IM 872: Infinite Jeffs - Why the Pope's AI Manifesto Matters

34:25There's not one negative. That's right. It really surprised me because, oh, there's a thing. I'll tell you a joke. I had a psychologist friend once, and he said to me, this was like 30 years ago. He says, Rick, if you were giving a presentation in Yankee Stadium, I don't know how many people there. Do you know how many people? Anyone know how many people in Yankee Stadium? 80 ,000. I don't know. 80 ,000. Yeah. If you're given a presentation in Yankee Stadium and 79 ,999 people stand up and cheer at the end, Rick Sam's the greatest. And after that, one guy stands up and says, I hated that. Rick Sam is no good.

35:14I'd go home feeling bad. Me too. I'm the same way. You're the same way, right? Same exact way. So on the AI class there, I was expecting some negative comments. None. I love it. I think because I make it fun and I look at it as fun. I'm not saying this is art. You know, you could be the next great artist. You can make a ton of money selling art. I'm just trying to show people, just like when I teach guitar. guitar, when I teach guitar, I start out by teaching people songs they know, so they get to have fun. And I think it's just, if I could bring a little fun into someone's life at this age, I think it's okay.

36:03Well, you brought some fun into ours today. Thank you, Rick Salmon. RickSalmon.com, S-A-M-M-O-N. Take the course, view the site, look at all these beautiful pictures buy some of his books uh supporting artists like rick salmon is really uh the best thing you can do to make the world a better place thank you rick we really appreciate it well thank you mike it's great great me and everyone else it was a ton of fun so much pleasure sam and everybody thanks rick uh we will have more intelligent machines in a bit with father robert balas here filling in for paris martineau and jeff jarvis our show today uh brought to you by trusted tech and this is a very timely advert because if you're managing microsoft 365 for your company you know you're the one in charge you're responsible for both the cost and whether it's set up correctly i hope you know this you must know this july 1st one month from now microsoft is raising prices so if there are any errors in your licensing they are going to get more expensive in 30 days 30 okay 33 days the clock is ticking most companies using microsoft 365 you see are either over licensed paying for unused seats and features stuff they don't use or under licensed which you know you might say well we're saving money but it also can create compliance and security risks and it's not unusual that accompanies both over licensed in one department under license in another department the result you're wasting thousands sometimes tens of thousands per year on tools either your team doesn't use or worse you're missing critical security features you thought you had trusted tech helps businesses understand what they have what they actually need and how to lock in the right setup now before costs go up their team ensures your m365 environment is well supported and aligned with how your business actually operates and if you need ongoing help they offer reactive support for your microsoft environment through their certified support services nobody denies including microsoft that their licensing is pretty confusing and it's changing all the time you got e3 versus e5 you got business premium you got add-ons there's the new e7 it it's mind-boggling easy to misconfigure and overpay it's no reflection on you but remember those licensing mistakes don't just cost money they can create compliance exposure that's going to get more expensive after july 1st so even if you think your licensing is dialed in now's the time to take a second look ask kevin turner you know his name former microsoft coo he was talking to trusted tech i got this i wrote it down he said quote he's talking to trusted tech you have an incredible customer reputation you have to earn that every single day the relentless focus you guys have on taking care of customers gives them value and differentiates you in the marketplace.

39:00After July 1st, you're stuck paying more. This is the last chance to fix your licensing before costs go up. Trusted Tech. Right now, they're offering a free Microsoft 365 licensing consultation. Visit trustedtech.team slash intelligent365. Get a clear data-backed view of your current licenses, what you're wasting, and how to lock in the savings before the price increase. Go to trustedtech.team slash intelligent365. Submit a form. Get in contact with Trusted Tech's Microsoft licensing engineers. Don't put this off. The clock is ticking. Back to Intelligent Machines. Father Robert Ballasier, the digital Jesuit.

39:42Everything good in your world, Robert? Aside from the fact that summer has arrived a month and a half early, yeah, pretty good. How well air-conditioned is the Vatican? uh well my office is very well air conditioned because when i first got here i salvaged an ac and installed it so it's completely not connected to the main house but the the main house is on energy safe mode so it's not nice uh so it was i think it was monday right the uh holy father leo the 14th issued his encyclical letter magnifica humanitas on safety humanity yeah i think we can understand what that that's a good latin it's very easy to understand on safeguarding the human person in the time of artificial intelligence um now this is 42 000 words jeff has read it in fact jeff you wrote a whole piece on buzz machine your analysis of this i spent a day and a half study i think it's a magnificent document.

40:43I spent a day and a half studying it and I could, I'll go back and read it again and again. Um, but I pulled out some, to my mind, the good bits. There are many, uh, and, um, I had to stop myself cause I would have ended up 42 ,000 words myself. And I also saw lessons here for others than just AI. I saw a lot of lessons for media here. Yeah. And so it's parallel. Let me, first of all, Robert, any disclaimer you want to make, this is the time to make it you are you are in the vatican city right now i am and people will always ask oh did you have a hand in this and i will just say there are some very very good people here who spent a lot of time putting together a document that represented the vision that pope leo had for the document and that's about all i can say on that it it reflects his point of view though yes correct that's fair to say yeah and that's the thing it's not a surprise if you look back to the the the uh the talk he gave when he was elected pope he brought up ai he brought up a lot of these same concerns and he's written three documents since then before this encyclical so those have just been expanded upon a lot of people were expecting this to be the anti-ai encyclical ai is the it's horrible it's going to destroy us they thought it was going to be a repeat of rayram navarum which was the document from the industrial revolution that was leo the 13th his predecessor yeah we kind of knew this was coming because he did chose choose the name very carefully leo he did right that wasn't that was not a coincidence that was not a mistake he understood that just as the industrial revolution changed the world it changed society and had a lot of very negative challenges that it brought uh the same thing is happening with the advancement of of technology not just ai but technology in general in his words technology is not counter to faith that's one of the very first things that you see in the encyclical so he wants to make it clear this is not the luddite encyclical we're not telling you to destroy your computers and disconnect yourselves was the original the 135 year by the way this was published on the 135th anniversary of the poet leo the 13th encyclical was leo the 13th kind of a luddite anti-industrial it wasn't and it's it's misunderstood people think that that was telling people not to participate in the industrial revolution but this leo and that leo recognized the same thing which is you cannot reverse time you cannot pretend like these things did not happen what you can do is you can be transparent you can safeguard those peoples at the margins and ultimately the entire thing the all of Catholic social teaching comes down to a choice.

43:30It's a question. Would you prefer a future where we are lifting up the innate dignity of every human, every man and woman, or are you okay with a future that commoditizes human? And that seems like a very simple choice, and it should be a simple choice. And then the entire encyclical is, if you want the first thing, these are the steps that we should probably be taking. Very simple. He begins with a contrast between the Tower of Babel and the rebuilding of the wall in Jerusalem. And it's a wonderful metaphor for it, where the Tower of Babel, as he presents it, sounds an awful lot like AI. A doom project was an impressive feat, a single language, a single technology, a single direction.

44:18However, the project concealed of profound danger. It was a project conceived without reference to God, supported by uniformity that eliminated diversity and that chose homogenization over communion. That's what sounds familiar. In contrast, the rebuilding of the wall in Jerusalem was done with prayer and permission and the collaboration of the people of the city, each bringing to it what they could bring to it. And clearly that was the choice that you just mentioned, Robert. That's the yes or no here. And I think it's a really apt metaphor for what we have to choose now. Oh, very much so. And one of the things that I really, really like about this document is he painstakingly, about a quarter of the document is stepping through each generation of Catholic tradition from Rerum Novarum.

45:09So, you go from Rerum Novarum, you go through Pius XI, you go through Vatican II, and the documents, the encyclicals of Pope Francis, Fratelli Tutti. So he's telling people, because he knows what the critics are going to say. They're going to say, oh, this is a radical pope and this is a woke pope. He's like, look, this has been our belief for more than 100 years. This is what we've been pushing for more than 100 years. So the challenge isn't new. It's just what we're focusing on is new. The AI is new. LLMs are new. But what they can bring, the negative consequences that they can bring, we know them quite well, and we know how to deal with them.

45:46If I may quote again, a little bit of length, technology has the power to heal, connect, educate, and protect our common home, but it can also divide, exclude, and generate new forms of injustice. In the abstract, technology in and of itself is not a solution to humanity's problems, just as it is not inherently evil. In practice, however, technology is never neutral because it takes on the characteristics of those who devise, finance, regulate, and use it. Therefore, the primary choice is not between a yes or no to technology, but rather between constructing Babel or rebuilding Jerusalem. Between a power that claims to dominate the heavens and the people who work together in the presence of God, rebuild the walls of fraternal coexistence.

46:26I love that the document emphasizes so much community and conversation and listening. What does he mean by rebuilding Jerusalem? Well, rebuilding Jerusalem is just, it's a metaphor for the fulfillment of the kingdom. that's it's it's not physically rebuilding jerusalem it's raising it back up uh that was supposed to be the start of the end time so forgive me because i don't mean to be insulting anyway but there are people who are going to say well this is a essentially medieval institution uh commenting on 21st century technology yes uh you know personally i think that that uh the pope has standing as a ethical as a as a leader in ethics as a leader in um you know uh not just spirituality but morality so i think he has some standing but it does it is a little bit jarring right i mean it's in latin the title is in latin ladies and gentlemen so what is the pope standing here oh i i get that i get that and uh i also understand that people are going well why is the Pope talking about this?

47:36Well, let's first of all say he is a head of state. Let us say that he is the head of a theological organization that deals with God, which many people don't believe in, and that's fine. And I would say, set that aside. Don't focus on the fact that it's coming from a man who may have a different belief system than you. Look at some of the primary questions that are asked. Like, for example, this is a big one. Is it a better world or a worse world if we concentrate power into the hands of about 12 individuals and i think most of us regardless of our theological faith background would say no that's not a good idea that's that's the quote that jeff read about um about concentrating power about right letting these systems reflect the bias of those those very few people and which is why there's a call at the end for transparency if you have transparency to show how these systems are actually being built to show what biases are being built into them, then you actually can say it's for the common good.

48:36And that has nothing to do with theology. It has nothing to do with faith. It has everything to do with, as a human person, do I want to be a disposable commodity or do I want to have my innate dignity recognized? Right. That's not a hard question. This is a document that really strikes me, Robert, that it's for generations. Yes. Very little of it is, I mean, it's halfway through before he really addresses AI fully and directly. He was about history first, 20 pages into 40 pages. And I think that there's kind of almost obligatory references to things like kids and screens. And there's obviously a need to address issues around work.

49:21And the Pope makes clear that the church has always supported labor. And war, an autonomous war, and that's current. What delighted me because of my own hobby horse here was the one thing that he called out currently is he called out specifically transhumanism and post-humanism. Yes. Which I talk about on the show often about the Tescrial bundle of these kind of crazy philosophies. There was a wonderful essay in Dietz Zeit this week by a German academic who said that AI is becoming a religion. And in some quarters, it is. and and and i was so happy that he that he pulled that out to call on that today and in essence without using the words he said um i'm trying to find it here without using the words it it it referenced the stench of eugenics that's my word not his um according to him these perspectives form the ideological background present in some centers of technological power and occupy the collective imagination in a simplified form, especially in the media and on social networks.

50:30They tend to foster enthusiasm for new technologies through a futuristic vision of an enhanced human being or human-machine hybrid. That's the eugenics. If the human being is treated as something to be perfected or surpassed, it becomes easier to accept that some lives are less useful, less desirable, or less worthy. That's a profound response to what you hear from a lot of the noxious AI boys out there today. And I think it also calls to the need to have others in the discussion and not just them. Yeah, yeah. And dovetailing into that is this idea that that's not just a theological question. That is a real practical discussion that needs to be happening right now in polarized society.

51:18Because if we went back 30 years and someone told you, oh, you know, I think that their human lives are less valuable than their human lives, there would be an immediate pushback. You can't say that. That's wrong. Of course not. I don't know if you'd get that today. I mean, you have people who are espousing that exact philosophy that there are parts of society, there are parts of the world, there are entire populations that are no longer useful. The world would be better off if we did not have them. That is such a dangerous territory for us to be in. So this part of this document is calling us back from that brink, back into sanity to say, no, no, no, no, wait, wait, wait.

51:58A monoculture is horrific. A monoculture where very few people control everything that we see and do is no longer a culture. And we are marching off the edge of that cliff for trinkets. and we don't have to. That's a good way to put it. One of my favorite parts of the whole thing, and this is very media perspective, is what he said that he talks about truth. Yes. And my favorite line is, understanding that the truth is a gift to be shared, not a possession to be monopolized. And there he's speaking again to the AI companies, but I think he's very much speaking to media and into my profession of journalism, that we think that it's a commodity that we can own and control and charge for.

52:47And whether it's in arguments with the AI companies or whether it's paywalls to the public or whatever it is, we, I think, lose sight of a higher calling here. And I'm reading too much into it for my world, for media, but I saw lessons here for us as well. Right. There's a very good comment in the Discord from L-R-A-U, I'm not sure how this is. Gold, L-R, Laurentian, A-U, gold. That makes sense. Got it, got it. Who says, look, we've already had this concentration of power. That's the human history. So all that's different now is that we actually kind of know who the people are, which is true, which is true.

53:26Except I would argue that the concentration we have now is far more dangerous. We always had wealthy people. We always had people who controlled a large percentage of the world's wealth. The difference now is 12 people, not 12 organizations, 12 people control more of the planet's wealth and resources than the bottom 70 % combined. And the wealth that they control is not just material wealth, resource wealth. They control what we see. They control our entertainment. They control our news. They control our communications. And more recently, they openly control our politicians and the policies that get implemented.

54:08That is an extremely dangerous combination. And we don't even have a political theory that really describes what that is. It's part oligarchy. It's part kleptocracy. It's the second Gilded Age. It really is. It really is. And so we are pushing ourselves back into a situation where we're essentially, our only option is to have a patronage. That's it. Because the regular economy is going to be gone. yeah we can only survive if you get somebody rich enough to support you yep oh maybe oh please give us some money so i can live uh that's that radical self-reliance that the vcs keep talking about they keep mentioning that and it's like do you know what that means i don't think that word you keep saying it i do not think it means what you think it means but interestingly And Pablo, I think it was clear in here, does not support the panacea of universal income.

55:06Really? Because he says that work has value. Work brings you dignity. And I think that's a really important perspective here. Don't think you can just buy people off. And that's a direct holdover from Ray Ramnevarum. So that comes straight through. And it's what FDR did in the Great Depression. uh he he realized the best thing you could do is give people work even if it's almost make work it's better to have that and we got a lot of interesting art and work out of that I'm curious what you both thought of of having the co-founder of Anthropic there that was interesting and by the way they were they were not the only AI company to lobby and to be and to have some you know input on this I don't think uh but he was sitting on the stage with yes yes and spoke Yeah.

55:54Yes, he was. What did you think of that? I think it was brave of him to come, honestly. I absolutely gave him kudos for showing up. But also... The media was very quick to pick up on some of the past slights. They said, oh, well, you know, on his social media accounts, he's very critical of the Catholic Church. He's a self-identified atheist. He has no use for theology. But remember, something that both Pope Leo and Pope Francis really emphasized was this idea of people of goodwill. We're not just looking for ourselves. We're not just looking for people who look and think like us. We are looking for people of goodwill.

56:33and those kinds of people can be found in every sort of life and of every sort of background and again some of them might be atheists some of them might be people of faith but if you have this belief that your work can be something that lifts up humanity we want to work with you and i think that was the message that yeah he absolutely telegraphed it and you know what that could be a very fruitful relationship for both of them because he does believe that his work is making the world better so he's not one of these he's not twirling his mustache in the in the background thinking of all the ways he's going to destroy the world he he really does believe that something good can come out of what he's doing what his company is creating uh pope leo could be a very good voice for him to say okay well if you want this this is what's going to happen let's let's let's think this out let's talk it through but it was also interesting that the the Pope does not see a moral machine.

57:34Correct. This notion of alignment, which is something that, of course, Anthropic pushes all the time, that we can create the machine that is aligned with human morals and ethics. He dismissed that. 100%. When I said that morality comes from human beings, it does not come from the machine. And it's another case of there's no quick fix there. So there was some irony, I think, in having Anthropic there because when they push their virtue, it's two things now. One is that we wouldn't let our machines be used to kill people autonomously. Hallelujah. But the other is we have a constitution and we're building alignment into this machine.

58:09I can always contend that that's an impossibility. And we better get our heads around that and understand that. The other interesting thing is that others who are critical of the AI companies, Timnit Gebru and Margaret Mitchell, criticized his presence. So why there should have been other people who should have been there instead. But I think you're right, Robert, that it was an open hand to say, you have the power, let's have this discussion. Right. And the corollary event to this actually happened before. Again, it was the Thursday before the release of the encyclical. It was World Communications Day, which we do every year.

58:45But this year, we actually did a workshop up at the Colegio Urbaniano, which is next door to us. It's literally the next campus over. And they had some heavy hitters in their panels. And what they made clear was this is the start of a conversation. It has to be a conversation. And it shouldn't happen just in the company. Correct. The Vatican can convene that conversation. It is killing me that I couldn't be there because of my damn back. Next one. We'll get you on the next one. It would have been an amazing event because you're exactly right, Robert. It's a conversation. And what the Pope makes clear is it shouldn't be held by a dozen people in control.

59:20It's got to be held by more. The other thing that I quite loved in this is that he doesn't define artificial intelligence, but he does define human intelligence and then shows that artificial intelligence cannot be sentient, cannot be – he doesn't use that word, but is not going to be conscious that human intelligence is a special and unique quality and artificial intelligence is not that. And I'm on team Leo there. Yeah. AI is limited to calculations of probability, which it is. It's a probability engine. And he says human intelligence includes things like self-awareness, morality, love, sacrifice, and this openness to transcendence.

59:59You cannot fake that. That's not algorithmically possible, or at least we don't know how to do that yet. So, again, he's not saying that AI is deficient and incomplete. He's just saying it is not the same thing as human intelligence. I think you can fake it. I think people fake it all the time. And I think there are a few AIs that also do a pretty good job of faking it. That's part of the problem of deciding whether it's conscious or not. He does say that, you know, governments have to get involved and there have to be legal institutions, but he's not very specific in his recommendations here.

1:00:37I would guess that that's probably typical, that that's how these things are. That's by design. Yeah. You cannot have an encyclical that specifically calls out a government. You do this. Yeah. Yeah. And actually, that was one of the criticisms. He's a moral leader. He's leading us in a direction rather than saying what to do about it. Have the conversation. And I don't think he could. If he were to offer the chance to write statutes right now, I don't think he could do it. Well, that's part of the problem. It is hard to know what to do. There's no instant solutions here, right? So the argument about whether Trump should have or shouldn't have signed the executive order, I don't think it's something you can just sign one piece of paper on and say, okay, we're safe now.

1:01:17The same way we can't have guardrails and say we're safe now. Same way we can't say that we trust this company with its constitution and we're safe now. No, this requires a kind of discussion that started here with 40 ,000 words. I thought it was interesting that Christopher Ola, who's the co-founder of Anthropic, who was there and who spoke, said that every frontier lab, including Anthropic, operates inside a set of incentives and constraints that can sometimes conflict with doing the right thing. in other words we're trying to make money here uh so he's saying the people who don't who aren't informed by those incentives uh should all who care about humans for instance should probably have something to say should have a voice in this right now though that is part of the problem is is the only people who have a voice in this are the frontier right ai labs and government i guess Yeah.

1:02:12I mean, Charlie Munger, famously from Berkshire Hathaway, rest in peace, he had that quote, show me your incentives and I will show you your outcomes. Yeah. And that's the problem because right now everything in AI is driven purely by profit. We have to become profitable, otherwise we die. Well, that's not going to give you the outcome you want. Ironically, that's what Elon Musk was thinking when he found it, at least he says he was thinking when he found it open AI. I believe it, actually. I think that actually Elon did want to have a non-profit, non-incentive by money AI and thought that there would be problems if the companies like Google that were driven by profit were the AI creators.

1:02:56Unfortunately, he kind of lost that fight and in fact is now taking the other side. Yeah, I'm not willing to give him that much credit. trying to make money. I'm not sure how much he lost that fight as much as he thought that if he could keep open AI as a non-profit committed to that different incentive that he, his ex-AI, could actually take that place. Yeah, we don't really know. That's what I said about faking morality. Humans do it all the time. If scale is everything, then you can't get to the scale you need to build what they think that they're building unless you have the money to do so.

1:03:32We can argue about scale there. I didn't see much discussion, a lot of discussion about transparency, about accountability, about the opportunity for appeal, about discussion. Are the words open source in it? Is that something that gets discussed? I don't remember reading that. I don't think so. I definitely didn't suggest it. Yeah. So tell me about how that might fit into this larger ecosystem. Well, yeah. So, again, I think you're getting too into the details that an encyclical definitely wouldn't touch on that. However, you could extend from, at the start of the encyclical, he talks about how AI is a natural product of human ingenuity and human inspiration.

1:04:19And therefore, it would be, we would consider it a grace from God. Towards the end of the encyclical, when he talks about transparency and the need for more people to be involved in this so that the biases are more easily seen and corrected, you could use that as an argument to say he's promoting open source. That's merely a technique to a larger goal, and he's dealing with larger goals. Correct. Correct. The other thing that I, in my point, I am going to make that the headline, the Pope endorses open source. I think that's exactly right. And I think he just didn't want to say it out loud. It's now Pope Floss.

1:04:58There we go. The thing that hits me too is that he talks a lot about power and justice with the technology and focuses. And he's not – if I were any of the technology companies, I would have this be required reading seriously and have discussions of this because I think it raises all the questions. But then he doesn't let the rest of us off the hook. And I'm going to read one more a little bit. At this point, however, a subtle temptation may emerge, namely the thought that the problems are too big and we are too small, and that our choices, therefore, cannot make a difference. This is a polite form of resignation, often disguised as realism.

1:05:35Certainly not everyone has the same power to make a difference. There are those who govern, make investment decisions, lead institutions, conduct research, educate, produce, or provide information. And then there are those who only seem to live their daily lives. Yet no one is without responsibility. We all have our own areas for action, and that is precisely there and nowhere else that we must choose whether to fuel the mentality of force, even if only through indifference, cynicism, lies, or hatred, or to preserve the mindset of peace with truth, moderation, closeness, and care. I think that was a charge to the congregation in the world.

1:06:09I would add to that.

1:06:15Paragraph 238, I think. It's where he is talking about education. This is actually, if you want to talk nitty gritty for priests and sisters and educators, he's specifically talking to educators and ministers. Because one of the biggest threats he brings up about AI, LLMs, is the ability to really amplify misinformation and disinformation. But he doesn't just say, oh, it's bad and therefore we need to regulate it. He says we need to do a better job of educating people. Let us invest in education, beginning with ourselves. We all need to learn how to engage with the digital world in a human way as an integral part of our education in the faith and in the life lived according to the gospel.

1:06:55Indeed, we must consider the digital world as a new continent to be evangelized, one that requires generous missionaries who are mature in the faith. That's a calling to the faithful. but it is, I think, a very important thing to say. We're responsible for educating the next generation and encouraging them. And this is just another way of saying the truth. This is the truth that Jeff was talking about. He says, today, accompanying children and young people and using technology for developing responsible relationships, helping them to recognize their risks and choose what fosters inner freedom is a concrete form of charity and will safeguard their dignity.

1:07:34I think that was beautiful. I think that was 238, right? Section 238. Very good. I'm impressed by your memory. It allowed me to jump right to it. We will quiz you soon. 99 was my favorite. Really? You want me to go back to 99? 99 I loved. Really? I sent it to people immediately. Section, like paragraph 99. It's nice that he numbered them. I thought that was very thoughtful. It's a thing. It's a thing. It's something we do ever since the months with the books. It's not possible to provide a single comprehensive definition of AI. However it can be stated is we must avoid the misconception of equating this type of intelligence with that of human beings.

1:08:17So that's what you were talking about. That's what I was talking about earlier. Robert, maybe you can't talk about this, but I'm curious. Leo is going to do this. He has people who are advising him. He has lots of reading to do. How much playing with – is he vibe coding? Is he getting demonstrations? Hopefully I was not vibe coding. I could pretty much promise you that. Is he getting demonstrations in Claude? I mean, he's reading about the news. He's keeping up on this. Does he say to people inside, show me what this stuff is? I got to get a sense of this. We have presented many, many members of the magisterium with very good examples of LLMs, AI, machine learning, services, hardware, and software.

1:08:59So they know what's out there. they see in fact i've got i've got an ai powered exoskeleton that was i just used in a demonstration for a bunch of partners oh yes it's this this this is that we're using this one oh yeah because it's it's it's a piece of equipment that is that uses machine learning to help those with especially the elderly who have trouble walking yeah um and it's like yeah so when we talk about this technology being used in a useful way that benefits mankind this absolutely is it so yeah it's i'm impressed so they get they get the the examples they they definitely get demonstrations yeah so what we know is is popelio vibe coding probably not i hope not honestly he's talking max and man does he does he have his own personal uh claude codbot running no uh i know of at least three closed models that are being used here yeah well i'm not surprised actually they don't touch the internet at all right they only work on our resources and we've set them so uh like the washington post llm it they it is a perfectly valid response to say i do not know right which is amazing for an ai to be able to say that i tell my ai to say that all the time but it's still reluctant.

1:10:23You can't make them. You can lead an AI to ignorance, but you can't make it admit it. And this I guess my motto. Robert, I'm so glad you're here. It's great to have you. We miss Paris. I'm sorry she missed this conversation. But it was a kind of very timely opportunity to get you on the show. Of course, yeah, talk about this stick around. There's lots more to talk about. You're watching intelligent machines. Father Robert Ballas here, the digital Jesuit is here. And of course, jesuitpilgrimage.app is the app he designed. I just downloaded it. Yeah. I actually would like to go on it. I mean, that sounds like fun.

1:11:00Oh, yeah. Yeah. I think there's great pizza on the Jesuit. I was working on replacing all... I did a couple of narrations in that app, but they were placeholders until we could get the real narrations. So I can't remember if you've replaced me. You're there. So yeah, I might be in there. I think there's still some robbery. Maybe we can get to the Pope in a Jaguar soon. Or does a Jaguar know? It's a Ferrari. A Ferrari. an electronic ferrari an e-ferrari so yeah after this message after this word from our sponsor i'm going to need a lot of these ads to afford a loose let me tell you or a luce as you might say our show today brought to you by zscaler the world's largest cloud security platform man you this is a good time to be thinking about security we are in the midst of a security apocalypse, a myth apocalypse, I guess you might call it.

1:11:49But if you're a company, you also realize, hey, the potential rewards of AI are so great, you can't ignore it. You can't turn your back on it. But you have to be aware of the risks as well. The loss of sensitive data, attacks against enterprise managed AI. It's no accident the Vatican's using closed models, right make sure that nothing gets out uh not most of us cannot do that we have to use public ai frontier models and that's risky but there is a solution there's also the issue of course of the bad guys using generative ai uh helping them to rapidly create phishing lures to write malicious code to automate data extraction and they can do everything 10 times faster than ever before the issue of exfiltrating information accidentally is a big one there were 1.3 million instances of social security numbers leaked to ai applications i bet a few of them and you might even have done this were people who took their tax return and fed it to an ai and say help me with this but everything a bad guy needs is in that tax return and if your ai has been compromised then you've been compromised zscaler is the way to stop this it's the most trusted ai security platform in fact 40 of the global 2000 use zscaler zscaler this is mind-blowing handles half a trillion transactions secures half a trillion with a t transactions a day every day with 9.4 000 global customers zscaler carries a net promoter score of more than 75 that's one and a half times higher than the average SaaS.

1:13:31But the best way to know how good Zscaler is is to go to their website and look at the testimonials from people like Siva here. The director of security and infrastructure at Zuora, they use Zscaler, and this is what he says about it, using Zscaler to prevent AI attacks. Watch. With Zscaler being in line in a security protection strategy helps us monitor all the traffic. So even if a bad actor were to use AI, because we have a tight security framework around our endpoint helps us proactively prevent that activity from happening. AI is tremendous in terms of its opportunities, but it also brings in challenges.

1:14:09We're confident that Zscaler is going to help us ensure that we're not slowed down by security challenges, but continue to take advantage of all the advancements. Thank you, Siva. With Zscaler Zero Trust Plus AI, you can safely adopt generative AI. You can safely adopt private AI to boost productivity across your business. Their zero-trust architecture plus AI helps you reduce the risks of AI-related data loss and prevents against those AI-driven attacks to guarantee greater productivity and compliance. Learn more at zscaler.com slash security. That's zscaler.com slash security. We thank them so much for their support of intelligent machines.

1:14:53buttons yeah this uh it's not really an AI car in fact uh Johnny Ive designed the luce the Ferrari luce its first EV to have switches because he says I don't think we should have too many buttons I actually just read that uh that he said this is the car we were designing at Apple oh so we finally did get an apple this might be the Apple car for 400 000 yeah it's maybe why Apple didn't continue on with it maybe they could have made it cheaper but Johnny said, this is kind of continuing the work I started with Project Titan at Apple, which kind of surprised me. I mean, it does look nice. Ferrari fans hate it.

1:15:35Because it's not the standard Ferrari Testarossa type style? Yeah. I don't know. Yeah, I don't know. Maybe they like that really loud noise coming out of the back. Yeah. I mean, we get that at F1 all the time. People wish we had the old engines back. Yeah, because it's a… Yeah, this new battery-driven F1. But this is objectively faster. It is. That's not a question. Yeah, EVs are actually amazing because their torque is 100 % at zero miles an hour. So they leap off the line. Yeah, it's$400 ,000, but if you buy two, it's$750 ,000. Really? They give you a discount? Look at that. An analog clock, for crying out loud.

1:16:18An analog clock. I'm not the biggest fan of that dash but I do like the rest of the interior well I mean look I wouldn't turn it down if somebody parked it in my driveway I'm not I'm not being that dismissive I wouldn't do yellow though that that yellow does not look good in that shape Chinese media apparently loved that according to Ars Technica and another thing I don't like the tire sizes are different in front and in back yeah isn't that weird I had a Model X Tesla did that with model x and it made it really hard to buy tires you had to you it was so annoying you can't rotate them yeah you can't rotate rotate them i went through four sets of tires in three years in my model x because of that so i just put up video in the chat of the pope uh touring the car oh really he got to see it oh yeah yeah yeah yeah i wasn't joking so i put it in the uh in the discord well it is a nice italian uh italian company uh are they i mean if they want a place look at the franca i think i i feel kind of bad for for the holy father to have to do this it's the new popemobile that's this is well you know what at least what i love about leo is you get the feeling he might have actually driven a car at once and oh he did oh he's like had a normal francis too Like they had normal lives before their elevation.

1:17:39I wonder how much Obama terribly missed driving a car. I wonder how much the Pope misses that. Although Obama's big miss was he was addicted to his BlackBerry. So when they took away his BlackBerry. I remember him complaining about that. Wired Magazine. Stephen Levy did a whole piece on Wired about the genesis of AI bots. he talked a little bit about claude bot uh and then and then he wrote this piece even if you hate ai you will use google ai search um a lot of people hate google's new ai focused uh search we talked about this last week how google essentially said we don't want you to leave the google page anymore we don't want you to click those links we want to give you the answers you need and keep you here which what steven is saying is that they're going to give you the answers and then you'll be happy and you won't care that you're putting ad the verge and everybody else and wired to add a business um i have to point out that duck duck go i just saw the stat today had a 28 increase in signups last week after google announced this uh there are you know there are people who hate ai and all i have to do is say there's ai in it yeah it's the thing to say now but the but the And the polls say that 51 % of Americans think it is the future is worse with it than better with it.

1:19:0870 % of Americans don't want a data center in their backyard, which I get. But usage stats keep going up. Yeah. Actually, I don't know about this. A Wall Street Journal article about the new way to create Google Docs with your voice. i am so i think google is really making a mistake forcing ai down people's throats coming soon to an airport an airplane near you the next time you're traveling to the atlantic new spreadsheet column one earnings uh yeah i'm i'm not going to log into the journal just to see this but but uh uh there that nicole nguyen who is their new technology personal technology or reporter actually uh wrote the article watch me try it so i guess if you really want to see it you can uh you can't do that on the paper in the newspaper you're gonna have to you have to go online shall i turn up i'll turn up the sound let's see here we go uh event called the trans tahoe relay and i want to create a comprehensive resource like so she's prompting it for me and my teammates it's our first time doing this event and so we don't really know what to make she says the bot blundered early i asked for an outline but the over eager ai composed a full draft instead i forced a do over that's exciting that you're participating in that for a comprehensive planning document we could cover things like the event overview rules and safety could you be more annoying google yeah it doesn't have to do that just do the thing i asked for yeah why is it cheerleading me yeah appetizers the wine uh anyway uh coming soon to uh google we use google workspace and it's just uh the user interface is terrible i don't want to summarize this data this is the show rundown i don't need to summarize it get your buttons out of my face i feel like will smith i want to slap it i have to admit that i am more and more and more when i hit a really long post but this is they summarize it this is jeff that's because the problem is teaching you to be oh no it's because this is because no what the what the web taught people to do was to write over long right and take 10 paragraphs to get to the damn point i was raised when we wrote a lead and a nut graph all that's gone it's gone right you got to go forever yeah there's a This started in the publishing industry where people decided that longer books sold better, right?

1:21:49And so you'd have a very thin premise stretched to 8 ,000 pages. Well, actually, in the early days, what you had to do was you had to do multiple volumes. A Twain book was sold in volumes. Was he being paid by the word? Well, they just said you had to stretch the book. You had to get 600 pages. That's crazy. And, yeah. Well, we've incentivized longer posts. We've incentivized multi-part posts because they bring in more clicks. They bring in more profits. So naturally, your AI is going to want to create really long posts. So it takes two sentences of actual factual data and turns it into 20 paragraphs, which is not great.

1:22:32So I feel nothing at all using AI to reverse that. That's a good point. What is the actual nugget of information I need from this article? And it's editorial ego. The New York Times and The Guardian both now have the long read. That's not a sale pitch to me. I'm busy. And by the way, Stephen Levy is the king of that. Yeah. He says he wants to do slow journalism, like slow food. Okay. Fortunately, he's a very good writer, and I love his stuff. I think we mentioned last week that people also get irritated in their office mates when they're whispering to the computer to create the document. Yeah.

1:23:10so what happens to type in anthropic uh who remember briefly was a supply chain uh risk yes aren't they still do they get rid of that no white house and it's not nearing nearing a deal for the nsa to use claude this was on the same day that the encyclical was released so it was it was a very strange confluence yeah yeah yeah uh the white house has also approved a nine billion dollar a request from our spy agencies to buy nvidia blackwell chips so uh they're going to run their own stuff everybody wants mythos right everybody wants the the the anthropic model that is good at finding uh it's the macguffin of ai sorry it's the macguffin of ai is it a macguffin we'll find out when we get to it steve gibson's convinced uh our security guy has spent a couple of shows now talking about mythos and the latest uh news about mythos uh is that it is uncovered now 10 000 zero days wow this isn't 10 000 bugs this is 10 000 cold zero abilities wait wait wait i gotta ask you he's not gonna answer but i gotta ask did they give mythos to the vatican they did not oh if they had i would be using it right now okay he's the guy by the way if you want to give me the guy to the Vatican call Father Robert he's the guy give me would you like to have it Robert uh I it actually is one of the things I very much would like to play with I'll bet I'll bet oh yeah who are we talking to that just got a mythos was it last week I feel like we were talking to somebody who just got this no this show I don't think um so the it i think it's pretty clear that mythos can do what it said on the tin uh it is also fairly clear that almost any um up-to-date frontier model is pretty good at finding security flaws the thing that is uh more and more and i said this when it first came out and it's more and more apparent to me is that the key skill in mythos is its ability to chain exploits and what that really comes to it wasn't designed to be a security model it's a design to be a general model but what it tells me is it has a lot of context it's got a big context window and it can so normally what you would do with any ai is you would find a zero day and that's it that session is done you got the zero day write the concept you know write the proof of concept write the you know validating code if you can maybe write a fix but you're not going to go farther than that because your context now is full.

1:25:58You're done. But Mythos can say, okay, good. I got that one. Let's remember that. Now let's do another one. Now let's put them together. Now let's really put another. And that's what really good accomplished hackers are doing is chains of, right, Robert? Chains of exploits. Correct. Yeah, so I mean, the advantage that Mythos has, because there have been companies that have tried to do this. I'm thinking of Tipping Point, you know, 15, 20 years ago. They were really good at it, but they had appliances. They had appliances that you would plug in and it would do this kind of work. The advantage that this has, that Mythos has, is it can essentially take the CVEs, which all use the same basic structure to explain how the exploit works, where it exists, and where you might find it.

1:26:43And so you just load it with all these CVEs and you let it do the drudge work. It's not doing something incredible. It's just doing something that would be extremely boring for a security professional. uh well it's also creating 27 year old bugs in fact steven steve yesterday read a piece from um the company that created cves 26 years ago saying we designed this as uh you know humans would find the flaws and then we would register the floor we designed a process that operated at human speed we are not prepared for 10 000 new cves yesterday we cannot handle this we need a new system It's broken, basically.

1:27:26I mean, I would love to run this on our enterprise because we've got literally hundreds of thousands of seats connected across our different networks across the globe. Because I know for a fact that many of our organizations have not patched their systems, and we will find hundreds of thousands of unpatched zero-day bugs. That's what's so scary right now. We know those bugs are there. Now, the other question… I'm envisioning a confession system. I have not patched my system. Forgive me, Father, for I have not patched. It has been 121 days since my last update. I'm sorry, I don't mean to be sacrilegious.

1:28:06Mythos found a BSD, an unpatched BSD exploit that was 27 years old, which means it's been running for 27 years, and no one even noticed that it was never patched. Can I ask a really accurate question? Because I don't understand security well. Is it checking against known exploits?

1:28:29There's two parts of it. One is it has the entire library of known exploits, known CVEs. And so that it just scans for them. But it can also find new CVEs and generate new CVEs. It's pretty amazing. Yeah. Steve was referring to a document from Cisco called Shields Up, guidance for defending in the age of ai enabled attacks in which they're basically saying stand back uh because you're going to see a huge number of new attacks now what steve said and i think this is really important is that at some point it will slow down that oh yes by using all of these tools we will get cleaner and cleaner and cleaner which is good news yes eventually Very good news.

1:29:19So this isn't going to be the new normal. There's going to be a tidal wave. We're going to have a lot of problems. We're going to have a lot of fixes. Or to Robert's second point, is it going to invent new ones, new exploitations? Absolutely. That we never imagined. But at a scale we couldn't have imagined. Until the code is bug-free. Is it ever bug-free? Yes. That was Steve's contention, is that with the help of these tools, you can make bug-free code. Now, there's certain... Well, let me finish, and then you can have Robert weigh in. There's certain things you're not going to fix. There's human error.

1:29:52You know, people are going to open phishing emails and enter their passwords into a bogus site. There are hardware errors like Rowhammer, which Mythos can't fix. Now, you could, of course, fix the microcode, and maybe microcode will be better designed down the road. But Steve was of the absolute opinion that with the help of these tools, not necessarily writing the code, but without these tools finding the bugs, that you can eventually eliminate bugs entirely. Now, that may be a little outlier of a kind of opinion. I don't know. I agree and disagree with Steve. The disagreement comes in the fact that if we're talking about using a solution like Mythos on top of what we're currently using in the enterprise, it will never be bug-free because the supply chain that we have for the creation of these products, correct.

1:30:43However, I think his position is this will drive companies to start using better languages to start. That's the second part. Yeah. That's the part where I think Steve is right, where if you allow the lessons that we've learned from using a tool like Mythos to go back into the supply chain, to go back into the actual creation of the hardware and the integration with the hardware and the software for the services, then you can get something that's bug free. Because from start to finish, you've thought about security. Yes. The problem is right now, our way of thinking is security comes at the end.

1:31:15Right. And if security comes at the end, it will never be bug free. Right, right. So this is the report that came out May 22nd from Project Glasswing from Anthropic, an update. And this is the list of what they've found. After one month, most partners of each, remember there are 50 companies that got Mythos, found hundreds of critical or high severity vulnerabilities in their software. Collectively, they've found more than 10 ,000. Several have told us their rate of bug finding has increased by more than a factor of 10. For instance, Cloudflare found 2 ,000 bugs, 400 of which are high or critical severity, across their critical path systems.

1:31:56And this is interesting. With a false positive rate, you know, in other words, a bug that really isn't a bug, a vulnerability that really isn't a vulnerability, with a false positive rate, the Cloudflare's team considers better than human testers. so that's important too uk ai security institute we reported on this uh said that the mythos preview is the first model to solve both of their cyber ranges simulations of multi-step cyber attacks end to end mozilla found and fixed 271 vulnerabilities in firefox 150. 10 times more than they found in firefox 148 with opus 4.6 so eventually i think it's fair to say firefox will be vulnerability free or at least very few i don't expect that every time another firefox fox comes out they're going to find hundreds more question then what can you i'm not a security guy um is it a bug or is it um just something you couldn't have seen that a person or a smarter system saw as a way to exploit Is it something you should have seen?

1:33:04Or is it something, oh, I never knew you could do that. Oh, geez. The number of novel zero days is dramatically decreasing. Most zero days that we see today have built up on previously found exploits. Oh, I didn't know that. That's interesting. Yeah. So the actual real zero days that need research, those don't happen nearly as much as they used to. I mean, even at DEF CON, when we're doing the bug hunts, you start with the CVEs because you know that if that was a vulnerability, there's probably vulnerabilities around it. They may have patched that particular thing, but that's a good place to start.

1:33:41That's actually a good sign. That means programmers are not writing buffer overflows. You know, they're not making, you know, off by one errors. They're not addressing an array outside of its range, things like that. Those are dumb mistakes. Extremely dumb. that programmers nevertheless often make, but I guess they're learning. One of the issues, this is the language issue, is that C and C++ completely allow those mistakes. There's nothing to stop it. But a modern language like Rust will not let you do that. So moving to better, more modern languages is another solution. I don't know. I still love my C.

1:34:23I love C too, but if you can address a random... arbitrary area of memory. That's not a good language. The Latin of programming. C's beautiful. It's elegant. It's simple. It's, it's wonderful, but it's a little bit. It will allow you to do anything you want to do, which is both awesome and extremely bad. That's what they call in the business, a foot gun. Expo, one of our sponsors, actually a pen testing company reports that mythos previews is a significant step over all existing models on its web exploit benchmark provides absolutely unprecedented precision on a token for token basis exploit bench and exploit gym two recently released academic benchmarks show mythos preview as the strongest performer and on and on it's the gold standard by far it's the gold standard the only thing that makes me feel bad about it is the fact that bug bounty competitions aren't going to be so much fun anymore.

1:35:24No longer be what it used to be. So I think that, you know, early on, Robert, there was some question, well, is this a marketing ploy on Anthropik's part? Are they, you know, that's what I thought? Yeah. I mean, honestly, they're leading up to an IPO. Do you still think that or no, no? No, no, no. It's real. It's pretty clear. It's real. It looked like, oh, no, this has got to be an exaggeration. This is marketing speak. But no, it really does do some amazing things. Here's an example. One example of an open source vulnerability that Mythos Preview detected was in WolfSSL, which is an open source crypto library that's known for its security as used by, get this, billions of devices worldwide.

1:36:07Steve talked about this on the show last week. Mythos Preview constructed an exploit that would let an attacker forge certificates that would, for instance, allow them to host a fake website for a bank or email provider. the website would look perfectly legitimate to an end user despite being controlled by the attacker this is cve 2026 5194 it is now patched is the only reason we know about it that's the other thing 10 000 vulnerabilities because there's so many many of those are not yet patched and so they're and so we don't know about them yet because they're waiting they won't talk about them yet but they shouldn't right they shouldn't until they're patched uh this is the flow chart uh uh 23 000 findings by 1900 reviewed by external security firms can you show my screen um 1726 confirmed valid which is 90 true positive then of those 467 reported to maintainers uh actually bypassing this whole triaging of 1129 of them were reported directly to maintainers by anthropic at the maintainers requests before they figured out if they were false positives they wanted to know right away 88 security advisories 97 patched upstream 1451 acknowledged by the maintainer so you know this is as of may 22nd we are going to see a tsunami about this model though is it would be so easy to weaponize this you give this model to the wrong people.

1:37:42Absolutely. Oh, my God. And so can you, and I think we'll get to this in a second, they're going to come out with a consumer version of it, or at least an enterprise version of it. Right. And I think the guardrails against odd creative things, you know, tell me how to make a bomb are hard to do. What about guardrails against this capability to find vulnerabilities in software? In a way, you don't want to guardrail it because you want people to find them. But on the other hand, you want to guardrail it because what are people going to do with it? yeah no that's how do you if you're anthropic how do you how do you manage that i mean you probably just follow this the same rules of responsible disclosure that we've been using in the security world for the past two decades um you you make sure again that are at human speed unfortunately yes that's that's true yeah this tool 100 it will make it easier for people to exploit zero days because because humans won't patch everything no matter how well you follow the guidelines for responsible disclosure.

1:38:41So the only real question is, is it still a net positive for you to be able to expose these vulnerabilities and do as much patching as you can? I think it is. I think it's worth losing the people who aren't doing what they're supposed to be doing with their machines if it means that the vast majority are going to be safe from exploits and WTF bugs. Does this bring back kind of a Y2K short-term industry of fixing old software? Steve actually called it. That's exactly what he called it. At the end of the report, Anthropic says, we're going to work with critical partners, including U.S. and allied governments to expand Project Glasswing to additional partners so that we're going to go beyond the 50 companies.

1:39:24And in the near future, once we've developed the far stronger safeguards we need, we look forward to making Mythos class models available through a general release. And some of the interest in that is not for security. if this model is that good it'll be good for software engineering as well right uh he's now they write at the end and this is the answer kind of to that question on the far side of these risks there's an encouraging world available to us one in which important code is hardened far better than it is today i mean that's what we're seeing is how bad this code is and which hacking is far less prevalent there are many obstacles but we're nonetheless confident the project glass one can help us get there i mean look i have mixed feelings about anthropic i think they're a little woo woo to be honest yes yes um i think like all the other they pretend not to be profit driven but i think like all the other this is like leo criticizing his father-in-law right i've moved on claude is his girlfriend i have moved on from my girlfriend but i've broken up came from this family i actually downgraded my uh my anthropic subscription to the did to the regular person 20 bucks a month wow what happened well you know ah deep seek deep seek happened uh-huh deep seek now i don't you know i'm really i'll tell you why in just a second you're watching intelligent machines this is uh the show all about AI and robotics with Father Robert Ballasier filling in for Paris Martin.

1:40:58Oh, so nice to have you, Robert. Perfect timing. And I'm glad you're doing well, too. It's great. It's just great to see you. We missed you a little bit. Jeff Jarvis is also here, professor of journalistic innovation and the author of Hot Type. So DeepSeek, and I really wonder why they did this, announced that they were going to aggressively cut the cost of their model which is a by the way for version 4 pro is i would say really close to sonnet if not opus quality it's down to 15 cents for a million tokens where you know anthropics maybe five ten dollars per million tokens in i mean huge cut so just as just as these american companies are going to ipo needing to say how they're great businesses and have revenue well and that's what i wonder um how come they're discounting this i don't know now yeah so much it's the undercut phase it's absolutely the undercut phase we know it costs them more than that for that amount of compute especially because first of all they're not using nvidia chips right yeah and they have unlimited power uh right but that still has a cost i mean it's a resource yeah um i yeah is government subsidizing it are they uh is the government is capturing the information you're sending to deep seek i bought harper reed who is basically a dangerous person to know convinced me to spend 64 on this little to do hickey it's an esp32 in here and uh all you do is you connect to your wi-fi and at no cost to you deep seek is in here and you can talk to it and it'll answer your questions and everything i'm not going to turn it on though right now so i can get spied on for only 35 a month well that's the question with your huawei phone you know so well i said that i said well i'm giving this my wi-fi password that means i'm putting it on my network he said leo is there anything else on your network made in china everything i'm waiting for my byd car i'm waiting for my huawei phone i mean i use ubiquity uh hardware for my router made in china it's american company but everything i'm looking at oh yeah so my phone what's one more one more thing made in china i don't know i uh yeah so it's but it's very good for agenda are you are you are you using deep seek yeah I'm my my Hermes my little Hermes have you tried Hermes yet Robert I have not we are gonna get Jeffrey Cannell back he was noose research he was on the show a few months ago when they were talking about their ethical models things like that but what we didn't know and he didn't tell us is that they had in the lab it had for some time a agent harness called Hermes that they were using in-house when openclaw came out they said well ours is better than that and they released it and you know what it is better it's fastest growing agent on GitHub right now okay all right and I moved to Hermes agent uh and I'm using it uh very effectively I think with the deep seek I wonder let me ask it uh about Tiananmen Square I wonder yeah that's the test when was that uh 1991 yeah with tank man hold on 1989 89 okay i knew it was in high school somewhere i didn't know what i got we i use this as a test because chinese models will immediately say no whoops uh apparently it will tank man even yep there's the photograph um yeah i had problem saying that even though i'm using a chinese model um tell me about the ouijers well also i mean that's an llm thing because once you start toying with the data set to remove things you don't want you kind of destroy the llm right but they were doing that notoriously they were doing yeah and so was musk and that's why his and so was elon with mecca hitler yeah and nobody trusts crock to this day as a result right well there's plenty of reasons not to trust chronic.

1:45:27Not just that. Yeah. So this is absolutely forthright about the Uyghur re-education prison camps in China. Oh, well, I'm sure that doesn't work the same way in China. That's probably the difference, isn't it? But see, they couldn't block the processing within the LLM. They'd have to block the result. Yes. Well, no, but I'm using a subscription to DeepSeek that is calling their servers. Oh, right. No, but what I'm saying is, so the LLM, the inner workings of the LLM. Oh, I see. Anytime you try to poke into it, it's going to mess it up. Oh, you're right. It's the post-training. You have to block the result.

1:46:04Yeah, so maybe my agent is not doing their bidding. I don't know.

1:46:12By the way, this thing is getting so smart. I can now say I hooked it up to, one of the things I like about Hermes, and we'll talk about this with Jeffrey Kahnel. He's going to be on June 10th to talk about it because I'm very excited. By the way, he's also a devout Catholic. And when we was on last time, remember, Jeff, we asked him about his faith and AI and technology. And I thought he had some very good answers. You know, he's, it just shows, you know, doesn't hurt. You can be a person of faith and a person of tech. They're not necessarily. You taught me that, Robert. You taught me that. And I'm very grateful to you for that.

1:46:50Let's talk about Muzak. Have you played with any of these AI music generators, Robert? We tried briefly because we wanted to get rid of one of our asset services that we use for some of our content. None of them, I mean, some of them are okay, but none of them actually made something I enjoyed, I liked. Well, okay. Now that's an interesting test, isn't it? Yeah. Yeah. Something you liked. Well, something because we needed something that we royalty free that would fit a particular mood. And I thought, oh, an AI generated tool, an AI tool that generates music from popular selections that might be noticeable to people, but copyright free was very, very enticing.

1:47:40We it was good, but it never really fit what we were trying to do. um let me just uh say let's see oh could i i could paste the encyclical in

1:47:55forgive me father for i am about to send 40 000 words and say make music from this uh probably shouldn't shouldn't put all 40 i'll just which paragraph 238 was that it yeah just do uh paragraph 238. Let me open this up here. Scroll down. And I'll have it be like a choral work. Right? Something beautiful. Oh, absolutely. I think it should be monks chanting. Oh, Gregorian in the style of Gregorian chants. Do you like Gregorian chants? I love Gregorian chants. You might hate them. Oh, they're beautiful, aren't they? And have you ever lived in a community where they do the even song and the we have had tasey here love tasey i went to a tasey service that is so cool it's originally french right yes and it's not catholic it's everything yeah this was at an episcopal yeah uh in the style of a gregorian chant okay let's see I have taken section 238.

1:48:58Program 238. This is Suno. I think it's actually not using my pro account. Or maybe it is. Yeah, I think it is. The title of this song is Let Us Invest in Education. Oh, my God. Oh, wait a minute. No, it's changed the title. To Adult Artisans. Here we go. Let us invest in education.

1:49:27We all need to learn how to engage with the digital world in a human way. Oh my god. As an integral part of our education in the faith and in a life living. This is better than me reading it. We must consider the digital world as a new continent to be evangelized. Oh my. Oh, that's amazing. Now I have to go back to those tools. Okay, because this. Oh. in a particular way we could use this at the next city we discover them I think this is good this is a Tuesday so tell us about the Tuesday

1:50:08I don't know what about the spoken word part interesting how it could switch to spoken word yeah yeah that is we don't do that in Gregorian chant there is no spoken word it was just too hard uh let me say no spoken word no i didn't that's enough i've played with this enough people hate it a gregorian chant that explains six seven that's that would be uh useful for today six seven to adults all right let's try that because at first let's understand the meme i don't know if it does i I don't know if it goes that far. It might just think the, yeah, the numbers. It might go six, seven, eight, oh, now it's nine.

1:50:52Actually, this is a good test. Is it meme aware?

1:50:59It's ramping. Yes, it's thinking. It's thinking.

1:51:06This is not a Gregorian chant. I don't think so. This sounds more like Enya. Yeah, this is from Vikings, I think, the TV show. Yeah.

1:51:19I think it just ignored your prompt. Yeah, it completely did. Let me, I'm going to say, write a full song. Explain six, seven, two adults.

1:51:34Now, the lyric writing may actually have a connection to a model. Okay. select this option now this is just that was just instrumental this is uh something that the a little thing i like to call there are people including harper reed who use don't listen to actual music listen to their suno generated music like it's so fast that by the time the first song's over it's got another one done that's a little weird If you heard it fly by and you missed the joke It's not a secret code, it's just a thing they wrote Kids say it fast like it means the room Half the time it's nonsense, half the time it blooms Don't overthink it, that's the trick It lands like a shrug, then it sticks a number with swagger a wink not a clue they say it for the vibe not to hand it to you this is pretty accurate that's the whole scene that's okay that's not actually not bad it's amazing for the sake of contrast with earlier errors Leo if you go to line 166 you can listen to Kmart Muzak I will get there I don't want to that was a pick that was your pick I don't want to spoil it it's okay it's okay you can use it i wasn't going to use it it is there is a whole so universal music and spotify have signed a deal to allow fan-made ai covers and remixes on spotify so it's going to be based on getting getting slopped up they are slopped up they are all the way slopped up uh but they don't have to pay those those creators and royalties correct and universal music so So Spotify is working with all the big music companies.

1:53:39But this is the first deal is with Universal Music. Remember that Spotify has an issue, which is their revenue is tied to what the music industry lets them make, basically. They don't control their own fate, which is problematic for a company like Spotify. Here's an article from the Financial Times, if i can log in spotify chief defends ai generated music streaming app strike steel allowing universal allowing subscribers to create controlled covered and so it's a way of controlling it i guess that's universal's point of view this is their new new ceo alex nordstrom um the tool will cost extra money it feels so disingenuous it really does all he should say is look it's gonna make us more money yeah well that's the bottom line on all this alongside AI narrated magazine articles for a dollar nine each alongside AI narrated books alongside AI created podcasts it's full-on slop well there's apparently a real problem on you audible has a real problem with YouTube now because people are the 11 labs and other voices are so good at reading books as really almost like 90 % of the way to professional readers that people are taking books taking the e-pub version of a book feeding it to an ai and they're putting the audiobook on youtube undermining uh of course i'm going to record my book on monday might save me the work right are you doing it on monday you're going they scheduled four days for how many thousand words i don't know you didn't count the words i forget more than 40 000 i think it was about 100 000 words so um 11 labs now has a new music generation model i'm not sure i really like it very much i'll play some for you 11 labs um they are the best at voice I mean nobody does better than Eleven Labs voice but their new Music V2 well let me just play some for you and you you you can be the the judge let's see this is their this track is an unedited one shot output actually that's not bad 11 on my wrist 11 on the rise 11 on the mic every level intro okay that's bad i think suno's got this uh market uh locked up let me see if i can find the um the samples yeah there's some samples on here so i just play a few of these and you can you can be a judge I had it do a I try to have it do a intelligent machines but never any quest for a theme song it was terrible this is country

1:56:51that's okay it's just generic right all this stuff is generic I mean when I in these days when I'm listening to music it's because I'm connected in some way to the creator, to the artist. I've been listening to a lot of 21 Pilots. I really like the vibe that they put into that music. I wouldn't have the same connection to an AI-generated piece of content. It's just, it's not there. You don't. There's no human in the mix. If you're in a dentist's office and you want something in the background that you don't notice, okay, AI can make that. Well, okay, so I was in the dentist's office this week, got my teeth cleaned, and I said, did you choose this music?

1:57:28The hygienist said, yeah. it was all the songs i played when i was a dj in 1984 adult contemporary and i said i can name every single song you're playing in the first two notes not only that i could tell you what cart it's on and i could talk it up too i could talk up the intro i little river band reminiscing like i haven't heard that song in 40 years but two notes in i go there's reminiscing yeah and those are the worst terrible songs i mean like bread baby i'ma want you baby i'ma need you terrible songs so it's not like humans can't make bad music on their own without any help but and would you would you remember an ai generated piece of music 30 no actually good point yeah okay that's a good it's all the same it's all generic i do kind of like that intelligent machines theme i wrote some do you remember that jeff it was pretty good yeah so when i worked at ponderosa steakhouse in my youth uh this is by the way robert so poor paris who is a young woman has to listen to me talk about playing little river band reminiscing in the 80s and and jeff has and has to bring up jeff's experiences working at ponderosa steakhouse almost every week in a cowboy hat with a string tie and a red check shirt was it a little did you look like like oh it's dorky as could be dorky as could be but the relevant point here is we figured out we had to wear a cowboy hat oh yeah oh yeah um we figured out how to hack the musak machine oh oh okay because musak was broadcast right private broadcast right to our to the thing we figured out how to make it turn into a real radio and we got in trouble and music wasn't real music although it was real melodies right well go to 166 and you can i know i have it because attention kmart shoppers the entire kmart music collection had the cassette is available on our internet archive is there one that you really like i just i just played the first one the tape-a-thon yeah this is 19 october 1989 ladies and gentlemen imagine yourself walking down the aisle of the kmart the blue light specials you want to scrape a little bit okay this is as automated music as you yes it's awful people people are playing those instruments though I don't think that's the sad part right in 1989 I don't think they were synthesis go to LA as a studio musician this is what you're playing so bored he's got a cigarette hanging out of his mouth yeah he's looking at the chart he hasn't practiced it you never saw it before right because it's just this see this is the same thing I must have heard this I must have heard this at some point in my youth I remember it not even a little bit it's wallpaper but that's what happened that's what's happened to a lot of music is it's become wallpaper Oh, wait a minute.

2:00:36Oh. Got to the jazzy bridge. See, I remember that vibe, that kind of vibe from a lot of TV shows in the 80s. Attention, Kmart shoppers. Blue light special. Yeah, this is just, so we're doing AI music. It's just this. It's going to fade from memory. And now back to lug you into insensibility. I guess that tries to wake up the staff once in a while. In their red vests. Oh, my God. but i gotta tell you 870 000 views on uh on internet archive and 664 people have favorited that cassette it's one i'm gonna favorite it too let's increase that number oh wait a minute i'm not walking i can't yeah you know what i do remember from that time uh a jingle which we can no longer play in the state of california do you remember the uh 1877 cars for kids oh thank goodness it's gone i am so pissed off yeah i donated my last car i we did too we did too and because i thought kids would get it who's who is the state of california has banned those ads now in the united states they're not the only state that has banned them why robert tell us why because it was fraud they were telling them oh yeah we're using these cars we're going to auction them off we're going to we're going to pay for programs for kids it was all going to send families and teens to Israel that was it that was they built a giant welcome center in Israel with 85 million dollars yep that's where my car went I'm sorry sorry but the jingle that jingle is in my head they had kids I'd bring it the radio every single time it comes on that's what comes on to me so it's amazing to to think of the jingles that stuck with me from the 80s like mostly there's cigarette commercials in my mind exactly they're all like have a lot have a lot today more pork sausages did you know do you remember the little kid sitting on the pier with a little plastic fishing rod singing the Oscar Mayer Bologna song.

2:02:53Oh, yes, of course. I wish there was an Oscar Mayer wiener. Did you know that that was an outtake? It recorded in 1975. It worked. The kid, the whole thing, they were just setting up lighting and stuff. I guess they thought they were going to do a line with him. My bologna has a first name. It was so good. the director said, cut. That's a take. That's it. We're going home, kids. I did not know that. That is a fun fact.

2:03:40And that's why at the end, the kid says, how's that? And then takes a bite of the sandwich. Well, we got the sound check done. bring in the actor no i don't think so what i don't think so um you're watching intelligent machines i don't know where i don't know how we got off on that tangent jeff i'm sure by now there is a picture of you in the ponderosa steakhouse wearing a little cowboy hat oh no in the discord i'm hoping i'm hoping oh no now they're doing the where's the beef ad uh by the way who directed that he died recently yeah martin scorsese no no but he was famous for making uh making that ad and many others ads with attitude ads he recently uh uh there is a picture of me at the ponderosa is there ah there we are i knew there would be yes i knew there would be i did not work at the ponderosa steakhouse i worked at mcdonald's uh but i was responsible for for uh breaking the uh the shake machine so i did have the mcdonald's hat i worked at mcdonald's as well um i gotta throw a little uh artistic feedback you know rick salmon kind of has helped me with this uh i really imagine it more like a little plastic woody hat like for toy story okay can you fix that just like a red check shirt yeah just make him like it make him look more like woody you can write you know the kid's name on his shoe if you want why my father tom are you father tom did you see the one where we're all priests yes i did i did i did that i thought so this one i thought was so fast oh they're so good um so we are all in clerical collars including our guest rick salmon at the top of the show but i thought each of us looked like a particular kind of priest now they didn't have to do a lot of work with you robert you were already wearing a clerical collar but each of us i looked like uh maybe i was secretly a drunk i don't know um was that the one that was sepia tone the breaking bad version uh let me see i'm scrolling back to try to find it i can pull it up this guy if you're not yet in the uh in the club you're missing out on some of the most well there i am as pope leo the 15th but i i don't think that's that i i wouldn't I wouldn't dream of that um let's see it was right at the beginning I don't know where you're going with that one it is such an active discord what is what is that uh yeah there I oh he was making a girl with a pearl earring that's right out of me I got it brandoid brand is very good that's actually a pretty good girl with a pearl earring uh yeah these guys discord uh power user tool is we have a slot board where all this goes to and you can just see all the images oh where is that oh there's a slot board of left general category oh how nice so if i uh not starboard the latest is we're all slot boards oh oh it's hidden for you why am i hitting jeff you know why because it's uh yeah i'll put a link in the live chat.

2:06:56Why am I not? Why did you not invite me to slop board? I think even I haven't. Well, that was a newer channel than like you probably have to opt in. Oh, okay. If you go to the bottom of the live chat, you'll see a link. Okay. If you're not yet a member of the club, twit.tv slash club twit, this isn't by any means the number one benefit, but it sure is a benefit. I mean, you get some really weird stuff in here. Here I am in the India video begging for rupees um there that's a that is a rembrandt lighting version of the show i like that one yeah um wow here we are in the tree of life these guys have so much fun in there uh twit.tv slash club twit the main i don't know what that is and i don't want to know the main benefit to this is dad if dad can't be there just put a cardboard cutout in and have the ai generate him i don't know what's going on uh in our club but if you're not yet a member you get ad-free versions of all the shows you get um all the specials we do here we are as priests see i look like i have a secret drinking habit uh jeff you look like you're just uh sad and rick salmon i think it's i think it's becoming rick salmon is the youth pastor and uh that's right the cool youth pastor the cool youth pastor uh and and father robert you're just father robert i'm father jeff looks like the uh the priest who's who drinks a bit too much is it is it him maybe it's that's it yeah that's a common person persona i'm the priest who is invested in plastics and I don't really need to be doing this anymore.

2:08:45By the way, will priests be mentioning the encyclical from the pulpit this Sunday? That's a good question. They should. They should. They should. But I mean, there's no requirements that they do. Are they encouraged to? There's no rule. So the way that it works is the Pope may promulgate the encyclical. However, it's still up to the local bishops to do what it is they want to do with it inside their diocese. There is that subsidiarity. So we don't tell them, oh, you have to release this. You have to do this. You have to make sure that you mention this. However, we can tell the priests, you have to read this.

2:09:23Whether or not you use it in your homily, that's up to you. But you have to read this. This is part of your job. Not on 40 ,000 words. Oh, they have to read it not out loud. No, yeah, no. Themselves. Oh, they have to read it themselves. Oh, yeah, good. Okay. Okay, get in for a long one, folks. cancel your broadcast it's well written it's good oh it's very very well written so um we had reported i think a couple of weeks ago that the president was considering a uh executive order saying i got to approve all ai models from now on this is in response to mythos i think uh fortunately cooler heads prevailed uh and some say it was mark zuckerberg mark says no i didn't know it wasn't me some say it was elon musk no it wasn't me but the wall street journal says it was david sachs the former ai and bitcoin czar cryptocurrency czar who made an 11th hour plea remember it was david sex one of the reasons david sex was there he's the host of the all-in podcast with a bunch of other besties one of the reasons he was there in the white house uh was because peter thiel uh really wanted to make sure that the white house didn't you know they actually reversed biden's ai regulations and they want to make sure there were no regulations on ai in the white house but mythos so scared the president he actually thought maybe we should be approving all ai which is talk about going the other way sacks apparently according to the wall street journal warned trump on a call last thursday that it could lead that it could slow down the industry in its race with the chinese competitors see deep seek sacks ordered argued the order would give a victory to the ai doomers uh trump who said he was already on fit on the fence anyway um decided uh all right they postponed the signing and uh added a paragraph at the bottom that said it would be voluntary anyway way you don't have to get approval we just would like to see it first that is what happens when you mandate things based on zero information yeah on zero reasoning because you can flip thought flip-flop back and forth right based on what is more popular there's no ground right or there's nothing it's an empty vessel yeah yeah I mean and also even if he had passed even if he had signed an executive order it would be so unenforceable that they'd basically just ignore it yeah right yeah it's not a law it's just a serving suggestion uh i put this in for paris she's not with us she'll be back next week i'm tired of talking to ai this actually got a lot of comments on uh hacker news uh who who is this guy i don't know uh he's a developer a team leader he's got 10 years in startups i guess he's anonymous he doesn't say uh he doesn't say who he who he is but he says he got really sick of talking about AI now let me click the link again because I lost the page I found GitHub repositories that were spreading malware I asked AI what to do about it it gave me nothing useful so I opened a discussion on GitHub someone replied it was the exact same text my AI had given me I called it out the comment was deleted.

2:12:46Then another person replied, it was the same AI answer again. I worked as a developer at a company, asked a business owner a question about a business task. He sent me a chat GPT screenshot with the answer. I replied, that had nothing to do with my question and everything there was wrong. A minute later, he sent me another chat GPT screenshot. He didn't even read the AI's answer. He just took a screenshot and forwarded it to me. Recently, someone messaged me on Reddit about my post. I replied, they wrote again. I replied again after a few messages. I realized I'm talking to an AI agent. I'm tired of talking to AI.

2:13:21I want to talk to real people. But even when I talk to real people, they forward my questions to AI and send me the AI's answer. See, this isn't just annoying. This is something that we covered two, three years ago, Leo, when OpenAI first released their first really good version of ChatGPT, which is if these models are scraping the internet, especially things like Reddit, for the data that they put into their training. And they're also contributing to the data that other models will be scraping. It's now become an Ouroboros of crap. It's just AI ingesting AI. And we all know that that's not going to lead to results that you want.

2:14:03You're building in hallucinations into your basic training model and data set. That's horrible. One of the books in my AI book series by Matthew Kirshenbaum, who's been a guest, is about the textpocalypse. Ah. You end up with a gray goo at the end of it. But you're the one, Jeff, who said you didn't want to read those long articles and just were going to feed it to the AI. I'll read a good book, but an egotistical blog post that takes forever to get to the point? Yeah, that's a good point. Some stuff doesn't deserve to be read. Right. Some shows don't deserve to be listened to. We're glad you don't think that about this one.

2:14:36That's a segue for you. All the way to the very end, you're listening to Intelligent Machines. before we get to our picks of the week though i do want to give you a chance uh jeff to mention a a mentor of yours a man who passed thank you for this recently uh you wrote about him in your blog and i thought we should uh memorialize don newhouse thank you for that don newhouse uh and his brother cy newhouse uh the all the family owns advanced publications which is newspapers and magazines uh cunning asked uh was random house for a while was cable systems uh he was the last grand gentleman of the newspaper business.

2:15:11He was a mogul who did good. Absolutely. And humble and kind and generous and soft-spoken. And I worked for his son, Stephen, who's now basically the chairman of the company. And Donald said to a friend of mine once, everything Jeff knows, he learned from Stephen, which is true. And Stephen in turn learned from Don, a wonderful, wonderful person. He died at the age of 95 and will be sorely missed. It's a good long life. What does Advance own still? Advance owns Conde Nast. Really? Yeah, so it's Vanity Fair, Wired, New Yorker, etc. It owns 20-odd newspapers across the country. Cy Newhouse founded this in the 20s.

2:15:51Sam did, yes, their father. Their father. With the Staten Island Advance. They never said Advance. It's always the Advance. I don't know why. I never got an explanation. The majority owner of Reddit? Yes. and whatchamacallit, the tech site that they own. Whatchamacallit. It's a good name for a tech site. Yeah, it is. And then they owned a huge chunk of Discover. They were one of the first investors in that. That turned a lot of money. And so they now own a chunk of Skydance, Paramount, et cetera, or Will. And then they've diversified out, but they own. What's one thing you learned? that stuck with you.

2:16:36Steve Newhouse taught me, we would go into a meeting in a place like Yahoo, we'd go out to California, we'd go to a place like Yahoo, and Steve would say, they're going to think that what they really want is content from Vogue. That's not the right thing for the internet. Let them think that they want that and hold back. But what Steve taught me was that the value of the internet is conversation, communion in the Pope's word, that it is connection among people, and that's the value. So Steve is the one who made me hyper-aware and obnoxious about community and conversation online. Nice. Nice. I also remember, the one thing I say about my little thing, is I remember a day early on when the Newhouses, and they were the managers of the newspaper group, sat around, and they used to make a fortune in classified ads, obviously.

2:17:24And no, it's not our friend Craig Newmarker got rid of them, but it was companies like Monster came along, and Monster had a bad quarter, and one of the Newhouses was kind of starting to grin, and Donald pounded the desk very uncharacteristically and said people it's not coming back oh yeah that was a moment that's hard that's a hard thing for somebody in his shoes to realize yeah it's a family business second generation yep you know they're that this is in their blood so it's really amazing that they were able to see that and say it out loud it was hard but they did it Don Newhouse RIP at the age of 95.

2:18:03He was preceded by his wife, who died a few years before him. Wonderful, Sue. It's a great run and proof that you can be a Titan and not a horrible stain on humanity. Yes. That is a good lesson. Because I'm really trying to be a stain on humanity. And I could be a stain on humanity without being a Titan. So it goes both ways, I'm just saying. It can go either way. You're watching Intelligent Machines. Paris will be back next week. Jeff Jarvis is here. Father Robert Ballasare. Oh, boy, we have some picks of the week. I think this is really good. I was at Jeff Atwood's house and I looked over on the kitchen counter.

2:18:48I said, oh, that's great. Who's reading Infinite Jest? And they said, no, look more closely. The cover is exactly right, but it's not Infinite Jest. It's infinite Jeffs. And let me tell you, the text is jeff, jeff, jeff, jeff. 776 pages of jeff, jeff, jeff, jeff. The blurbs in the back are jeff, jeff, jeff. The blurbs in the back are jeff, jeff, jeff, jeff. The description on Amazon is jeff, jeff, jeff, jeff, jeff, jeff. It's by jeff, jeff, jeff, jeff, jeff. And it is, I tell you, for$18, actually 18 pounds, i don't know i'm apparently on the wrong page here uh but uh you can get it in america too uh published last month infinite jeffs and i thought what a great apparently it was just a joke somebody realized they could go to amazon and have it published and they did i think i'll wait for the audio book jeff you could read that i could i could but don't do the uh don't do the ai synopsis because it'll be just jeff jeff be too short yeah infinite jeffs all right do you have um gemini on your page ask it what it thinks of this book no i could ask you want me to ask my ai what it thinks of thanks for this book infinite jeff what's good what do you mean like should i give it the text no it's okay never mind what do you think about the book infinite jeff it's going to think i mean i typed it wrong it's often corrects me it's going to think oh you mean infinite jest no no i don't uh father robert has a pick of the week for us i do and it's actually something that i saw on mike elgin's blue sky page three days ago and immediately bookmarked and liked it it's called is ai profitable someone just probably vibe coded a page that looks at the earnings of different companies uh who are involved in ai spoiler alert nobody is making money except the company selling the shovels and the pickaxes for the gold rush who's losing the most amazon then alphabet microsoft meta oracle open ai anthropic way down at the bottom um oh there's the green video they're in the they're in the green yeah by the way and this is for some reason is not losing much money I don't know why well it's I think that's actually a dearth of information and actually I don't think that the xai uh line is quite right because there's been so many deals that are leveraged back and forth that that loss is actually much bigger it's it's 20 billion that they spent on their infrastructure which by the way they're essentially renting out to anthropic so right for a billion months or something for some huge amount of money yeah they bought a rapidly depreciating asset to rent it out for pennies on the dollars to a direct competitor i i don't get that business model but yeah it's a nice way to visualize i did ask and i showed this the other day my uh my agent uh should i buy spacex stock and it said no as a matter of fact you told me specifically never to buy any spacex stock and then it gave me all the reasons why it was probably not the best investment by the way since we have been on this page is ai profitable yet ai has spent 2.5 million dollars oh well i mean they're consistent yep it's going fast i love this uh actually i bookmarked it too because i thought uh again i thought paris would quite enjoy this thank you robert appreciate it how about your pick of the week, Jeff.

2:22:37Do we use them all up? Oh, no, no, no, no, no. Appropriate to Robert being here on our topic, I'm going to mention a couple of things. One is the Wall Street Journal just came up with an editorial about the encyclical. Oh, yes. Pope Leo Zan manifesto, and the subhead amuses me. His defense of human agency is welcome, but not his faith in the state. I don't think that's where his faith is. No. No, I don't think that's where it is. So then the other thing - It's interesting, though, that the establishment rag Wall Street Journal would criticize the Pope. Yes. For his faith in the state. Yes. Maybe you should have more faith in the stock market.

2:23:17It's very selective reading. Really? Extremely selective reading. It really is. Yeah. So then I was the criticism. I was hard to find criticism, actually, of the New York Times had a column pissed off that he wasn't attacking all of technology. Timnit Gebru and Margaret Mitchell, I think I mentioned before, were unhappy that Anthropic was there. But this was what amused me most. Less wrong, which is always wrong because it's a bunch of doomers, crazy AI doomer cultists who probably were unhappy with their call out in the encyclical. They accused the Pope of using AI to write the encyclical. and then they linked to someone else a lynch zong who went through a more detailed analysis and you can guess claude wrote it claude wrote it because there's a lot of you guess you're right m dashes oh this is why i put no stock in ai detectors at all no this is exactly what proves it all full of crap yeah i personally know the team that was most involved with this encyclical and i i can't wait to tell them tomorrow that they're ai yeah i think i thought you'd enjoy this pan graham which is the one that everybody seems to like uh said that some paragraphs are between 40 and 100 ai while most paragraphs appear to be zero percent ai okay so uh zero percent of paragraphs in past encyclicals are registered as ai but the declaration of independence is almost entirely ai so right yeah i mean naturally nobody's italian version is flagged as ai more so than the english version which makes which kind of cast doubt in the whole process remember how jeff and i were talking about how ai models take the bait the biases of the people who create them and turn them into an institution yeah you just saw an example of that exactly what's happening exactly wow same with blogs yep um i think infinite jess is a conceptually perfect and practically unreadable no exactly the point okay got it it got it it's a joke object not a book 776 pages of jeff as a parody it's wonderfully stupid as a physical artifact it's even better the sheer wasteful commitment is the gag very dada very internet the street finds its own uses for print on demand wow you know this is good this is why i love hermes actually his name is quicksilver but i just call him q i'm just so depressed i'm so upset for claudia that you left her it had to happen oh she was she was eating me out of house and home i couldn't afford her do you mind saying what was it what was the by the way bill by the way because this is my agent so it knows me it says would i read it absolutely not would i place it reverently on a shelf next to infinite jest absolutely yes would i give it to jeff jarvis only if i wanted to start a minor theological incident i don't know why it said that maybe it saw you in the priest's outfit i don't know oh that's funny so it knows about you baby you're not a oh you're not a closed book privacy my agent knows all uh i actually the next project i'm going to do i'm downloading my genome i got my entire genome scanned uh we had uh i had rich robert uh church um what's his first name i can't remember the um a geneticist on and he has started he started some some years ago he started a company called nebula genomics that they don't it's not like 23andme where they do a partial genome they do the whole thing it's many many gigabytes and i'm downloading it and i'm going to give it to my ai along with health information and we'll see what happens you're going to give ai your genome um yeah so we can all create oh the Leo model yeah the Leo genetic model oh it is my ultimate goal I I'll admit this now to clone myself to completely replace myself or at least to take my brain and put it in a machine in such a way that I can live with you forever see the Vatican we we perfected cloning decades ago I'm not the original Robert Balliser there's been many deaths oh i know opus day has a hundred of you in the basement i know that i know robert you're the best i love your sense of humor i'm so glad you're here so perfect i'm so so when i saw this here today it made me so happy oh i love the show now this is great this is fantastic well we'll get you back anytime you've got time uh you and jeff are actually going to be on a twitter episode coming up yes big boy table as jeff talks calls it so there well a little bit less on the encyclical but with it will be a repeat of this performance i'm pretty sure ai is still going to be a topic there will be more ai news yeah although it's interesting because i am kind of i now have a place to stuff the ai news so it doesn't quite overwhelm overload everything yeah yeah because it's what we talk about all the time uh thank you robert great to see you my pleasure what time is it oh it's after midnight you get to bed you gotta get yeah that's right yeah or do you or do you have to ring the bells now do you go up to the bell tower no i i have the 6 a.m mass tomorrow no no no no that's normal that's totally normal that's what happens get out of here and it's sorry thank you no you have to feed the cats before or after god bless you oh i've already fed the cats okay the vada cats jeff jarvis always a pleasure don't forget jeff's book hot type is now available for pre-order his website jeffjarvis.com you will be back with paris martineau next week our guest actually is somebody that you are bringing to the table I believe you said Robert you're sick yes you I mentioned him and you said let's get him on so all right futurist innovator author part is sir MTV history entertainment history yeah so he will have lots to say he's telling people to in the Hollywood to calm down about AI.

2:29:45Calm down. Calm down, he says. So Robert Tersek will join us next week. The following week, Jeffrey Canel, as I mentioned from Noose Research, I'm really excited about this. It's his company that created my agent, the agent I use, Hermes, and I love it. So some good interviews coming up. I hope you will be here. We do this show, Intelligent Machines, every Wednesday afternoon, starting at 2 p.m. Pacific, 5 p.m. Eastern, 2100 UTC. You can watch it live if you want. club members get to watch inside the discord the club to discord but everybody gets to watch anytime on youtube twitch x.com facebook linkedin or kick uh after the fact on-demand versions of the show appear on our website of course twitter tv slash im there's audio and video the video also makes its way to a dedicated youtube channel for intelligent machines great way to share clips and of course you can subscribe in your favorite podcast client and get it automatically club Club members, by the way, now this is a new feature we're adding.

2:30:42We need your help. We're adding chapter markers for the club feeds. And the reason we can do that is there's no ad insertion after the fact. After the fact, that insertion screws up time codes. So we can't do chapter markers on the non-club shows, alas. Not accurate ones anyway. But we can do it on the club shows. So if you're a club member, try it in your favorite podcast app. And we're looking for input on which podcast apps support it and which don't, because not all do. And there is a list of ones that do at our website, twit.tv slash club twit slash chapters. So you can see what we know about, what we don't know about.

2:31:27And if you can help us out, go in the club and tell us if your client works or doesn't work. There is a list here, though, of podcasts that will work with audio. Some work with video, even, and some don't. But we are putting those chapters in, in a format that is, I guess, close to industry standard. But not everybody supports it. Spotify does not. Of course. Nor does YouTube. Nor does Amazon. But we're really looking for help with the Android apps. We don't have a lot of Android app on here. Pocket Cast.

2:32:07Thank you, Jeff. Thank you, Robert. Thank you, everybody, for watching. We'll see you next time on Intelligent Machines. Bye-bye.