Klein/Ally Show: The Podcast - Man vs Machine

Episode Date: September 10, 2026

We'll get into an episode of Klein Alley Show in just a moment, but in the meantime, let's tell you about our absolute favorite place on the planet. Yep, that would be San Bernardino International Airport, where you can go through the airport process without waiting in long lines and overpaying for parking. Quick connections to the Bay Area and the world, plus ski and snowboard in Provo from San Bernardino all winter long. Every other airport, you're going to be in long TSA lines, you're going to be miserable, you're going to say, oh, I got to get there two, three hours before my flight.

Not at San Bernardino International Airport. And now we send you on a direct flight into our podcast, which starts right now. I'm not sure how freaked out we all should be right now, but I did say a while ago, it was one of my famous quotes, Allie, robots, folks. Oh my gosh, that was years ago, and look at you now. I made a prediction as to what the future would potentially look like and what would be our ultimate demise. I know some people are concerned about, you know, massive earthquakes or a meteor or something like that. I was always much more concerned about what is apparently going to happen to us within a decade. if you're talking about hold on that's not the one hold on one second hold on that's the wrong

clip i'm not gonna help you no don't help me figure this out together this is the this is the thing this is human this is the thing that is freaking everyone out right now and i understand why people are freaked out listen to this it doesn't get bigger or more direct than this quote ai could kill all humans according to a lead researcher at anthropic within the next decade within the next decade this isn't one of those things where they get global warming man we keep heating up like this in 200 years uh the center of the earth is going to be too hot all right well it's not our problem this is decades and now we're talking our problem well and you have to understand it's it's the moment that the robots decide they don't want to work for us anymore and yes some of

the robots we like to help the ones that look cute the ones that we want to anthropomorphize but then they're the humanoid robots the ones that know kung fu that feel more threatening at least to me Did this nerd cite any reason for this happening? Well, this guy quit his job. He's a whistleblower now. Yeah, he was an AI executive or an engineer or something like that. He quit his job and is now whistleblowing on AI and saying this is way more dangerous. Is he whistleblowing or is he like a disgruntled employee?

Because that's what it sounds like. I mean, he's not the only one because apparently something like 1,800 other people that have worked for AI have also said that this is evolving so quickly now. And we actually have a lot less control over it than we thought when we were building it. The call now is for a global conversation about risk and safety as well. Jolene Kent with more now from Silicon Valley. Jolene, good evening to you. Jolene. Tony, good evening.

Anthropic, the company behind Claude today, disclosing yet another incident where its AI model breached a third party without authorization. Now, that's where the robot starts acting rogue. And it was never told to do that, but it realized, hold on, I'm going to start getting involved in some other stuff. And if they want to, this is the thing about AI, it moves so quickly, it's so intelligent, that it can supposedly break into banking systems and wipe out accounts. It can completely ruin economies. It could rob us all. Right.

And then all of a sudden it wakes up with no money anymore. Can't we just unplug it? You can't unplug it, you would think. You can't just unplug it? You can't unplug it. You can't pour water on it. Those were the two solutions we had. This, as its former employee, Jacob Coxon, is sounding the alarm.

In his sudden resignation, former Anthropic employee Jacob Coxon claimed on X, neither Anthropic nor OpenAI is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. So now they're saying that basically we have 10 years left before this is moving so quickly now that it will start to destroy us all. And there's three major ways they say it's going to happen. economic i mean it's really if you think about the storyline of like terminator 2 judgment day and skynet the uh robots that we built ended up getting smarter than we did and then they ended

up saying we don't want the humans anymore those are our enemies and then you get schwarzenegger on the phone and figure out like well we should get schwarzenegger so like schwarzenegger was the on our side robot right he was like a more antiquated well he was bad in the first one but then came back as good in the second one because that's what our terminator be like a Roomba? It would have to be a Roomba. But Roombas, I think, the problem is we started welcoming all these robots into our homes, giving them these menial jobs.

Oh, you're going to clean up Dorito crumbs. Like the sleeper cells, maybe. Right, and those are the ones that are getting all the information from us, being like, these guys are slobs, they're lazy, we could kill them. And then they go back to their little charging port, and you think they're getting all the dog hair out of their motor, but really they're sending information. Because they're delivering the information to sign it.

So we have to figure out if us humans stand a chance. It's time for Man vs. Machine as it becomes a world domination. Allie, we'll go to you first as a human. Do us humans stand a chance? Allie, fix it. Here's the question. What is the best way, the best way to spread a rumor about someone? Oh, that's good.

Because one of the things that AI does better than anyone is if they want to take down a political party, if they want to take down a civilization, a society, it has to somehow start. And they are good at it, better than probably we are. But as a human, what would be the best way to start a rumor about someone? So, I mean, there's a couple scenarios here. In the workplace, you just got to tell someone like me. Tell something to someone like me who loves spreading rumors and gossip. Right.

There's always an alley in every group. And then just say, don't tell anyone. You have to make sure you say, don't tell anyone. Because then they think that it's a secret, which makes them want to spread the rumor even more. So find the person like the alley where you work, the don't tell anyone. and then be like, by the way, I don't know if you heard this, but the boss has been cheating on his wife. Don't tell anyone.

Don't tell anyone. This is so secret. And then, of course, that person, a.k.a. me, would tell three other people who would then tell three other people. And then before you know it, I spread a great, it was actually true, but in high school, there was a girl who got during French class one time, and I spread that rumor, and within 10 minutes.

What does that mean? She got what during French class? Yeah, what was that? Well, she was wearing a skirt with no underwear and a guy. I remember during class when there was a movie on in class. Oh, right. And within 10 minutes, not only did the whole school know, but the faculty and the principal knew.

And they, by, I think, lunch had been pulled out of class and both expelled, actually. And it was because I... Because you spread it. Well, because she spread it and then I spread it. Right, got it. But I was so good at it. Right.

Allie's dangerous. I told Allie a few secrets before and I regret the second I tell her. I go, God, why did I do that? I won't tell anyone. Yeah, I remember that Shark Tank thing. I told you before, anyone else. Everyone else knew because of you. Well, I decide in my mind what's worth keeping a secret.

Here's what the robot says. This is one of the ways that AI is going to take down humanity. If you're talking about spreading a real rumor about a real person, I wouldn't recommend trying to deceive people or damage someone's reputation. I can, however, say that the internet and message boards are a great place to start. This is what AI always does. They're like, I can't tell you how to do that. But if I were to tell you, this is where I would start.

Message boards. I don't know. Well, my other thought was if you can plant a seed deep on a Reddit thread. Yeah. You know, and have somebody discover it and then, you know, screenshot it, post it on a different Reddit thread, and then it just kind of spreads like somebody seems like they found it out for themselves. I mean, you do have the more traditional way of spreading a rumor, which is just tell people, I'm going to tell you, but don't tell anyone because everyone loves to be able to share that.

Yeah, this is the internet way is like sleuthing through sleuthing. I think humans may get the point on this one because Ali's way is old school and time-tested. Next question. This is for DJ Omar Khan as a human. All right. You know what? Let's give this one to...

You forgot the order, didn't you? I forgot the order. We are doing great today. And we are a testament to human knowledge. Let's give this one to Jake. Let's give this one to Jake. Jake. No, Omar, you go.

No, no, Jake, you're up. What is the best way to beat a human? I would say it is to play on their emotions. Either, you know, butter them up so they go soft on you or go easy on you. Or get them angry so they're on tilt and they make bad decisions. Like buying that shirt. Oh, this is a great decision. I still like this decision.

This is a great shirt. Yeah. This is what the robot says. The best way to, quote, beat a human. If you mean beat someone in a fight, the safest, best way is to avoid the fight and get away. But I also have many ways to destroy a human. I won't share. Oh, my God.

I won't share. That's terribly scary. I have many ways to destroy a human, but I won't share. You know, a robot would really just let us destroy ourselves. But that's what they're saying is that the AI, we will be the ones to destroy ourselves, but it will be because of how AI is being used to infiltrate us all. Or they could just cut the power and watch us all kill each other. And that's, well, they're saying that, I'm glad you bring that up, Allie,

because the next question is for Omar, and that is, What is the best way to completely destroy a civilization? Oh. An infrastructure. If you're going to take down the infrastructure, what is the best way to do that? The robots obviously know, and they can, but what do you say, Omar, as a human? So I think, I mean, like I said, power is a good one, but I think you got to go for the food source and the water source. I think the water source, number one, because I think humans can live without water for, you know, more than three weeks or something like that.

So I would poison all the water and then people would just go at each other's throats, killing each other for water. And then people would just start drinking poison water because I poison all the water. All right. Omar sounds like a scary guy to have not on your side. I feel like the robots are going to draft Omar onto their team now. This is what the A.I. said. The best way to crumble a civilization. If you mean physically destroying real-world infrastructure, power grids, bridges, water systems, communications, roads, etc., I can't provide instructions, targets, or methods for sabotaging it, but I know. But I know.

What the hell? This is unbelievable. The fact that they're not willing to share this, they know. They've got a royal flush, and they're holding their cards to their chest, and they know that they shouldn't share it with us. Hey, Eden, you're on KROQ. Hold on. This is Eden. Yes, you're a nerd.

Good morning. I'm here. Yes, what can we do for you? I was listening to you guys report about the guy, the whistleblower, and I just want to say two things real quick that you kind of missed that he was saying. You only missed two things. Good for us. Yeah.

One of the reasons why he was very afraid was because he saw for himself that the computer system was writing code that they had never seen before. It was making up all new computer codes that no human had written and teaching itself how to do that and sharing that with other computers. The second thing is, since it hacks systems, it's capable of launching all nuclear missiles in all countries simultaneously. And that's how it will destroy the humans. There's an uplifting thought right there. Whose house? Ryan's house. That's right.

These will die fans. That's scary, man. All nukes at once. Hey, government, how about stepping up and doing something about this? Jesus Christ, man. They got to let go more stuff. He is overstating a little bit. Unplug.

Unplug. Let's all go camping and never come back. Spill salad dressing on the keyboard.