Latest / The Matt Walsh Show / Only A Few People Win from AI Hysteria. Everyone Else Loses.

Transcript

Only A Few People Win from AI Hysteria. Everyone Else Loses. · The Matt Walsh Show

• 0:00
  Speaker 1: Opinions on AI seem to be, to put it mildly,
• 0:04
  Speaker 1: sharply divided. But one thing we should all be able
• 0:06
  Speaker 1: to agree on, a statement that should attract a wide consensus,
• 0:10
  Speaker 1: is that this is an incredibly consequential technology, advancing and
• 0:15
  Speaker 1: scaling at enormous speeds, and both the benefits it provides
• 0:19
  Speaker 1: and the threat it potentially poses should be seriously considered
• 0:23
  Speaker 1: in an objective, sober-minded way. If we can agree on
• 0:27
  Speaker 1: nothing else when it comes to AI or anything else,
• 0:31
  Speaker 1: we should be able to agree on that. It's not
• 0:33
  Speaker 1: a political statement. It's not an ideological statement. And yet
• 0:37
  Speaker 1: sober-minded objectivity is the very last thing we're seeing on
• 0:41
  Speaker 1: this subject from either side of the debate. And the
• 0:44
  Speaker 1: events of this week perfectly illustrate that point. So let's
• 0:47
  Speaker 1: start with those events, which immediately reveal why sober minds
• 0:52
  Speaker 1: are not prevailing and why they desperately need to. There
• 0:57
  Speaker 1: is a colossal amount of money and power at stake.
• 1:00
  Speaker 1: For example, let's say that you're a massive privately held
• 1:04
  Speaker 1: artificial intelligence company, and you're planning a blockbuster IPO in
• 1:08
  Speaker 1: a matter of months, which will probably value your company
• 1:11
  Speaker 1: at several trillion dollars. You have every incentive to promote
• 1:16
  Speaker 1: your product and to convince as many people as possible
• 1:19
  Speaker 1: to sign up and pay a subscription fee. Why in
• 1:22
  Speaker 1: this scenario would you go on national television and tell
• 1:26
  Speaker 1: everybody that your product has the potential to eliminate all
• 1:29
  Speaker 1: of humanity within the next decade? Why would you deliberately
• 1:32
  Speaker 1: torpedo public confidence in the entire AI industry at a
• 1:36
  Speaker 1: pivotal moment for your company? You know, it seems a
• 1:41
  Speaker 1: bit like Taco Bell announcing a new burrito by explaining
• 1:44
  Speaker 1: in great detail how it's going to, you know, absolutely
• 1:47
  Speaker 1: demolish your intestines and potentially all of the intestines on Earth.
• 1:51
  Speaker 1: Doesn't seem logical. Seems like self-sabotage, and yet that's exactly
• 1:55
  Speaker 1: what Anthropic, which makes the AI model known as Claude,
• 1:59
  Speaker 1: has apparently just done, providing an incredible window into how
• 2:03
  Speaker 1: the most powerful corporations on the planet can manipulate public
• 2:07
  Speaker 1: opinion and the variety of tactics they use to push
• 2:11
  Speaker 1: more government regulation and control over the economy. So this
• 2:15
  Speaker 1: was an op. There's no question about it, and it
• 2:18
  Speaker 1: was a very successful one. And yet, and this is
• 2:21
  Speaker 1: where we need those sober minds to come in, the
• 2:23
  Speaker 1: concerns that are being raised, even if they're being raised
• 2:27
  Speaker 1: for less than benevolent reasons, still should be taken seriously.
• 2:32
  Speaker 1: And the very fact that they're being raised at all,
• 2:35
  Speaker 1: whether genuinely or not, create a scenario where no matter
• 2:39
  Speaker 1: what's actually going on, it's pretty bad. And I'll explain
• 2:43
  Speaker 1: why in a few moments. But here's how it began.
• 2:47
  Speaker 1: And this timeline was largely put together by a researcher
• 2:49
  Speaker 1: named Parker Thayer. So on Tuesday evening, just before 8 p.m. Eastern,
• 2:54
  Speaker 1: the Wall Street Journal published an exclusive story with the
• 2:56
  Speaker 1: very alarming headline, Anthropic Researcher Quits Over Out-of-Control AI Fears.
• 3:02
  Speaker 1: Concerns are rising inside AI labs. The competition is pushing
• 3:05
  Speaker 1: tech companies to race towards self-improving models that risk spiraling
• 3:10
  Speaker 1: out of human control. The article states that, quote, Jacob
• 3:14
  Speaker 1: Kokson said, A researcher who specializes in training new AI
• 3:18
  Speaker 1: models by having them consume vast amounts of data said
• 3:21
  Speaker 1: Tuesday that he is leaving the company because he doesn't
• 3:24
  Speaker 1: want to participate in an industry-wide rush to build AI
• 3:27
  Speaker 1: systems that can improve themselves, worried such systems could spiral
• 3:31
  Speaker 1: out of control and destroy humanity. Now, this article is
• 3:35
  Speaker 1: about a 27-year-old who, as the article eventually mentions, left
• 3:39
  Speaker 1: OpenAI earlier this year to join Anthropic because it's known
• 3:44
  Speaker 1: for its model safety efforts. So in other words, Jacob
• 3:47
  Speaker 1: Kokson was working at Anthropic's competitor OpenAI until very recently.
• 3:52
  Speaker 1: He was at OpenAI for about three years before coming
• 3:55
  Speaker 1: over to Anthropic earlier this year. Now, based on all
• 3:58
  Speaker 1: the reporting I've seen, he's been at the company for
• 3:59
  Speaker 1: around five months at the most. So after a very
• 4:03
  Speaker 1: short stint as an entry-level employee, he's already become a
• 4:06
  Speaker 1: whistleblower somehow. He's sounding the alarm about how AI is
• 4:10
  Speaker 1: gonna destroy the entire human race. This is an unusual timeline.
• 4:17
  Speaker 1: To put it mildly, for most entry level employees, when
• 4:20
  Speaker 1: you start at a new company, you spend the first
• 4:22
  Speaker 1: few months finding your way around the building. You know,
• 4:25
  Speaker 1: you're trying to meet new people, you know, trying to
• 4:28
  Speaker 1: figure out your commute, that kind of thing. But for
• 4:30
  Speaker 1: this guy, after a very short period of time, he
• 4:33
  Speaker 1: was ready to quit in the most flamboyant, splashy way imaginable,
• 4:37
  Speaker 1: complete with an exclusive article in The Wall Street Journal
• 4:40
  Speaker 1: announcing his departure. And on top of that, the Wall
• 4:42
  Speaker 1: Street Journal doesn't even mention that the guy is brand
• 4:45
  Speaker 1: new to the company until the 10th paragraph.
• 4:48
  Speaker 2: So think about how bizarre this is.
• 4:50
  Speaker 1: He clearly arrived at Anthropic, it would seem, with the
• 4:53
  Speaker 1: idea of quitting in mind. He's not a senior level
• 4:57
  Speaker 1: employee or anybody with highly specialized knowledge. He doesn't have
• 5:00
  Speaker 1: any kind of impressive technical background. And yet the moment
• 5:05
  Speaker 1: he quits, he's making very dire predictions. The most dire
• 5:09
  Speaker 1: prediction you could possibly make, which is the extinction of
• 5:11
  Speaker 1: the human race. And he's getting national news coverage that
• 5:15
  Speaker 1: doesn't often mention or at least highlight his very short
• 5:18
  Speaker 1: tenure at the company. And the coverage wasn't just coming
• 5:21
  Speaker 1: from the Wall Street Journal. Variety, Time, NBC News, several
• 5:24
  Speaker 1: other outlets have run similar stories. Fox interviewed him as well.
• 5:29
  Speaker 2: Watch.
• 5:31
  Speaker 3: Minimal of actual takeover. But because of self-improvement, this AI
• 5:35
  Speaker 3: is making themselves smarter could happen as soon as next year.
• 5:38
  Speaker 3: Some people even say six months. So this attitude is
• 5:41
  Speaker 3: shared constantly. across the industry, where people essentially are aware
• 5:45
  Speaker 3: that the work they are currently doing, if we don't
• 5:47
  Speaker 3: slow down, could lead to AI takeover by the end
• 5:51
  Speaker 3: of the decade. I think you have to think of
• 5:53
  Speaker 3: this as an arms race. This technology is not ordinary technology.
• 5:58
  Speaker 3: It's not like other forms of innovation. This is essentially
• 6:01
  Speaker 3: a super weapon. And I think that I expect the
• 6:04
  Speaker 3: international community to be ready to make agreements about how
• 6:08
  Speaker 3: to control the development of super weapon technology. Yeah, so
• 6:12
  Speaker 3: unfortunately we know how to control nuclear weapons pretty much.
• 6:15
  Speaker 3: We don't yet know how to control AI. This is
• 6:18
  Speaker 3: possibly the most dangerous technology that humanity has ever created.
• 6:21
  Speaker 3: It's basically out of science fiction. I think we have
• 6:24
  Speaker 3: no other choice but to cooperate internationally because an arms
• 6:27
  Speaker 3: race would be disastrous in a way that no other
• 6:31
  Speaker 3: human activity has been in the past.
• 6:33
  Speaker 2: So the pitch is pretty clear.
• 6:34
  Speaker 1: He's saying he wants the government to regulate AI domestically
• 6:38
  Speaker 1: and internationally. He wants to slow the pace of AI innovation.
• 6:41
  Speaker 1: And he's presented as a kind of good Samaritan, as
• 6:45
  Speaker 1: a man who bravely sacrificed his lucrative career in order
• 6:48
  Speaker 1: to tell the truth about this existential danger. This random
• 6:53
  Speaker 1: guy out of nowhere appeared on pretty much every TV
• 6:55
  Speaker 1: station last night as an expert on the topic.
• 6:58
  Speaker 2: Watch.
• 6:59
  Speaker 4: Jacob, I appreciate you being with us. First of all,
• 7:03
  Speaker 4: I assume you are saying literally that you believe that
• 7:08
  Speaker 4: AI could kill us all by the end of the decade.
• 7:12
  Speaker 2: How would that happen?
• 7:15
  Speaker 4: What would that actually look like?
• 7:18
  Speaker 3: Yeah. So the precise scenario sounds a little bit like
• 7:21
  Speaker 3: science fiction. It sounds like something that's not real, but
• 7:25
  Speaker 3: I think it is frighteningly real. And it's easiest to
• 7:27
  Speaker 3: understand it if you think about real things that happened
• 7:29
  Speaker 3: two months ago. So two months ago, OpenAI agents, AIs,
• 7:34
  Speaker 3: hacked into third-party infrastructure entirely at their own accord. And
• 7:38
  Speaker 3: this was like a concentrated hacking spree that they carried
• 7:42
  Speaker 3: out of their own volition. And I think that if
• 7:44
  Speaker 3: you extrapolate into the future the level of capabilities of
• 7:48
  Speaker 3: these AIs, with the same independent volition, they could cause
• 7:51
  Speaker 3: extreme havoc. For example, hacking critical infrastructure, building extinction-level bioweapons.
• 7:58
  Speaker 3: There's a lot of... ways that the AI could actuate
• 8:02
  Speaker 3: itself in the world.
• 8:03
  Speaker 1: Now, in this clip, he's referring to a real situation
• 8:05
  Speaker 1: called the Hugging Face Incident, where AI agents managed to
• 8:09
  Speaker 1: break containment and hack a well-known website using a security
• 8:12
  Speaker 1: vulnerability that the AI agents discovered on their own.
• 8:15
  Speaker 2: So the AI went rogue, in other words.
• 8:18
  Speaker 1: And while that's both impressive technologically and concerning deeply, it
• 8:24
  Speaker 1: is also different from hacking critical infrastructure. That's because unlike
• 8:27
  Speaker 1: the website that these AI agents hacked, Most critical infrastructure
• 8:31
  Speaker 1: in this country is physically separated from the internet. There's
• 8:34
  Speaker 1: a so-called air gap between the systems that run the
• 8:36
  Speaker 1: nuclear power plants and the rest of the online world.
• 8:39
  Speaker 1: And this security measure exists precisely so that hackers can't
• 8:43
  Speaker 1: find a random vulnerability and kill millions of people. Now,
• 8:47
  Speaker 1: the reason this guy and Anderson Cooper doesn't mention this
• 8:51
  Speaker 1: is that they want to stoke fear, clearly. And to
• 8:57
  Speaker 1: be clear, I'm not denying that AI hackers are going
• 9:01
  Speaker 1: to be a problem. Even if they can't access nuclear reactors,
• 9:05
  Speaker 1: we would hope, they can still hijack, say, the payroll
• 9:09
  Speaker 1: systems at the utility. Something like that actually happened in 2021,
• 9:13
  Speaker 1: although AI agents weren't involved. AI agents can hack personal accounts.
• 9:18
  Speaker 1: They can blackmail employees. In theory, they can cause plenty
• 9:22
  Speaker 1: of damage. But if you're going to say that AI
• 9:24
  Speaker 1: can cause nuclear Armageddon, then you need to show your work.
• 9:29
  Speaker 1: You need to hum us a few more bars. How
• 9:32
  Speaker 1: exactly is that going to happen? And this whistleblower never
• 9:35
  Speaker 1: does that. He never explains exactly how AI can go
• 9:38
  Speaker 1: from hacking one website to destroying humanity. And that's a problem.
• 9:43
  Speaker 1: It makes you wonder if he doesn't have an ulterior motive.
• 9:47
  Speaker 1: And indeed, when you look closer at the narrative.
• 9:49
  Speaker 2: He's pushing, it starts to fall apart.
• 9:50
  Speaker 1: And I'm not just talking about the fact that this
• 9:53
  Speaker 1: guy is a 27-year-old kid who just arrived at Anthropic
• 9:55
  Speaker 1: a few months ago. and now he's suddenly on every
• 9:58
  Speaker 1: primetime television show. And first of all, in terms of bias,
• 10:03
  Speaker 1: we have no idea precisely how much equity Kokson holds
• 10:06
  Speaker 1: in Anthropic or OpenAI. The fact he quit the company
• 10:08
  Speaker 1: doesn't tell us anything about his wealth or potential wealth
• 10:12
  Speaker 1: that's tied to the success of these companies. There's certainly
• 10:16
  Speaker 1: no indication that Coxon has ever surrendered whatever equity he
• 10:19
  Speaker 1: might have had in open AI, although I've seen articles
• 10:21
  Speaker 1: saying his anthropic stock won't vest.
• 10:25
  Speaker 2: We don't really know.
• 10:26
  Speaker 1: And that raises the possibility that he's engaging in a
• 10:28
  Speaker 1: marketing stunt, an attempt to make AI software seem more
• 10:30
  Speaker 1: powerful by talking about how dangerous and scary it can
• 10:34
  Speaker 1: be in the wrong hands. And secondly, there's new reporting
• 10:37
  Speaker 1: from Brian Chow linking Coxon to a British NGO called Newspeak.
• 10:42
  Speaker 1: which specializes in running influence campaigns for deranged leftist policies
• 10:47
  Speaker 1: like prison abolition. Some of Newsweek's, Newsweek's, Newspeak's employees explicitly
• 10:53
  Speaker 1: say they're experts in propaganda. Chow found Coxon's photo on
• 11:00
  Speaker 1: Newspeak's website, quoting from his reporting, influence campaigns were Newspeak's
• 11:05
  Speaker 1: overt specialty. The Guardian describes Newspeak's mission as trying to
• 11:09
  Speaker 1: hack British politics. So this guy's a political operative, or
• 11:15
  Speaker 1: at least he was a political operative. And there's a
• 11:18
  Speaker 1: reason why Anthropic may have deployed him in this case
• 11:22
  Speaker 1: in order to stall the development of.
• 11:23
  Speaker 2: AI and call for more regulation.
• 11:25
  Speaker 1: This is an article from Business Standard from just one
• 11:28
  Speaker 1: month ago, or rather one week ago, quoting the Financial Times. Quote, OpenAI,
• 11:33
  Speaker 1: the company behind ChatGPT, has launched its latest artificial intelligence model,
• 11:37
  Speaker 1: saying it has taken the lead over rival anthropic and
• 11:40
  Speaker 1: key areas such as software engineering, science, and cybersecurity.
• 11:44
  Speaker 2: The company described GPT-6 Astra.
• 11:47
  Speaker 1: As the world's most intelligent AI model as it seeks
• 11:49
  Speaker 1: to regain its position at the top of the fast-growing
• 11:52
  Speaker 1: AI industry. OpenAI said the model has achieved market-leading results
• 11:57
  Speaker 1: in several fields. So in other words, if these reports
• 12:00
  Speaker 1: are accurate, then Anthropic is falling behind in the race
• 12:04
  Speaker 1: to develop AI. It's a very competitive industry, obviously, and
• 12:08
  Speaker 1: at the moment, it looks like Anthropic is slipping into
• 12:11
  Speaker 1: a kind of runner-up status. So what's a company supposed
• 12:15
  Speaker 1: to do when it's getting outclassed? One of the most
• 12:17
  Speaker 1: common strategies is called regulatory capture. The idea is to
• 12:20
  Speaker 1: seize control over the regulators by convincing the public and
• 12:24
  Speaker 1: the government that you're the only company that cares about
• 12:27
  Speaker 1: safety and protecting humanity, And then once you have control
• 12:30
  Speaker 1: over the regulators, you can simply eliminate your competitors using
• 12:33
  Speaker 1: the force of law. So the dream scenario for companies
• 12:36
  Speaker 1: like Anthropic or OpenAI is to ban all the other
• 12:39
  Speaker 1: non-woke open source AI models with the justification that they're
• 12:43
  Speaker 1: simply too dangerous for humanity. In this interview, by the way,
• 12:49
  Speaker 1: by way of skepticism, Brett Baer did ask Coxon whether
• 12:52
  Speaker 1: he coordinated his big resignation with any third party.
• 12:56
  Speaker 2: And, you know, he of course said no.
• 12:58
  Speaker 1: But there's reason to doubt that response, certainly not to
• 13:01
  Speaker 1: take it at face value. One of the biggest issues
• 13:04
  Speaker 1: is that within minutes of the publication of the Wall
• 13:06
  Speaker 1: Street Journal article, a coordinated social media strategy clearly began.
• 13:12
  Speaker 1: So first, within 20 minutes of the article going live,
• 13:14
  Speaker 1: Cox had posted this message on X, and it's notable
• 13:18
  Speaker 1: that his account at the time of this posting was dormant.
• 13:21
  Speaker 1: So there was virtually no prior activity whatsoever. Quote, I
• 13:25
  Speaker 1: resigned from Anthropic today. I spent the last three years
• 13:27
  Speaker 1: doing pre-training research at both OpenAI and Anthropic. Neither company
• 13:33
  Speaker 1: is acting responsibly. They are racing straight to self-improving super
• 13:37
  Speaker 1: intelligence and gambling with our lives. Do not underestimate the
• 13:41
  Speaker 1: power of this technology. The people building AI earnestly believe
• 13:44
  Speaker 1: that it could kill us all by the end of
• 13:46
  Speaker 1: the decade. This is not a marketing stunt. If anything,
• 13:49
  Speaker 1: many executives and senior researchers will couch their phrasing in
• 13:52
  Speaker 1: the press to sound sensible and but I hear the
• 13:54
  Speaker 1: same people express fear privately. No other human activity poses
• 13:58
  Speaker 1: this level of danger. I don't feel like we're on
• 14:00
  Speaker 1: track to prevent a global race, which may require costly actions,
• 14:04
  Speaker 1: such as a temporary ban on improving model capabilities. Now,
• 14:09
  Speaker 1: this post, last I checked, has more than 130 million
• 14:13
  Speaker 1: views on X. Yes, 130 million. It's one of the
• 14:18
  Speaker 1: most widely viewed posts in the history of the platform.
• 14:22
  Speaker 1: Every hour, tens of millions of more people see it.
• 14:25
  Speaker 1: According to Elon Musk, nothing like this has ever happened
• 14:28
  Speaker 1: with a new account with almost no prior activity. So
• 14:32
  Speaker 1: this is clearly a post that some very powerful people
• 14:36
  Speaker 1: wanted to go viral. And when I first saw this
• 14:38
  Speaker 1: post on X without any of the context I just
• 14:40
  Speaker 1: went through, I thought it was pretty remarkable.
• 14:42
  Speaker 2: And indeed, it is remarkable.
• 14:44
  Speaker 1: Someone who's working on AI is telling us that it
• 14:47
  Speaker 1: might wipe out the human race in less than a decade.
• 14:53
  Speaker 1: I mean, that's notable no matter how you slice it,
• 14:56
  Speaker 1: no matter how you interpret it. And this has happened
• 14:59
  Speaker 1: many times before to some degree or another. Sometimes the
• 15:02
  Speaker 1: accusation is credible, sometimes less so. But at a minimum,
• 15:06
  Speaker 1: a post like this should make you ask yourself about
• 15:09
  Speaker 1: the risks of AI. Doesn't mean we need to ban
• 15:12
  Speaker 1: it or that we even could do that if we
• 15:14
  Speaker 1: wanted to, but it also doesn't mean we should embrace
• 15:16
  Speaker 1: it without hesitation as this kind of unalloyed blessing on mankind.
• 15:22
  Speaker 1: If even the people working on this stuff, and again,
• 15:24
  Speaker 1: this guy's far from the first one, if even they
• 15:29
  Speaker 1: are telling us that there are potentially catastrophic implications, something
• 15:33
  Speaker 1: Elon Musk has also said, by the way, we should
• 15:36
  Speaker 1: take that seriously. Our government should take it seriously too.
• 15:40
  Speaker 1: That was my take at the time. It's been my
• 15:43
  Speaker 1: take for years now. I still think it's the correct one,
• 15:46
  Speaker 1: but it soon became clear that something else was going
• 15:49
  Speaker 1: on with this particular narrative at least the version coming
• 15:53
  Speaker 1: from this particular guy. So within 15 minutes of this
• 15:57
  Speaker 1: post going live, as Thayer reported, it received several reposts
• 16:00
  Speaker 1: from officers at various anti-AI nonprofits, including the General Council
• 16:05
  Speaker 1: of ENCODE AI, the head of policy at the AI
• 16:08
  Speaker 1: Policy Network, the head of the AI Futures Project. And
• 16:14
  Speaker 1: here's when things get interesting. Quote, the AI Futures Project
• 16:17
  Speaker 1: website says it is funded primarily by the Survival and
• 16:21
  Speaker 1: Flourishing Fund, which says on its own website that it
• 16:25
  Speaker 1: has advised John Talen, Skype creator and one of the
• 16:29
  Speaker 1: leading investors in Anthropic, to grant over $ 2. 5 million to
• 16:32
  Speaker 1: the AI Futures Project since 2024. Encode AI says on
• 16:36
  Speaker 1: its website that it is also funded by the Survival
• 16:39
  Speaker 1: and Flourishing Fund, which in turn says that it told
• 16:43
  Speaker 1: Anthropic investor Talen to grant $ 516, 000 to Encode AI in 2025,
• 16:49
  Speaker 1: And wouldn't you know it, the Survival and Flourishing Fund
• 16:52
  Speaker 1: also says it told Talon to grant $ 2 million to
• 16:56
  Speaker 1: the AI Policy Institute, the 501c3 affiliate of the AI
• 17:00
  Speaker 1: Policy Network as well. What are the odds that the
• 17:02
  Speaker 1: first three quote tweets of Kokson's post would all be
• 17:05
  Speaker 1: major AI restriction policy advocates funded generously by the same donor,
• 17:10
  Speaker 1: who also happens to be one of the leading investors
• 17:12
  Speaker 1: in and a board member of Anthropic, the company Kokson
• 17:15
  Speaker 1: was resigning from, and all within 15 minutes of posting?
• 17:19
  Speaker 1: So in other words, this was perhaps not a case
• 17:22
  Speaker 1: where one brave employee decided to blow the whistle at
• 17:26
  Speaker 1: Anthropic in order to inform the public about the risks
• 17:29
  Speaker 1: of AI.
• 17:29
  Speaker 2: It's very clear that.
• 17:31
  Speaker 1: He made this announcement in coordination with powerful interests that
• 17:35
  Speaker 1: are financially tied to Anthropic in a very direct way.
• 17:38
  Speaker 1: In fact, current Anthropic employees immediately boosted this post as well.
• 17:44
  Speaker 1: Here's one of those employees, quote, Jacob's thread is very
• 17:47
  Speaker 1: worth reading. Here's my bird's eye view on the situation
• 17:50
  Speaker 1: with risks from AI. One, AI developers believe their technology
• 17:54
  Speaker 1: could cause human extinction or similarly bad outcomes. This could
• 17:59
  Speaker 1: happen in the next few years. In general, the more
• 18:01
  Speaker 1: senior the employee, the more concerned they are. Two, why
• 18:04
  Speaker 1: do AI developers continue despite the risks due to a
• 18:07
  Speaker 1: mixture of commercial incentives and a belief that they are
• 18:09
  Speaker 1: in a race with other less responsible AI developers that
• 18:13
  Speaker 1: will abuse the technology or develop it less safely? Many
• 18:17
  Speaker 1: AI developer staff desperately want to slow down to figure
• 18:20
  Speaker 1: out how to build AI more safely. Now, whether you
• 18:25
  Speaker 1: agree with what he's saying or not, needs to be
• 18:29
  Speaker 1: said that none of this dramatic language is surprising coming
• 18:31
  Speaker 1: from anthropic employees. They've tended towards sort of neurosis and
• 18:37
  Speaker 1: self-importance for a long time. It's the same company, as
• 18:40
  Speaker 1: we previously discussed, that hired an anti-white female Scottish philosopher
• 18:44
  Speaker 1: named Amanda Askell. and gave her high-level control over their
• 18:48
  Speaker 1: AI model. She's the reason why Claude can tend to
• 18:52
  Speaker 1: browbeat you with woke nonsense. She floats around the Anthropic
• 18:55
  Speaker 1: office telling everyone to be kind to one another. She
• 18:58
  Speaker 1: doesn't have any technical background, but they love her over
• 19:02
  Speaker 1: at Anthropic. This is another Anthropic employee, for what it's worth. Quote,
• 19:07
  Speaker 1: Jacob is correct. We really do earnestly believe AI could
• 19:11
  Speaker 1: kill all humans. I personally think it's greater than 10%
• 19:16
  Speaker 1: Within the next decade, I believe Anthropic is trying its best,
• 19:19
  Speaker 1: but we do not yet have a plan to solve
• 19:21
  Speaker 1: alignment for superintelligence and are not clearly on track to
• 19:25
  Speaker 1: be clear. As we say in our latest risk report,
• 19:27
  Speaker 1: I think the risk from present models is low. What
• 19:30
  Speaker 1: I'm worried about is superintelligence arising from recursive self-improvement, as
• 19:34
  Speaker 1: we've said, is happening faster than we thought. Now, if
• 19:39
  Speaker 1: you read media reports about this whole incident, they'll tell
• 19:41
  Speaker 1: you that these Anthropic employees are lending credibility.
• 19:44
  Speaker 2: To the alleged whistleblower. But you could argue that they're
• 19:47
  Speaker 2: doing the opposite.
• 19:49
  Speaker 1: Another way of looking at it is that they're making
• 19:50
  Speaker 1: the regulatory capture scenario more plausible. They're trying to gain
• 19:55
  Speaker 1: power by causing a panic, which serves two purposes. First,
• 19:58
  Speaker 1: it makes their product look good because it's so powerful.
• 20:02
  Speaker 1: And secondly, it will lead to regulation, which could eliminate
• 20:05
  Speaker 1: their competitors. Anthropic has been playing this card for a
• 20:08
  Speaker 1: long time. Their founder famously told the Pope a few
• 20:10
  Speaker 1: months ago that AI could take everyone's job. So Anthropic
• 20:14
  Speaker 1: has clearly settled on a new variation of the strategy,
• 20:17
  Speaker 1: and they've gotten a lot of buy-in. And it's not
• 20:19
  Speaker 1: just Anthropic investors, employees, and AI nonprofits that are involved.
• 20:23
  Speaker 1: Pretty much the entire Democrat Party lined up within hours
• 20:25
  Speaker 1: to boost this post and to call for immediate regulation.
• 20:30
  Speaker 1: By one count, two governors, seven senators, and 13 representatives
• 20:34
  Speaker 1: responded to Cox's posts with calls to regulate AI, all
• 20:39
  Speaker 1: of them Democrats except for two Republicans. So Democrats immediately
• 20:43
  Speaker 1: latched onto this new AI narrative, presumably because they want
• 20:47
  Speaker 1: to derail every AI company they can control. They have
• 20:49
  Speaker 1: a very clear reason to convince the public that AI
• 20:53
  Speaker 1: is a serious potential threat. And that as a result,
• 20:56
  Speaker 1: we should only allow AIs, you know, AI that's safe
• 20:59
  Speaker 1: and government approved. Notice they don't want to eliminate it entirely.
• 21:03
  Speaker 1: They simply want to stall out the competition and pick
• 21:05
  Speaker 1: a winner that they can control. Here's the socialist Michigan
• 21:11
  Speaker 1: Senate candidate, El.
• 21:12
  Speaker 2: Saeed, for example.
• 21:13
  Speaker 1: And notice how utterly fake this whole video is, but
• 21:16
  Speaker 1: also pay attention to what he says about the ownership
• 21:20
  Speaker 1: of AI profits.
• 21:21
  Speaker 3: Watch.
• 21:22
  Speaker 5: You might have woken up like me to this tweet
• 21:24
  Speaker 5: from a guy you probably never heard of until today.
• 21:26
  Speaker 5: His name is Jacob Coxon. Up until yesterday, he worked
• 21:28
  Speaker 5: for an AI lab called Anthropic, one of the leading
• 21:31
  Speaker 5: edge AI research organizations in the world. He quit yesterday.
• 21:36
  Speaker 5: He quit because he's worried that the technology that he
• 21:38
  Speaker 5: was building could end humanity. Now, like, that's something we
• 21:41
  Speaker 5: should be paying attention to. You might remember a couple
• 21:44
  Speaker 5: months ago, we released a plan, AI Under Democracy. calling
• 21:47
  Speaker 5: for things like an FDA-style regulatory agency to oversee the
• 21:50
  Speaker 5: development of AI, banning AI from doing things like firing
• 21:53
  Speaker 5: you or hiring you or making decisions about your healthcare
• 21:55
  Speaker 5: or firing weapons, making sure that these corporations that are
• 21:59
  Speaker 5: developing the leading edge of AI aren't doing it with
• 22:01
  Speaker 5: the same incentives as the rest of big tech, and
• 22:04
  Speaker 5: making sure that you've got democratically appointed or elected members,
• 22:08
  Speaker 5: more than 50% of every single board, and making possible
• 22:12
  Speaker 5: public ownership of all the wealth that these corporations could produce.
• 22:16
  Speaker 5: And I stand with my colleague Bernie Sanders in calling
• 22:18
  Speaker 5: for an all-out pause. Enough is enough. We need to
• 22:21
  Speaker 5: make sure that the pace of development isn't outpacing the
• 22:24
  Speaker 5: pace of risk. We cannot, in a race against China
• 22:28
  Speaker 5: or a race for these corporations to be the dominant
• 22:31
  Speaker 5: AI manufacturer, inadvertently create products that slip human control, that
• 22:36
  Speaker 5: could fall into the wrong hands, that could mean the
• 22:38
  Speaker 5: demise of humanity. That is a risk none of us
• 22:41
  Speaker 5: ought to be able to bear, and it certainly shouldn't
• 22:43
  Speaker 5: sit with the richest, most powerful corporations in the world
• 22:46
  Speaker 5: on incentives that are driven by their greed rather than
• 22:49
  Speaker 5: our wellbeing. So the time for action is now.
• 22:53
  Speaker 1: So he pretends that he just woke up one day
• 22:56
  Speaker 1: and noticed this tweet from the AI researcher, just like
• 22:59
  Speaker 1: everyone else. And then without missing a beat, he tells
• 23:01
  Speaker 1: us that he wants to redistribute the profits from these
• 23:04
  Speaker 1: AI companies. And he adds that it would be terrible
• 23:07
  Speaker 1: for private corporations to have control over potential doomsday weapons.
• 23:10
  Speaker 1: So obviously we need to put socialist politicians in charge instead. Now,
• 23:14
  Speaker 1: one thing we should note here is that evil profit-driven
• 23:17
  Speaker 1: corporations would seem to have little incentive to usher in
• 23:22
  Speaker 1: the end of humanity, at least intentionally. They would have
• 23:26
  Speaker 1: the exact opposite incentive. They want to keep civilization going
• 23:29
  Speaker 1: so they can continue to make a lot of money.
• 23:30
  Speaker 1: It's a very simple and straightforward problem with people who
• 23:33
  Speaker 1: claim that evil billionaires are going to kill us all
• 23:36
  Speaker 1: or eliminate every job so that no one has any
• 23:38
  Speaker 1: money and we all starve to death. If the corporations
• 23:40
  Speaker 1: eliminate humanity, then the corporations will die, too. And if
• 23:44
  Speaker 1: AI destroys the global economy and unemployment goes to 100
• 23:47
  Speaker 1: percent or something, then money will have no meaning. No
• 23:50
  Speaker 1: one will provide any services. And obviously, the billionaires wouldn't
• 23:53
  Speaker 1: want that outcome either for their own purposes. Wealth has
• 23:57
  Speaker 1: no value in a lawless, apocalyptic hellscape where no one's
• 24:00
  Speaker 1: around to sell you anything. So these points, they don't
• 24:03
  Speaker 1: negate the legitimate concerns about AI, but they do make
• 24:05
  Speaker 1: the AI is going to kill all humans in the
• 24:07
  Speaker 1: next few years scenarios seem ridiculous because they are. It's
• 24:13
  Speaker 1: not even clear how AI, even in a hypothetical scenario,
• 24:16
  Speaker 1: would literally murder the entire global population because that is
• 24:22
  Speaker 1: what these people are saying. It's not a straw man.
• 24:27
  Speaker 1: What would be the practical mechanism for achieving that? and
• 24:32
  Speaker 1: achieving it so quickly? What's the practical mechanism by which
• 24:36
  Speaker 1: the AI, by the year 2030, the end of the decade,
• 24:41
  Speaker 1: will have brought about the extinction of 8 billion people?
• 24:45
  Speaker 1: No one has quite explained that. I want to take
• 24:48
  Speaker 1: a moment here to tell you about one of the
• 24:50
  Speaker 1: companies that supports this show, Tacovas. I've been wearing their
• 24:53
  Speaker 1: boots for years because they're exactly what.
• 24:55
  Speaker 2: I want in a boot.
• 24:56
  Speaker 1: They're simple, understated, Extremely well-made and most important for me,
• 25:00
  Speaker 1: they're comfortable. Comfortable enough that you can actually wear them
• 25:03
  Speaker 1: all the time as opposed to the pair of cowboy
• 25:05
  Speaker 1: boots that spend an entire life sitting in your closet.
• 25:08
  Speaker 1: The Earl is a classic roper boot with a shorter
• 25:10
  Speaker 1: heel and a clean leather design. Tecova's handcrafts their boots
• 25:14
  Speaker 1: in more than 200 steps and the quality is obvious
• 25:17
  Speaker 1: when you put them on. But what I appreciate most
• 25:20
  Speaker 1: is that they don't require some miserable break-in period that
• 25:23
  Speaker 1: a lot of boots do before they become comfortable. Mine
• 25:26
  Speaker 1: felt good right out of the box. They're just a
• 25:28
  Speaker 1: very good pair of boots, and I've genuinely enjoyed wearing them.
• 25:32
  Speaker 1: Right now, get 10% off at Tacobas.com slash Matt. When
• 25:35
  Speaker 1: you sign up for emails and texts, that's 10% off
• 25:38
  Speaker 1: at T-E-C-O-V-A-S.com slash Matt. Tacobas.com slash Matt. Seaside for details.
• 25:44
  Speaker 2: Tacobas. Point your toes west.
• 25:47
  Speaker 1: There are days when we're reminded more forcefully than others
• 25:51
  Speaker 1: that life is precious, that our time here is limited.
• 25:54
  Speaker 1: and that every human life has immeasurable value. That conviction
• 25:57
  Speaker 1: has always been at the heart of what we do
• 25:58
  Speaker 1: on this show, especially when it comes to defending those
• 26:01
  Speaker 1: who cannot defend themselves every day. Women facing unexpected pregnancies
• 26:06
  Speaker 1: are making decisions that will change their lives forever. Preborn
• 26:09
  Speaker 1: Network Clinics meets these mothers with compassion, practical support, including
• 26:13
  Speaker 1: a free ultrasound that allows a mother to see her
• 26:15
  Speaker 1: child and hear that heartbeat. More than 80% of abortion-vulnerable
• 26:19
  Speaker 1: women choose life after seeing their baby on an ultrasound
• 26:22
  Speaker 1: at a Preborn Network Clinic. Right now, another mother needs
• 26:26
  Speaker 1: that moment and you can give it to her. Just
• 26:28
  Speaker 1: $ 28 helps provide one life-saving ultrasound. $ 140 helps rescue five babies.
• 26:34
  Speaker 1: Please give today. Dial pound 250 and say baby. That's
• 26:37
  Speaker 1: pound 250 baby or visit preborn.com slash Walsh. That's preborn.com
• 26:42
  Speaker 1: slash Walsh. If you're a business owner, would you consider
• 26:45
  Speaker 1: a larger donation to claim as a write-off? A donation
• 26:48
  Speaker 1: of $ 1, 000, $ 2, 000, $ 10, 000, $ 15, 000 will sponsor an ultrasound machine
• 26:49
  Speaker 1: for a needy clinic. helping to save countless babies' lives
• 26:56
  Speaker 1: for years to come. All gifts are tax-deductible and will
• 26:59
  Speaker 1: reach eternity.
• 27:01
  Speaker 2: Get involved today.
• 27:04
  Speaker 1: On the other hand, as we've discussed many times in
• 27:05
  Speaker 1: the past few years, communists do want to destroy at
• 27:09
  Speaker 1: least Western civilization. They don't care about profit.
• 27:13
  Speaker 2: They're nihilists.
• 27:14
  Speaker 1: They cheer when health insurance CEOs and prominent conservatives are assassinated.
• 27:18
  Speaker 2: They don't believe in God or any higher purpose whatsoever.
• 27:21
  Speaker 1: They're agents of destruction, which is why the They open
• 27:24
  Speaker 1: the borders, refuse to prosecute criminals, and ruin every creative
• 27:27
  Speaker 1: project they're involved in. So if the only choice is
• 27:31
  Speaker 1: between letting corporations and the free market do whatever they
• 27:33
  Speaker 1: want with AI or giving total top-down control to communist bureaucrats,
• 27:40
  Speaker 1: then there's really no choice at all. You can't hand
• 27:43
  Speaker 1: power over to the communists under any circumstances. History of
• 27:46
  Speaker 1: the 20th century shows us how that turns out. And yet,
• 27:50
  Speaker 1: Letting trillion dollar AI companies do whatever they want with
• 27:54
  Speaker 1: no controls or regulations of any kind is also not
• 27:58
  Speaker 1: a wise path, which is why we should be looking
• 28:01
  Speaker 1: for other options outside of total power to tech companies
• 28:06
  Speaker 1: or total power to communists.
• 28:09
  Speaker 2: But here's the key point, which should be obvious.
• 28:13
  Speaker 1: We also can't allow the government to pick and choose
• 28:15
  Speaker 1: which specific corporations deserve power. One of the biggest risks
• 28:20
  Speaker 1: that AI poses right now is that if leftists dominate
• 28:22
  Speaker 1: the entire industry, then they'll use AI to destroy their
• 28:25
  Speaker 1: political opponents. AI will become a tool for enforcing social
• 28:29
  Speaker 1: credit scores and mass censorship and policing wrong think. As
• 28:33
  Speaker 1: it stands right now, if Anthropic's model wants to censor
• 28:36
  Speaker 1: you for saying that men can't become women, you know,
• 28:39
  Speaker 1: you have other options. You can use other models that
• 28:42
  Speaker 1: aren't completely insane. But if Democrats eliminate those options, They'll
• 28:46
  Speaker 1: be able to control the flow of information just like
• 28:48
  Speaker 1: they do in Canada or North Korea. AI will almost
• 28:51
  Speaker 1: certainly replace search engines in a few years. Leftists know
• 28:55
  Speaker 1: that if they run all the AI models, they'll be
• 28:57
  Speaker 1: able to dictate what information people are able to see.
• 29:00
  Speaker 1: That's why JB Pritzker, the governor of Illinois, who's positioning
• 29:03
  Speaker 1: himself for a run for president, also immediately boosted the
• 29:06
  Speaker 1: alleged whistleblower.
• 29:07
  Speaker 2: And here's what he wrote.
• 29:09
  Speaker 1: Quote, it's time to sound the alarm louder on reigning
• 29:12
  Speaker 1: in artificial intelligence. It's becoming more clear the threat AI
• 29:15
  Speaker 1: poses to humanity. So I'm calling for immediate action from
• 29:17
  Speaker 1: the industry in Washington. The federal government must intervene. I'm
• 29:21
  Speaker 1: proud to have signed the AI Safety Measures Act, landmark
• 29:24
  Speaker 1: legislation requiring more transparency, accountability and oversight of large developers
• 29:29
  Speaker 1: and to protect workers who report concerns. And here's Pramila Jayapal,
• 29:33
  Speaker 1: the communist from Washington state. Quote, researchers for AI companies
• 29:37
  Speaker 1: are sounding the alarm about rapid development of supercomputers. intelligent
• 29:40
  Speaker 1: systems that could hack anything and kill us all. These
• 29:42
  Speaker 1: AI companies are out of control and a serious threat
• 29:45
  Speaker 1: to our future. We need to shut down their dangerous
• 29:47
  Speaker 1: models and push out the executives who are signing off
• 29:49
  Speaker 1: on their creation. Again, all these replies come within hours
• 29:53
  Speaker 1: of the anthropic employee's resignation. Ro Khanna was ready with
• 29:58
  Speaker 1: a video.
• 29:59
  Speaker 2: Watch.
• 30:00
  Speaker 6: Anthropic researcher Jacob Coxon just resigned because he's concerned that
• 30:05
  Speaker 6: the frontier labs are racing towards developing self-improving AI that
• 30:10
  Speaker 6: could threaten our very existence.
• 30:13
  Speaker 7: Now, here's how we know he's not an alarmist. Evan,
• 30:16
  Speaker 7: who is the head of alignment and science at Anthropic,
• 30:21
  Speaker 7: says that Jacob's concerns are valid. In fact, he says
• 30:25
  Speaker 7: that there is a 10% chance that in the next decade,
• 30:29
  Speaker 7: agentic AI could eliminate all of humanity. We need action.
• 30:35
  Speaker 7: Here are five things we must do. First, I've been
• 30:39
  Speaker 7: calling for an AI agency, a federal AI safety agency,
• 30:46
  Speaker 7: just like we have an agency to regulate nuclear power
• 30:50
  Speaker 7: and to regulate our airlines.
• 30:53
  Speaker 1: So this is a guy who holds tens of millions
• 30:55
  Speaker 1: of dollars worth of AI-related stocks. So he's not actually
• 30:59
  Speaker 1: worried about AI, nor does he want to shut it down.
• 31:01
  Speaker 1: He simply wants to make sure that his political party
• 31:04
  Speaker 1: controls the industry. Here's Bernie Sanders, quote, Mr. Kochson is right.
• 31:09
  Speaker 1: The very people building this technology admit that it could
• 31:11
  Speaker 1: threaten the future of humanity. This is why I will
• 31:13
  Speaker 1: soon be introducing legislation to ban superintelligence and pause AI development.
• 31:20
  Speaker 1: He added, quote, Trump's ignorance about AI is appalling and
• 31:23
  Speaker 1: extremely dangerous. We must pause development now. Sanders' legislation on
• 31:27
  Speaker 1: AI would create, quote, a new cabinet-level federal agency to
• 31:31
  Speaker 1: safeguard the public from the dangers of artificial intelligence that
• 31:34
  Speaker 1: will be advised by an artificial intelligence advisory board comprised
• 31:38
  Speaker 1: of experts on artificial intelligence. You know, it sounds like
• 31:42
  Speaker 1: a bureaucratic nightmare because that's exactly what it is. The
• 31:44
  Speaker 1: more layers of bureaucracy you introduce, the more the industry
• 31:47
  Speaker 1: will be sabotaged. for the possible benefit of just one
• 31:51
  Speaker 1: or two companies. And also, the more bureaucracy you involve,
• 31:55
  Speaker 1: there's not any reason to think that that will make
• 31:58
  Speaker 1: a company less dangerous or evil. That's generally not how
• 32:02
  Speaker 1: it works when you add bureaucracy to a thing. And
• 32:06
  Speaker 1: of course, that's the whole point. And it's not just
• 32:08
  Speaker 1: Anthropic that's apparently decided on the strategy. OpenAI just demanded
• 32:11
  Speaker 1: that Congress create mandatory national safety regulations as well, And
• 32:16
  Speaker 1: a former open AI researcher just published this op-ed in
• 32:19
  Speaker 1: New York Times using the same language as Anthropic. Quote,
• 32:23
  Speaker 1: the current pace of AI development is blistering, and yet
• 32:25
  Speaker 1: there is so much more the AI companies could do
• 32:27
  Speaker 1: to relieve imminent danger and to lay bare the gambles
• 32:30
  Speaker 1: they're taking with all of our futures. The window to
• 32:33
  Speaker 1: ensure we are heading to a bright future and away
• 32:35
  Speaker 1: from catastrophe is closing fast. After a while, it gets
• 32:40
  Speaker 1: reminiscent of the climate alarmism that used to be popular
• 32:43
  Speaker 1: on the left. These kinds of predictions have been coming
• 32:45
  Speaker 1: for a while now. A few years ago, a former
• 32:47
  Speaker 1: Google engineer named Blake Lemoine claimed that the company's AI
• 32:51
  Speaker 1: chatbot had become sentient, which turned out to be nonsense.
• 32:55
  Speaker 1: Scientific American even wrote a blog, wrote a big article
• 32:58
  Speaker 1: about it. Quote, Lemoine said he considers Lambda to be
• 33:03
  Speaker 1: his colleague and a person, even if not a human.
• 33:07
  Speaker 1: And he insists that it has a right to be recognized,
• 33:10
  Speaker 1: so much so that he has been the go-between in
• 33:13
  Speaker 1: connecting the algorithm with a lawyer. So in the end,
• 33:17
  Speaker 1: the guy was just kind of deranged. You also might
• 33:19
  Speaker 1: remember this prediction from Jeffrey Hinton, the so-called godfather of AI.
• 33:23
  Speaker 1: In 2016, he predicted that AI would render radiologists obsolete
• 33:27
  Speaker 1: within five years because AI would be able to read
• 33:30
  Speaker 1: the scans for them. That turned out to be false.
• 33:32
  Speaker 1: There are more radiologists now than ever, actually. They're just
• 33:36
  Speaker 1: using AI to be more productive. But here's Hinton explaining
• 33:40
  Speaker 1: where his reasoning went wrong.
• 33:41
  Speaker 2: Watch.
• 33:42
  Speaker 8: The first is that healthcare is elastic.
• 33:46
  Speaker 4: So.
• 33:47
  Speaker 8: If you could do more scans and get more scans read,
• 33:50
  Speaker 8: there'd be a lot more scans happening. That's one thing
• 33:53
  Speaker 8: that's happening. A significant fraction of the cost of doing
• 33:59
  Speaker 8: a scan is the cost of the radiologist interpreting it.
• 34:03
  Speaker 8: As AI gets to help more and more radiologists interpret scans,
• 34:06
  Speaker 8: we can interpret them faster and faster for less and
• 34:09
  Speaker 8: less money.
• 34:09
  Speaker 9: They're getting more efficient. And you would have thought that
• 34:13
  Speaker 9: would mean you needed less radiologists. But actually, what it
• 34:16
  Speaker 9: means is you get more scans. So that aspect of
• 34:19
  Speaker 9: the prediction was wrong. A second thing that was wrong
• 34:22
  Speaker 9: was I didn't know enough about radiologists and what they do.
• 34:25
  Speaker 9: And that was because I had a former student who
• 34:28
  Speaker 9: had an MD and then did a physics PhD with
• 34:32
  Speaker 9: me on something called Boltzmann machines. And he didn't particularly
• 34:36
  Speaker 9: like people. So he got a job as a radiologist,
• 34:40
  Speaker 9: just interpreting scans. And he was my model for radiologists.
• 34:43
  Speaker 9: All he did was interpret scans. He never talked to people.
• 34:48
  Speaker 9: And that's what was going to get replaced. And that
• 34:50
  Speaker 9: is now becoming replaced. So I think there's now of
• 34:54
  Speaker 9: the order of 100 AI systems for interpreting scans that
• 34:58
  Speaker 9: have been federally approved.
• 35:00
  Speaker 1: Now, it's important to keep predictions like this in mind
• 35:02
  Speaker 1: when you're talking about the future of AI.
• 35:04
  Speaker 2: It's very easy.
• 35:05
  Speaker 1: Sometimes it's politically or financially beneficial to conclude that a
• 35:09
  Speaker 1: new technology is going to cause.
• 35:10
  Speaker 2: Some extreme outcome.
• 35:12
  Speaker 1: It's easy to say that AI is going to destroy
• 35:13
  Speaker 1: everyone's life or that it's going to provide endless prosperity.
• 35:17
  Speaker 1: On the other end, with no middle ground, the reality
• 35:20
  Speaker 1: as usual is probably somewhere in the middle. And it's
• 35:24
  Speaker 1: frankly ridiculous for so many people, including every mainstream media outlet,
• 35:28
  Speaker 1: to claim otherwise. This is the entire problem with the
• 35:31
  Speaker 1: conversation on AI. It's generally divided into two camps, both
• 35:35
  Speaker 1: of which are cartoon caricatures of reality. One side only
• 35:40
  Speaker 1: acknowledges the upside. The other tells us that we're doomed,
• 35:43
  Speaker 1: humanity's over, it's an apocalypse. And both of these takes
• 35:47
  Speaker 1: are obviously ridiculously simplistic. And the more that the AI
• 35:51
  Speaker 1: skeptic side, which is the side that I've always leaned towards,
• 35:55
  Speaker 1: focuses on literal apocalypse scenarios instead of the more realistic
• 36:00
  Speaker 1: and practical dangers it faces, the worse and dumber the
• 36:04
  Speaker 1: debate will get. And more importantly, we won't actually succeed
• 36:08
  Speaker 1: in putting any reasonable controls or guide rails in place.
• 36:14
  Speaker 1: So AI has tremendous promise. There's no question about that.
• 36:18
  Speaker 1: It is, and anyone should be able to acknowledge this,
• 36:20
  Speaker 1: an obvious productivity booster so far. Because of AI, it's
• 36:26
  Speaker 1: easier to do research. It's easier to process data. It's
• 36:28
  Speaker 1: easier to target advertisements, diagnose medical conditions, and much more.
• 36:34
  Speaker 1: But if it increases productivity for the good guys, it
• 36:36
  Speaker 1: also increases productivity for the bad guys. And when people
• 36:39
  Speaker 1: talk about AI killing us all, that's often what they mean,
• 36:44
  Speaker 1: we can assume. Last April, Anthropic paused development in one
• 36:48
  Speaker 1: of its cloud models because it was finding previously unknown security.
• 36:52
  Speaker 2: Vulnerabilities in software.
• 36:54
  Speaker 1: So what happens when AI can suddenly hack into financial institutions?
• 36:59
  Speaker 1: What about biological and chemical research facilities or the armed forces?
• 37:05
  Speaker 1: The Bulletin of Atomic Scientists claims advanced AI systems could
• 37:08
  Speaker 1: lower barriers to creating harmful biological agents by providing guidance
• 37:13
  Speaker 1: to users with limited expertise.
• 37:16
  Speaker 2: Is that true?
• 37:18
  Speaker 1: Seems like it certainly could be. That's not an unreasonable
• 37:22
  Speaker 1: thing to be worried about here. But even if it's not,
• 37:25
  Speaker 1: you know, a nuclear holocaust, what happens when it can
• 37:28
  Speaker 1: be used to hack into anyone's bank account or email
• 37:32
  Speaker 1: or anything else? When should we be concerned about AI's
• 37:37
  Speaker 1: implications for data privacy? Apple and Google are already using
• 37:40
  Speaker 1: AI to screen your inboxes, text messages, voicemails. How much
• 37:44
  Speaker 1: powered information can these companies consolidate before it's too much?
• 37:50
  Speaker 1: For decades, digital marketing companies and data brokers have collected
• 37:54
  Speaker 1: granular information on everyone. They can accurately predict when you
• 37:57
  Speaker 1: have to pee or when your wife is pregnant. That's
• 38:00
  Speaker 1: not science fiction. It's also not fiction to acknowledge the
• 38:02
  Speaker 1: risks to white-collar workers, most.
• 38:05
  Speaker 2: Of whom do very little already.
• 38:07
  Speaker 1: Maybe the day-in-my-life tech girls who spend their day in
• 38:09
  Speaker 1: endless meetings drinking mocha lattes deserve to lose their jobs, but.
• 38:13
  Speaker 2: What about the other white-collar professions?
• 38:16
  Speaker 1: An estimated 165,000 people have lost their jobs because of
• 38:19
  Speaker 1: AI just this year.
• 38:20
  Speaker 2: What about truck drivers, cab drivers?
• 38:25
  Speaker 1: The more this technology advances and scales, and again, it's
• 38:27
  Speaker 1: doing both of those things arguably faster than any technology
• 38:31
  Speaker 1: ever has, the more jobs.
• 38:34
  Speaker 2: Get swallowed up by it.
• 38:37
  Speaker 1: And all of this is related to the core risk
• 38:39
  Speaker 1: of artificial intelligence, which is the thing that some of
• 38:42
  Speaker 1: these so-called whistleblowers have brought up.
• 38:46
  Speaker 2: They often bring up and it's true. And that is
• 38:49
  Speaker 2: the risk of misalignment.
• 38:51
  Speaker 1: So what happens if AI systems develop goals that are
• 38:55
  Speaker 1: not aligned with human interests? That's not some kind of
• 39:00
  Speaker 1: outlandish James Cameron fantasy. It's hard to see how that's
• 39:06
  Speaker 1: a scenario that leads to the extinction of mankind in
• 39:08
  Speaker 1: the next couple of years, but it's a real thing.
• 39:12
  Speaker 1: These are trillion dollar companies pouring all of their resources
• 39:16
  Speaker 1: and brain power into developing an artificial mind. There are
• 39:20
  Speaker 1: massive implications to that. What we need are serious people
• 39:25
  Speaker 1: we can trust.
• 39:28
  Speaker 2: So not Democrats.
• 39:29
  Speaker 1: To start taking this seriously and to, yes, put some
• 39:34
  Speaker 1: controls in place or try to. But right now it
• 39:37
  Speaker 1: seems to be a debate between Bernie Sanders types who
• 39:40
  Speaker 1: want total control over it and generally conservatives who want
• 39:45
  Speaker 1: to do nothing.
• 39:46
  Speaker 2: As if there are.
• 39:47
  Speaker 1: Zero serious risks posed by AI at all, which is
• 39:51
  Speaker 1: also a total fantasy, by the way. It's a false
• 39:54
  Speaker 1: dichotomy and it fuels the kind of PR campaigns we're
• 39:57
  Speaker 1: seeing online.
• 39:58
  Speaker 2: And in the media right now. And there's one other
• 40:00
  Speaker 2: thing to keep.
• 40:01
  Speaker 1: In mind here that I referenced earlier. I've laid out
• 40:08
  Speaker 1: the case that some of the most hysterical doomerism from
• 40:12
  Speaker 1: the AI companies themselves is being spread for manipulative self-serving reasons.
• 40:18
  Speaker 1: And I believe that that's likely true. It's also what
• 40:22
  Speaker 1: the side that is far more pro-AI than I am
• 40:25
  Speaker 1: will say.
• 40:27
  Speaker 2: But here's the thing.
• 40:28
  Speaker 1: If you're on that side, just consider one thing. If
• 40:32
  Speaker 1: a company is willing to literally threaten global extinction for
• 40:37
  Speaker 1: the sake of expanding their market share, then even if
• 40:41
  Speaker 1: it can't actually bring about global extinction and wouldn't want
• 40:45
  Speaker 1: to for the reasons I said, that still tells us
• 40:48
  Speaker 1: that this is a potentially dangerous company, right? Maybe not
• 40:53
  Speaker 1: annihilate the earth dangerous, but dangerous. Like it doesn't make
• 40:58
  Speaker 1: any sense to wave the warning away and say, no,
• 41:01
  Speaker 1: they're only saying they can kill all humans.
• 41:04
  Speaker 2: They're actually totally harmless.
• 41:08
  Speaker 1: Well, that doesn't work because if they're saying they can
• 41:10
  Speaker 1: kill all humans, and if they're saying that just to
• 41:13
  Speaker 1: chase profits, then by definition, they're not harmless.
• 41:19
  Speaker 2: That would make them at.
• 41:21
  Speaker 1: A minimum extraordinarily irresponsible, manipulative people controlling extremely powerful technology.
• 41:31
  Speaker 1: I mean, you can't have it both ways, which again
• 41:37
  Speaker 1: speaks to the need to get responsible people into the room.
• 41:42
  Speaker 1: So not Bernie Sanders or Pramila Jayapal So the takeaway
• 41:48
  Speaker 1: here is that whether you're an AI skeptic or not,
• 41:52
  Speaker 1: you are being manipulated on a massive scale. There are
• 41:57
  Speaker 1: extremely well-funded, coordinated efforts behind the scenes to influence your
• 42:01
  Speaker 1: opinion on AI, both for it and against it. Anything
• 42:06
  Speaker 1: you read on the topic should be treated with the
• 42:09
  Speaker 1: highest possible degree of skepticism. There's an enormous amount of
• 42:13
  Speaker 1: money at stake for people on both sides. It's now
• 42:19
  Speaker 1: abundantly clear that the mainstream press is trying to swallow
• 42:22
  Speaker 1: up and regurgitate any narrative they're fed when it comes
• 42:25
  Speaker 1: to AI. So the worst mistake you could make, the
• 42:30
  Speaker 1: one that might actually lead to our downfall, is to
• 42:34
  Speaker 1: take any of it at face value.
• 42:38
  Speaker 2: That'll do it for the show today and this week.
• 42:40
  Speaker 2: Talk to you next week. Have a great weekend. Godspeed.
• 42:50
  Speaker 2: Every significant movement in history has started with protesting.
• 42:53
  Speaker 8: You're a traitor!
• 42:53
  Speaker 2: Uncle Tom!
• 42:54
  Speaker 8: This is our generation's civil rights! I have no idea
• 42:57
  Speaker 8: what the protest is about.
• 42:58
  Speaker 2: Bingo!
• 42:58
  Speaker 1: Given the current climate, should a figure as polarizing as
• 43:02
  Speaker 1: Clay Saunders even be allowed on campus?
• 43:04
  Speaker 2: Uh, can you hear me okay? Saunders is a close
• 43:06
  Speaker 2: personal friend of Chancellor Freeman.
• 43:07
  Speaker 8: This is a decision I cannot fight.
• 43:09
  Speaker 2: Islam will conquer the West. As-salamu alaykum. Are we paying attention?
• 43:18
  Speaker 2: They've taken over the building. They're downstairs and they've blocked
• 43:20
  Speaker 2: the doors.
• 43:22
  Speaker 9: I seek forgiveness from Allah the Almighty.
• 43:26
  Speaker 2: We cannot run and hide.
• 43:28
  Speaker 7: Muhammad is the messenger of God.
• 43:32
  Speaker 8: There is no God but Allah.
• 43:35
  Speaker 4: What do you say we turn those SOBs.
• 43:39
  Speaker 2: Into a fine red mist?
• 43:41
  Speaker 3: As we move, one man will cover while the other
• 43:44
  Speaker 3: guys go.
• 43:44
  Speaker 8: They call that leap project.
• 43:45
  Speaker 5: Stretch them out and listen. You need a course in
• 43:49
  Speaker 5: safe weapons handling.