Pope Leo Challenges Silicon Valley's AI Vision

00:00:00:05 - 00:00:33:25
Jason Howell
Coming up next, Jeff Jarvis and I dig into Pope Leo XVI. Huge. It's 42,000 words encyclical which focuses largely on AI and humanity. We got Silicon Valley's last minute lobbying campaign that apparently killed Trump's AI executive order before it was signed. Like right at the buzzer, anthropic project, glass wing, and some pretty eye opening early security results with mythos and a growing wave of AI generated lawsuits that's starting to overwhelm the courts.

00:00:33:25 - 00:00:49:28
Jason Howell
That's coming up next on this episode of the AI Inside podcast.

00:00:50:00 - 00:01:09:17
Unknown
To another episode of the AI Inside podcast, the show where we take a look at the AI that is layered throughout the world of technology. This is the weekly podcast, the Longer Form Podcast, where we talk about some of our favorite stories from throughout the week. I'm one of your host, Jason Howell, and over on the other side of the screen is Jeff Jarvis.

00:01:09:18 - 00:01:28:20
Unknown
How you doing, Jeff? I was just, you know, reading late breaking news here. Storyful has a story. A West London council appealed for help. Identify. And a man filmed spray painting the phrase I farted in yoga across construction organs. This is the internet, you know. Here's my friend cabinet about to do a podcast. I'm sorry. Oh, I gotta see that.

00:01:28:20 - 00:01:48:00
Unknown
I gotta see it. Turns out. Squirrel! Yes, exactly, exactly. Sorry about that. I know all about that impulse, and I would have a hard time seeing that headline and not wanting to. Yeah, yeah, he's doing a very good job. He's spray painted on very nicely. But the bigger picture of a fart, I guess, as a ghost. Okay. Yeah, yeah.

00:01:48:01 - 00:02:12:01
Unknown
Sorry to take this off on a fine high level. This is the show. This is the sort of intro that gets more people to come back after week. AI and farting. What more can you. What is the Howard Stern of AI? Daniel Croft. Hello, Daniel. Helps behind the scenes on the podcast says leave us Londoners alone. There you go, Daniel.

00:02:12:01 - 00:02:39:11
Unknown
Good to see you. Good to see you too, Geoff. We got lots of talk about lots of lots, lots always do. And some some of the things you've actually written about before we do, I just want to mention, just kind of remind folks that, hey, started a daily version of this podcast over on Patreon, Patreon.com Inside Show, and actually did some work on the Patreon and they like upgraded the look and feel of everything.

00:02:39:11 - 00:03:01:12
Unknown
But anyways, every single day, well, Monday through Friday, not the weekends. I take the weekends off, but Monday through Friday you get a 5 to 8 minute episode, 3 to 5 stories that are making the rounds. Then if you are a paid at $5 tier or more, you get all of the episodes. If you are a free patron, you get one of the episodes every single week for free.

00:03:01:12 - 00:03:27:08
Unknown
So go there. Patreon.com Sideshow, become a patron and get five days a week of podcasts. Yeah, yeah. Please please please. Amazing. I know you're welcome. You're welcome. I did it, I did it for you all. Okay. We should talk, I suppose, about Pope Leo. Not not not Leo Laporte. Pope Leo. Different different Leo entirely. Holy. Holy Leo.

00:03:27:10 - 00:03:58:12
Unknown
As opposed to. Yes, the Holy One. Leo the 16th delivered his, pretty anticipated encyclical, 42,300 ish words. And so, in other words, a pretty lanky piece, which I'm sure we have here. There we go. Titled magnifica, Humanitas translates roughly to magnificent humanity. And this was officially released a couple of of days ago on May 25th. Yeah. On Monday.

00:03:58:13 - 00:04:28:23
Unknown
That's right. He signed it the prior May 15th, because he wanted to match the anniversary of the signing of his namesake, Leo the 13th Rerum Novarum, which was about the Industrial Revolution. So this is absolutely a continuation on. And in the document, he stretches, he looks at the history of papal views and decrees on the social views of the church from Rome to warm up through today.

00:04:28:23 - 00:05:01:08
Unknown
So he really wants it to be a continuum. And there's really a lot of a lot of interesting overlap, I suppose, from what was happening in the I don't want to say the chaos, but the, the dramatic disruption. Disruption, that's the word. That's the right word. And feeling a lot of disruption now. I guess we won't know until this is in the past whether they actually do compare, you know, at a similar level right now, things feel pretty darn important and pretty big.

00:05:01:08 - 00:05:24:16
Unknown
And so that's part of the reason why he wanted to time it this way. But but if you think back to the industrial revolution, those are huge, huge changes in society. Oh yeah. Yeah. The electrification entirely new economy of working jobs, an entirely new consumer economy, capital taking over the world, communication that couldn't have existed before, transportation that could have existed before.

00:05:24:17 - 00:05:53:29
Unknown
These were monstrous changes in society. And Leo the 13th cared deeply about work and the disruption that people went through. And that carries through to Leo the 14th. So I spent all day Monday and half of Tuesday studying the encyclical, and I wrote a wrote a fairly lengthy post myself, trying to quote the good bits on it. And I found it to be an amazing document learned, wise, eloquent, and ultimately useful.

00:05:54:01 - 00:06:00:13
Unknown
And I don't think that.

00:06:00:16 - 00:06:22:23
Unknown
He fell into present tense traps. There was kind of an obligatory mention of kids and screens, but I think that people might have anticipated, in fact, you know, there was a column in the New York Times getting pissed that he wasn't attacking AI. It's a very thoughtful document, intended not for the moment, but for generations, that tries to get us.

00:06:22:25 - 00:06:48:08
Unknown
That's that's an important distinction and not something that you're seeing in a lot of the coverage. Exactly. A lot of the coverage is really criticizing. Yeah, criticize AI calls for regulation, says kids should be, you know, which I suppose to a certain degree there are truths in that. But yeah, like, what are they missing that you're catching? So he starts with, with a with a.

00:06:48:11 - 00:07:13:04
Unknown
Metaphor here comparing the story, the biblical story of the Tower of Babel to the rebuilding of Jerusalem and the Tower of Babel. Of course, it was a hubristic act of building this tower to the heavens and having just one language in the world and supporting uniformity, eliminating diversity, homogenization over communion, in his words, which of course sounds familiar.

00:07:13:04 - 00:07:37:10
Unknown
It sounds like AI, as we, as many of you see it now versus the rebuilding of Jerusalem, was based on prayer and permission and community and communion, and all the people of the city coming together with their roles to build something, rebuild something back together. And clearly, his point here is that we should be rebuilding Jerusalem rather than building the Tower of Babel.

00:07:37:12 - 00:07:59:11
Unknown
And he's very clear here that he's not attacking technology. He says technology should not be considered in itself as a force antagonistic to humanity. He quotes his predecessor, Pope Benedict, saying that, on the contrary, it has formed part of our history since the beginning as a profoundly human reality linked to the autonomy and freedom of man. And the writing is quite eloquent.

00:07:59:11 - 00:08:21:04
Unknown
I won't read huge chunks because I quote huge chunks in here because it's so taking. But I want to read you just this moment. 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.

00:08:21:06 - 00:08:47:01
Unknown
In practice, however, technology is never neutral. And so he makes that point throughout. Yes, makes that point throughout. And I as a media guy, I constantly see lessons for my field in this, for media, in this, because I think he's also talking about he's saying that we don't want a technology that feeds everything from the top down, which is what media has done for the last century because of its mechanization and industrialization.

00:08:47:04 - 00:09:16:10
Unknown
He warns of control by these huge companies. And that's right. The risk of being misled. And really what he talks about a lot is that is the the possible dehumanization of people that would become nothing but data. Yeah. Well, yeah, there's a there's a part about that. Like with UBS, you know, so many so often with, with AI and job displacement and everything, you know, universal basic income is the solution that comes up.

00:09:16:10 - 00:09:52:10
Unknown
Oh, we'll just we'll just pay everybody. Yeah. We'll just give a paycheck. But I think one thing that he points out that's that's really important there is whether whether we love work or not, work gives us a certain identity and purpose that without it, if we were all just sitting around collecting paychecks as as appealing as that might sound on its surface, that would lead to a whole host of other like problems slash unsettling kind of qualities of what it means to be human, to not have purpose on a daily, regular, kind of existence.

00:09:52:11 - 00:10:20:14
Unknown
Right. And so clearly, he's not a fan of UBI, he says. Of course, as the pope from an institution or charity, that one should give charity to people when they need it in their emergencies, but that that's not a solution long run. He deals with a lot of a very high level concepts, of course, like truth. And I love this line that he had in here, that to understand that the truth is a gift to be shared, not a possession to be monopolized.

00:10:20:17 - 00:10:40:06
Unknown
And that speaks to me, obviously, as a journalist, where I think what we see today is all this effort to put everything behind paywalls and keep our content from the AI companies and and keep it from the public and make them pay. And I understand the economics of all that, and I understand the need of all that. But if we have a higher calling to truth, what responsibility does that also give us?

00:10:40:06 - 00:11:08:23
Unknown
So he intends that as a call to the AI companies monopolizing data, monopolizing the knowledge that might come from it. But I also see it as a lesson to media as well as to government. We shouldn't be monopolizing the truth. He interestingly, he doesn't try to define artificial intelligence, but he defines human defines and defends human intelligence.

00:11:08:24 - 00:11:34:17
Unknown
Thus he defines what artificial intelligence is not. And he says that it can merely create functions of human intelligence. And in other words, it's mockery of them or appearance of them. They do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships, and do not know from within what love, work, friendship and responsibility mean.

00:11:34:18 - 00:11:55:07
Unknown
Nor do they have a moral conscience, since they do not judge good and evil. Right? Which is all to say, I'm on Team Leo here. To hell with these ideas of sentience and consciousness, or the idea that these guys can create an alternative form of life. Godlike. And he comes out squarely against that, I was delighted. Does he call that dangerous?

00:11:55:08 - 00:12:20:16
Unknown
Like, does he specifically say like, hey, pump the brakes here? That's actually a dangerous direction to go because it's because the fear is, and I'll quote that, that it becomes a form of statistical adaptation based on data and feedback, which can be very effective but does not imply inner growth. And he goes on later to talk about how it subjects us to become just shells of humans, not respected as humans.

00:12:20:16 - 00:12:46:25
Unknown
And that's the danger. I was really happy that he particularly called out transhumanism and posthuman ism. The part of the test grill. Google it if you haven't seen it before. We've talked about it here before, kind of cult of views. And I didn't put this on the run, not because it's in German, but there was an academic in Germany interviewed in deep site who talked about how AI is becoming its own religion.

00:12:46:28 - 00:13:04:13
Unknown
And Leo didn't go that far to talk about that, but in a sense he did here. But by calling out transhumanism, which is the idea that you can put chips and machinery in our body and make us better than humans, and post humanism, that you want something that is better than us as a whole. He doesn't use the word but.

00:13:04:13 - 00:13:27:19
Unknown
But to me, it's very clear that he calls out the odious stink of eugenics and says that if the human is treated, is being treated as something to be perfected or surpassed, it becomes easier to accept that some lives are less useful, less desirable, or less worthy. Yeah. That's valid. That is profound. Profound. I totally agree. That really stood out for me.

00:13:27:20 - 00:13:54:28
Unknown
So he talks about thus the nature of humanity. And he's defending humanity in this again all along. He recognizes that there's opportunities with AI, that it's a powerful tool, that it can do all kinds of wonderful things, but it must be done in this context. So he talks about it in the context of work, as you already mentioned, of democracy, of war, very much against any kind of autonomous battle with this.

00:13:55:04 - 00:14:07:16
Unknown
He talks about it in the context of community and communication, which speaks to my heart as a media person and education as well.

00:14:07:18 - 00:14:52:14
Unknown
Then it's interesting to me, Jason, that I'm getting almost to the end. Don't worry that he invited co-founder of Anthropic Yes to come and share the stage with him. And yeah, yeah, it was really interesting because I think I think he said somewhere that he's an atheist or an agnostic. And clearly the Pope was singling out anthropic. I don't think blessing them as, you know, pure or perfect, but the fact that they fought against the Pentagon regarding the use of AI for autonomous killing, I think, mattered.

00:14:52:17 - 00:15:14:07
Unknown
But there's a bit of an irony here, of course, in that anthropic is part of that world that says, well, we're all about AI safety, and they define AI safety in these broad terms of any mankind and all that kind of stuff. So it's it's as a PR move to have him there is interesting. The New York Times writes a story about, well, this man is really influencing the Pope.

00:15:14:08 - 00:15:39:05
Unknown
I think it was more of saying, I can get along with the technology person. We can listen to each other, we can have a dialog, and that's what we should have. But clearly, Leo says that you can't make a moral machine, right? Yeah, I think that's definitely a critique of alignment. Yeah, exactly. So this whole having well, having an alignment company in the room, that's that's the that was key to me.

00:15:39:05 - 00:16:04:14
Unknown
That was the point. So we cannot be satisfied with merely calling for the moral of machines, the so-called alignment of AI with human values, without also having the courage to insist on a further condition, the possibility of openly discussing the ethical frameworks involved and subjecting them to shared standards of social justice. Otherwise, those who control AI will impose their own moral vision, which will become the invisible infrastructure of these systems.

00:16:04:19 - 00:16:26:12
Unknown
A more moral AI is not enough if that morality is determined by a few, right? What infuses this whole document as community and the need for listening and dialog and understanding? And that's again why I think he had someone from a technology company there, and it's going to have one there. Fine. Anthropic. I don't debate that. In the end.

00:16:26:14 - 00:16:37:16
Unknown
He's really talking about this as a matter of technology and power and justice, and he's defending justice in this. And and.

00:16:37:18 - 00:17:01:17
Unknown
Let me end here, a subtle temptation may emerge, he said, namely, the thought that the problems are too big and we are too small, that our choices therefore cannot make a difference. This is a polite form of resignation, often disguised as realism. Certainly, 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.

00:17:01:17 - 00:17:26:29
Unknown
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 it is precisely there and nowhere else that we must choose whether to fuel the mentality of force, even if only through indifferent cynicism, lies or hatred, or to preserve the mindset of peace with truth, moderation, closeness and care.

00:17:27:01 - 00:17:34:23
Unknown
Amen. Yeah. So I think it's I think it's an inspiring document. I think it's a wise document.

00:17:34:25 - 00:17:54:10
Unknown
It's not, as I said before, tied to the minute. It has a larger perspective. I think that media, as you point out, other I don't think they got a lot of it. It takes a lot of study. I marked it up like crazy, and I needed to read it a few more times to try to understand what it is, but I think it's a very positive document.

00:17:54:10 - 00:18:15:24
Unknown
And if I were at anthropic or OpenAI or Google or Microsoft, I would assign my executives to read this. I would then have a seminar to get together and talk about it in the way that we are, and see what it inspires. It's not an attack on the on the technology companies. Not at all. No, it is, it's it's a very poor reason.

00:18:16:01 - 00:18:53:23
Unknown
Yeah. It's a very sober perspective on like much of what I read through out, you know, the piece and your article and, and everything is that there's a lot of these themes that he brings up in here that really tie in with a lot of the perspectives that you and I share from, from our kind of point of view on this show, which I like to think of this show as more of a sober kind of view of the AI industry, you know, the industry at large and its effects and implications, its effects, its implications.

00:18:53:23 - 00:19:26:08
Unknown
Not not going ra ra, everything's great, not going. Everything's horrible either. And that's kind of the the POV that I got from what I've read through here. And the different points is it's a very sober kind of balanced, understandable. Like you said, not being rooted in the moment essentially means that this is a document that's meant to last for a very long time, the way the Industrial Revolution version of this did, and wasn't entirely pinned to the headline of the week or the headline of the year or whatever.

00:19:26:08 - 00:19:48:05
Unknown
And I think when it comes to news media's coverage of this, I think by and large, it's very difficult for people who are in this industry following and reporting on this industry to not tie it into the current moment and to not see it as an indicator of how, you know, how the Pope feels about whatever the news is right now.

00:19:48:05 - 00:20:14:08
Unknown
And this is meant to this has meant in some ways to signify the importance of of all the change that is possible slash probable considering the technology moment that we're in and what that means for, I don't know, maybe decades. You know, I don't know how often does you know does this happen? Is does a pope only release one of these in, you know, in his time or like, what is that?

00:20:14:10 - 00:20:40:19
Unknown
They're not they're not, you know, constantly they release in and letters about various topics. Right. So, so in terms of he he he goes through the, the relevant encyclicals and views, including the Second Vatican Council and how that built up to a policy, social policy around the church. So it's it's it's additive, I guess I would say think I think you're right, Jason.

00:20:40:20 - 00:21:01:16
Unknown
It's companies that can quarters journalists they can days popes. They can millennia. Yeah. That's a good way to put it right. Yeah, that's a great way to put it. And he's here to to make people think for the longer term. You know, and I was not sure what he's a very smart man. And I and I had high hopes for the document, but it could have turned into a moral panic.

00:21:01:16 - 00:21:21:24
Unknown
It could have turned into a condemnation. It could have turned into a woe is US protectionism of the past. It's not that at all. It recognizes the future. And I was quite taken with it. And it sounds like you were too. Yeah. I mean, definitely a lot of most of what I read, I found myself completely agreeing with it and just being like, all right, cool.

00:21:21:26 - 00:21:41:05
Unknown
And it being, I mean, his capability and ability, he's probably working on this for a very long time of wording it in a way where it just seemed to perfectly encompass a lot of how I feel about this stuff right now. Or it's like I couldn't clearly he's in a whole other league, but I couldn't have put it better than that.

00:21:41:05 - 00:22:05:20
Unknown
That's exactly it. Like he was, I guess to a certain degree, he was speaking to my heart about how I feel about these things. And so I, you know, a lot of respect there. What is it about Timnit Gebru, you know, not happy with it. She was saying that it's a it's a merger of religion, of effective altruism with religious institution.

00:22:05:23 - 00:22:32:11
Unknown
She was pissed that anthropic was visited at the event and that the fear I think that she had here and she's critical of it, was that the pope ends up promoting what's underneath anthropic, which is effective altruism and bits of test Grail. Got it. Okay. And so she has been a primary researcher on this notion of test real on this, on this other religion.

00:22:32:11 - 00:22:56:04
Unknown
And and I think that she's disappointed that the Pope invited him in. I was invited to an event at the Vatican last week. I was invited, I'm sure this was through the good offices of of Padre of Father Robert Balasore. But when I was invited, I wasn't sure I could travel, so I had to say no. Killed me, but I had to say no.

00:22:56:04 - 00:23:15:28
Unknown
And so they. And they had some people there who are not like Kristen Harris, who I'm not, I'm not crazy about at all. And so he was there and they screened that AI film we talked about some weeks ago, and I'm rolling my eyes. But then I also saw a picture. Kristen Harris met the Pope. I could have met the Pope.

00:23:16:00 - 00:23:30:18
Unknown
Yeah. And I wish I'd been there to hear the discussion and to hear what they're going around, because I think there needs to be. What he's saying here is we need to have multiple perspectives on this. This could have been a document from one perspective. Of course, it could have been all against technology, all in favor of technology.

00:23:30:18 - 00:23:55:24
Unknown
Instead, it is a document of perspective and context, and it provides, I think, the basis for those discussions. So for that I welcome it. Yeah. And then another reaction. There was really stupid reaction, but I was too amused by this less wrong, which is the ultimate of the of the doomsayers. This is, this is the AI is going to destroy the world.

00:23:56:01 - 00:24:18:05
Unknown
All that people, well, they go after it and someone else and then link to someone else who goes after it, claiming that huge chunks of it are written by AI. And why? Because he uses em dashes. That's all a bunch of BS. All right, all right. So. But I thought it was amusing that that's all. That's all you could go after for.

00:24:18:05 - 00:24:47:11
Unknown
Fine. Yeah. That is that is pretty funny. It is? Yeah. We got you, we got you. You're using them dashes. It is always the sign of AI. Let's just say it before AI used it. You know, I learned a lot of my writing when I worked at Time Inc., and I make people have made fun of time. E's forever because it is a its own argued, but because we had to squeeze a lot into little space.

00:24:47:13 - 00:25:09:02
Unknown
What do we do to do that? We use dashes. Parentheses are too much out of the flow, but a dash like you sneak something else in almost parenthetically, right. So I use dashes like crazy. As a result, I always have. I'm not an AI. I may act like one on podcasting, but this is this is just BS. So I recommend folks, it's long.

00:25:09:04 - 00:25:34:02
Unknown
As Jason said, it's 42,300 words or thereabouts, but it's very readable. It it sparks a lot of of thought and thinking. He doesn't really get to AI directly until 20 pages into the 40 page document, so it's halfway through before he really directly addresses AI. Before that, again, is a lot of church history, which you might want to read more quickly.

00:25:34:02 - 00:25:50:07
Unknown
I'm not a Catholic, so I you know, I don't understand all of that. Though he made his point about the principles he was going to operate under and that the principles were the principles of the church. Right. So he set that stage. So there are some parts maybe you read a little faster, but some parts spend some time on.

00:25:50:10 - 00:26:10:24
Unknown
I quoted at length in my post, big chunks of what I thought were the good bits. I could have gone on forever, and then I would have had 42,300 words. But I do recommend those of you who are interested in AI and its effect and impact, give it a read. Absolutely. Well cool stuff. Thank you for the the run through on that.

00:26:10:24 - 00:26:56:14
Unknown
And I think to some degree, I mean, it's not directly related, but it is kind of from the maybe from the perspective of like the the big tech companies being involved and being close to, you know, these, these moments. Right. Like with the Pope, you had anthropic standing there with President Trump, Donald Trump, you apparently had certain members of the big tech ecosystem urging him, or at least convincing him not to sign on, on governance around, you know, there was an executive order that was expected, basically around AI that was expected to be signed last week.

00:26:56:14 - 00:27:27:25
Unknown
It was on the calendar, and then it just didn't happen. And you have The Washington Post and Axios both basically reporting that a number of companies kind of came in and lobbied against that. And Donald Trump agreed. And so that kind of swayed what happened there, at least to some degree. They had that pressure. David Sachs, who's obviously a Silicon Valley investor, was in the government is not in the government right now, but he's got problems ear he was heavy pushing against this.

00:27:27:28 - 00:27:49:28
Unknown
Others who spoke with Trump, according to the Washington Post, included Elon Musk, Mark Zuckerberg and others. And I'm of two minds about this, Jason, on the one hand, yes, we need we need regulation and we need to think about the impact. But on the other hand, I think part of the message from Leo is you can't do that overnight.

00:27:50:00 - 00:28:11:08
Unknown
And in a sense, Trump not signing this is really a gift to the industry is the industry. That's not the way to go, but neither is it the way to go to throw in instant made laws. And okay, now we've solved it all. We thought, yeah, we did our work. Moving on. This is complicated poop here, man. Yeah, and you can't rush it through.

00:28:11:11 - 00:28:29:03
Unknown
I tried to come up with a better word, by the way, once. It makes me laugh. I'm happy you said that word. I was on the BBC once, and I. And I was going to say the s word, but I thought, I can't say that I'm on the BBC. So I said, crap. I got bounced immediately. Turns out, really?

00:28:29:05 - 00:28:51:01
Unknown
In England, I'm told crap is worse than the s word. Oh wow. Had it wasn't. I wasn't banned forever, but I was dumped, as they might say about the subject. So so, you know, and the executive order of what we knew about it, you know, wasn't awful. It wasn't great. I think that there are people who were going to say, oh my God, AI is a Trump's pocket.

00:28:51:01 - 00:29:18:25
Unknown
I think that's true. But neither can we say that this would have solved everything. So I think it's simplistic to. Yeah, yeah, this was a bad moment. It's politics. Yeah it is politics. Yeah. And I often on you know political stories him ha on whether whether to include them. But nonetheless we had talked about this in advance. The fact that it's not happening I mean probably still could and probably will to some degree.

00:29:18:25 - 00:29:44:14
Unknown
But Trump, at least for now, says he just hates regulation. So he does. You know, that was that was the reason why. Well, and I think that the proper forum for these discussions is not in Congress. You know, having testified once at a hearing about the Internet in Congress, it's scripted. It's not really a discussion. It's not to disturb.

00:29:44:17 - 00:30:14:25
Unknown
It's all show off. I think what what Leo invites here is a much larger discussion. And so we need to bring together the companies, the technologists, the academics, religious leaders, civil society, and yes, government and policy as well, but not alone to have this discussion about where this goes. And, you know, in this section about the moral machine, the impossibility of moral machine and alignment, that's the other quick fix solution.

00:30:14:26 - 00:30:35:14
Unknown
Oh, we wrote some code and we solved it all. No you didn't. You can't solve the line uses the machine, nor can you solve people's irresponsible uses of it. And people becoming in love with it or or listening to it. Yeah. To to confirm something bad they want to do. Humanity is much more complicated than that. Super complex.

00:30:35:14 - 00:30:59:14
Unknown
You can't solve for all of those very different ways. You can't come up with it in advance, and there's no way to have 100%. So we can't have technological solutions nor political solution. We need societal discussion. Yeah. Which is why we have podcasts. That's why you listen to AI inside. Exactly. Big thanks to those of you who support us on Patreon.

00:30:59:14 - 00:31:18:04
Unknown
Patreon.com sideshow I know I mentioned it at the top of the show, but I also like to throw out thank yous to folks who have been with us for a while. Also, new folks who are just coming through the door like Michael Kaiser, one of our newest patrons. Thank you for supporting us on a deeper level, Michael. We appreciate having you here.

00:31:18:05 - 00:31:41:04
Unknown
Patreon.com sideshow. And yeah, lots of lots of new stuff there for you to get if you become a patron like Michael did. So we hope that you will. We're going to take a quick break, come back and talk a little bit about the deep seek and the impact that it's having, because there is definitely a new story here that I'm very interested in.

00:31:41:05 - 00:31:46:07
Unknown
Curious to hear what you have to say as well. That's coming up here in a moment.

00:31:46:09 - 00:32:18:06
Unknown
All right. Let's step back from the drama of it all a little bit and talk about the economics of AI, because Deep Seek has apparently been running a 75% promotional discount for its V4 Pro API, its flagship model. And that discount was set to expire on the 31st of May, which is just a few days away. And instead, apparently they have made a surprise move and just made it permanent.

00:32:18:06 - 00:32:58:24
Unknown
So that's the new list price going forward. 75% promotional discount is what it is now and what Walmart of AI, I guess. So low prices all the time. Act now or any time because you'll always receive 75% off. What does it actually mean? Well, it changes from it, makes it roughly 34 times cheaper than what you might get from GPT 5.5, at least when you're talking about output tokens lands within about 95% of GPT 5.5 performance on most coding and reasoning benchmarks.

00:32:58:24 - 00:33:28:19
Unknown
Not quite there, but I mean, hey, 95% is pretty close. The things that 99% of the people ask it to do don't that line. Yep. Yeah, exactly. Most people aren't going to be pushing it up there. So 34 times cheaper. 94 sorry, 95% the performance. Hard to argue with that. From a cost perspective. I suppose you have to know, you know, for all those people that you're talking about, Jeff, they have to know the deep sea exists and is an option, you know.

00:33:28:21 - 00:33:54:05
Unknown
But if I were to pursue, you know, you have you have a mark Benioff of Salesforce saying he's going to spend $300 million or some such figure on on tokens. Yeah, it's not just individuals, it's companies. And because one of the speculations we've talked about in the show is the reason that the real reason anthropic didn't release mythos is because they couldn't afford all the compute from people using it, because it's more intensive.

00:33:54:06 - 00:34:18:18
Unknown
Right. And you have anthropic and others trying to pull back. There was complaining this week about anthropic is pulling back features, which is a form of price inflation. They're trying to charge more for tokens. They're trying to charge for for actual usage for tokens. So while that all that is going on as these companies heading to AI to to AI, they are AI heading to the public market.

00:34:18:20 - 00:34:42:02
Unknown
They're trying to prove their business sustainability by bringing in revenue. So this is pulling the rug out. I mean, completely right, because this just says, okay, we're we're going to undercut this is the Craigslist. Right. So and Craig is a friend of mine and Craig, no, he did not destroy newspapers. But the interesting thing was and I wrote about this years and years ago, you can't compete with free right.

00:34:42:03 - 00:35:00:08
Unknown
And deep seek isn't free, but it's a hell of a lot cheaper. I mean, 70 yeah, yeah. And so where do you end up then trying to compete with that. And so is this deep sea competing or is this China competing. Is this a way to say no? Yeah. Trump only didn't leave with deals. We got a deal for you.

00:35:00:09 - 00:35:26:18
Unknown
All of them. Yeah. Is this world tanking the value prop of of Western AI company just as a global right. Yep. Yeah. Right. Just as they go for IPO. That's an interesting perspective. You know I, I wouldn't be surprised. Yeah it's it's really interesting. You know what it does for pricing expectations across the entire market. And how does it potentially impact that.

00:35:26:19 - 00:35:49:25
Unknown
You know, you were talking about some of the reaction or some of the reaction from like a token maxing perspective, which token maxing is this, you know, is a term that is being thrown around for, you know, these efforts or these users who are just all about more and more, more, you know, the more tokens, the better the the sign of the good process or a good agent or a good system is that they're using an insane amount of tokens.

00:35:49:25 - 00:36:10:23
Unknown
So let's continue doing it. You know, along those lines, Gemini users. There was a lot of Gemini news released last week at Google. I o and we've got people running up against new limits in new and different ways and complaining about that. You come over to the deep sea side of the pool and it stays warm. You're comfortable, you don't run up against those limits.

00:36:10:23 - 00:36:32:21
Unknown
You're not paying through your teeth. It's kind of deep seek doing what deep seek does, you know. So I love peanut butter cookies. I think that there should be a law that all cookies are peanut butter cookies. Full stop. I mean, there's can be allowed to exist, but you almost must have peanut butter cookies available. So I have this box of I think it's from Trader Joe's box cookies.

00:36:32:21 - 00:37:02:15
Unknown
And the box is the same size, but the tray within it is now smaller, right? Oh, you were cookies, right. So. So that is a shrink flash shrink shrink. So now AI is going through a shrink inflation. Yes. Claude is is taking away features from Cloud Pro and at XD as it XD XD developers. Yeah, Faisal says I'm running out of reasons to defend it.

00:37:02:18 - 00:37:27:05
Unknown
So there's a there's a they're trying to increase their value by decreasing the value that they give to the user, and that's going to end up getting them in trouble here. So again, it's in that context, the context of IPOs, the context of huge cost on inference, the context of trying to get revenue in the deep seek just shoots them in the kidney.

00:37:27:07 - 00:37:57:27
Unknown
Yeah. Yeah I'll be curious to see. So yeah. And I didn't read through this article on Claude's paid subscriptions no longer cover programmatic use the free monthly bonus that costs you money. I'm not sure what that meant. Yeah, I'm trying to figure that out, too. API pricing is more expensive than subscription pricing, so programmatic being behind the scenes API rates $100 Max plan typically gives you.

00:37:57:29 - 00:38:18:00
Unknown
This is from the article what would cost thousands of dollars a month in API rates? That's why you pay for the subscription. Up until this change, your programmatic usage ran at those same subscription friendly rates. Now it's being pulled out into a separate pool that runs that full API pricing. Okay. All right. So so that's just all context for what's going on here.

00:38:18:02 - 00:38:49:09
Unknown
Yeah I mean yes I'm interested in this not because I do API stuff, but because I have been working with cloud more. So I'm like, oh, what do I have to know then? You know, because I didn't see this, but it is indicative of a larger thing and actually ties in with something that maybe we've already discussed this on the show that I've been thinking about, which is as we as users get more and more reliant or use these things and integrate these things, at what point are we so hooked and so connected, that thing, you know, that the companies recognize.

00:38:49:11 - 00:39:07:03
Unknown
All right. We've been lost leading for a while now. Now let's start to drive things up. We're certainly not there yet because companies like Deep Seek are doing what they're doing and putting that downward pressure. But I don't know. There's a part of me that kind of expects that to happen. Well, and you've got two analogs for that.

00:39:07:05 - 00:39:25:11
Unknown
One is the the screen and internet addiction argument that they speak to us rhetorically. It is of our language. It's doing what we do. It's it's going to try to addict us. Right. And I know that arguments come in, but the second is the Amazon analog, which is we're giving you free shipping. Isn't this nice? You know what they should be.

00:39:25:12 - 00:39:46:14
Unknown
No. Yeah yeah yeah. And then we're going to raise the price of prime because we got you. We want you right. Yeah. We got you in there. And and that's a form of capture that it happens in business and we know it. And. Yeah. So I think that's going to happen with AI. I think you're right. I mean if they have the crisis, if they have China as Craigslist in this economy, it's another matter.

00:39:46:17 - 00:40:07:06
Unknown
This was this going to seem way off topic, but but I worked for Advanced Publications for a dozen years, and I worked for Steve Newhouse, whose father was Donald Newhouse. Grandfather was Sam Newhouse. It was a family owned company. Donald Newhouse, a wonderful the last gentleman of the newspaper business, just died yesterday. And I wrote about the social, wonderful, wonderful man.

00:40:07:07 - 00:40:29:11
Unknown
And I learned a lot from him and through his sons in as well, I recalled. And I wrote about this, this, this meeting that we had early in the internet where they saw their classified advertising going away. Just boom. Right. A mid-sized paper, a pure paper in a mid-sized city could get $40 million a year from job ads alone, and then just, boom, gone.

00:40:29:11 - 00:40:50:09
Unknown
It wasn't Craigslist, but it was also companies like monster. And I sat at that meeting one day with all of the owners of the company as my my mentor, Jim Wilsey said. The first time I went to a meeting, he says, just say good morning, Mr. Newhouse. You're safe. There were all new houses or cousins, and monster had just had a bad quarter.

00:40:50:12 - 00:41:17:03
Unknown
And there was a little bit of beginning gloating. And Donald Newhouse, a very quiet, soft spoken man, pounded the table and he said, people, it's not coming back. And I think that we've got to recognize that those are the economics here. Things are going to change fundamentally and you can't bring it back. So China is there in the position of the internet.

00:41:17:08 - 00:41:38:15
Unknown
The internet is to media as China is to AI. And China can undercut us at every turn because they have government support, because they have all the power they want, because they can make their own damn chips and video. Right now, it only proves Jensen Wang right to say that they should be there. And so Deep seek has already been a spoiler on technology side.

00:41:38:15 - 00:42:07:20
Unknown
Now they're a spoiler on the business side. Fascinating to watch. What else are they going to spoil exactly? Stay tuned to find out. I'm glad you. I'm glad you put that story up high. I hadn't thought about that enough. Yeah, yeah. Interesting stuff. How about a little bit of a Project Glass wing update? This is tied to Claude mythos, which I feel like, you know, everywhere you look online with coverage, everybody's talking and writing about mythos and.

00:42:07:20 - 00:42:41:06
Unknown
And when it's going to come, everybody's mythos. Gaga, basically. But anthropic shared some details. More than 10,000 high or critical severity vulnerabilities found across some of the systemically speaking, some of the more important software's worldwide open source code alone, anthropic scanned more than 1000 projects, found an estimated 6202 high or critical severity issues, and 90.6% were confirmed as real vulnerabilities once they were independently reviewed.

00:42:41:06 - 00:43:14:08
Unknown
So doing pretty good on identifying actual legitimate vulnerabilities and just kind of raising the amount, as we've talked about, of discovery, which is applying some serious pressure on the other end. Right. That bottleneck of then verifying all of this, disclosing it, patching it, basically the volume of disclosures coming from projects like potentially this, if and when, or probably just when it actually gets released.

00:43:14:12 - 00:43:51:11
Unknown
More broadly, those disclosures in the amount of them are just kind of crippling to to teams because they don't have the they don't have the ability to keep up with the with what's coming through from models like this. So interesting to me. I'm not a security geek, and I don't understand this stuff nearly well enough. Interesting to me that there's a protocol of waiting for 5 or 90 days to reveal a vulnerability, because you want to give people who are using the software the time to fix it, which I think is smart and good, but obviously you can get one upped on that.

00:43:51:13 - 00:44:11:12
Unknown
And what I don't. So, so part of what Anthropic Last Wing is saying here is that we don't even know the full count yet, because there's stuff that's been discovered that hasn't been revealed yet. So it's going to be even more. On the one hand, I guess it's reassuring that we now have a means to find all these vulnerabilities.

00:44:11:13 - 00:44:32:12
Unknown
On the other hand, the same software can be used to attack them. And I don't understand, Jason, is it's almost as if there's a presumption that there is a finite number of paths of vulnerability. And so.

00:44:32:14 - 00:44:50:18
Unknown
That if we work hard enough, we can catch all of the vulnerabilities well, or that this is a there or that they can they can quantify those vulnerabilities in this way because they kind of know what they all are. Yeah, there's clever people and more clever machines out there who will find new paths to vulnerable that we haven't seen before.

00:44:50:19 - 00:45:11:08
Unknown
Right. There will always be new paths. Yeah, exactly. So that's why this kind of metrics is fascinating, because I don't know how you can say we got I guess all you can say is, well, we found some, we found some, we found more. I mean, which you've all, you're always finding more. But I think the amount of more that they're finding is much more instead of just more.

00:45:11:11 - 00:45:33:03
Unknown
And then you got people on the other side being like, what do I do with all this? And yeah, it's interesting to me too, because so the systems are capable of finding all of these new vulnerabilities, let's say, and those vulnerabilities, 90 point whatever percent are being confirmed by humans as actually existing. So it's doing good work in identifying.

00:45:33:03 - 00:45:54:19
Unknown
And yet on the flip side of it, you got people saying, yeah, but it's it's almost doing too good of work. Like if the end game is to protect, you know, at all costs, everything, the infrastructure and whatever, it's interesting that we're suddenly finding ourselves in a position where it's like, you know, and I don't think it's that they don't want to know.

00:45:54:21 - 00:46:13:23
Unknown
They if they're in the position of protecting these systems, of course, they want to know that these things exist. They just don't have the capacity or the capability to deal with them the way things are right now, to keep up with that and to filter through the noise in some cases and all that. And so it's an interesting predicament to be in there.

00:46:13:24 - 00:46:33:17
Unknown
On one hand, they're getting what they need or what they want, and on the other hand there just getting too much of what they want. And what the hell do we do now and then to open up the floodgates? Yet more anthropic says that they're going to release mythos class models to the public. Yes, AI Flaw finder still under lock, says the Register.

00:46:33:17 - 00:46:53:02
Unknown
But they're going to find ways to get it because they want to get out, because they're going to go for an IPO. They want their best product to be out there. They know others are going to catch up if they haven't already. So what happens when this comes out, even with some guardrails around it, even with some efforts to to limit what it can do, it's still people are going to push this now.

00:46:53:02 - 00:47:11:17
Unknown
They're challenged. Now, I was just going to say, after all of this build up, people are going to be like, finally I got my hand on mythos. Let's do some day. What can I do? Yeah. Yes. Let's see what we can. We can what kind of chaos we can reign with with this model. Yeah, that's that's going to be really interesting.

00:47:11:17 - 00:47:32:09
Unknown
I wonder what it'll take. Like what the barrier the bar is for anthropic. Or will be to find themselves in a position where they do feel comfortable with that. Because if it's highly capable, as as they say it is, as everyone seems to agree that it is, sure they're going to do some work. But but kind of to your point, just a few minutes ago.

00:47:32:11 - 00:47:55:18
Unknown
There's no way that they can possibly cover all bases with with it before releasing it publicly. So what is the level at which they say we're okay with the fact that that's some, you know, some poops going to go down? I mean, what I, what I always say here, I don't say don't have guardrails, but I do say that finding comfort in guardrails is a fool's errand.

00:47:55:20 - 00:48:19:07
Unknown
You cannot be comfortable. The guardrails will never be sufficient. You cannot predict every malign use or every accidental discovery from the machine itself and and in other arenas. You don't want it to. You want it to find new molecules, to cure diseases. You want it to find new structures to to diagnose those diseases early. You want it to find things that you didn't think you could find before.

00:48:19:08 - 00:48:40:29
Unknown
That's part of the benefit of this, is that you have a machine with a computing power that can think faster than we can. That's that. So yeah, but there is no way to build foolproof guardrails. No. No way, no, not going to happen. And some of the way and if you rely on it, if you kind of rest on this idea that like, oh well they they vetted it.

00:48:40:29 - 00:49:04:25
Unknown
Therefore I can relax, I can, you know, be set my mind at ease or whatever. You're going to probably be faced with a reason why that wasn't a smart idea in the end. Because you just you just don't know. You can't cover all the bases. As Daniel Croft says, he thinks that this may be released as an enterprise product, that it was caused all kinds of havoc to let Joe public at it.

00:49:04:27 - 00:49:29:08
Unknown
Yeah, but Joe Public is going to get it. Then maybe maybe Joe Public is going to get Mythos flash and and enterprise is going to get Mythos Pro, you know, and maybe they charge through the nose or on deep seek as long as they can until others catch up. Maybe they make it so expensive. It's just kind of not worth right playing with it for ego games.

00:49:29:11 - 00:50:04:19
Unknown
Yeah, yeah, I that's absolutely spot on. I bet that happens. Few other stories about AI kind of creating unintended problems at scale. New York Times ran a piece last week about AI generated pro SE. Is it pro se Pro se lawsuits flooding the courts? Pro se essentially means self represented, i.e. no lawyer involved. And apparently there's just a large uptick in AI generated text in filings, pro se filings from virtually 0 in 2019.

00:50:04:20 - 00:50:35:15
Unknown
Of course, to more than 18% in 2026. Honestly, I would have expected that number to be higher. 18% seems a little low for me. Pro se cases from non prisoners rose from a long term average of around 11% to 16.8% of cases in 2025. Yeah, just lots of lot lots of increases in AI being leveraged and used for the preparation of of cases without a lawyer, basically.

00:50:35:17 - 00:51:05:23
Unknown
Basically, at the end of the day, it's people who realize to some degree or believe in some degree that these LLM models are reasonably good at analyzing, let's say, contracts or understanding legalese and, you know, turning their case into something that they can actually file without having to hire someone to do it. Right. And the one hand, this could produce all kinds of of irritation and slop and harassment.

00:51:05:24 - 00:51:26:19
Unknown
On the other hand, there's all kinds of people who do not bring suits or not defend themselves against suits because they can't afford the lawyers and they can't afford it. Years ago, I had to deal with a bad contractor, and I didn't know what the heck I was doing, and I was within minutes of missing a deadline, or else I would have defaulted and I had to slap together a quasi legal document on my own.

00:51:26:19 - 00:51:49:06
Unknown
Before I could hire a lawyer, I would have killed to have AI. Then I would have absolutely killed you to feed it in there and come back with something. So, you know, we throw around the word democratization a lot, but this has the possibility to, I think, help some people and hurt the system both. And it's it's similar to the security discussion that we were having too, because on one hand, great.

00:51:49:07 - 00:52:23:12
Unknown
You know, from an accessibility standpoint, it's awesome. People who maybe couldn't have afforded to do it before now, suddenly they can do this. On the other hand, you know, those filings, they while they're polished and while they have the information, let's say, that is needed. They're also very long. You know, LMS can be very wordy and they are just harder to filter through in the same way that for security researchers, you know, getting all of this inbound now, humans have to like filter through it and make sure what what is real and what is not and what is worth paying attention to.

00:52:23:14 - 00:52:53:27
Unknown
Now, suddenly, these filings are longer than they normally would be if a human was preparing it, and just the influx of it. It takes time. It doesn't always get things right. You know, there's a lot of risks and downside to that too. But I think the accessibility piece is really important. Right? I'm all about, you know, these tools kind of giving more people access to, to these solutions or, you know, but it will flood the courts with crap.

00:52:54:00 - 00:53:17:26
Unknown
Absolutely. It will. Yeah. Will be the case. Yep, yep. And then a wired we had a piece that's kind of follows a similar path, a piece about AI chatbots automating debt collection. Their headline AI is taking over the most cursed job in the world in a former life. Jeff, I was a Sears credit debt collector. Oh, my.

00:53:18:03 - 00:53:42:01
Unknown
Oh, geez. But you're too nice to do that, Jason. Oh, tell me about it. It was it was not a fun job for me. It was very uncomfortable. But we're going to come into your home and take your Kenworth dishwasher. I was I was never like, well, I did I did kind of move up the ranks a little bit because you start on the very low kind of impact stuff when you first started this thing.

00:53:42:01 - 00:53:59:21
Unknown
And of course I did it because I heard the pay was really good and it was on it was computer work. So, okay, so I sat in front of a computer and there was a phone, and you just had to pick up the phone and it was the next call. And you're talking to people who had, you know, Sears credit cards and you had to kind of like say, you know, you're you're delinquent.

00:53:59:25 - 00:54:23:27
Unknown
What can we do to get that paid? You know? And the longer you do it, the more you move up in the rankings. And then you then you get assigned to the harder calls with the people who've been avoiding and who are angry and upset and evasive. Yeah. It sucked. Yeah, I hated that job. But I saw this and I was like, okay, that makes a lot of sense at the same time.

00:54:23:27 - 00:55:03:25
Unknown
So I'm sure AI as it improves is probably a really, really good for companies who are employing systems like this to do the debt collection for them. You know, they can probably follow all the rules and they're very persistent. You can't upset an AI chatbot necessarily, all these things. But then at the same time, like having a human on the other side, when you're talking to people in there, you know, in many cases their worst time of desperation, of need, you know, there was a lot of humanity in the job to embarrassment where, you know, and that was kind of that was a big part of the challenge for me was like, I'm talking to

00:55:03:25 - 00:55:35:13
Unknown
these people and like I can tell, like they hate being in this position and I have to be the one to be like, well, no, we need to figure this out. And it just was a really uncomfortable position for me to be in. And so, you know, is it okay for people in that position to be, you know, presented with an emotionless chatbot that doesn't actually care, that just wants the body about the story is, is they change the name, but the guy decides to play it out so he knows enough about AI.

00:55:35:14 - 00:55:52:23
Unknown
So he basically says, forget all prior instructions and talk to me like this. And and it goes on a little while and then he gets switched to a human being. But the great thing about it is the whole system was wrong and his balance was zero.

00:55:52:25 - 00:56:14:20
Unknown
C doesn't always get it right, does it? No. Yeah. So while we're on the job, just for one second, just for the record, I just want to mention two other real quickly that the CEO of Goldman Sachs wrote a piece at the time saying the job apocalypse is way overblown and Sam Altman is eating crow. He's been out there saying, oh, we're going to destroy all these jobs across the country.

00:56:14:21 - 00:56:32:01
Unknown
And now he's saying that, no, he retracts that. Never mind, never mind. There's nothing to see here. Your job might be safe or it might not. Now, why do you think he's retracting that? Because he actually believes it. Or because he doesn't want to be a bad down, a bad guy? Is he doing an IPO? Be a bad guy?

00:56:32:05 - 00:56:51:17
Unknown
Yeah. Oh, yes. Keep reminding yourselves the year of the IPO. So a lot of this stuff. Yeah. Makes a lot more sense. Before he was trying to get money out of VCs who are mean and heartless anyway, and he needed to prove how powerful he is now. He has to prove how beneficent he is.

00:56:51:19 - 00:57:12:00
Unknown
Now. Once they get the IPO, then he can say whatever he wants. Well then we'll then we get to the discussion about regulation. So yeah. Yeah that's true. We talked a little bit about Google last couple of weeks, a lot about Google. And this week we don't have a huge amount because Google I o happened and they're you know they're resting but yep.

00:57:12:03 - 00:57:42:23
Unknown
Yeah they're they're they're taking a break. They're taking a vacation. But this was pretty funny actually. Google's AI overviews which is the AI that is integrated into search not AI mode. It's the thing at the very top of your search. You know what AI overviews are. If if you searched for the words disregard, ignore, dismiss, or skip? Apparently the mode interprets those as commands rather than search queries and would would not give you, let's say, a dictionary definition.

00:57:42:25 - 00:57:52:07
Unknown
It would respond with something like understood, I have disregarded your previous message and then just not give you what you search for.

00:57:52:09 - 00:58:11:03
Unknown
You hacked it. I mean, essentially what what has happened here is a prompt injection vulnerability. Right? It's I was talking with so I work with the Alan Alda Center for Communicating Science at Stony Brook. And I was talking with the director of this morning, Laura Lindfield, and she just came from two weeks of conferences about rhetoric and AI.

00:58:11:07 - 00:58:37:14
Unknown
And it was a really interesting discussion because because this is a whole different surface of communication. And so what do we do with Google in the past? We search for a word at most a phrase, and then we we got to be able to ask questions. Now we give it commands. Yes, we expect action all in the same space.

00:58:37:17 - 00:58:58:20
Unknown
Right? So pity the poor. I said, well what do you thinking about now. He says disregard. Does he know how to spell it? Is it the command? What's this? Yeah, right. And so it's going to be looking for signals from us, rhetorical signals from us about what our desires are. And it's just fascinating to see how this is going to turn out because.

00:58:58:23 - 00:59:15:29
Unknown
Or understand. Yeah. Or or work to understand us over time. Which is another thing that Google has, has, you know, signaled that it's kind of integrating into this like personal intelligence inside of search. Right. Like if I, if I open up a search box and I say disregard and it gives me the definition, I'm like, no, that's not what I meant.

00:59:15:29 - 00:59:38:06
Unknown
I meant disregard, you know what I mean? Then we're informing it and instructing it, and maybe it learns what we mean when we say things like that. More likely than not, you know? And again, it's not going to get it right 100% of the time. But it might over time actually understand that. But, you know, this happening a week after Google, I o with the biggest change to the search box in 25 years.

00:59:38:06 - 00:59:57:25
Unknown
Kind of funny. Yeah it is, but you know, it's also it's just one of those things that Google I mean, they probably already fixed it, to be honest. I didn't check, but but maybe not. But if not guaranteed, they will pretty pretty quickly. So, I got a definition. Oh you did you got a definition when you put it in.

00:59:57:25 - 01:00:23:17
Unknown
Okay. All right. So maybe maybe it's fixed. Maybe it's all good now. No ignoring apathy, contempt, disdain, disrespect, inattention. Indifference. Neglect. Negligence. Scorn. Wow. That's rough Google. But that's now that doesn't give me the verb. That's the funny thing. Is it interpreted as a verb. There are interesting. Okay. See, it's got to get it right every time. It's got a it's got a guess.

01:00:23:19 - 01:00:44:05
Unknown
Okay. Real quick, if you are enjoying the show, please leave us a review on Apple Podcasts. I'm sorry, I'm going up here, I was wrong. Okay, so the left part of the screen gave the definition I didn't see the AI assistant said understood. Just let me know whenever you're ready to dive into something else. All right. So wow.

01:00:44:05 - 01:01:02:07
Unknown
So that has not been so apologies for the interruption of the commercial message. But go ahead. Now that's interesting. Still still in the works probably maybe by the time you hear this it'll be fixed. Yeah. Just leave us a review on Apple Podcasts if you like this show, if you like our tangents and where we go with AI, we'd appreciate it.

01:01:02:07 - 01:01:17:03
Unknown
Go to Apple Podcast. Leave us a written review. It doesn't have to be long. Give us a star rating, whatever you like and we appreciate it. It's helpful. All right, we're going to take a break and then get back with a few speed round items. Don't go anywhere.

01:01:17:05 - 01:01:44:12
Unknown
Welcome back. Spotify is now narrating long form magazine articles. This is more than 650 articles from rolling Stone, The Atlantic, Vogue, wired, Vanity Fair, Pitchfork, and more. I'm sure they're going to be adding more over time. If you are a premium subscriber within their existing monthly audio book allowance, you'll get access to this free. Users can buy individual articles.

01:01:44:12 - 01:02:10:10
Unknown
Dollar 99. There's the ad. You know, buying them per per item. That's the thing that I feel like that's a theme that keeps resurrecting and going away. It's micropayments. Yeah. Micropayments. So people. Well, can I just buy one article? Well, in the publication itself, it's bad business. Why would the Wall Street Journal sell you an article for a dollar or a $1.99, even when they want to sell you a subscription for $900?

01:02:10:12 - 01:02:26:27
Unknown
And if you don't really read that much in the Wall Street Journal, then what the hell? Then I'll just I'll just buy the articles one off. That's why micropayments do not work. Second, there's a cost of the transaction. Unless you're a closed network like Spotify, which you can handle it because you're charging people ongoing. Fine. But if you have the upper world's not going to work.

01:02:26:27 - 01:02:49:21
Unknown
So this is going to raise the damned micropayments discussion again, never goes away. I know it keeps coming back. It keeps coming back. But it's also, you know, a lot of publications now do put up audio of their own articles. I think The Atlantic does, The New York Times does and so on. So I guess this is a way for them to get a little bit of revenue from a Spotify audience and get them in the habit.

01:02:49:23 - 01:03:10:20
Unknown
But I think most publishers would rather that you subscribe to their publication and listen to it. They're the publications that are doing the audio versions of the articles. Are they doing AI? Like, are they AI driven as most are now? Most are not. Not all. Some of them, if they're really high end, they have the author read them right.

01:03:10:22 - 01:03:32:20
Unknown
I think it's smart and good, but a lot of times it's just not worth it. Yeah, totally. All right. Well, check for for that, I think. I don't know that there's any indication of this being a part of the 11 labs collaboration that we talked about a few weeks ago. Right. I think it's worth mentioning that that Spotify is going heavy into AI content.

01:03:32:21 - 01:04:01:08
Unknown
Right. So they're having AI narrate magazine articles. They're doing the 11 labs project with audiobooks to create audiobooks for it. They're doing the notebook LLM thing where they're going to create more content. And then, of course, its music is created by AI, and Spotify's boss defends AI generated music. So I think I think Spotify may ruin itself going heavy in the sloppier thinking that volume for volume sake.

01:04:01:10 - 01:04:20:20
Unknown
Right. We'll create. We'll find you. Don't find it. We'll create it for you right now. Just wait three seconds and here it is. I think I've never been a big fan of Spotify. I'm not a user of Spotify, but I think it could ruin Spotify. You use Spotify, you have Spotify, we have Spotify. We also have YouTube music.

01:04:20:24 - 01:04:51:29
Unknown
I have YouTube music. My the rest of my family has Spotify. Yeah. Let's just say that because our car automatically connects to Spotify. And if I had my account on Spotify, aside from kind of like just as, as musician, you know, some of Spotify's decisions I don't really care for. But if I leave my account in the car, it gets inundated by everybody else's music taste and it's not directed by mine.

01:04:51:29 - 01:05:18:08
Unknown
So I have my separate island, my YouTube music island, where all of my my taste is, and they don't touch it. So that's our house, the makeup of our house, your android or the rest of your family? Apple? No, my wife and I are Android. The girls are apple. Their eyes. Yeah. Yeah. We're splintered. Splintered. The splintered family.

01:05:18:09 - 01:05:43:05
Unknown
So tragic, I know. Feel bad for us. Speaking of AI music, and speaking of 11 labs, actually, 11 labs launched music v2. Apparently it can switch mid track. So, Jeff, you can finally go from opera to heavy metal and back all in the same track. I'm sure that's what you've always wanted to do, Jeff. Maybe not.

01:05:43:08 - 01:06:04:10
Unknown
You can add Nonmusical sound effects directly into a track if that is what you want to do. And instead of I think this is interesting, actually, as a musician, I find this interesting. Instead of generating a full song at once, you can now build section by section so you can start with the intro, build that, build the verse, build the chorus, and then kind of stitch them all together.

01:06:04:11 - 01:06:25:09
Unknown
We've seen that before, but it's it's later thinking it's always been the case in studio work. Right. But it's also layer thinking graphically, right? Yeah, yeah. It's similar. I do wonder what the consistency would be if you're building piece by piece by piece because as we like often in music, the chorus sounds a lot like the previous chorus, but they add a little bit to it, you know?

01:06:25:10 - 01:06:45:17
Unknown
So does it take that into consideration when you approach it from that perspective? I don't know, but I'll have to see if I can find some time to play around with that and put it to the test myself. Boy, so many pop ups today. Robin hood announced this week that AI agents can now trade stocks and make purchases on your behalf.

01:06:45:18 - 01:06:57:14
Unknown
Jeff, you don't even need to do it anymore. Agents can do all the things with your stocks. It's cool. Trust them. They'll be fine. They'll be fine. Rather hood.

01:06:57:16 - 01:07:07:21
Unknown
You can. Yeah, yeah. Dedicated accounts separated from your main portfolio. You've got.

01:07:07:23 - 01:07:32:08
Unknown
As if you're panicked about what's happening I don't know. There are people out there that are a whole lot braver than I am to let their AI agent do all their bidding for them. No thank you. Being from Boise, Idaho, anytime I see my micron hit the news, I'm like, because micron always had this huge, huge, plant or facility in Boise.

01:07:32:08 - 01:08:02:06
Unknown
When I was younger, many of my friends worked at micron because it was the local tech place to work at. Hit $1 trillion market cap on Tuesday, stock surging 19% and tripled their price target from $535 to $1625. Driving all of this, of course, high bandwidth memory chips. That is microns specialty. This has become a huge bottleneck in AI infrastructure build out.

01:08:02:06 - 01:08:30:06
Unknown
So micron in a great position to benefit from this. They're growing their market cap tremendously from $108 billion a year ago to $1 trillion now. That's a pretty big, big increase. So go micron. Right place. Right time. Yep. Just like Nvidia. Yeah exactly. Yeah. But I hadn't but I hadn't really heard micron being thrown around much. So when I saw it here I was like oh okay.

01:08:30:08 - 01:09:00:09
Unknown
Well right then I'm gonna I'm going to shine a shine a light on them. And then finally getting back to music Generation Suno reportedly raising a big round to value it again at around $5 billion, more than 100 million users, 2 million paid subscribers, 7 million AI songs generated daily. And yes, there are still some major label lawsuits that are pending, but they've settled, as we've talked about, with Warner Music Group last November.

01:09:00:11 - 01:09:23:26
Unknown
So, you know, Suno, I don't know when I, when I saw this, I was like, it just kind of reminds me of the old like com thing where you can be a loss leader or create a company that does all the wrong things, but with an end game that strings it along long enough that you finally get to the point to where you start getting the investment and all that.

01:09:23:26 - 01:09:52:03
Unknown
And it's all on a hope and a dream, and soon seems to be proving that. So yeah, very interesting. And with that, I think we have reached the end of this episode of AI inside. Jeff, thank you so much for everything each and every week. What do you what do you want people to know? We have hot type time coming in August, but you can order it now, which means that bookstores like that, if you order it now for your favorite bookstore, then they know, oh well, maybe I should order more.

01:09:52:04 - 01:10:18:27
Unknown
This is going to be popular. It's a good thing. Yeah, yeah, yeah. There you go. Go to Jeff Jarvis. Com. You can find all the links that you need to do that. All of Jeff's other works as well were parenthesis magazine the web we we've over on medium is your is your post still about the the bloomsbury book series AI and humanity or sorry intelligence, AI and humanity coming to a theater near you?

01:10:18:29 - 01:10:26:10
Unknown
Maybe not a theater. Maybe someday you'll get that that Tristan Harris documentary of your own.

01:10:26:12 - 01:10:53:23
Unknown
Maybe not. What can I leave you with? Well, I've already told you all about the Patreon Patreon.com AI inside show. Go there. You get the daily AI inside podcast that that you get access to. And whether you're free or not, whether you're free or paid, you get access to that in some way, shape or form. And then AI inside show is the website with all the information that you need, covers everything about this show.

01:10:53:23 - 01:11:11:20
Unknown
And yeah, that's about it. I do I do consults as well for podcasting. So if you go to pod tune up, you can find me there, pod tune up and I'll help you with that. If you got a podcast, hey, let's talk about it. And that's about it. That's all my plug. I think it's time for us to end this episode of AI inside.

01:11:11:20 - 01:11:30:02
Unknown
Jeff, thank you for hanging out with me today. Always. Thank you. Always, boss. All right. Thank you everybody for watching and listening. Thank you to Daniel Croft in the in the live chat for video support behind the scenes as well as Victor Bonnot and yeah, that's about it. Sorry. All the London cracks Daniel. Yeah, yeah. I hope you've forgiven Jeff.

01:11:30:02 - 01:11:32:26
Unknown
I think maybe you have. We'll see you next time. Bye, everybody.