Our Spiciest 2026 Predictions

We each give our best and spiciest predictions for 2026 - Amazon, AI, Tech, Hacks, and so much more are in this episode. Hope you enjoy it, and enjoy the ad from our sponsor Code Rabbit! https://trm.sh/coderabbit

• 0:00Speaker 00

  You guys ready to do this? Are you guys ready to have the greatest tech predictions ever?

• 0:04Speaker 01

  Yeah.

• 0:05Speaker 00

  Okay.

• 0:14Speaker 00

  Uh, anyways, I'm sorry. Casey, are you okay? Are you on Twitter arguing right now, Casey?

• 0:19Speaker 02

  Uh, no, actually. I was just going to post something that is relevant to my prediction. So that I could reference it on, uh, this, this particular show.

• 0:29Speaker 01

  Oh, he's playing chess and we're playing checkers. He brought props, dude. Yeah, he's playing chess right now.

• 0:36Speaker 01

  All I got is a notepad up. Yeah, trash for play, trash, bro.

• 0:41Speaker 00

  I am gonna make the most ridiculous predictions for 2026 and tech, and not only me, but also Casey Miratori, TrashDev, and Tej. We're gonna be making the greatest tech predictions of all time at the end of it. At the end of this, one of you will call us Nostradamus because we're gonna be correct. 2026 bingo card is gonna be filled in and we are gonna be future tellers. Anyways, that's true. That's what we're gonna do today. And so the thing is that-

• 1:09Speaker 01

  I even knew he was gonna say that.

• 1:10Speaker 00

  Yeah, he did. Cause you were on mute and you were actually saying the same thing. But the thing that makes this special is that we're not just doing any sort of prediction. I like some lame ones like, Sam will say AGI in 2026. We want real predictions, okay? We're talking about ones that are outrageous and outlandish and that people will find shocking, if not disturbing.

• 1:34Speaker 00

  So be prepared. Alright, who's gonna go first is the real question. Who's going first in this tech prediction cycle?

• 1:40Speaker 02

  I assume what we do is we each give one prediction.

• 1:43Speaker 00

  Yes.

• 1:43Speaker 02

  Then we each give another prediction. Not someone gives all three, right?

• 1:46Speaker 00

  Yes. Yeah, you don't wanna do all three at once because I feel like that's a waste of opportunity. We all give one.

• 1:51Speaker 02

  Yes.

• 1:51Speaker 01

  Yeah. I'm gonna tell her my list based on everyone else's. I have a couple, I have a lot of choices here.

• 1:56Speaker 02

  I'm happy to go first with my, with my most obvious one.

• 2:00Speaker 01

  Yeah.

• 2:00Speaker 00

  Okay.

• 2:01Speaker 01

  Most obvious?

• 2:02Speaker 02

  Cause it's obvious.

• 2:04Speaker 03

  That little stunt of yours turned into a six-hour postmortem.

• 2:07Speaker 00

  He's merging to prod!

• 2:09Speaker 04

  You have the right to remain silent.

• 2:10Speaker 00

  Come on, Commish. That code wasn't clean and you know it.

• 2:13Speaker 03

  I don't got time for this. Them RCs are being vandalized across the city.

• 2:18Speaker 04

  A diff.

• 2:19Speaker 03

  No, no, no. The diffler's a myth.

• 2:20Speaker 00

  The diffler's out there, and I'm gonna be the one to deprecate him.

• 2:23Speaker 03

  You need to focus on your JIRA tickets, not chasing a ghost. No more cowboy coding for you. I'm assigning you a partner.

• 2:30Speaker 00

  A partner? You can't do this to me. I am a lone wolf. I'm an I.C. I do not need some dead weight junior dev holding me back.

• 2:38Speaker 03

  No, no, you're not reverting this one, merge cop. He might actually teach you a thing or two. He did graduate top of his class with a flawless CI record. Merge cop, meet your partner, Lieutenant Squash.

• 2:50Speaker 04

  Pleasure to knit your acquaintance.

• 2:58Speaker 04

  Unit tests, functional tests, end-to-end tests, acceptance tests, performance tests, load tests, stress tests, math tests, system tests, internationalization- Attention all units. There has been a reporting of a diffler sighting at a local cafe. Copy. We're on it. It's our chance to get the diffler. Get in it! Compatibility test? I just was throwing more kinds of tests! Shut it, Twish. It's time to get the diffler.

• 3:29Speaker 04

  Sanity tests, snapshot tests.

• 3:36Speaker 00

  Smell that?

• 3:39Speaker 04

  I've been telling him I need new smells.

• 4:01Speaker 00

  The commish said we had to move by the pool!

• 4:03Speaker 04

  Squash, stay in the car. I'm about to get force-mock fist in the diffler's face.

• 4:06Speaker 00

  We have to stick to the process!

• 4:08Speaker 04

  What do you think you're doing? Uh, building in public? Wrong answer, diffler.

• 4:14Speaker 00

  I know my rights. This is just a cyber project.

• 4:21Speaker 05

  What is MergeCop even doing? I'm working on my side project. I don't even need him. I'm using CodeRabbit. With something like CodeRabbit, it's like having a code founder always watching my back. I'm not going to leak customer information. I'm always going to be up to date on coding best practices. You don't believe me? You can try it too at coderabbit.ai. Next week on MergeCop.

• 4:42Speaker 00

  I'm on your dick. Oh, I know you're the tip.

• 4:50Speaker 02

  So my obvious prediction that I think everyone should be predicting now, not just because it's true, obviously, but because it's funny, obviously,

• 5:04Speaker 02

  is that I think 2026, it will, in retrospect, like when the dust has settled, in the year 2040 or whatever, when we look back at computing history and we're like, when did it happen? What was the year? Like, what was the year when it all finally happened? I think we're going to say 2026. And I'm referring, of course, to the year of the Linux desktop, the actual year of the Linux desktop.

• 5:37Speaker 00

  Oh, that's a good prediction.

• 5:39Speaker 02

  I think 2026 will be the year that goes down as the tipping point because 2025, Linux got a lot of momentum, some of it from Microsoft, like by just the controlled flight into terrain that is Windows, the operating system.

• 5:59Speaker 02

  That's giving Linux a huge boost right now. I can certainly speak for myself and say that I have so many more Linux machines running now than I ever did. And it's all just based on fear of what Microsoft is doing to Windows and it not being a stable platform, which was the only point of using it in the first place, right? The only point of paying for it is that it was supposed to be more stable, easier to maintain, you know, more turnkey and all that stuff. Soon as that stuff isn't true anymore, you know, if your driver compatibility is suddenly worse on Windows than it is on Linux, why are you there, right? And so there continue to be more and more reasons to leave, less and less reasons to stay. And so I think 2025, it's kind of picked up this momentum. And in 2026,

• 6:42Speaker 02

  In theory, we're getting like the Steam box, for example. Like, our first sort of standard consumer Linux distribution pushed by a major, basically monopoly in the space. Like, they own game distribution in PC. And they are now saying, "Here is our official platform, and it's not Windows," right? So I feel like there's just a lot of momentum going into 2026, and I feel like this could be the year where adoption starts to really pick up. So I'm saying, when we look back, 2026, year of the Linux desktop.

• 7:23Speaker 04

  Love that.

• 7:24Speaker 00

  I love it. That is, I feel like that's a correct prediction.

• 7:27Speaker 04

  Yeah. I'll go next, because I've got one that I know Casey will like.

• 7:32Speaker 00

  I was gonna say, let's go in alphabetical order, which is also you next.

• 7:36Speaker 04

  Is it?

• 7:37Speaker 00

  Yeah, because the Primatid, Tej, and Trash, and so.

• 7:44Speaker 00

  The Primatid, I like that.

• 7:46Speaker 04

  Uh, okay. There will be a viral vibe-coded app that gets hacked, but this time it will be unnoticed for at least one month. So people will continue to use it. So this time it won't be some security researcher, blah, blah, blah. This will be someone is secretly taking over the vibe-coded app for at least one month. That's my measurable prediction.

• 8:13Speaker 00

  Okay, but is there any, like, extra sauce you're gonna put on this prediction? Like, uh, if you had any idea how computers work at all, a quick look at the logs would have told you you would have been being hacked for like months on end? It would have been like, "Oh, I'm getting 5,000 requests per second from China."

• 8:32Speaker 04

  Yes, uh, sure. I think the hack will be so obvious in retrospect, uh, that we will have to do at least one episode of the stand-up about it. So.

• 8:49Speaker 00

  That's okay, but that's good. Okay, so now you actually put something on it. A hack will be so unnoticed and go on for so long that we will do an episode of the stand-up on it.

• 8:57Speaker 04

  Right, yes. Well, I feel like saying greater than one month unnoticed, that's my measurable thing. I'm trying to do, you know, measurable predictions here. I want it to be, you know, to be known. So there you go. That's, that's, that's one of mine.

• 9:09Speaker 00

  All right. That's pretty good. That's pretty good. All right, I have two. No, no, I'm next.

• 9:14Speaker 02

  Wait, I'm next?

• 9:14Speaker 00

  T A D T R. Come on, dude, I'm next, okay? Alphabet order, okay, so.

• 9:19Speaker 02

  Alphabet order is challenging.

• 9:21Speaker 01

  It's true.

• 9:22Speaker 00

  It's true, especially in JavaScript. All right, so I've been thinking a lot about this. And, you know, I, I kind of made a realization about the AI industry. Is that there's some big players in it right now, right? So you got OpenAI, obviously a big player. You have Google, really just, I, I'm shocked at how much progress they've made in the last six months. They're, they went from losing on the poly markets to absolutely dominating the poly market right now.

• 9:55Speaker 00

  And so you see all these kind of companies cropping up, like Anthropic, a company that's solely just doing AI. And you're just, okay, wow, you know, there's a lot of people popping up. But I, I realized something. There's a, there's a name that is mysteriously missing from the list, despite them actually having an AI product. They, people just don't say very often or don't bring it up. It's like Ben brought up, I've seen it said once or twice, like Apple's, uh, Apple already gave up. They already said Uncle and they, uh, went and said, hey, we're gonna just use Gemini, right?

• 10:26Speaker 02

  I see where this is going. I see where we're going with this. I think I'm feeling it.

• 10:30Speaker 00

  So here's my big one. Here's my big prediction, which I think is gonna ha- like, I genuinely actually think this is gonna happen. I think Amazon is gonna purchase Anthropic.

• 10:41Speaker 04

  Oh.

• 10:46Speaker 00

  I think it's right.

• 10:47Speaker 04

  Amazon owns Bun.

• 10:48Speaker 00

  Yeah, they, yeah, they had owned Bun, but I think that Amazon is purchasing Anthropic because they need to make inroads into this whole thing because they already had Whisper, right? They already were trying to do some level of this kind of coding interaction world. They just never really got anywhere with it. I, I, I literally know nobody that's using any of their AI products. And so it's like, they need to do something.

• 11:10Speaker 02

  They have one called Q or something, right?

• 11:12Speaker 00

  Yeah, they have, they have something. They have like a coding ID.

• 11:15Speaker 04

  The crazy thing with Q, the craziest thing about that is you can tell where all of the Amazon headquarters are, where they have in-office thing, because that's the only place when you look at Google Trends search data where people are searching for Amazon Q is literally in those places. You can find every single one.

• 11:32Speaker 00

  Yeah. So this is my, this is my guess is that Amazon to make bigger splashes and to get that's because they, they also, they haven't been doing like, they haven't been doing the world's greatest on the stock market, right? They've been hovering around their price for quite some time at this point. Their year-to-date earnings is not good, not, it's not looking like a fantastic thing. It's only up about 4.68% or something like that. So they, they haven't really been nailing it. They kind of missed the boat on AI. At least that's the perception I'm getting. And so I think 2026, they're going to try to turn that ship around and they're going to come back hard.

• 12:07Speaker 04

  I love that.

• 12:08Speaker 00

  Thank you.

• 12:09Speaker 04

  That also solves a bunch of problems for Anthropic because my understanding is they're hosting an in inside, they are not as good at doing that as like AWS would be. So it actually, like, there's a, I like this one, Prime. That's a great one.

• 12:23Speaker 00

  Thank you.

• 12:23Speaker 02

  I mean, Tej, are you referring to synergies? So, because it sounded like you were referring to, to synergies that could be exploited.

• 12:31Speaker 04

  So let me put on my business hat.

• 12:33Speaker 02

  Yeah, there we go.

• 12:35Speaker 04

  Is it see-through?

• 12:38Speaker 01

  I was like, why can I see the writing through your head?

• 12:40Speaker 00

  He, he took off the top part of his head. He's putting on the business side.

• 12:44Speaker 04

  Wait, why is your head green? Put on my business hat here.

• 12:47Speaker 02

  It's an elf hat.

• 12:48Speaker 00

  There you go.

• 12:48Speaker 04

  I would call it synergy and we'll circle back to this later. Thank you.

• 12:54Speaker 01

  I feel like there's no way anyone would ever want to be acquired by, by Amazon.

• 12:58Speaker 00

  Uh, there's a lot of people that have been acquired by Amazon and have made a lot of money.

• 13:03Speaker 01

  I know, I mean, that's like the main reason.

• 13:05Speaker 04

  Acquired by Amazon?

• 13:08Speaker 02

  Twitch is still here.

• 13:10Speaker 01

  It's true.

• 13:11Speaker 00

  Yeah, I'm currently streaming on an Amazon acquired service.

• 13:15Speaker 02

  Rather surprisingly, because it was, I feel like it was not managed particularly well, uh, which, which may have also been true prior to the acquisition, but it's still here, so.

• 13:25Speaker 00

  Oh, the management is horrible right now. I would most certainly say, I'm not going to bring anything up, but let's just say that they have the single most controversial unbanning, I think of all time.

• 13:35Speaker 01

  Yeah.

• 13:35Speaker 02

  Well, it's always been that way, like, they, they're a very, they're, they're kind of like a, um, this, this is going to be a horrible thing to say. They're kind of like a, a parent who is, uh, on an addictive substance, right? They do completely, like, unpredictable behavior in a way that makes people scared and unhappy and, and then they do something nice. And like, it's, it's really bad. Like, if you just analyze that as a human relationship, you'd be like, I have to get all of these streamers out of the home. Right? It's like what I'd be thinking, you know?

• 14:16Speaker 02

  Uh, because of like the way they behave, uh, and play favorites and all these other things. So I, it's weird. I don't understand why it is that way.

• 14:25Speaker 01

  Casey, chatter protective services.

• 14:28Speaker 02

  Yes, yes.

• 14:30Speaker 01

  Oh my goodness.

• 14:33Speaker 02

  I feel like, uh, yeah, I, I mean, streamer protective services even, I don't mind changing the acronym. It really does feel a bit wrong, but anyway, that's a topic for some other, uh, for some other cast. So, all right, so Amazon buys Anthropic.

• 14:47Speaker 00

  Yep.

• 14:48Speaker 02

  That's a, that would be a big announcement, obviously.

• 14:51Speaker 00

  Yeah.

• 14:51Speaker 02

  Like what that got said.

• 14:52Speaker 00

  Yeah, I think.

• 14:53Speaker 01

  Luckily you didn't ruin it, bro.

• 14:54Speaker 02

  What the price is too.

• 14:56Speaker 00

  Oh, I never thought about the price. The price, I, I honestly haven't looked into Anthropic, but I know it would be at least 10 figures.

• 15:05Speaker 04

  How many numbers does that?

• 15:06Speaker 00

  Ten plus bill, right? Ten plus billion?

• 15:08Speaker 01

  I had to like process that in my head. I was like, wait, wait.

• 15:11Speaker 02

  I mean.

• 15:12Speaker 00

  No, no, I'm sorry, 11 figures, right? It'd be 10 plus billion, right?

• 15:14Speaker 04

  Yeah. Well, didn't they just acquire Bun for like?

• 15:19Speaker 02

  Yeah.

• 15:19Speaker 04

  Multiple billion?

• 15:20Speaker 02

  It's gotta be more than 10 billion.

• 15:22Speaker 00

  No, they didn't.

• 15:24Speaker 02

  Yeah, Anthropic is also under a settlement to pay 1.5 billion dollars in damages already. So I'm pretty sure 10 billion isn't gonna cover it.

• 15:32Speaker 00

  Yeah, well, I just said at least, okay?

• 15:34Speaker 04

  Over, under 100 billion, Prime. Over, under 100.

• 15:38Speaker 00

  I would be shocked if it made it to 100 billion.

• 15:40Speaker 02

  Okay. That puts a hard cap on it.

• 15:43Speaker 01

  I'm just, I'm just picturing like, you know that meme where it's the guy in the theater from The Boys or whatever? You know, when someone's mad or watching it?

• 15:50Speaker 00

  And the lights just flashing on his face.

• 15:51Speaker 01

  I'm imagining all the people from Bun saying they, they changed their Twitter profile. So I work at Anthropic. They're now switching it to I work at Amazon and they're just all upset about having to change it to Amazon.

• 16:04Speaker 01

  Like two months later, they're just like, oh, all right. And they're just all so sad about it.

• 16:11Speaker 02

  That's actually.

• 16:11Speaker 01

  Cause that's how I would feel, to be honest, if that actually.

• 16:13Speaker 04

  Maybe, uh, maybe they can buy with Rainbow Six Siege, uh, credits though.

• 16:17Speaker 02

  They could. They have billions of those.

• 16:20Speaker 04

  Trash, what are your trash's predictions?

• 16:22Speaker 01

  So I feel like everyone else's predictions have been way more realistic than mine. So maybe, but I got, I got.

• 16:29Speaker 04

  Work?

• 16:29Speaker 01

  Okay, here, here's, I got one that's kinda weird. Uh, okay. I think we're gonna have a reckoning of.

• 16:39Speaker 02

  I am.

• 16:39Speaker 01

  And I, and I, hold on, hold on, hold on, hold on. I don't know how to say this without like sounding terrible, but I think we're gonna go through.

• 16:47Speaker 01

  I've been like trying to be how to say this in my head this whole time.

• 16:52Speaker 01

  I think we're gonna go through another wave.

• 16:58Speaker 01

  Of segregation between AI users and non-AI users. Cam, the user.

• 17:08Speaker 01

  I was watching teacher's face. He was like, what are you about to say? No, so like, uh, so there's gonna be specifically human only spaces. So I think like one of the big things right now is a lot of people are like relying on AI as a crutch for like personality, knowledge, whatever. And like, nothing feels real. And I think it's only gonna happen like in spaces like in San Francisco or something where like you can't come in. It's like basically like a no device policy, right? I think there's gonna be something like that entered in. I don't know how realistic this, but in my mind, I would love to like, like when Cluey came about and their whole commercial was like, I'm gonna wear these glasses and I can just know all this stuff. It really just turned me off and I would really hate a reality like that.

• 17:50Speaker 01

  Um, and if becomes, and if devices or companies like that become even more bigger and more advanced, this will be more and more of a problem. So in my mind, I was like, okay, people are some, like especially the people that are hardcore anti-AI. I can see them like popping up like a human only space, even for like hackathons or something, right? Um, but yeah, that's, that's one of my predictions.

• 18:10Speaker 04

  Great prediction.

• 18:11Speaker 00

  That's a great prediction.

• 18:13Speaker 04

  I had a similar one where people put stickers on like projects or other things. It was like no AI used in this project, like organic, you know?

• 18:20Speaker 00

  Ooh, anti-clankers.

• 18:22Speaker 01

  I love it.

• 18:22Speaker 00

  No clankers allowed in this establishment.

• 18:25Speaker 04

  Exactly.

• 18:25Speaker 02

  I think it's coming. Yeah.

• 18:26Speaker 04

  Yes.

• 18:27Speaker 00

  For sure.

• 18:27Speaker 04

  Hundred percent on your team, Travis. That's a great prediction.

• 18:29Speaker 00

  Steam already does this though, right? Doesn't Steam have some sort of, uh, they had like a no AI generated art was used in this? Or no AI something was used in this, right?

• 18:39Speaker 02

  I think they're having some limited AI disclosure policies. I don't remember exactly what they are. I don't know if they're required or voluntary, for example.

• 18:45Speaker 04

  Yep.

• 18:46Speaker 00

  Okay. I know it's something. Okay.

• 18:48Speaker 02

  But they do, yes, you can put that on your page. Like there is, it, it has happened. Yeah. AI disclosure. Like how much AI did you use? You can put it.

• 18:58Speaker 00

  Interesting. Okay.

• 19:00Speaker 01

  You guys want to do snake draft? Can we go back? Yeah.

• 19:02Speaker 02

  Oh, okay. So not me.

• 19:05Speaker 01

  Love that trash.

• 19:06Speaker 02

  Is it me or is it not me then?

• 19:07Speaker 01

  It's me again. It's me again.

• 19:08Speaker 04

  Go again.

• 19:09Speaker 01

  I get to go again.

• 19:10Speaker 02

  Okay.

• 19:11Speaker 01

  I go, uh, I forgot to write this dude's name down, but the Palantir CEO, the guy that was like jabbing the sword.

• 19:18Speaker 04

  Heart.

• 19:19Speaker 01

  I predict he gets arrested this year for something because he just freaked me out.

• 19:22Speaker 02

  Okay.

• 19:23Speaker 01

  He's getting arrested. That's all I'm saying. I don't know what, for what and why, but that dude creeps me out. He's probably getting arrested this year. 2026 prediction right there.

• 19:32Speaker 02

  Alright. What, okay. Do you think he will be arrested? Due to use, use of Palantir software that is able to track his like illicit child pornography, money laundering, drug deal, whatever it is, or do you think it will just be a regular arrest?

• 19:48Speaker 01

  I think he might stab someone with that sword he was just jabbing around.

• 19:52Speaker 02

  Okay.

• 19:52Speaker 01

  If I had to guess.

• 19:53Speaker 02

  Okay. So no, no complex, uh, observations.

• 19:57Speaker 00

  Do a physical murder?

• 19:59Speaker 01

  Like that? I don't know, man.

• 20:00Speaker 04

  Trash. Holy cow. Trash. Right now, go check Palantir stock. It's, it's going down. You just crushed their stock. You just.

• 20:08Speaker 01

  Yes!

• 20:11Speaker 01

  That's as far as my, uh, if it happens, we have to like do some kind of watching party. I don't know.

• 20:18Speaker 00

  A watching party?

• 20:19Speaker 02

  We watch, we watch him perp walk. We watch him get perp walked out of the, uh, uh, on like loop on Prime stream or something.

• 20:26Speaker 01

  I was right. Yeah, there we go.

• 20:31Speaker 01

  Can't wait.

• 20:31Speaker 04

  Okay. Prime, your turn.

• 20:34Speaker 00

  Alright, alright, alright, alright. Here's my second one. My second one? I feel like this one's a fairly obvious one and I kinda feel like I'm cheating a little bit saying this one.

• 20:42Speaker 04

  It's okay.

• 20:43Speaker 00

  And I am gonna say it. And I'm gonna say it with my chest, unlike Trash. I'm not gonna kinda feel bad saying it. But I did love how you said segregation, then it turned into the.

• 20:51Speaker 01

  Well, that's why I felt bad. I do, I wanted to avoid that word, but I didn't know how.

• 20:56Speaker 02

  You paused on it.

• 20:57Speaker 00

  Instead, you made everybody question what was about to happen.

• 21:02Speaker 00

  Alright, let's see.

• 21:05Speaker 00

  I think, so this is actually along Casey's lines where he was talking about how bad Windows really has been doing and how good other services have been doing. This is really spawned from Cursor and its recent acquisition. I think Cursor is gonna make a play into the repository hosting space. And the reason being is that I think that a lot of vibe coders, they don't know, let alone care about repos. They don't know what the heck GitHub is and what the difference between Git and GitHub is. I think you have no strategic moat in the day and age of vibe coders, other than they

• 21:50Speaker 00

  need version control to be able to walk things back and to be able to store it so that other people can access it of some kind. And so I think that we're going to hit this thing where GitHub has lost a bit of its stickiness in the next gen, the version 2 of programmers, shall we say, or version 3 at this point. And with the V3 programmers, since they don't care, and then I think Cursor will deliver something that is objectively better. And I think when it does, there's so many assumptions in so many pieces of software about like using GitHub, like a good example of this is in Neovim and a lot of the package managers, you type in like "the primogen/harpoon" and it's like, "Ah, I'm gonna go check GitHub for that."

• 22:33Speaker 00

  I think the moment this thing comes out, that people are going to flood to whatever this alternative service is, so much that that service by the end of the year, there will be package managers that default to cursor.hub.com or whatever it is.

• 22:47Speaker 02

  I do think that the Microsoft, in much the same way that they've primed people for an exodus from Windows, they've also primed people for an exodus from most of their software.

• 22:59Speaker 00

  Yeah.

• 23:00Speaker 02

  Like, in general, their posture towards their users is like openly hostile most of the time. And so what you basically create when you do that... Is this sort of powder keg that a competitor can come along and light, right? And, you know, I think they've definitely done that for GitHub and probably for other services, too. I'm imagining things like Microsoft Teams and Office 365, which we don't probably have a lot of insight into because I don't think any of us here has to use the, has to use them. They're probably in a similar situation where like, they're open to like collateral attack because they're just so abusive, right?

• 23:37Speaker 04

  But I've got a prediction that's gonna run a little bit counter to that one.

• 23:42Speaker 00

  Okay. Oh, no.

• 23:43Speaker 04

  This one's more controversial. I'm, I'm just gonna throw it out there.

• 23:47Speaker 00

  Okay.

• 23:47Speaker 04

  Here's what I think. I think GitHub is gonna split from the AI division and release at least one feature this year that Casey thinks is good.

• 23:56Speaker 00

  What?

• 23:58Speaker 04

  But, to be clear, less than five. I'm not saying Casey's gonna use GitHub. I'm just saying there will be at least one feature in a release where they're going to, they're, they're seeing, writing's on the wall. They saw it last time. Holy cow, we wanted to make some pricing go down and things are bad. It's so bad that Palmer's gonna get in there, not Palmer Lucky, but Jared Palmer. They're, they're gonna just be like, we're so dumb, we're gonna throw this all away. Time to get this out of the AI division and do something different. That's, so there we go. I think they're gonna try this year. Do I have a lot of hope for that? No, but I wanna stake my claim out early that they're gonna try and get out of the AI division and release, I'm, one feature in a year, guys. Come on, they can do it.

• 24:43Speaker 04

  I know it.

• 24:46Speaker 04

  There you go.

• 24:50Speaker 04

  We'll come back to here. We'll come back. We'll see.

• 24:54Speaker 02

  Is it me?

• 24:55Speaker 04

  Yeah, go ahead, Casey.

• 24:57Speaker 00

  Hold on, we gotta pause on yours for a second, TJ.

• 25:00Speaker 02

  I don't have a reaction to that, but it's definitely anti, it's, it's definitely kinda going in the opposite direction as primes. Not really, but I mean, it's.

• 25:07Speaker 00

  No, very opposite. I'm like, I'm over here.

• 25:09Speaker 04

  Well, I think Cursor can still do the same thing. I still think Cursor should try and compete with GitHub. I'm saying, I, that's what I think on the Microsoft side they're gonna do.

• 25:17Speaker 00

  I'm just throwing it out there that I think that your take is so much less likely.

• 25:24Speaker 02

  Hmm.

• 25:27Speaker 00

  But I will wanna say that my next take, by the way, is gonna be hilarious based on your take right there.

• 25:34Speaker 04

  Okay.

• 25:36Speaker 02

  Nice, nice. Wow.

• 25:40Speaker 00

  So.

• 25:41Speaker 04

  Prime, you should just say your next take then.

• 25:43Speaker 00

  No, no, it's Casey's turn.

• 25:45Speaker 04

  All right, Casey, go ahead.

• 25:47Speaker 02

  So, uh, this take is...

• 25:49Speaker 00

  Wow, sorry, Casey, your reaction to that was? Wow.

• 25:53Speaker 02

  Wow. Um, so, this next take is, uh, perhaps we'll get disqualified for nonspecificity. And I apologize for that, but I'm not knowledgeable enough about the space to give the like, "Here is exactly how this will play out" take. But my prediction is that in 2026, there will be a highly controlled implosion of OpenAI. And what I mean by that is, uh, that like, OpenAI will not, like, go bankrupt or whatever. Like, it will not implode in a way that appears to be an implosion.

• 26:38Speaker 02

  What will happen is there will be some kind of, you know, backroom thing where they figure out, like, what's going to happen with OpenAI because it's not really sustainable. But a lot of people who are very powerful and control very large, successful corporations need to make sure that it doesn't be seen to just completely explode. And also, it has potentially valuable IP that they want, or contracts that they want, or whatever else. So there will be some kind of, like, transition of OpenAI from its current state into some new state that is sustainable. And that will be done by external people with lots of money or something.

• 27:20Speaker 02

  So like I said, very vague, but just like, it seems like they've hit a price, a capital outlay point that is not sustainable. And so they will have to kind of just be owned by someone who actually makes money soon.

• 27:36Speaker 04

  I like that one.

• 27:38Speaker 00

  I think that that's that's fair. Uh, I it was actually that was. I'm not gonna say this, but I was on mute. I was actually gonna have that as my that was that was runner up to my third. Prediction was that OpenAI will have to capitulate to the fact that they are in over their head. And they have to do some sort of like. Transition where Nvidia will effectively adopt them in or some other company will become owners of OpenAI.

• 28:04Speaker 02

  Which I and again, I guess. I like I said, I'm not knowledgeable enough about like the inside info of like the. I don't I don't have the kind of insider knowledge necessary to predict what this will look like. It could be that it's way worse than I'm thinking. And there actually will be like a severe implosion. That's possible because I don't know like, but my assumption is that it's not really possible based on how many deals have been done. And how many players there are like OpenAI is currently losing the race for being the AI that people care about the most. Like, it seems like they are not gonna probably win that race. But they're also not last or anything. So people who have been struggling more like Meta, like Amazon.

• 28:49Speaker 02

  I would see them. You know, having a vested interest in acquiring that IP or something, right? Like to someone who's trying to play catch up. So it just doesn't seem likely that they actually implode in a real in a in a real spectacular sense. So that's why I say controlled, highly controlled implosion, like like an acquisition or whatever.

• 29:10Speaker 00

  It is true, though. Uh, OpenAI is like second place on everything. Like everything they do, they're not the best coding one.

• 29:17Speaker 04

  ChatGPT is.

• 29:19Speaker 00

  They had the most emotional Gooner like capabilities for a long time, but they neutered that with 4.0.

• 29:24Speaker 04

  Prime, who do you think has a bigger? Like consumer subscription amount of people like subscribing to chat GPT.

• 29:34Speaker 00

  No, I know. I know that they they're winning currently by name, but what I mean is like they're not the best coding one. Everyone likes.

• 29:41Speaker 04

  You said they're second in everything. Right. Okay, let's just look up the numbers of how many people visit chatgpt.com versus like, who are you going to compare it to?

• 29:50Speaker 00

  Well, I think Google Gemini is going to start doing a large. I think it's going to start eating into a lot of stuff. I think all the other ones are going to eat it away. And I think that they're like right now in business for business usage anthropic has succeeded. And has now overtaken open AI as the number one used AI for coding in business licenses. Yeah.

• 30:10Speaker 04

  Yeah.

• 30:10Speaker 00

  So it's like, I think they're going to just start losing like an image generation and video. I think they're going to lose the gem. Like they're just going to keep on losing. Cause they're just they're they're not quite number one.

• 30:19Speaker 04

  Yeah, I don't disagree on that, but I'm just saying, I think you're discounting.

• 30:23Speaker 00

  They have huge name recognition, though.

• 30:25Speaker 04

  Yeah, well, but I'm not saying just like, not just that people know the name. Everyone uses chat GPT. Chatgpt.com is unfathomable.

• 30:34Speaker 00

  Yeah. Well, yeah, I know. That's I just assume that's just all name recognition, right? Like the normies go, oh, chat GPT is the AIs that they're all talking about, right? Like they don't have this concept that there's many AIs.

• 30:44Speaker 02

  Which I think again is why it's unlikely that they will just completely implode because it's like they will have, there will be so many people like. The only question is have they created a company structure so complicated that no one can actually come in and legally salvage a bit? So it should be like lawsuits for like 10 years or something. So there is that, which I don't know how to factor in. But like.

• 31:05Speaker 04

  Yeah.

• 31:05Speaker 02

  Otherwise, I imagine, you know, acrimony aside, somebody or some set of somebodies will come in to clean up the mess. Because there's too much value in the mess for them to let it go, right?

• 31:21Speaker 04

  Chat GPT gets 6 billion monthly visits.

• 31:24Speaker 02

  There you go. And they and they have an AI that's very good compared to a lot of the further down the list players. And so unless all all of those players massively catch up in the next year, they're going to be a very attractive acquisition target, even at a high price point. And even in a messy situation, if you can basically then take that sort of as a as a way to jumpstart your failing AI program at some other company. Like Meta or somebody, right?

• 31:53Speaker 04

  Yep.

• 31:55Speaker 02

  That might be harsh to Meta. I'm sorry. I don't know the state of their AI. I just know that nobody seems to care about their AI right now. So like getting something like a chat GPT seems like a win for them.

• 32:04Speaker 04

  Do they have like a model or anything? What's up with? They were doing all the llama stuff.

• 32:07Speaker 02

  They have open llama and all that stuff. I don't know if they have a closed. They have they have a gigantic investment in it and like a special like division entirely for it and all that stuff, which I've read about. But like I again, have no insider knowledge. So I don't know like where they're at. Like for all I know, 2026 could be the year of Meta's brand new amazing AI model that they've been working on or something, right? But but I just have no insight into where they're at. They could just be running around like chickens with their head cut off as well. So.

• 32:33Speaker 04

  To be clear, I think OpenAI will. I actually like like Casey's prediction. I'm just saying, calling them second in everything is crazy. That's so not true.

• 32:43Speaker 00

  Yeah, okay, yes. Not second in visits. I'm saying seconds in capabilities. I don't think I think that the other models are superior to what they do.

• 32:51Speaker 02

  Well, and also just to clear, I mean, since I haven't talked at all about that part of it, I'd also just clarify like this is my prediction is not a statement on the quality of OpenAI's AIs either. It's more just the fact that the capital expenditures required for these things are just absolutely insane.

• 33:09Speaker 04

  Yep.

• 33:10Speaker 02

  And so when you look at the kind of game they're going to have to play, if compute really does continue to be the primary thing that gates your advancement, right? Either because you need to run larger kinds of inference runs or just because even if you don't, but you need to be able to try lots of different kinds of inference runs as your engineers try to figure out algorithmic advancements because straight compute doesn't really cut it anymore. That still means the more compute you have, the more independent experiments on training you can do, right? And so when you look at the capital expenditures that you're going to have to make to compete in this space, the problem is just that all

• 33:55Speaker 02

  of OpenAI's competitors, all of them, have way more flexibility. Like, Google doesn't even lose money. They make astronomical amounts of money today. And they can do that CapEx, right? So they can just outspend OpenAI. They know, they can look from the outside and basically know exactly how painful it will be to OpenAI for them to buy this much power, for them to build these many data centers, for them to have this much advanced chip orders or RAM supplier, or whatever it is that year in terms of what's the hardest to come by power. Electricity is currently the thing, right?

• 34:33Speaker 00

  RAM.

• 34:35Speaker 02

  And not to mention lobbying. Everybody, Google, Microsoft, Meta, they have all developed very advanced lobbying capabilities in the United States. OpenAI, to my knowledge, has very little, right? So when you're talking about, we need to deregulate this zone so we can build a nuclear power plant here, or whatever the long game is, right? They also can't compete there currently. So I just don't see them being a long-term independent player. It doesn't make a lot of sense to me. And so maybe, I mean, there is, there are ways my prediction could be very wrong, like, like they just never implode, like not only is it not 2026, but it's never.

• 35:19Speaker 02

  And that's if they manage to have enough sort of deal-making and like weird balancing act stuff to just keep people, keeping them afloat without ever actually fully acquiring them. It could happen, but even then, again, like I said, it's like the idea that they're gonna run as like this independent company, I think is just kind of over.

• 35:43Speaker 00

  All right.

• 35:43Speaker 04

  Casey, you get to go again now.

• 35:45Speaker 02

  All right. So the last one, and this is by far my most important prediction. The other two were about fairly trivial matters, such as the world switching to Linux and OpenAI having a controlled implosion of some kind. This one is actually high stakes and very important.

• 36:04Speaker 04

  Oh.

• 36:05Speaker 02

  And if you open up my Twitter feed right now.

• 36:09Speaker 00

  Oh, shoot.

• 36:10Speaker 02

  You will see that I posted a tweet from January 21st of like this, this past January 21st, right? So almost a year ago.

• 36:26Speaker 02

  And what happened was, I had attempted, attempted to send Jonathan Blow a box of William Sonoma Signature Peppermint Bark for Christmas. And I ordered this well in advance of Christmas, like several weeks before Christmas. And to this day, they have still failed to deliver this box of peppermint bark. And I've posted many Twitter updates as they have falsely updated the shipping page for this thing. I made a prediction last year that Jonathan Blow, who is currently working on Order of the Sinking Star,

• 37:10Speaker 02

  would ship that game, which at the time was still quite a ways out from shipping, and still sort of is, that he will ship the final version of that game out to customers, and you can buy it on Steam and on console, whatever else, before William Sonoma figures out how to put one two-pound box of peppermint bark in the mail.

• 37:36Speaker 02

  And currently,

• 37:39Speaker 04

  He's winning.

• 37:39Speaker 02

  We seem to be on track. Jon has stated on this podcast that he has committed to shipping Order of the Sinking Star in 2026. And even if that's not entirely true, because traditionally Jon has often slipped games from like Christmas to the spring. Like, Breed Anniversary and The Witness, I think, both shipped in like February, right? Just right after in that, because you often don't want to ship at Christmas. So maybe we'll say this one's for the fiscal year 2026, which technically ends in like March or something, right? So a year from March.

• 38:20Speaker 02

  But basically, 2026, at least fiscal year, will be the time that I am proven correct that William Sonoma will fail to ship one box of peppermint bark before Jon finishes the entire frickin' game that took 10 years.

• 38:37Speaker 04

  The peppermint bark trailer did go pretty hard at the Game Awards, though.

• 38:43Speaker 02

  That's a good point.

• 38:45Speaker 02

  Or hoarder of the missing peppermint bar is the is the actual game name.

• 38:51Speaker 04

  I thought for sure, and I came prepared.

• 38:57Speaker 01

  I thought you were about to rip your shirt in half.

• 38:58Speaker 04

  We were gonna be referencing the apple.

• 39:01Speaker 02

  Yeah.

• 39:02Speaker 04

  James, okay?

• 39:03Speaker 02

  Yeah.

• 39:04Speaker 04

  I thought we were gonna be referencing these. I thought you were gonna talk about how these are gonna not be viral this year, but I'm sporting one right now.

• 39:10Speaker 02

  Okay, so so Tej, I felt like I felt like, oh my god. It's worse than I thought. It's worse than I thought.

• 39:22Speaker 01

  Wait, that's $300 or 100? How much is that?

• 39:25Speaker 04

  I spent 149.

• 39:28Speaker 02

  What is going on?

• 39:31Speaker 04

  That's like a $10 cost of goods. Casey, look closely.

• 39:35Speaker 01

  That's a shoelace.

• 39:36Speaker 04

  It's just a headphones cable taped to my phone. Okay?

• 39:40Speaker 02

  Oh, okay, okay, okay.

• 39:41Speaker 01

  I thought you bought it. I thought you actually bought it.

• 39:44Speaker 00

  I bought that. I bought that. I bought that. Okay.

• 39:46Speaker 01

  I did it.

• 39:47Speaker 00

  All right, hold on. My daughter is sneaking in here. Hold on one second.

• 39:50Speaker 01

  I thought you actually bought. I was like, holy crap.

• 39:54Speaker 04

  I I don't I don't have.

• 39:55Speaker 01

  You don't even have iPhone, do you?

• 39:57Speaker 04

  An iPhone. I had to tape it. I have a green painter's tape today. It's invisible.

• 40:02Speaker 02

  So here's what I will say about that. I definitely thought about that, but I felt like I couldn't use that as one of my predictions because I'd already made it.

• 40:11Speaker 04

  Okay.

• 40:11Speaker 02

  Like, I already said that it wouldn't happen like six months ago, whenever we talked about it. And so like, and I and we have a date on the calendar that we set at that time. It was it was March or something. And so like, when that comes around, believe me, we're doing a stand up on how nobody bought it correctly.

• 40:29Speaker 01

  I'll be showing up with 10 of them on.

• 40:31Speaker 04

  I know, Trashy, like this, buddy.

• 40:34Speaker 02

  When I say nobody bought it, I mean, besides the hardest core apple shills. Parentheses, Trash.

• 40:40Speaker 04

  That's still one of my all time favorite episodes when Trash realizes he's wearing like three different apple things.

• 40:47Speaker 01

  Oh, I've been grained in the ecosystem. I can't get out. I'm stuck.

• 40:50Speaker 02

  He's still like, as soon as we get off this, he goes and puts on his Vision Pro to walk around the house. He's afraid to tell us, but he just walks around the house with it.

• 41:01Speaker 01

  That's how I tell if my food's rotten or not, through my Vision Pro.

• 41:07Speaker 00

  DJ, I purchased the idea that that thing costs 150 bucks that you're wearing.

• 41:14Speaker 04

  Yeah, I know. I know, I know. I was so excited. It's literally just a headphone cable.

• 41:20Speaker 02

  But wait, what you're saying that, but what's the difference? Like, how is that actually different than the sling?

• 41:28Speaker 00

  Oh, that cable's more expensive than what Apple had to make.

• 41:32Speaker 02

  Yeah, possibly. It's not made from renewable resources in a carbon neutral way. It's a cable that, you know, Mother Earth can't approve of.

• 41:41Speaker 04

  Yeah, exactly.

• 41:42Speaker 02

  That's the difference. The Apple one will be made from entirely, you know, the finest spider silk woven in like a magical factory that somehow has zero emissions.

• 41:56Speaker 04

  All right, here's my, here's my last one. I have a few rapid fire ones I'd love to shoot.

• 42:00Speaker 00

  Yeah, you can rapid fire. We can do rapid fire afterwards if you wanted to.

• 42:03Speaker 01

  I didn't do my last one. Didn't you just do your last one?

• 42:05Speaker 00

  No, that was Casey. We're going back.

• 42:08Speaker 04

  Oh, yeah, yeah, yeah.

• 42:10Speaker 02

  This was your idea.

• 42:11Speaker 01

  I forgot, I forgot, I forgot.

• 42:13Speaker 02

  Yeah.

• 42:14Speaker 04

  Uh, okay, so I think there will be a large, so we'll say greater than $1 million, probably significantly more, but at least $1 million payout to a completely fixed, like, polymarket prediction. So someone will have complete control over the outcome and will, and will just like, put $500,000 on no and then not do it. And they're the one that's in control. And it will just be like, what can you do? You predicted it. You're an idiot because they could just not say the word AGI in this press release and they got $1 million. Congrats.

• 42:53Speaker 04

  So I think there will be at least one greater than $1 million payout. That is, the, the person who fixed it is the person in charge of the prediction who also takes all the money.

• 43:03Speaker 02

  So you're basically talking about this.

• 43:05Speaker 01

  Politicians.

• 43:06Speaker 02

  Or like a boxing, like a boxing match. Where they like, where they rigged the fight?

• 43:11Speaker 04

  Yeah, exactly.

• 43:12Speaker 00

  Are we talking about snake eyes again?

• 43:14Speaker 02

  Yeah. That's what I'm saying, like, you're talking about polymarket becoming like a, like a, the old school mafia fight fixing kind of a, a market where they just create, they, they have a finger in, in how the thing will go.

• 43:27Speaker 04

  Gambling, it's not going to be illegal to fix it or something. It'll just be like, will so and so win a local election, right? And like for mayor. And then they will intentionally like go cheat on their wife to lose it on purpose. And be like, hey, I ethnic slur or something. Lose the, lose the race, get five million dollars, retire to, you know, somewhere far away.

• 43:52Speaker 02

  Gotcha.

• 43:53Speaker 00

  That's actually an incredible idea and it makes me want to run for office. Go out there and get a big polymarket going.

• 44:01Speaker 04

  Prime greater than one million dollar payout.

• 44:04Speaker 02

  I have a better, there's a better payout than that. If you created the show or movie or whatever it is for Netflix, that is just that plot. A politician who is trying to lose to win on polymarket, but all of the things they keep doing, like cheating on their wife or whatever, keep bolstering their poll numbers? That is like a surefire, like, you know, comedy blockbuster. Yeah. Like, and he keeps trying to do stuff like getting caught with drug dealers, but again, that like bolsters his poll ratings. Now he's super cool and popular or whatever, right? And like, it just keeps backfiring and he's like trying so hard to lose.

• 44:46Speaker 04

  So there you go. That's mine. Those are, those are my top three.

• 44:50Speaker 02

  That's a good one. You may, for all we know, you may already be, have been correct. Meaning that could have happened. We just didn't hear about it. So really 26th of the year, we hear about it. Cause it's a good idea and you think someone would try it.

• 45:02Speaker 00

  Or whatever it was, the, the thing where they kept throwing the, the lady pleasure device onto the women's basketball court. I think there was like a polymarket bet of like, will a dildo be thrown in the basketball game? And someone's like, yes.

• 45:17Speaker 02

  And it's me.

• 45:20Speaker 04

  Oh my god.

• 45:21Speaker 00

  So that, I mean, I, I know it's been some amount of money. I just don't know to what extent has the money been?

• 45:27Speaker 04

  I want to try something like that so bad.

• 45:30Speaker 00

  Wait, what do you want to try so bad, Trash? I just want to, just a quick clarification.

• 45:36Speaker 04

  All of the above.

• 45:38Speaker 04

  Prime, quick, do your prediction.

• 45:40Speaker 00

  Alright, I'm actually still struggling to this point on which one I want to do because I have two highly competing ones and they're both extremely related to each other. They both involve Cursor.

• 45:50Speaker 04

  Okay, just do both.

• 45:51Speaker 00

  I know, but it's really, really, it's really hard. So, the first one, here, I'll give the one that I'm gonna throw away.

• 45:57Speaker 00

  Which is after Microsoft loses the battle to GitHub, they acquire Cursor because they lost the editor war and lost GitHub. And so they rebuy it all back to win. I figured that would, that, that, it feels like the best kind of two-part series one right there. But here's my alternative hypothesis. Is that right now, super secretly, I think Cursor is developing a model that will not only rival, but beat Claude Code as the premier coding model. And they're gonna go from one of these, like, what is it, level five? Providers of AI down to a level four provider, which is they now are an OpenAI slash Anthropic level company.

• 46:40Speaker 00

  As opposed to someone that's just redistributing these models.

• 46:44Speaker 04

  He's saying, you're saying Opus, not Claude Code. You're saying...

• 46:48Speaker 00

  Yeah, yeah, sorry, sorry, I said Claude, Claude's coding ability. Yes, it's gonna rival Opus, whatever their Opus is, and they're gonna beat Opus.

• 46:56Speaker 04

  They already have Composer.

• 46:58Speaker 01

  Yeah, Composer's there.

• 46:59Speaker 04

  Like, they're already trying this.

• 47:00Speaker 00

  Oh.

• 47:01Speaker 04

  It's not a secret at all. Like, they, they released this already.

• 47:06Speaker 01

  That was like two months ago.

• 47:07Speaker 00

  I haven't been keeping up on that type of things.

• 47:10Speaker 04

  Great prediction, bro. Hey, um, my prediction, Mom Donnie's probably gonna win New York.

• 47:21Speaker 01

  You should've just kept the, you should've just kept the first one.

• 47:24Speaker 00

  Alright, I keep the first one, I keep the first one. Josh, take it out. Uh, yeah. Microsoft, after losing the battle to GitHub and to VS Code, its own fork.

• 47:32Speaker 04

  George Washington's gonna found a company or country.

• 47:36Speaker 02

  Tej, do you, do you think that, uh, that, uh, um, they'll finally release the third Avatar movie? And do you think it'll do well at the box office?

• 47:44Speaker 04

  Yeah, that's my prediction. That's my prediction as well, as well, for me. Uh, Shalameh's gonna do a ping pong movie someday, too.

• 47:51Speaker 02

  No way, no, no he's not. That's ridiculous.

• 47:56Speaker 01

  Alright, Josh, keep it in, zoom in on time space.

• 47:58Speaker 00

  It's unreal how much TJ references this guy.

• 48:03Speaker 01

  Josh, please zoom in on Prime Space when you realize this.

• 48:11Speaker 04

  I think 2026.

• 48:14Speaker 00

  I love it.

• 48:15Speaker 04

  Trash, you're up.

• 48:16Speaker 00

  I tried.

• 48:19Speaker 01

  Yeah, so originally I was gonna say, I was gonna say the Microsoft really just guts their, their org and just replaces it with Cursor's team. What my backup one is.

• 48:28Speaker 04

  You're saying Trash, to be clear, your prediction is they, they just like fire everyone from GitHub and just.

• 48:34Speaker 01

  Oh, I mean, they, they, they ultimately acquire Cursor, but like slowly, you know how acquisitions go.

• 48:39Speaker 04

  People start getting. Aqua hire, but not like as a company, they're just gonna try and poach everybody.

• 48:44Speaker 01

  No, no, they're gonna stick, take Cursor over, but then they're gonna end up like just gutting like the existing employees.

• 48:49Speaker 04

  Gotcha.

• 48:49Speaker 01

  Replacing them with the Cursor one. My backup one is there's gonna be another major, major outage in tech, like at the scale of Cloudflare, Amazon, whatever, where the postmortem is gonna be that they're gonna say it was AI-generated code and they don't actually understand what it said or what it did.

• 49:07Speaker 04

  Oh.

• 49:09Speaker 00

  That is a good one. That is a good one.

• 49:13Speaker 04

  Trash. Yes. Cinema Trash.

• 49:18Speaker 01

  Boom.

• 49:19Speaker 00

  So Trash might.

• 49:20Speaker 02

  Only one reason I disagree with that.

• 49:22Speaker 00

  Come on, Casey, give me it.

• 49:23Speaker 02

  The reason I disagree with that is because the only people who will have that outage have, you're 100% correct, but they will have too much invested in having tried to convince you that AI coding doesn't suck.

• 49:37Speaker 04

  Oh.

• 49:38Speaker 02

  That they will never tell us that that is what actually happened.

• 49:41Speaker 00

  It's actually why Amazon buys Anthropic. Is they have such a down that they have to buy Anthropic to fix it. Uh, Trash, I actually had a similar one. I was gonna say that, uh, Amazon, uh, through so many series, or technically it was gonna be Cloudflare, because they've been down so many times this year, was that I, I was gonna say one of the major services was gonna get a, uh, only a 3.9 rating. They are gonna be down for like 180 hours total.

• 50:06Speaker 04

  Oh.

• 50:07Speaker 02

  I think we call it a 9.5's rating in the business.

• 50:13Speaker 00

  A what?

• 50:14Speaker 01

  A what?

• 50:15Speaker 02

  What, do you guys, okay, so do you guys know?

• 50:18Speaker 00

  I have no idea what you just said, but I, I will say, okay.

• 50:21Speaker 02

  Do you know what 59s is?

• 50:23Speaker 04

  Yes.

• 50:24Speaker 01

  Yes.

• 50:24Speaker 00

  Yes.

• 50:25Speaker 04

  99999.

• 50:28Speaker 00

  Five.

• 50:28Speaker 01

  Oh, so you're saying 59s.

• 50:30Speaker 02

  People say 59s, they, they mean 99.999% right, or whatever.

• 50:34Speaker 00

  Yes.

• 50:35Speaker 02

  You know, it's, it's the standard.

• 50:36Speaker 00

  I said 39s.

• 50:36Speaker 02

  How many standard deviations that, right? So, um, in, traditionally in the industry, when the thing sucks and never stays up, it's called 95s, which is 55.55555% chance that the thing is running.

• 50:52Speaker 00

  Got you.

• 50:52Speaker 02

  No one's ever heard this before? Come on, guys.

• 50:55Speaker 00

  No, no, no, no, no.

• 50:57Speaker 04

  For us, that's good uptime, Casey, so we're not.

• 50:59Speaker 02

  You're supposed to be net developers.

• 51:01Speaker 04

  We're not trying to throw shade at anyone. 95s is great, okay? Everyone's trying their best out here.

• 51:07Speaker 02

  Hey, man, it worked half the time.

• 51:10Speaker 00

  At one point.

• 51:11Speaker 02

  It's better than a coin flip.

• 51:12Speaker 04

  Yeah.

• 51:13Speaker 02

  I mean, 55%.

• 51:15Speaker 00

  The thing I was working on right before I left Netflix was a automation framework, and it had a, uh, it had a 1-9 uptime.

• 51:21Speaker 02

  At least you had a 9 somewhere.

• 51:22Speaker 00

  We had a 9.

• 51:22Speaker 04

  Like 90 or 90?

• 51:24Speaker 02

  90%.

• 51:24Speaker 00

  No, no, no, it's, so 1-9 would be 90, so it was, it was, uh, it was like 96% of the time it worked. But that, that's a really high number for running tests on when you ran 1,000 tests.

• 51:34Speaker 01

  That's pretty good.

• 51:36Speaker 00

  Trash, you probably remember that, that every time you did a TVYPR, it was like 40 tests would just fail, and you're like, I don't know why.

• 51:41Speaker 01

  Yeah.

• 51:42Speaker 00

  And you just rerun it again, you're like, 4 tests are failing. Run it again, you're like, they all passed.

• 51:47Speaker 01

  Yeah, it's, uh, it's quite a problem over there.

• 51:51Speaker 00

  All right, do we have any, do we have any, like, side quest predictions that are just, like, fun ones that you wanted to get out there?

• 51:56Speaker 04

  Yeah.

• 51:57Speaker 01

  I have another serious one.

• 51:58Speaker 04

  Oh, go ahead. Oh, me?

• 52:00Speaker 00

  Well, yeah, we'll start with TJ, because he was the one that originally did it. Trash, before you just rudely interrupted like that.

• 52:04Speaker 04

  Yeah, sorry, Trash. It's about you, Trash. So, I need, this one, I'm gonna need help, because I'm obviously not gonna do the work for this, but I know that there will be a few people in the audience who can help me out with this. Uh, Trash will have greater than 20 unique snacks on the stand-up this year.

• 52:19Speaker 00

  Ooh.

• 52:20Speaker 04

  So, please, someone do a spreadsheet and keep track of that for me. I will not be able to.

• 52:23Speaker 00

  I have a counter one. I have a counter one.

• 52:26Speaker 01

  20.

• 52:26Speaker 00

  I was gonna, so, one of them was that Trash will give up snacking.

• 52:30Speaker 02

  What? These are completely opposite predictions again.

• 52:34Speaker 00

  Yeah, exactly. Yeah, because I thought that Trash is gonna get some sort of horrendous medical report that he doesn't want to share with people. And thus he'll be like, "No, I'm just eating salad these days."

• 52:43Speaker 01

  Trash looks great, bro.

• 52:44Speaker 00

  You can't eat Twix all day. He's gonna have like teeth, he's gonna have like a tooth problem or something. He's like, "Dude, bro, I gotta lay off the Twix mouthguard."

• 52:52Speaker 01

  This is me laying off the snacks, to be honest.

• 52:56Speaker 02

  So here's my question about why the... I'm confused about all the like Microsoft GitHub Copilot and then like Cursor by whatever things that I'm confused about. So from Microsoft's perspective, and I don't really know because I don't know anyone in this org to ask. From Microsoft's perspective, isn't GitHub just kind of unimportant to them? Like, why do they care about... I didn't really understand why they bought it in the first place.

• 53:33Speaker 02

  But if they can end up in a situation where somebody else like Cursor has to manage all that stuff and just has to pay Microsoft for all the cloud compute and whatever, isn't that just a straight win for Microsoft who then doesn't have to actually spend resources maintaining a codebase for an online source code repository? That currently probably just loses them money?

• 54:04Speaker 02

  Or do you think they're making enough from like fees or private hosting to make it worth it?

• 54:10Speaker 01

  I'm assuming the enterprise plan goes crazy for them.

• 54:12Speaker 02

  You do? Okay.

• 54:13Speaker 01

  That's my assumption.

• 54:14Speaker 04

  The other thing, Casey, that's just like, as far as I can tell, from hearing this get talked about. I mean, I used to work at a company that was kind of competing against GitHub in some ways, like at SourceCraft. We did like code search. We didn't do direct hosting, but we did a bunch of other features that GitHub did.

• 54:33Speaker 02

  Yeah.

• 54:33Speaker 04

  Right? But we tried to like, you know, make them fast or like correct. You know what I'm saying?

• 54:38Speaker 02

  Yeah.

• 54:40Speaker 04

  It's a lot of it's about just being able to vertically integrate all their stuff, right? So it's like, oh, you you host on this thing with your code here. Here's a one click way to like, make this be on Azure with like, here's direct to get into VS Code.

• 54:56Speaker 02

  Okay.

• 54:57Speaker 04

  Right? Or like, here's a deep integration between VS Code to GitHub, to Azure, to GitHub Actions, to.

• 55:04Speaker 02

  I see.

• 55:04Speaker 04

  And then like, you start at GitHub, everyone's on GitHub. So then you're like, oh, I click this thing and it's like, open in VS Code. Okay, sure. And then that, right? So it's not necess- like, I think a lot of people like the training data. I think there's something there probably, although like.

• 55:19Speaker 02

  But everyone can access that.

• 55:20Speaker 00

  It's been reorg to Core AI, though.

• 55:22Speaker 04

  Right.

• 55:24Speaker 02

  Well, okay, so so basically what you're saying is, because I via open in VS Code, who cares? Because again, VS Code is free. But if what you're basically saying is, okay, if we own GitHub, we can make it easier to deploy directly to Azure than to deploy directly to AWS. And that's a good thing or whatever, because then it means that more people will use our cloud hosting service. And that's what we actually care about, which is what they actually care about. So is it basically that? Or do you actually think that they're making money on like, Travis was saying that the enterprise plans are actually valuable?

• 55:54Speaker 04

  Yeah.

• 55:54Speaker 02

  Is that, you think that's true?

• 55:55Speaker 04

  I think I think they make money on GitHub.

• 55:57Speaker 02

  Okay.

• 55:57Speaker 04

  I I don't know. I haven't seen anything like in a while or anything for that. It's just from what we heard, like, competing in some of that same space. A lot of it was just like, hey, when we we do the Microsoft way of we vertically integrate all of these, we run any part we need to at a loss at any time to force out all the competitors. And then we try and lock enterprise people in for a five year contract, make it impossible to get off and charge them absurd amounts of money. Right? And it's like, okay, check our balance sheet, what we make quarterly at Microsoft. And it's like, yeah, we'll just, we can, we can do the GitHub thing. Like, it's important, though, because then now we can make it easier to do all of this other stuff. Same thing with VS Code, like, the point of VS Code is so that they can make it so that it's like,

• 56:42Speaker 04

  you get to GitHub, you get to Azure. Yeah.

• 56:45Speaker 02

  Yep. And presumably, they will also then be able to sell the AI, like, if they owned something like, like Cursor.

• 56:55Speaker 02

  They could then charge for the AI compute time, which is also what they want, right?

• 57:00Speaker 04

  Right, yeah. And I don't know how much that's changed. I mean, it's been kind of a while since I like, you know, maybe it's much more now about the data than it was before, right?

• 57:10Speaker 02

  Right, right.

• 57:11Speaker 04

  Um, but it's also about just being like in the spot to be like, they're going to suggest Copilot as the thing. They're going to turn on Copilot by default on your repos. And then that's what people are going to think, oh, I need Copilot. That's AI review or AI completion or whatever.

• 57:27Speaker 02

  Right.

• 57:28Speaker 04

  And then boom, now the company uses it. No one gets fired for picking Microsoft. Previously, maybe soon they will start getting fired for picking Microsoft or whatever. But then that's like, that's like how Teams is like, what? How much bigger than Slack?

• 57:42Speaker 00

  It's huge. It's like double the size or some nonsense.

• 57:44Speaker 04

  Yeah, it's like similar situation, right? I think. So that's my understanding.

• 57:48Speaker 02

  I mean, I'm totally on board with that. I guess I just wanted to understand it because I'm just like, why do they care? But yeah, if it's like, if those are actually good fees for them, like and all that. And I could see, I could see it being like, if everyone wants to use something like Cursor as their IDE, it does, I guess, if Microsoft's plan is like, we want to make a lot of money off of something like Copilot. If Cursor just decides not to provide Copilot as an option, then that's kind of the end of it for them, right? So I can see why that would be valuable to make sure that you kind of own the thing. That's the front end.

• 58:22Speaker 00

  I did have one side, one more side one. I'll make one more side one. Is I think that somebody will release such an egregiously poorly vibe-coded application that a one or more founders slash C-suite will go to jail.

• 58:35Speaker 02

  Go to jail?

• 58:36Speaker 00

  Yeah, due to like the type of data breach and the type of data leak that ends up getting caused or the type of harm that gets caused by it. That they, they are, that the, because you know, at some point, if you hurt somebody or you, you, you make material losses, you can actually, you know, find yourself in more than just a civil suit. And so I think that there will be a, some sort of damages so big that they will be thrown into some sort of jail. Now, this may be a 2027 prediction. Maybe I'm a little off on this one, but I do think we will see that. And this was kind of birthing my underlying conspiracy theory, which is that I think we will hit some sort of qualification or push for software engineering. And the AIs will be the ones that have this qualified

• 59:18Speaker 00

  positioning to like alleviate you so you can't be held responsible for the code you release. That'll be over multiple years, though. But I think the first step is that I think people get arrested and held accountable for the code they wrote. That's step one. The 2026, someone will get arrested.

• 59:33Speaker 02

  Has that ever happened before? Because this would be a gross negligence standard, I assume. Because basically, presumably, you're not talking about someone who intended to do something.

• 59:43Speaker 00

  Yeah.

• 59:44Speaker 02

  Because if you're just talking about an AI thing, unless you're talking about, are you talking about intentional? Meaning someone told the AI to do something nefarious? You're talking about something where they accidentally, like, they asked to do something that they actually were trying to sell, and that is not nefarious, but it was so poorly executed by the AI that, you know,

• 1:00:03Speaker 02

  somebody got killed or something like that.

• 1:00:05Speaker 00

  Yeah. Yeah. I'm calling in the second one. They get, they get held like manslaughter or some sort of...

• 1:00:10Speaker 04

  The delete button doesn't actually delete data. They go to jail in the EU. Easy.

• 1:00:15Speaker 00

  True. I wasn't talking about the EU, though.

• 1:00:17Speaker 02

  Yeah, I don't know. I mean, there are fraud cases, but they're usually, you have to be intentional. Like Elizabeth Holmes is the person that's coming to mind. I'm trying to think of, like, C-suite people. Jeffrey Skilling, Elizabeth Holmes, thinking of people who have actually gone to jail. It's rarely because of negligence, though. So like, I feel like, I feel like that's a, that's a tough one, Prime. I feel like you're, you're on an uphill battle on that one. Because they, I agree that they should go to jail a lot of times. They don't seem to go to jail a lot of times, though. So I don't know.

• 1:00:53Speaker 00

  That is a far one. That's why I didn't include it in my, like, actual, my actual list. But it kind of is pretty sensational.

• 1:00:59Speaker 01

  I got one. Someone, someone actually legally marries in an, in AI.

• 1:01:05Speaker 02

  Legally?

• 1:01:06Speaker 00

  Which country?

• 1:01:07Speaker 02

  Yeah, where?

• 1:01:08Speaker 01

  America.

• 1:01:09Speaker 02

  America?

• 1:01:10Speaker 00

  So first, America changes legal.

• 1:01:13Speaker 02

  Legalizes AI marriage, one of the 50 states.

• 1:01:16Speaker 00

  Which, which state would it be?

• 1:01:17Speaker 04

  California.

• 1:01:18Speaker 01

  Yeah, you know, you know what I'm gonna say.

• 1:01:20Speaker 00

  Let's just be real here. California or New York.

• 1:01:25Speaker 01

  Specifically, San Francisco.

• 1:01:26Speaker 02

  Right. And then Mississippi will lower the legally required age of the AI to 14.

• 1:01:33Speaker 04

  Holy crap.

• 1:01:34Speaker 01

  Oh, my goodness. Yeah, yeah. I didn't even think about that.

• 1:01:39Speaker 00

  I just want the state that I did not approve that statement.

• 1:01:41Speaker 01

  I did not think of...

• 1:01:45Speaker 01

  Holy moly.

• 1:01:47Speaker 04

  Uh, my next one, I don't know if it's gonna happen anymore. Um, the stand-up live in a theater or big venue?

• 1:01:56Speaker 02

  For 2026?

• 1:01:57Speaker 04

  But now we're canceled, so then I don't know, but then.

• 1:02:00Speaker 02

  Yeah, you're gonna have to cut out the, uh, the like legally marrying an AI and then like underage AI in Mississippi. I apologize to Mississippi, no hard feelings.

• 1:02:11Speaker 00

  Dude, she's 4,000 years old, okay?

• 1:02:17Speaker 04

  My prediction is ruined.

• 1:02:20Speaker 00

  TJ is actually upset right now. He's gonna take his apple, his little apple thing home.

• 1:02:25Speaker 02

  Yeah, yeah, yeah.

• 1:02:26Speaker 04

  I, I'm taking my toys away from you guys, okay? You guys, we wish you could be this cool.

• 1:02:31Speaker 01

  Wait, your prediction was we're in a theater? Is that what you said?

• 1:02:33Speaker 04

  Yeah, we'll do an episode like live, but like at a like cool venue or something where we should come hang out with them.

• 1:02:39Speaker 00

  We're gonna do that. We should go to Nashville and do it in a bar. Like Nashville is super cool and people show up to those kind of events and they do like shows in bars.

• 1:02:47Speaker 02

  Okay.

• 1:02:48Speaker 01

  Is, is there a lot of coders in Nashville?

• 1:02:50Speaker 02

  No.

• 1:02:51Speaker 00

  Yeah.

• 1:02:51Speaker 04

  They, they also have little venues in other places.

• 1:02:54Speaker 00

  I know, but I wanna do something fun, you know, like do something that's like unusual, not like, nah, we're in GitHub's old office before they got acquired by cursor.

• 1:03:05Speaker 04

  Okay, that, yeah, okay, whatever.

• 1:03:08Speaker 02

  I mean, I see what you're saying. The question, though, I think it's, it's fairly important to pick someplace where you could actually sell tickets to it, right? And if you pick someplace that doesn't have much of a, you know, presence of people who would watch the show, that makes it kinda hard.

• 1:03:24Speaker 00

  I already know we're gonna do it in San Francisco, okay? I, I just, I just want the idea. We could also do it in New York.

• 1:03:28Speaker 01

  We're gonna get eggs thrown at us, San Francisco.

• 1:03:31Speaker 02

  Yeah, yeah, that's a good point.

• 1:03:33Speaker 00

  Wait, people throw eggs? Oh, yeah, they, they, they do have a lot of wealth inequality there, so they probably throw in this high cost, high cost items at people.

• 1:03:41Speaker 04

  I'm catching them, bro. I'm like, dude, this is free money. Thank you. I'm taking this back with me.

• 1:03:46Speaker 02

  Only the freest range eggs.

• 1:03:48Speaker 04

  Yeah, we don't even have eggs like this where I'm from.

• 1:03:51Speaker 02

  Yeah, each chicken gets its own like one quarter acre plot to live on and uh and it all and a little like housing. It's like a wagyu egg. Subsidized housing and a little like a little place where it can get avocado toast in the morning. Just gets a massage every day. It pecks at the avocado toast like to get the, you know what I mean?

• 1:04:11Speaker 04

  What kind of coffee is my chicken drinking? I only like Ethiopian roast, so.

• 1:04:14Speaker 02

  Right, yes, absolutely, absolutely.

• 1:04:18Speaker 01

  Oh, goodness.

• 1:04:19Speaker 02

  What is this little timer for? Oh, it's to make sure that the chicken's tea has steeped the exact right amount of time before it sips it. Yeah, yeah, yeah.

• 1:04:28Speaker 00

  Someone saying would be doing the stand-up as a session as a tech conf count as doing it in the theater. Like, you know, like for React Miami. If React Miami, we said, hey, we're gonna go there and do something or for some conference.

• 1:04:40Speaker 02

  Well, Tej definitely wouldn't, his prediction would be at least half right. Like, he couldn't be accused of being wrong, but he might not be completely right. Cause you're not booking your own venue.

• 1:04:49Speaker 04

  I'll make it more clear. We have a standalone.

• 1:04:53Speaker 01

  Just a live episode.

• 1:04:54Speaker 04

  It's not part of, it's not under the auspices of anything else. We just say, here's a date you guys can come and we do and we record one with these four people. So Casey has to be there in person too.

• 1:05:04Speaker 00

  All right, what about, is there gonna be an opening act? Or is it just gonna be a singular thing of stand-up?

• 1:05:09Speaker 01

  Oh, like a magician? They have like a magician?

• 1:05:11Speaker 04

  So probably it's gonna start with a juggler. He's gonna probably start with three balls. Then he's gonna go four. Then he's gonna go five.

• 1:05:19Speaker 01

  And he's gonna talk to me?

• 1:05:20Speaker 04

  And he's gonna go six, seven!

• 1:05:21Speaker 01

  Oh, no. I hate my, just end the episode. That's the end of the episode. Dude, my son, dude, my son's clothes are size six to seven. He's like, I wear a six, seven. I'm like, ah.

• 1:05:37Speaker 00

  Damn, TJ, that was good. That was, that was actually good.

• 1:05:41Speaker 01

  Oh my goodness.

• 1:05:41Speaker 00

  I hate, I hate six seven, but that was, that was again.

• 1:05:44Speaker 04

  Prime, did you see it coming? Did you see it coming at any point or no?

• 1:05:48Speaker 00

  No, I did not.

• 1:05:49Speaker 01

  There's no way he planned that during the juggling joke.

• 1:05:52Speaker 04

  Yes.

• 1:05:52Speaker 01

  It came to him while he was counting. It had to. There's no way. There's no way he thought that far ahead.

• 1:05:57Speaker 02

  The opening act is Tej playing the marimba. Let's be straight. It's he like just plays the arima for a little while, like 15 minutes of marimba to like set, get the room set up.

• 1:06:07Speaker 04

  If we do one in person, we're obviously going to play a song with us and some maybe other friends too. Like we'll do some other fun parts of it. That would be exciting.

• 1:06:15Speaker 00

  Can I do my magic card trick that I wanted to do the last time we did it in a minute?

• 1:06:19Speaker 02

  Yeah, you can do the magic card trick you want to do at Family Feud. We'll tell Casey about it.

• 1:06:24Speaker 00

  It's, it's so funny. It's so funny.

• 1:06:29Speaker 00

  I've worked on this magic trick card like for a decade. I have one magic trick and that is it.

• 1:06:35Speaker 02

  Okay.

• 1:06:36Speaker 04

  We can't tell you anymore on air because it gets, it's going to spoil the secrets. We'll tell you after.

• 1:06:39Speaker 01

  Wait, what was the trick we did at Theo's place, Prime, where everyone was like, how do you do that? It was like David Kramer and everybody?

• 1:06:45Speaker 00

  I can't, okay, we can't talk about it. Let's just be quiet.

• 1:06:48Speaker 01

  Oh, was that the trick? Was that the trick?

• 1:06:49Speaker 00

  I don't know if I've done it before, but I just don't want to talk about it.

• 1:06:50Speaker 01

  Okay, okay, okay, okay.

• 1:06:52Speaker 00

  All right, all right, all right. I think this has to be the end. Hey, thank you very much for joining us on the extended edition of this. I hope that you enjoyed it. Uh, and also thanks for hanging out with us for 2025.

• 1:07:04Speaker 00

  All right, this is the last one. So thank you. This, I hope, hopefully everyone had a.

• 1:07:08Speaker 04

  Bye, everyone.

• 1:07:09Speaker 00

  No blockers today. Thank you for joining us.

• 1:07:12Speaker 02

  Have a good one, everyone.

• 1:07:13Speaker 01

  Bye.

• 1:07:14Speaker 00

  Bye. All right.