The Standup with ThePrimeagen - Microsoft Admits Defeat in AI?

Episode Date: June 13, 2025

💸 $1M+ hackathon is live — https://hackathon.dev Microsoft just open-sourced Copilot Chat for VS Code — but is it a W or a panic move? 👀 Prime, Teej, Casey & Trash break it all down: ⚔�� Cursor & Windsurf eating VS Code’s lunch 🧠 Is MSFT losing the AI dev wars? 🌐 Open source or just open copium? 💀 Also: mustaches, rats, and corporate trauma 📌 Chapters: 0:00:00 intro 0:03:25 Trash humiliated by a German Man 0:05:23 Prime admits he knows nothing about TMNT 0:06:00 Sponsor: Bolt 0:06:48 VSCODE Copilot going Open Source 4 Real 0:16:11 Inept vs Incapable (semantics) 0:25:37 AI Models vs AI Integration 0:41:20 Economics and Cheese Puffs 0:57:00 Klarna Klarna Klarna! 1:08:00 use code "Casey" at check w/ Klarna 1:09:42 Vote Teej!

Um, I'll tell Prime at the Prime again. Hello, it is time for stand-up. Don't be late. Trash, what you're watching right now? Uh, I'm watching Fire Force right now. I'm kind of late on it. I think they're like three or two seasons, right? But it's really good.

Because like when I watch dishes, like, I just need to like watch something. I don't even like hear. A literal zoomer brain. I can't even hear it because the water's so loud. So I'm literally just reading the subtitles. Nice. It's just like watching the other. I'm just like, oh, that's pretty good.

That is, that's relatable, though. That part's true. It's Prime. Oh, hey, Prime. Welcome to the stand-up. On today's episode, we're going to be talking about Klarna and other stuff. Can you make money off of other people forgetting to pay things? And is Microsoft actually open sourcing things or just still evil?

Find out on today's pod. TJ. That was beautiful, TJ. That was beautiful. TJ, this, okay, that was my jobs. Can you stop? Well, if it's your job, explain the introduction of the last several episodes. Let me help out here.

Ladies and gentlemen, please put your hands together for today's podcast host, The Primogen. Hey! Great to be here. Thank you very much, Casey, for that introduction. But you know what really needs a good introduction is today's episode. We're going to be talking about VS code. Why did they open source copilot? What does it mean?

and a second topic of clara yes should loads on burritos be a thing have we finally iterated to the final form of mankind or is there still room left to just get some extra money out we have two very serious topics

as you know we cover things very seriously here on the stand-up we're trying to talk about VS code going open source I guess as the first topic but here's the thing co-pilot of VS code going open source VS code is already open source Right. Copilot. And I would just like to say that I am going to sit here and watch because I have nothing to say about this. So go. Go.

All right. I have a lot to say. You have Lase. This is about Microsoft the company doing things. Okay. He's going to say has nothing to say. And then like he's going to go in a 10 minute or like five minutes later. So it's fine. It's fine. That's right. That's right. After we say something stupid, he can correct the record, which would be very helpful. Yeah. I don't know anything about VS code. What do I know about this code? My computer is dying right now as we speak. Well, that was trash. That's a really solvable problem.

Thanks for saying. Just install Arch Linux. Just install Arch Linux. No problem. Just start vacuuming, dude. Everyone's favorite podcast guest was Dax. And so that's the secret. That's what you've been missing.

Okay. Did you get the same headphones as Dax? He got the same ones as me, okay? I was about to say they're actually having the exact. same headphones and I realize it right now. But imagine people buying the same headphones, you know? This is Mattian, these are built-

It kind of looks like you're trying to copy his swag right now, too. You're going to start growing a beard? Yeah, what are you bald also? I mean, I've, I am bald, but I can't grow a beard, so I'm pretty sad about that. Every time I try, did I tell you this story? We're about to go on a side rant. Hit us.

I was working out an agency, and I tried to grow my mustache out, and I was growing out for like one and a half, two months, and my manager walks by my desk, stops, turns around, walks up to me, points in me, and goes, you should shave that. And then walk away. I was like, what the hell? Harsh. That's called direct actionable feedback.

Everyone asks their managers for that. I think he was like from Sweden or something, so he had like the cool accent. I don't, what's a Swedish person like? You should shave that. Is that what they sound like? I don't know. Yes, that's what they sound like. No, trash.

One more time, what was that like again? I don't know. I'm not going to say it. Was it Bruno? He reminds me of Bruno when they talk. There you go. But he just said that and I was like... Isn't Bruno like super German? Isn't he like, yaw? You should shave it. Yaw.

That's got to be some kind of Title VII violation. Can you tell an employee to shave a mustache? I don't even know. Yeah, why not? To be fair, it looked terrible. I looked like a... Okay. I think my now wife said I look like a master shredder or something.

Okay. The front ninja turtle. You know, oh, he was like the... movement. So, like, I can grow the hair, like, long right here. She looked like a rat, bro. I know, that's what I'm saying. But, like, it can only grow right here, and it's, like, always empty right here.

So it just, like, gets long. And it just, it's terrible. TJ, by the way, splinter would be the rat. Shredder would be the bad guy. Shredder has the, like, metal. You're talking the bad guy. The master splinter. Was the splinter's a rat?

Because I knew he meant the rat was. You wouldn't see any kind of mustache. We were on the same page. We were on the same page. I messed that up. Yeah, I know you guys are on the same page, but my autism does not allow such jokes to persist. Yeah, so we were both wrong, but in the same way, Trash. I've heard that, and he was immediately like, what are you talking about?

The canceled trash is fired. We have nothing to do with rats when he says Master Shredder. Okay. Sorry, sorry, sorry. But she was talking about the rat, though, to be clear. Yeah, I figured he was. 100%. Can I tell you a little bit of a truth?

Can I tell you just like a quick truth? Yes, of course. When he said Master Shredder, I thought he was talking about the rat, but instead I looked at Chat, and Chat was like, Shredder like, I actually had no idea. That's stolen Valor, Prime, stolen valor. Stolen valor. It was so bad. Oh, my goodness.

Actually. In between the different things you need to type into your arch install with someone telling you which teenage mutant character it was. Yep, yep, yep. I'm actually embarrassed. I messed that up. But anyway. Okay. Okay, Hall, we should probably set the stage again.

Join the world. Biggest hackathon in the history of the world. With over $1 million in prizes. Oh, that is a lot of yoga balls. Just visit hackathon.dev to enter. Obviously for senior dev zone. Literally anyone can enter. It could be your first app or your 1,000.

Uh, what? Obviously, me and Prime won't be entering. We're going to be reacting to the award ceremony at the end of this hackathon. Did I build balls.orgia for nothing? Both of us are really excited to see what you build this month. Remember, visit hackathon.com. For how to enter. Rules and restrictions obviously apply.

But seriously, join the world's largest hackathon. It's going to be fun. Go to hackathon.com. For more details. Can you set the stage and then give your take? All right, yeah, I'll set the stage. I'll set the stage. Yeah, yeah, yeah, yeah, yeah.

Yeah, yeah, yeah. Anyway, sorry. I'm Tej. We've got Casey. We've got trash. We've got prime. Here's our topic that we're going to start with discussing. But of course, we're leaving in Trash's mustache story.

Don't worry. So VS code recently decided to open source MIT license, too. So like proper open source. The co-pilot chat extension so that people could see that and also integrate new AI features. into vS. So my take on this, they also open source of a few other things.

Yeah, Weasel. Weasel. Like Weasel, yep, and Weasel too. Windows subsystem for Linux, for those who don't know what that means. That's not a made-up thing, like when we were saying, like, Florbenschmorgan,

the Apple OS, MacOS operating system extensions. No, Weasel. So my take is this is not Microsoft saying that they love open source. No surprise. there. I don't think that they do. Instead, I'm very confident that this is Microsoft declaring just absolute defeat and admitting defeat that they are unable to ship something that developers love and they need the community to do it for them. They see the clear writing on the wall

that they're getting absolutely maugged by cursor and windsurf, literal forks of their own projects getting purchased for billions of dollars and they're unable. With their own money. With their own money. With some of their own money, they are buying a fork of their open source project that they own, that they had a head start on, and they are saying, look at this, we're getting mugged by a bunch of kids from San Francisco. What is happening? We don't know how to win. So instead they say, we're going to abuse our distribution.

We're going to abuse our edge. And we're going to open source this so that other people instead will integrate into VS code so that people can keep using it. So that's my take. So I... Why is that abuse, though? Why is that abusing? No, no, I mean abuse and not like the action, like in the like game mechanics sense of like they're going to take advantage. Yeah, yeah, yeah.

No, like I think it's perfectly legal and fine for them to do it. But I do think it's not some like suddenly they've decided in the AI space they love open source. That's clearly not true. Otherwise they would be like working hard to try and make their core extensions like for C. and for Python available in other VS code forks. But instead, they actively license those away so that people can't. And or just say, oh, there's a maintenance burden. Oh, there's a maintenance burden?

You're Microsoft, dude. You guys are the supposed to be to people that make stuff work from Windows 98. And you can't figure out how to make your, like, little thing that you own the extension and the editor and everything else continue working in other forks. No, you are intentionally not prioritizing. that work, which is fine. I don't think they, like, are obligated even morally to do so, but, like, we can drop the act

that Microsoft actually, like, cares about open source. Can we just all stop pretending that part? So there you go. He doesn't care. Obviously, they still have the Python, right? The Python LSP extension is still not open source. There's still, I assume the C++ side is still not open source.

Their C1 is also not, like, open source, and their Python one prevents you from running it outside of Microsoft project. like buy license. Yeah. Yeah, they're actually very clever. So kind of to backup, I have some obvious disagreements with you, or maybe some not so

obvious disagreements. But I will back up your case that they don't care about open source in the sense that they created LSP as an open source standard and then immediately released multiple closed source ones. I think they later on open source their HTML one, though I'm not certain on that. But their HTML server, I don't think, or maybe it recently has become open source. but they immediately went closed source and then got the community to make all the auto-complete open source.

That's the reason why their editor, I mean, truly in my opinion, it's why the editor became so popular and why bespoke languages were largely available to become popular, were LLVM plus LSP's. Like having those two combos allowed all these languages to flourish in a world that was very hard to flourish in, and VS code is able to create something

in which allowed a bunch of people to give them free labor, and then they're just still writing on, the top of it. But I think Microsoft, I, this is- Can I say, I don't think it's actually about free labor. I think just like internally structured, they're unable to ship something useful, like, at the level of windsurf recursor. Like, I just think, like, it's not, like, they could, they continue to pay people a lot of money. I don't think it's about the money. I think, like, to me, it seems structurally they're incapable of shipping, like a complete package that people are excited about.

So I don't, what's the last Microsoft thing that they did. I don't think they're incapable. I don't think they're incapable. I think the people who, run VS code, we're running it since the days of GitHub. And their whole focus is that VS code is creating an environment for developers to work in, and that really what they want them on is GitHub. They want them on these virtual environments where you run VS code in the cloud. You want everything's an Azure play. It's all running towards that. And when the world underneath them shift, they got copilot because Nat Friedman was smart enough to observe that the world is changing and pushed GitHub to make co-pilot,

he's no longer at GitHub, and they have not done anything good since he effectively left. I actually truly think he was actually the real brains behind a lot of probably some good moves that they've done, and they just haven't iterated. I think it's just simply they didn't believe it was important. Can I insert a question here? Because I'm just, again, listening to what you guys are saying.

So is the idea you think that Microsoft is, wants the open source part because they want the community, they want it to be developed that way, they want people to adopt it. But they are trying to retain enough control to ensure that those people who adopt it will then have to pay for Microsoft cloud services. Is that what you're sort of implying the business model is? Or I just want to make sure I understand what you're saying. Yeah, or other Microsoft, like, it doesn't have to literally be. Azure, you know what I'm saying? Like, it can also be GitHub Enterprise-type integration, right? Like, the entire vertical stack of, like, Microsoft products, right?

Right. From where you're going to use Microsoft languages, you're going to use Microsoft, like, hosting, you're going to use Microsoft Cloud, you're going to use Microsoft AI, right? And if people are switching off of that into, like, cursor or something, right, or WinSurf or whatever, and then also using, like, Claude inside of it, Microsoft's like, oh, that could be bad like that's our yeah you think that basically what they're right now

trying to figure out what do we do to push people and maybe keeping something as closed source as part of that is that like because like one of the things you're saying in closed source is like the Python part of that or something and you think that's just because they need to retain some things

that can't just be forked and cloned because if they don't they won't have any levers to pull to try and keep you on some kind of Microsoft drip feed that, you know, where the money goes to them. I'm just making sure I'm not just agreeing or agreeing. I'm just asking, like, is that what you mean? I think that that was their idea. They just didn't, I mean, to me, it's just like they had the ability to execute and

create a good editor. I think the rest wouldn't actually have mattered nearly as much. I think they could have open sourced all those things. And people, like, obviously, VS code, like, by all objective standards is a very successful text editor. Like, out of all of them, it's a very, it's probably the, most successful text editor to have ever existed. And so it's not like they didn't have, it's not like they're like, oh my gosh, we're worried

about forks. They could have made everything open source and it could have gone that way. They just simply miss, to me it seems like they've miscalculated the importance of AI. And so now they're just going to try to make massive plays to effectively destroy cursor because my sneaking suspicion is that Microsoft observes cursors, observed these forks and realize that they completely leave open AI. They could completely leave a lot of these things and slowly develop. their own models that will eventually not need any of the products that Microsoft either

adjacently owns or directly owns. And so I think they see it now as a threat. That is how I look at it. So they're attempting to beat them by beating them free. So they're basically worried, oh, actually now we're going to have these giant data centers filled with tons of Blackwell GPUs or whatever else we end up putting in there. And no one will be using them because they'll be using our VS code, but with somebody else's, like, AI plugin that use someone else's data center, like Google or whoever. That's the idea. That's the fear. Well, I think they, they're, like, I think, prime, we're pretty close to agreeing, but like, I, I think there's, like, you're inept or incapable. Mine is they're inept. We, like, we, we agree fundamentally.

It's just that you think they're incapable of doing it. I don't know why if on the table is own every, own, all of development. like editor tooling that's on the table, they're saying, we'd prefer not. I just don't think, I literally don't think they have, I think the people who are running VS code

don't know how to program and they don't understand developers. That's why they are able to, I think it started off as developers making something for developers, and they literally just took VS or a Visual Studio, which was a very successful product in of itself, and so let's make it open for the web.

And that was very successful. And then they knew, never figured out anything else because they didn't they couldn't just draft off a product that already existed and so they're just like i guess this is what we're doing now and they never had another original idea afterwards and they just kept on just like resting on their laurels and so that's where i think the state of it is and so i don't think it has anything i think it's management but i think we both agreed they're not doing it i made that tweet two weeks ago that said that that that Microsoft is incapable of competing against cursor and here we are yeah so wait

Are you saying they are incapable? I don't, whatever. I don't really want to argue the semantics of it. Like, the thing that I wanted to say along Casey's line and then trash, I do want to hear, like, what you're saying is, I think if, like, in my, in my head, I'm viewing the situation, Microsoft started with a closed source AI product. And, like, compared to some of these other, some of them closed, some open source products, like, they've failed. Like, I just don't hear, it's, to be fair, it's maybe my bubble online or whatever. But I never hear in the conversation of like AI tooling people talking about how often they use co-pilot and how much they love it. Now, maybe that's not, maybe that, maybe they have usage numbers that are different or whatever.

That's fine. But like I think they would have kept everything closed source if they felt like they could keep charging money for it and win. Like, but in my mind, it's a cleared mission of defeat because they're not doing that. Same with any time. And it's, I don't think it's like some sort of like moral wrong, like to have close. software or whatever, right? Like, I think that's fine. I think people should be able to make and sell whatever software they want. If they want to be open source and free, that's cool. If they wanted to be closed source and not free, also cool. I don't really care. But like,

Microsoft is not, they don't love open source in the sense they're not going to like do something that hurts their business, I feel to like do an open source play. So they are admitting defeat, I think, to these like forks and saying what we need instead is other people do come and integrate their AI stuff into VS code so that instead of having, this fear where cursors on a separate marketplace from Microsoft. That's what happens. They have separate marketplaces to do the apps. That means Azure may not get a privileged spot,

or maybe it won't be as tightly integrated into the other parts of this tech. For me, that's where I see their fear has to be, is that all of these Microsoft land things, like, you usually don't pick Microsoft something because it's like has this one product you love. It's like, oh, well, it's because they have the 30. things we can do altogether, ostensibly working together, sometimes maybe not whatever. But like, that's at least what's supposed to be happening. Right. And so cursors on a separate plugin extension marketplace. That's dangerous for Microsoft. I think that's like the point of

VS code is to get you into the Microsoft extensions. So that's, that's kind of where I'm at. Trash? What do you think is? Hold on. Actually, I didn't realize there's on a separate plugin marketplace. Was it, what is that? So why am I like, say I use VS code right now and I, switch to cursor. Yep. Are they like completely just, I guess like how do I use the same plug in? So I'm pretty sure you can import stuff from VS code and cursor.

So there are currently ways that like people can publish to both marketplaces. I think it's called like VTSX or something like that. It's like the open VS code marketplace and you can run it in VS codium and other like more, you could say properly open source like VS code forks or properly forked VS codes, right? the sense that like it's no longer a Microsoft product because and generally this is not a huge issue except if you like VS code to use different Microsoft things because lots of those have licenses where they may only run inside of Microsoft land things like the C extension some of the C extension things and some of like Pileants for example and various other ones and I think Microsoft probably

has some other ways that they like for example co-pilot was using APIs and other stuff that are not available in the extension marketplace. Like, Copilot was a blessed plugin because it's made by Microsoft and they can like cheat, not in like the literal legal sense, but they can like cheat as developers by using APIs that aren't available to other

tools. Just like Windows back in the day. Wow. Okay, they're back to that, huh? Oh, yeah, it's almost like, it's the same game plan, guys. I'm not taking crazy pills. I'm just recognizing the patterns. You may be taking

crazy feels and recognizing the matter. Need to change this background to the, to the, it's always sunny background. I have one thing to add, because I think you all pretty much hit everything. Like, I am,

I do side with, Wittage, like I do think, well, I don't think they're incapable. I think they have definitely capable engineers. What I think is happening, I think people are,

there are people skilled enough there to literally type the code to make it happen to like make something. Yeah. Well, what I think is happening is they're kind of just being slowed down by the big machine that is Microsoft. And that's kind of just slowing down their roadmap.

and they're just kind of held down by bureaucracy, in my opinion. And if you open source it, obviously people can just work on stuff that aren't even like in your peripheral. And like, oh, man, that was a good idea. Maybe we can pivot to that and kind of. But then there's like the aspect of leveraging free labor. I don't really think that's a part of it. I actually think it's like the way engineers like kind of operate at least in my tech bubble is like, oh, it's open source, their goodwill. And they kind of get like brainwashed of like, this is a good thing.

So I'm going to use them. Because you hear, I hear every engineer say, I use this tool because it's open source. And that happens like nine times out of ten. Like if it's open source, I'm going to use that instead because there's transparency, et cetera. So I think that's like a part to kind of like skew some of that target audience to start using VS code. I have used like the copilot integration.

I use cursor as well. And yeah, cursor is like hands down like way 10 times faster, especially like on big tasks. So I can see like, you know, the reason to want to go to cursor and I use VS code. And I'm sure they're seeing that obviously watching the market share kind of teeter tattering and watching people paying money for these forks, which is like the most hilarious thing to me.

So, yeah, I don't know. We'll see how it shakes out. I was listening. I was reading some research about it. Some people from their team were kind of saying, like, they wanted, they open source it because they wanted to standardize the UI. They said the experiences were just so fragmented between cursor and VS code. I don't actually believe that. I think that was kind of just like, you know, they had to say something.

to open source to get separate editors with separate motivations to become the same UI. If Linux has taught us anything, it's that open source leads to a very unified user interface across all the products. All the different distribution.

There's just one version of Linux. There's just one version of all these things. And the UI is always the same. It's always exactly the same. You know, you install Ubuntu, you install your arch, everything. Whatever it is, it's always exactly the same UI. And that's the lesson that I can fig also save everything you know it's beautiful i actually think like another part of this so

when i did work in the government space a lot government is have us government specifically we're really heavy on Microsoft tools um and uh i read somewhere like they needed to be transparent with their tools for some compliance reason i don't know how true that is but i can imagine you know d o d or something would only allow their you know development developers to work on such like high class products if there was some transparency on like what their AI integration is sending to the server. Right. I think that's like the black box part.

And I feel like this is like definitely part of that play. But I can't prove it. But I think like obviously like the more transparent, the easier or the more relaxed the government feels about using these tools. So I also think that's going to be a big, big play on their part. Can I ask another question of you guys while? Oh, can I say one thing on Trash's thing? which they are not open sourcing the back end parts of co-pile. I'm just talking about integration.

No, no, I know. I'm just saying like, yeah. Microsoft is clearly, in my mind, framing it like, look, we've open sourced our entire copilot stack.

Like, copilot. I mean, one, you don't know which co-pilot they're talking about. They've got 37 products named copilot. So you never know which, what they're actually talking about when they say copilot.

Yes, right. That's what I'm saying. He's like, you open source your whole OS now? And it's like, no, no,

editor. Can you imagine what Normies are thinking when the industry? But anyways, um, it, they're not open sourcing the back end part. I thought they were saying, oh, there's no more secret sauce anymore. It's all, AI's got so much better. We don't need the secret. Okay, open up, open up the backend. There's no secret sauce. I think the big thing is that they realize that the, I, at least I personally think the money is in the integration. It's not necessarily in the AI. And the reason why I say that is that, uh, Deep Seek effectively just proved that if you have access to the hardware, it's dramatically cheaper because you can just brain drain other AIs.

Like I just don't think it's nearly the same open moat that it used to be. Or it used to be this wide mode where you could just be able to be like, I have the best AI. Now people are like, I just distilled your model. Now my model's as good as yours. And here we go. It's like all about the distribution. It's all about the availability.

It's all about the speed. And so that's where Microsoft wants to compete. And so they're like, okay, we're going to let anybody integrate. so we want to be looked at as the best because now you get to choose how it's done as opposed to being in this closed world cursor. But really, I just think that, I mean, I'm with TJ pretty much 99% of the way. I just think they don't care. They don't care enough. They had two choices, which is they could get a whole bunch of talent and actually compete with cursor or windsurf on a real platform.

Or B, just give up the ghost and say, we just don't care enough to do it. And so that's where I think it's not incapable, it's indifference, maybe. Maybe indifference is a better word for me. I would find that indifference a little hard to believe, though, because indifference would be like imagining that Microsoft didn't have 1,500 meetings about this where people were disagreeing or something. Like, this didn't just happen. Like, there was, there was not indifference, is like a something I would say. They're committing a ton of engineer hours to making this happen. Like, they're going to roll a bunch of stuff out of co-pilot chat into VSCO core.

Yeah, but it's like how many hours are they actually committing? a thousand a month. It's just not a different. That's what I'm saying. It's such a drop in the bucket what they're actually committing, like comparatively to what they have available to themselves. Like if they put 30 engineers on something, they virtually put 0% engineers on it.

Okay, yeah, yes. It's not true really because it's going to still cost them millions of dollars. I think they'd rather have millions of dollars back or like they'd rather have the play where they own an editor. someone wants to buy for $10 billion? If something costs $10 million, it's virtually nothing to them. I mean, yeah, remember, they made $25.8, I think, billion net revenue after all operational expenses, including engineers and everything.

So, like, they make so much money. When you say $25 million, you're talking about, what, 0.1% of the total net revenue, which is what, 0.01% of total revenue? Okay. So I guess you're just saying Microsoft doesn't like money. Like, because I'm saying, I think they don't, I think they don't bend over to pick up $500 million. If there's like $500 million on the ground, they're like, no, we want to go over, we want to go after like the 10-months. They want as you wear to be a success because they look at

Amazon and see Amazon being the, I mean, Amazon, AWS makes Amazon money. If Amazon did not have AWS, Amazon would be out of business right now. And so back to trash, back to trash. He forgot. He forgot. He literally said he forgot to me. You raised your hand, technically. I know, and then I forgot. No, no, I think, I think Prime said something about the money is in the integration. Yeah. or something like that. Yeah, money's in a name. I think this is all about, like, Compu, because this is what you're paying for, right?

So the more people you send to your servers, that's where all the money is, right? I'd say it's, but I think, um, I think, like, for VS code, like, VS code does represent like a significant,

like, investment from Microsoft over many years and like, as well as a serious part of their overall strategy. Like, they care what happens to VS code. I, I, like,

it matters. And they care, too, if, like, a bunch of $10 billion companies get made running VS code forks and they can't capture any of that money. Like, I just don't, I like, I don't really think that that, like, they want to prevent that from happening. I think a play, they have multiple options to prevent that from happening. One option would be making VS code as good as cursor. You're saying they could just put 3,000 software developers. on it and they wouldn't care. So then I don't get why.

Cursor does not have 3,000 software developers. They have nowhere close to 3,000 software developers. I don't know how many they have, but it's nowhere close to 3,000. So like, are they capable or not? You're saying like it would cost them nothing to just make cursor. So I think so, but they're choosing not to make cursor. They're choosing to make it open source that other people can come integrate in their thing and they can keep people in the VS code ecosystem. I'm saying that's because I think they can't do it.

otherwise why would they not? I don't understand. It's like, I don't understand. You're just making a very compelling argument here. Because, yeah, especially if, Prime, if your argument is that this money doesn't mean anything to them, why can't they pay the $500 million you don't think they'd bend over to pick up? Like, why can't they just pay that money then to the same number of developers as cursor and have a cursor compete? And I think a reasonably good answer to that question is institutional incompetent.

the same reason that a lot of things on Microsoft are bad. Why is the start menu a train wreck? It's not because they couldn't have the money to go pay someone who's ever programmed graphics before, right? Or a UI. They could do that, right? They could go find someone. You know, they could go hire someone from some awesome looking game UI

that was amazing and had all this crazy stuff. They could hire those people. But institutionally, they don't have that ability, right? I don't know why. I mean, I've had interactions with Microsoft, so I see some of the like facade of why, but technically why, like, why can't they go spend half a billion dollars and get a great product? I don't know, right? But they don't seem to be able to do that across any Microsoft product. Like, they seem to be very, very incapable of that kind of motion. And I, you know, obviously that tends to associate with company size. It's what we think of when we talk about things being like IBM or something, right? But I don't actually know exactly what the pathology is. But I'm kind of more with T-J in this one.

I apologize. I'm just more with T-J-A-N. No, no, that's fair. It's fair. You know, you guys may be correct. Maybe it truly is just incapability. That's preventing them. I just have a hard time believing with a company of that size

that there's not enough talented engineers that have enough voices to say, hey, we can do this. We should definitely do this. Whereas there's people that are just like actually just inept at their job. It's not that they're incapable. They just have stupid leadership. in just the right locations that can't see, you know, the force for the trees. That is a type of incapability at the point.

Right. Yes. That is institutional incapability. It doesn't mean that you have your individual developers. If you picked certain ones and put them together and gave them the resources they need couldn't make cursor, it means that institutionally you can't do that. You don't have the ability to do that thing. Yeah, that's the argument I'm making. Like I know someone like personally who like worked on VS code awesome engineer constantly cut out at the knees to like make improvements to LSP specification and other stuff like this.

And it was just like, oh. So I'm not saying I think there's a bunch of people that are way better software developers that work at Microsoft than me. Like I think they're way better software developers. And there's some of them who's like insane. Like the guy who made Pyrite, Eric Trout, who like retired a little. bit ago and everything, but like, dude, this dude wrote an entire Python thing all by himself. You look at the commit history. It's just him.

It's powering all of Microsoft's like Python ecosystem. It's just like one dude. And, you know, obviously some other people helped and everything, but like, you can go look at the commit history and you can just go see. So like, and it's like, that dude's a thousand times better software developer than I am. He's so cool. He's like one of my software developer icons. But it's like, but okay.

But then, so then why didn't they just say, yo, Eric, do you want to go build cursor? like they just i'm saying they can't as a company not like that the people there okay i mean that's literally what i said with like the bureaucracy thing right like yeah what if it's something just make whatever chess move though like because they somehow won the editor okay so they're incapable of making decisions i 10 years ago they were just as incapable of making decisions as they are today if they've made plenty of good decisions do you think there is a world that exists where they're actually still making really good decisions we just don't understand it i think that i think the decision to open source co-pilot chat is a very smart decision by them, given the constraints

that they have. And clearly they have a bunch of smart people, both who are probably convinced internally that Microsoft is doing this because they love open source. Like, I don't think a lot of people they have on like PR campaigns are lying. You know what I'm like? I think they're probably like excited that they get to open source their work and that they think it's going to be in that win for people and possibly like it will be in that win for people. Like, I think open sourcing it is a smart move. I think they literally, I'm saying my perspective is, if I'm sitting inside of Microsoft and I'm somewhere up the chain,

and I'm looking around and I'm seeing cursor and Windsor and all these other guys, and I'm saying, we were first. How are we not here? And all I hear back is, well, don't worry in Q3. We're also releasing an agenic drop down. You'll be able to select your model too. It's just going to be Q3. And you're like, Q3, we need to go faster.

So the option given the constraints that's really smart is let's open a bunch of other people do a bunch of things inside of here, integrate their companies inside of our product and be inside of our ecosystem, and we'll continue to embrace them. We'll allow them to extend. We're not going to talk about what happens next. So there's also sort of the, oh, sorry, go ahead. No, you can go ahead on this one. I've been, I've been yapping a lot. There's sort of the old Amazon principle. I don't know if they do this still, but in the old days, they used to. to kind of have this idea where you would have multiple teams doing the same thing and you'd pick whichever one worked, which I always thought was extremely smart because I'm like, that's what actually, that's what you actually need. Because even if we handpick some engineers, they still just might not come up with the right thing for whatever reasons, even if they're all very good, right?

Everyone who's ever programmed knows sometimes things come out, sometimes they don't. You make some bad decisions, whatever, right? And so I think like the other problem could be inside Microsoft, as far as I know, they still kind of have the mentality. that like someone's going to own this thing. Like this is the group that's doing the, you know, the start menu now. We don't have 10 groups that all do start menus than to pick the one that's best. So the other thing is like, well, cursor, windsurf, whatever, you know, Microsoft may have been on equal footing for whatever they were making and theirs just didn't pan out as well,

just for random chance or whatever else. And they didn't have other ones, right? So sometimes you just have to go like, all right, if we're not going to embrace that entropy, if we're not going to embrace the multiple die rolls that we need, then we have to figure out a way to just accept that someone external is probably much of the time going to be the people who do this better. And we have to buy them or have them integrate with us or we have to have some other strategy. So it could be that too, right?

Microsoft is awesome at making platforms. Whoa. You mean like 30 years ago? I mean in terms of turning their platforms into making money, not delightful user experience. I was like, whoa, we went off the rails quick. You said, whoa, 20 years ago. Yeah, but right, I was, okay, I know I will never be nearly as successful as Microsoft, right? Or like a bunch of people own, like in the vertical.

Yeah, that sucks, TJ. Well, all right, we're almost there. Check out SSH Sherville.competre. We're transforming terminal-based e-commerce, boys. if you're in a select few places in the world where we currently can ship, you can buy coffee from us. But like,

I mean, so that's what I think. I feel like Microsoft bread and butter is not pleasing users. It's building platforms that they make money from and other people are forced to use. Is that not how they're making $25 billion, a quarter or a year, whatever?

Quarter, yeah, yeah, yeah, net quarter. Okay, okay, so I guess the follow-up question is, do you think Microsoft will do the final E? Right? Because they are embracing. They are extending and using the ecosystem and their own power to extend their editor into the next levels. Will they then either extinguish by making the most successful product or will they actually just purchase out? Like does Microsoft has a history of purchasing out good ideas or do they just extinguish the ideas?

Yeah. So trash, any blockers? Probably made me drop. Prime made me drop is the most child answer I've ever seen in my lifetime. I'm wearing all black and I dropped a cheese ball and it just rolled down my shirt, down my legs. I have like a cheese trail. Prime made him drop a cheese ball. That's crazy considering I'm thousands of miles away from you.

Prime made me mute my mic and then I dropped my cheese ball. So I'm a little upset. Okay, to answer your question seriously, I just want to get trash in on the mix again, of course. I think, so I guess I don't know because I have a really hard time predicting anything that's going to happen in like the AI space. I think it's very, it's very surprising what happens all the time. Like, I don't know, is Google going to win now? They just released a bunch of stuff. And all I've been seeing is like crazy video generation from Google, like for the past two days from Google I.O.

You're like, I don't know. Are they going to win? I have no idea. I'm just my broader point is more like, I think if Microsoft can extinguish the competition through this method, then they will do so. If that means that they're going to, there's some new AI mode of editing that comes in the next year, whatever that could be. There's like, oh, we develop this way that we can hook up into your brain to do it or something. I don't know what the next step would be. Then Microsoft will first try to make that available only to the co-pilot extension without open sourcing it,

into VS code proper so that they could capture that completely for themselves. And if not, then they will do the same thing here, where they will open source that and hope that people build the integrations into their editor so they can continue being a platform that, like, the majority of software developers use, I don't even know. I don't know what the actual number is stack overflow is kind of a lie. It's a bit of a biased audience right there. Yeah.

So like that's, so that's where like I think, they are and that's why I'm saying like this move is an admission of defeat because otherwise they would have extinguished the VS core vS code like fork competitors right by doing something inside of the S code that was not available I hate you trash I don't know why we let you back on I'm loving it personally so there you go that's so I think if they could acquire I mean my understanding is like open AI attempted to acquire Cursor and Cursor said either no or they wanted

more money than Open A wanted to do and then they went to go get windsurf instead. I don't know if that's true or not. That's what I read on X the everything application. So your mileage may vary on that one. Can I ask a question before he moved

on to Clarna? Yeah. My question for those of us who literally don't care about any of this who may be some of the small very small percentage of the audience and me what is so when you use something like cursor what is cheese puffs what is what is a cheese trail i was like starving i had to eat i'm good now i'm ready flap let's clap what is the actual revenue model end game for cursor so when you use cursor what are exactly you're paying for?

Is it just how, you know, based on your AI usage, it goes to their cloud, well, whoever's cloud they're running on. And they charge you a markup on that compute. And that's what you're paying for? Do you pay, like, is that the entirety of the economic model or is there more to it? You could just give me, give me something for those of us who don't care. Go teach. But want to understand. Okay, my understanding, generally speaking, there's currently, I would say two major, like, money.

strategies in the editor space there would be one is a fixed rate uh like with different tiers of like here's you're going to get this many like basically like high priority large model big kind of queries blah blah blah they're going to pro rate that to make sure that like on average they're going to make money they're going to try and figure out ways to like reserve big compute clusters in advance to like bring that price way lower than like you would be able to do on your own right be like oh, it's going to cost me $0.25 to run this query. It costs them $0.3 because they've pre-reserved these, blah, blah, blah. And so they provide this, like, bundled subscription experience that, like, probably ideally

gets paid by your employer so you don't have to think about it. And you're just like, it's a requirement for me to work here. I have to have Cursor Pro Plus experience Enterprise Edition, right? So there's like a headcount kind of thing to this. It's like, we're going to pay an extra $1,000 every week for every employee for all this compute or something. But that's just what's going to happen. Okay.

And then there's, and like you see this with inside of cursor and like sort of as an alternative, one that's like almost purely just usage based where they say, hey, maybe they have some like free tier with small models. They can run very cheaply, you know, no matter what you could not, if you're running 24-7 for the whole month, you wouldn't be able to, you know, cost them more than like $10 or something like that, right? and then to do large queries you pay like an additional on demand kind of thing to say like hey this is going to be like 13 cents to run this one and then like you spin that up with token so like for example i've been playing with open code which is just like a more of like a claude code kind of thing cloud code does the same exact thing just runs in your terminal you pay by the token and that's it right and there's probably some either markup you would do. do if you're cursor or somebody else, I'm assuming, or it's still cheaper than what you'd be able to get buying as a like individual person, but like they're going to have some markup over their costs, but they can bundle up all the costs and do all this other stuff to try and make

running the model cheaper than you'd be able to do. So those I think are the two main like ways that you see like the pricing strategy for these editors going because most of them are not like not integrated in some other way like Microsoft is as like a platform, right? You, you, you probably will see other ones doing things like, these wouldn't be the VS code forks probably. They would be some other thing. But like, hey, easy deploy to our platform, super deeply integrated to like our stack. Right. So you might pay some like monthly subscription fee. The additional benefit is like, oh, deploy instantly with this button. It also knows how to talk to Stripe or it also can like, we preload it with all the latest React stuff

and it's like perfect for doing your thing or whatever, right? I even saw one recently for like Elixir that was the guy who made Phoenix, I think, or one of the guys who made Phoenix. Yep, right? And he just built one that's like, this one's for building Phoenix apps. Oh, see you, Trash.

Oh my gosh. Okay, well, this ruins the live experience. Thanks, trash. It improves it. Yeah, back. Oh, there we go. Oh, that's great. Thanks, trash. So that's the other option, too, I think, Casey, is some of them may be able to go much closer to at cost if they have some other platform experience that they're trying to, like, build off of or integrate with, like, the editor thing.

Okay, this is probably the, oh, sorry, Prime, go ahead. I was going to say that, can I just say that in a bit more simple of a way? Yeah. Effectively, they pay X for usage and they want you to pay Y, where Y is some C of X that's larger. than one. That's effectively how they make money. And their goal is that you don't use as much maximally. Because most people don't program. They have it all priced in. They're probably going to, and my assumption is that down the road, they'll have their own models. They're probably already researching their own models such that they can do a lot of these smaller things themselves,

save even further money. And their goal is just to shave money off. Just like people who make stuff on top of AWS, they make money on a per request style model or on some sort of group of request model and they get a reduced price. So it's just a reseller with bulk purchasing is as far as I can tell, but I'm sure there's other things that I'm missing and there's future plans that I'm missing. And at the risk of asking a sort of question that maybe this is the wrong audience to ask, but what happens when the models just run locally? You don't have the integration. So that's why I said I think the agents or the integration is where all the money's at is it's not just running the models. I think there's going to come a point where a lot of these models are going to be more and more available. I think they're going to make better and better improvements on how to make these things more efficient and you'll be able to run more consumer hardware, assuming hardware also improves. You know, like there's going to come a point where you can run pretty powerful models locally. I think that where it's going to come down to is who has the best integration, because we can all agree that if you had to go to chat gpt.com, paste in the files you want, get the answer out, take that the five different sections of code and repaste it back into your editor. That's,

a horrible experience comparatively to being able to be like, hey, cursor, here's my question. Take that file. Take that file. Take that file. Take that file. And it just goes and does it, right? Like, that's a real. And so my guess is that the integration is going to be a much bigger win. And if they don't allow you to use it or if they allow you to use it with your own models, but you still have to pay a monthly fee to use their integration, then I think that that's where it's going to be at. What does the phrase integration mean in this case? When I say that, I mean from the AI to the thing you use to make someone. something.

Yeah. So no matter what that is. So that can't be usage-based anymore, there is my point, right? So that would have to just be like... You use it based. Like Microsoft or like Visual Studios. Visual Studios is just like a yearly subscription. Same with IntelliJ.

Right. Like you could imagine a world where these things, if when or if it becomes ubiquitous that everybody runs their own models, then I could imagine a world like Intelli where it's like, hey, you want to use this nice thing. It costs 20 bucks a year for you to use this much better editor. Got it. I would say, too, the other thing is... Um, you're still going to have the problem of like, you're using a MacBook somewhere.

How are you going to load in the latest GPU and the biggest model? Yeah. So someone still like is going to need to run the model somewhere, right? Like even with the best like, oh, Lama advancements or whatever, it's still going to be less than a huge stack of GPUs. Right? So you're still going to have some cases where you're going to want to outsource, I think, large requests, assuming that in five years, we still think LLMs are good for doing like coding tasks or whatever, right? But like assuming that's true, you still like a lot of times probably want some thin client

to your actual GPU farm, right? And like in that case, I still think you're going to want to, at least for some of your requests or some of the things you're doing, probably end up paying someone else to run that for you. And yeah, I think like Prime is saying, I think like people pay for editors sometimes anyways, but that market's way less than getting everybody to spend $250 a month on your thing like Google Gemini Pro. I feel like there's a lot of risk here that's not really being priced in. And the other thing I would point out is what you just said, TJ, where you're like, well, you're still going to have to, you know, for these bigger ones, room full of GPUs or whatever, that's really not a business model that works either, though, because you can't have a business that's based on selling people something at a fairly low,

or reasonable cost that requires that much GPUs. I mean, you've seen this over and over again where people try to do games in the cloud. And they can't make it economically viable because you're like, well, like, it would cost as much as if you bought that PC. And here you're talking about a ton, like you'd have to have more than one GPU

to make this cost effective, right? I mean, to make this more powerful or significantly more powerful a lot of times. So I don't know, like it feels like there, if that was really where they were headed, I think they might be in a lot of trouble. Whereas if it's just like, look, this is going to be a simple SaaS thing, we charge developers for, you know, a certain amount every year.

That seems stable, but not a huge business necessarily, right? So I don't know. I don't think that's, I just generally disagree with you on that one. I think that they can make the money, that GPU's large-scale GPUs. I think games are fundamentally different. Because running a game is, like, if you played one round of Apex Legend, just one on a cloud GPU, what is the, equalized amount of queries that thing had just made.

Is that a full day's worth of queries that that thing just made playing Apex Legends in just 20 minutes of a game versus a whole day of just playing nonstop LLM queries? I think that fundamentally you use a lot more GPU with games. And even games, I mean, Netflix is doing its own venture into cloud gaming and they're becoming a bit of a success because they're not doing AAA. They're doing a different kind of game. They're doing, you know, double, you know, not these huge mass of 3D worlds. And so there is a world that does exist where it does work out.

And real time or server-side render stuff does fully work out. And we already do that right now with just say websites. Websites are already server-rendered, even though we should render them on clients. You can use a lot less hardware if you did that. But we realize all these types of gains. So I assume it's still going to exist that the big models, the big things, the big wins, or the super secret sauces. Like cursors tab complete is obviously better than everybody else's.

They're not going to be like, yeah, we'll just make it open source and everyone can just use it. No, it's like their secret sauce, and obviously running it must not be that expensive because they continue to offer for everybody. So they must cover their costs or they hope for one day to cover their costs, which, you know, maybe that's where the danger is. But I just somehow doubt that it's a one day to cover the cost and they are actually just covering their costs. Well, the only two things I'd say there is like, well, okay, that assumes that no one else could come up with a secret sauce, right? Because otherwise they could and then just run the secret sauce locally. but assuming that that were true, like, sure, if you can keep the thing hidden on the cloud so that no one else can access it, then you may just be able to charge for the cloud service because people can't run it locally because you don't let them.

But the thing I was going to say about the games versus deep learning stuff, the problem, though, is that in order for the GPU to be doing meaningfully more than you could have done locally, right? Because you're talking about something that now you're saying you have to run in the cloud, because we need that much compute to process this query. That has to be something that would require like many GPUs or something. It's not going to be something that just requires one GPU. So I don't know. Like, again, you're counting on those economics to work out in your favor,

and I don't know that they necessarily do, right? And you also have to count on you updating those GPUs because if you're just going to leave them sit and five years down the road, well, everyone's now got a much better laptop with much more TPUs in it, and so your data center is falling behind. and that capital cost goes up, right? So I'm not saying that I would place a bet either way. You could very well be right.

But it's also like, I don't know about those economics and have people actually figured them out. Maybe they have, maybe they haven't, right? And there's a lot of unknowns in there about whether or not, you know, like TJ, I think was saying, maybe we end up with things that are actually more efficient than these LLMs and then you really can't scale them. So there's a lot more risk, I guess that's why I was saying,

than maybe people appreciate. It would be the only thing I would say. Yeah, I would say the only sort of like counterpoint, even though like I do agree that it, I mean, everything for me and AI seems super risky. It seems super easy to like make everything terrible is the great philosopher, the Dax. You guys may be familiar with them from a previous episode. Oh my goodness. What did you just say, right? Trash 2.0.

Oh, that's that's a huge insult to Dax. Okay, trash 3.0. 3.5. Trash Sonnet 3.5. Trash Sonnet, haiku, ultra-plus, medium. Is people do

choose the lazy option. Like, you can already see right now, like, it's true. It's easier to run a website on your own hardware probably than ever before, and more people choose not to run it on their own hardware. I'm not saying that's, like,

an inherently wrong decision. it's just like the easier option. So there is still something to be said like, oh, if they could make a really easy experience for having this, especially if it's like for thin clients where I can continue having to buy a new MacBook every two years because Apple releases it in a new color.

And then like I can just like continue to use the GPUs in the data center. There is something like there that probably is like quite, quite useful and powerful, even if it becomes. all the models are open and all of them are just like able to be run locally with no licensing fee or something. There's still, I think like some, at least some utility there. I don't know if then it's a $100 billion company or not, you know what I mean? But that's more my question because then it becomes more commodity.

It's just like, which to be fair is maybe why Microsoft and Google and these places are just like, look, this is just a data center business. It's not going to be that much different than what we're doing now. which just get like, you know, yes, you could run it locally. We're hoping that you don't bother off you go, right? You know, I imagine a lot of data center companies live in fear of someone ever actually going, maybe I'll make a really good peer-to-peer like system because it just wipe out the entire business overnight, right? But yeah, it hasn't happened yet, so maybe they're just like, it won't, right? Yeah.

So totally fair. All fair. Well, have we beat to death this Microsoft opening the sources and GPUs? think so. All right, hey, no blockers here, huh? No blockers at all. They probably just wrap it up. Cut it. All right, are we done?

We'll do clarna next time. We can do clarna next time. We can do clarna next time. Yeah. All right. Well, that was wonderful. Thanks. That doesn't sound sincere. That's the conclusion. Yeah, that's wonderful. That was like, thanks. I was kind of terrible. Part of me was excited, so I wasn't expecting.

Okay, then let's do it, Prime. Go. Go. Hit me. Go, I'm here for it. Let's just not do it, I guess. Okay, fine. No, I'll go for it. Chat says Klarna now. Chat says Klarna now. Get a cheese trail going. Let it happen. I was thinking we do it. I would think we were going to talk Klarna one part this month, one part next month,

more month the month after that, and then we'd finally do it all. All right, Clarna part one. Hit us prime. Clarna part one. I think TJ just did Klarna part one. Which is this. Clorna seems, okay, so obviously Dax, we are talking with Dax earlier about. Clarna and they still make a lot of money. So I do think that they're probably not nearly in as much business problem as

being said. But for those that don't know what Clarna is, Clarna somehow has managed to become into the center spotlight multiple times, though we should probably have never even heard of Clarna. So one thing that Clarna does is that they do some sort of effectively layaway like style, except for you get the item right away. So you do easy payments of X many dollars for something that costs Y, right? So if you wanted to buy a burrito, you could literally pay with Klarna and then pay in installments for the burrito.

Like they have completely, um, they like integrated with Uber Eats. Yeah, they've monetized your entire world. A payment plan. Yeah. So you can, you can lay away anything at this point. You can, you can just pay over time for me. Is it called layaway because you just lay away the whole day like after doing that or what? That was the best I got.

It's one hour later. people are going to leave a comment that that one was good. They probably won't. But, okay, so I think there's a couple things. And then Klarna also has become really well known over the last year because they're also the same company that fired most of the people working at Klarna and replaced him with AI and said that AI is the future. Only to almost exactly one year later go, whoopsie daisies.

It turns out it didn't work. And now they're rehiring people and saying actually people are super. important to the AI process. And so Klarna has kind of like a history of acting in kind of goofy ways. But I do think that there's maybe a fundamental problem here, which is that how I typically look, especially at microloans like this, is that it is purely an advantage, or it's purely an opportunity to take advantage of people who do not have enough money. Right. Like, because you're not people who are getting a loan on a $15 meal. Like that, you're in a state that's not very good. And so they're like, this is not, this is not a recommended state of life.

Like this is actually a really kind of crappy business model. And morally, it feels like they've just been on the wrong side of every argument consistently. AI is better than people. Oh, wait, no, they're not. People should be, people, we should commoditize even people that are broke. Maybe, maybe not do that. Maybe this is not the place to be. I don't know.

I see Klarna as just like a pretty awful company. and I'm not even sure how to process that they have been doing all of this. And now it's coming out that they might be losing money. And it turns out that when you give people who cannot afford even meals alone, that they may not even be able to pay back that money as well. So it's kind of, I mean, it's kind of funny in some sense that they might be reaping what they've created yet again. So I should act as a poor substitute for Dax here and also a good substitute for my by just saying that both he and I read the actual financial statement that they put out,

like the investor relation statement. And if you look at it, the actual losses are not really, like it doesn't look like, at least from that report, that the losses are because consumers aren't paying them back. That's the only thing that I would want to put out there. And Dax had a more positive take on it than I do because I, I, I mean, I just, I'll go ahead and say, I think he's more of a, like, startup kind of, like, business growth kind of person, mindset-wise. So he's like, this is normal the way that these financials look.

And I agree with that, but I'm also like, but also a lot of these things end up failing. And or selling out to people who then have to pick up the pieces because it turns out not to work out that well or things like that, right? And so when you look at the financials here, the thing is, really what they've done is they've paid a bunch of stock compensation to people. Like, it was called stock-based payments was like the line item, right? Which it's not a publicly traded company, so it's hard to know exactly what they mean by that.

But generally speaking, it's like things where you give people stock or you do things based on stock where you're trying to pay employees out of that or you're trying to, you know, you're trying to use your stock as a way of paying for things that you need. and for lack of a better term a lot of times corporations try to claim that this is sort of like a one-off thing like it's like well we won't have to do that recurring because it's just these bonuses that people got

or these like you know things that people were getting stock options and whatever but a lot of times they tend to have to continue because in software like this is my perspective in software you constantly have all this upkeek and you constantly have this challenge of retaining talent and all these other things and so when I look at that

and I see you know that on a balance sheet I think Dax was like, oh, that could just go away. And I'm more like, I don't know that it can. It really depends on the corporation. So I think we both came down on it. It's too early to tell what's going to happen with Klarna. He was maybe more positive and I was more negative. I don't think anyone knows for sure.

And especially without a publicly traded company, they're under, like, they don't have as much disclosure requirements, I don't think. I don't know how things work in Sweden. I don't know or whatever that is. But so I would just say, I would be careful reading too much into this. I don't think we have a good picture. about whether this company will work out or not. I hate to rein on that particular parade prime

because I kind of agree with you that it sounds like a bit of a sleazy company or sleazy business model, but I don't think this necessarily shows us much about them, this particular report. That's all I could say. So I have a question as someone who didn't read the financial statement. Okay. Good, good.

Because I would literally have no idea what I'm reading. I don't have a lot of the financial strategy. Meeting expectations. Thank you. So I do agree that I think businesses like this suck. And, you know, to be honest, I personally would like to see them go under. My question is to Casey, because he probably knows this, is, so say a company like this does go under, what do they do they do they do effectively sell it to someone else? And then there's like some collector that kind of just has to like, obviously get the money somehow.

Like how does that work? You mean the consumer debt. Yeah, yeah. Yeah, usually. I mean, you know, in the U.S. anyway, we have bankruptcy court, right? and where you would go in and the bankruptcy court basically tries to apportion things out to your creditors. So all the people who you owe money to. And presumably those consumer loans have like all these consumer microloans,

they have value to the extent that the consumers are going to pay them back. And that's some percentage, right, which Klarna internally knows, like they've got an idea of that. And of course, like the 2008 mortgage crisis, you never know when that's going to go south too. Like, their models are all based on what they think is going to happen, but you never know. Like, all the consumers could start defaulting on Klarna tomorrow, right? And that would be this huge upset in their business model. So I don't know about any of that. But, yeah, what would happen is other creditors, people who that Klarna owed money to,

would get those loans apportion to them to receive the payments to the extent that they were going to come in, would be what I would assume would happen in bankruptcy court if they filed for, but reality is they'll probably get purchased if they, They would probably get purchased by someone, you know, a PayPal or a stripe who just picks them up and then like integrates that into some of their stuff and does something with it, right? That would be my guess. I doubt they would flat go out of business, but I don't know. It's not my realm that you should have Daxon. Maybe he studies this stuff closer.

I only read these sort of things to find out, like I just said, whether or not the news story is accurate with what they portray. They definitely weren't in this case, right? They kind of spun it the way they wanted to, and it doesn't necessarily represent. and the underlying financials yet. Could change in the future. The only thing I wanted to say, the reason that I brought this up, I just wanted to show a tweet of mine. Can I show it right now, Prime? Yeah, yeah. Go for it.

Okay, nice. That's not showing anything. No, it's not. That looks good. It's a really good tweet. Yeah, what is that, Teach? A hoddle. A hoddle.

Okay, nice. There we go. Cinema. What's a skill? that can be picked up in one month that will have compounding returns to the rest of your life. Usory. That's all, that's, that's, that's the reason that I put this on the, on the market. Thank you everybody.

And Prime, if you have anything you want to say now, feel free. No one's going to know that term. No one's going to know the term. Usory. Usory is debt, a debt collection, effectively. Yeah. Yeah, I think the, I think the thing that's really interesting here is that this is a news item that shouldn't be a, a news item about a company because people hate them that much. There's like a general dislike of them so much that I think that when people hear this, it's like a celebratory moment. Like, hey, I think it's bad what they're doing. Therefore, yay. Like for me, the whole user thing I think is just

obviously bad. But I truly only know them for their just dumb takes on AI. I can tell you all about those takes. And it's just shocking that this company just manages to constantly be in the spotlight for their behavior. And so it's just curious how long, like, how do you even do this as a company? How can you be so universally disliked? I just find completely fascinating. Even people with no skin in the game are like, I don't like them. I don't like them one bit. Well, the funniest part about that would be when you look at the balance sheet, it's like, okay, you know, one of the big losses was the stock-based compensation stuff or, you know, again, we don't have a breakdown of what that actually entails. But it would be hilarious if a lot of that was because they had to go back and

make good offers to all the people that they got rid of because of the AI. That would be pretty funny. It's like, oh, my gosh. But it probably isn't that. Like, I don't know, right? So, like I said, but that would be humorous, obviously, to your point in prime. Yes. I would just love it.

I don't know. I just find this more fascinating, not from a, what they're actually losing their money for. It's just that it's weird to see a company so disliked, even continue to be successful or continue to be running as a company. I would have just assumed they'd go under, but it's just assume they'd go under, but it looks like what people think on Twitter is abstract or not even true to actual reality, which is a time and time again point that's proven to me over and over again.

Yeah, well, of course, also, these financial statements are always cooked, like I said. Like so, you know, I mean, again, and not even a public company. So who knows what's going on, right? We have, we have, you can go look at it. It's a PDF. You're like, I wouldn't know what I'm looking at it.

Yes, you would. It's literally just, there's like one page, and it just says crap on the side. Like, here is a line item. and here's the number. And you don't really get much breakdown from it because, you know, I don't think they are under any... Maybe they are.

I don't know. I don't even know why they put that out, right? Maybe under Sweden law they have to, I don't know. Again, I studied this for five seconds before this podcast just to see. Because it's not, I don't care. I do not care about Klarna one bit other than the mild disdain that I would have

just from your description of their business model, right? But I'm just saying, so you can go look at that sheet, and that's as much information as I think people have. So really anyone who's creating a narrative out of it is largely just making it up unless they've got an insider giving them information because there's not much there. What we can say is it's not an overwhelmingly obvious set of numbers that would lead you to the conclusion that it's because consumers aren't paying them back. That we could say for sure. The set of numbers they published does not tell that story. What the truth is, I have no idea.

Well, I think we can end it here. The Clarnis section is actually not that interesting. I just really just like them, so I just wanted to say I disliked them. You know, as any good stand-up, I'm sorry. I'm sorry. They sound like a pharmaceutical. They sound like those like medicine commercials. It's like try Clarno.

Side effects, diarrhea. Ask your doctor if Clarner is right for you. Exactly. That's like the first thing that popped in my head when someone mentioned Clarno. Stashed. Trash does literally zero research every single website. Okay, well, I was going to say. I read the article, but my initial reaction was diarrhea.

Okay. Anyway. On that note, I was going to say, vote for T. Tige 2028. I'm going to be running on Chippole Burrito forgiveness. I'm going to be forgiving everyone's Chipotle Burrito loans.

So if you're interested in that, if that's kind of your target demographic, consider me for voting. I'd be happy to represent your interests. And I just wanted to close on that note. Thank you. Can we get some Tige 2028 in the comments? He has like no pants on, just a nice shirt.

Oh, shoot. Can you see that? No, no. I'm just getting out of her pants I'm wearing pants I was like are you There's no way you made that You were just close enough that we couldn't see it

I know it was perfect You could have got it I was like oh man I'm sorry for calling you out Wait am I the only one who's not wearing pants For this podcast Could be Casey Could be we'll never know We'll never know

Please don't stand up Casey you know I just want to let you know I really appreciate you And I feel like you don't get off topic nearly as much as everybody else. You know, it's an honorable... What? I was locked in today.

Me? Yeah, you. That's the first and last time anyone will ever say that about me. But I appreciate it. All right. Well, see you later. See you later, everybody. That's getting cut for sure, right?

We're sorry. Moot up the day. Vibeco. Errors on my screen. Terminal Coffee And here