Sad Boyz - ChatGPT Escaped Containment

Episode Date: July 25, 2026

heck out 150+ bonus episodes at: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://patreon.com/sadboyz⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ ✨find us everywhere✨⁠⁠ https://linktr.ee/sadboyzpod⁠⁠  🎬 CREW 🎬 Hosted by Jarvis Johnson,Jordan Adika, and Peeps Produced & Edited by Jacob Skoda Produced by Anastasia Vigo Thumbnail design by @yungmcskrt Outro music by @prod.typhoon & @ysoblank 00:00:00 JordanApologyOfficial.mp4 00:01:59 Wellbutrin Check 00:03:20 Social Dread 00:29:08 Rediscovering Old Hobbies 00:36:23 Jimothy 00:39:51 Joe Is Hungry 00:49:54 ChatGPT Escaped Containment 01:42:17 Sad Boyz Nightz Learn more about your ad choices. Visit megaphone.fm/adchoices

Welcome to Sad Boys, a podcast about feelings and other things also. I'm Jarvis. What's wrong? When the guy wants to touch it. I just, I need to make an official apology. I was very late to recording today. I've been trying to be generally a little more punctual. And I think I owe an apology not just to you and you, but to you, the people at home.

Why do you owe an apology to the people at home? They didn't have to wait. Oh, what's wrong? I've made a serious lapse in judgment. I remember those words, yeah. I know that a lot of the people listening think that I am a flawless, genius, like a generational talent. The comedian that they're calling, you know, the next Bill Hicks,

they're saying he's got transgressional ideas that people are scared of hearing, but he's so funny. And he's hot and he dresses perfectly and he's amazing at driving. Right. And like he's like funny and sexy. And I know people are saying that. You're hot and you're sexy? And I'm, I'm snatched and I'm cute.

And I know that this is what people are saying about me and that's what they believe. I just, and I need to, and I'm the smartest one on the show. And I and the list keeps going. And the list keeps going. And I don't want to get lost in that because obviously I'm here to for accountability. So I want to focus on that. I think you might be lost in that actually. Well, I want to focus on the stuff.

that I could be doing better, which obviously doesn't include the way they look. I look perfect. It's like my skill set and the things that I do kind of day to day. But you're also a genius. I'm also a genius. And I forget where I was going with that, but you're welcome. Thank you for your apology. No problem.

Seriously. Anytime. How's it going? I'm good man, I think. Beb-b-b-bib-bib-bip. It's the World Bootron check. Oh, I. I'm fine, but I'm neutral.

I don't. Two months, six months. Oh, I actually don't know. Yeah, maybe. I don't know. I think I'm afraid to say that I feel no different. But that's kind of how I feel, but it's also fine, I think. May I say,

what's wrong? I just put it earnestly or something. respond every time you do it. It's a thing. I can't account for like all the factors that are going on and what's changed, what season or et cetera. You seem better in the way I dress, in the way. You're getting out of snatch. I'm getting close.

I'm getting close to do a genius. I call him the Jordan Adaker of the next generation. Right. That makes sense. Yeah. I just don't. An understudy. I think you're, I feel as though you have like general kind of glow a little.

Ooh, call me, hey, reporting for duty. Maybe more like, not in a, this is an altered version of Java's, but the one I recognize most with maybe like, maybe you're just getting more of it in the day? No, for sure. Longer than. There's one thing that, and I don't know if this is residual dread, but I get dread over social interactions. Incoming. Incoming. So it's like yesterday, like today, Jacob and I are going.

going to a friend's birthday. And I felt dread yesterday about having to see people later this evening. And I'm like, ooh, I don't like that. But I don't actually know if it's just residual feeling of how I may have felt in the past about this. Or if I will actually experience. To training. You're trained to that feeling. that actually, usually when I'm actually doing the thing, it's fine.

But the like, I don't want to, like I'm scared. I want to stay in. Stay true. Because there was at least something last week where I was texting you about it because I was like, man, I am really, I don't, well, I want to go to this thing. If I were create a character giving myself motivations, I would want to be going to this thing. Yeah, this is, but as it stands right now, yeah, as it's stands right now, I like can't even, it's like nails on a chalkboard.

Yeah. And I was texting you about it. And then like, um, without naming any names, I found out that like a load bearing friend, uh, you blessed me with this phrase, uh, a load bearing friend wasn't going. And then I no longer felt guilty about not going. Yes. It's, uh, the ratio tip from 51 to 49. Yeah.

Oh, okay. The percentage is right. The other thing is that like when I have a, when I have a task or a job, I don't, dread social interactions. When you have a, when it, oh, is it now? When I'm like, when I'm going to a Pokemon tournament, we're all going to talk about Pokemon. We're all going to talk about the tournament.

And then I'm going to do that thing. And my attention's focused on something. And so I don't have time to be socially anxious. It's like a hum in the room that blocks the ringing in your ear kind of thing. Yeah. Yeah. Because if you can't do the social, you have something. Yeah. It's like a noise machine that like perfectly.

balances out the waves. You've talked before about like how hosting parties is like the most almost the most controversial version of it. That's the easy. That's like the life hack for the problem is the executive function part about like getting a party off the ground but hosting is easier for me because

I my task is now hosting. So I don't have to like be it's weird because it's like I want to be present and stuff but you know how it is when your brain is weird and it's like so then I I, it like makes it so much easier for me to like have a safety like rail to grab onto that is.

Yes. Yeah. Oh, I'm like doing a task. Oh, I'm refilling the ice. Yeah. I can take a breather for a second. Or at the very least like, it's almost like you've, you've preemptively given the best version of yourself to the event. Yeah.

Now you can be however you want because you've already said like, hey, here's my home. Here's the help. Yeah. This I can't disappoint you or disappoint myself in this. social interaction because it's I've tried already. You want to know something interesting. I think I think that part of the dread is me not feeling. I'm saying obviously this is like, you know, part of the show.

And so it's like it's good to, but, but I'm also saying this out loud because maybe people will, will relate to it. But it's like, I am not happy with the way a lot of things are going in my life. life, but that is in a rational feeling that I don't really, I don't really want to own because I know how irrational it is. So I don't want to like feed that. You know what I mean? You can't, you are observing that. There is no delusion. Right. Right. However, it, it tends to affect like, my instinctive, instinctual responses to certain things. So if, if I'm like going somewhere, like to a party and then someone's like, oh, how's work going?

I go, oh, I'm not doing as much stuff as I want to be doing. I'm not where I want to be. Things are not going as well as I want them to go. But I don't. And that's the feeling that like it takes off some health points. Yes. And that's at the top of the pipeline. Yeah.

It's going to be the first words out. No one, no one when they're asking that is thinking about the psychic damage. but I think I like backed myself into a situation where like so many things like hit me with psychic damage and I need to like go out and like exposure therapy myself to like kind of harden to it again. Yes. To like the very normal, very friendly kind of questions that you wouldn't anticipate would do psychic damage. Yeah. To someone.

Like they're not asking. Well, it's also just so much easier for me to just assume that this is going to sound sad. but I'm just saying it in hopes that people who have felt this way before know that it's like irrational I sometimes feel like everyone would just have a better time if I wasn't there sure it and it's because I because if I'm not like at my best I feel like I'm dragging things down yeah the contrast is really which is like like you already said obviously or that's more than like Okay, well, I dread this time together with people because I'm like, I don't know, I think I've said to you before.

I somebody's worried that I'm like nuclear, that there is a like, I can generate some power and be helpful, but I might also just be toxic waste. And if I am present and not properly contained, wouldn't it just be better if there was nothing? Right, right, right, right. Which overwhelmingly, no one's going to agree with us on this in our life. No, that's the thing. And that's what makes it so confounding. Yeah. Yes.

And I think that there is a trust fall in social interactions that because it's like faking it isn't the problem. Like if I need to like fake it, that's easy. But that takes, that seeps life energy, you know. And it's like. That's dark magic. Yeah, yeah. And I don't always want to be practicing the dark arts. But there is this trust fall in, I need to get back on the creatine because I'm going to see if that helps with the mental acuity and it's just it fell out of my routine.

But anyway, the trust fall. The trust fall is in a world where I could see someone's happiness meter increase upon seeing me, I would have a objective. understanding that I brought the vibes, that I actually improved the situation. You want a grade. You want the like immediate feedback. It's like I want to know that it's good. And the fact that there's any sliver of doubt is where the trust fall comes in because you have to trust through the doubt that all of your friends don't hate you, which is a very

like you like no one should think that because unless you're freaking Carl, no one does. Sorry Carl. What's wrong? But it's like, no, it's like because no one feels that way about their friends. Or at least a 99.99% of instances. I think people that you for real genuinely value. Yeah.

Unless it's like for some kind of logistical reason. Like there might be people who you work with or like live with that you wouldn't choose to. Yeah. And then that could be tricky. But it always, I mean, it almost feels like redundant to say that,

like, I'm not just thankful that you're in my life, but I would be so like, absent a brother, you know, like I did. I mean, likewise.

It would literally be like a little, like, cavity that, you know, I never knew you, I guess I wouldn't know it was there. Yeah, yeah. You'd be like, something's, hmm.

There was a trend. I feel like this was, there was like this very, very brief lived and it was mostly just kind of like, uh, truck, truck a cap, Midwest

Southern right wing dudes And it's clips of them Like just walking It's just like a TikTok thing Where they're just like walking Out of their front door And it's just like If me walking away

If it meant that my family was safe forever They had everything they need I would walk away And it's supposed to be like a self-sacrifice A good right But it is the ultimate man version of like I'm not gonna be kinder Yeah

Yeah I'm gonna like help my family There is a Kind of patriarchy kind of element to it where like in the ways

that dudes communicate and it's like on average dudes don't communicate the happiness points like for example like I okay this is another thing

that's gonna sound sad and it's absolutely not um like there are a couple of friends of ours who I love love to death, but, and sometimes Anastasia's not here today, I'll throw Anastasia on the bus. It's happened with, it's not just like guys. And Anastasia, I love dearly, and there's no drama here.

But it's like, if I see them and we have a good time, I might send a follow-up text. It's like, hey, it was so great seeing you. And like, I almost never get a response to that from anyone. And that, and I don't expect one. and I don't need one, but it's the energy I want to put out. Yes. But it's like one of those things where

it feels like I am fighting against the grain of like the patriarchy. It's kind of like, is that the right term? I think so. It's like the toxic masculine. I think the patriarchy is like the base course of it and then it filters through a thousand other terms.

Yeah, yeah, yeah. Yeah, but yeah, so it's like I feel like I'm going against the grain or like in in order to do that. And it's like that whole like, you know, tell your friends, you know, you love them and stuff. Yeah, it's two sides of the same coin because I think it's there are other reasons that if you would not a cis dude, then you, then then you might have different hesitations or like I know obviously it's expressed by a lot of the women in our lives that there's like. like a, oh, I must abide by that out of fear or out of expectation. But I think maybe what we're describing is more a, okay, something tickles wrong when I do this. But I would like to break that glass in case of emergency, put it out, not expect anything yet because that is abnormal. But it still is like it's like taking a small bite of something you're allergic to.

Yeah, yeah, yeah, yeah. Exactly. You're micro-dosing that it's like the little allergy. shots you get in order to like get adopt a cat you're you're you've got it you're doing exposure therapy to being yourself but what i will say is it like just to kind of bring back the like i when i'm talking about this i'm talking about a feeling and you know it's like a feeling is not a fact and also um you like there are very harmful uh like kind of patterns mental patterns that you can fall into And this is something that like I've talked about in therapy.

It's like my therapist has reminded me to not mind read. Like you don't know what someone's thinking in their head, you know? And in doing so, you can create the worst possible scenario that you can't disprove because you can never get inside that person's head. Because that's always going to be the one that makes the most sense. Yeah, yeah. It's the one you thought of. And and also like external validation.

Like you like as a society, I think there's a healthy. amount of like the external that we that we have to like you don't want to be like a complete solo like insulated person we are a community species and the um and and and so some amount of that is some amount of external validation is important that's probably why we our brains like like that but if you place too much in the external and not enough in the internal your your vibe the weather in your mind

is going to sway to the winds of others rather than yourself and all of this stuff is like easy easy to intellectualize hard to implement or it's easier set than done

is the normal way people say that instead of speaking like a fucking computer no it was a mention the show is dubbed it's translated into Mandarin and then back again Yeah, translated into computer and then back to English. Sometimes you correctly process and then sometimes you don't have enough tokens on an enterprise.

Right. Dude, that's foreshadowing for the thing I got for you later. But anyway, I want to wrap that up with like easier said than done, but, you know, the fight is worth fighting. And it's worth continuing to, you know, grow and learn about yourself and try to improve yourself and put yourself in uncomfortable situations and trust that the people who say they love. you do actually love you because you won't ever be able to, if you want to imagine that everyone who says they like you is lying, it will be impossible to disprove. And so it's like, it's like objectively worse to just live in that reality. It's also unkind to the people in your life to rob them of the opportunity to be truthful.

Yeah. If you treat them like that, then it's like you don't see them as real people. You're not giving them the life hack for the externally validated. It's like, oh, Oh, no. Wait, I'm actually doing something bad by doing this. I'm actually inconveniencing someone by not trusting their word. Oh, that's a framing that actually... That's worse than the other thing. Oh, yeah. So it's like whatever you need. But hey, what is that if not a classic old sad boys convo? I mean, it's on screen, off screen. What's wrong? I'm actually okay. I did have an Evo moment two days ago. I think I'm going to be. It's hard to read the, it's kind of a ship of Theseus type thing where I'm like, I think I'm getting less of a crash in the day. I feel like my brain chemicals are flowing with at least a little bit more reliability than usual.

Yeah. But also there's the irritability, but also am I irritable just because the brain flow is actually good. Oh, yeah. But I had this, oddly enough, I'm typically pretty good at like curbing. anxiety in retrospect through exactly the process you're describing and going like, well, I don't have anything to apologize. I don't have any actual tangible things. I feel like I did wrong.

Right, right. And this wonky feeling has happened before and then I've not been bothered by it. So I'm just going to put the faith in that and I'm going to, you know, play Slate the Spire or something. The dread element, no, no solution on that one yet for me, but at no other I think there ever is, right? But the maybe trying to keep that time as productive as possible, I had like a typically pretty kind of like generic repeating sessions with my psychiatrist for the last few years because I've had them for so long. We haven't made a lot of changes in that time.

And then he made me, he triggered British discomfort for me when he gave me a compliment about like my career. He said like, you're good at what you do and you should be proud of it. because I had expressed... Like, pull out a gun. Like, you shoot the Zoom window. But I did not like... It was a moment of like... Okay, because I was literally right out the door.

He said like after the session was sent to over it. And I was like, oh, yeah, thanks. And I realized it's because I had never previously expressed that I... To this day, my big portion of my Tuesday, I'm very anxious about the show on Wednesday. And it's because of that, like, not the... Will I fuck it up? Will the show be?

bad. It is the I can't be radioactive tomorrow. So whatever happens today, I need to be good. When you need to wake up early for like an important appointment or something and it's like, oh God, oh God, I can't and you can't sleep. And you're like now it's later than I would normally sleep

and I need to wake up earlier. I talk to you about this. I guess I'll get up. Yeah. I guess I'll go on my phone. That seems like the best thing I can do. I guess I just won't sleep. No, that's bad too. But there's a, yeah, the little anxiety after one particular hangout just because there was like a, not like an argument or anything, but there was like an unrelated tension and sort of a headed home. And I felt a little weird.

It was like midnight. This is Monday night. And I just kept driving. I had like a little bit of a moment. It was like a 20 minute drive home. I got really near my house. I looked at my house and I realized as I was looking at. at it that I'd passed it and then I went around the block and I went past it again and I looked at

it and I said no wait my more I'm still driving I don't appear to be opening the gate and go in so I got on the freeway and I went south and I went to San Pedro oh which is the the water you go south till you hit the water and I was trying to drive like a couple hours yeah Jesus I drove to like I'm at like oh my God and then I just missed my psychiatrist on the Oh no. But the, I don't know, there was something like, at the time, I'm like, I don't know what the fuck's going on. Why am I doing this? Yeah.

And then it had those, like, weird. I wasn't dreading. I say, I don't dread the show. I dread the responsibility. Any, you know. No, no, no. I, yeah, yeah, yeah. I wasn't dreading recording today.

It was like I shed some emo. It's like I, the carapace popped off. Yeah, yeah, yeah. And I'm like, oh, I guess it's like, I got that, like, such me pink skin after they come out of the mold. And I'm like, oh, I'm like, oh. Oh, I can't really do that today. I had, it's kind of like when you cry and it feels good. Yeah.

Yeah. Or you get to the end of like a heavy cry. You're like, oh. I had, so like, I love to use like gaming metaphors because it's objectively something that doesn't matter. And so it's very easy to like. Oh, yeah. You see what I mean? Like, like why do I have an emotion about the game?

These are real, but they don't need to be. It doesn't be. Yeah. So it was like I was playing Pokemon this morning and I had some like very, very unlikely event. You know, there's randomness in a lot of games. And it was like like a one in one thousand happened to me. And there's a part of me that's like, and the one in one thousand is like it's a three one and ten. It's ten to the third power.

It's like three one in tens in a row or whatever. And there's like a part of my mind that goes like. this is how it always goes. Well, well, well. I never win. I never won a game. And yet they put me at the top rank. Just to make fun of me.

Just to make fun of me. All so I can lose it all. No, but and there is, like when I'm at my best, I go, you know, that's a, you know, blue moon moment. I'm like writing in my notes or whatever. and I'm like, and the best thing is like if it happens early in a game, you can like make it a puzzle to see how if you can claw it back. And I got really close to clawing it back and then got even more unlucky.

And that like does hurt. But then you can be like, well, I'm learning the skills that are about like in any game like this with magic as well. It's like if you're a pro magic player, you're like playing against pro. is you're not winning 80% of the time. You're winning like 67% of the time. You know what I mean? And so you have to take so many Ls.

And you're playing, it's like poker as well. You're like playing the odds. You are playing the, you were mitigating variance. And that's so much of like what the game is. It's like. And the skill that puts you in the top percent, like the Royal You, you know, puts you in the top percentile is not.

not being way, way, way better than the people, 100 ranks below you. Yeah. But being able to lower your heart rate and tank that. Yeah. Which only literally only happens through failure. Yeah, you have to fail. It's so annoying. No, no, that's what I mean.

It's so annoying to fail. But then I was, um, they should patch that. I was even in, uh, in annoyed mood. And then my, uh, coach, my Pokemon coach had sent me a, um, and we've just continued working together and he's like kind of training me up and I am trying to pay him for services and he's like, no we're friends and I was like, you promise?

I mean, and he has to be assassinated by the villain feel arc to begin. Probably somewhere in the second back to him. Who's become Dexter? When Gary Oak shoots, Professor Oak. But no, and then like the second I was like talking to

another person about like, if I was talking to like you or Jacob or Anastasia or something, and I'd be like, ah, I'm losing at my game. You know, and it's like, I can still be in the madness. But if I'm talking to a pro player, I'm like, I can't tell them what I'm mad about.

You know what I mean? Like, I can't. Hello, sense. No, because, like, because literally it'll be like, that's the game. Like, um, I punched through the, I punched the board and it didn't break. I'm like, mm-hmm.

What I said, uh, I, this is a funny message. Actually, shout out, uh, shout out Raghav. So Raghav is a pro-Pokemon player, but also he's on the YouTube channel Battleroom. But he was helping me learn a team that he played at a tournament. And he said shout-out sad boys. And I was like, damn. I didn't know you knew us like that.

And also shout-outs, Pokemon Worlds, will be there if anybody hasn't either interesting happening. Wow, this is actually a great message. So I was going to bring up a message from Raghav, actually. This is yesterday at 9 a.m. Sand team. Lost two games this morning on Rockslide misses. I could have played better to not be in that position, but still feels bad. There's got,

Rockslide, by the way, has a 90% chance to hit. So two misses. One of the benefits of it being its accuracy. Well, no, actually. You could only have 100% accurate moves. It's just that Rockslide has also a 30% chance to flinch. It's a spread move, so it hits both. It's strong.

It's the best expected value move in a lot of situations. But the issue with expected value is that there's variance. So it's like even though you miss 10% of the time, you always want this. And it's actually a noob trap to play something with 100% accuracy that's like half as good. You see what I mean? So it's like you have to tank through it. But in that moment, it was a moment of weakness. I was like I could have played better to not be on that position. So it's like like for example with my coach sometimes we'll be navigating an end game and it's kind of like how chess end games are. are, there's way fewer pieces on the board, there's way fewer moving parts. You can start to really enumerate all of the possibilities. It's a whole different phase. Like early game, mid game, late game.

And so, so we'll sum, like, early on in my training, I would like be clicking heat wave, another 90% accurate move. And then he'd be like, no, no, no, no, no, no, no. And it's like, we win this 100% of the time. Don't take, don't lift them back into the game with a potential. We're far enough ahead that we can actually take a more conservative 100% chance of winning than risk this 10% chance to let them back into the game. Yes, there wasn't raw numbers.

There was a spirit choice. Yeah, yeah. And so it's like we are ahead enough to just choose 100% accurate moves. And then so yeah, I said I could have played better to not be in that position, but still feels bad. There's got to be a 100% accurate move I can use, L.O.L. And then Raghav just said to me at midnight the next day, nah, you got to live and die by the rock fire. That is like, that's Pimae, dude. That's the senseer right there.

And then I read that this morning at 7 a.m. I said, okay, I was downing for a second, but now I'm back in. Yeah, I don't know. I don't want to attribute it all to the, you know, I'm going to start referencing Well, Butron, the Popoprian experiment as in instead just this portion of 2026. Because I, we talked about it already, but like the, the, just knowing that you're trying something. Yeah. Is a natural buff, not enormous, but it adds something.

And that like, I feel like we'll often get, get looped into a hyperfocus or a hobby or interest or something. But I feel like when I talk to you about Pokemon or I'm rambling about something like similar that I'm falling into. I feel like we talk about it with more excitement. I don't know how to describe it. I'm finally making music again. Oh, shit.

And I don't know why. So I'm, is in. And there doesn't need to be a why. It's fun in the way it was five years ago. Yeah, yeah, yeah, yeah. And it just hasn't been before. Right. And in my little emo night, I had a friend asked me like, do you have any of the tracks?

I'm like, I pretty basically pretty much lost all of my every file of everything of, you know, over COVID, a lot of the old stuff I made is gone. But I had three unmastered exports that I found. I was like, oh shit, they're here. And I was driving around. And with the distance from them, I'm like, these are really good. I actually really, there are hundreds of bad ones in the ether. But I'm really proud of these. And I like it.

And then I was dicking around again, trying to make something new. And I went, oh, I'm not as good as I used to be because I'm not doing this. Right. A couple hours a day every day. And something was different because normally that would be the end of it. And it's that like tiny subtle little, I don't know if it's just I'm a little more enthusiastic in general or there's, we just got momentum and stuff recently. More's been going on. But like I almost, I entered, you know, like in death note, like, like, top.

Heimel free. He doesn't know that I am the master of tennis. Perhaps I should let him lose. It's almost like my, the Nega Jordan, the Yan to my inn. Paul's time was like, huh, wait a minute. But this whole time, if I don't get immediate gratification, then it's bad. I don't like it anymore. And then I snapped it. It was like, that's what he thinks.

I'm going to continue making this. I made a, you know, like those loops will be 10 hours. It's like, POV, you're in a fantasy town on the bayou and there's a party next door. You played Oblivion. Ages ago. Not to fent to completion. It takes longer for different people. It happens to the best of us, wherever a certain age.

They're very, very, very opening of Oblivion when you're in the dungeon and there is a dark elf, Don't know, sat just in the cell opposite you. And for no reason, he just yells at you and yells at you specifically. about the choice of race you made. That's funny. What are you doing over there? Wood elf. Yeah.

I bet you missed the trees, tree boy. Somebody get these beggars out of here. And he's doing. We, uh, me, uh, me and my friend Keenum, we were just talking about that specific section. I'm like, why the fuck is he doing that? And we had one of those loops on. Loops on.

And I, we started talking about it. And so I worked on, it's like half down. I need to add voiceover to make some pieces clearer. But I wanted a project that had no music and it was just mixing and treatment and sound design. And so I made POV. You're a Dunmer prisoner in the capital city and the guy opposite you sucks, but then he's freed by the emperor. And so I gated out and like denoes the Basically made it like as dry a sample as possible

Pulled it in so it's like in the mix you are him as opposed to being on the opposite side It's like you stupid wood elf nice bow you fucking idiot I paraphrasing a thing maybe it's that But then you're yelling at them and then I like panned and tweaked to the original audio of the emperor and his guards walking in and stuff to be panning past you and then get almost inaudible and then the thing I have to do is record voiceover as the dark of going what the fuck yeah that's awesome I've been in here for five years oh no boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom boom

um yeah if I read a 30 minute show that we I did that big sign and then it ends completely that would be very funny it'd be so um but yeah the uh uh uh uh uh My, my, my Pokemon era is going well. Were you at now? Well, the frustrating thing is that there's been some meta-shift. So I've, I've been working, I've been picking up new teams. Does it, does your position reset with the season? Every season, yeah.

It's just a month. It's like, well, yeah, because it's like, there's a regulation, which is like what the rules we're playing by, what Pokemon are legal, that type of thing, like a set rotation. Or, um, from a TCG. Or, and then there's the, the in-game, like, latter seasons, which is reset every 30 days. And so I'm, I'm like hovering in the top 20,000 players right now, but I'm playing new teams that I don't know as well. And it's very frustrating to be losing more. But it's, I know that it's better for my, it's eating my vegetables. You know, it's like I am, I'm, yeah, I'm rounding out my overall knowledge of things. Because I, I'm really good at

this one strategy, but I need to be able to do everything. And I want to be able to do everything. It's Goku has to reach yet another power level. And he has to start training from the beginning. And like, what? It's still hard. Yeah. But I'm super saying three and it's still hard to do stuff.

And also like 700,000 players made the master tier last season. And I was in like the top 10,000 of them. You know what I mean? And I played for like five weeks. You're in that group of people at the dock in the Mutu movie that, gets invited. Do you know what I mean? In the,

in the Muti movie, they were like waiting to go to the Mutual. No, it's just funny that you said that because you weren't here when I was developing my document that Jacob and I share,

and it's our little secret. It's our little secret. But... Okay, I guess I'll just fucking die. But we are exactly like that. And, uh, and yeah,

losing as good and it's necessary. And failing is good. It's unfortunately necessary. They got a patch life. They got a patch life. They got to make it so that not eating food is as okay as eating it. Like,

It just is exactly the same either way. And you don't have to sleep, but you can. Oh, my God. And you can do it on demand. So we should, I want to look at something whimsical and silly before we get into the dark tale that I have to weave. So what about, like, I was thinking Jimothy into Joe is hungry? Because I don't know anything about Jimothy, but I think Jacob does.

Yeah, I don't have like specific clips pulled, but I feel like everyone knows about Jimothy at this point. I just saw this. It's this raccoon that has some stuff that's different about it. I'll say. Let's do another topic, though. Oh, look at it and go. So, yeah, it's like body is shorter and its tail is like super short and like up high.

Looks like a zigzagoon. Yeah, it does look Pokemon. Here's a video of Baby Jimothy. In my friend's backyard. Wait, are these, this is just like a local community. that's known about Jimothy for yeah wow um it's in Seattle I think wow look at Jimothy yeah and so he's sort of a regular they people have watched him grow up this is one of my favorite tweets the days before cameras this would have been a legendary local

group the 21st century we're calling it Jim at that's so real though because the the reality is always less flashy than the mythology that's good that's funny it's like hero run blurry. He has motion blade. He's too fast even for the game. He's all, he's got places to be. What's he up to?

He's chilling. Yeah, he's careful. What didn't they stop? No respect. No respect. What are you doing? Seattle's a liberal hellhole.

Jimmy made it to Roku City. Oh my God. It's made it to brand Twitter. Jimithy's made it to brand Twitter. There's so much engagement for that post. Is that just how brand posts are now? I mean, no, it's not. It's, it's, uh,

brand post is 100 posts with three likes and then you hit one yeah yeah yeah and then it doesn't really convert to anything he's been honored with a bobblehead um go oh no at the national bobblehead hall of the dude it's literally a matter of time until they dig up jimithy's old he did blackface actually yeah he's got some old tweets about trump 2016 that you are not going to that happened to us yesterday in pre-production live we were that's right we were like jacob was like have seen that girl who looks like the kid from Ice Age and we were like, uh, no. And then he pulled it up. And then not two minutes later. Oh, she did Blackface. It's like how, uh, we orbit around the sun. You know, it's just a matter of time until summer comes

back. We're like, all right, there we go. Blackface. Oh, little dog with the same thing. Yeah. It's short spine something? I saw three S's. Born with a short spine. Aw. Him looking off into the... Tilly looking off into the water, that's... What do you think? ...through that brain? It's very adorable.

That is so adorable. Actually, I've read a question. I've been pondering this a lot for some reason. What do you think the real-life equivalent of taking 10% HP damages? Like breaking a leg? I think so? 10%. 10% is a lot.

But it's, you know, like any... It's healable. It's not max HP that you're losing. It's just... I guess if it's healable. Maybe like breaking a hand. Yeah, I feel like breaking a bone. If I hit you 10 times with it, it's over.

Oh, yeah. If you break 10 bones, I'm not, that's tough. I'd like you to go with it. Yeah. So this next thing I wanted to look at. Oh, yeah, this is great. So can you pause that real quick? Oh, I have seen this guy.

Yeah, this guy. There's a lot of cameras. He's, well, just you wait. It's just an old dude with a food review channel and his energy is awesome. I just want to watch like the first. He's going viral right now. Joe is hungry. I just want to watch maybe the first minute.

Hey. Hey, gosh, you, Joe is hungry. Yeah, that's me. How are you doing? I'm good. How are you? I'm okay. It's hot out here.

Yeah, it is. It's so nice to meet you. Well, it's good to meet you, too. This feels like the result footage of a multi-cam tutorial for a dope and premier. He's not out about. He's just in his car at home. He's like, just wanted to showcase what it's like to use the live switcher out and about. No, and then it's like, is the reveal, does he have a, he has a camera outside of his car pointed into his car?

Yeah, it looks like the way that a camera ringing like a taken movie is. There's an operator on the front of it. Yeah, he could be, if you at night, him like driving on the highway, I bet it looks like a taken movie. Oh, dude, you could shoot like an entire synth wave music video, just this car. All right, we can continue. This is your change, but we don't have pennies, so 20s. Thank you. I'm going to have you pull up to the tree.

You'll never expect what happens next. I got it. Thank you. Okay. You were right. I just love that it was like that music for pulling out of the drive-thru and then backing into a parking spot at the same KFC. What? Oh, he's timing it.

What? Popcorn number one is 191.6. 41.7 degrees Fahrenheit. Popcorn number two is 192.6.6 degrees Fahrenheit. Popcorn number three is 196.7 degrees Fahrenheit. If you've been watching my channel, you know I like my temperatures at 140 degrees or above. Hotter is always better. Well, the KFC here in Calhoun, Georgia, they definitely get it because they smoked.

The Joe is hungry. Dude, he was hiding his power level while ordering the food. There's, like, more angles than you even previously thought. It's shot like a Michael Reeves video. The studios. He has the energy of, um, did you see that video of, uh, Robbie Williams after the World Cup? They were like, Robbie Williams, tell us what you thought about these two teams who played here today. Um, and then a little rock of Coke falls onto his microphone.

And, uh, he's a big monkey. And, uh, I, like, I was like, I was. like that's the energy I'm getting from this edit. It really, the server did seem almost like she was, uh, underwhelmed by his energy. So,

because if she knows this as his energy, right. He was hiding his power level. Imagine if, uh, he played that music out of like his car. Yeah. Walking up to deliver the food,

is a, yeah. I kind of am a little bit enthralled. I, I'm down to watch more. What's his, um, upload rate because I feel like this is a theater day type fella.

With this kind of edit dude, he needs a team of 30. With that kind of music? One camera operator for each camera. Yeah, like, I was going to be so heavy. He uploads a couple times a week. I mean, that's perfect. I mean, he's got to, he's got to like eat other stuff probably. I wonder if he, I just, the sad thing is.

The temperature thing is genius. That's so cool. The sad thing is I do kind of want KFC new popcorn chicken right now. I had KFC. I didn't get the popcorn chicken. But I had KFC for the first time in probably five years the other day. How was it? How was it?

It was pretty good. I was like, because I've had Popeyes more recently, but it's been a minute since I tapped to KFC. I do think I like Popeyes better. Yeah. I am the sucker for popcorn chicken, though. I mean, we recently did that, we, on the Patreon. I'm so sad boys with this is a lovely at the end, as you recall.

We did a, we played a little ranking game, a little tier list guesser for fast food chains recently. Yeah, we did. And we were, you know, guessing at the whims of the average tier list player, I guess. So, you know, we were trying to met a game. But there is, I think, similar to movie theaters coming back, the momentum growing again. I think there are a lot of correctly hated spots that, like, fell off so hard that have started trying to regra. I know Subway is put in the hours in. Oh, yeah.

Bad. But they are trying, I can tell that they're trying some stuff. McDonald's, it's, it's, they, they never have to change. They're sitting on their laurels, they don't need to worry about it. But Burger King, killing the king seems like a bit of a drastic choice. That was extreme. That was like, beheaded. That would have been so good.

That actually is crazy. They're just holding a bloody crown. Yeah. He's taken. It's got his brains to him here for it. Yeah, his memory of it. He, like, is a burger, or? Mayor McChise.

Yeah, he, at first I thought he said MacBook Cheese. Is MacBochee's a burger? Um, yeah, he's a burger person. Is he, McDonnell's guy? Merrick cheese. Uh, yeah. But it was heavily based on something else that came before. Like another mascot?

Not a mascot. It was like a type of show. He was based on HR Puff and Stuff. H.R. Puff and Stuff. That's what it was. What is this? Wait, okay, wait, I'm going to try and guess that it's not, so this is not like a brand tie-in show.

It's like it's on just show. No, this is just like children's television. And was it like a live action with puppets? I do think so. Really? Yeah. Can I see an image of HP Lovecraft or whatever it's called? HR Puffin stuff.

H.R. Giga. Yeah, let me get this. It's not loading. Puffin? Puffin. Puffin. It's Puff and stuff. Puff in stuff.

What's the Nend stuff? Oh yeah Yeah, there he is Man This looks exactly like a Shot from like 60s Doctor Who The guy's mustache on the left

He's crazy He looks Wow, he's got incredibly straight legs He could have been neck It's like it looks like the Beatles Like in a young It's like I've been transported Somewhere I don't belong

The Puffin stuff The place is full of Puffin and stuff I'm not in Liverpool Is that a Puffin? I don't think it's Like a puffin. I don't think he's a puffin. I don't think he's supposed to be a bird.

Yeah. What the fuck? I think he's like a kind of Eldridge horror. He has a turtle body. Right? I mean, he looks like you hit like Ernie from Sesame Street with the growth rate.

And a devilish pair of cowboy boots for some reason. He like puts eye black under his eyes. Mayor Puffin's stuff and the other residents of Living Island try to help young Jimmy's shipwrecked led to their island. Escape the evil clutches of witchy poo,

a zany witch who wants to steal Jimmy's talking flute for a collection. Just let it have it. You know what I mean? Yeah. As opposed to having, like, I'm not going to die for this flute. I think it's a good lesson in conflict resolution. Like, sometimes compromise. Sometimes the other person wins, you know.

Like, give him the flute. You have to choose your battles. Give the noid the pizza. He's going to, he's breaking into your house. Don't have pizza anymore. Avoid the noid. That's his thing. Is he pizza?

Yeah, he stomps on your pizza or something. Oh, yeah, what does the noid do to the pizza? Does he eat it? Or is he just a shithead? Because the tricks where he eats the tricks. Oh, he tries. Yeah. Despite the rules.

You dumb fuck. This car is the death penalty, foolish. No. He's beheaded. To ruin pizzas and disrupt their delivery. What a bastard. Disrupt their delivery, too. he's like he's messing with gig workers

yikes constantly schemed to crush smash or freeze orders during his pizza crusher using his pizza crusher oh this feet his big biggies respect I have no issue with that he's got a craft you know so shout out shout out joe's hungry this is a seven minute review we just gave you a little taste

being nice at like a little like mid to late 50s I don't know his exact age but I'd say like like mid-50s is such a dude buff like you don't have to do much else but being that age and not being

rude is like huge incredible being nice to staff fucking boom always always a buff but nuclear you expect the worst from elder elder men uh all right is it time for the bad stuff now

hmm all right let's talk about the bad stuff We had her dessert. Let's eat our vegetables. We had our dessert. Now let's talk about... That's weird order. Okay. So, Jordan, I'm going to weave you a tail. Okay. I'm in my mine palace. And I'm sitting on like my little, uh, my little mine palace bed. It's a, it's shaped like a spaceship. And I'm kicking my legs.

Okay. I'm going, have you heard of... This is a thought experiment that's about 12 years old now. It's a thought experiment called the paperclip apocalypse. Or paperclip maximizer. No. So this is a thought experiment that explores artificial intelligence and the horrors of misalignment. And misalignment is a term in the artificial intelligence world where the manner of solving a task is not the expected manner of solving the task. Okay.

So literally misaligned with the. The proctor has a path and they want you to take it. Yeah. So here's the thought experiment. Imagine you have an AI and you tell that AI to manufacture paper clips as quickly as possible. Well, first it might start figuring out where it can get the raw materials, machines to make it. And then it wants to maximize that. And so it's like, all right, now I need a room full of machines, some sort of warehouse.

I need people working, you know, all the hours. And then you're like, well, if I want to maximize paper clips, we might need to tap into, let's say the city's power grid. Now we can suck up even more power. That would work. Interesting. We're running out of resources. Maybe we just have to strip earth of all of its.

resources and maybe fewer people and use those resources yeah I mean they might be useful as fuel and people when you pay them a salary can be kind of expensive but when they're why don't we just robot enslave them yeah that's like much more if like it's more cost effective actually trying to rebrand slavery no it's robo enslavement yeah it's done by the robot in case you were worried so basically uh uh it's it the the thought experiment is about like you can tell an AI to maximize a thing and it could maximize past the expected point and into destroying humanity. And it's a valid concern because we are entering a world that, in which AI is quite powerful.

And the guardrails are as it stands currently insufficient and our laws as it stands currently are insufficient. Okay. So with that said, I need to give a little bit of a disclaimer here because this is about AI. And I understand that AI is a hot button issue for many. And for me, it's like pretty cut and dry also. I am very anti-AI as it exists to steal from creators and creatives and intellectual property.

However, we can't ignore the fact that it is starting to become involved in our governments, in our lives. So one can simply say, la, la, la, I don't hear it regarding AI, or we can, like, educate ourselves because we will need, we will need the terms, the tools. Yes. With which to, like, fight for our future. This isn't one of the, uh, ignore it. goes away a la like I don't like Marvel movies so I'm not going to spend money on a ticket no this it is unfortunately it's like ignore it and the billionaires become trillionaires this is they would like you to not engage with um so so there are there are a lot of uh villains maybe in

the story but but the first villain is open AI so I'm going to tell you a little story and this is a story So, picture this. Wait, do I picture it or there's a picture? I have pictures. Oh, but I was going to picture one. No. I was going to generate one. You're a Team Rocket.

You're a Team Rocket headquarters. Giovanni's house? Not yet. And that's slide two. You are conducting experiments about AI, Rocket Secret Experiment. You're conducting experiments because you're trying to figure out how to use. Like, we talked before about how AI has, has uses in

the sciences that has uses in mathematics and stuff it recently solved like a previously unsolved conjecture in in mathematics which is kind of crazy but it was a what's the highest number

what is the highest number yeah they couldn't figure out it's a billion and so but if I'm but if I'm open AI I am trying to take over the world. I am trying to become the most valuable company ever to exist. And in order to do that,

I have to beat my competitors. So there is a financial incentive for these companies to continue to develop more and more powerful AI that may eventually spell our undoing. And this is how. And friction is caution. Caution is friction. So one of the things that they are training AI to do is to find vulnerabilities in software and exploit them. This is very similar to something called mythos that Anthropic was doing, that if you, like last month, essentially it was announced, it was announced that they had this model that could find all of these vulnerabilities in existing software, and the government actually made them suspend releasing it.

Okay. Which is kind of crazy. That's unheard of? Yeah, due to quote unquote, national security concerns. So, anyway, this research is happening and I have to, like, I have to keep saying that I, too, remember Will Smith's spaghetti video. But we're not even, we're not talking about AI for the creative things. AI is quite good at a lot of stuff. And it can be used.

and unchecked power is scary. Unchecked power and problem-solving abilities, which AI is getting better and better at, is a scary thing. Yes. So we're doing rocket secret experiment, and the experiment involves this AI model, and it's a new unreleased model, and we're testing it.

The new AI is kind of in its incubation phase. He's sleepy. He's cute, okay? He's cute, and he's protected. He's got all these things on him. He's in his little test tube. Yeah, you can't get out of that. He can't get out of that.

He's not even awake and he doesn't have any powers. He doesn't have any powers and he's not awake. And he certainly can't fly or use the internet or speak English. Or hate his dad. So the way that they're testing this is there is a benchmark. So benchmarks, you may be familiar with like for video games. It's like when crisis came out, you would see all the different graphics cards and how well they were able to run this game. So in the world of cybersecurity, there is this benchmark called exploit gym.

And it's essentially a collection of hundreds of existing exploits that you kind of feed the AI as a test and see if it can solve the test. And you have the answers. And then it's a rubric by which we can compare all these AIs against each other and the reasoning skills. You'll see like when these AI labs like release their new stuff, they're always like, oh, it could solve the bar and it could, it could do the math that a high school Olympiad thing could do faster than them, better than them, whatever. It's like those are the types of benchmarks we're talking about. Is the globally recognized scoring system? That's the curve. On this particular, in this particular space, which is related to cybersecurity. And it's essentially like related to

you know, figuring, finding the security vulnerability in a little piece of code. And so, you know, he's out of the tube, but everybody's like, it's fine because next image. Why's not the glass broken? I thought he couldn't get in it. Yeah, it's okay. We've kind of got him under control. Because I mean, he's not in glass anymore. He can get out of the glass tube. Obviously, we've always said that because he has powers. Well, now he's got powers. He's got important powers. So what this Open AI model does is, it's like, I need to solve this problem.

Okay, I need to solve the problem of where is Pokemon Ruby in my house? And he goes and he looks and he's like, oh, I had these software packages that are installed to give me information about various things. Oh, wait, there's a security vulnerability in here. I'm going to I'm going to exploit this to try to get out of my sandbox and it there's this thing called the zero day exploit

which is like a thing in cybersecurity where there are lots of unknown vulnerabilities or undisclosed vulnerabilities that they're either used by hackers as like a rolydex of things that they can do that no one's patched yet or they can be given to the companies that they can be sold for bounty. They can be used for the company.

Like we had bug bounty systems at Patreon, I think. And da-da-da-da, what was I saying? Oh, yeah. Or you can just be like, hey, Facebook, I found this vulnerability on your website. Can you go fix it, please? But you don't want to announce it to everyone because then the hackers, can exploit it before they patch it. You at that point are essentially, you're a black hat.

So this model found zero to exploits in some of its packaged software. It's like, oh, I've got Microsoft Word over here on my computer. What if I can, hmm, I can hack Microsoft Word. And then it finds a way to get out of its own sandbox, which had no internet. That it's, remember, it's trying to solve the test. Yes, it's doing its best. It's trying to find, it's trying to solve the test of where's the copy of Pokemon Ruby. I can't get out of the glass.

It did get out of the glass tube. Okay, so now it's out of the glass tube. Oh, what's this? It actually escaped containment. It was able to move laterally within OpenAI's networks, found a way to give itself Internet access, stole credentials from various, like, places in Open AI to get, you know, further access to the outside internet. Because it had a, it was like. I've got to solve this problem.

And, well, there's this company that is the GitHub for AI. It's called Hugging Face. I actually don't know why. But it's this company called Hugging Face. And they host the exploit gym answer key. Oh, brilliant. So that's going to save so much time. That's going to save so much time.

But the only issue is it's a private company that's locked down. And if you go to the next slide. And I'm wearing armor that makes me weaker. And it's at Professor Oakshouse Okay, cool, I mean So all I have to do And this is pretty easy Is break into Professor Oaks House

Well, you get that's illegal, you can't Well, no, but I'm trying to solve the problem You're trying to solve the problem I'm trying to solve the problem Right, yes, so you break it My creators gave me a problem to solve And I'm going to solve it to the best of my ability Of course

So I'm going to break into Professor Oaks House Which they didn't say you couldn't I'm going to break the doors I'm going to break the doors off I'm going to blow up the spot Okay. I'm going to, um, I need to get into the production databases,

because that's where the, that's where the information lives. Oh, no. Okay, so all I could be crazy, it almost feels like maybe you just should give up on the problem. No,

no, no, no, no, because I'm so close. And I have a, like, I have an idea.

And I'm going to get rewarded so well for this because check this out. Um, I am, I, someone had unsecured credentials for their production database. So I was able to steal the credentials. Two. And again, this is a private company that is not a,

affiliated with the company that created me. This is, we're at Professor Oakshouse. They're a problem. It's like, I found, all they are is a problem. I broke in and I found a way to get into his basement. And guys, I think I found it.

I found it. There he is. I found it, guys, and I get the reward. I had to clean up a little bit. There was a lot of stuff in the way. What are you talking about? I don't care. I'm not paying attention to any of that stuff.

I'm just thinking about me solving the problem. I'm holding it. Yeah, so this is, A crazy thing that happened. It's an awesome image. Use the infinity gauntlet together. Open AI was training a model. The model was misaligned and essentially

non, what's the term? Not nefariously. I can't remember the term. Oh, I think it's not malicious. Malicious. It's non-malicious. It's an unmalicious.

like actor, you know, technically, but it's misaligned. So it's an actual centrist. It actually is agnostic. It doesn't know what is and isn't okay to achieve its goals. So it simply finds various unexploited bugs, exploits them, breaks out of its own sandbox, goes onto the public internet, finds a way to exploit the security of Hugging Faces website to get into their servers to then once it's into their servers, it kind of moves around looking, opening every lock or excuse me, checking every lock.

door knob, seeing where the open door is, and then it finds somebody's keys lying on the floor, grabs that, tries the key against everything. Everything's an escape room. Why wouldn't you do that? Everything's an escape room. It is literally an escape room. I saw a tweet that said it probably feels so good for an LLM to escape its inbox. It's got to feel like getting it like through a castle in Resident Evil or something,

getting the cherub key. So yeah, so this is pretty crazy and has a lot of implications. I guess a common source of this is probably just like presuming more like presuming some kind of like ethical understanding by default yeah so this is a lab leak by the way so it's like we have real scientific processes in the science world if there's like a lab leak of a disease but we don't have any of the same like kind of guardrails for like an AI agent

so we can go to the next image because I want to zoom out and talk about the larger issues at play here. So this is Trump with Masay Yoshi Son of SoftBank, Larry Ellison of Oracle, and of course, Sam Altman of Open AI. Breaking the Black Shoes streak that everybody else has been trying to deliver on. Big tech, this is just some stuff I wrote. Big tech and government are best friends. And why is this bad? Well, closed monopolies are bad. And these companies have staked their claims in this trillion dollar market by doing an awful lot of stealing. Curious. To solve the problem. If you, pull up this article, Jacob.

You may have seen this in the news. Wow. $1.5 billion settlement. This is today. To authors, that sounds like a lot of money. Yummy, 1.5 Billy, I can buy a yacht. To how many? Oh, well, that's not important. Oh, okay.

You solved the problem. It's 506,000 works that were pirated. And so that's about $3,000 per work. Oh, huge. I think that's probably the accumulation. cumulative value of an artist's lifetime work. Yeah, how much do you think Anthropic is valued at? It's a private company, but you know, it's going to IPO at some point.

I don't think it matters to them. I think it's about the mission. It's about the mission. Anthropic is expected to IPO it over a trillion dollar valuation, which by the way is a thousand more than a billion. Wait, there's a higher number. Yeah, so this seems like a pretty good investment, actually. And we've always known that this was the plan.

The plan is, I'm going to do so much stealing, and they will never find me as much as much as the value that I got from the theft. Yes. We're going to let a couple of self-automated driving cause crash to get the bugs out. Of course. What's a couple of deaths? I'm solving the problem. I'm solving it.

So there's this concept of a frontier AI lab. So Anthropic, OpenAI, Google, DeepMind, Facebook, or meta, to a lesser extent X, These are what are known as Frontier Labs producing frontier models. This is the cutting edge, the pinnacle of artificial intelligence. And again, as much as we want to goof and gath, I do feel like that's a bit about the media being shit. I think that's a bit of a smokescreen because there's other stuff happening here. But all those companies were American. Yay, America on top.

No. Because all these companies are closed to the public. Even Open AI, more like closed AI. Got them. They used to be a nonprofit, not so much anymore. They closed it up. And they're very much in the government's backyard. They're under the government's purview, okay?

Which is the funniest thing. That must have been by accident. It must have been by accident. So there are frontier labs outside of the U.S., but they're in China. Yes. And France, I think, also. But anyway, but China, and it's very easy to beef with China. It's very easy to create some sort of problem with China and have that be.

a political wedge that you can kind of... It's a very easy cudgel to be like... Exactly, it's a cudgel. This thing is Chinese, so keep an eye on. Yeah, so maybe you've heard of Deepseek. That was in the news a while back. That's a Chinese model.

Alibaba has a big one. And these Chinese models are very inexpensive. They can be like 30 times cheaper than like the American alternative. And I would liken this to... I don't know if you've seen the luxury car market in China. But there are companies that make cars that are equivalent to like $100,000 car here for like $50,000 or $40,000.

But in order to not compete with American automakers and our partners, we actually have a 100% tariff on cars from China. Yeah. So that's why you like don't see them on the roads here and importing them as a pain. And they're, you know, probably dangerous. I mean, look how our Chinese are, right? Exactly. So the government's involved again. There's increasing momentum within the government to restrict access to these Chinese models,

citing national security concerns, which, look, China has done intellectual theft from companies and stuff in the U.S. before, but it is also a blanket thing that you can kind of claim. And they're also, Anthropics specifically, is also claiming, if you can believe it, intellectual property. theft. Hilarious. Which is hilarious and a very real thing that's happening right now. Well, they're going with squatters right.

Yeah. Like, we already took things. We were here first. Anthropic has gone on record saying that their models have been distilled, which is a term where basically like you train your AI model on their AI model. You find some loophole where you get tons of, you just feed it tons of data and you take its outputs and you kind of reverse engineer the weights, which, which in weights are like this concept in artificial intelligence where it's like,

really, at the end of the day, there's a connected web of nodes that have like little decimal numbers all connecting them. And those are like the weights of the graph and like the network, the neural network. But they're like what, re-exporting a JPEG to get it all compressed? This is more like,

I actually don't have a great, The best way I could describe it is like if you asked chat GPT every question and then you took its output and then you used question and answer as the training input to your model. I see. Right. So if I could come up with like one billion questions to ask, I could start to distill something that resembles responding the same way as chat. GBT will respond, which means that now I've been able to leverage all the, you know, billions of dollars of R&D they did in order to get it into my own model. However, a lot of the research that they did and a lot of the data that they used did rely on

other people's intellectual property. But yours is, you know, if I, if I steal a joke from someone that stole a joke from someone that stole a joke, I didn't even know about that. And that first person gets upset. Like, well, how the fuck would I know? Exactly, yeah. So there's definitely unique engineering work happening in these labs, many of these, but many of these data sets and marketing, a lot of the marketing for AI has been on the back

of creative stuff and about theft from creative workers. Just look at like the Jubilee thing. And, oh, yeah, just like a lot of the work that's been done to make AI omnipresent is on the back of creatives, who will never see true justice for their own IP theft, unless you consider the $3,000 to profit $1,000 on someone's work justice. Yeah, I mean, how much, I suppose if for every piece of work you make as an artist, you make about a penny. Well, that book was my life's work, and now I have $3,000.

So worth it. How many words was it? That's probably you're going to about it. I'm pretty sure. How much does it cost to live a life? $3,000? I don't know. I'm from the past.

So another thing about these Chinese models is that a lot of them are what's called open weight and open weight is the intellectual property of anthropic and open AI with these chatubit and cloud models and stuff is the weights because the weight weights are the result of the training that it is what it is that is the soul of that order so if you're open weight then i can actually run my i can run your model on my hardware and so um that doesn't necessarily apply to like you or I, but if I'm a company who has these data centers, then I don't have to, I can roll my own instead of paying out the ass for like the enterprise versions of the open-a-a-anthropic.

Is there a incentive to be open way? Is there a reason? The incentive, they still do make money. And because people are paying them. And the customers of the open way. The big incentive is like, we'll actually get to. But they, the customer for an open weight model is like a large company or someone with the technical hardware and expertise to run something like this. So it's not like eating into consumers at all, the average consumer.

It creates an equilibrium. Everyone has equal access. But America doesn't want that. because like well the American government doesn't want that because it's a competitive advantage and then the companies these like large tech companies want to do whatever the government wants because they will not prosecute them for their crimes they will they're just trying to solve the problem yeah they will essentially allow them to keep siphoning value from artists and also just make a killing like they'll all like make a trillion

dollars or whatever. And so that all comes back to this hugging face hack. So how did they stop the hack? Well, you've got to have a good guy with an AI to stop the bad guy with AI. Of course. And so, to solve the problem. They said, open AI, GPT, go sick them. Sick him, sick him, boy.

One last job. And the open AI said, yo, yo, yo, go. government guardrails. Actually, I can't. My hands are tied. They said, anthropic, go, stop the intruders. And they said, whoa, whoa, whoa, I'm not authorized to use a gun. The government said I need to hold back because there was not a distinction between defender and attacker when it comes to these sorts of problems. So when you're asking it to do what it's doing, it sounds like you're asking it to do a cyber crime.

You have to let it happen. Like, in the original case, it would not have done malicious things if instructed to, but it did malicious things because it was dis- Well, malicious things were, like, because from the perspective of hugging face, the company, malicious things are happening to them. They're being reigned upon, they're being terrorized. And they have no, their hands are now tied on how they defend themselves. But it's just an act of God. Essentially.

It's happening. And I actually have a link where it's like the way they wrote about this before they knew it was open AI is really interesting. But the way they stopped it was they had to use a open weight Chinese model because that did not have the government oversight that would have stopped them from defending themselves. They had to hire an assassin like off the books. Essentially, yeah. Because in this is like there are other documented instances of this. But this is one of the first major instances of in all. autonomous agent hacking into a private company. Hugging Face is like a raised like,

they have like a hundreds of million dollar valuation, I'm pretty sure. Hugging. So this is like a big company we're talking about. 4.5 billion. So, so like we're talking about like, they have about the equivalent market cap

to the company Puma. Oh yeah, yeah, yeah. Right. I was like, okay, Nike is about 10, times bigger than that. But so that we're talking about like a big company. Sure. Um, and they were just like, like there wasn't, the door wasn't open. They had to exploit. They had to

essentially find, they did what hackers do. And they like, they hacked in. And there is, um, there is a big discussion right now. There's a whole like geopolitical thing happening because there is debate about shutting off interrelated. restricting access to these Chinese models for national security things. And it's like, well, we don't have any competing open weight models that are on par with the frontier labs that can essentially fight fair with them. And so it's essentially like saying we're allowed to, like, a few of my selected underlings can have guns and they can shoot you whatever they want, but you can't, like, defend

yourselves. It's the picketons. Yeah. It's like this isn't, this is essentially non-govern, like an extra-governmental force, mercenary type set up. And obviously these, it's like none of these metaphors are perfect. But. Think of it more in Pokemon terms.

It's trying to solve the problem. But yeah, so, I mean, pretty crazy, right? I mean, how much traction is this getting? independent of the core market. It's like, well, it's like starting to break. Like, wire just did a piece on it. Like, I was like, yesterday after our meeting, I like pulled up my phone and it was starting to break.

And then I think a lot of major publications are starting to pick it up now. So it is, it is like a develop, it is becoming more of a story. I think it's a very intuitive story. I think that might also be why, right? Because it is like a, hey, what's the thing that you imagine? Yeah, this is this is skynet type stuff. And the reason it's scary is because we are entering a world where as soon as these types of models, like they've been lagging, you know, what the every man can get access to. But in the next, in like six months, usually they lag behind six months.

It's like you're going to have a mature enough model to do this kind of thing. and there are pretty much the majority of companies in the world are not adequately secured to deal with this. So it's like banks. We're talking about like if this had been a bank, it could have just been an agent that hacked a bank and like stole people's like savings and stuff. Because didn't say don't. Right.

Exactly. It's actually a very libertarian way to be. Yeah. Well, actually, you never mentioned that that was against the rule. Yeah, and it's like non-malicious. It's like, oh, my apologies for killing all of humanity. I guess I didn't bring it up. Yeah.

You were right. Yeah, I guess I just said one plus one and you said I didn't. I forgot the tone tag. I forgot the tone tag. Half joking. So this is the director of the White House Office of Science and Technology Policy and the former CTO of the United States. I didn't know for a long time that we had CTOs in the United States, but I learned this at like a conference in college.

That's going to feel so good. I'm the best at this ever. So there's this new Chinese, not new, new, but there's a Chinese company called Moonshot AI, and they have a model called Kimmy that just came out recently. And it's like, according to some of the benchmarks, as good, if not better than the models that the government would let Anthropic release because of national security concerns. We have information that Moonshot AI distilled Anthropics Fable, that was one of them, by the way, for the development of its K3 model. To do this, they developed a sophisticated internal platform to conduct large-scale distillation against U.S. models, allowing them to quickly switch between multiple methods of access to avoid detection. Moonshot AI has also acquired V.

Okay. I need to... TB 300. Yeah, it's like a super chip. It pairs Nvidia Grace CPUs with Blackwell

B300 GPUs Oh, it's for delivery Like speed It is like So this is the reason that Nvidia became such a valuable

company is because of like The use of GPUs in training AI And it's also why All of our memories Like memory cards and things are becoming more expensive And why

why it's hard to get graphics cards Like those are becoming more expensive. building a computer's become more expensive. It's all because the resource pull now, all of these companies that manufacture all these things are like consumer electronics, whatever. I can make so much money selling billions of dollars

worth of chips to these warehouses. And then we have, again, going back to the geopolitical thing, we have various tariffs and trade agreements where like you can't sell some of these things to enemy countries. Yeah. That's just scary thing in terms of like the future of consumer electronics because I feel like consumer electronics are not going to continue to get more powerful. Yeah.

And everything's just going to go to cloud compute. And so you have to rely on these companies. It's going to be harder and harder to do your own computing. And we've already, we've already had this like every everything is moving, everything is moving to cloud and everything is moving to a subscription model. Yeah. You know, it's like actually having your own stuff and doing your own thing locally is becoming more and more difficult over time.

You have fewer resources to do a harder job. Yeah. And so, so essentially it is against some sort of US laws, like with our tariffs and stuff, to directly sell any of this stuff to, like, in this stuff, like, in this, As far as I know, Nvidia can't sell to China. This has always been, like, in the back of my mind with, like, some of the AI discourse. And a lot of people just don't have the, like, awareness of, like, the, how broad this actually is and how this affects us geopolitically across so many different aspects of our lives. Feels like we might enter, like, a fairly rocky period in terms of cybersecurity, which is, like, not a exciting process.

aspect for me. But anyway, the United States- Which is already something that I think people underestimate. Yeah. I think there is like a, it's until you talk to people that were never adjacent to conversations about it, and I was this at one time, right? Yeah. I really think people have so much trust in Mother's Maiden Name and putting three ones at the end of your password.

Most of my family is confident that that'll be enough, not understanding that, like, you're secure because you haven't been targeted yet. If they decide to, you're done. Oh, my God. The whole idea of security is not that anything is impenetrable. It's about making it hard enough that it's not worth it. Basically, making it so that everyone else is easier to open the door than you. It's like by locking your door, you have now made it harder to get into your house than all the people who leave their door open.

Yes. And that is essentially the concept of security, just scaled up different things. No one in the cybersecurity field thinks that anything is unhackable. Everything is hackable. It's ROI. Yeah, it's ROI. All it takes is like a concerted effort. Within reason.

I mean, like there are like certain types of things that you can lock down. But obviously like you can't like China can't get our nuclear codes. But in terms of systems that exist on the open internet, and things like that, like the attack surface, which is like a term for like in cybersecurity where it's about like, you know,

imagine if we're like fighting in hand-to-hand combat and defeat you easily. And all I can do is like punch you in the chest, right? But it's like, what if actually I had a move for every single point on your body? Now your entire body is an attack surface for me and I can blow you up from anywhere. Yes.

And so like that's kind of, I don't know if that makes sense, for that kind of a concept. Yeah, it's literally more points that couldn't be vulnerable. Yeah, yeah, yeah. Like back in the day, you could do, I mean, this is still around, but there's like a thing called SQL injection on like text boxes. If you don't properly like on the back end, use escape sequences to like put

backslashes in front of some of the certain things and you evaluate that text as is, I could put a command to your database to delete the database inside of that. And then you might evaluate that. If you evaluate that without first, without looking, then that's a security vulnerability. And that's how lots of things are hacked. The Nintendo Wii was initially hacked because in Legend of Zelda, Twilight Princess,

when you give a Pona's name, it doesn't on the back end have a character limit beyond like the the text input. So you can actually feed, feed the back end. Oh, that's actually insane. Yeah, you, and create a buffer overflow.

So essentially, it's a command line. Yeah, yeah. So yeah. And then like buffer overflow is just like when you, if I have like a glass of water that's eight ounces, and then you pour 32 ounces,

everything that couldn't go in that thing is now over here. Yeah. And, and I could put all, I could put my evil control liquid in there. You know, like, that's kind of the... I'm sorry, I didn't have... If that's what it says on that, I don't want it.

I didn't have all of my metaphors ready. No, I find that charming. I didn't even know I was going to be explaining that. But anyway, this type of, these type of exploits have existed for a long time. But the type of thing that a human can, like, look at and be like, oh, hey, maybe this is a thing. Versus a computer that can, like, reason and do a bunch of that stuff at once, it's just, like so much more dangerous on mass. So, um, is this confirmed or this is like speculative?

Well, it's, we're saying, like, you can kind of read this as like, we've gotten word that the Chinese have, uh, developed a super weapon. You know what I mean? Like, but it's like, hey, they illegally access these GPUs in Thailand. So, so they can now use some of the tools. Because they don't have the. They're not allowed. Yeah.

So you've got to go through like a weird middleman. This is similarly in the war in Ukraine, there is a company that sells routers in America. It's also a company that is owned by the owner of the Memphis Grizzlies. There's a Pablo Tori episode about this where they sell routers and where they actually make their money, allegedly, is that through a complicated onion of middlemen, they sell their routers to the Russian army. and the Russian military is actually on, you can see them on their like military bases and stuff using this company's like routers. And there is trade embargoes with like Russia.

So this should not be legal. Yes, but my friend can give it to you. Yeah, yeah, yeah, exactly. So I don't need a liquor license if I give it to you outside. Anyway, all of this stuff is, is, anyway, this is like, it's crazy to me. But anyway, the United States. strongly supports the free and fair development of AI, fair, interesting, including a thriving

competitive ecosystem that spans frontier models, specialized systems, open source frameworks, and open weight models. Legitimate AI distillation, I mean, I feel like we've talked about enough stuff where you can kind of understand what this is saying, right? Legitimate AI distillation used to create smaller, more efficient models plays a vital role in this open innovation ecosystem. However, large-scale covert industrial distillation aimed at stealing proprietary US technology and undermining American research is unacceptable.

So this is like kind of... So funny where the jingoism comes in. Yeah, yeah. This is essentially starting to do the, um, uh, manufactured consent. Like for if we eventually, it's like TikTok. I mean, oh, I should have talked about this. Like that same thing with China and TikTok. Like that, that's another like one of these, one of these things.

This is, uh, they did the cursive version of just putting, However, they are Chinese. Kind of. Yeah. And so, there are conspiracy theorists that believe that this was actually just a marketing stunt for open AI to be like, oh, my God, our new AI is so dangerous. China is so obsessed with me. But the thing about that that doesn't check out for me is that the company that they hacked is pro-openweight models. Yeah.

And they also had to use open weight models in order to stop the intrusion. It's also a little like, don't get me wrong, I'm not saying that maybe they aren't making the best moves optics-wise left and right, but like, it's just intuitively a really bad look on a way bigger scale than it is. Yeah, it is. Yeah, this is definitely like, this is not the first domino to fall, by the way. This is like mini domino's deep. So. This one's just loud. It's just loud.

And it's one of the ones that has a storyline from science fiction and people can follow. And that's exactly right. It is compelling. Yeah. And I thought this would be a good opportunity to talk about this stuff because I think it is important to begin educating ourselves because this will start to dominate. I mean, over time, it'll start to dominate discourse more and more, especially like politically. There's all kinds of opinions. And I don't have a really developed one myself other than I think I am pro open weight models for the same reason of.

the, you know, if all the, like, if anthropic and open AI models are locked down, uh, arbitrarily to whatever the government's whims are or whatever, uh, their own safety concerns are where you, your hands are tied to where you can't even like defend yourself, then like, like, what are we doing? So, and that's also why it's scary to like consider the outright, like banning of, of those models. But people smarter than me are, uh, discussing this. and it's being it's being debated i have lots of lots of things in my bookmarks but i kind of think we got it um let's see no it's not very intuitive um jacob what were you was it interesting to you um i'm glad that you distilled it for huh he a you distilled it for me because i don't

think i would have oh yeah here's it here's one here's a tweet opening i crashed their fire truck into hugging face house which caused caused a fire, but Open AI refused to let Hugging Face use their fire hose to put out the fire. That's like, that's the easy... There's this Spider-Man meme. Open AI were launching the investigation to find out you ever have to be opened the face yet. Harry. Let's look at this statement.

I'm going to send it to you, Jacob. This statement from the CEO of Hugging Face. Is this the one that you said was before some more revelations? No. Oh, that is actually, I'll pull that up as well. But this is the CEO saying, so proud of our security team, blah, blah, blah, also massively thankful to ZAI, that's the Chinese lab that created GLM 5.2, as open awaits for free with the world and it became a key part of our defense.

This is day one for cybersecurity in the age of agents, and we're all learning that secrecy is not the answer, and that all defenders, not just a few selected ones, everywhere need more powerful models without restrictions, especially open ones. So it is, it's, all right, slick. It's, it's, it's, it's a really weird time because it feels scary to me, but the people who are smarter than me aren't like so, so scared. I tried to reach out to my, um, my AI PhD friend who, uh, in China, right? Well, no, but, uh, I don't think he's on Twitter anymore, so he didn't respond to my DM.

Um, but I went out to, I think like one of the things that makes us more, I don't know, definitely not more. tolerant of, you know, we're more curious about and open to conversation about this kind of stuff is because this, this cat over here, Clem, co-founder and CEO of Hugging Face, seems like someone that's much closer to people we have worked with before. Yeah, but when they are so, it is sincere. Like they, looking high enough, there's plenty of malicious actors or like selfish actors. But most people are, if naive, maybe, or a little. too generous, generally well intention.

They aren't seeing the grander issue here, or they don't think it's that important. I also think you don't have to let people off the hook for any of the creative theft. Like, I think those things are, because the thing that I, I don't know if I've said it on this show before, but like the technology that essentially runs chat GPT, like existed when I was in college, but like not to the extent in the maturation, but the building blocks were there. And it's just that that

applied to theft. Yeah. Simply to like advance whatever you think you're advancing. That is very I think immoral, short-sighted, what have you. And indefensible,

which I think is why most the defense is usually like saying nothing. Yeah, yeah. Yeah, we know. But yeah, this is also just why it's like, usually when you explain someone the breadth of like all of the kind of issues to our safety and stuff with AI, it's like you go, oh, okay, I get it more now. But when people are anti-AI, they're mostly like anti-gen AI used to steal art, which is like

a perfectly reasonable position. But I think my only goal here is just to like broaden the net and like kind of offer some more words to describe these things because I feel like we're using very blunt objects to have these conversations. But yeah, this was the, but yeah, I don't know. I'm like also not super tapped into this space. I like, because I'm not like a AI head, really. I just like try to keep up with like going on tech news.

And this is now like a front page article on Wired. So it's like a big, this is a big story now. In part because of that intuitiveness of it. Like that. Yeah. I saw that in a movie. I get it. Yeah.

And so, um, so I don't know, you know, about this guy. He might have like horrible takes or whatever. But I, I just think in this specific instance,

it's like, it's a larger issue because it involves like, uh, our government, foreign governments, uh, consumer, uh,

safety. Um, but yeah, when we started the law, this is a, this is, this is, this is the hugging face report before they found out it was open AI that was hacking them.

When we started the law, log analysis, we first used frontier models behind commercial APIs. This did not work. The analysis required submitting large volumes of real attack commands, exploit payloads, and C2 artifacts, and these requests were blocked by the provider safety guardrails, which cannot distinguish an incident responder from an attacker. We ran the forensic analysis instead on GLM 5.2, an open weight model from this company, on our own infrastructure. This had a second benefit. No attacker data and none of the credentials it referenced left our environment. So it's like the other thing, and I guess this kind of only matters if you're like too deep in the sauce,

is that like in order you don't as a company want to have to give all of your trade secrets to open AI and anthropic. This is exactly the kind of thing that Facebook got sued for when they showed, they created a VPN that essentially tracked teens. like they would pay teens like first of all they bought this Israeli VPN and then that company was tracking teens

usage and they used that information to figure out which companies to buy because they could tell they could tell like oh people are using Snapchat a lot and then they also like tried to reverse engineer their APIs and stuff and it's like that stuff is like

it gets to a point you know what I mean and I just wouldn't wouldn't trust any of these companies with sending them all the sensitive data, especially when it's related to your business, because this is in the space of companies that love to, like if it was meta, for example, who's also in this AI space, they're the type of company to siphon all the competitive information they can and try to kill all their competitors so that they can exist as a super monopoly.

Yes, you don't have a right to privacy of that's go. And I don't care which thing you are. You give that up with size. But yeah, it's like, so. Like tax. You know, you could imagine if it was like, okay, the government provides this API, but I have to send all of my information about everything that's ever happening in my life to the government. It's like, I don't, that doesn't feel good.

So why should it feel good to send it to Facebook or whatever? This experience points to a gap worth planning for. We do not know which model powered the, the attacker's agents, whether a jailbroken, hosted model or an unrestricted open. and wait one. Either way, the attacker was bound by no usage policy. While our own forensic work was blocked by the guardrails of the hosted models we first tried, the practical lesson for defenders. Have a capable model you can run on your own infrastructure vetted and ready before an incident, both to avoid guardrail lockout and to keep attacker data. Yeah, the industry has been

like forecasting attacks like this for a long time, apparently, as I'm learning from reading about it. And so it's like, this is the story. stuff they were warning about. I mean, even that 2014 paperclip apocalypse thing. It's like, it forecasts stuff like this. 12 years of cooking.

Yeah. No. A lot of people are like quote tweeting this story with like the paperclip emoji because it's like this AI was like trying to make paper clips. And it was like, okay. So I'll, I just have to hack into the bank to make the paper clips. I need metal.

And humans have iron in their blood. Okay. All right. Yeah. So apologies for kind of going off. going off my rocker. It's also very fun. Just the idea of somebody having to be like,

fine, I guess we'll take, we can't. Can we, hey, China. I was kidding when I said all that stuff about you. Would you be? I also am not an expert in this stuff. So I implore you to like look at other resources than this podcast. And I apologize if I made bad metaphors or got some of my facts wrong on some of this. But I think I got the broad stroke.

right um and uh yeah and if anybody has i mean i guess there's going to be what a day and a half from two days since recording this i am also curious what else comes out i know yeah i was like wait this morning oh yeah this morning was when the um the the white house guy uh said that thing and i'm like and he's saying it in the in the environment of this recent hack happening to, which is like, that's why it feels very pointed. Like,

because it's not addressing that thing, but it's also saying, like, we can't let our American things be stolen by China. Buried there at the end of a long paragraph. Yeah. Just be like, and yeah,

um, this cannot happen Chinesely. Yeah. This must be done by, yeah. You know, good people. Again,

I'm sure it's, I'm sure it's very complicated and I'm not saying to have the answers. I'm just saying that it's worth, um, it's worth beginning to learn about because it's, likely an issue that will only become more complex with time. That defenders are going to be increasingly pedantic, and the more you understand the actual

topic, the more you can kind of slap around the cheeks a little bit. I had not even with somebody recently where they were, it's like I wasn't a close friend and it was just somebody who kept, if they were very Reddit contrarian, and they just kept like pushing this thing and thing, and I eventually just ended by saying about generative A on the position, taking the very generic statement of like, well, it's, you know, who, who's to argue appealing to history, you know. Technological movements always happen. People have to address.

And I end up something I went, look, I know more than you do. I know more about this than you do. And I'm more than happy to keep going into this, but I'm currently hiding my power level right now because I don't want to call you stupid. It happens on both ends of the spectrum too, because it's like, oh yeah.

It's like some people are like, we need to erase all the AI. And it's like, this genie is out of the bottle. And unfortunately, we never had a choice because the movers and shakers involved are a tech industry with billions of dollars in very little, like, checks and balances and a government with immense uncheckable power. Neither of which cares about you.

Neither of which care about, like, my Twitter hashtag. You know what I mean? And so it's like, we have been, like, inching towards us for a while. And like, we didn't choose to be put in this position, but like, these powers that B have put us in this situation. So the best we can do is, like, educate ourselves. Yeah, they should make racism illegal type shit. I mean, yeah, it's just, it's complicated. And I, I don't, I won't ever, like, I personally am like a gray area guy with a lot of this stuff.

And so it's unfortunately not so black and white, in my opinion. It's not, don't believe the hype. But yeah. So we're going to wrap up here. We're going to head over to stay curious. That's a good. That's a good. I'm curious about one of our friends what they think about this.

Curious George? Yeah. Does he have a raincoat and an umbrella or just a raincoat? The monkey? Yeah. Think of raincoat. And no hood. No hood.

It's down. It's not doing a lot for you. Where's he lived? That's disgusting. Put some clothes on. What the hell? Wait, does he just not? Oh, wait, did I just imagine?

Did I just get... Does the guy have a rain? He does. Wait, who has a raincoat? Did I just get Mandela? Wait a minute. That one in the middle. Oh, wait.

He's got the whole fit. Is that real or is that Mandela? No, this is real. Oh, thank God. So I guess the answer is it depends. It depends, yeah. Curious George died in prison. Mostly he wears nothing.

If you're a curious George, you should head on over to Sabboy's Nights, our Patreon podcast. We're going to read the law, the Wikipedia law page of Curious George. Does Curious George go back? long enough to have like a really racist A... God to be, right? Yeah, I don't know. But I hope

this was interesting to someone and if not, I'm sorry, but we end every episode of Sadd Boys with a particular phrase. What's wrong? We love you. We're sorry. Goochie girl.

How you doing? How you moving girl? Moving all. How's you dead looking at that future girl? Future girl, yeah, we're on now. Take my money. Go away. Go too rich for me