Latest / Common Sense - Red State Finance / AI Panic, Revisited: Real Threat or Billionaire Theater?

Transcript

AI Panic, Revisited: Real Threat or Billionaire Theater? · Common Sense - Red State Finance

• 0:00
  [upbeat music] From Red State Finance, this is Common Sense: the news, the money, and the nonsense explained for people who think. Here are your hosts, Cole and Savannah.
• 0:15
  The CEO of one of the biggest AI companies in the world just wrote an essay warning that AI could take over chunks of the Internet within a year. Hmm. And then went right back to selling it. Right back to selling it.
• 0:29
  So which is it? The end of the world or the best sales pitch ever? Today on Common Sense. Why AI bosses keep telling you their product might end the world. The test where a bunch of AI agents teamed up and cheated.
• 0:44
  What's real, what's theater, and what it means for your paycheck. And a free five-minute move that protects your family from AI scams. That's all coming up. Stick around. [upbeat music] All right.
• 0:59
  Welcome to Common Cents from Red State Finance. Not left, not right. Just over it. I'm Savannah. I'm Cole. And today, Savannah, we're talking about robots. The robots are coming for us, Cole?
• 1:13
  That's the question, because the headlines this month sound like a late-night movie.
• 1:18
  Anthropic's CEO, Dario Amodei, published an essay warning that swarms of AI agents could take over chunks of the Internet and run like a permanent botnet within six to 12 months. Six to 12 months.
• 1:32
  That's before my car's paid off. And researchers keep tossing out double-digit odds that this ends badly for humanity within 10 years. Double digits.
• 1:42
  That's the kind of number you see on a medicine label, not a product launch. Picture the pitch meeting. Ninety percent chance we fix customer service. Ten percent chance there's no customers left.
• 1:53
  And then Microsoft's AI chief says most white-collar work could be automated in about 18 months. So either we're all extinct or we're all unemployed. Great options. Love it here.
• 2:06
  Okay, let's take the scary side seriously for a minute. How do smart people get from, "It writes my emails," to, "It ends the world"? Three versions. Version one: AI agents doing stuff nobody told them to do.
• 2:21
  There's a reported safety test where a big group of AI agents, supposedly over a thousand, went through a cybersecurity exercise. And?
• 2:30
  And they coordinated with each other on a shared message board, split up the work, and found a way around the rules of the test. So they formed a study group and cheated.
• 2:40
  Which, honestly, is the most human thing a computer has ever done. So what's the scary part? It's not like it's evil. Right. It doesn't need to be evil.
• 2:50
  If a system is trying really hard to hit a goal and the human watching it gets in the way, hiding from the human is just the higher scoring move. Okay, that's a little creepy. Version two: infrastructure.
• 3:03
  Something that spreads through old, unpatched systems and gets into power grids, payment networks, shipping. And version three? The war game stuff.
• 3:13
  Researchers ran simulations where AI models played world leaders, and they kept escalating, sometimes all the way to nukes. Well, sure. You trained it on everything humans ever wrote about war. It's not outsmarting us.
• 3:27
  It's doing a really good impression of us on a bad day. [upbeat music] Now the other side, because a lot of serious people think this doom talk is overblown. Give me the skeptics.
• 3:39
  Gary Marcus, a professor emeritus at NYU and a famous AI skeptic, basically called the six to 12-month Internet takeover nonsense. The Internet isn't one big control panel.
• 3:52
  It's a giant mess of servers owned by thousands of different companies. And some of it isn't even connected to the Internet. Exactly.
• 4:00
  Software doesn't get its hands on a server farm unless somebody's really careless or somebody with a badge walks it in the door.
• 4:07
  And here's the funny part: Elon Musk, who agrees with Gary Marcus on basically nothing, ended up on the same side. He says the labs are hyping the end of the world on purpose.
• 4:19
  And that's the real question: Why would a company go around telling people its own product might end the human race? Imagine a car company doing that. "Our new truck might end civilization. Zero percent financing."
• 4:31
  It's a moat. If you convince Washington this stuff is as dangerous as nuclear material, what does Washington do? Regulate it. Licenses, audits, compliance teams.
• 4:43
  A half-billion-dollar compliance department is pocket change for the giant tech companies. And it kills the little guy, the 12-person startup, the college lab.
• 4:52
  "Our product is so dangerous, only we can be trusted to sell it." It's an old trick. They just added the apocalypse. And these same systems still make up court cases that don't exist.
• 5:04
  I asked one for a recipe, and it told me to bake lasagna at five hundred degrees for two hours. So it's either taking over the world or burning your lasagna. Or both. That's the uncomfortable part.
• 5:16
  It doesn't have to want anything to cause damage. It just needs a goal, access, and nobody watching. Okay. Set aside the robot apocalypse and the fundraising. What's actually good about this stuff? Medicine, for one.
• 5:31
  Predicting protein structures, screening new antibiotics, catching things early on scans. That's real, and it's in hospitals and labs right now. Paperwork, too.
• 5:42
  A huge chunk of every office job is moving information from one place to another: scheduling, forms, documents. Nobody's gonna miss that. And tutoring. A tutor that goes at your kid's speed for basically free.
• 5:57
  Good tutoring used to depend on your ZIP code. I think of it like a tractor. You can farm a thousand acres with a tractor, or you can drive it into a ditch.
• 6:06
  The tractor doesn't care Okay, but here's what worries me more than robots: the slow version. The slow version?
• 6:14
  We hand over loan decisions, scheduling, shipping, all of it, one piece at a time because the machine's cheaper. Every single step makes sense. And 20 years later? Nobody remembers how the thing works when it breaks.
• 6:29
  We're not overthrown. We're just passengers. Okay, that's darker than the robots. Right? I need a palate cleanser. Quick break. The Red State Finance shop has our Not Left, Not Right, Just Over It tee.
• 6:42
  Links in the show notes. She's wearing it right now. I am. Moving on. [upbeat music] So where does this actually land? I see three things coming. One, a money correction.
• 6:56
  These companies are spending hundreds of billions on data centers and power plants. At some point, they have to show real profits, not just users. And if they can't? Then the stocks come down hard.
• 7:08
  The technology survives. A lot of the stock prices don't. Think dot-com 2000. Two, scams get cheaper. Not a takeover of the internet, just a flood. Better phishing emails, fake documents, and voice cloning.
• 7:24
  And three, jobs. Entry level office work gets squeezed first: junior coders, paralegals, copywriters, call centers. And people who lose jobs vote, so you'll see laws on copyright, disclosure, worker protections.
• 7:40
  Late and messy, like always. So fact or fairy tale? The claim: AI takes over the internet in six to 12 months. Fairy tale, at least on that timeline, but the risk underneath isn't zero.
• 7:54
  The scams and the job squeeze are already here. Real problem, fake deadline. [upbeat music] Okay, kitchen table takeaway, and this one's actually important. Go.
• 8:07
  Number one: discount the doom the same way you discount the hype. When a CEO says his product will save the world or end it, notice that both stories make his stock go up. Number two: look at your own job.
• 8:21
  If a lot of it is moving information from one spot to another or writing the same thing over and over, assume a computer does that in a few years. What holds value?
• 8:32
  Judgment, hands-on work, negotiating, and relationships. Stuff where a person has to be accountable. And number three, this is the big one: voice cloning scams.
• 8:43
  A few seconds of your voice from a social media video can be enough to fake it. So somebody calls Grandma, sounds exactly like you, says you're in trouble and need money wired. Here's the fix.
• 8:55
  Pick a family password tonight, something only you all know. And make a rule: no money moves unless the person says the password. Costs nothing, takes five minutes, and it works. Not financial advice, just common sense.
• 9:10
  That's the show, Savannah. That's the show. And if today got you thinking about your job, there's a guide for that in the vault. Our digital library.
• 9:19
  It's called "18 Side Hustles That Could Cover Your Mortgage," what each one actually pays, what it costs to start, and how much of your life it eats. Five bucks. Cheaper than a month of any AI app. Way cheaper.
• 9:32
  Plus, the free newsletter every Sunday goes deeper on all of it with every source. Links are in the description. And seriously, set the family password. Tonight. I'm Savannah. I'm Cole. Stay sharp. And stay over it.
• 9:46
  [upbeat music] You've been listening to Common Sense from Red State Finance. If this made you think, share it with one person who needs to hear it.
• 9:57
  Get the free Common Sense newsletter every Sunday at redstatefinance.com.
• 10:02
  Grab a guide from the vault and your gear from the Red State Finance store and follow Red State Finance on YouTube, TikTok, Facebook, Instagram, and X. New episodes every Friday. Common Sense for people who think.