09:35.500 version and it i was looking at it with matt before this and it was like kind of hard to 00:09:35.800 --> 00:09:41.360 find that version without digging through eight layers of node modules to a random file that had 00:09:41.360 --> 00:09:45.459 the version and maybe there's a better way to do it but it's just kind of weird that it 00:09:45.480 --> 00:09:46.820 Ships at some random version. 00:09:47.460 --> 00:09:48.440 It kind of pissed me off. 00:09:49.120 --> 00:09:49.460 Nextjs is, 00:09:49.620 --> 00:09:49.880 s***. 00:09:53.060 --> 00:09:53.480 Hot take. 00:09:53.640 --> 00:09:53.860 Noted. 00:09:55.240 --> 00:09:55.460 Noted. 00:09:56.080 --> 00:09:56.920 You can bleep that. 00:09:56.920 --> 00:09:58.940 Coming from the guy who still works with it. 00:09:59.600 --> 00:09:59.920 Noted. 00:10:00.580 --> 00:10:01.500 Please bleep that. 00:10:03.160 --> 00:10:03.920 Oh, I can bleep that. 00:10:04.680 --> 00:10:08.100 Well, we thank the React team for giving us an early Christmas gift of this vulnerability. 00:10:08.660 --> 00:10:09.400 God, I'd love to see that. 00:10:09.960 --> 00:10:11.380 Earlier in this week, it's amazing. 00:10:11.540 --> 00:10:14.340 This week feels like it's been a month long, but it's only Wednesday when we're recording. 00:10:14.840 --> 00:10:26.060 But earlier this week, there was an announcement that Bun, everyone's favorite JavaScript runtime, package manager, test runner, et cetera, extraordinaire, has been acquired by Anthropic. 00:10:27.260 --> 00:10:29.300 Did you guys have this on your bingo card for 2025? 00:10:29.980 --> 00:10:33.320 Were you expecting to see Anthropic acquiring a JavaScript runtime? 00:10:34.220 --> 00:10:35.600 Absolutely not, Matt. 00:10:37.139 --> 00:10:40.580 JavaScript, this is putting JavaScript or TypeScript on the map right here. 00:10:40.780 --> 00:10:41.540 I'll tell you that. 00:10:42.220 --> 00:10:46.360 I mean, if it wasn't a hot commodity language to learn, now it is. 00:10:46.940 --> 00:10:50.380 If I'm doing a little bit of digging, isn't most of Bun written in Zig? 00:10:50.940 --> 00:10:52.560 Is that really what's happening here? 00:10:53.100 --> 00:10:55.000 Is Anthropic acquiring Zig? 00:10:56.620 --> 00:10:57.780 Or is Zig-based programming? 00:10:58.780 --> 00:10:59.980 They're not necessarily. 00:11:00.060 --> 00:11:03.060 I mean, the Bun team, from what I've seen, I guess, 00:11:03.600 --> 00:11:07.440 they've tried to make proposals to the Zig language, 00:11:08.040 --> 00:11:08.940 but they've been shut down. 00:11:09.260 --> 00:11:12.140 I think Jared had tweeted something around Thanksgiving 00:11:12.160 --> 00:11:17.000 that they had more shut down proposals for the language 00:11:17.170 --> 00:11:18.760 than approved proposals for the language, 00:11:19.170 --> 00:11:19.760 which is kind of crazy. 00:11:20.050 --> 00:11:21.000 But that is interesting to say, 00:11:21.030 --> 00:11:22.180 like just the other week, 00:11:23.020 --> 00:11:25.020 Zig had announced that they're moving off of GitHub 00:11:25.460 --> 00:11:26.500 as their source of truth 00:11:26.550 --> 00:11:28.560 or like sort of code hosting solution 00:11:28.650 --> 00:11:29.300 and moving to, 00:11:30.000 --> 00:11:31.300 I don't know if they announced exactly 00:11:31.330 --> 00:11:32.280 what they're moving to, 00:11:33.500 --> 00:11:34.340 but they're moving off of GitHub 00:11:34.640 --> 00:11:36.760 because of Microsoft and AI 00:11:37.220 --> 00:11:38.560 and shitification and whatnot. 00:11:38.880 --> 00:11:40.940 But it is interesting that, yeah, 00:11:41.120 --> 00:11:43.740 Anthropic's going after Bun and Bun is written on Zig. 00:11:43.740 --> 00:11:46.960 And meanwhile, OpenAI is using their codex agent 00:11:47.070 --> 00:11:48.160 as being rewritten into Rust. 00:11:48.780 --> 00:11:51.520 So it's an interesting space. 00:11:52.339 --> 00:11:54.200 Big win though for JavaScript. 00:11:54.740 --> 00:11:55.520 It's a big win. 00:11:55.940 --> 00:11:57.660 I also think another one of the big pros 00:11:57.750 --> 00:11:59.940 is they're keeping the whole Bun team, 00:11:59.990 --> 00:12:01.140 a very talented team, 00:12:01.720 --> 00:12:03.140 and they're expanding that team. 00:12:03.380 --> 00:12:08.100 So it sounds like Anthropic likes the product enough 00:12:08.150 --> 00:12:09.140 and the team enough to say, 00:12:09.340 --> 00:12:11.880 we're going to acquire you and let you do what you do best. 00:12:12.080 --> 00:12:18.440 It's nice to see that from a company for someone who works at a corporation or has worked at 00:12:18.680 --> 00:12:19.300 multiple corporations. 00:12:19.960 --> 00:12:23.540 It's nice to see when a corporation like really trusts the team that's working on something 00:12:23.570 --> 00:12:25.120 and keeps it, keeps it the same. 00:12:25.150 --> 00:12:26.980 And they understand that and have value for that. 00:12:29.580 --> 00:12:29.660 Yeah. 00:12:29.660 --> 00:12:35.840 I think another interesting aspect here is like Bun maybe famously so far has like, you 00:12:35.860 --> 00:12:39.040 openly admitted that they don't have a path towards profitability. 00:12:40.350 --> 00:12:43.940 You know, they, they, everything they make is free and open source right now. 00:12:45.660 --> 00:12:48.480 You know, a lot of people speculate early on in the year or early on in 00:12:48.800 --> 00:12:52.760 Bun's development of like, you know, when is like Bun deploy or Bun cloud or, 00:12:53.160 --> 00:12:54.040 you know, whatever coming out. 00:12:55.760 --> 00:12:58.600 It's interesting to see, you know, maybe this accelerates some of that stuff. 00:12:59.200 --> 00:13:02.880 Maybe, maybe they don't have to worry about it because Anthropic is printing 00:13:03.160 --> 00:13:04.180 some money, I assume. 00:13:04.740 --> 00:13:07.140 You know, maybe they're not necessarily, I don't think they're, I don't think 00:13:07.380 --> 00:13:11.480 Anthropic is making a profit at the moment in the sense of like, they make money, 00:13:11.610 --> 00:13:15.460 but you know, they earn money, but they're probably spending more than that right now. 00:13:15.940 --> 00:13:21.100 Um, and so it'll be curious to see how that, like how this, you know, sort of 00:13:21.200 --> 00:13:24.440 handles, uh, or, you know, comes to fruition in the coming months or years. 00:13:25.560 --> 00:13:25.760 Yeah. 00:13:25.860 --> 00:13:30.580 It's not like a competitor kind of, I guess, like Deno where Deno has plans when using it. 00:13:30.740 --> 00:13:34.580 So like, I guess one of the thoughts I had was kind of getting into my cons. 00:13:34.620 --> 00:13:38.660 list that might be jumping ahead, but I know they, I believe the article stated that like, 00:13:38.740 --> 00:13:44.160 it'll still be a free open source tool, but does that mean new features become price gated? And 00:13:44.330 --> 00:13:50.640 the other con I have here is now that it's, it's no longer driven by like an open source community 00:13:50.800 --> 00:13:57.440 building for open source or open source consumers really, is Anthropic going to change the priority 00:13:57.800 --> 00:14:04.200 of what gets built for Bun? Are we going to see only features that Anthropic needs? I don't know 00:14:04.020 --> 00:14:09.460 that's a bad thing could be it could be a great thing but it changes the audience of who's using 00:14:09.580 --> 00:14:13.720 it a bit so it'll be interesting to see where bun goes from here i do think this is a net pro 00:14:14.220 --> 00:14:19.560 for bun and and overall maybe we'll we'll see more people adopt one in the future 00:14:20.030 --> 00:14:24.920 due to this major change but there are some ambiguities i'm interested in seeing how they shake out 00:14:25.300 --> 00:14:32.560 yeah kind of agree with your points me shared i think also what might not be very clear 00:14:32.580 --> 00:14:38.660 to a lot of people when they hear this announcement is that um you know cloud code heavily uses bun 00:14:38.780 --> 00:14:44.820 like they ship as a bun compiled uh sort of single binary uh when you're running the cli 00:14:45.380 --> 00:14:50.000 but also all as far as i know all the runtime code when you're using cloud code in the web 00:14:50.380 --> 00:14:56.240 or through the mobile app uh it's running on something that's running bun right so it could 00:14:56.240 --> 00:15:01.040 be so you know i don't know it could be some container in a kubernetes uh ecosystem that 00:15:01.060 --> 00:15:06.460 Anthropic is running maybe on AWS and, um, and it's running bun under the hood. 00:15:06.780 --> 00:15:10.380 They had already had these pretty solid investments into the 00:15:10.380 --> 00:15:11.840 bun ecosystem, as I understand it. 00:15:12.660 --> 00:15:13.640 It is interesting. 00:15:13.680 --> 00:15:17.020 I, you know, I wasn't necessarily expecting to see this announcement. 00:15:17.540 --> 00:15:20.900 I was actually kind of thinking about it as like, I wouldn't be surprised if 00:15:21.120 --> 00:15:24.520 Cloudflare would be interested in buying Anthropic eventually. 00:15:25.220 --> 00:15:27.700 They've been buying some sort of AI companies. 00:15:27.820 --> 00:15:31.020 I don't know if they're necessarily in the frontier model space versus the 00:15:31.040 --> 00:15:37.800 just run models sort of space 00:15:31.040 --> 00:15:37.800 cloudflare buy anthropic you think that would happen 00:15:37.880 --> 00:15:42.480 i think there's a lot of cross-pollination there like i think they they like the Claude team like 00:15:44.060 --> 00:15:49.500 mostly with cloudflare containers and like from the people i follow on twitter very limited subset 00:15:49.980 --> 00:15:54.120 both both companies but there seems to be some cross-pollination they said that and it was like 00:15:54.280 --> 00:15:58.420 anthropic to me sounds like their valuation it feels like their about their valuation is probably 00:15:58.420 --> 00:16:02.660 super, super high to the point where I don't even think Cloudflare could afford it. 00:16:03.370 --> 00:16:05.640 But maybe that's just how it feels. 00:16:05.920 --> 00:16:06.740 Yeah, that could be the case. 00:16:06.830 --> 00:16:12.420 But I think also, I think there's growing tension in the market of are these companies 00:16:12.660 --> 00:16:13.280 properly valued? 00:16:13.610 --> 00:16:15.420 The frontier model companies, are they properly valued? 00:16:16.920 --> 00:16:19.020 Is OpenAI actually worth $500 billion? 00:16:19.650 --> 00:16:21.240 Or are they actually worth $50 billion? 00:16:23.820 --> 00:16:28.380 I think there's also some news, maybe it's just leaks at this point, of maybe Anthropic is 00:16:28.400 --> 00:16:30.260 gearing up for an IPO in 2026. 00:16:31.100 --> 00:16:34.300 So like that sort of cloud for buying them is probably less likely. 00:16:35.740 --> 00:16:36.540 Scott, you brought up Deno. 00:16:36.760 --> 00:16:41.820 What if open AI goes after and buys Deno and uses Deno as their sort of opinionated runtime 00:16:42.080 --> 00:16:45.820 for running coding tasks via codex? 00:16:46.480 --> 00:16:46.960 That'd be interesting. 00:16:47.880 --> 00:16:53.260 Not sure it'll make them faster, but well, I mean, I'm not, it's not fair for me to judge 00:16:53.420 --> 00:16:53.920 Deno like that. 00:16:53.920 --> 00:16:57.140 I haven't really actively used it much. 00:16:57.320 --> 00:17:02.200 I just felt like for something that kind of rewrites how things work and behave, 00:17:03.180 --> 00:17:04.520 there's not a lot of support. 00:17:05.220 --> 00:17:09.220 I think one time you remember I was like building out a monorepo with it and 00:17:09.959 --> 00:17:11.959 they have like deno.json files. 00:17:12.839 --> 00:17:15.660 And I thought I could swap out all my package on JSON files, but it didn't 00:17:15.740 --> 00:17:21.939 support a monorepo and I thought like, oh, well, I don't really want to go backwards here. 00:17:22.180 --> 00:17:27.280 So, I mean, I really got to give Deno a run in an app and see how I really 00:17:27.300 --> 00:17:33.340 it but bun like was a very seamless integration it's all it's not the same as using npm or yarn 00:17:33.820 --> 00:17:38.680 or pnpm especially when we were using it early there there's a lot less tooling that existed at 00:17:38.740 --> 00:17:43.540 the time but at least like you knew what you were doing with it as as a runtime and it kind of felt 00:17:43.660 --> 00:17:48.820 more like a seamless transition i really enjoyed that with bun and bun is just incredibly fast i 00:17:49.320 --> 00:17:56.400 think Deno is as well but bun has been a seamless carryover from those other runtimes 00:17:56.920 --> 00:18:02.620 Yeah, I think, I think that was maybe Bun's sort of stated goal when they started of like, try to be as compatible with node. 00:18:02.750 --> 00:18:06.260 Whereas like Deno maybe intentionally started with, we don't want compatibility. 00:18:07.130 --> 00:18:08.800 We want to like make it good. 00:18:10.020 --> 00:18:20.140 And then maybe the past couple of years, two years, three years, Deno is like maybe shifted focus to say, actually, we do care about compatibility because enterprises using node care about compatibility. 00:18:21.030 --> 00:18:25.180 And that was like their path to profitability is like get enterprises using Deno. 00:18:26.220 --> 00:18:27.320 And yeah, I don't know. 00:18:27.420 --> 00:18:32.320 It'll be interesting to see if other frontier model companies that offer services like this 00:18:33.140 --> 00:18:35.940 decide to buy other runtimes or other languages also. 00:18:36.280 --> 00:18:39.340 Like, you know, you can't really buy a language, but, you know, I don't know. 00:18:39.540 --> 00:18:44.300 Like, yeah, maybe Elixir or the main team behind Elixir or something like that gets bought up 00:18:44.440 --> 00:18:50.160 or the, you know, the primary contributors to effect get bought up by someone and used. 00:18:51.380 --> 00:18:52.320 I said, yeah, I don't know. 00:18:53.500 --> 00:18:55.420 It seems kind of an interesting move, 00:18:55.450 --> 00:18:57.460 but I think it'll pan out pretty well for Anthropic. 00:18:59.220 --> 00:18:59.680 This was great. 00:19:02.040 --> 00:19:02.240 All right. 00:19:02.330 --> 00:19:04.680 Any further thoughts on either topic? 00:19:05.200 --> 00:19:08.060 Otherwise, we can transition into our stand-up updates. 00:19:09.360 --> 00:19:12.820 Dillon, you got any Bun thoughts here? 00:19:13.220 --> 00:19:16.160 You guys using Bun over there at the big old Whoop Factory? 00:19:16.580 --> 00:19:17.420 Dillon's like, what is Bun? 00:19:17.700 --> 00:19:19.320 Who uses Bun besides Matt? 00:19:21.140 --> 00:19:22.120 And Claude, apparently. 00:19:22.320 --> 00:19:23.340 We didn't use Bun at. 00:19:23.380 --> 00:19:29.800 at the big old w did we i use bun all the time i don't i default to bun that's great 00:19:29.800 --> 00:19:35.740 before this was announced like i had no idea that anthropic had any ties to to bun but that maybe i'm just not 00:19:36.880 --> 00:19:43.760 tied into that news like as much as map 00:19:36.880 --> 00:19:43.760 i mean i would kind of expect that for a cursor daily 00:19:43.850 --> 00:19:48.380 user instead of a Claude code daily user so i would agree 00:19:43.850 --> 00:19:48.380 i'm using claude code now 00:19:49.820 --> 00:19:51.500 It's about time we've converted you. 00:19:51.840 --> 00:19:51.940 Yeah. 00:19:52.200 --> 00:19:53.340 It was just because I was given access. 00:19:53.540 --> 00:19:55.200 I'm not paying my, my access. 00:19:55.639 --> 00:19:56.440 Superior product. 00:19:56.580 --> 00:19:57.040 Money for it. 00:19:57.260 --> 00:19:57.840 Superior product. 00:19:58.400 --> 00:19:59.640 Anthropics, sponsor our ads. 00:20:01.220 --> 00:20:02.460 I will say it's interesting. 00:20:03.240 --> 00:20:06.720 You know, this is like sort of getting further away from the Bun side of things, but it's 00:20:06.860 --> 00:20:10.720 interesting that the frontier model companies are like kind of maybe falling into different 00:20:10.940 --> 00:20:11.280 niches. 00:20:11.680 --> 00:20:14.580 You know, I'm sure they don't necessarily want to pigeonhole themselves this way, but like 00:20:14.680 --> 00:20:17.440 Anthropics is definitely going after coding and like. 00:20:18.400 --> 00:20:25.180 enterprise like agentic user use whereas open ai is maybe still very consumer focused you know i 00:20:25.180 --> 00:20:28.860 know they have obviously they have enterprise agreements and contracts and whatnot but 00:20:29.980 --> 00:20:35.420 but you know as a layman in the industry it's like i see open ai as the consumer name brand 00:20:35.760 --> 00:20:40.700 you know like the apple for example and then anthropic more as the windows of like you know 00:20:40.700 --> 00:20:46.840 the enterprise sort of offering or microsoft rather now 00:20:46.860 --> 00:20:53.940 i feel like those are excellent analogies with like the the open ai browser and it's a lot about consumption of of it uh and they 00:20:53.940 --> 00:21:00.720 they create the products as opposed to anthropic being the developers tool i i really like that 00:21:01.080 --> 00:21:06.160 thought i hadn't had that thought great great call on that 00:21:01.080 --> 00:21:06.160 it's also maybe interesting to say like i 00:21:06.160 --> 00:21:11.220 think even i think before this announcement there were some threads online saying that uh open ai 00:21:11.180 --> 00:21:17.420 has internally codified a code red, what they call it, basically saying like all hands on 00:21:17.580 --> 00:21:24.040 deck, we need to figure out why we're losing to people like Anthropic or other sort of 00:21:24.960 --> 00:21:25.780 model provider companies. 00:21:26.800 --> 00:21:32.360 It does feel like OpenAI has slipped in some regards of like, not necessarily in terms 00:21:32.400 --> 00:21:35.120 of model capabilities, but just in terms of the product offering. 00:21:36.240 --> 00:21:46.500 Like, I feel like they haven't really, besides the browser, they haven't really announced any, like, new, exciting usage of ChatGPT or the models behind it. 00:21:48.320 --> 00:21:48.760 So, I don't know. 00:21:48.940 --> 00:21:50.920 It's, yeah, kind of curious to see. 00:21:51.020 --> 00:21:53.280 Like, I feel like OpenAI was the dominant market player. 00:21:53.380 --> 00:22:01.840 I think Anthropic has, like, been slowly, not necessarily slowly, but, like, has been climbing back up to compete in the same weight bracket as OpenAI. 00:22:03.840 --> 00:22:11.700 Is it clear that market share of enterprise companies using these different APIs, where it stands? 00:22:12.160 --> 00:22:16.120 Because I feel like, especially where I work, we're pretty heavily invested into OpenAI. 00:22:16.390 --> 00:22:22.420 And it might be kind of hard for us to just completely switch over to Anthropic. 00:22:22.740 --> 00:22:24.420 It sounds like they're doing their job then. 00:22:25.480 --> 00:22:29.080 That's what you want to be, sticky and make it difficult for these companies to get out of it. 00:22:29.110 --> 00:22:29.560 Go ahead, Matt. 00:22:29.640 --> 00:22:39.060 No, I mean, to pull out a Scottism, it kind of depends on the use case of how your company has, you know, like how are you leveraging that contract with OpenAI, for example? 00:22:39.240 --> 00:22:45.220 Like, is it maybe the sales team using OpenAI to help, I don't know, do what, you know, whatever? 00:22:45.600 --> 00:22:48.900 Or is it engineering using the models from OpenAI and Codex, right? 00:22:49.559 --> 00:22:59.340 Like, you know, that doesn't necessarily speak to the stickiness, but it just speaks to like, yeah, like organizations might leverage one company or the other for different use cases. 00:22:59.480 --> 00:23:03.440 So like, for example, HubSpot, you know, I don't know necessarily if it's private information, 00:23:03.620 --> 00:23:08.460 but, you know, we, I think we've shared often enough that we partner with both OpenAI and 00:23:08.620 --> 00:23:09.780 Anthropic for a lot of different things. 00:23:09.880 --> 00:23:13.980 So it's like we're paying for contracts for both of those things, but for different use 00:23:14.100 --> 00:23:18.780 cases internally, primarily the Anthropic for the models for coding and then OpenAI for 00:23:19.540 --> 00:23:23.040 things outside, like sort of day-to-day use, AI usage. 00:23:25.340 --> 00:23:25.860 All right. 00:23:26.180 --> 00:23:27.120 Completely unrelated, Matt. 00:23:27.340 --> 00:23:33.480 Did HubSpot ever get one of those plaques for using 1 million tokens? 00:23:34.100 --> 00:23:35.160 I saw those on Twitter. 00:23:36.299 --> 00:23:36.980 I don't know. 00:23:37.720 --> 00:23:37.880 Maybe. 00:23:38.820 --> 00:23:41.320 My company received one a couple of months ago. 00:23:41.400 --> 00:23:41.760 That's pretty cool. 00:23:41.960 --> 00:23:42.940 It's kind of a weird thing. 00:23:43.580 --> 00:23:46.860 Thanks for spending a million dollars with us or whatever. 00:23:47.900 --> 00:23:50.880 Thanks for giving us 10 million over the past couple of years or whatever. 00:23:51.320 --> 00:23:52.760 I wouldn't be surprised if we did receive one, 00:23:53.020 --> 00:23:56.500 but we're also such a big company that we probably could have received one 00:23:56.540 --> 00:23:58.740 and not like someone shared it in a channel, 00:23:58.920 --> 00:24:00.180 but like not everyone's in that channel. 00:24:00.300 --> 00:24:01.000 So who knows? 00:24:02.340 --> 00:24:04.020 In their desk and it's forgotten about. 00:24:05.220 --> 00:24:07.040 Well, our CTO was like a big proponent, 00:24:07.300 --> 00:24:09.820 like big supporter of their company from the get-go. 00:24:09.880 --> 00:24:11.960 He was using it before it was like a consumer app. 00:24:11.960 --> 00:24:13.180 He was using it just as the API. 00:24:13.880 --> 00:24:16.600 So like, I'm sure he himself used million 00:24:16.860 --> 00:24:19.340 or like has consumed or generated trillions of tokens 00:24:19.460 --> 00:24:19.800 or something. 00:24:20.780 --> 00:24:21.340 Yeah, I'm not sure. 00:24:21.580 --> 00:24:22.000 Yeah, probably. 00:24:22.220 --> 00:24:24.060 We probably have a plaque sitting somewhere collecting dust. 00:24:24.920 --> 00:24:25.300