00:41:13 · Maya, Anika
[00:41:13] S0: Was it Or like Devin, the first software engineer. That's also a viral moment that they had. [00:41:18] S2: Yeah. So, like, they they created Devin, which was the first software engineer. And then the whole drama with Windsurf was that at the time, OpenAI was gonna acquire Windsurf. Everyone thought it was a done deal because they were, like, dropping these, like, little nuggets that it was gonna happen. And then Google, for $2,400,000,000, got basically the CEO, Varun Mohan, the co founder, Douglas Chen, and key engineers from Cognition. But then left to, like, kind of this, like, headless dead dead company people thought behind. But then Cognition, like, basically bought the rest of the company and it was like a big acquisition And they've continued to grow. [00:41:57] S0: Yeah. I mean, actually a great turnaround story for the Windsurf engineers too. If they Yeah. You stayed through it. I know a couple people didn't because like it was just tough probably mentally to go through that too. [00:42:05] S2: Well, also Cognition's work culture, I think is insane. [00:42:09] S0: Yeah. I mean, Windsurf's was too. I... But like... [00:42:12] S2: But I heard a lot of people say like Cognition's like the worst. Well, apparently, they were just like, you have to be... I think the difference is like, Cognition was like, you have to be in for... During the weekend. [00:42:21] S0: Oh, I see. [00:42:22] S2: I feel like that's really tough. [00:42:23] S0: That is. Yeah. That is I mean, that's
00:42:23 · Maya, Priya, Anika
[00:42:23] S0: That is. Yeah. That is I mean, that's [00:42:24] S1: just your entire... Like, even if work is your whole life, that's, like, a whole another level. Dude, I can't do that. Their revenue has profited from this, though, because it's increased from 37,000,000 to 492,000,000 in twelve months, and they're aiming to cross 1,000,000,000 [00:42:42] S2: Yeah. [00:42:42] S1: This year, which is pretty incredible. [00:42:44] S0: Like token economy. Yeah. [00:42:47] S1: And, like, I mean, us talking about, like, whether software engineering can really be replaced or not. Apparently, 90% of their code is now written by Devin, and that number was 13 in December. So just, like, six months ago. And my assumption is that it's because of Opus 4.6. Yeah. I don't Is that what you think too? [00:43:03] S2: Yeah. Yeah. Like, I [00:43:05] S1: feel like Opus 4.6 dropped in February. It's the... By far, the best coding model that we've seen so far. And so I'm assuming that that just, like, increased their numbers significantly. Yeah.
00:43:05 · Priya, Anika, Maya
[00:43:05] S1: feel like Opus 4.6 dropped in February. It's the... By far, the best coding model that we've seen so far. And so I'm assuming that that just, like, increased their numbers significantly. Yeah. [00:43:14] S2: Yeah. And they basically are seeing a pretty strong market wedge right now because Cursor is in talks with xAI to get acquired as... Or like SpaceXAI to get acquired for $60,000,000,000. It's like on the table. SpaceX gets to decide, but everyone thinks they're gonna do it. And the reason why that matters is Cognition will be left as pretty much the only well known independent coding harness, which means that they can switch between different models. And there are a lot of people that value that and, like, that's why there's still people that use Cursor as well. Yeah. And then I think... [00:43:44] S0: And enterprises value that actually too. Like enterprise contracts, the ones that they like... The big like banks that sign them for their software engineers. [00:43:52] S2: Yeah. Like one of the reasons they value that is like a lot of really big enterprises don't really wanna be locked in with one provider. Yeah. And
00:43:52 · Anika, Maya, Priya
[00:43:52] S2: Yeah. Like one of the reasons they value that is like a lot of really big enterprises don't really wanna be locked in with one provider. Yeah. And [00:44:00] S0: we can go into like what a coding harness is. So basically like you have the model, you can give us an input, which is like a set of tokens or a set of words, and then you get an output, which is also another set of words. And these set of words could be like Python executable, like language and stuff. And so there's so much that you need to build around that to make it work. And a harness is everything surrounding the model. And so it's things like giving the model tools or telling the model, or like basically dynamically putting prompts together. So you can ask the... You can, you know, add the prompt like you're a crazy software engineer. Now your task is to do this, or now you have this tool available to you. And so essentially what I think about a harness is just like ways to... Like if you're an input, like ways to optimize that input to get a better output. [00:44:47] S1: And so... And know your environment really well too. Because like every single enterprise is super different in how they use things and what softwares they use, like in general. And so it's basically like all the scaffolding around like the model itself.
00:44:47 · Priya, Maya, Anika
[00:44:47] S1: And so... And know your environment really well too. Because like every single enterprise is super different in how they use things and what softwares they use, like in general. And so it's basically like all the scaffolding around like the model itself. [00:45:01] S0: Yeah. And so the Claude Code is example of a harness, Cognition has a harness, Cursor is an example of a harness. Even things like open evidences like ask bar, like all these like chat applications, like ChatGPT is also an example of harness. Like, Harvey is an example of a harness. There's a lot of application layer companies that wrap a model in general. [00:45:25] S2: And I feel like one of the reasons people have been talking about harnesses a lot also is as people talk about not just having like a bunch of slop that can, like, affect your security or affect your entire, like, code base. Because one of the examples that gets given is, like, LLMs are not deterministic. And you can actually, like, see this. Like, sometimes if you, like, tell an LOM, like, hey, don't do this. It's not like an actual rule and it'll, like, ignore you, which I [00:45:53] S1: was gonna say. Honestly, he gets really upset about this. [00:45:57] S2: Well, it's just, like, insane. Like, I literally, like, written in my clod in, like, 20 different places, like, just stop using Updashers and, like, it'll keep doing it.
00:46:04 · Priya
[00:46:04] S1: Definitely, as we see more AI code that's generated... Because, like, cogeneration is so cheap now. Like, we all know that, and you can literally have anything do it. Things move to trusting the code that's written, like, verification. Like, it's the entire reason that you're, like, I'm still spending so much time before something can be shipped. Yeah. Because there is so much that goes into it on, like, the trust and verification layer. And I feel like there's been, like, Horizon incidents, like Horizon, like, so many things across, like, every company. Andrej Karpathy also said that vibe coding is passe. I don't know if you guys saw that quote. He was like, vibe coding is passe, and the real value now is in systems that can plan, execute, test, and iterate. And the really interesting thing... I don't know if you guys read it. I sent it in our chat. But MongoDB came out with this paper, and it was basically about, like, agent harnesses and the importance of them. And one of the examples was of Vercel. And so Vercel deleted 80% of their agent tools, and success went from 80 to a 100% in, like, outputting code. And the latency fell from seven hundred seconds to one hundred seconds, and they were way more efficient with the exact same model. And so that shows that, like, if you keep the model the same, like, the infrastructure that you build around it has, like, a significant difference in, like, what actually gets shipped to, like, production code bases, how fast it can be done, and also the cost that you're spending on it. Well, I