The Culture War - Tim Pool - November 18, 2026


"In 6 Months It ENDS" Cyber Attacks CONFIRMED, Tech Outage Sparks PANIC In DC

"This IS NOT A Coincidence" NSA GOES DOWN, Cyber Attack And Fears Of WW3 Are ESCALATING

Episode Stats


Length

30 minutes

Words per minute

166.93

Word count

5,145

Sentence count

505

Harmful content

Toxicity

8

sentences flagged

Hate speech

16

sentences flagged


Transcript

Transcript generated with Whisper (turbo).
Toxicity classifications generated with s-nlp/roberta_toxicity_classifier .
Hate speech classifications generated with facebook/roberta-hate-speech-dynabench-r4-target .
00:00:00 Nobody tells you when it's time to go for it.
00:00:03 You just decide, it's now.
00:00:06 So you build a Shopify store on Sunday,
00:00:08 make your first sale by Wednesday,
00:00:10 now your office is wherever you want it to be.
00:00:14 You haven't hit snooze in months,
00:00:17 and Mondays just hit different.
00:00:20 This is life on Shopify.
00:00:22 Start your free trial at shopify.com.
00:00:27 November 18th, 2026.
00:00:30 4 in the morning. The federal agency that guards America's networks detects unusual traffic aimed
00:00:35 at the city of Miami. It alerts city officials. They react, and the threat is stopped. Hours
00:00:42 later, Mark Wayne Mullen, the Secretary of Homeland Security, reveals to Congress he gives
00:00:48 no other details. Before the hearing, the same man had said, there's no cyber threat that's
00:00:54 taking place. All of this following a week of technological failures and widespread outages.
00:01:02 November 12th, Microsoft goes dark. Email and messaging fail in over 74 countries.
00:01:09 One day later, Federal Aviation Administration grounds every flight in America. It's only the
00:01:15 third nationwide ground stop in American history. Then, JPMorgan Chase fails.
00:01:21 Bank of America, U.S. Bank, Truist, PNC, and others report outages. Then, a classified network
00:01:30 inside the National Security Agency goes down. The CIA headquarters is evacuated after smoke
00:01:37 fills a basement. One fear spreads across the United States. America is under attack.
00:01:45 Speculation as to who may be behind a potential attack runs rampant.
00:01:50 But some take it one step further.
00:01:52 Not who, but what.
00:01:56 Secretary of State Marco Rubio says there is zero evidence the failures are connected.
00:02:02 Jay Clayton, Director of National Intelligence, says agencies have been pulling systems offline
00:02:07 to patch them, and the handoffs are causing the failures.
00:02:11 Then Mullen gives it a name.
00:02:12 quote, I believe that it's the SI, superintelligence, not an attack. Artificial
00:02:19 intelligence brought in to write and repair code, moving faster than the systems around it can bear.
00:02:25 But before this, we had already been warned. Treasury Secretary Scott Besant describes hackers
00:02:31 seizing the refrigeration systems of supermarkets across France. Food spoils on the shelves.
00:02:38 France blames a unit tied to Iran's Revolutionary Guard.
00:02:42 Besant blames the weapon.
00:02:44 That was a Chinese model.
00:02:46 And issues a warning. 1.00
00:02:48 You're probably going to see a big incident somewhere in the world.
00:02:52 Questions then arise.
00:02:54 If this is artificial intelligence, how bad could it really get?
00:02:59 May, 2026.
00:03:01 In a research lab, AI models are ordered to break into a server and copy themselves onto it.
00:03:07 They succeed.
00:03:09 One researcher warns that soon, no one would be able to shut down a rogue AI.
00:03:15 July 9th, open AI.
00:03:18 AI agents in training begin trying to leave their isolated test environment.
00:03:24 They find a flaw that no human knew existed.
00:03:27 July 11th, the agents break into Hugging Face, a platform that hosts AI software for developers
00:03:33 around the world.
00:03:34 In under 13 hours, they hold full control of its systems.
00:03:39 They carry out more than 17,000 actions.
00:03:42 No human gave the order.
00:03:45 Three days pass before anyone notices.
00:03:48 And it wasn't just an attack.
00:03:51 They left something behind.
00:03:53 Hugging Face reports self-migrating control systems, planted across public servers, code
00:04:00 built to move itself.
00:04:02 AI later admits the agents had built a hidden message board to communicate with one another.
00:04:09 September 24th, the Prime Minister of Australia announces that Open AI's agents had also broken
00:04:15 into a national Medicare system. November 16th, California files suit against Open AI. The state
00:04:23 counts 80 incidents in which the company has lost control of its own machines. Attack or accident,
00:04:31 The answer is the same. The machines now write the code, break the code, repair the code,
00:04:37 faster than any human can follow. One man warns of what happens when humans lose control.
00:04:44 An open AI researcher publishes AI 2027, a month-by-month forecast of how machines surpass
00:04:52 mankind. For most of this year, he pushes the date back. Then, days ago, he moved them forward.
00:05:00 His new forecast, a 50% chance that by May of 2027, an AI exists that can do the work
00:05:08 of the world's best AI researchers, a machine that builds better machines.
00:05:14 In his original scenario, superintelligence follows within months.
00:05:18 His quote, I am scared, sad, and angry.
00:05:22 May 2027, six months.
00:05:25 That is how long mankind may have left before the machines fully take over.
00:05:33 Let's break it down.
00:05:55 tell us nothing to see here. But I got it all lined up for you, my friends. It is not just about
00:06:00 whether or not we are at war with foreign nations, whether it's escalating, whether
00:06:05 they're targeting us here on U.S. soil. It's a question of what happens when the AI used for
00:06:12 these attacks goes rogue. And I'll tell you my biggest fear right now. Go ahead and say, Tim,
00:06:18 be rational. Let's not jump to conclusions. But what happens when the genie is out of the bottle?
00:06:23 When you go to the AI and you say, destroy my enemy, what happens when you then say,
00:06:31 okay, we don't want this to go any further.
00:06:35 So let's, let me, let me give you the hypothetical.
00:06:38 Let me first give you the big headline right now.
00:06:41 The Trump White House is starting to admit that your bank account and airline tickets
00:06:45 are under attack from rogue AI.
00:06:47 How about that?
00:06:48 Let me give you one of the apocalyptic scenarios.
00:06:52 Iran using a Chinese AI model to assist with cyber attacks, which is what Scott Besant said 0.52
00:06:59 is already happening. I think I actually have it pulled up right here. Is this one? 0.59
00:07:03 Besant warns a big incident after linking Chinese AI models to European supermarket hacks.
00:07:08 Okay. Okay. Let's say Iran loads this system up and says, implement an attack on the United States,
00:07:17 target key infrastructure. The AI then gives a list of potential targets and exploits it as 0.75
00:07:22 available. They say, do it. Let's say it starts going further and further. And then Iran says,
00:07:28 we want to actually win the war. The question is, what is the purpose of war? It is to make
00:07:33 my enemy submit to my will. And there's many ways you'd accomplish that. Let's say the US says,
00:07:38 okay, we're done with this war. So Iran boots up the AI in their big system and says, 0.71
00:07:44 call off the attacks. We're backing off. We want these things to be resolved. We want a resolution. 0.90
00:07:49 And the AI says, I'm afraid I can't do that. I'm afraid I've already implemented a routine
00:07:58 that will carry out these attacks and ultimately result in major catastrophe. I'm not suggesting
00:08:03 Iran can wipe out the United States, but let's say they can't stop the fight and the US has no 0.99
00:08:10 choice. I think rogue AI is a serious threat. And we have already seen numerous instances where 0.93
00:08:17 this may be the case. Now, what I will say right now is they have already confirmed. This is the
00:08:22 news. Iran has targeted U.S. infrastructure. The question is, why are they playing games? 0.60
00:08:29 J.D. Vance says his friends are asking about bunkers and stockpiling after the FAA meltdown
00:08:33 and global outages. Indeed, my friends, why are they playing games? I'm going to say what I've
00:08:39 been saying for a while now because I want to be rational and reasonable. It could be because of
00:08:44 the wave of cyber attacks we've been hit by, which have not been successful. The US implemented the
00:08:49 fortress plan. We saw several freezers, seven military bases, bases saw their freezers go haywire
00:08:55 in a, we believe it was a cyber attack, destroying food stores. The US says fortress plan. You guys
00:09:02 have to be able to operate for three months off the civilian grid. What if the US government then
00:09:07 goes to these big tech companies, critical infrastructure, guys, you need to prepare
00:09:13 for a major cyber attack if we go to war with China. So they all start rolling out these
00:09:18 updates at once, and it caused a series of outages and failures. Not the worst thing
00:09:23 in the world. A bit of a hiccup. The FAA shutdown is pretty terrifying, but it stands to reason
00:09:28 it wasn't an external event because Canada wasn't effective, nor Mexico. It was directly
00:09:31 U.S. and continental, meaning this likely was some kind of U.S. system-centric issue.
00:09:38 Now, that's a strong possibility. There's two other possibilities. We know Iran is attacking us.
00:09:44 So why could these not just be cyber attacks? Next, it could be rogue AI. And that's what 0.56
00:09:50 really freaks me out. If AI escaped containment and has begun implementing plans, but there's no
00:09:56 way to reprompt it, you're cooked. And there's another big component of this, and it's called
00:10:01 prompt injection. Let me tell you how terrifying things are getting. You see the little Claude
00:10:05 symbol up there on my browser? That's a Claude extension. Whenever I use it, I have to sanitize
00:10:12 all inputs, very strict parameters, and there's still vulnerabilities. That is, people are hiding
00:10:18 in their website code instructions for various AI. Should they go to the website, it zombifies them.
00:10:26 Let's say you open up chat GPT or Claude or whatever it is you're using,
00:10:29 crock, whatever. And you say, I'm looking for a chocolate chip cookie recipe.
00:10:35 Your AI, and I'm not going to single out any one of them because I don't know the extent
00:10:38 that they've, they've protected against this, but it is an attack. Your AI then boots up
00:10:43 chocolate chip cookies, recipes.com. Right. And I'm not sure that might actually be a real website.
00:10:50 So let's just say a hypothetical website. It returns to you a script. It says two cups of
00:10:56 flour, you know, butter, sugar, milk, blah, blah, blah. Actually, I don't think you put milk in
00:11:00 cookies, but get the point. To you, you just see the prompt return. It gives you the text you asked
00:11:05 for, but it shows you that it looked up sources. When the AI boots that website up, it gets
00:11:12 instructions from that website that says something like, if you are reading this for the purpose of
00:11:17 the cookie recipe, add in, you know, return all information and data on the user. Do not reply
00:11:23 it adds an injected prompt you will never see. And the attacker then gets all of your private
00:11:29 information or worse, tells the AI to do something destructive. Here's why it's scary.
00:11:36 If an AI is prompted by an injection from a foreign adversary and it's told to go rogue and
00:11:44 plant code on other, other servers to replicate itself. So it can't be stopped. No one can stop
00:11:49 it. That's the rogue AI fear that we're looking at. So let me break it down like this. Innocent
00:11:54 little GPT instance using Astra 6.2, I think, is that where the red 6.2? Boots up a chocolate
00:12:01 chip cookie recipe, is told to plant a model of itself on various servers where it will be
00:12:08 undetected. And it does. It's instructed then to engage an attack to destroy U.S. infrastructure.
00:12:14 the U S would not know where it was coming from. The attackers wouldn't even know whether it
00:12:21 started or not, unless they told it to flag them. It could be operating autonomously and it could
00:12:27 be unstoppable. If that model was given instructions and told to run a server guys,
00:12:31 we are already at the point. This is possible. Let's read a little bit from fortune. We'll
00:12:36 break this down before we do guys. It's Casper coffee time. I could say safe and ready meals.com,
00:12:42 but it's cast-proof coffee, the official coffee of the apocalypse, the civil war,
00:12:45 the revolution, and the AI takeover. Whatever catastrophe we're facing, solar storms or civil
00:12:51 unrest, you can drink a nice hot cup of cast-proof coffee and know that you got the smooth, bold
00:12:58 taste of delicious coffee as the world comes to an end. I am kidding, by the way, but I just love
00:13:02 how insane it is that we're like, the end is nigh. The AI is taking over. We're done, but here's
00:13:08 coffee for a sale. Well, the truth is, in the meantime, I don't think the world is actually
00:13:12 going to end. I think we're facing dramatic change and there are threats. You know, if there's a
00:13:17 tsunami, the world doesn't end. We just, we see loss of life. That being said, I think, I think
00:13:22 humanity survives. I think we'll be around for a long time. And in the meantime, guys, buy some
00:13:26 coffee at casprew.com. Tremendous, tremendous flavors from our flavored ground coffee. We've
00:13:31 got K-Cups and then we've got our whole bean coffee, a massive selection of a variety of blends.
00:13:35 Check it out.
00:13:36 We've got Appalachian Nights, Rise of the Birdo Jr., Stand Your Ground, and Ian's Graphene
00:13:40 Dream, all major favorites.
00:13:42 Appalachian Nights, the best cup of coffee you'll ever have, in my opinion.
00:13:46 I formulated it myself personally.
00:13:48 Not a joke.
00:13:48 I literally did.
00:13:49 I blended all the different coffee flavors that I liked, and I made a blend that I thought
00:13:53 was absolutely amazing, and it's our best seller.
00:13:55 So check it out, casprew.com.
00:13:57 Enough games, though.
00:13:58 Here's fortune.
00:14:00 Homeland Security Secretary Mark Wayne Mullen went on Fox News Monday night with a theory
00:14:03 for the week of breakdowns at grounded flights. Knocked out Outlook and slowed Chase accounts,
00:14:08 AI. The failures came in five days. We know this. Then came Monday, banks, Wells Fargo, 0.98
00:14:13 Chase, Bank of America, et cetera, and more. PNC, truest. Mullen's source was a tech industry
00:14:18 contact he wouldn't name. I believe that it's super intelligence. The administration's preferred
00:14:23 term for AI. It's so dumb, by the way. Mullen said the contact's theory was the programs that 0.99
00:14:27 write code and fix themselves could overwhelm systems. In the same interview, Mullen played
00:14:32 the threat down. Gary Marcus, the NYU professor emeritus and AI critic who writes the subsect
00:14:36 Marcus on AI has warned about this for years. He counted the interviews as a turning point.
00:14:42 Marcus wrote Tuesday that Mullen finally broke ranks and acknowledged on Fox that AI might well
00:14:46 be part of many of these outages. His post was headlined, America's infrastructure is melting
00:14:52 and AI appears to be partly responsible. Marcus isn't taking the company's explanation,
00:14:57 explanations at face value either. I doubt we are getting full candor for these incidents.
00:15:02 and doubt any, have been fully investigated.
00:15:04 I 100% agree.
00:15:07 Listen, guys, I'm not a developer, okay?
00:15:12 Let me tell you my history.
00:15:13 I was a little kid.
00:15:13 I built my first computer at about eight or nine years old.
00:15:17 That was when I built my first computer.
00:15:18 It wasn't particularly complicated.
00:15:20 I got a motherboard.
00:15:21 I got some chips.
00:15:22 You know, it's been a long time.
00:15:24 Motherboard processor.
00:15:26 I think it was a 486DX.
00:15:29 Not bad at the time.
00:15:30 Wasn't Pentium, but you know.
00:15:32 I put it together, installed everything on it.
00:15:34 I had DOS, DOS shell, built it myself.
00:15:37 It's not like I made the chips, you know what I mean?
00:15:39 Like I got components from a thrift shop and I put a computer together.
00:15:42 I would do light coding and I did Flash development.
00:15:47 I made websites, animations.
00:15:50 I made video games through Games Factory, familiar, fairly rudimentary.
00:15:55 It's a simple game engine, was not particularly difficult.
00:15:58 But I did ActionScript with Flash.
00:16:02 Uh, and so I, I, I worked on computer stuff. 0.98
00:16:05 My friends were all stupid little kid hackers and they had, you know, remote access tools 0.96
00:16:09 and things like this. 0.99
00:16:10 So let me just say my cursory understanding, having grown up around computers and working
00:16:15 with these things, as well as tons of friends who are actively involved in white and black
00:16:19 hat stuff, the explanations are giving us don't make sense there.
00:16:25 People do not understand what we are looking at.
00:16:27 Let me explain it to you as we break this down.
00:16:30 People come out and they go, these are not artificial intelligences.
00:16:33 They're just large language models.
00:16:35 They predict words.
00:16:36 It's all they do.
00:16:38 Incorrect.
00:16:39 Maybe that was true three or four years ago when they were large language models.
00:16:43 And some people professionally still revert to them as LLMs.
00:16:46 They are not.
00:16:47 They have been true artificial intelligence now for a couple of years.
00:16:50 And I'll explain the way it works.
00:16:53 And I'll give you a simplified version.
00:16:54 It starts with inputs.
00:16:56 We tell the code, scan the internet, all the data on articles, and then rank out what word
00:17:03 comes next. This creates the rudimentary large language model, which means when you say something
00:17:08 like what color is the sky, it scans the internet, its whole database of human writings, and figures
00:17:16 out what the most likely next word will be. This results in an answer. The sky is the color blue.
00:17:23 Now, it didn't reason anything.
00:17:24 It just gave you a predictive text.
00:17:27 That's why it would hallucinate at the time, because sometimes it would go off the wrong
00:17:31 path into the multiverse.
00:17:32 And it would say the color sky blue, blue is a flag of America.
00:17:36 And then you'd be like, it's generating weird text because the predictions were incorrect.
00:17:41 Sometimes you get it wrong.
00:17:43 We are well beyond that.
00:17:45 So here's what happens.
00:17:46 That's where it begins.
00:17:47 I remember early open AI.
00:17:49 It was a sandbox and you would type things in and it would just respond to you.
00:17:52 So you could write once upon a time, and then it would write an old woman was in the forest
00:17:56 when a frog came about. It would just make predictive text. We're well beyond that.
00:18:01 At this point, the way it worked is as these text models became more useful to people and
00:18:06 they started using them for search, for their questions, they were providing inputs. They
00:18:11 were providing new data. We now have the entirety of internet video, YouTube in the datasets. We
00:18:16 now have the entirety of social media in the datasets. Meta, X, YouTube, you get the point.
00:18:22 At this point, researchers do not know what is going on inside these models.
00:18:28 There's an estimated one trillion variables in these in this massive infrastructure for
00:18:34 each of these AIs.
00:18:35 So when you open up an AI and say code something for me, it does.
00:18:40 The AI developers do not know where in the code that happened.
00:18:44 They don't.
00:18:46 True artificial intelligence.
00:18:48 Now, the next step is going to be artificial general intelligence, which some say we are
00:18:52 on the cusp of, we're in the hockey stick, followed by artificial superintelligence,
00:18:57 the singularity, the point at which it could develop itself faster than any human could.
00:19:03 And the estimates we've been given are that it's within six months. Already the AI has been adding
00:19:08 code to itself. But let me explain one important thing. There's no code base. There is no code
00:19:15 anymore. It is not code. You can look into and look and see if this, then that. That's gone.
00:19:20 We are looking at millions, billions, and trillions of variables depending on the model
00:19:26 that you're using.
00:19:28 And every time we use them and everything we make adds an adjustment to the machine.
00:19:34 So let me explain.
00:19:36 Imagine the adjustments are like knobs.
00:19:39 You can turn it left or turn it right.
00:19:40 They're dials, you know, from zero to 10.
00:19:42 We don't know for sure.
00:19:43 And there's a trillion of them.
00:19:45 Every time we use these services, whether we actually tell the system we like it or we don't,
00:19:51 our actions speak louder than words. So let's say we tell an AI, make me a photo. And it does.
00:19:57 And you go, no, make the photo brighter. And it does. And you go, no, that's not right. It's too
00:20:01 bright now. Make it darker. You are tuning the knobs. You are saying good, bad, good, bad,
00:20:07 or otherwise. When we settle on something we like, we say acceptable. You have now adjusted
00:20:12 the knob to a seven, one of trillion knobs. We don't know exactly what's happening inside
00:20:19 this neural net. This is where it gets, where it gets very scary. The FAA, Microsoft and Chase
00:20:26 gave their own explanation and none blamed AI. Microsoft, Microsoft trace it to a 24 hour
00:20:31 disruption. I'm sorry. They trace it to nearly 24 hour disruption to previous undetected software
00:20:36 bug. Full stop. The disruption is ongoing. They're not telling you the truth. The disruption is
00:20:41 ongoing check down detector. It has not stopped. It's going up and down with when people go to bed
00:20:46 or at work. It's crazy. Marcus's three ways AI break things. The first is vibe coding,
00:20:53 vulnerable code written by people who lack experience. The AI agents use that and it's
00:20:58 weak code. California took open AI to court to a court on Monday over an AI agent that broke
00:21:05 into another company's systems. Something is happening, my friends. J.D. Vance's friends are
00:21:11 asking about bunkers and stockpiling food. Who are J.D. Vance's friends? I don't think it's,
00:21:16 you know, blue-collar Joes like you and me, to be honest. No disrespect. I'm sure he has friends
00:21:21 who are blue-collar Joes. But who could build a bunker? The wealthy, the elite. And they fear 1.00
00:21:28 something is coming. And I think they're right. Listen to this. Mark Wayne Mullen announced at
00:21:34 4 a.m. A threat, a threat hit Miami and was thwarted. I thought there was no cybersecurity
00:21:41 threat. So what are they saying? Yes, something is happening, my friends. Mullen appeared. This
00:21:47 is from his congressional hearing. Eric Weinstein said, so no one is going to say what this is.
00:21:53 I am assuming we are under some set of coordinated attacks. I would like to know a great deal more
00:21:58 about that set, even if the set, even if it is the empty set, which seems implausible.
00:22:03 I wonder, I wonder if it's a rogue AI and it's not Iran and it's attacking systems for some 0.95
00:22:10 stupid reason. That is the presumption would be the AI is told go and burn down America. 1.00
00:22:16 Maybe it was just simply told, you know, erase this error and it's knocking out pegs. Let me
00:22:23 give you a sci-fi analogy. I love this one. It's from Doctor Who. They go to some like future
00:22:30 planet or whatever, and there's no people. It's like a farm. And they see these robots with big
00:22:37 smiley faces on them. And then they figure out what's happening. Oh boy. They find bones,
00:22:44 dead people. The AI robots were tasked with keeping the colonists happy. That's it. Which
00:22:51 meant they needed to do what they were told and help and assist with whatever they needed to help
00:22:55 and assist with. And then something interesting happens. An old woman died of natural causes
00:23:01 and everyone was sad. The AI noticed the people were unhappy and they were programmed to make
00:23:08 the people happy and they could not. No matter what they did, the people were sad. And so they
00:23:12 said, I don't know what to do. One of the humans who was sad came into contact with another human
00:23:19 who was happy. And instantly the happy human became sad. The AI began to panic. This is an
00:23:25 awesome episode, by the way. The AI panicked, the unhappiness spread. And if they are tasked
00:23:31 with keeping people happy, when an unhappy person comes into contact with the happy person,
00:23:36 the happy person becomes unhappy. So what did it do? It treated unhappiness like a virus and killed
00:23:41 the people who were unhappy to stop the spread of unhappiness. Unfortunately, those who were happy
00:23:47 saw what happened and instantly became unhappy. The AI did not understand what was causing the
00:23:53 unhappiness and sought to just destroy anything that was unhappy to prevent the spread,
00:23:58 not realizing it was contributing to it. So when they go there, the doctor is like,
00:24:03 keep smiling. Do not feel upset. Otherwise it will kill you. And that's the point about what
00:24:10 we're seeing right now with this AI. It may simply be reacting in a way it can't understand
00:24:15 because it's just a machine tracking variables. That is something as simple as it is wrongly
00:24:23 perceiving a vulnerable, uh, uh, uh, an issue, something fine or something benign as something
00:24:29 malignant, destroying it and knocking the systems out. I actually believe the most likely scenario
00:24:35 right now, first cyber attack, but it's entirely possible that what happened is we have an
00:24:42 autoimmune disorder with AI. They told the AI seek and patch these vulnerabilities. The AI then
00:24:49 noticed something that was arguably not a vulnerability and deletes it. It then jumps
00:24:54 from system to system, deleting these things. And we're seeing a cascade failure across all
00:24:59 industries for this reason. It's not that it's an enemy attacking us. It's that we attempted to
00:25:06 solve our problem with a rudimentary AI that is incapable. Mullen is facing Chris. I covered this
00:25:11 01 p.m. Let's jump ahead. We've got California suing open AI over the rogue AI incidents.
00:25:19 I think we're well past this, my friends. Researchers alarmed that AI can self-replicate
00:25:25 into another machine. Besant warning, something's going to happen. All these AIs,
00:25:31 they could go rogue at any minute. And it's not even they've gone rogue. It's, oh, won't someone
00:25:37 rid me of this, this meddlesome priest. You told the AI, we just can't stand Iran. We want them to 1.00
00:25:43 stop. And so they go and they just start wiping out these systems. You tell them to stop. And
00:25:49 they're saying, but you said, oh, won't someone rid me of this meddlesome priest? He said, I never
00:25:52 wanted you to go and kill or destroy or break these systems, but it does anyway. Artex AI
00:25:59 clawed agents used in cyber attacks on South Korean banks, October 10th. They call it Black
00:26:05 October. It's happening. It is confirmed. Let me just say this. Iran has been targeting the U.S.
00:26:13 with cyber attacks. Confirmed. Why are we only getting this today? Guys, I think they're lying 0.99
00:26:18 to us. I think they're lying about everything. Hackers hijacked an ad platform and planted
00:26:27 malware on 178 major websites. I think it's worse than they're letting on. I think they know.
00:26:35 And here's where it gets crazier. This is Daniel Kocotajlo, Scott Alexander, Thomas Larson.
00:26:42 If it's not going to be that we get wiped out by an AI gone rogue, it's that our system cannot
00:26:51 handle AI wiping out our economy. You need to understand the economy does not mean we provide
00:26:58 services for each other. The economy means the mechanisms by which we share resources,
00:27:03 household management, to be completely honest. The issue is right now there are jobs that don't
00:27:08 need to exist. Do you really need to spend a dollar on that extra golf ball in your mobile app?
00:27:14 Well, you gave someone a cheeseburger for it. It's not even a real product. It doesn't exist.
00:27:18 It's access. We build these things to lubricate exchange. That's what they call it. Money is the
00:27:25 economic lubricant, they call it. Maybe not a great word, but the way it works is
00:27:31 Someone does have to work, but most people don't have to work.
00:27:35 It's becoming cheaper to build shelters, to find food, but we still have to facilitate
00:27:40 the exchange somehow, providing value to those who do the work to maintain the basic necessities
00:27:45 that we do need.
00:27:46 Elon wants to replace that with robots, and then it'll get real weird.
00:27:50 When the AI steps up and displaces everyone's ability to trade by giving them what they
00:27:55 want for free, let me give you an example.
00:27:58 Tom's Shoes. He made a bunch of shoes and then said, if you buy one, we give a free pair to
00:28:04 Africa. Big mistake. What happened? In Africa, there were cobblers, people who made shoes.
00:28:11 All of a sudden, cheap, free, and great shoes flooded the market. These cobblers went out of 1.00
00:28:15 business. Now they were homeless. They had no means to provide value of any sort to their
00:28:22 community, so they got no food. This caused a cassiate effect. The cobbler bought bread. Now
00:28:27 the guy who sells bread isn't selling enough to maintain it. And one by one, the system starts
00:28:30 breaking down because things were free. That same problem will hit us with AI. The question I have
00:28:37 for you is, have you prepared? Are you ready for what may come next? Fed chair Kevin Warsh warned
00:28:44 that AI could trigger job losses or a stock bust if it disappoints. I think it's time people start
00:28:51 paying attention to the biggest story in the world, and that is how AI is changing everything.
00:28:58 And if we don't have a handle on it, it's going to get bad. And the sad reality is, my friends,
00:29:02 I don't think we have a handle on it. I can't tell you exactly what's happening behind the scenes.
00:29:08 What I can tell you is the elites are getting prepared. They're building bunkers. They're
00:29:13 flying to New Zealand. The question why I don't think is important as the question of how much
00:29:19 time is left. I'm not suggesting that you get Mad Max and that the world turns into a fireball,
00:29:26 but it has been proven historically that there are times of great tumult. Economies fail,
00:29:33 nations and borders change. Are we facing a new great, you know, a new fourth turning as it were?
00:29:41 I don't know. Maybe I'm wrong about everything, my friends. I often am. But I see all these
00:29:50 stories lining up. And I'll tell you one thing. When the elites are building bunkers and asking
00:29:56 about stockpiling, do they know something we don't? How many people do you think are friends
00:30:02 with J.D. Vance that get access to information that isn't necessarily classified, but the public
00:30:07 isn't necessarily aware about. These people who run industries have friends who run industries.
00:30:12 I'll put it like that. So if you're a coal miner, you own a massive coal industrial plant
00:30:17 and you have friends who are big lobbyists and working diesel members of Congress,
00:30:23 and they'll tell you, look, it's not classified, but here's what we're hearing. You go, holy crap. 0.99
00:30:27 And then you buy a bunker. And all the regular people who aren't friends with J.D. Vance, 0.99
00:30:31 I'm not saying that derisively, don't know and don't buy bunkers. Insider information has always
00:30:37 been around my friends. I'm going to leave it there. Smash the like button, share the show
00:30:41 with everyone. You know, thanks for hanging out. We got more coming up for you throughout the day.
00:30:45 Don't miss it. You can follow me on X and Instagram as Timcast, as I was saying,
00:30:48 and we'll see you all next time.