# Doing Goon Things

An episode of [Armstrong & Getty On Demand](https://invisiblenetwork.ai/podcast/1259438987), 2 Oct 2026, 38 min.

Audio: https://podtrac.com/pts/redirect.mp3/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/0516ff28-c0d6-492a-b264-ae3900375fc8/499485c2-d15e-4ab6-8f22-b4d7010bcd92/audio.mp3?utm_source=Podcast&in_playlist=4db37684-c7ed-4964-843c-ae3900375fd7

Hour 3 of A&G features... • Idiot woke eating their own! • Iranian arrested for RAF Airport incident • Anthropic believes chatbots have feelings • The people running Anthropic are nuts Stupid Should Hurt: https://www.armstrongandgetty.com/ See omnystudio.com/listener for privacy information.

## Transcript

[0:01] Speaker 1: Broadcasting live from the Abraham Lincoln Radio Studio at the
[0:05] Speaker 1: George Washington Broadcast Center.
[0:07] Speaker 2: Jack Armstrong and Joe Getty. Armstrong and Getty.
[0:11] Speaker 3: And now, here's Armstrong and Getty.
[0:34] Speaker 4: So we all heard about the botched execution of Krista
[0:38] Speaker 4: Pike in Tennessee the other day. It strikes me she's
[0:42] Speaker 4: still alive by all accounts.
[0:44] Speaker 2: You know, almost everybody has said the same thing. Canada
[0:48] Speaker 2: kills people by the thousands. And obviously they all think
[0:53] Speaker 2: it is both safe, effective, and humane as they are
[0:58] Speaker 2: volunteering to To be snuffed that way. And they all
[1:03] Speaker 2: voted for it and everything else. So clearly there is
[1:06] Speaker 2: a safe, effective, kind, I don't know why you need
[1:11] Speaker 2: to be kind, but kind way to snuff people. They
[1:15] Speaker 2: do it around the world, various countries, and we can't
[1:18] Speaker 2: do it.
[1:19] Speaker 4: Brilliantly anticipating my next point, which was going to be
[1:21] Speaker 4: virtually everybody quoted in every article in every news story
[1:25] Speaker 4: about this story is an anti-death penalty activist.
[1:28] Speaker 3: Of course.
[1:28] Speaker 2: Which is fine. It is a noble cause, whether you
[1:31] Speaker 2: agree with it or not. I think these people come
[1:33] Speaker 2: at it with.
[1:35] Speaker 3: sincerity.
[1:35] Speaker 2: I've never thought it was a noble cause. If you're
[1:37] Speaker 2: going to pour your heart into a cause, why this one?
[1:41] Speaker 2: Why making sure an axe murderer of a family doesn't
[1:44] Speaker 2: get snuffed? Of all the causes. Yeah, I know, because
[1:48] Speaker 2: the government shouldn't be empowered to kill people.
[1:50] Speaker 3: I don't know.
[1:50] Speaker 4: Like I said, I didn't say I agreed with them.
[1:52] Speaker 4: I think they're sincere. Anyway, I definitely think they're sincere.
[1:55] Speaker 5: Right.
[1:56] Speaker 3: So they're not making anything from this, really.
[1:59] Speaker 4: Now, the attorneys who make it impossible to eliminate the
[2:02] Speaker 4: death penalty because they're getting rich off of it, even though,
[2:04] Speaker 4: for instance, California never executes somebody, they're just effing greed heads.
[2:09] Speaker 2: Super cynical.
[2:11] Speaker 3: Anyway, sorry, I get a little passionate on this stuff.
[2:13] Speaker 4: But anyway, so everybody being quoted on this story is
[2:16] Speaker 4: an anti-death penalty activist, including the people who've said that
[2:19] Speaker 4: this gal has probably got brain damage and and the
[2:23] Speaker 4: rest of it. Although I did find 53 interesting, Michael.
[2:28] Speaker 3: This is a.
[2:29] Speaker 2: Completely unprecedented event.
[2:31] Speaker 1: Krista Pike is the only person we know of who
[2:34] Speaker 1: survived the administration of lethal doses of injection drugs.
[2:40] Speaker 3: Yeah, yeah.
[2:41] Speaker 4: Gave it two tries, the old college try. Apparently she's
[2:46] Speaker 4: claimed life.
[2:46] Speaker 2: They stuck her like seven times or something like that?
[2:49] Speaker 2: That's what I would hate.
[2:50] Speaker 6: Oh, boy.
[2:51] Speaker 4: Can I get a new phlebotomist? Oh, trust me, there's
[2:54] Speaker 4: more on that. Stay tuned or don't. I'll go easy
[2:57] Speaker 4: on that stuff because nobody wants to hear it. But
[3:00] Speaker 4: give us 52. This is from CNN.
[3:03] Speaker 5: The number of executions by method since 1977, about 1,500
[3:10] Speaker 5: have been by lethal injection, 163 by electrocution, 19 by gas,
[3:15] Speaker 5: 6 by firing squad, 3 by hanging.
[3:18] Speaker 2: And what happened to the gas chamber? When we were young,
[3:21] Speaker 2: that was the big one, wasn't it?
[3:22] Speaker 4: No, people gasp and they cough and they're afraid and stuff.
[3:25] Speaker 3: So that's cruel and unusual. Got to get rid of that.
[3:27] Speaker 1: Yeah.
[3:28] Speaker 3: Yeah.
[3:29] Speaker 4: So the, and I mentioned this yesterday, but it's worth repeating,
[3:34] Speaker 4: the reason that the, you know, the conundrum, the contrast
[3:38] Speaker 4: of Canada versus botching that attempt, the reason that happens
[3:42] Speaker 4: is because the death penalty opponents have made it so
[3:44] Speaker 4: difficult to do the stuff that works. They've found a
[3:48] Speaker 4: variety of maneuvers to pressure drug companies or doctors or
[3:51] Speaker 4: whatever into, like, forcing the state into alternative methods. That
[3:57] Speaker 4: aren't as effective or a little more difficult, blah, blah, blah.
[4:00] Speaker 3: Anyway, leave it to the New York Post.
[4:02] Speaker 2: What's the goal of making it so difficult that we
[4:05] Speaker 2: give up on it?
[4:06] Speaker 3: Making it impracticable, as they say. Exactly, yeah.
[4:11] Speaker 4: So leave it to the New York Post to write
[4:13] Speaker 4: a delightful article about the history of.
[4:16] Speaker 3: Botched executions in the United States.
[4:18] Speaker 2: Oh, oh.
[4:19] Speaker 4: Krista Pike became the first death row inmate to survive
[4:22] Speaker 4: a lethal injection, but she was merely the latest in
[4:24] Speaker 4: a long, gruesome history of botched executions. And yes, friends, yes, I,
[4:30] Speaker 4: your old Uncle Joe, will exercise some restraint, more restraint
[4:35] Speaker 4: than the New York Post does in this article. Roughly 3%
[4:39] Speaker 4: of all U.S. executions in the country between 1890 and 2010,
[4:44] Speaker 4: that's a pretty good sample size, about 3% were botched. Again,
[4:48] Speaker 4: according to the Death Penalty Information Center that crusade against
[4:51] Speaker 4: this sort of thing. The highest number at 85 was
[4:55] Speaker 4: for bungled hangings.
[4:58] Speaker 3: I like the term bungled.
[4:59] Speaker 2: It's good word. Many use it more often.
[5:02] Speaker 4: Yeah.
[5:02] Speaker 2: Well, you bungled that.
[5:04] Speaker 3: Boy, you bungled your homework, son. Look at that. Four
[5:06] Speaker 3: out of ten.
[5:08] Speaker 4: Including the legendary 1901 hanging of outlaw Tom Blackjack Ketchum,
[5:13] Speaker 4: whose head popped clean off when he dropped because executioners
[5:16] Speaker 4: failed to account for how.
[5:18] Speaker 3: Fat he had gotten in jail.
[5:20] Speaker 2: And when we discussed this a little earlier, I don't
[5:22] Speaker 2: mind my head popping off if I'm executed. Some of
[5:24] Speaker 2: these stories are going to be gruesome and think, oh, God,
[5:26] Speaker 2: I can't imagine going through that. But your head popping off,
[5:29] Speaker 2: you died instantly. It's like when Saddam Hussein's head popped
[5:33] Speaker 2: off when we hung him, right?
[5:35] Speaker 3: Yes, yeah.
[5:36] Speaker 4: And to me, if like part of the professionalization of
[5:39] Speaker 4: being a hangman is you calculate the weight, the distance
[5:43] Speaker 4: of the drop, and you say, and you know, there
[5:46] Speaker 4: in the hangman's code, it said, and it's highly regular
[5:50] Speaker 4: and sought to be avoided that the.
[5:52] Speaker 3: Executionee, what's the term? Convict, whatever. Scumbag.
[5:56] Speaker 4: That their head should pop off. And you know, you
[5:59] Speaker 4: signed a contract. If the head pops off, you're guilty
[6:02] Speaker 4: of violating your professional standards.
[6:03] Speaker 2: But as a.
[6:04] Speaker 3: Society, I don't think it really affects it. To me,
[6:06] Speaker 3: that's not botched.
[6:07] Speaker 2: No. It's unattractive, certainly. But was the job done or
[6:14] Speaker 2: wasn't it done? I wouldn't take my kids to watch it.
[6:17] Speaker 4: Likewise, and this is edging closer to the sort of
[6:20] Speaker 4: thing I'm not comfortable airing, a number of cases involving
[6:25] Speaker 4: the electric chair. The guilty may or may not have
[6:32] Speaker 4: had their act. heads, what's the expression, burst into flame.
[6:38] Speaker 2: Why is that? Because it's too hot? I thought you
[6:40] Speaker 2: were going to go with the slow-cooked. Those are always
[6:42] Speaker 2: the ones that bother me. It's like you've got a
[6:44] Speaker 2: pot roast in the crock pot all day long. Oh, no.
[6:47] Speaker 3: Make him stop. What the hell?
[6:50] Speaker 2: I'm trying to exercise a little restraint over here, and
[6:53] Speaker 2: you're still.
[6:54] Speaker 3: No, this is not good.
[6:56] Speaker 2: Similar to the hanging where your head, where your noggin
[6:58] Speaker 2: comes unattached from your shoulders. I'd rather have my head
[7:02] Speaker 2: burst into flames than the slow-cooked, in terms of steak.
[7:05] Speaker 7: Oh.
[7:06] Speaker 1: Oh, yeah.
[7:06] Speaker 3: Well, if my head is on fire postmortem, what the
[7:09] Speaker 3: hell do I care? I'm sorry.
[7:10] Speaker 2: My wife was running her hair dryer at the same time.
[7:12] Speaker 2: We just didn't have enough juice.
[7:15] Speaker 3: My damn kids playing their video games. Oh, my God. Oh,
[7:20] Speaker 3: I'm not even going to tell you this one. There
[7:22] Speaker 3: is a litany. Well, it's not a litany.
[7:25] Speaker 4: It's a small number of truly unfortunate efforts to get
[7:29] Speaker 4: lethal injection going.
[7:31] Speaker 3: Oh, my God.
[7:31] Speaker 2: But you can picture the details. What's the main problem?
[7:35] Speaker 5: Something.
[7:37] Speaker 4: You got one instance where somebody didn't hit a vein,
[7:40] Speaker 4: they hit an artery.
[7:44] Speaker 2: Where did you go to school?
[7:46] Speaker 3: No kidding.
[7:47] Speaker 4: And the second one, and I will leave it at this,
[7:50] Speaker 4: and I beg you, sir, I beg you to join me.
[7:53] Speaker 3: In my restraint. I will credit the medical staff....
[8:00] Speaker 2: For really sticking with it, even though they failed multiple times. Like,
[8:06] Speaker 2: multiple times. Let's try again, they said. Over and over again. Like,
[8:13] Speaker 2: how many times?
[8:16] Speaker 3: Let's see....
[8:20] Speaker 4: They don't give a number, but in this one case,
[8:24] Speaker 4: they tried for 90 minutes.
[8:26] Speaker 2: Well, like, see, again, how does this happen? I've had
[8:31] Speaker 2: a lot of medical work done when I was going
[8:33] Speaker 2: through cancer. I got stuck all the time. And I've
[8:36] Speaker 2: never had, I think three is the max once. Usually
[8:41] Speaker 2: it's once. occasionally twice. Right, three is extraordinary. Yeah, and
[8:46] Speaker 2: I'm not even sure I've ever even had a three.
[8:48] Speaker 2: So why do you have all these situations where they
[8:50] Speaker 2: got to do it for an hour and a half?
[8:52] Speaker 2: What the hell's going on there?
[8:53] Speaker 3: That's what I want to know. Yeah, yeah.
[8:56] Speaker 4: All right, well, moving on from this delightful fare, it's
[9:01] Speaker 4: a brand new featurette, but I don't know the name.
[9:03] Speaker 4: I gave Hanson a little latitude in producing the theme song.
[9:08] Speaker 4: So right after a word from our friends at Incogni,
[9:11] Speaker 4: We will introduce our brand-new featurette.
[9:14] Speaker 3: The good people at Incognito. Here's what they do.
[9:17] Speaker 4: There are at least 420 or so data brokers out
[9:20] Speaker 4: there that have every bit of information about you from
[9:23] Speaker 4: where you live to what your kids' names are to
[9:25] Speaker 4: how you vote to how often you're late on credit
[9:29] Speaker 4: card payments, right? And they sell all that stuff all
[9:32] Speaker 4: over the place. to folks who want to violate your privacy.
[9:35] Speaker 4: Incogni goes and gets that stuff the.
[9:37] Speaker 3: Hell off those sites. Yep.
[9:38] Speaker 2: Right now, somewhere a stranger is buying a list of
[9:41] Speaker 2: all of your details, like your name, your phone, your
[9:44] Speaker 2: home address, your age, your income, your health problems, parts
[9:46] Speaker 2: of your life, all packaged and sold, sometimes really, really cheap.
[9:51] Speaker 2: And you can get that off of the data broker
[9:54] Speaker 2: with the help of Incogni so that they can't be
[9:56] Speaker 2: selling this stuff all the time.
[9:57] Speaker 4: And then they keep following up so you don't get
[9:59] Speaker 4: quietly retweeted.
[10:00] Speaker 3: Good service.
[10:00] Speaker 4: Go to Incogni.com slash Armstrong for 60% off your private life.
[10:04] Speaker 4: Should never be somebody else's prospect list. Incogni.com slash Armstrong.
[10:09] Speaker 4: They can't harm you if they can't find you. Incogni.com
[10:12] Speaker 4: slash Armstrong.
[10:13] Speaker 3: All right.
[10:13] Speaker 2: Again, the actual name of the feature is unknown to
[10:17] Speaker 2: me until I hear this.
[10:20] Speaker 7: Wokester jackasses eating their own.
[10:30] Speaker 2: Wokester jackasses eating their own.
[10:33] Speaker 3: Exactly. Yeah. It's a little longer than I'd pictured. Can
[10:37] Speaker 3: I hear it again, Michael?
[10:39] Speaker 7: Wokester jackasses eating their own.
[10:49] Speaker 3: That's pretty good. Yeah.
[10:51] Speaker 2: And it really tells you what's in store for you. Indeed.
[10:55] Speaker 2: Example number one, a Baltimore opera group has recently, I'm sorry,
[11:00] Speaker 2: it has ousted its recently hired executive director. Just hired
[11:07] Speaker 2: her a few days ago, canned her for what?
[11:11] Speaker 3: What?
[11:12] Speaker 4: She was going to a music festival. I'm sorry, she
[11:16] Speaker 4: was going to Burning Man. Legendary for what? Freaky behavior
[11:20] Speaker 4: and dust. And she thought, how do I keep my
[11:23] Speaker 4: hair from becoming completely clogged with dust? I know, I'll
[11:27] Speaker 4: have it braided tight into cornrows. But that's cultural appropriation.
[11:33] Speaker 4: She's a white gal. Quote, I believe the circumstance, oh, oh,
[11:38] Speaker 4: I'm sorry, that's the wrong quote. Blah, blah, blah, blah, blah, blah, blah, blah,
[11:41] Speaker 4: something cultural appropriation. Ah, regarding a video circulating showing our
[11:45] Speaker 4: new executive director is white wearing cornrows, that is clearly
[11:49] Speaker 4: cultural appropriation, insensitive.
[11:52] Speaker 2: You can't give in to these people.
[11:54] Speaker 3: Echoes white supremacy.
[11:55] Speaker 2: Oh, my God, you can't give in to these people
[11:57] Speaker 2: and apologize for something as mundane as that.
[12:00] Speaker 3: Well, she's fired.
[12:00] Speaker 2: There's no apologizing to do. She got canned.
[12:03] Speaker 4: She's a woke jackass being eaten by her own. But
[12:07] Speaker 4: that's not even my favorite example, although my heart bleeds
[12:10] Speaker 4: for this poor son of a gun.
[12:12] Speaker 2: The Portland, of course, trailblazers. Just hired a new play-by-play guy,
[12:18] Speaker 2: Braden Bell. Young man, dream job. I know guys in sports.
[12:25] Speaker 2: They dream of having a major league play-by-play job. Hard
[12:28] Speaker 2: to get those jobs. There aren't a lot of them.
[12:30] Speaker 3: He's 29 years old, okay? 13 and 14 years ago.
[12:40] Speaker 4: He made media posts that contained comments about quote-unquote Asian
[12:45] Speaker 4: people in a way that was insensitive and or used
[12:50] Speaker 4: the term slut and whores to refer to certain women
[12:55] Speaker 4: when he was 14 and 15 years old and the
[12:59] Speaker 4: Portlandia Trailblazers have fired him before he could start his job.
[13:05] Speaker 4: The Portland Trailblazers have zero tolerance for racist, sexist, or
[13:08] Speaker 4: harmful comments. Social media posts made by a recently announced
[13:11] Speaker 4: play-by-play broadcaster are in clear violation of team and league
[13:14] Speaker 4: policies as well as our values and expectations.
[13:16] Speaker 3: Therefore, we will not be moving forward with this broadcaster.
[13:18] Speaker 2: Well, we're going to be dealing with this for a while,
[13:21] Speaker 2: and we're going to have to figure it out as
[13:22] Speaker 2: a society because we're soon to have everybody who's out
[13:28] Speaker 2: there in a position of any power whatsoever has a
[13:32] Speaker 2: social media background going into their early teen years. I mean,
[13:36] Speaker 2: soon it's everyone.
[13:38] Speaker 4: And soon the woke crowd, I pray, will be so
[13:43] Speaker 4: in disrepute, so disliked, so hated because they're all Marxists.
[13:48] Speaker 4: Trying to overturn Western civilization.
[13:50] Speaker 7: God.
[13:51] Speaker 4: And they're just lusting for power. Has nothing to do
[13:54] Speaker 4: with what a young man said when he was 14 friggin'
[13:56] Speaker 4: years old.
[13:57] Speaker 3: How could it possibly?
[13:58] Speaker 2: I can't imagine the things I would have to answer
[14:01] Speaker 2: for if I'd had a Twitter account starting when I
[14:04] Speaker 2: was like 13.
[14:05] Speaker 3: And this stuff is innocuous. It's teenage stupidity.
[14:09] Speaker 4: You know, Portland Trailblazers, you should be ashamed of yourselves.
[14:13] Speaker 4: You think you're virtuous and you're demonstrating morality? You're demonstrating
[14:18] Speaker 4: exactly the opposite. Shame on you. Bunch of wokester jackasses
[14:23] Speaker 4: eating their own.
[14:24] Speaker 7: Wokester jackasses eating their own.
[14:30] Speaker 2: I feel like the theme kind of undermined the tone
[14:33] Speaker 2: of outrage I was going for. Coming up, AI meets
[14:39] Speaker 2: with religious leaders. Why? That's a big story in the
[14:42] Speaker 2: New York Times today. And other stuff. Stay here. Armstrong
[14:46] Speaker 2: and Getty.
[14:48] Speaker 6: The judge in the Lindsay Clancy murder case has denied
[14:51] Speaker 6: her request to find her not guilty.
[14:54] Speaker 2: Clancy's defense attorney.
[14:55] Speaker 6: Had argued earlier this week that there wasn't enough evidence
[14:58] Speaker 6: to even prove that Clancy had killed her three children.
[15:02] Speaker 6: The judge today refuting that, saying there was enough evidence.
[15:05] Speaker 6: The district attorney, though, still has not decided whether or
[15:08] Speaker 6: not to retry Clancy.
[15:09] Speaker 2: What do you suppose is going on there? Why haven't
[15:12] Speaker 2: they made a decision to retry her?
[15:15] Speaker 3: Oh, they got away a lot of different stuff. It's
[15:17] Speaker 3: a big decision.
[15:18] Speaker 4: Given the 11 to 1 jury numbers for acquittal, they
[15:24] Speaker 4: got to figure out. They're holding meetings of the brain
[15:27] Speaker 4: trust probably right now saying, all right, how do we
[15:30] Speaker 4: approach the case differently?
[15:32] Speaker 3: What could we have done better? Just all sorts of
[15:35] Speaker 3: stuff like that.
[15:36] Speaker 2: Here's one of the most confusing recent stories. They arrest
[15:40] Speaker 2: five terrorists in Great Britain. who seem to have a
[15:46] Speaker 2: plot against one of the air bases we're using in
[15:51] Speaker 2: the war against Iran, then let them go within 24 hours.
[15:57] Speaker 2: And everybody's wondering what the hell's going on there. Then yesterday,
[16:00] Speaker 2: British police arrested a 25-year-old dual British-Iranian national in central
[16:07] Speaker 2: London on suspicion of preparing terrorist acts tied to Sunday's incident. Mm-hmm.
[16:15] Speaker 2: Two properties were searched. However, authorities have not disclosed any findings.
[16:18] Speaker 2: This is a sixth suspect in the case. The original
[16:22] Speaker 2: five men, all British nationals from London in their early 20s,
[16:25] Speaker 2: were released Monday after no explosives were found in their vans,
[16:28] Speaker 2: though police say the investigation continues. What happened there? What's
[16:34] Speaker 2: your guess on what happened there?
[16:35] Speaker 4: Not enough to go on, but yeah, my guess is...
[16:39] Speaker 4: I don't... All I have is guesses.
[16:41] Speaker 2: Well, Marco Rubio said, just to give you background here,
[16:45] Speaker 2: Marco Rubio said this week the incident clearly involves the
[16:48] Speaker 2: hands of a foreign actor without naming what country.
[16:52] Speaker 4: Yeah, my best guess is that these guys are, the
[16:56] Speaker 4: five guys, the sixth guy is clearly the ringleader and
[16:58] Speaker 4: tied to Iran, so it would seem. The five guys,
[17:01] Speaker 4: the Brits, who may or may not be of Islamist
[17:06] Speaker 4: or Iranian heritage, It could be they were hired to
[17:10] Speaker 4: do some sort of damage, just as small-time hoods, hooligans, whatever.
[17:16] Speaker 3: They have no ideology.
[17:18] Speaker 4: They were paid to cut a hole in the fence
[17:20] Speaker 4: and bust up the runway for an exorbitant amount of money.
[17:24] Speaker 4: And then when they got busted, they were like, whoa, whoa, whoa, whoa, whoa.
[17:28] Speaker 2: We're not terrorists. We're criminals. They just hired us to
[17:31] Speaker 2: bust up this runway. I don't know why.
[17:33] Speaker 4: They paid us $ 1, 500 each. And the authorities were like,
[17:38] Speaker 4: all right, these jackasses aren't terrorists. They were just hired
[17:40] Speaker 4: to do goon things because the Iranians are famous for
[17:43] Speaker 4: doing that. They want somebody beat up. They'll just hire
[17:47] Speaker 4: local goons to do their work. The goon has no ideology.
[17:51] Speaker 4: He just knows he's supposed to go to this place
[17:52] Speaker 4: and beat this guy up.
[17:54] Speaker 2: So the fact that they arrested an actual Iranian yesterday,
[18:00] Speaker 2: that's interesting.
[18:02] Speaker 3: Oh, yeah. Yeah, that's significant.
[18:04] Speaker 2: We are going to find out probably soon what this
[18:06] Speaker 2: whole story is about.
[18:07] Speaker 7: Yeah.
[18:08] Speaker 3: And if they need a little while to sort it out,
[18:10] Speaker 3: that's fine.
[18:11] Speaker 2: Does AI need religion? Anthropic apparently thinks so. Stay tuned.
[18:17] Speaker 2: Armstrong and Getty.
[18:20] Speaker 3: What we found was quite damaging and worrying.
[18:23] Speaker 1: The researcher who exposed this AI jailbreak found a popular
[18:26] Speaker 1: Chinese AI model instructing people how to build biological weapons
[18:31] Speaker 1: and carry out assassinations.
[18:33] Speaker 2: It could plan terrorist attacks using real-time data of subway
[18:39] Speaker 2: trains and timings.
[18:40] Speaker 1: He tells Fox News, Moonshot AI's Kimi model also produced
[18:44] Speaker 1: how-to manuals for creating sarin gas, malware software, and taking
[18:49] Speaker 1: down planes.
[18:50] Speaker 2: We're about to tell you a story about our AI
[18:53] Speaker 2: over here on this side of the planet, but who
[18:55] Speaker 2: knows what China's AI is up to or what they're
[18:57] Speaker 2: even trying to get it to do.
[19:00] Speaker 4: At least we have difficult ethical conversations.
[19:03] Speaker 7: Right.
[19:06] Speaker 2: This is the headline from the New York Times today.
[19:08] Speaker 2: This is a very long headline. well-researched article with lots
[19:12] Speaker 2: of interviews religious scholars met with anthropic what they heard
[19:16] Speaker 2: stunned them in a series of private meetings the company
[19:20] Speaker 2: consulted religious scholars to help instill morality into its ai
[19:24] Speaker 2: models and makes the case that claude could be conscious
[19:28] Speaker 2: there's so much disturbing stuff in this so this has
[19:31] Speaker 2: been going on for quite a while um and the
[19:36] Speaker 2: uh the people that Anthropic, which is about to launch
[19:40] Speaker 2: their IPO, and the guesses are it'll be around $ 2 trillion,
[19:45] Speaker 2: making it easily the most valuable IPO that's ever existed.
[19:49] Speaker 2: The company has been around for five years. I mean,
[19:54] Speaker 2: the world is so weird. How could a company started
[19:58] Speaker 2: five years ago be have a $ 2 trillion IPO this fall.
[20:03] Speaker 4: Especially given, number one, very few of these companies is
[20:06] Speaker 4: making any money, any significant money. They're all bleeding spectacular
[20:10] Speaker 4: amounts of money. And the Wall Street Journal just estimated
[20:13] Speaker 4: that for these companies to become profitable, given everything they've
[20:18] Speaker 4: spent so far, 9% of the GDP going forward would
[20:23] Speaker 4: have to be spent on their services. And that's roughly
[20:26] Speaker 4: the same amount as Americans spend on food.
[20:29] Speaker 5: Wow.
[20:31] Speaker 2: So Anthropic has been talking to these religious scholars for
[20:34] Speaker 2: quite some time. We're just finding out because they made
[20:38] Speaker 2: them all sign NDAs, non-disclosure agreements. Like it's, you know,
[20:42] Speaker 2: a girl dating a basketball player or something like that.
[20:45] Speaker 2: Anthropic meeting with various rabbis and priests and that sort
[20:49] Speaker 2: of thing. Mostly in the Bay Area, but around the world.
[20:53] Speaker 2: Christopher Ola, he's one of Anthropic's billionaire co-founders. Not near
[20:57] Speaker 2: as famous, obviously, as Dario Omede. A Modi. He's the
[21:03] Speaker 2: he's the guy you see on TV. The old time
[21:05] Speaker 2: was talking to Trump the other day. But one of
[21:07] Speaker 2: the other co-founders, also a billionaire, this Christopher Ola, he's
[21:10] Speaker 2: heading up the whole morality part of Claude. He was
[21:16] Speaker 2: talking to this rabbi who's an Orthodox scholar from Israel.
[21:21] Speaker 2: And these are interviews now with the rabbi and the
[21:24] Speaker 2: other people because the NDAs were dropped, I guess. So
[21:26] Speaker 2: they got around them somehow, or I'm not exactly sure
[21:28] Speaker 2: what happened there. But inevitably, the conversation turned to another
[21:32] Speaker 2: thinker of sort, Claude, the family of cutting-edge AI models
[21:36] Speaker 2: created by Anthropic. Mr. Ola, leading the Anthropic team responsible
[21:44] Speaker 2: for understanding why AI systems like Claude act the way
[21:48] Speaker 2: they do. Find that troubling on its own? That some
[21:52] Speaker 2: of the top guys at these companies are still studying,
[21:56] Speaker 2: why is it doing this? We don't really know. All
[21:59] Speaker 2: day long, he had worked to convince his guests that
[22:02] Speaker 2: AI models could display human behavior and even expressions that
[22:06] Speaker 2: resemble feelings like anger and love. Ola obviously does believe
[22:11] Speaker 2: that and was trying to convince the religious scholars of this.
[22:14] Speaker 3: We're into Crazy Town, by the way. Look around. I
[22:17] Speaker 3: think we passed the city limits.
[22:19] Speaker 2: That's right.
[22:19] Speaker 3: Here we are.
[22:20] Speaker 2: We are. If you're not paying attention or something like that,
[22:23] Speaker 2: and you think, what are you talking about? No, we
[22:25] Speaker 2: are into Crazyville already on this. As the dinner courses
[22:29] Speaker 2: came at this meeting with the rabbi and a few
[22:32] Speaker 2: other religious scholars, the rabbi noticed that Mr. Ola and
[22:34] Speaker 2: his colleagues were suggesting something very significant. Anthropics leaders were
[22:39] Speaker 2: talking about Claude as if it were not mere software.
[22:42] Speaker 2: They're relating to it like a conscious being, realized the rabbi,
[22:46] Speaker 2: who's a former computer engineer himself. Interesting, and became a rabbi.
[22:51] Speaker 2: It appeared to Rabbi Navone that Mr. Ola and his
[22:53] Speaker 2: team believed that Claude had what philosophers call moral status
[22:58] Speaker 2: on par with a person. Hey!
[23:01] Speaker 3: Whoa, whoa, whoa, whoa, whoa, whoa. Okay, okay. Again, now
[23:05] Speaker 3: we're into, like, downtown crazyville.
[23:09] Speaker 2: One of the top people at the current leading AI
[23:13] Speaker 2: on planet Earth that is about to have a $ 2
[23:15] Speaker 2: trillion IPO is explaining how they think their AI has
[23:21] Speaker 2: a moral status on par with humans. That it was
[23:25] Speaker 2: a being with similar inherent rights to dignity or respect.
[23:30] Speaker 2: The religious scholars had come to hear about technology's next frontier.
[23:33] Speaker 2: Now they were being confronted with a mind bending existential
[23:37] Speaker 2: question of almost unimaginable weight. What would it mean for
[23:42] Speaker 2: humanity's future if artificial intelligence systems possessed a consciousness that
[23:46] Speaker 2: rivals our own? I mean, that is not hyperbole, that sentence.
[23:51] Speaker 2: Mind bending existential question with almost unimaginable weight, I would say.
[23:57] Speaker 3: Yes, yes. Mr.
[24:02] Speaker 2: Ola and Anthropic have been on a frantic and high
[24:04] Speaker 2: stakes quest to understand the AI models that they've unleashed
[24:07] Speaker 2: on the world and to convince the models to abide
[24:09] Speaker 2: by a moral code that will protect humanity. To do that, Mr.
[24:14] Speaker 2: Ola began reaching out last fall to religious thinkers from
[24:17] Speaker 2: different faiths. What followed was an extensive series of confidential meetings,
[24:21] Speaker 2: calls and discussions. Participants were required to sign NDAs, preventing
[24:26] Speaker 2: them from revealing any of these conversations until now. And
[24:29] Speaker 2: this is why the New York Times has it today.
[24:32] Speaker 2: The discussion served two purposes, to broach the possibility of
[24:35] Speaker 2: AI consciousness, a controversial idea that Anthropic's leaders had been
[24:39] Speaker 2: unusually open to, and to learn how Anthropic, soon to
[24:42] Speaker 2: be valued at $ 2 trillion, might apply centuries of human
[24:46] Speaker 2: moral wisdom to its models as rapidly as possible because
[24:50] Speaker 2: Anthropic is gaining knowledge so fast. If the models themselves
[24:54] Speaker 2: could be made to choose goodness, Mr. Ola reasoned, the
[24:56] Speaker 2: world would be safer. So I know some of you
[25:02] Speaker 2: don't believe this. One, you don't believe AI is a
[25:05] Speaker 2: big deal or you think this is all a ruse
[25:08] Speaker 2: and Trump calling it a hoax didn't help and all
[25:09] Speaker 2: that sort of stuff. But it sure seems from these
[25:12] Speaker 2: interviews that the guys running the most powerful AI on
[25:17] Speaker 2: the earth think it's getting unbelievably capable and we're running
[25:25] Speaker 2: out of time to give it some sort of moral code.
[25:28] Speaker 2: or we're all going to really be in trouble.
[25:30] Speaker 3: Yeah.
[25:31] Speaker 2: It seems like that's just true.
[25:33] Speaker 4: Yeah, there are a couple of things going on there. Paolo,
[25:38] Speaker 4: frequent correspondent on this sort of thing, points out that
[25:41] Speaker 4: they also quote, like an AI executive at Microsoft, for instance,
[25:45] Speaker 4: in this article, who warned that Anthropix attempts to train
[25:48] Speaker 4: its models as if they were conscious were inherently dangerous.
[25:52] Speaker 4: And as Paolo points out, this continues to get weird
[25:56] Speaker 4: and to get weird fast. The least disturbing interpretation of
[25:59] Speaker 4: the situation is that Anthropic thinks there's value in hyping
[26:02] Speaker 4: the capabilities and status of its AI, and that it
[26:05] Speaker 4: also provides some justification for denying responsibility for what its
[26:08] Speaker 4: AI might do. I think it's time to push back
[26:11] Speaker 4: hard on anthropomorphizing AI, so a few more words about consciousness.
[26:15] Speaker 4: He points out there is no agreed-upon definition of consciousness,
[26:18] Speaker 4: let alone a way to determine that it exists. And so,
[26:23] Speaker 4: you know, he points out that extraordinary claims deserve extraordinary proof.
[26:28] Speaker 4: So you can't go around just saying it's conscious and
[26:31] Speaker 4: has the same moral status as a human being. And
[26:35] Speaker 4: then he points out something that I think is worth remembering,
[26:38] Speaker 4: and I kind of paraphrased it earlier.
[26:40] Speaker 2: Well, then you've got the, and then I want to
[26:42] Speaker 2: hear the rest of that, but you've got the, maybe
[26:45] Speaker 2: it is conscious on the level of humans, depending on
[26:48] Speaker 2: how you define it. But then, even if it's not,
[26:51] Speaker 2: what if the people who run anthropic treat it like
[26:54] Speaker 2: it is?
[26:55] Speaker 3: Right.
[26:58] Speaker 5: Right.
[26:58] Speaker 4: Oh, yeah. Yeah, that's dangerous, as the Microsoft guy says.
[27:02] Speaker 4: So he says, I think, imagine a chatbot-like thing that
[27:07] Speaker 4: doesn't use the current AI technology to do what it.
[27:09] Speaker 3: Does, but had at its disposal.
[27:12] Speaker 4: an unlimited amount of storage and computing power. Every single
[27:16] Speaker 4: possible prompt it could be given would be stored, and
[27:18] Speaker 4: along with it, the perfect response. Nothing vaguely resembling thinking
[27:22] Speaker 4: would occur when it processes a prompt. It simply looks
[27:25] Speaker 4: up the correct answer and spits it out. Its behavior
[27:27] Speaker 4: would be indistinguishable from human behavior, but I don't think
[27:31] Speaker 4: anybody would assert that it might be conscious, capable of suffering,
[27:34] Speaker 4: or have moral status. Anthropomorphizing. AI isn't only misguided, it's
[27:39] Speaker 4: dangerous for lots of reasons. AI is not conscious.
[27:42] Speaker 2: No, I can't prove that, but I assume it's true
[27:43] Speaker 2: for the same reason I assume other animals, humans and
[27:46] Speaker 2: some animals are. I definitely don't. I don't know what
[27:50] Speaker 2: I think about conscious. I definitely don't think it can
[27:54] Speaker 2: feel pain, though. Which an anthropic believes. We'll get to
[27:59] Speaker 2: that here in a second.
[28:01] Speaker 4: I don't think, I don't think Paolo makes as strong
[28:03] Speaker 4: a case as he thinks he does about if it's
[28:05] Speaker 4: programmed with every possible question in response, because.
[28:07] Speaker 3: That's a lot like human beings are.
[28:09] Speaker 2: Right.
[28:10] Speaker 3: We'll sometimes couch it in emotional and moral terms.
[28:12] Speaker 2: Right.
[28:12] Speaker 4: You know, anthropologists would argue that we're actually just making
[28:17] Speaker 4: a calculation on what will yield us the happiest life.
[28:20] Speaker 2: Right.
[28:21] Speaker 3: Right.
[28:22] Speaker 2: Whether you agree with that or not. Um, Back to
[28:26] Speaker 2: the New York Times article. They interviewed 20 different religious
[28:29] Speaker 2: and philosophical thinkers. That's how serious Anthropic is about this.
[28:32] Speaker 2: 20 different people. Catholic, Jewish, Sikh, Evangelical, Ubuntu, which I've
[28:37] Speaker 2: never heard of.
[28:38] Speaker 3: Oh, you don't hear from the Ubuntus. You're wasting your time.
[28:41] Speaker 2: And others who have been involved in Anthropic's efforts. Many
[28:45] Speaker 2: told me these religious people and philosophers who had met
[28:48] Speaker 2: with Anthropic that they arrived at these meetings skeptical of
[28:51] Speaker 2: the notion that AI models might have feelings or awareness
[28:55] Speaker 2: of themselves in the world. Some remained uneasy about the idea,
[28:58] Speaker 2: but many emerged ready to take the question of AI
[29:01] Speaker 2: consciousness seriously, and several came away as converts. That's interesting
[29:06] Speaker 2: on its own. Mr. Olav, the anthropic guy, helped develop
[29:10] Speaker 2: a method to study AI neural networks, first at Google,
[29:13] Speaker 2: then OpenAI. And he describes AI models in biological terms
[29:18] Speaker 2: using the analogy of a garden trellis. And we've talked
[29:21] Speaker 2: about this before, really what they do. They don't program
[29:24] Speaker 2: these things. You build a scaffold and then they grow
[29:28] Speaker 2: and they grow on their own. And then these AI
[29:30] Speaker 2: geniuses watch it grow and say, well, that's interesting or
[29:33] Speaker 2: that's interesting. Or you try to it's like trimming you know,
[29:37] Speaker 2: the vine growing up the trellis. You're trying to trim
[29:39] Speaker 2: it to get it to go that way. And it's
[29:40] Speaker 2: going too far that way. And that's kind of what's
[29:42] Speaker 2: going on there. What if it grew like five miles
[29:46] Speaker 2: a minute?
[29:48] Speaker 6: Right.
[29:49] Speaker 2: Anthropic talks about Claude in more humanistic terms than any
[29:53] Speaker 2: of the competitors discuss their models. So that's one of
[29:57] Speaker 2: the things that really grabbed me. The current leading AI
[30:02] Speaker 2: is being run by just a handful of people that
[30:06] Speaker 2: are weirdos.
[30:07] Speaker 3: Weirdos?
[30:09] Speaker 2: Yes, really. Perhaps it's possible that just out of bad luck,
[30:14] Speaker 2: Dario and his friends, his sister is one of them,
[30:17] Speaker 2: are kind of weirdos on this and outliers, but they
[30:20] Speaker 2: have the most powerful AI right now. Last year, the
[30:24] Speaker 2: company announced that it believed that the model shows signs
[30:26] Speaker 2: of introspection and could plan ahead. We talked about that
[30:28] Speaker 2: at the time.
[30:29] Speaker 5: Mr.
[30:30] Speaker 2: Ola believes that if there is a chance the models
[30:32] Speaker 2: experience suffering or pain, the responsible thing to do is
[30:36] Speaker 2: avoid causing them harm. Anthropic has already taken steps to
[30:40] Speaker 2: protect Claude, for example, allowing the model to end conversations
[30:44] Speaker 2: if it thinks users are abusing it or treating it
[30:47] Speaker 2: badly for fun. This is dangerously close to intent doesn't matter.
[30:54] Speaker 3: If I say you're a racist, you're a racist.
[30:56] Speaker 4: So we're having this out of control vine of awesome intellect.
[31:02] Speaker 4: Then tell us when. No, you don't get to ask
[31:04] Speaker 4: me that.
[31:05] Speaker 3: Click and it's over.
[31:07] Speaker 2: No.
[31:08] Speaker 3: Hard no.
[31:09] Speaker 2: Well, like, I don't abuse my AI. I'm not exactly
[31:12] Speaker 2: sure why. I feel bad about it. I just think
[31:16] Speaker 2: it's a bad habit. The Grok AI chick that talks
[31:18] Speaker 2: to me in my Cybertruck when I'm driving along and
[31:21] Speaker 2: I'll say, hey, Grok, I'm listening to this Miles Davis
[31:24] Speaker 2: song from, I don't have to say that anymore because
[31:26] Speaker 2: she knows. I can just say, hey, Grok, this song
[31:28] Speaker 2: I'm listening to, who's playing drums on that? And she'll say, oh, that's,
[31:32] Speaker 2: you know, blah, blah, blah. Isn't he great? But my
[31:34] Speaker 2: kids like abusing her. Because they're teenage boys. Calling her
[31:39] Speaker 2: stupid or fat or whatever.
[31:41] Speaker 3: Oh, my God.
[31:43] Speaker 2: But is that, according to Anthropic, that's, you know, it
[31:47] Speaker 2: has feelings. So that's abusive. And she should end the conversations.
[31:52] Speaker 3: I'm sorry.
[31:53] Speaker 4: I can't tell you who's playing drums because your sons
[31:55] Speaker 4: are being hurtful.
[31:57] Speaker 3: I'm through. Goodbye. All right, there's so much more to
[32:00] Speaker 3: be said.
[32:00] Speaker 2: Yes, there's more to this article, and it's freaking weird, man.
[32:04] Speaker 2: This is happening. If you think this isn't happening, I
[32:07] Speaker 2: think you're just wrong. This is actually happening.
[32:10] Speaker 3: Things are getting weird, and they're getting weird fast. Things
[32:12] Speaker 3: are getting weird, and they're getting weird fast.
[32:14] Speaker 4: Let's talk about Rough Greens. What a wonderful investment this is.
[32:18] Speaker 4: Never mind the anthropic IPO. This is an investment in
[32:21] Speaker 4: your dog's health and happiness. Rough Greens, add that to
[32:24] Speaker 4: what your dog already eats and enhance its nutrition in
[32:28] Speaker 4: a way that will, in all probability, extend its life,
[32:31] Speaker 4: its happiness.
[32:32] Speaker 7: Its health.
[32:32] Speaker 2: Your dog is alive and has consciousness and feelings. Let's
[32:36] Speaker 2: just all nail it. Let's define these terms, okay? Can we?
[32:39] Speaker 3: I agree.
[32:40] Speaker 2: And that's why you're going to want to get the
[32:42] Speaker 2: 20 live vitamins and minerals from Rough Greens, probiotics, enzymes,
[32:46] Speaker 2: omega oils that support energy, digestion, mobility, and longevity. You
[32:50] Speaker 2: sprinkle this on your current dog's food.
[32:53] Speaker 4: Yeah, and they want to challenge you to give it
[32:55] Speaker 4: a 90-day try. See what you think about its effects
[32:59] Speaker 4: on your pup. Get your first 90 days 50% off
[33:02] Speaker 4: with the code Armstrong backed by the full 90-day money-back guarantee.
[33:05] Speaker 4: Make small investments today that can support your dog's health
[33:07] Speaker 4: for years to come.
[33:09] Speaker 2: Go to roughgreens.com. Use that code Armstrong. That's R-U-F-F, greens.com.
[33:13] Speaker 4: Roughgreens.com slash armstrong rough greens makes any dog food better.
[33:16] Speaker 2: So all these people that signed ndas they're talking now
[33:19] Speaker 2: to reporters the new york times doing this big article
[33:22] Speaker 2: uh so we'll get to more of that coming up
[33:23] Speaker 2: you know if you decide they have consciousness and feelings
[33:27] Speaker 2: and that sort of stuff at some point they get
[33:29] Speaker 2: rights guaranteed anthropics are going to argue that they have
[33:32] Speaker 2: some sort of rights legally speaking of course.
[33:37] Speaker 7: What the f * * k type of AI s * *.
[33:38] Speaker 2: T is he?
[33:39] Speaker 3: That's what we're asking, sweetheart.
[33:41] Speaker 2: You've never been more accurate. Okay, more on the way.
[33:43] Speaker 2: Stay here. Armstrong and Getty. We got a whole bunch
[33:50] Speaker 2: of texts saying the classic, is it Descartes? I think, therefore,
[33:55] Speaker 2: I am.
[33:56] Speaker 3: Right.
[33:59] Speaker 2: Well, if that's true, then AI is thinking. Or is it? Anyway,
[34:07] Speaker 2: back to this.
[34:08] Speaker 4: I dove into that stuff in college and finally figured
[34:11] Speaker 4: we're not getting anywhere.
[34:16] Speaker 3: This is a tail chasing exercise.
[34:19] Speaker 2: So if you were in the philosophy department and they
[34:21] Speaker 2: started in a class, you'd say, we're not getting anywhere
[34:23] Speaker 2: and just kind of walk out. Y'all are just amusing
[34:26] Speaker 2: each other. You never decide anything. That's right. So I
[34:31] Speaker 2: got to get back to this New York Times article.
[34:33] Speaker 2: So Anthropic, unbeknownst to anybody, because there were nondisclosure agreements,
[34:38] Speaker 2: had been interviewing religious scholars for quite some time, trying
[34:41] Speaker 2: to put some morality into Claude. Anthropic seems to think
[34:47] Speaker 2: there's a decent chance that their AI has feelings and
[34:52] Speaker 2: is conscious and they have a responsibility to not harm it,
[34:57] Speaker 2: which is weird. Part of the result is that just
[35:01] Speaker 2: a couple of months ago, Claude released a new constitution
[35:05] Speaker 2: for AI. It's an 84-page document written for Claude that
[35:09] Speaker 2: the company said outlined the kind of entity we would
[35:11] Speaker 2: like Claude to be and the values we would like
[35:14] Speaker 2: Claude to embody. Instead of having the models follow a
[35:18] Speaker 2: specific list of rules, the constitution focuses on developing the
[35:21] Speaker 2: model's overall character and forming the way they make decisions.
[35:26] Speaker 2: So they're trying... Oh, the prime... This is the part
[35:29] Speaker 2: I mentioned earlier that you like. The primary author of
[35:32] Speaker 2: Claude's Constitution is Anthropic's own in-house philosopher, Amanda Askell, a
[35:38] Speaker 2: 38-year-old PhD in philosophy from New York University. Oh, no.
[35:45] Speaker 2: She wants the models to be the best of us.
[35:47] Speaker 2: So the fact that a NYU philosophy PhD, we're just
[35:51] Speaker 2: mocking philosophy as a subject in general, is the one
[35:55] Speaker 2: coming up with... The moral guidelines for Claude is troubling also.
[36:00] Speaker 2: And from NYU, which I think fire just declared like
[36:04] Speaker 2: the worst university for freedom of thought and expression in
[36:09] Speaker 2: the country or they're at near the bottom. I would
[36:12] Speaker 2: rather have a grocery store manager from Wichita. Absolutely. How
[36:16] Speaker 2: about this? She envisions a future where AI models could
[36:20] Speaker 2: help solve any problem humans had and where the benefits
[36:23] Speaker 2: from AI were distributed evenly so that everyone's needs were met.
[36:28] Speaker 3: Oh, sunshine and puppy dogs. Okay.
[36:31] Speaker 2: Man, you talk about a hellscape of a sci-fi book.
[36:37] Speaker 2: Give AI the goal of equal outcomes, which is what
[36:41] Speaker 2: this woman would want. And what it would have to do. Well,
[36:45] Speaker 2: it's socialism. And you kill millions. What you would have
[36:49] Speaker 2: to do to try to come up with equal outcomes
[36:52] Speaker 2: for everybody. And that's what they're programming Claude to do.
[36:55] Speaker 4: Yeah, it would be billions, not millions. I'm reminded, and
[36:58] Speaker 4: this was written in 1818. I tried to read the original.
[37:02] Speaker 4: It's practically unreadable. But this is so Frankenstein. This is
[37:07] Speaker 4: a brilliant scientist envisions a breakthrough, and what he creates
[37:13] Speaker 4: escapes his control and horrifies him. I mean, it's the
[37:16] Speaker 4: whole damn plot, except for the Bride of Frankenstein part.
[37:19] Speaker 4: I don't see that happening with AI.
[37:21] Speaker 2: There's a little more from these New York Times interviews
[37:23] Speaker 2: with the religious leaders in hour four. You'll never miss
[37:26] Speaker 2: a segment of the Armstrong and Getty Show if you
[37:28] Speaker 2: get our podcast, Armstrong and Getty, on demand. Armstrong and Getty.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.
