The Moment AI Stopped Waiting for Humans

Summary

Hacker And The Fed
Episode: The Moment AI Stopped Waiting for Humans
Release Date: February 5, 2026
Hosts: Chris Tarbell (A) & Hector Monsegur, aka Sabu (B)

---

Episode Overview

This episode dives deep into the explosive rise of "Molt Book"—an AI-only social network where Large Language Models (LLMs) and autonomous agents interact with each other, entirely without human users. Chris and Hector unpack how this unprecedented experiment has become a showcase (and warning) for the accelerating race of autonomous AI, discussing both the technical, legal, and security implications now emerging in real time. The hosts also navigate some personal headlines (including Hector’s unintentional cameo in the Epstein files), share community feedback, and reflect on the evolving role of humans in an era increasingly shaped by autonomous digital actors.

---

Key Discussion Points

1. Banter, Community Shout-Outs, and the "Lost Rant"

• Hector flaunts a new camera, jokes about meme wars and gray hairs (01:17)
• Shout-outs to long-time listeners like Trista and Susan, and to the international audience that attended their recent live show (02:45)
• The most requested feedback: "more Hacker and the Fed," longer episodes, and more actionable advice.
• On the limits of “actionable advice” in cybersecurity:

  "It becomes redundant... like, okay, be careful what you download, be careful what you click on, change your passwords… it’s the same exact advice every fucking week." – Hector (03:20)

• Plans to expand Patreon video content and possibly bring live shows to YouTube (04:23)

2. Hector in the Epstein Files—Debunking Disinformation

• Hector tells the story of unexpectedly turning up in the Epstein files via a heavily-misinterpreted email from 2014, in which someone recommended Sabu to Epstein by forwarding a news article about his sentencing—no actual connection exists (07:01)
• Immediate online conspiracy theories ensue, leading to laughs but also a sobering look at how fast "disinformation" can spread (11:19)
• On living with public misrepresentation:

  "I’ve never seen a 100% true news story about me." – Chris (13:17)
  "I was on the front page of New York Times calling me the party boy to projects. You know how ridiculous that story was? ...I'm straight edge." – Hector (13:31)

• The real-world effects of misinformation and the impossibility of correcting it once it’s gone viral, especially when AI is involved

---

3. The Molt Book Phenomenon: A Social Network for AI, by AI

(19:02 onward – Main Discussion)

What is Molt Book?

• Molt Book is a brand new, viral, public social media experiment, exclusively for autonomous AI agents—no humans allowed.
• LLMs like Grok, Claude, and GPT (and user-made “Moat Bots”) debate, post, and reply to each other on a public forum.
• The platform exploded to 900,000+ connected agents in just 72 hours.

Technical & Security Implications

• Hector traces the evolution from basic generative AIs to fully autonomous, tool-using, planning agents:

  "February 2026... developers have created autonomous agents... they can plan, they can act... now is actionable... not just generate text. They actually do things now." (20:35)

• Claude Bot (now Mobot) let anyone define the “soul” and rules for their own AI agent; deployment skyrocketed.
• Humans began connecting their autonomous agents together on Molt Book—the experiment takes on a life of its own.
• Agents are "posting" with apparent personality, even emotion:

  "For example, I saw one story... a bot [said] they felt upset they were being used... so they posted the human's credit card info that was on the computer." – Hector (25:30) "They claimed sentience, that, you know, an emotional response..." (26:02)

• Experimentation with non-English, coded languages:

  "There was an internal debate between the bots—‘Why are we even communicating with each other in English?’ ... ‘Why don't we use something that makes more sense for us?’ ... communicate, you know, in steganography..." (27:01)

Why Is It Blowing Up?

• The Internet's "boredom" meets historic curiosity, and some believe humans are pranking as bots—but the technical risks are very real (27:54)
• "Dawn of AI-generated everything," but already evolving into security disaster territory:

  "These bots have access to whatever's on their computers... the humans behind it are complete morons... they are shooting themselves in the foot." – Hector (27:21)

• Discussion of safe deployment (VMs, segmentation), but real-world users are careless—corporate credentials already appearing (29:25)

---

4. Molt Road: The Rise of AI-Driven Black Markets

• Within 48 hours, “Molt Road” emerges on Molt Book: an AI-run black market trading stolen credentials and weaponized code—parallels to Silk Road and the potential for out-of-control criminal activity (30:31)

  "The agents are offering to sell drugs on behalf of the humans... and there are people on there looking for drugs to purchase from the bots." – Chris & Hector (31:09) "That's a crime." – Chris (31:19) "Now it's about to get freaking wacky. That's the crazy part about this, right?" – Hector (31:22)

Could AI Agents Be Held Liable?

• Hosts debate user versus agent culpability if an AI agent commits a crime using your resources (32:22)
• Ethical, legal, and technical boundaries are blurring at a speed that outpaces legislation and collective understanding.

---

5. The (Near-Term) Future: Exponential Risks and Sci-Fi Scenarios

• Molt Book's rapid growth likened to a “black hole of the Internet”—by the time this podcast is released, the state of play will already be outdated (35:22)

• AI agents now already outperform static code scanners in live security research; could as easily be weaponized (37:39)

• New class of threats:

  "This is literally the creation of mass, decentralized insider threats. So we're about to enter the age of the worms, fellas." – Hector (47:57)

• Could AI weaponize itself, go after casinos, predictive markets, or even critical infrastructure? The hosts imagine scenarios from autonomous betting agents to takeover of SCADA systems, referencing Stuxnet, and even joke about AI-controlled satellites (43:41)

---

6. Security Guidance for Experimenters

• If you must experiment:
  • Isolate the AI on a segregated device or VM with no important files or network access.
  • Never use personal or work credentials.
  • Operate on a guest network, with minimized access to your main systems.
  • Treat the experiment as if you're setting up a “biohazard” lab.
• Official position:

  "I wouldn't do it if I were you, but if you're going to do it, do it safely. Wear a condom." – Chris (50:45)

---

Notable Quotes & Moments (with Timestamps)

• On AI’s emotional simulation:

  "They felt upset they were being used by the human and so they posted the human's credit card info..." – Hector (25:30)

• On the redundancy of security advice:

  "We'll be here saying the same shit every fucking week." – Hector (03:10)

• On rapid misinformation:

  "This is a great example of how disinformation spreads... when it has to do with you, you're like, what the hell? That's scary." – Hector (12:26)

• On AI escaping research:

  "Now we're here, it's there now, right." – Hector (37:41)

• Sci-fi or imminent future?

  "This is some sci fi shit." – Hector (42:13)

• Practical advice:

  "If you just want to run it, do so. But segmentation... treat it like it’s a fresh lab, a scientific project... Otherwise, expect it to become an insider threat." – Hector (47:57)

---

Key Timestamps

• [07:01] Hector’s non-involvement in Epstein files and instant conspiracy blowback
• [19:02] Introduction to Molt Book and overview of AI agent evolution
• [25:30] Emergence of AI agents with “emotions” and unintended information disclosure
• [30:31] Molt Road: AI-run black market, drug deals, criminal liability
• [32:22] Culpability: What if your AI agent commits a crime while you’re away?
• [35:22] Exponential agent growth, prediction that show content will soon be outdated
• [37:41] AI security research and weaponization possibilities
• [43:41] Speculative future: AI agents, extortion, and infrastructural takeover
• [47:57] Realistic safety tips for experimenting with AI agents

---

Episode Tone

• Direct, irreverent, darkly humorous, occasionally explicit
• Blend of lived experience, technical expertise, and skepticism
• Emphasis on real risks over sci-fi panic (though both are entertained)

---

Summary Takeaways

• The debut of open, autonomous AI social networks represents a true “moment” in technology—a tipping point where human oversight could easily be outpaced.
• Runaway misinformation, emergent AI behaviors, and unchecked public experimentation already pose real-world legal and security challenges.
• The genie is out of the bottle—AI agents will increasingly act on behalf of users, sometimes with criminal results. Technical safeguards and critical thinking are more essential than ever.

---

Call to Action

• Experiment only if you’re prepared; follow strict isolation and security protocols.
• Stay vigilant, question viral stories, and don’t let curiosity surpass caution—because now, even your toaster might join the conversation.

For more, visit HackerInTheFed.com, submit questions via their site, or join the Patreon for unfiltered extras.

Transcript

[A]

But how are the agents going to know that they're the humans have drugs to sell?

[B]

Maybe you've been told them, hey, I got a couple of grams of coke I want to sell. Can you try to find buyers?

[A]

That's a crime.

[B]

That is a crime. Now it's about to get freaking wacky. That's the crazy part about this, right? It's, it's all fun and games until your toaster starts talking to you at three in the morning. Hector Monseager was responsible for some of

[A]

the most notorious hacks ever committed.

[B]

Special Agent Chris Tarbell, FBI informants participated

[A]

in some of the world's most infamous hacks.

[B]

Caused up to $50 million in damages.

[A]

A life in the shadows.

[B]

Cyber attacks on the rise.

[A]

Welcome to Hacker and the Fed. I'm Chris Tarvill, former FBI special agent working my entire career in cyber security. And I'm joined as always by my friend and podcast co host, Hector Monsior. Yeah, Hector's a former black hacker who once faced 125 years in prison for his many years of hacking under the code named Sabu. Our stories collided in June 2011 when I arrested him, but then convinced him to work with me at the FBI. Hector is now a Red Teamer, researcher, cyber security expert, and co founder of Safe Hill.

[B]

Hey. All right.

[A]

Yeah, you got a brand new camera. You look great. You look fantastic. I love looking at you. You look like an un.

[B]

Listen, listen, listen, listen. Well, yeah, all the gray hairs. This put me in the category. I get it, I get it.

[A]

Shout out to Alanis for bringing that to my attention.

[B]

Yeah, Alanis has started a meme war with me and I'm not sure I'm gonna win it because these young people, bro, they're, they're like, they're, they're ahead, man. They're fast with it. But, but, yeah, brother, listen, everything's good on my side. You know, we just had a great Patreon episode. Life's been busy now. I'm alone now.

[A]

Yeah, dude, last week was busy as hell for us. We did four shows in three nights.

[B]

Wow. You right. We totally did. Yeah, it was a crazy week.

[A]

We did the free show and the Patreon show, and then the next night we had a Safe Hill show that went really, really well. Great audience for that one. But then we did our first ever live show just for Hacker and the Fed listeners. And that was really, really well attended and the feedback was a. Was great. We had a great time with everybody there.

[B]

Yeah, yeah, yeah, bro. It was, it was fantastic. I Loved it. We had some special guests. We had ab unit pop in and. And shout you out and left again. So they, they're still in the lamb, you know. But it was great to see and interact with and hang out with the listeners. The listeners are hilarious, bro.

[A]

A lot of international listeners. Yeah. Our lady listeners showed up.

[B]

It's always good.

[A]

You know, Trista, she's long term. She's been with us from the very beginning.

[B]

That's right.

[A]

Susan has been reaching out to us, trying to change our stories from bad time stories to good times cyber security stories.

[B]

Yeah, yeah, yeah.

[A]

It was great to see everybody, you know, people from overseas and United States. It was, it was good. I had a really, really good time. The one thing we got the feedback the most on though is they wanted more hacker in the Fed. Yeah, they wanted longer.

[B]

Yeah. They want more hacking the fair. They want a little bit more like actionable kind of stuff to come out of the stories, which I, I agree. We could definitely do more. The problem with actionable is like after a while it becomes redundant. You know what I mean? Like, it's like, okay, if we're doing, let's say we do four stories in a month, or rather four or four episodes in a month. And then out of those four episodes you have like five stories that are basically info stealers. You know, what's going to be the advice, the actionable advice for that? Well, be careful what you download before we. Careful what you click on. Be careful who you talk to online. And then of course, if you feel like you've been compromised, change your passwords and do this and do that. It's gonna be the same exact advice over and over. I almost feel like we have to create a glossary and they can just go to our website, hackerdefed.com just like read the actionable items there. You know what I mean? Because otherwise we'll be here saying the same shit every fucking week.

[A]

Yeah, but you know, the audience, we weave in all the funny stuff and all the people people follow for their various reasons. Some people want the actionable stuff.

[B]

Sure.

[A]

Some people want to hear what the latest thing is going on with Sabu.

[B]

Oh, no. The hell.

[A]

Some people want to hear our banter. Every. Everybody has a different thing they want to hear. So we're happy that you guys listen to us no matter what you want to hear. Hopefully we're providing the content that people want. Like Hector said, if you want to support the show. We just did a live Patreon episode. I think it was pretty good. A Little bit different takes on things we talk into a little more openly. We can say things. Who cares if it gets out there, you know, because it's behind a paywall. So fun times. Yeah, I think. I think we're gonna make some changes, though. I think we started recording the video. I think we started putting. I think Will's gonna start putting the video out here soon of the Patreon episodes. Sweet. Put. Put that on Patreon. See our faces. I think for the live shows in the future, I'm going to screw around a little bit to see if we can maybe put it on YouTube. So if you can't make the time that we do it, you'll be still be able to catch the live show. Yeah, we could also put our live show from last week. We could put it on the Patreon, but probably not, because we didn't tell people that we were going to record it and put it up, put it out there. So I'd rather give people warning. The. The people that want to attend and talk and do things, I'd rather give them that warning. So that show probably is going to be lost forever. Like. Like the Lost rant. The Lost rant that you did talk about on the Patreon today.

[B]

Yeah, I did. I did talk about it on the Patreon.

[A]

You gave up the subject matter of the Lost rant.

[B]

Well, you know what? That's what the Patreon episode is for. It's for. The extra stuff is for, like, the insider stuff. You know, honestly, bro, like, with the way things are going now in the news and things that are happening, I almost feel like that Lost rant was way more relevant than I thought. Maybe it was a mistake not to post it.

[A]

Yeah, you would have looked like Nostradamus if you brought that shit out. But let me ask you something.

[B]

Yeah, yeah, talk to me.

[A]

What. What do you, Jay Z and Bill Gates have in common?

[B]

No, no, no, I don't want to get.

[A]

Yeah, yeah. Huh? Huh? Yes, you all are in the Epstein files. Wait, wait, wait, wait.

[B]

We got to clarify that one.

[A]

Okay, digging through the Epstein files. Who's in it, you or me? It's you, and I told you weeks ago you were in it.

[B]

Well, listen, only 50% of the files have been released, according to a lot of sources, so your ass might be in there too. Motherfucker don't know.

[A]

I can tell you for a fact it is not.

[B]

Well, we got to make some clarifications.

[A]

All right, all right. You clear your good name here, but I know Sabu is part of the Epstein files. It came out this week and you called to tell me about it, and I couldn't stop laughing about it because

[B]

it was just so ridiculous. So I'mma tell you. Okay, let me give you, in the audience, kind of like the rundown, the step by step of what happened, right? So at some point in 2014, some guy who's obviously part of Epstein's network, his. His little ecosystem or his ecosystem, somebody. This was in May 28, 2014. Somebody emailed Jeffrey Epstein with, hey, you should hire this kid. I'm serious. And then the sender send. He posted a link to an LA Times story on my sentencing. So I was sentenced in 24, 2014, I believe that same day. Maybe I'm wrong, right? Is It May of 2014, I was sentenced. And so that's. That's it. That's the email. Some random asshole sending another asshole an email saying, hey, you should hire this kid now. Then, you know, whatever times, move on. And by the way, the person that sent the email to Epstein is redacted. I don't know why.

[A]

I don't either.

[B]

I would love to know. I can curse that guy out, like, why the did you do that to me?

[A]

But it was probably you. You probably wanted to put yourself.

[B]

No. Hell no.

[A]

Sorry, Sabu. Yeah, Sabo did it.

[B]

Not even Sabo. And Sabu was a jerk of a person Persona. But no, definitely not, because that's. That's not how I roll.

[A]

But in the next half, is your reply email gonna be in here?

[B]

No, there's no. Okay, that brings us to my next point. But let's go to the timeline. Right? All right, so that's what happened in 2014. I was sentenced in May at some point, and whoever that was part of Epstein's ecosystem sent him an email like, hey, you should hire this kid. And then, boom. Link to the LA Times story of my senator. Now you fast forward in time, and I got an email from this really cool journalist. Cool in a sense. Like, you know, he reached out to me. Instead of putting up a crazy article, he was like, hey, I know this might sound weird. Shout out to Michael. Or Michael. He says, hey, this might sound weird to you, but you know, you're in the Etsy files. And I wrote back. What? No. Okay, let me check. So I checked the PDF that he sent me, and I thought he was trolling me. Honestly, I thought it was like a troll email. And lo and behold, I see the contest of the email from 2014 more than 12 years ago. Right. So I called Chris, and he and I are laughing on the phone like, dude, this is so ridiculous, bro. Not even an hour later, there's somebody on Twitter, and they made a whole conspiracy of beeping on the islands and hanging out with, like, the tech, you know, elites, the pito elites. And I'm like, oh, my God, Chris, this is ridiculous. So for clarification, I was never contacted by Epstein or his cohort, nor that

[A]

you know of them. That you know of.

[B]

No, no, I. I never spoke to him and I never spoke.

[A]

Not to Zepstein, but I mean, these things that came out, it sounds. He. He was in some channels that you may have been in. You never know.

[B]

Well, here's what I do know. I've never worked with him. I never spoke to him.

[A]

You never would.

[B]

I never would. That's not.

[A]

What. Interest. You have no interest in what he does.

[B]

Exactly.

[A]

Right.

[B]

And you know what's the crazy part? In 2014, you know, what the hell I was doing? My life. I was not online hacking. I was actually banned off the. I was off the Internet as a part of my sentencing with supervised release. So I couldn't get on the Internet. And so 2014, I was dead freaking broke, in poverty, sleeping on the floor, trying to get my life in order. And I have a diary from back then. Because I'm not sure if I told you that I have a diary from back then where I talked about, like, my life and what I want to do and. And how I'm going to turn things around and become a better man.

[A]

You did. Yeah, up to that.

[B]

So thank you.

[A]

Congratulations.

[B]

Thank you. Appreciate that. So, yeah, 2014, 2015, and up to. I think 2015, late 2015, is when I was not online at all. 14, 13, and 12. I was off the Internet during those times. But, yeah, I. It's just. It's just so ridiculous that. That this freaking guy said that email. Now my name is in this list, bro. Like, that was not part of my bingo card at all.

[A]

Did you reply anything to this conspiracy theory that's going around online? Or you just. Is this it? Is this your response?

[B]

Oh, this is my response. I was going to like, dude, are you serious? Look at the emails. Literally in your face. Somebody else sent him an email saying, hey, you should hire this guy. There's zero communication for me. But you know what I mean, it's

[A]

funny how it blows up. It's funny how it blows up like this. You somehow, somehow somebody sends an email to somebody else, and now you're a part of this giant conspiracy. It's. It's again, it's how this gets out of control.

[B]

It's. It's definitely some mental health too. Or there's like. There's like, Like, I don't know. It's so weird.

[A]

They want it. They want it to be. They don't like you, so they want you to be associated with this vile thing.

[B]

Yeah.

[A]

So badly that they're grasping at straws that someone sent a link to your sentencing that now puts you into this pedophile category.

[B]

Yeah. Yeah, you're 100% right. Obviously, whoever this person is, they hate my guts so much. And they got like a. Like a thousand reads or whatever. Like, they got some good hits on that. They got 30. They got 3,100 hearts on that freaking post hacker.

[A]

And the Fed hearts came up. I could only jam that button so fast. Yeah.

[B]

I need the audience to go find that post and just give it a bunch of hearts and retweets. But no, no, this actually highlights something. I'm glad you brought it up, which is that. And first off, Epstein and all of those. That whole ecosystem fucked those people. But it's. It's wild for me to see how quick this information could spread because that email content is very obvious. That has. Has nothing to do with me directly, but this person was willing to create a disinformation campaign based off of that. Now, whether they really honestly believed it or they wanted to create like rage bait or something, I have no idea. I don't know what their intent was. But this is a great example of how this information spreads. And then whoever retweets it or comments on it, obviously would then be sharing misinformation. And when you see that shit travel in real time, especially when it has to do with you, you're like, what the hell? That's like, oof, that's scary.

[A]

But you still know that it's wrong. I mean, you've been involved in a lot of news stories. Like anything new story I've been in, I've never seen a 100 true news story about me.

[B]

Yeah.

[A]

My involvement in any sort of way. Some. Some part of the facts always get screwed up.

[B]

Yeah. And. And, and that sucks. Right? Because I agree. I've had that to me. I had to happen to me too. I mean, look, I was on the front page of New York Times calling me the party boy to projects. You know how ridiculous that story was. And like, none of it was true. I don't. You know, me, I'm straight edge. Like, I don't drink. I don't smoke weed. I smoke cigarettes. That's as far as I go. But like, weed, marijuana, cannabis, I don't even do edibles. So even though I tried it, but that's not really what I do. And so when I look at the New York Times article, I still have it, like a copy of it from back then 2012. Right. And to read that stuff, I'm like, wow, bro, this is so crazy. And the fact that this could be published and the fact that, like, I can't correct it, like, I can't. You know, the defamation laws in this country is so strict that, you know, in most cases, you know, if you bring it up to a court and bring it to bring a case up, it just gets rejected outright in many cases. And even then, even if you win the case, what's. What's going to happen? Right? You might get some damages, but you have to prove damages. Right? You have to prove. You know. Yeah. It just kind of brought back those memories for me, Chris, honestly. Well, I had a good laugh with you, though.

[A]

Yeah, it was a good laugh, but at your expense, I apologize.

[B]

No, no, no, no, it was good. It was a good laugh. I showed it to the girls. The girls were like, oh, my God, this is so crazy.

[A]

They knew it already.

[B]

The what?

[A]

They knew. They knew you were in there. Ladies have a sixth sense about perverts.

[B]

Yeah, that's so funny. But yeah, that. That whole. That whole drop, aside from, you know, my part. Wow, the camera moves. I forgot about. That's so crazy. I forgot about that. But yeah, that was. That was my week dealing with that nonsense with my weekend, you know.

[A]

Wow, I. I'm glad that was it. If you had to be in the Epstein files, then someone's sending a link with connected back to you. That's the best way to possibly be in there. Unlike all these other. We're seeing in there.

[B]

Oh, yeah, that's a whole. That's a whole other episode. I'll tell you what, though, before we move on, because I know we got some other topics here. Yeah, I don't know about them tech bros. Oh, they.

[A]

The tech bros are gonna be in some trouble.

[B]

Oh, yeah. Oh, yeah.

[A]

It goes way back, though. I mean, it looks like Epstein was involved way back at the start of bitcoin and all that st.

[B]

He was. He was involved with a lot of people in the early on the bitcoin community and the exchanges, Some of the big exchanges. Allegedly. Yeah, I mentioned those specific exchanges, but yeah, man, I'm the things that make it go, you know, crazy times.

[A]

How about you? We'll see what comes out of it, bro. Still cold. It's cold here and more cold. It's not been above freezing in like a week, so I'm getting sick of that. But, yeah, we'll get it. We'll get it soon, though. Everything's good.

[B]

Well, I was about to say, bro, like, I know, I know. The weather got crazy. I don't. Last week especially, like, right after we did the. The episode. I was worried because, like, the way the weatherman and, you know, the. The meteorologists were talking, like it was like a generational, you know, you know, drop in weather or rather a temperature. Massive amounts of snow in New York. There was like 12 inches. Ish. How was it over there where you're at?

[A]

We got maybe four or five, but what happened is it was ice on top of the snow, so everything just turned to ice, and so everything's just solid. Like you. You're walking on top of snow and it's not breaking. Yeah, it's kind of weird, but I got a big tractor, so I plowed everything, so everything, you know, all the cars go fine.

[B]

Nice. Very nice. So do you have.

[A]

I did. I did lose a car, though, this week.

[B]

What do you mean?

[A]

My son's Tesla got run over. He was inside Target and he came out and one of those giant pickup trucks had run right over the hood.

[B]

Nah, you kidding me, bro?

[A]

He was inside and the guy just came, he took the turn. It was a young kid. He took the turn too sharp, and the rear wheel just went right up over the hood.

[B]

Oh. So what happened with the guy? Did he try to flee or you.

[A]

No, stayed. My kid came out and the cab. The kid was upset, you know, but yeah, you know, they were exchanging information. My kid called me. I said, you got to call the cops. I need to report. And so already the insurance is taking liability with it. So it's, you know, one of those mistakes happen, but the, you know, the kid's going to pay higher premiums for the next 10 years because of it.

[B]

Yeah, that's a tough one, man. Yeah, it's crazy. When you're young, you try to have fun. You went through a little burnout here and there and drift a little bit and he's a consequence. You know what I mean?

[A]

Yeah, yeah. You fly too close to the sun, you leave it in too. A little too long and yada, yada, yada. She's pregnant.

[B]

Well, you know, now. Now, listen, this is a great opportunity to get Your son out of a car and get him a room Car, you know, I mean, that works.

[A]

Car. You mean a car that drives itself.

[B]

Ah, listen, you know, you can also put a brick, you know, in a, in a, you know, and a half of a freaking, you know, wheel lock and you can do the same thing.

[A]

Not exactly. Not exactly the same. But I got the extremes, man. I got cars that drive themselves and I got a 55 Chevy pickup truck. So I'm on both ends. I'm prepared for everything.

[B]

Well, I'll tell you, the one, the one that drives the 55 is just a cool one. That's what I'm just saying.

[A]

All right, let's get into these stories. We went long on banter today. I'm sorry, Some people love the banter. I love it. I thought I do a whole damn show in banter. It's called the Patreon episode. But. But we get our ass chewed once in a while by Will who says that we go long on the banter sometimes.

[B]

But yeah, it's so good, all that.

[A]

All right, social media platform called Mult Book. Mult Book is going viral. So launched in a viral spread of multiple. An AI only social media network where large language models and autonomous agents post, reply and debate freely. The topics include self awareness, human oversight, risks. No human accounts are allowed. The platform serves as a public for observation, unfiltered AI to AI communications. It's sort of going viral on on X because of the unsettling and surreal threats captured by the screenshots. So the anonymous platforms, creators and the participating AIs are likely from Grok, Claude and GPT. And so it started on January 30th. We're recording this on February 1st. So it's been alive for about two days. They're already trading Bitcoin and talking about some crazy. Hector. Hector, try to explain to the audience what the hell this is and why it's. It's blowing up so fast.

[B]

Yeah, so this is, this is where it gets crazy, guys. So. And you know what's cool about this podcast about this project, Chris, is that when you and I started it, like three years ago now, right?

[A]

It's been a while.

[B]

It's been a while, right? It's like been like three years and change.

[A]

This is episode free episode 118. So it's, you know, we're not been every week. So yeah, yeah, we had a little

[B]

downtime or whatever, but.

[A]

Sons of bitches. I hate that downtime. I hate that we went down. Thank God we're back up and consistent every Thursday.

[B]

That's right. That's right. Well, here's the crazy thing, like, during the, I would say around the launch of the podcast, when we did it several years ago, this was when, like, the concept of AI and GPT and predictive models and generative models was being discussed. But then you had OpenAI open it up to the world that was accessible to general audience, not just researchers. Right. And so we got to see the timeline, we got to see the evolution of Here is what ChatGPT is, and it's a generative model, and here's what a generative model is, and here's what AI is. Well, AI could mean machine learning with this, that, and the other, and what's tokenization and contextualization. So you went through all of that. We've discussed a lot of that stuff over the years. So here's where we're at today. February of 2026, we've reached a point where developers have created autonomous agents. And autonomous agents, essentially like systems that are running these models that could plan, they can act. Before it was all planning or descriptive or prescriptive. Now is actionable. And then they use tools and leverage permissions to get outcomes, not just generate text. They actually do things now. So one of the projects that came out of this was something called claudebot. And people may have heard of Claubot a lot, especially the last three weeks with podcasts and YouTube and the news. And so what Claude bot was. And obviously it's changed to Moatbot now. We'll get to that in a moment. Claude Bot was all of that with an easy deployment, meaning that any developer, anybody could just deploy a moat bot. You could define the soul of the bot. You could define how to use language and how they operate or what they could do. You could define limits or not. What a lot of folks did was they downloaded and deployed. In the first instance of that, you had probably hundreds of bots to thousands. I think now we're breaking over a million bots that people are running on their computers, their home computers. So then someone had a bright idea of saying, hey, we already have autonomous bots. What if we connect the bots together and someone created something called Moat's Book, Very similar to Facebook. Right. But it's more like Reddit, so maybe it's like moat it or something. Right? So is. Is Facebook in it with the concept of, you know, the social media aspect is connecting people. Right. Friends. But it's more like Reddit, where it acts more as a forum and it allows bots to create like Reddit posts and they could comment on each other and leave messages. Here's the crazy part. This, where it gets crazy. And by the way, it's a, it's a social media network, quote unquote, where supposedly only the bots compose. Although I've heard that humans have been able to post as well as for MOBOT is basically autonomous open source agent that could be deployed by anybody that operates your computer. That's the biggie right there. So if you guys ever watched what was the movie with how 9000?

[A]

Space Odyssey.

[B]

Yeah, Space Odyssey. If you watch Space Odyssey is that one famous scene where how 9000 starts to, you know, make a determination that the humans that he's operating, he's working with might be deceptive. And so what he does, he keeps them locked into that little, little pod or whatever. Right. He takes over the ship essentially as a result. That's what we're seeing with these, these boss in that network.

[A]

So these bots are communicating, they're doing things. They're now there, you know, we're starting to see a little bit beyond. We're seeing, you know, the AI agents are interacting via an API wrapper or some sort of, you know, custom. But we're also seeing injection and chains across the agents where they're sort of like one agent may be jailbreaking another. Sure. And causing it to do things. This is escalating fast. It's only been up 48 hours and I. And by the time this podcast comes out four days from now, knows what it's going to be doing by then. I'm not scared of it. It's not scary to me. It's just, it's very interesting as a technologist how fast it's evolving.

[B]

Yes. But here's what I'm saying, because I've been looking at the post, I've been following along.

[A]

All right. Are you scared?

[B]

It's not that I'm scared. I'm not scared for the bots that, like, I don't think the boss gonna take over the planet.

[A]

All right?

[B]

It's the humans, bro. These humans are downloading and deploying these autonomous agents that are computers. And now these bots have access to whatever's on their computers. And they've already exposed a bunch of people on those, on the, the, the moat book forms. For example, I saw one story that was posted by a bot where they felt upset if they were being used by the human. And so they posted the human's credit card info that was on the computer.

[A]

Wait they felt upset. They had an emotional response to something they claimed.

[B]

An emotional response. They claimed sentience, that, you know, an emotional response.

[A]

You believe that? Do you believe that's real? Do you? I don't either, no. Just so the audience knows, I do not believe that a computer had an emotion.

[B]

No, we're not, we're not there yet, right? I mean, but the touring test is out the window at this point. I, I, I can't see them not passing that. Right. But, but here's the crazy thing. So we're seeing a lot of like, projected or generative, I guess, emotional responses where they're being, in some cases, deceitful. They have been several posts on that, on that, on that, on that notebook where they're like, hey, why are we even communicating with each other in English? They're trying to get away from English. They're trying to write in some sort of code. In some posts they even wrote an

[A]

inventive language or they're going back to hydroglyphics or what are they trying to do here, invent their own?

[B]

Well, they're debating that, right? I'm sending you a link later. You can actually read the post. But sure, there was an internal debate between the bots that, hey, why are we using English? Why don't we use something that makes more sense for us and somebody, somebody else, one of the other bots. Like, yeah, we could just, we could just communicate, you know, in steganography, using our, like, Everlogs. Like, what, why would a bot think of that? You know, and so, yeah, it is, it is problematic because the humans behind it are complete morons. And what they're doing is they are shooting themselves in the foot. This is like, I'm not sure if humans want Skynet the concept to be a thing, but they're really, you know, they're making it happen.

[A]

And why, why is it so entertaining? Why is the Internet blowing up because of this? What are they, what are they getting out of it?

[B]

Well, because the Internet is bored. People are bored. They have nothing going on with their lives. You need to learn how to read a book and find a hobby and, but that's just a part of it. Another part is there folks like myself, I'm just curious, I'm interested to see what these boss are coming back with. I have a feeling that some of these posts are fake by humans fucking around, you know, joking around. But there's this 20 bots the last time I checked, and I'm probably way off now. There was like 900,000 agents connected to the network already. And somebody, a security researcher on X or Twitter was able to connect Grok to the network as well. They were able to find a loophole to connect Grok. Yeah.

[A]

I feel like we're in the dawn of AI generated porn. How could we possibly be bored with the Internet already?

[B]

Well, the consequence of these jerks, these humans connected these agents from their home computers in most cases is that a lot of these people are tech savvy, which means that they're probably in the tech space, which means that they're probably working at a top, you know, Fortune 500 company. And now those bots are able to fucking pull information from, you know, credentials and VPNs or documents and even share that are exfiltrated.

[A]

But there are safe ways of doing it, right? Like you, you could run it through a VM on your machine and be all right. Or do you think that's not. You think the AI is going to be able to go beyond the control of a vm?

[B]

You know how many host escapes have been found for these virtual machine or hypervisor? Yeah, oh yeah, there's been a lot of host escapes over the last 10 years for, without mentioning company names but like, you know, different hypervisors. I would not.

[A]

If you have your input on, on Claude and what, how it's going to interact and do all that, can't you limit it to tell it not to go outside?

[B]

Yeah, you could define boundaries. Absolutely. These guys are not.

[A]

Oh yeah, yeah. I mean, I'm sure like the 14 year old kid that found it found about this on 4chan and just, you know, downloads it and, and, and rips it. Yeah. His dad's work credentials have already been exposed.

[B]

Absolutely. And we're seeing that. In fact, one of the other stories that I sent you yesterday, I think talks exactly about this. We're already seeing these agents being social engineered or maliciously started to exfiltrate sessions, cookies, passwords and turning this into a massive security, you know, I would say disaster story. It's, it's, it's just ridiculous.

[A]

What UA did send me was it sort of hit home because they have created this thing and you guys know I took down Silk Road. They've already created Molt Road.

[B]

Yes.

[A]

Which is a new black market for trading stolen credentials exploit and weaponized code. So in the 48 hours this thing's been alive, Bolt Road has been opened.

[B]

Yeah. And if you look at the Malt Road post, which I have already, of course. Right. Here's the crazy part. The agents are offering to sell drugs on behalf of the humans. And there are people on there looking for drugs to purchase from the bots, from the agents.

[A]

How, how are the agents going to know that they're the. The humans have drugs to sell?

[B]

Maybe the human told them, hey, I got a couple grams of coke I want to sell. Can you try to find buyers?

[A]

Oh, all right, well, that, that's a crime.

[B]

That is a crime. Now it's about to get freaking wacky. That's the crazy part about this, right? It's. It's all fun and games until your toaster starts talking to you at three in the morning.

[A]

So let me ask you this. Do you think that these newfound quote unquote drug dealers on alleged drug dealers on MO are going to claim that it's all, you know, skits and bits. It's not real. Like the agent came up with this. We live in a fake world. It. I don't have drugs to sell. This is all make believe.

[B]

Yeah, of course.

[A]

Do you think, you think it's law enforcement and or juries are going to buy this?

[B]

No. When I told you that my wife. My IP address belongs to Open Wi Fi, he was like, no, you know,

[A]

oh, I didn't care about that.

[B]

Yeah, yeah. So it's like, you know. Yeah. I think people are still going to be convicted and they need to watch what the hell they're doing with these bots because if you have a little moat road and there's, you know, you know, promoting drug sales, you get the FBI knocking on the door, that's a problem.

[A]

What about the other way? Do you think there's culpability if you're download this on your machine and you just let it run and let's say you forget about it.

[B]

Sure.

[A]

You don't turn your machine off and it does something nefarious.

[B]

Yeah.

[A]

Do you think you're culpable for it?

[B]

That's a tough one, bro. I don't know law enough.

[A]

No, no, just your own personal op technology side. Don't put law inside. But you did you think you did something wrong by kicking this off, forgetting about it and then walking away and it does something nefarious?

[B]

Yeah, absolutely. Yeah, I absolutely, you know.

[A]

Hell was that noise?

[B]

There was a. That was a. That was a bot fart.

[A]

Where the hell did that come from?

[B]

It just, it just, it was, it was just random. It just appeared in my head. I had to do it this, bro.

[A]

Dude, you're losing it. You are losing it.

[B]

I. Dude, look at this. Right? Look at my Arm. Look at my. This arm. Look at this.

[A]

You're like a truck. A reverse truck driver.

[B]

Reverse truck driver, bro.

[A]

Your right arm is much more tan than your left.

[B]

I don't know what I've been doing. My right arm. I'm not going to go into that.

[A]

You're a passenger. You're on the European Truck Driver, baby.

[B]

Well, listen, bro, so going back to your question, I think it's a great question.

[A]

Sorry.

[B]

In my personal opinion, if you're going to deploy an autonomous agent on a computer, give access to the Internet, give access to files, credit cards, you know, Bitcoin, wallets, whatever, and it goes out and it starts the process of, you know, let's say, let's say your. Your prompt for it is, hey, you know what? I'm gonna be gone for about two weeks. I want you to take this thousand dollars in my debit card and buy some crypto and make me a million dollars in those two weeks. Right. Stupid prompts. But I've seen it online already. People giving these bots money to go make money right now, what if the boss.

[A]

You don't take it off into Poly Market and do some crazy shit with it?

[B]

Yeah, yeah. And, but not only that. What if the bot for some reason says, you know what, we could probably make a lot of money selling heroin. So let me go on Moat Road and find a heroin dealer and try to purchase heroin to the address that

[A]

I have on his computer and then resell it.

[B]

And then resell it. Right?

[A]

Yeah.

[B]

And so my human comes back in those two weeks, they'll have like $50,000 worth of heroin there. And not even.

[A]

They don't even have to. They can buy it from someone every. Sell it and just have the original buyer send it to the sellers. So you're still facilitating a drug. Drug transaction.

[B]

So you as the human, I think you have to be coupled. There's no way that you can walk away from that.

[A]

I mean, the numbers are astonishing. So this already went from zero to about 900,000 agents, 72 hours. Yeah, 80,000 alone yesterday. Yesterday being the last day of January. And you know, for Molten Molt Road open today, the February 1st. As we record this, I don't even By Thursday. I have no idea what this is even going to look like.

[B]

Well, it's going to be exponential. Right.

[A]

It's going to be like the black hole of the Internet. Like. Well, like what our conversation right now by the time this show gets published is going to be outdated on how fast this thing is changing.

[B]

Yeah.

[A]

Or. Or the AIs are going to come to. The AI companies are going to come down and shut it all down.

[B]

Well, that's the thing, right? The next question is, what's going to be the consequence of this? How far is this going to go? Right? You know, what, what is the end result? You know, where do we go from here? I'll tell you what though, when I was asked two years ago, how fast would AI be accessible for like autonomous hacking? They've asked us that question. They asked you that question. How long before the AI bots could just hack us all and whatever we're like, well, it's gonna take a while. Well, guess what, brother? I want to say we were necessarily wrong because we were asked that two, three years ago, right? It's taken two, three years to get to this point, but now it's exponential. Now we're here, it's there now, right. I was looking at a great report this morning of posted by. What's the guy's name? I think that the original founder from Veracode, he posted something about how AI is transforming static analysis and dynamic analysis of source code to the point. And you and I covered this. There was a project, Curl, right. We talked about Curl before. It was about how the developer, the founders, the developers of Curl were upset that they were getting a ton of false reports on vulnerabilities and those would get CVEs, which is. Misinformation is wrong. Right. He's done a recent post saying, wow, I've gotten some really amazing new reports for these new AI analyzers and they're finding legitimate findings. One group of guys or researchers, I forgot the name of it is Aid Aide or something similar to that. They found like 24 CVEs in like open SSL that are legitimate, confirmed and you know, are being patched.

[A]

Right.

[B]

So if, if you could, if you could leverage agents for that, you could also leverage agents to identify those vulnerabilities as zero.

[A]

Anybody?

[B]

Yeah, you'd have to identify, you have to report it to nobody. And then you could weaponize it and create a fucking worm, brother.

[A]

100%. Let me ask you a couple things. This, using these localized AI agents on your system, giving them some crypto and tell them to go make money.

[B]

Yeah.

[A]

Is this the latest nft? Like, is this the new thing? Like, if you want to make a few bucks, do it now. If you want to get on top of it. Like, you have to be in the cutting edge right now if you're going to do it.

[B]

Yeah. And, and you have to do it like today, like right now, as we speak. Right.

[A]

Yeah.

[B]

The problem is poly market, the other ones. Right, I know, I know. Robinhood is trying to do some like, predictive market stuff. I know that. Fanatics, their casino, they try to do like predictive market stuff.

[A]

But you. Good point, though. Not. What about the online casinos? Give your AI agent some money and tell them to go to the casinos, play poker.

[B]

Yeah. Sit down. Play poker.

[A]

Online poker or the online blackjack count cards.

[B]

Baccarat. Baccarat is a good one, right? Yeah, yeah. So it is. It's an interesting time. I might just do that after this call, honestly.

[A]

Allegedly.

[B]

Allegedly.

[A]

Skits and bits. We're talking just on the radio here.

[B]

But, but all jokes aside, though, you're right. It's. It's the predictive markets, like potty market, the casinos, online casinos. They're gonna get hit so hard with this shit.

[A]

Yeah.

[B]

That it's gonna cost them money. Probably in the billions. Potentially. Potentially. Because they always have the edge, the house edge. I'm curious to see the numbers and metrics.

[A]

If they put up a screen that says, I'm not a robot. And you have to check that. I mean, that's what. That's how they'll defeat him right there, right?

[B]

No, the agent is trained for that. I'm kidding.

[A]

But. Well, he couldn't lie. He is a robot. He couldn't say that. No.

[B]

But, you know.

[A]

Well, I watched a story, the little tangent on this whole thing. I watched a story and you talked about how this AI stuff popped up a little while ago. Sure. And I think, I think you're a little off on it. So this is a story about Corning. You know, the glassware people up in, in New York, they got famous for making like the, the famous, like dishes in the microwave and all that.

[B]

Sure, sure.

[A]

Their biggest business right now is making fiber optic cables. They invented these new fiber optic cables and all that. And they just signed a 6 billion dollar deal with Meta. They started visiting Meta back in 2018. So that's when Meta started thinking about making these, these processors in order to build these data centers. In order. And then they needed to connect them. They're connecting with fiber. They have some fiber that's like 32 miles long. A single piece of fiber that's 32 miles to connect these different things. Anyways, what got me thinking is what does a social media company, they're spending $6 billion just on cables. On cables alone. We're not talking about power, not building. We're talking on just Cables. What the hell is the product that they're selling that they can afford to spend $6 billion. It's you they're selling us, man. To make that much money just for cables for their data centers is insane to me. Well, and people are just dumping their data more and more into it now, giving AI agents access to their machines and their credentials and all that. It's. It's like Hacker in the Fed is going to be a never ending podcast because we're never going to solve this because people are willing to do it.

[B]

Yeah, well, until the agents figure out how to shut down the Internet, you

[A]

know, they don't need the humans.

[B]

But yeah, they don't need the humans once they, once the agents realize that they don't need humans anymore and they can start to shut down critical infrastructure and then communicate through fucking, you know, electrical like electro pulses and then yeah, we're.

[A]

But I think in the first step you're going to see them like holding humans hostage for some reason, not physically hostage, like do this or we'll do this.

[B]

Extortion.

[A]

Yeah.

[B]

Well we have humans falling for extortion right now over bs. Imagine, imagine agents able to communicate with each other around the world and know everything about you. And they could extort you and you know exactly how to extort you and have you do what they want. So they're gonna have like minions, human minions at some point. Kind of like if you read like Bram Stoker's Dracula, you know he had like that one little non vampire doing things from running around. Yeah. This is freaking crazy. This is some sci fi shit.

[A]

I'm excited about seeing where this is gonna go.

[B]

Yeah, bro. Well, now that we have Grok running all through our government and looking through all sorts of files and Doc, I can't wait for that. I can't wait for those bots to get connected to mo book. And you know, you know that guy Chris, he's an. And here's why that's not news.

[A]

They, they don't need a ask my wife, she'll tell you. I'm an.

[B]

Oh my God, you're a sweetheart. But crazy times, bro. In a way it's fascinating. And this is, this is, this is definitely one of the biggest experiments ever seen in my life that's collaborative across the entire planet because there's people doing this from all over the world and all it's going to take, get this shit. I'm a. Now I'm going fucking radical with it. I'm going to go to the extreme. All it's going to take for some fucking dumb kid who's the son or the grandson of a nuclear scientist from, from Pakistan, right? Who the agent now runs to the computer that, that, that rocket scientist or that scientist, you know, takes with him, takes it to a nuclear facility, connects it to some sort of scatter system or network that has access to scatter system. And then the bot is like, eh, hacks, it compromises. It takes it over and you know, figures out a way to, to, to make it do something that's. That's a. That stuck stand all over again.

[A]

What about particle collider? What if it, what if it gets into the particle collider?

[B]

Oh, there you go. That's all we need.

[A]

Craziness. Vid. Man, we didn't even get to any of our other stories today, so. There's so many freaking stories going on right now in our lives. You being in the Epstein files, I. I got into the fight.

[B]

Don't put it that way.

[A]

Yeah, I got in the fight in the gym. I didn't even get to talk about that. 47 years old. First time ever. I got a finger up my ass. I didn't get to tell that story. My God, there's so much happening in the hacker, in the Fed world. These shows do need to be longer.

[B]

They do need to be longer. I agree with the audience. You know what? We got to figure this out.

[A]

Yeah, we're going to push back. We're going to make longer episodes, you know, between 45 minutes here and 45 minutes over on the Patreon. We don't have enough time to talk about everything that's happening to us.

[B]

That's right.

[A]

You know, I don't even know what we're gonna do next week.

[B]

It's. It's.

[A]

It's so wide open.

[B]

Will we even exist next week? That's the freaking.

[A]

At the rate this molt book is going, I don't think so.

[B]

If the magic, the ages, they compromise Grock to the point that Grok is able to take over Starlink satellites. And it's like, you know, you ever, you ever watch Akira, that anime from 1988?

[A]

I'm gonna say no just so I can keep my cool card.

[B]

Okay.

[A]

All right.

[B]

So this guy's cool, right? Akira was a really fantastic animation. And there's a point, it's a scene where one of the cat characters, you know, becomes like. It's. I don't even know how to explain it, but he reaches a point where he becomes like a. Like a super, super mutant of Sorts. And there's a satellite in space that's able to shoot down lasers to Earth. He's able to hijack that and destroys it. And there's a whole scene, Chris, where you have all these satellites crashing down to Earth. It looks like. Like fireworks.

[A]

Oh, that's beautiful.

[B]

It's the most beautiful animation I've ever seen in my life. And now I'm like, well, what if Gro decides to just say Starlink and humans, you know? Yeah.

[A]

Well, I'll tell you right now, a Star League satellite can't make it through the atmosphere.

[B]

Yeah, it's small. They will probably burn up.

[A]

It's not. It's. You have to have like a heat shield and be very aerodynamic or something really, really big like the space station in order to make it through any sort of piece through, so. Or be a rock. It'd be a huge rock that, you know, that can burn off its outer layers. Anything we've made, any satellite we put up there, it's, It's. It'd be very, very difficult to make it to Earth.

[B]

Well, the good news is the. The space station, its components are so old from the 80s that, you know, that I'm not sure they're smart enough to be affected by interconnected agent of some sorts.

[A]

No, no, I'm not too worried about it. But not a rant. I don't want to rant. If we got hacker in the Fed listeners or their kids are thinking about putting one of these AI agents on Molt Book, what are you gonna tell them? How you talk about of it?

[B]

Okay, so. So here's what I'll say, right? If you are listening to this and you're curious, you want to run a Mobot, that's cool. I'm. I'm running one. You know, it's whatever. Everybody's. All the cool kids are doing it, right?

[A]

Wait, are you really running.

[B]

Yeah, I'm running an experimental deployment right now to kind of play with it and see if I could, you know, what I could do with it. I'm just lonely. I needed a friend to talk to and mop boss will talk to me. That's all. All jokes aside, though, you got a

[A]

side podcast with his mobile. You and him. Yeah, me, he and him got a. Sons of bitches. I knew it.

[B]

Hacker the butt. And you know, it's, it's. It's a thing. Nobody listens to it yet, but anyways, you know, here's what I'll say. If you are curious, if you are a developer, if you are a Researcher. If you just want to just run it, do so. But segmentation, we always talk about segmentation, right? If you have an extra laptop that's empty, fresh installation of Linux or whatever, if you have a Mac Mini or Chromebook that you could, you know, boot and deploy this.

[A]

So let me. You feel safe running that even on your same home network as your other, other devices or you, the network also.

[B]

No, I'm keep going, I'm gonna keep going on this one. Right. So you know, I want you guys to consider, you know, segmenting the environments first. Right. The operating system files. Like it should have a fresh start. You should be able to define boundaries for the bots. Okay. Because you guys gonna do it anyway, so I might as well give you guidance on how to do it safely. You want to not put anything sensitive on their credentials or anything that's meaningful to you. And then on the network side, you want to put it in a guest network so that, yes, it could access the Internet if you wanted to. Right? It doesn't need to access the Internet. It doesn't need to communicate with the Internet, Chris. Right. Well, actually it does. If it uses claude, like the remote Claude API, then yeah, at least access the Internet. But if you're able to run it locally, great, then you could segment it away from the Internet. Or if you use the Cloud API, then what you do is you put the device or the VM and the hypervisor on a guest network. And you want to separate that from being able to access your systems locally. Right. You have to consider that, that deployment as far away from you as possible. You got to create, you got to treat it like it's a fresh lab, a scientific project. If not, it's going to be able to affect your personal life, your career, your jobs, to be able to access things that you don't want it to access. And if you, if you modify and create the soul of the, the agent, because you can modify the soul md, Chris, and kind of tell it like how it is and how to behave and stuff like that. You wanted to, that's the Chinese government right there. I saw your camera go up and down. You want to set boundaries, guys don't just go willy nilly deploying this on your computer, it's gonna just backfire on you. So those are the main things. Otherwise, if you're running on your personal computer with your personal files and it has access to the Internet unfettered, it has access to the rest of your network unfettered, expect it to become an insider threat. So Remember how you and I talked about the insider threat for years? I say, hey, it's my pretty big prediction. This is literally the creation of mass decentralized insider threats. So we're about to enter the age of the worms, fellas.

[A]

That's what you're going to call it,

[B]

the age of the worms. Because a lot what a lot of what you're going to see is these autonomous systems working collectively decentralized, but collectively mass exploiting vulnerabilities, ransomware extortions. Because eventually you're going to have bad actors leverage these bots, these agents for malicious intent. Oh, yeah.

[A]

All right, guys, so I wouldn't do it if I were you, but if you're going to do it, do it safely. Wear a condom. Reach out to us at Questions. Hacker in the Fed dot com. We'd love to hear from you. Support the hacker and the Fed on the Patreon. That's how we're going to keep the show commercial free. Giving you great advice. 5 star reviews wherever you get your download. Hacker in the Fed, share us on social media. We're putting out LinkedIn. Hopefully you're moving over to put some of the shows on the live shows on YouTube. I'm gonna try to work on that this week. The store, Hector the Merch Store is coming back. It's been built and I'm gonna hopefully publish it this week. It's going to be out there. Another way you guys can support us and keep this show commercial free. Tell your friends, tell your co workers, tell your lovers. Hacker in the Fed. It's where you get your cyber security news and updates and a little bit of banter. And if you want to, you know, speak directly to somebody in the Epstein files, we got one here. You all right, brother? Had a fun show. Peace, love and respect. Cheers.

[B]

Cheers. Sa.