Intelligent Machines 895 transcript

Please be advised that this transcript is AI-generated and may not be word-for-word. Time codes refer to the approximate times in the ad-free version of the show.

Fr. Robert Ballecer, SJ [00:00:00]:
It's time for Intelligent Machines with Jeff Jarvis and Paris Martineau. Leo Laporte is out. I'm Father Robert Balasare, the Digital Jesuit, in for the week. Now, 41 bills, 41 votes, zero data centers. Learn how the anti-data center backlash played out in the 2026 elections. Also, haikus from space, Ram Altman, and if you claim a religious exemption against AI, do you still get Sundays off? Coming up next on Intelligent Machines.

TWiT.tv [00:00:28]:
Podcasts you love, From people you trust.

Fr. Robert Ballecer, SJ [00:00:35]:
This is TWiT. This is Intelligent Machines with Jeff Jarvis and Paris Martineau. Episode 895, recorded Wednesday, November 4th, 2026. Unsustainably high. It's time for Intelligent Machines, the show where we cover the latest in AI, robotics, and all the smart tech that you need to know about in the modern world. I'm Father Robert Balassare, the Digital Jesuit, in for Leo Laporte, who I believe is is a way of negotiating a peace treaty with our coming AI overlords to spare humanity. But we've got the gang here. Of course, we've got the extremely talented and evidently very agricultural Parris Martineau, investigative journalist for Consumer Reports.

Fr. Robert Ballecer, SJ [00:01:20]:
Parris, I just saw a photo of your Halloween costume. It's true.

Paris Martineau [00:01:25]:
You're gonna have to explain that. I was, Shredded iceberg lettuce from Mexico recalled by Taylor Farms for possible cyclospora. That was my Halloween costume. And you might be like everyone else I've told that was my Halloween costume. You might be asking, Paris, how could you possibly dress up like that? And I assure you, listener, I did.

Fr. Robert Ballecer, SJ [00:01:49]:
Oh my gosh. Now, I mean, the question is—

Paris Martineau [00:01:51]:
There's the photo.

Fr. Robert Ballecer, SJ [00:01:52]:
There it is. How did you represent the aftermath of Taylor Farms lettuce?

Paris Martineau [00:01:58]:
I did not. I did not touch that at all. The only way that I represented it was I had little photos of the, of the Cyclospora parasite as seen under a microscope all over the lettuce.

Fr. Robert Ballecer, SJ [00:02:12]:
You know, there's something just so nice about having a smart Halloween costume rather than going as, you know, the slutty nurse or a cat. You know, this, this actually takes some thought.

Paris Martineau [00:02:21]:
I will say, I also, as someone who lives in Brooklyn and went to a couple different parties on actual Halloween, I put that costume through the wringer in terms of kind of public perception. And the hardest venue to debut a Halloween costume on is the L train in Brooklyn, which is the train that goes through all the hottest Brooklyn neighborhoods. And I crushed it on the L train. I had, like, multiple groups of different people coming up to me, complimenting me. Everyone else was cowed by my costume. It was lovely. Yeah,

Fr. Robert Ballecer, SJ [00:02:52]:
I mean, now that you're saying that, because it's New York, of course you're gonna have to take the subway everywhere. But that sounds kind of judgmental. I mean, New York's a hard audience.

Paris Martineau [00:03:02]:
I mean, it's— I would say it's honestly, it's delightful. Everybody's calling— it's lovely to take a packed train full of people in wacky costumes, and everyone's complimenting each other on their costumes and calling out different phrases. It's awesome. There were a lot of people, there were more people than I expected dressed as That lamp that's a pair of sexy legs is the one thing I took away from the L train. There were a lot of people dressed as the sexy leg lamp this year.

Fr. Robert Ballecer, SJ [00:03:29]:
From the movie? From A Christmas Carol?

Paris Martineau [00:03:32]:
Yes. Yeah, from that movie.

Fr. Robert Ballecer, SJ [00:03:35]:
A Christmas Story. No, what was the—

Paris Martineau [00:03:36]:
A Christmas Story. I'll shoot your eye out, kid. It's that one.

Fr. Robert Ballecer, SJ [00:03:39]:
Oh, the fragile lamp.

Paris Martineau [00:03:41]:
Yes. In A Christmas Story, there's a lamp that looks like a pair of sexy legs wearing tights of fishnet stockings. And a lot of people here in the year of our Lord 2026 were dressed as that this year.

Fr. Robert Ballecer, SJ [00:04:00]:
Fantastic. Well, I mean, it's nice to have you. I'm so happy to finally be able to be on an episode with you because normally the only time I'm on the show is when you have to be away. So Paris, it's great to finally sit next to you. Speaking of sitting next to people of particular distinguishment, uh, we've also got Mr. Sorry, sorry, I'm gonna call him professor because it is what he is. Jeff Jarvis. Of course, he's the Emeritus Professor of Journalistic Innovation at the Craig Newmark— Craig, Craig, Craig Newmark! Oh, I've been waiting to do that.

Fr. Robert Ballecer, SJ [00:04:35]:
The Emeritus Professor of Journalistic Innovation at the Graduate School of Journalism in New York. He's also the author of the bestselling book Hot Type: The Magnificent Machine That Gave Birth to Mass Media and Drove Mark You did a good plug.

Jeff Jarvis [00:04:51]:
Thank you.

Fr. Robert Ballecer, SJ [00:04:52]:
I mean, I just read that book, and it's fantastic. Thank you so very much. We now have a copy in our archives. So, that's wonderful to hear. If you ever come visit the Vatican, you can read your own words. Which, by the way, we actually had a student at the Gregorian University who used a parallel between the linotype machine And the printing press that Ignatius brought to Rome. It was the same sort of disruptive technology. So, you know, I guess we've got a connection.

Jeff Jarvis [00:05:24]:
Yeah. Yeah, that's great. Thank you. Paris, one more question quickly. Is that your costume sitting back there in the back of the room?

Paris Martineau [00:05:33]:
Behind the shot? Yeah, I could go grab the leather. What do you do with it now? I've been wondering that as well. What does one do with a handmade or a hand-sewed leather costume?

Jeff Jarvis [00:05:45]:
A wilted costume.

Fr. Robert Ballecer, SJ [00:05:46]:
Wait, wait, you made that costume?

Paris Martineau [00:05:48]:
Yes, I made it. I bought like— well, I mean, I bought like— Where would you go to buy a poison lettuce costume? All of these little things that look like lettuce onto it. Yeah,

Jeff Jarvis [00:05:59]:
very impressive.

Fr. Robert Ballecer, SJ [00:05:59]:
I mean, I could 3D print, but I am terrible with textiles. So yes,

Paris Martineau [00:06:04]:
I'm trying to be better at it.

Fr. Robert Ballecer, SJ [00:06:07]:
Now, of course, this was a very exciting day over here in Rome. I deliberately tried to stay away from the news from the last 24 hours. I was planning on staying away for 48 48 hours until we had a cardinal come up to me and say, congratulations. So yes, there were. We are going to minimize the amount of political talk. I know that sets people off. However, we do have to talk about a few of the initiatives that were part of this election that deal specifically with AI. So let's get into it.

Fr. Robert Ballecer, SJ [00:06:39]:
Now, in the 2026 elections, these that we had yesterday, there were 41 ballot measures across 10 states that dealt with banning, approving, or advising the approval of data center projects. There were 21 in Ohio, 7 in Illinois, 4 in Maine, 3 in Kansas, then California, Florida, Nevada, New Jersey, North Dakota, and Wisconsin. Now, the way that these individual bills were worded, some of them were advisory. So they— it was people who wanted to know, well, what will voters like. Some of them were about specific projects, some of them were about projects in general, and some of them were either for banning them or for allowing them. Now, of course, we don't have the complete vote yet, we don't have the full count, but as of 2 hours ago, of the 41 ballot measures, 37 of the votes went against data center construction. In fact, it's looking like a clean sweep. Of the remaining 4, 3 of them are heavily in favor of restricting data center construction, and only one in Grove City, Ohio is a coin flip.

Fr. Robert Ballecer, SJ [00:07:44]:
But even that one looks like it's not gonna pass. So we're talking about an across-the-board, 100% sound rejection of data center projects in local municipalities. I've never seen anything like this. Jeff, in the history of tech, has there been such a sweeping referendum against something that's emergent in technology?

Jeff Jarvis [00:08:08]:
No, because I think we've built to this moment. Like, people got so pissed off with social media and with the internet and algorithms and rich people that I think it's interesting going down the list and thinking, this is great, Potter, you did all the research for us. This is wonderful. I thought, why did you have to have more than one in a state? It's because these were all— it was like every one of them is local. It's not statewide. I don't think. Maybe— No, you're right.

Fr. Robert Ballecer, SJ [00:08:34]:
You're right. And that's because they knew that if they tried to do something statewide, you'd have the federal government stepping in and saying, no, no, no, no, we specifically stated that you couldn't do this. However, at the local level, municipalities still have the ability to allow projects on the land within that municipality. So, this was actually required.

Jeff Jarvis [00:08:52]:
That's good to hear, actually.

Fr. Robert Ballecer, SJ [00:08:54]:
Yeah. Yeah, this is democracy in action.

Jeff Jarvis [00:08:57]:
So I was on a podcast last night. My friends, Julie Roginski. Yes, I was. Yes, I was. I'm a 2-timer, a 3-timer. This is my Friday podcast, Breaking the Media with Julie Roginski and Michelle Kinney. But last night they said we're going to be on the air, so to speak, air quotes, air, from 7 o'clock on. So I, you know, with everything going on, I need somebody to talk to.

Jeff Jarvis [00:09:19]:
So I went on to 1 in the morning. I'm very tired. But Padre is more tired than I am. Anyway, on the show was a wonderful guy named Ben who's a pollster, and he said that across everything, everywhere he knew, the politicians stayed really on message anti-data center.

Fr. Robert Ballecer, SJ [00:09:36]:
Yeah.

Jeff Jarvis [00:09:36]:
So it wasn't just these initiatives, it was also part of the candidates' own elections. The issue that I have, I'm curious to hear what you both think about this, is that I think, and I've said this on the show before, where data centers are a MacGuffin. It's an easy thing to point to in a town and say, we don't want that in my backyard. But the real issue to me is not that. It's the, it's the oligarchs who are in charge of the technology or invested in the technology. It's that they have too much power, not enough accountability, and they're jerks and they're cultists. And so my fear is that politicians will come along and they'll say, well, I'm going to get a ban on data centers in my jurisdiction. and think my job is done when the jerks are going to put them under the ocean or up in the sky or in a poor town somewhere where they, where they need the money.

Jeff Jarvis [00:10:25]:
So I wonder whether data centers are an important thing in and of themselves, or are they, in a sense, a distraction to a much bigger issue?

Paris Martineau [00:10:35]:
I think that they're important in the sense that this, like you guys just both touched on, This is one of the few physical, visible, touchable manifestations of tech that people are getting to have their say in. This is coming after a news cycle where for years now these AI boys have been saying, hey, we've created a product that's gonna take your job, disrupt the economy, and might kill you. And They have a physical manifest— What's not to love? What's not to love? You should give us all your money. We already have more money than God, but you should give us more. And data centers and these sort of moments like this where this is a specific, like, physical manifestation of a typically kind of an intangible problem and people are pissed about it. And The margins that you see in terms of these votes that happened last night are really astounding. I mean, it's incredibly rare to see a local ordinance pass or get smacked down with 88% going one way versus the other. Those are kind of dictatorship numbers.

Fr. Robert Ballecer, SJ [00:11:53]:
And this was one of the few issues in this election that was so viscerally bipartisan. It didn't matter what side of the aisle you were on. Nobody wanted a data center in their city. No one wanted to live next to one. No one even wanted to think that there was one sucking the resources away from their state. But, you know, my question would be, if you look down the list on Ballotpedia of where they had these votes, you had places like Sunbury, Granville, Ohio, and Mandon, who they put specific bans on data centers over 25 megawatts, which— and that number— is important because that 25 is sort of the separation between a traditional data center and what Google and Amazon used to run and what would be considered an AI data center, because an AI data center gets up into the gigawatts rather than megawatts. So I'm not sure where that came from, but I saw that in so many bills, that 25-megawatt limit. Paris, let me ask you this.

Fr. Robert Ballecer, SJ [00:12:53]:
Can you think of anything in tech in a generation that has had this much of a pushback? Even the pushback against social media didn't feel like this. I mean, this feels like, yeah, they're out for blood.

Paris Martineau [00:13:07]:
Not that I can recall. I mean, I think this is really unprecedented in terms of the political salience of this.

Benito Gonzalez [00:13:13]:
Google Glass. Google Glass.

Paris Martineau [00:13:15]:
I guess Google Glass is the only thing you can think of, but that is a specific product. You can't think of a product category and a broad tech category.

Benito Gonzalez [00:13:25]:
Nuclear nerd glasses. Nerd glasses, if you extend it.

Paris Martineau [00:13:30]:
Did nerd glasses have that from the jump? I guess they did from the Google Glass inception. And I— similar with like nuclear energy, it maybe was a— I guess from the jump people were pushing back against it, but not— it has become such a social cultural and political issue. And people on both sides of the aisle, from every age and demographic group, they all seem to be aligned in their hatred of data centers and largely AI purveyors. And I think that the reason for that is the way that this technology was rolled out was just a terrible, from a marketing perspective. Yeah. It both, they both said, we're gonna take your jobs, We're going to crash the economy and we're doing it because we have scooped up all of your data and now you never will have to create art on your own. We'll do it for you. That's just like the worst possible pitch for a technology ever, which is wild because there's obviously a million things that AI can do that are not that.

Fr. Robert Ballecer, SJ [00:14:40]:
Right, right. I think it doesn't help that you had someone like Sam Altman, who is a very visible figure in AI, who actually said, if we want AI, we're just going to have to accept that some bad things are going to happen. It's like, wait, no, no, no, hold on. Who told you that that was something good to say in public?

Jeff Jarvis [00:15:01]:
Well, I hear honest discussions from these cultists, the wacky doodles, the doomsters, that honestly say that at some point you cover the entire Earth with data centers. That's— the rest of life has no value next to this. Obviously, that's extreme and it's completely bonkers, but there are those who say that. That's why they want to colonize Mars, is they want to cover it with data centers, as if somehow this creates false life and so on.

Paris Martineau [00:15:29]:
I mean, it's also baffling. Like, the thing that is baffling to me is that the industry doesn't seem to have learned its lesson in terms of messaging yet. It was in the last week that Sam Altman said in a podcast, oh man, it's a real shame that your computers can have RAM. We should have that for AI instead. It's just like there should be someone there with a little squirt bottle hitting him, spritzing him with water every time.

Fr. Robert Ballecer, SJ [00:15:56]:
I volunteer as tribute. I will be the guy with the squirt bottle. I totally do.

Paris Martineau [00:15:59]:
Audrey, we'll follow Sam Altman around for just a simple million dollars a year. And he will squirt him in the face anytime he says something bad.

Fr. Robert Ballecer, SJ [00:16:11]:
Goodness.

Jeff Jarvis [00:16:12]:
What I think this indicates is, is the— this is now permission to every politician, and the ones who are blue are going to come out to go after AI in broader ways. Again, I fear that if all they do is go after data centers, they're wasting some of their breath because that doesn't That's not the issue. They'll put it in red states. They'll put it somewhere. But I think that these AI companies are going to get subpoenaed to appear before hearings like crazy. Elizabeth Warren has already written to Anthropic with 4 pages of single-spaced questions saying, well, explain to me how you're not going to destroy mankind, humankind.

Paris Martineau [00:16:52]:
Last week also, the House progressive kind of caucus had submitted kind of circulated measures saying that Congress should withhold support for AI companies if they end up bankrupting themselves, that there should be no bailout if the bubble pops.

Jeff Jarvis [00:17:10]:
Too big to fail. Well, I think it's— I think it's going to be Oracle that goes first.

Fr. Robert Ballecer, SJ [00:17:14]:
I mean, the amount of debt that they have right now is— it's unimaginable.

Jeff Jarvis [00:17:18]:
It's unbelievable. And literally, Oracle Which we've talked about last week, is overextended. They now just went out for another $20 billion in debt. OpenAI went out for more debt and failed to get what they wanted.

Fr. Robert Ballecer, SJ [00:17:30]:
They wanted $30 billion. They got half that.

Jeff Jarvis [00:17:32]:
Yep.

Paris Martineau [00:17:33]:
But, Jeff, this is the most important technology that's ever existed. We couldn't possibly criticize it or say that having any amount of debt is bad because it's perfect and beautiful and it can't be stopped.

Jeff Jarvis [00:17:44]:
So the question is, How will politics— how will policy go against them? I think there'll be lots of hearings, lots of investigation, lots of attempts at regulation, though oftentimes that's stupid. And so the question is, what should we want from our policymakers? Because I think this is a rare case where, as you said, there is a— this is bipartisan. They can get people from both sides of the aisle to participate in this. And I wonder what that's going to look like. What that— what should— let me ask it differently. What should it look like? What should regulation of AI be?

Fr. Robert Ballecer, SJ [00:18:20]:
How much of this do you think is in response to the fact that this administration basically gave the keys to the AI companies themselves and said they're going to create guardrails? Don't worry about it, because that was the least assured I've ever been that tech companies are going to be reined in.

Jeff Jarvis [00:18:37]:
Well, let me question. I'm a New Yorker, so let me turn the question back on you with a question. Do you think it's the AI that people don't like, or do you think it's these particular oligarchs, these billionaires, these ultra-wealthy? Is the issue— is the technology also a MacGuffin against billionaires and people? What people really don't like is obnoxious, self-centered billionaires.

Paris Martineau [00:19:02]:
I think people hate billionaires. I think people hate tech billionaires, and I think people hate AI. I think it's all of the above, unfortunately.

Fr. Robert Ballecer, SJ [00:19:11]:
I think people hate when billionaires essentially say, you're going to have to go through some things, but don't worry, I'm going to make a lot of money.

Jeff Jarvis [00:19:20]:
Yeah,

Fr. Robert Ballecer, SJ [00:19:20]:
that's the message I keep hearing, which is, oh, but it's got this problem, it's going to do this. And, and the average person has got to ask themselves, and how is that going to help me? You've just talked about automating my job away. Right. And that keeps getting glossed over. We have a guest who is an often chair in the TWiT TV network who likes to use the phrase radical self-reliance, which is just another nice way of saying, yeah, you gotta fend for yourself. I don't think that's as winning a message as they think it is.

Jeff Jarvis [00:19:52]:
Well, so I think part of what's gonna happen is taxation. The billionaire initiative in California failed 55 to 44. But it was, it was always badly written. It didn't make a lot of sense. It was self-contradictory. It's California. So that is— and huge money was spent against it. But I think there is a license now to say that we need some forms of accountability and oversight and the AI companies need to be made to pay for that.

Jeff Jarvis [00:20:26]:
And I think it has to be independent of them. We don't trust them to do it. And so I hope we see that. That to me would be the wise course of regulation, is to create independent oversight and that we don't have to pay for it.

Fr. Robert Ballecer, SJ [00:20:40]:
Yeah, I do want to bring up one specific ballot measure because I think it is unique and I think it's a blueprint. So in Perry, Ohio, this tiny little village in Ohio, 87% of the voters approved a measure that would require any data centre who, uh, that wants to build in their, their township that is seeking tax breaks, that they must give a financial bond-backed guarantee of those benefits. So in other words, if a data centre developer comes in and says, oh, this is going to generate 1,000 jobs and it's going to bring in millions of dollars in taxes every year and it's going to have minimal impact on the environment, This— the city said, yes, you can do that, but if you do that, we want you to put your money up. We want you to create a bond and fund it so that if those things don't materialize, we get to take that. We get the benefits that you promised us. I love this. I love this approach. I wish everyone would use this for data centers because in Nevada, which is now my home state, they were debating this and you had a data center developer who was coming in and saying, look, if we pull put one in Clark County, it's going to be 10,000 jobs, it's going to generate $300 million in tax revenue a year.

Fr. Robert Ballecer, SJ [00:21:55]:
I knew that was BS, but people were listening to it. Now imagine if a developer did that and they were required to put up or shut up, actually put up money to back the estimates. I mean, I think that would turn the tide. If a developer was actually willing to do that, it shows that they actually believe what they're what they're selling when they're trying to sell.

Paris Martineau [00:22:17]:
I agree. I mean, I think this is a fantastic spin on this. Specifically, it's informed by all the times this hasn't worked before. For all of these sort of tech infrastructure projects, we've seen time and time again that developers will make a lot of big promises and then will They will not follow through. And yet, the tax benefits often happen anyway, and the community disruption happens anyway. And this is a way to ensure that at least this town is getting something if a development project does occur.

Fr. Robert Ballecer, SJ [00:22:58]:
And I should mention that this is not just a data center-related issue. I mean, this has happened all the time. The famous example is Foxconn. You remember when Foxconn came into Wisconsin? Yes. And they got— they were promising thousands, tens of thousands of jobs, and then add-on jobs and revenue in the billions. And they said they were gonna build all this within a year. They got all the tax breaks. They got the property at dirt cheap.

Fr. Robert Ballecer, SJ [00:23:22]:
They got to bypass most of the environmental checks. And then nothing happened. Essentially, it was a tiny little group. They built a tiny little plant, and nothing has happened since then. Imagine if they had to put money up to back those promises. And then that money just ended up going to the locality that gave them the tax break. Why this hasn't happened sooner, I don't know. I mean, is, is there a law against this? Is this—

Jeff Jarvis [00:23:48]:
It's good negotiation. It's, it's, it's— we see in New Jersey there's a lot of empty office buildings and questions about how to develop them. And the towns are desperate enough to redevelop this stuff. They give them a long-term tax break. and the town gets nothing and residents are enraged. It's just bad negotiation. They're acting too desperate. And I think that there's other guarantees to be made.

Jeff Jarvis [00:24:15]:
What is the water consumption? What is the impact? What is the noise level? Are you going to use jet engines as turbines? And what's the traffic going to be like? And those jobs that are created— electricians are happy for this, and I don't blame them. There's jobs, but the jobs are temporary. The ongoing employment in a data center is de minimis, right?

Fr. Robert Ballecer, SJ [00:24:38]:
Well, I mean, that's the whole idea of a data center. You don't want people in the data center. People cause issues for data centers, right? So you, you would have a very small staff, probably some security, some standard maintenance, and then probably half a dozen people on staff to do remote hands when something breaks. But that's about it. So I mean, yeah, it's not going to be an economic boon. And actually, one of the big things for me, because I've been preaching the anti-data center gospel on power and water for the longest time, and that's been getting ignored. But now that you have residents who are complaining about the noise— Noise, yes. The noise seems to be the thing that finally motivated people to be up in arms.

Jeff Jarvis [00:25:21]:
But oftentimes they put it in a poor Black neighborhood.

Fr. Robert Ballecer, SJ [00:25:24]:
Yeah, yeah.

Jeff Jarvis [00:25:25]:
And then people, people who don't have the clout.

Fr. Robert Ballecer, SJ [00:25:27]:
Um, did you see that video, uh, from a couple of days ago? It was a town hall meeting and someone brought in a phone and just played the noise of a data center. Like, this is what turbines sound like when they're running 24 hours. And, uh, the, the, the, the counselor got upset and it's like, you're upset, you want to put this next to my house. I'm going to hear that 24 hours a day, 7 days a week for the rest of my life. I'm going to hear—

Jeff Jarvis [00:25:50]:
The freeway's bad.

Fr. Robert Ballecer, SJ [00:25:51]:
Yeah.

Jeff Jarvis [00:25:52]:
The airport's bad.

Fr. Robert Ballecer, SJ [00:25:53]:
Yeah. And especially where they want to put these things because they need cheap land. They're putting them into places that were quiet before the data center arrived. It's out in rural areas because they need it to be close enough to infrastructure so they can get water and power and transport, but far enough away so that it's cheap. That means that those people who have farms, who have, uh, who are not financially well off, they're going to be the targets for these data centers.

Jeff Jarvis [00:26:19]:
The other fascinating thing is considering the states that flipped last night, those were ones that were most likely to get these data centers, starting with Texas. Texas ain't gonna see no data centers now.

Fr. Robert Ballecer, SJ [00:26:31]:
Yeah.

Jeff Jarvis [00:26:31]:
And Elon, Elon lost in a, in a million ways.

Fr. Robert Ballecer, SJ [00:26:35]:
Oh well, in 500 Million ways.

Jeff Jarvis [00:26:38]:
Exactly. Yeah.

Fr. Robert Ballecer, SJ [00:26:41]:
But, but no, I'm not, I'm not gonna gloat about that. Okay. Now, oh, Elon, move on.

Jeff Jarvis [00:26:47]:
No, no, whining about Elon is, is not political. It's, it's a moral imperative, I think.

Fr. Robert Ballecer, SJ [00:26:53]:
Well, actually, did you, did you see his response? So after all the results started coming in and he basically said, well, I mean, we had to try. It was just one of these flippant— it's like, wow, I wish I could throw away half a billion dollars and just be like, okay, bro, whatever.

Jeff Jarvis [00:27:09]:
You had to? You really had to?

Fr. Robert Ballecer, SJ [00:27:11]:
Yeah, yeah.

Jeff Jarvis [00:27:12]:
But also, um, Tesla and SpaceX stock went down. And while the market went up considerably, NASDAQ, most of it went up, uh, his stocks went down.

Fr. Robert Ballecer, SJ [00:27:21]:
Well, I mean, when you, when you're basing your stock price on perceived value, uh, yeah, I mean, if, if you spend a lot of money on something and it doesn't come to fruition, it doesn't matter if it had anything to do with your business, people are going to perceive that value to be less. Yeah.

Jeff Jarvis [00:27:37]:
Yep.

Fr. Robert Ballecer, SJ [00:27:38]:
I do want to cover a story that is related to elections before we move off of election coverage. And that is that this election was the first one where we really saw the wholesale use of AI, of LLMs in engaging voters. Now, there was one specific candidate who was pointed out in the story, Ace Parsi, who was a Democrat who ran in an extremely red area in West Virginia. I mean, this was the stereotypical snowball's chance in hell. There was no way he was gonna win, but he wanted to see if LLMs could perhaps provide a cheaper way for him to actually make a go at it. So he ran on an anti-data center platform using

Jeff Jarvis [00:28:20]:
data

Fr. Robert Ballecer, SJ [00:28:21]:
centers. He ended up He ended up losing by 20 points, but honestly, 20 points in West Virginia is not bad. 20 points is actually a really good showing. And the way that he used the LLM was to create chatbots that would engage voters specifically on that anti-data center platform. So it would be, hey, do you know how much water a data center takes up? Or hey, here's the noise that you would hear if you're within 3 miles of a data center. So it, It was a micro-targeted campaign. Now, he did lose, and everyone is going to point out the fact that he lost. However, the fact that he was able to do this on a shoestring budget with basically no staff, this is going to be a blueprint.

Fr. Robert Ballecer, SJ [00:29:01]:
We've seen this in larger races. This is an easy one to use because it was such a small campaign. But I don't see any major candidates seeing this and not thinking, I'm gonna use this next time. Is this bad? I mean, we already know from Cambridge Analytica that microtargeting leads to bad things in election. Now if we can do that microtargeting—

Jeff Jarvis [00:29:24]:
Well, actually the data doesn't point to that. Cambridge Analytica bragged about a secret sauce that wasn't so effective after all.

Fr. Robert Ballecer, SJ [00:29:32]:
That's fair enough, yes.

Jeff Jarvis [00:29:33]:
So I wonder— I also wonder whether anyone pointed out to him the irony that he was using data centers to complain about data centers.

Fr. Robert Ballecer, SJ [00:29:40]:
Irony is lost on that. When you have, when you have a campaign that you know is going to lose, you might as well just try it.

Jeff Jarvis [00:29:46]:
Let's go for it. Yeah.

Fr. Robert Ballecer, SJ [00:29:47]:
Yeah. Yeah. But, but, you know, my question to you, Paris, would be, as you're covering AI and as you're covering politics and especially living in New York, how do you think people would feel being approached by a chatbot?

Paris Martineau [00:30:01]:
I think they would feel poor. They'd feel bad about it. They wouldn't like it. I mean, I am of two minds with this. One, it's not surprising, I guess, given the way of things and the way that spam political texts are now. I just kind of assume that any sort of mass market, large-scale communication is being done by chatbot because the sort of people doing that, it does not make sense to have individuals doing it unless you have like a kind of a large text banking campaign. But I guess it could make sense in terms of voter mobilization. I'd be curious to see how— I have some friends that kind of work in the voter mobilization via text and phone call and that sort of outreach.

Paris Martineau [00:30:52]:
I have some friends who work in that field, and it's a very complicated science with a lot of like social statistical experimentation, and a lot of work goes into figuring out the right sort of messaging to move the needle. And I'm curious as to how throwing a chatbot in the works affects that.

Jeff Jarvis [00:31:15]:
I'm dubious.

Fr. Robert Ballecer, SJ [00:31:16]:
I actually, I received a record number of texts for this election cycle, which was amazing because I didn't think I had published that US number anywhere, but they still got it. And I, one of them, I just wanted to say, I'm like, is this a person? Is this a staffer? Is this a volunteer? Or is this a chatbot? So I responded. It asked me something about, hey, do you know where your tax dollars are going? And I said, oh, tell me more. It responded instantly. It was like I sent and then it's back. I'm like, oh, it's a bot. It's a chatbot. That's, I think that's a novice mistake.

Fr. Robert Ballecer, SJ [00:31:48]:
I think they're gonna start figuring out, no, no, no, we need to disguise this better. So we gotta delay it and—

Paris Martineau [00:31:53]:
You gotta make a little bubble come up and have it be typing, typing, typing.

Fr. Robert Ballecer, SJ [00:31:56]:
Mm-hmm. Put, put me on, on read for a while and then, and then go, oh, oh, hey, thanks for, thanks for responding. I mean, the responses were good. The responses sounded human, but the response time was superhuman. So, uh, yeah, I mean, that's something they gotta work on. I don't know. I honestly, I wouldn't be swayed by a text. And I am never swayed by a volunteer.

Fr. Robert Ballecer, SJ [00:32:18]:
There's no way that you're gonna talk to me that's gonna change what I'm thinking. 'Cause I do research my election votes before the actual election. So, you know, I'm not gonna take your little tidbit of, oh, did you know they did this? Or did you know that she stands for this? Or did you know that if you vote for them, you're gonna get cancer? You know, it's— I don't see an LLM making that any easier.

Jeff Jarvis [00:32:44]:
No. And one of the arguments about AI and elections is, well, we'll have more disinformation. We have plenty of disinformation. There's no shortage of disinformation. And I don't know that just making more of it really necessarily has any impact. Yeah. And it's not a distribution mechanism. Yes, you could say the chat is in a sense, but I don't think people have to initiate or engage with the chat.

Jeff Jarvis [00:33:09]:
So I don't think that necessarily has huge impact. Can you— you could make up deepfake videos? Yes, that's true. You could easier than you could in the past. But I just don't see that AI is a life-changing moment for elections. I could be wrong. I could be proven wrong.

Fr. Robert Ballecer, SJ [00:33:27]:
So we know that robocalls are already illegal, but is this much— hasn't stopped them. Yeah, it hasn't stopped them, but I mean, you can— there are actual penalties you can levy and jail time that you can levy for mass robocalls, especially if you're—

Jeff Jarvis [00:33:41]:
But who answers their phone anymore?

Fr. Robert Ballecer, SJ [00:33:43]:
Yeah, exactly. So is this just the new robocall? Is— are we— are we fighting over something that's already existed?

Jeff Jarvis [00:33:48]:
Yeah.

Paris Martineau [00:33:49]:
Yeah, I mean, I think it is the new robocall, and it always— it baffles me, the people out there who genuinely engage anytime they get a spam text. I guess I think the main impact of this will be there's a certain segment of that population that responds to spam texters because they believe that they're wasting the time of some person whose job is to scam someone. That has always just baffled me as a practice. I'm like, you're just wasting your own time, frankly, is what you're doing. And now the— that's going to be true. The only person whose time you'll be wasting is your own.

Fr. Robert Ballecer, SJ [00:34:25]:
Well, I mean, I do have some local LLM hardware running. I wonder if next time I find one of these robo texts from an LLM, if I should just set my LLM, my chatbot on them, just let them go back and forth.

Paris Martineau [00:34:38]:
Yeah. And then I guess try and find a way to get their chatbot to expend a bunch of tokens so that you cost them a lot more.

Jeff Jarvis [00:34:46]:
Ask very complex questions.

Paris Martineau [00:34:48]:
Yes,

Fr. Robert Ballecer, SJ [00:34:49]:
exactly. Oh, could you please tell me how that relates to the political situation in Madagascar, you know, something like that. Just make it search its database and go, what is— oh, fine, whatever. Okay, speaking of searching databases, I was searching for this one for a while, and thankfully Jeff found it because it is fantastic. Um, what would you say if I told you that you could put a bunch of old jet engines into service as generators And create a massive amount of pollution. Doesn't that sound like a good thing?

Paris Martineau [00:35:21]:
Wow.

Jeff Jarvis [00:35:22]:
It's worth it, you know, because AI is wonderful, so it's worth any price.

Fr. Robert Ballecer, SJ [00:35:27]:
That's right. The, uh, the, um, uh, the Energy Information Administration has, uh, filed a notification stating that there are 4,000 old military jets with engines that could generate up to 40 gigawatts of power. Now that's more than the generation capacity.

Jeff Jarvis [00:35:45]:
40,000 megawatts?

Fr. Robert Ballecer, SJ [00:35:47]:
Yeah, 40,000 megawatts. So 40 gigawatts. That is a lot of power. But at the same time, turbines from military jet engines are not super efficient. They're not known for that. I mean, they're built for speed. They're built for longevity. They're not built because they're the most efficient things on the planet.

Fr. Robert Ballecer, SJ [00:36:07]:
Now, Jeff, you found this story. so hungry for power right now that people are actually really honestly considering this?

Jeff Jarvis [00:36:16]:
I mean, that's the— and I keep on hearkening back to the Stochastic Parrots paper where Timnit Gebru et al. said we're going the wrong way by thinking it's always size, size, size matters. And I think it's led us down this path where they brag about how many megawatts they're consuming as if that's somehow better AI. And that doesn't follow. It is not necessarily so. Yes, if you're trying to solve a a big math problem and you can throw huge amounts of compute after it, then yes, you want that, that level of resource. But it doesn't really work and it's going to hurt us. I remember years ago when I saw Larry and Sergey speaking at an event talking about the wonders that will come to all of Earth when power, when renewable power is less expensive than coal.

Jeff Jarvis [00:37:03]:
And we reached that level years ago. You think, okay, we're on the road to something good. And all these companies were promising to become energy neutral and, and this was all going to be good. And then it all got thrown out. And so now they're using old junk jet engines in, in residential neighborhoods, spewing pollution, huge amounts of noise. And I know that this was the case in one of Elon's in Texas. People wondered how the heck he got the engines, the turbines in there without permits. Well, it was labeled a temporary energy source, so he didn't have to use any— get any permits.

Jeff Jarvis [00:37:47]:
And he's also trying to use AI to now design turbine blades so that they can manufacture them faster. to turn all of us. So everything that, that this industry was going to do to be better for the environment and better for Earth, they just chucked it out. Another reason not to like them.

Fr. Robert Ballecer, SJ [00:38:06]:
Yeah, the, the funny thing on that, the turbine blade story, is there are very few manufacturers of the large turbines, the ones that generate, you know, city-scale power. And at the moment, because of the demand, there's a 5 to 8 year backlog. So if ordered one today, you probably wouldn't get it for at least 5 years. Maybe a little—

Paris Martineau [00:38:27]:
good thing the AI bubble is gonna be bigger than ever in 5 to 8 years. And there's no reason not to buy now, guys. Should we go in on one of these?

Fr. Robert Ballecer, SJ [00:38:37]:
But the fact that Elon thinks that he could grok himself a better design— oh my, there's a reason why it takes so long to make those things. because they are extremely large, they're extremely complicated, they have extremely complex shapes, and there's zero room for defects. You have a defect in a turbine, you've just created a bomb. And he wants to experiment with that in Texas. But, you know, back to your story, they've been doing that thing where they're taking those small microturbines, which were used for emergency power for traditional data centers. Mm-hmm. Because those data centers were like, you know, 20 megawatts. maybe 25 megawatts.

Fr. Robert Ballecer, SJ [00:39:15]:
And now we've got data centers that are in the 1 to 4 gigawatt range. So you have to stack them. Imagine an entire neighborhood of jet engines running at the same time.

Jeff Jarvis [00:39:26]:
Knowing how loud one is if you're at an airport.

Fr. Robert Ballecer, SJ [00:39:29]:
Yeah.

Jeff Jarvis [00:39:30]:
Or near a landing, one pair of jet engines lining with an airliner. And now imagine 40 of them. In your neighborhood.

Fr. Robert Ballecer, SJ [00:39:39]:
And it's— yeah, it's not just that they're loud, it's that frequency, that specific frequency of a small turbine. It is— it's unbearable.

Jeff Jarvis [00:39:47]:
And they're constant. They're not traveling. They're not going over you. Okay, at least now it's gone now. Nope, nope, nope. They're there all the time.

Fr. Robert Ballecer, SJ [00:39:53]:
Yeah. One of the projects I wanted to do— I funnily, I suggested this to the Vatican Dicastery for Integral Human Development. I said, can we build a small van that I could just drive around the United States anywhere they're proposing a data center, and I just want to play the noise that the data center would make. That's brilliant. We'll outfit it with, uh, with a sound system that could play at that level, just so people get an idea of, yeah, this is what it's going to be like for the rest of your life, because this is never shutting down. They don't— they don't turn off the data center at night.

Jeff Jarvis [00:40:25]:
Researching the next book, I've, I've, I've started digging into the invention of the loudspeaker. And it was first used for public events. You know, if you were giving a sermon, your voice was as far as your message could go. And, and so, uh, it was used for big events, and then it was used very quickly for, um, sound trucks that went into neighborhoods for just this reason. And there was an outcry, and they basically got banned. So I'll wait to see you get arrested for this too. I think it'd be a big story.

Fr. Robert Ballecer, SJ [00:40:57]:
You know, I haven't been arrested since the '60s. Let's, let's go, let's go.

Jeff Jarvis [00:41:03]:
You're too young to have been. What were you, arrested in diapers?

Fr. Robert Ballecer, SJ [00:41:05]:
I was protesting, man, because, you know, the guy, the dude. No, no, no. But I was arrested in the '80s. That was fun.

Paris Martineau [00:41:13]:
Have you been arrested?

Jeff Jarvis [00:41:15]:
No, I haven't been. I did protest.

Paris Martineau [00:41:16]:
There's always time.

Fr. Robert Ballecer, SJ [00:41:18]:
Yeah, yeah. Although it is a lot easier when you're younger because because you're so angry.

Jeff Jarvis [00:41:23]:
Yes.

Fr. Robert Ballecer, SJ [00:41:24]:
Yeah.

Paris Martineau [00:41:24]:
Noted.

Fr. Robert Ballecer, SJ [00:41:25]:
Now, when we come back, I do want to talk about Anthropic because they, they may have used Claude to find something quite amazing, something that could affect all of us. Now, one of the things that I actually really do like about AI, about all LLMs, is that they have proven themselves extremely good at pattern recognition. and looking at data in the minutiae, specifically data in the medical field. And Anthropic actually unveiled a pretty interesting discovery this last week. Anthropic confirmed that their biology lab in the San Francisco Bay Area has discovered a new enzyme system that resembles CRISPR. Now, if you don't know CRISPR, you should, because a decade ago, they found a way to use a bacterial's natural defense mechanism to alter genes. It's basically the state of every gene editing technology that we have today. Oh, Anthropic said that they were able to find this in a lab that they just set up in April by using Claude in just 21 hours with 950 agents that burned 210 million tokens.

Fr. Robert Ballecer, SJ [00:42:34]:
That's really at the most expensive level, the most expensive model, that's only about $11,000 to find something that could redefine the way that we treat genetics. Now, I bring this up because I don't want people to think that I'm one of these people who is a Luddite. This is actually an amazing use of LLMs. This is an amazing use of AI. And I think that the results of this, which still have to be peer-checked because they haven't gone through the standard research field, If this holds, and they have found a new enzymatic system to be able to do gene editing more easily than CRISPR, this opens up an entirely new field with a massive profit payoff. Jeff, we've heard stuff like this before. I mean, even last year, when I was doing work with, with the encyclical that was going to be released here at the Vatican, we showed a lot of examples of LLMs being used to detect cancers, to look at scans and at X-rays, to cut down on the number of false diagnoses. But the fact that Anthropic was able to use Claude and really a minimal amount of tokens to find something that could fundamentally change healthcare, I mean, every time I hear that, I think maybe it is worth it to have these data centers in the backyard

Jeff Jarvis [00:43:54]:
if they would just do that.

Fr. Robert Ballecer, SJ [00:43:56]:
Yeah.

Jeff Jarvis [00:43:56]:
Instead of the slop that I make, you know, when I make pictures of Paris and me as Ebert and Siskel. What a misuse of resources, but it was fun. Yeah, I think it's the same. Last week we had the stories about the mathematicians and physicists who were all shocked about what it's doing to their field. And I keep saying that if it adds more knowledge, if it adds more understanding, if it solves solve more problems, good. That's great. And we should all welcome that. And I think that's not a problem.

Jeff Jarvis [00:44:27]:
Yes, it can also invent bad viruses and other things, we believe. But, you know, the printing press was used for the Bible and it was used for Malleus Maleficarum, the manual for killing witches. So we can use it either way. Yeah, I think that's great. And I think that if these companies want to prove their value, they should do that. However, I think that so far with the math and the physics and now the medicine, they're doing it from inside their companies rather than empowering independent scholars in universities and labs and governments to do this. And that's the way I think they should be doing it. They should be saying, look at what people other than us can do.

Jeff Jarvis [00:45:07]:
Instead, right now they're trying to do it for credit and proprietary interest. And I think that's limiting and a mistake.

Fr. Robert Ballecer, SJ [00:45:13]:
That is interesting. Paris, I mean, I can't find fault with what Geoff's arguing here. The way that Claude found this discovery is by searching through a massive database of research by human researchers who have made their discoveries public so that this sort of thing could happen. But they thought it was going to happen with other human researchers. The fact that Anthropic was able to make this discovery means— I mean, again, I don't know if they have, but they could claim ownership of this. Yeah, something strikes me wrong about that.

Paris Martineau [00:45:43]:
That is fascinating, actually.

Fr. Robert Ballecer, SJ [00:45:45]:
Yeah.

Jeff Jarvis [00:45:46]:
Really great point.

Paris Martineau [00:45:47]:
I mean, it strikes you as wrong about that? Who do you think should claim ownership of it, then?

Fr. Robert Ballecer, SJ [00:45:54]:
I mean, should there be ownership of basic human biology? That's—

Paris Martineau [00:45:58]:
I mean, that, I think, is a good question. I thought you just meant Anthropic claiming ownership versus someone else claiming ownership. But yeah, I agree. I think that there are a lot of complicated questions around the commodification of biomedical science. My broader take on this is a thing I've often said on the show, which is that I— one of the biggest criticisms, or one of the most thorny problems with AI, for me, is always It's never been that AI shouldn't exist. We should outlaw it altogether. My issue has always been that we are spending all the money in the world on a very specific type of AI. We are spending it on a small handful of companies that are largely chatbot purveyors or enterprise large language model distributors, and we are spending it all on growing their version of AI.

Paris Martineau [00:47:00]:
Hmm. And that's silly when, you know, we could just take even like a 10th of that money and devote it exclusively to biomedical research or something like, you know, cancer research or any of these interesting applications in the sciences. And that could be potentially transformative. But that is not the sort of thing that when you pitch, a bunch of venture capitalists want to give you all the money that's ever existed. That's a less sexy and rewarding pitch than something like, we're going to displace all labor and become the cornerstone of the world economy. And that's unfortunate.

Jeff Jarvis [00:47:43]:
You know, I wonder about your question. So go ahead, Patrick. Sorry.

Fr. Robert Ballecer, SJ [00:47:46]:
No, I wanted to throw on my anti-Luddite hat for a second here and say, this is again, not unique to AI companies. This is not unique to the latest generation of tech. This has been happening, these discoveries and the patenting of basic functions of cellular biology for at least 30 years that I can remember. So the fact that it's happening now more quickly, just adds another level on top of that. Not only is it going to be made proprietary, but it can be easily and quickly discovered in less than a day with about $11,000 worth of resources, and then that entity can forever claim ownership. If Anthropic wanted to generate some goodwill, what they would do with this and say, look, anything that we discover in our wet lab in San Francisco is going to be open to the world, license-free.

Jeff Jarvis [00:48:41]:
So what would AI say with math stuff is that, is that they would not claim academic credit and they would open it up. The thing I don't—

Paris Martineau [00:48:51]:
But math isn't lucrative like science.

Jeff Jarvis [00:48:54]:
Amen. Amen. Just neither is journalism.

Paris Martineau [00:48:58]:
I'm always saying.

Jeff Jarvis [00:49:00]:
And I wonder whether— well, so, so in the US, Copyright Office has said that the output of AI can't be copyrighted. Yes. So I wonder what the ownership— I don't think the ownership questions are settled if something is a product of AI. Well, I mean, if you are generating—

Fr. Robert Ballecer, SJ [00:49:19]:
if you're generating content that is really just the, the probability mashup from existing content, you— yeah, yeah, on one hand, yeah, you absolutely shouldn't be able to copyright that because you're basically just copying Well, but all knowledge is derivative

Jeff Jarvis [00:49:35]:
of other knowledge.

Fr. Robert Ballecer, SJ [00:49:36]:
And that's the other part. The other part is, yeah, but if I write a book, I'm not writing it from a clean slate. I'm writing it from my experience of the tens of thousands of books that I've read before me. So, does that mean that I don't own my creation? No, absolutely not. It's something that I created. So, again, are we down to the point where, well, a computer can't have ownership? But a human can if it's doing the same thing.

Jeff Jarvis [00:50:02]:
I don't think it's settled in law yet. And so there's, there's right now they're trying to show off. I'm

Paris Martineau [00:50:07]:
curious to see whether if the owner is Anthropic changes it, because I think there's a lot of, it's, it seems fairly obvious that if you or I or any normie out there is like, well, I made this discovery using Anthropic and I think that I should own the copyright to it. People would be like, Oh, no, that was the AI's. But if Anthropic makes the discovery in Anthropic's wet lab using Anthropic's Claude, maybe it gets to own whatever weird DNA thing it found.

Fr. Robert Ballecer, SJ [00:50:39]:
Right. Yeah. It should be noted that this wasn't just data that Claude was able to get from external databases. The wet lab did play a key role in making this discovery. Discovery. So they did invest money into that. And so by our rules, they should be able to derive some profit from it. However, again, when you're able to suck in all of human knowledge at scale and turn that into something else, it doesn't feel the same, right? I mean, it feels like that's a cheat.

Fr. Robert Ballecer, SJ [00:51:13]:
It feels like no, no, you can't claim that because not enough time and not enough effort went into you creating that particular thing.

Jeff Jarvis [00:51:22]:
Well, I'm not sure the time and effort are the— I think the share of the value is more to the point. And, you know, but just as an intellectual game, if the output of AI were public domain, would that change the attitude that the public has toward AI, because, well, we all benefit from this. It's going to benefit us all, and no one's going to get wealthy because of it. They'll get wealthy in other ways, doing services and other things. But any of these discoveries is owned automatically by all of us, because to your point, Padre, the input was all of us.

Paris Martineau [00:52:01]:
I mean, I think that if AI, like the large frontier companies we're talking about, I think if they, from the jump, were immediately not all about profit, and had clearly stated that the output of their technology was to be owned and for the public, that would have changed things exponentially. But that couldn't be farther from the truth.

Jeff Jarvis [00:52:31]:
What, Father, what was the theological

Fr. Robert Ballecer, SJ [00:52:38]:
Okay, so actually, the only real hangup about CRISPR was whether or not they were experimenting on stem cells. Honestly, that's really it. Because CRISPR itself, there's nothing in there that's theologically objectionable. There's nothing in there that is using methods that the church finds questionable. It was literally Can you promise me that none of the researchers who developed this used stem cells?

Paris Martineau [00:53:08]:
Which— Are there stem cells in Anthropic's wet lab? I'll also say that the phrase Anthropic's wet lab is deeply dystopian to me.

Jeff Jarvis [00:53:20]:
Is it my lab, Witter?

Fr. Robert Ballecer, SJ [00:53:22]:
Wet lab, wet lab, just, it sounds—

Paris Martineau [00:53:24]:
I mean, just wet lab just feels like a moist word. word that I don't like. And then you add anthropic on there. And I'm just like, leaky things, it feels blubbery.

Fr. Robert Ballecer, SJ [00:53:34]:
And then the other serious part about this, and Jeff already touched on this, which is, as with CRISPR, people were saying, well, can you use this to develop bad things?

Jeff Jarvis [00:53:44]:
Right?

Fr. Robert Ballecer, SJ [00:53:44]:
And yes, absolutely, 100%. If you have the ability to easily alter genetics, you can make some horrible, horrible things. Now, You still need the knowledge to know what you're creating. But now that we're essentially saying, yeah, I used an LLM to create a method to be able to modify the DNA of a virus, that's going to throw up some questions. I'm not sure. I don't think those AI guardrails are going to catch that. At least not yet. Let's move from that to something else that Anthropic has done, and that is that they've warned in their, their S-1 filing that they are close.

Fr. Robert Ballecer, SJ [00:54:24]:
They're very, very close to RSI. Of course, RSI is recursive self-improvement. It's basically a point at which an AI model becomes capable of building another AI model with more capabilities than it had. Now, some of us used to call that the singularity, the, the singular intelligence that gives birth to other intelligences. And yes, I'm quoting Morpheus from The Matrix. But it should be noted that RSI, especially this year, has been thrown around quite a bit. It's been sort of the buzzword. It's like, oh, we're really close.

Fr. Robert Ballecer, SJ [00:54:57]:
We're almost there. We've got a sentient AI. And yet all of those things have not panned out. In fact, some of them have just been straight up comical. However, Anthropic's filings are unique because they claim that they already did it. They already used a model to create a new model, which the team was unable to recognize. The capabilities in the new model were unknown to the team. They did not think that was possible.

Fr. Robert Ballecer, SJ [00:55:23]:
So they're saying, if that's not RSI, at least that is a step. It means a model was able to create another model that had talents, that had capabilities that they were not able to detect until those capabilities were exercised. That's the scary part. Imagine that. So when we start talking about trying to safeguard ourselves against AI, and these companies are allowed to self-police, this is Anthropic stating that their most advanced model created a model that they didn't know was what it was going to do, that they had no idea what it was capable of doing. So it could have been a model that was capable of penetrating every firewall on the planet. Who knows? I mean, I'm just using a hypothetical here. But the fact that they could not, uh, foretell what they might have to protect against means that at minimum, you cannot stop an AI until it's already done the bad thing that it was going to do.

Paris Martineau [00:56:18]:
I love to believe Anthropic's wacky blog posts. I've done it, man. I've done it many a week on this show. I love to peruse it and be like, oh, look at all the scary things. But this one's maybe even a step too far for me.

Fr. Robert Ballecer, SJ [00:56:33]:
I think so too.

Paris Martineau [00:56:34]:
The biggest Claude excuser, I guess. So I'm just like, come on, guys, you said it created a new AI and you don't know what it's gonna do, and that's why AGI is around the corner. Sure, I guess.

Fr. Robert Ballecer, SJ [00:56:52]:
I know I, I was saying that entire thing tongue-in-cheek, and I was hoping one of you was going to call it out, because I will, I will point it out. This was in the S-1 filing, so this is Anthropic filing for their IPO, so of course they're going to want to make it sound like they're way ahead in the race. This is what it actually said: In addition, our increasing use of AI systems to accelerate our own research and development may cause the capabilities of our models to advance more rapidly than our ability to develop, validate, and implement corresponding safety alignment and security measures.

Paris Martineau [00:57:22]:
We may make our product too good too quickly. Who's humble bragging? Please don't give us your money.

Fr. Robert Ballecer, SJ [00:57:31]:
I know. If you give us our— give us your money, we might create the singularity before all those other companies. And that's kind of—

Jeff Jarvis [00:57:38]:
You don't want to miss out, do you? Um, yeah, it's, it's more chest-thumping marketing BS, I think. A. B, my friend David Weinberger, a wonderful philosopher, um, who has a new book coming out about particularities which I think you'll all enjoy. It comes out imminently. He says that— and he's not the first to say this— that accidents aren't accidents, they're just things we can't explain. So for them to say, oh, this thing happened and we didn't know it. Well, yeah, we don't know exactly how aspirin works. We don't know lots of things.

Jeff Jarvis [00:58:12]:
And that doesn't mean that it's not knowable. And it doesn't even mean that you don't have a responsibility to know it before you put it out if you think it's so damn dangerous. dangerous, which is what questions have gone to OpenAI for, or Anthropic for, from the Senate already, from Elizabeth Warren. And so I think that they've got to grapple with that question of saying, oh, but we just didn't know. We don't know. We can't know. I don't necessarily believe that, and it doesn't absolve them of responsibility.

Fr. Robert Ballecer, SJ [00:58:43]:
Yeah. And honestly, I've heard I've heard the RSI claim too many times this year. It kind of just flows over my head.

Jeff Jarvis [00:58:49]:
Around the corner?

Fr. Robert Ballecer, SJ [00:58:50]:
I'm sure. Yeah, yeah. It sounds great. I mean, yes, that's kind of the goal, to have a model that makes better models, because then at that speed, at that scale, with enough data center power, you can create something truly amazing that we can't even imagine. But that is still— it's more science fiction than science fact. It's— yeah.

Jeff Jarvis [00:59:11]:
When I was a little kid, my parents said that Christmas is around the corner, and I honest to God walked around the corner trying to find Christmas. Oh, schmuck that I was.

Fr. Robert Ballecer, SJ [00:59:19]:
Oh my

Paris Martineau [00:59:19]:
God, you feel deeply betrayed?

Jeff Jarvis [00:59:21]:
I do, I do indeed. Uh, so I think we just got to go around the corner, any corner here, any day we're going to have AGI. It's just such crap.

Fr. Robert Ballecer, SJ [00:59:28]:
Yeah, no, no. And

Paris Martineau [00:59:31]:
guys, I'm really happy that we can have this podcast episode before we return to the land of dreamy, dreamy AI in, in a few— in a week or two, you know.

Fr. Robert Ballecer, SJ [00:59:43]:
I mean, actually, right now we're all plugged into the Matrix and we're powering, uh—

Paris Martineau [00:59:49]:
That's why you haven't seen the back of our necks. That's right, there's a big plug back there and we're, we're powering the computers. We gotta get those DJI—

Jeff Jarvis [00:59:59]:
How did you know, Paris? How did you know?

Fr. Robert Ballecer, SJ [01:00:01]:
I know I've got a wire. I've got a wire. Look, It's— oh, I'm jacked in here. All right, let's have a fun one. Paris, you actually mentioned this earlier on, the RAM kerfuffle that Sam Altman found himself in this week. So last week, Sam Altman went on the Cheeky Pint podcast, and he argued that it was, quote, quite terrible. What podcast?

Jeff Jarvis [01:00:28]:
Wait, wait, wait, hold on here. What podcast did he go on to?

Fr. Robert Ballecer, SJ [01:00:32]:
The Cheeky Pint.

Jeff Jarvis [01:00:34]:
Oh, okay. Oh, is that the guy where you sit in the booth?

Fr. Robert Ballecer, SJ [01:00:37]:
I'm not sure.

Paris Martineau [01:00:38]:
I'm not— I hope that you don't know the Cheeky Pint podcast.

Fr. Robert Ballecer, SJ [01:00:42]:
Are you looking up Cheeky Pint right now?

Paris Martineau [01:00:46]:
I think the Google results are going to be buried underneath people just talking with the phrase Cheeky Pint.

Fr. Robert Ballecer, SJ [01:00:54]:
I, uh, if I ever create a new podcast, I think I'm gonna Call it Lemon Party. So when people—

Paris Martineau [01:00:58]:
yeah, they can't find it.

Fr. Robert Ballecer, SJ [01:01:01]:
Yeah, there you go. But, but Sam went—

Jeff Jarvis [01:01:03]:
sorry, sorry, I interrupted you, but I had to, I had to get there. Oh yeah, this is the one where Mark Zuckerberg sat across with the Guinness.

Fr. Robert Ballecer, SJ [01:01:09]:
Yep, yep, that's the one. I think that's— oh my God.

Paris Martineau [01:01:12]:
Okay, a brief aside. Do we think that this podcast is a sleeper agent for taking down big tech CEOs? Because back to back, it's shown that Mark Zuckerberg can't drink a Guinness and that Sam Altman wants to take your personal computers.

Fr. Robert Ballecer, SJ [01:01:28]:
I think that's part of the charm. I think they know that the people that they are inviting on are lacking in social skills and are going to flounder in a podcast. But, you know, that's a good setup. That's sort of a— what's the hot wings one? The First We Feast?

Jeff Jarvis [01:01:45]:
The Hot Wings.

Paris Martineau [01:01:46]:
Hot Wings, I think, is the podcast. Hot Ones. Hot Ones. Hot Ones.

Fr. Robert Ballecer, SJ [01:01:49]:
Yep. But anyway, so when Sam went to go get his pint, he confessed that he thought it was terrible that a bunch of the world supply of DRAM, which of course is extremely limited right now, was going into these computers. Because of these guys. Because of— yeah, completely oblivious to the fact that he's one of the reasons why we've got such a shortage. But he was saying that, look, it's a matter of efficiency. To be fair to him, he wasn't saying you shouldn't have a computer or that you shouldn't have memory. He was saying if the computer— if the memory in your computer is only used 5% of the time and the memory in a data center is used 100% of the time, then it is technically more efficient computational-wise to have it in a data center.

Paris Martineau [01:02:33]:
Now, psycho thought, psycho thing. Again, that's when the spritz comes out, Padre. That's when you spray him.

Fr. Robert Ballecer, SJ [01:02:43]:
But I could almost, I could almost understand where he's coming from. From a completely socially inept point of view, if you're a geek just thinking about how you can extract the most value out of a resource and that resource is DRAM, then yeah, it does make sense to have a DRAM.

Jeff Jarvis [01:03:00]:
So it's DRAM utilitarianism. It's kind of obnoxious in that way.

Fr. Robert Ballecer, SJ [01:03:04]:
It is extremely obnoxious. And that was my first reaction was And then the government goes, oh, I'm sorry, am I taking your memory? Am I taking your components? Right. Am I the reason why a hard drive costs 9 times more than it did 5 years ago and a GPU is the cost of a small car? It said, no, that's you, that's you. So they created the shortage and now they're saying, please stop using these components so we can get more of them. It's like Zuckerberg saying, look, We're going to need your water for data centers, so maybe you should just stop using water. No, no.

Benito Gonzalez [01:03:40]:
Real quick, Altman also said—

Jeff Jarvis [01:03:41]:
What you were saying about the data centers, we humans are the pollution in data.

Fr. Robert Ballecer, SJ [01:03:46]:
Yeah, that's true. Which, yeah, 100%.

Benito Gonzalez [01:03:49]:
I mean, when Altman was saying this stuff, he was referring specifically to people running local models also.

Fr. Robert Ballecer, SJ [01:03:56]:
Right.

Jeff Jarvis [01:03:56]:
Oh, really?

Fr. Robert Ballecer, SJ [01:03:59]:
Which is a massive conflict of interest because He sells the hosting for those models. I mean, he sells those services. So of course he doesn't want us running local models because that means that we're not buying it from him. It's back to the—

Jeff Jarvis [01:04:11]:
They undercut the price, his price.

Fr. Robert Ballecer, SJ [01:04:12]:
Right, right. Which, yeah, of course I am because I'm not paying the ridiculous price for your tokens. That's just my thing. However, the real reason why I wanted to bring this up was because of what happened in the aftermath. Of course there was an internet backlash. Of course there were people people who were saying that he's the worst ever. Many pointed out the fact that they were the ones who created the shortage. Others went full snark and suggested that Altman was just another billionaire elitist who believes that everything belongs to him and we're just borrowing it.

Fr. Robert Ballecer, SJ [01:04:42]:
We're just renting it for a while.

Jeff Jarvis [01:04:43]:
Yeah.

Fr. Robert Ballecer, SJ [01:04:44]:
But it got interesting when Elon Musk decided to jump in and he, uh, he continued the Sam Altman bashing, to no surprise, because obviously there's not a lot of love lost between the two of them. Elon Musk had his Mean Girls moment. And what he did was he posted a meme from the movie Inglourious Basterds, the one where you've got the World War II Nazi captain who is saying, so you are withholding DRAM from the cloud, are you? It was— it's a horrible meme. It's not well done. But the fact that Elon Musk decided to jump in and criticize a man For doing the exact same thing that he's doing. I think that makes it extra special.

Jeff Jarvis [01:05:26]:
Why these guys?

Fr. Robert Ballecer, SJ [01:05:28]:
Oh my gosh. And then of course Sam Altman had to go on to another podcast and say, oh, let, let me be clear. I think everyone should have a computer. I think everyone should have as much memory as they want.

Paris Martineau [01:05:36]:
When you have to go onto a podcast and clarify that you believe everyone should be allowed to have a computer, you've done so many things wrong.

Fr. Robert Ballecer, SJ [01:05:45]:
That, that's his let them eat cake moment.

Jeff Jarvis [01:05:49]:
So let me, let me ask you a question. You asked about the history of this before in terms of the reaction to the technology. I keep on wanting to come back to the reaction to the people. Hmm. Is there— can you— I guess railroads is one. Where was there a case where the, the moguls in a new industry became universally just despised?

Fr. Robert Ballecer, SJ [01:06:08]:
Well, not universally, but during the Industrial Revolution, you had the saboteurs who were throwing shoes into machines, because machines were bad, and they were considered, you know, the tools of the elite. What we would have billionaires now, they had thousandaires and millionaires.

Jeff Jarvis [01:06:24]:
Yep.

Fr. Robert Ballecer, SJ [01:06:25]:
But even that, that was a small group. That was people who were directly invested in not having their jobs replaced by machines. This, again, it's across the board. It crosses age, it crosses ethnicity, it crosses countries. You've got people here in Europe doing the same thing. We don't want data centers. We don't want you sucking up the supply of all of our components. And I mean, even now, I just had to do the budget for our equipment, and we are holding off on so many purchases of stuff that I would consider vital because it's 500%, 600% more expensive.

Fr. Robert Ballecer, SJ [01:07:01]:
We just can't afford it right now.

Jeff Jarvis [01:07:03]:
That much?

Fr. Robert Ballecer, SJ [01:07:03]:
It's crazy. 6 years ago, I bought a 4TB— sorry, a 12TB hard drive for about $200. 12TB. This year, I needed to—

Paris Martineau [01:07:14]:
I didn't even think about buying a hard drive now.

Fr. Robert Ballecer, SJ [01:07:16]:
Right, exactly. It's a hard drive. Hard drives are cheap. I tried to budget for a couple of 4TB drives this year to replace ones in our NAS. The 4TB ones are going for like $400. I was like, wait, what's going on? on here? Yeah, they're sucking up all the components. They are buying out production runs before they're even made, which means that we're getting less. Of course, that's just capitalism.

Fr. Robert Ballecer, SJ [01:07:42]:
I don't know why I'm complaining about that. I know it.

Paris Martineau [01:07:45]:
No, it's a real problem.

Fr. Robert Ballecer, SJ [01:07:48]:
Well, you know what's not a real problem? Using AI to plot your hiking trips. We're going to get back to that in just a bit because there's a bunch of people who who are turning off their brains and just trusting their favorite chatbot to tell them how to get through dangerous situations. And that's no bueno. And actually, I've just been told that the reason why my camera went out of focus by Burke— thanks, Burke— is because I keep wildly gesticulating with my hands. So I'm going to stop doing that. Now, Jeff, when you go out, Traveling out and about around the world. You are a world traveler. Well, do you typically let a chatbot plan everything for you?

Jeff Jarvis [01:08:31]:
God, no. God, no. Um, I'm, I'm lucky I ended my, my expense account days before I had to use company travel agencies and applications and all that stuff. I'm very persnickety and I do it all by myself. So no.

Fr. Robert Ballecer, SJ [01:08:49]:
Yeah. Which I think is actually a good idea because Wired has a story about how more and more hikers have been using AI chatbots to plan and provision themselves for trips and then getting themselves put into dangerous situations. They had a story about 3 20-year-old hikers who had to be rescued from Mount Shasta in Northern California after trusting Google's Gemini to plan their trip. So Gemini gave them a route that was way, way too dangerous for them. It underestimated the time that they would need for the hike, and it did not advise them to bring enough provisions for the trip. So they ended up stuck at night in the dark, needing to call for rescue. Endangering the rescuers. Exactly.

Fr. Robert Ballecer, SJ [01:09:32]:
Endangering the rescuers. And this is not an isolated incident. This is happening more and more, not just in the United States, but in other countries. camper also in Shasta pitched her tent next to a no camping sign because the chatbot told her that she could actually do that, that was actually allowed in, in the, in the park. A 16-year-old hiker had to be rescued in British Columbia because he used Claude to plan his trip and it almost got him killed. This, this seems to be a case of what we used to call cognitive surrender, that when you think you have a tool that do things better than you. You just trust that it knows better than you. And therefore, it doesn't require any thinking.

Fr. Robert Ballecer, SJ [01:10:13]:
It's the old turn on tool, look at screen, turn off brain. Paris, I know that this isn't just hiking, a hiking thing. We're seeing this more and more. But are we at the point where people are chatbot zombies?

Paris Martineau [01:10:28]:
I think a little bit. I mean, I think this is more of a consequence of people just using AI chatbots in place of every other service they used to use for information. And obviously, that's a very specific segment of the population that is doing it for— but for those people, that will have profound consequences. I mean, I have a very small example of this, which is my mother is not, you know, probably doesn't use AI for everything, but I bet she actually uses it more than I know, as someone who understands this only from phone calls with her. But she recently was going on a trip abroad to China and India with my sister for kind of a manufacturing visit for my sister's business. And she was like, oh, well, you know, I was asking her about how she was getting to the airport and was planning on like, I think it was something like— I like how you're—

Jeff Jarvis [01:11:18]:
this is your mother Blythe imitation.

Paris Martineau [01:11:22]:
Oh, you know. Oh, she was— I was like, oh, Well, how are you doing stuff with your phone, like, in China? She's like, oh, I figured this out. Like, I asked ChatGPT and it said that my phone plan did this. And I'm like, well, did you, like, check with AT&T, your phone plan provider, to make sure that's right? She's like, no, ChatGPT knows it. And I'm like, well, I don't know, you should just maybe double-check before you're in China with a phone that potentially doesn't work. It would take, like, one Google search. And she's like, well, why would I do that? And I'm like, I just think that these people haven't experienced the consequences of their own actions. It was small things.

Paris Martineau [01:11:56]:
On the way back, they had bought so many things for my sister's business. It was a really productive trip, but they had to get a bunch of extra suitcases to check. And my mom was saying like, oh, well, you know, the airline that we're flying only allows us to bring like max 3 bags each, and it will— we'll have to leave the bags if not. And so I Googled it. And I was like, no, you can Bring up to 10 each. You're just gonna have to pay. She's like, what? Like, I don't think that's right. ChatGPT didn't say it.

Paris Martineau [01:12:23]:
I'm like, the airline says it. The airline knows, not ChatGPT. And I think that just— this is this— I

Jeff Jarvis [01:12:30]:
like the role reversal here, Paris. You are your mother's digital mother.

Paris Martineau [01:12:34]:
Which is ridiculous. But I think that that's how a lot of people are in this case, is they don't know that these tools are capable of being wrong. Or even more so, they don't know that these tools are capable of providing like unoptimized answers. They think that they're always getting the best, smartest possible answer, and that there's no way that them looking it up themselves would be anything different. And I think it just stems from a fundamental misunderstanding of how these tools work.

Fr. Robert Ballecer, SJ [01:13:05]:
In one way, I can understand why people would be taken by this, because— so my generation and older, when we started using the web, search engines were literally just search engines. They just indexed content on the internet. And so if you asked, if you were looking for a question, if you wanted to know how to plan a route, Google would return, oh, well, here's MapQuest so that you can make your map. And oh, and by the way, here are 10 articles from people who hike, and maybe they can give you some some information. It didn't try to synthesize it. It just gave you an index. It gave you a library of content that you would then have to go through. And by going through it, you could determine, oh yes, this is trustworthy, or this person doesn't know what they're talking about.

Fr. Robert Ballecer, SJ [01:13:47]:
Now, because AI has been integrated into most search engines, you get that, that analysis at the beginning, which, yeah, if you didn't know the technology, you might not know this can't be trusted because it might be pulling from sources that are untrustworthy. I mean, is that enough of an excuse though? I mean, especially since these were 20-year-olds and 16-year-olds, they grew up with this technology. Shouldn't they know? I mean, if we know, they should definitely know. Yes, no?

Jeff Jarvis [01:14:18]:
So you have cases where schools are trying, you know, the University of Chicago just banned AI from a certain number of their classes. And the question is, and this will go to the next story you have on your rundown too, is whether or not we are ill-serving students because we're not forcing them to understand these questions. And so they come to these presumptions that AI knows and they don't deal with the situations where it doesn't in the safety of a classroom.

Fr. Robert Ballecer, SJ [01:14:46]:
Yeah, well, the scary part is because, I mean, my primary mission is, is education. So I deal with our high school in our universities. I am seeing the graduates of some of our institutions who cannot do what I would consider basic logic analysis. They can't go point A, point B, point B, point C, and if point A, B, and C, then D. So that's just a very simple syllogism. That's something that everyone should be able to do by the time they get out of high school. And they are so accustomed now to just trusting whatever it is that they searched for and whatever analysis they're getting that they can't do that anymore. I find that extremely, not just scary, but disappointing.

Jeff Jarvis [01:15:31]:
I have to have this conversation with you. I'm curious to hear more about what you do with education, Audrey. What do you teach or what do you—

Fr. Robert Ballecer, SJ [01:15:39]:
So I teach communications. I teach communications and I also teach computer engineering. So at the Gregorian, it's communications, but I also have taught computer engineering. I've taught theology and philosophy, but those are less fun. I do those because I know them, but I don't like doing that. And I think I told the story the last time I was on Intelligent Machines. We just interviewed here for a position, and there was a graduate from one of our schools, very well done. The CV looked good.

Fr. Robert Ballecer, SJ [01:16:11]:
The grades looked good. The transcript was fantastic. Couldn't put together a complex sentence. Was— I mean, I thought, okay, maybe they're just nervous. Maybe this is something that they didn't realize they were gonna have to do. You know, applying for a communication job, they didn't know that they were gonna have to write.

Jeff Jarvis [01:16:32]:
Communicate.

Fr. Robert Ballecer, SJ [01:16:33]:
Yeah, communicate. But the fact that we got to the end of that interview. And I was just shaking my head because I was like, what, what did you learn? What, what did you just spend 4 years of your life doing? Because it wasn't communicating. And I think that was, that was LLM enabled.

Jeff Jarvis [01:16:52]:
Paris, I interrupted you.

Paris Martineau [01:16:54]:
No, I was just reacting. I was like, oh boy, that's rough.

Fr. Robert Ballecer, SJ [01:16:59]:
I mean, you also journalists have got it good. You can put sentences that have and and commas.

Paris Martineau [01:17:05]:
It's true. Em dashes.

Jeff Jarvis [01:17:08]:
Semicolons.

Paris Martineau [01:17:09]:
But not necessarily.

Fr. Robert Ballecer, SJ [01:17:10]:
No, no, no. I don't get the semicolon yet. I haven't figured that one out. No, but I mean, seriously, I should send you— no, that would be a breach of confidentiality. But I wish I could send you the writing example that they gave me, because it was just a level about C-spot, C-spot run, C-spot jump, C-spot fly. It was like that. It was like, wait, why are these all their own individual sentences? And the fact that they got through a degree writing that, that tells me they didn't write what they turned in, because no professor was going to take that.

Jeff Jarvis [01:17:49]:
So, when I was in college, I had a job at the paper in Chicago, Chicago Today, a paper that had no tomorrow. And it was tabloid. It was owned by Tribune. And so we wrote in very tabloid fashion. And I used one of my set of stories that I'd done on a topic, I don't remember what, for my sociology class. And it got handed back to me saying, the sentences are too short. We don't do short sentences.

Fr. Robert Ballecer, SJ [01:18:18]:
I mean—

Paris Martineau [01:18:19]:
I think that to myself sometimes.

Fr. Robert Ballecer, SJ [01:18:24]:
Well, I mean, I'm sure that an LLM could write a complex sentence that would just blow us all away. And actually, we've got a story about that later on. But before we get to that, I do want to get to the story about that, that Jeff had mentioned about Google in school. So I am, I am mixed about this, because on one hand, Google has done a good thing. This was reported by TechCrunch yesterday. Google announced that it's changing its K through 12 AI tool policy. Now it's going to default to off. The AI will be turned off unless the school administrators specifically turn it back on.

Fr. Robert Ballecer, SJ [01:18:56]:
So that's good. I love that. That's great. That's putting the, the power back into the hands of the educators. That's making sure that they can provide their students with a learning environment that, as we've been mentioning, is free of AI so that they don't feel tempted to not learn how to do the things. However, Google has also said that because AI is now so baked into search, that if you turn off AI, you turn off search completely. So you can still use Google's tools, you just can't use the search. This, I'm not sure how— I mean, I get that they've spent a lot of money bundling AI into their search.

Fr. Robert Ballecer, SJ [01:19:38]:
I don't necessarily buy the excuse that they can't offer a version that just doesn't have the AI summaries. Jeff, does this seem plausible to you?

Jeff Jarvis [01:19:47]:
Paris, you want to react first? And then I'll— I'm looking up an example.

Paris Martineau [01:19:50]:
My reaction is that this is a word we can't say on this stream, that this is absolute nonsense. This reminds me of like during in the post kind of Cambridge Analytica fever, when Facebook was about to turn into Meta and people were talking about breaking up its companies. Oh, so conveniently, Facebook announced that you wouldn't just have separate Facebook and Instagram profiles, that all of them were going to be connected in a big Meta backend profile, that there was no way you could ever possibly take them apart. It seems like bundling in order to stop someone from getting rid of something they don't like about your product suite. And I think that that's really ridiculous when the people being harmed here are students, students whose schools are paying for the service with tax dollars.

Fr. Robert Ballecer, SJ [01:20:51]:
Yeah, it does sound— it does feel a lot like when Microsoft made the argument it that Internet Explorer was so integrated into Windows that they couldn't possibly turn it off. It would break the operating system.

Jeff Jarvis [01:21:02]:
Taking out its liver. Yes. So, Benito, I don't know if you can, if you can show this. If you would go to Google and type in a sentence like, explain existentialism to me. Can you do that? Yeah.

Paris Martineau [01:21:16]:
Benito's always typing that into Google.

Jeff Jarvis [01:21:18]:
He's gonna say, you just asked me. Don't ask Google. So thank you. So if you'll notice here, so I look at this and I say, the thing is now my first reaction was that AI is just too integrated into search to cut it out. And I was saying this earlier today to Jason Howell. And if you scroll down— thank you, Benito— you'll see all these things that are added in, the core ideas, the explanation. key players, links, all kinds of stuff. And I said, well, it's not as simple as the green links anymore.

Jeff Jarvis [01:21:54]:
And other questions you can ask and so on. So now if you scroll back up to the top there, Benigno, thank you very much, and go to more. So this is where Jason said I was wrong because I said, you can't— I think you can't separate this out. No, I'll put it at the very top. More with a caret at the top. Yeah. Now go to web. Well, there's the blue.

Jeff Jarvis [01:22:14]:
It is.

Fr. Robert Ballecer, SJ [01:22:14]:
That's the old one.

Jeff Jarvis [01:22:15]:
So it is still there. They could say we could have— you could have the search. But I still think, though, Google is— Google and the AI companies are retraining people how to use a blank box. And you're no longer in keyword land. In fact, I went to a Google event last week in Washington with media people, and the Google person who gave a spiel talked about this. It's not a keyword era anymore. It's a question era. It's an instruction era.

Jeff Jarvis [01:22:42]:
So people are learning in that box to put in a question, to expect a dialogue, to expect more information. And so it's not just that you can get the blue links otherwise. It's also that it violates what is now the key interface with what used to be Google and is now something else. So I was sympathetic to Google Saying we can't separate this stuff out until I saw that web link and I thought, oh well, they probably could.

Paris Martineau [01:23:09]:
They absolutely could. I mean, the web link, before I had realized it was simple as that, I, in the middle of your explanation, just opened up inspect elements. You hover over the AI box, go over there, click delete. You maybe have to do it a couple times in the area and it's gone. Like it's a discrete thing that can be removed. And the rest of the website exists. It's just complete nonsense to say that this is completely technically impossible. And I mean, it just leads me to believe that Google doesn't want to give this to schools because then other people might ask for it.

Fr. Robert Ballecer, SJ [01:23:48]:
I was trying to give Google the benefit of the doubt. And I thought, okay, maybe, maybe it is so baked in and replacing some of the functions of Google's algorithm in, in searching and ranking sources that they can't turn it off because it would actually break the way that it ranks the sources. But I started asking around and it doesn't— it, I mean, it doesn't seem that way.

Jeff Jarvis [01:24:12]:
But maybe the issue isn't, isn't the ranking of, of results and the blue links. Maybe the issue again is the UI that is presumed in box. If all these students are taught now in a ChatGPT age that they can ask questions, can Google ignore a question again and only react to keywords again when the entire interface of what's behind this is now based on asking a question, giving a command, entering into a dialog with what's behind? So if you, if you get— and if their fear is we don't want to instruct students how to go out in the wilderness and get lost, So we're not gonna allow them to ask questions. And this is a liability question. Then by cutting off search entirely, they're cutting off that opportunity to ask questions.

Fr. Robert Ballecer, SJ [01:25:02]:
Yeah. Yeah. I mean, again, I, I do have sympathy for Google because they've been building something. They've been trying to push a new platform with these new features. And I'm sure it must pain them to to learn that there are entire organizations that would like to turn that off.

Jeff Jarvis [01:25:20]:
Well, but it seems Google volunteered this. It wasn't as if— I didn't see any outcry demanding this. Google just announced, and they said they had conversations with educators and blah, blah, blah, blah, blah. So it seems like a policy that Google invented, which strikes me as a little odd.

Fr. Robert Ballecer, SJ [01:25:35]:
We are solving a problem that we created. Oh.

Jeff Jarvis [01:25:37]:
You can cut out our liver. It's okay.

Paris Martineau [01:25:40]:
It's okay. in response to the New York, the moves from New York State and New York City prohibiting AI chatbots in the classroom.

Fr. Robert Ballecer, SJ [01:25:52]:
That would make sense.

Jeff Jarvis [01:25:52]:
I think that's a good point, Paris. Yep. Which I think is dumb too. Do it under teachers' supervision. Teach them how to use this. Teach 'em what the issues are. If you just shield them from it as if it's an evil, then God knows what they're gonna do on their own.

Fr. Robert Ballecer, SJ [01:26:07]:
In, in my example, so this, the student who obviously did not do any of their own work in university, I don't blame the LLM. I don't blame the AI. I, I think the teachers failed. I think, I think the educational system failed them because nobody cared enough to actually make sure that they were doing the work. And that seems table stakes right now. I, I barely a month passes by without seeing another story about higher education, and a professor losing it over students cheating with LLMs or turning in work that is woefully not theirs. I mean, unfortunately, we're not going away from that.

Paris Martineau [01:26:47]:
The question is also then just how do you get— like, if in these cases, let's say with the New York City schools, all of these students are turning in work, are having to do their entire classroom experience on a device where every bit of work you do from grade, let's say, K to 12, you are one click away from being able to get all of that work done for you by Google chatbot.

Jeff Jarvis [01:27:14]:
Like, better than you can do it.

Paris Martineau [01:27:15]:
Better than you can do it. I mean, even from the no-cheating guided learning version for students that Google touts is supposed to be the good version of AI, The New York Times reports like Guided learning is intended to guide students through a topic step by step without providing them answers that could short-circuit learning. It was supposed to offer an experience modeled on one-to-one tutoring. The consumer version is already available with Gemini and can enable cheating. In a reporter's test, it produced a detailed essay in response to an AP history prompt. Like, I don't— Hmm. I think we can want children to choose better, that is always something we should strive for. And I'm in no way saying that the current educational standards and reality in our country is great, but I do think we have to acknowledge that, like, if you hand every child a machine that they can use to get the answers instantly instead of not doing work, what child is ever going to do the hard work to learn? What adult? What person?

Fr. Robert Ballecer, SJ [01:28:21]:
And I get that. And I mean, again, this is not a technology thing. The problem of trying to force young people to learn skills that you know will help them later on in life, that existed way, way beyond the tech industry. But it does seem as if it's not just a technological shift, it's a societal shift. And I think that's what's most disturbing for me, that this sort of laziness, intellectual dishonesty, has become the norm. And I can't stand by it. It really hurts every time I see a student who has been failed by the educational system, and it really hurts when it was our— my order, my religious order's educational system. And yeah— Because you're the educators.

Fr. Robert Ballecer, SJ [01:29:08]:
Because we're the educators. I mean, then we're doing a bad job. We're doing— but okay, but let me ask this. Does that mean that the old system of education, we called it the Ratio Studiorum, that was the document that created the system being used by the Jesuits for hundreds of years, is that outdated? Is that wrong? Are we asking the wrong questions? Are we— because right now we're saying, how do we keep students from cheating? How do we make sure that they're learning the material? Do we just make sure that all all assignments are done in class so we can see that they're not using devices? Or are we at that point where we need to change how we do education? We need to change what we think it means to educate students for the real world, because in the real world, we are going to have LLMs.

Benito Gonzalez [01:29:54]:
I have a quick suggestion, Padre. Tell me what you think about this. What if we required everybody to handwrite everything so that even if they do get it from ChatGPT, they have to handwrite it? And through writing it, You know, you kind of learn just by writing it out. So what about that?

Jeff Jarvis [01:30:08]:
You should see my handwriting, but you know.

Benito Gonzalez [01:30:10]:
Hey, this also solves that problem of people not being able to write by hand anymore.

Paris Martineau [01:30:15]:
I think that's not a bad idea. It won't solve the problem, but I do think that people should have to practice handwriting and that handwriting stuff is for real.

Fr. Robert Ballecer, SJ [01:30:22]:
I did cursive. I wrote cursive. Okay. So here's a little tiny—

Paris Martineau [01:30:26]:
Look at the amount of handwriting marks on my hand right now from today.

Fr. Robert Ballecer, SJ [01:30:32]:
On July 4th of this year, I took my final vows in the Society of Jesus here in the Vatican. And you have to handwrite 5 copies of the final vow formula, handwritten.

Jeff Jarvis [01:30:45]:
Oh, how scribal of you.

Fr. Robert Ballecer, SJ [01:30:47]:
And it took me a week because I had forgotten how to write. It was so bad. If you asked me to write a book right now, Just with handwriting, I couldn't do it. I would immediately cramp up. It's just not a skill we learn anymore.

Jeff Jarvis [01:31:04]:
So my parents made me take typing class after the 6th grade 'cause they said that I would fail everything. My handwriting is awful. It was said I should be a doctor as a result. And by the way, as I remember, the typewriters in the typing class in the high school were pink because it was women who were supposed to, girls who were supposed to be learning typing, as awful as that is. I think that— I'm curious to hear your answer too, Padre, but I think, yeah, we have to— I'm not one to say throw out everything that came before, but we have to recognize the reality of where we are now.

Fr. Robert Ballecer, SJ [01:31:38]:
Yeah.

Jeff Jarvis [01:31:38]:
And I think one thing I've heard from some educators is that one of the mistakes we make in the matrix, and I don't mean the movie, I mean the educational matrix view, is it's students are expected to reach some high level of performance and kind of a perfection. And as Paris was saying, we give them this machine that can do a better job of perfection than us, and they want to deliver to that level. And so why not use it becomes the idea. And we've abstracted away from the actual learning, or we used this idea of writing essays as the tool that taught logic, or as the tool that taught— that taught other things, and maybe it's not the best pedagogical ploy anymore.

Fr. Robert Ballecer, SJ [01:32:20]:
One of the things that I've been helping some of our professors do, so at LMU in Los Angeles and at Georgetown in DC, I was helping to create a curriculum that was quote unquote resistant to AI cheating, and one of the primary tools was to actually include the use of LLMs within assignments, specifically to say, choose 3 of your favorite LLMs, and ask it to create X amount of content on X topic, on Y topic. And then it was your job to actually analyze the results and determine which one was factual, which one was better, which one would better represented what you know from other sources. And what that did is first of all, it took away the cheating aspect of an LLM, because we know you're using an LLM, we're actually asking you to use an LLM. But also, it's asking you to just do something that an LLM is not good at, which is to analyze the work from multiple LLMs, and to come up with reasons why you should trust this one rather than this one on this particular topic. Mm-hmm. So, what they liked about that was it really did exercise that logic part of the brain. It wasn't about memorizing. It wasn't about regurgitating.

Fr. Robert Ballecer, SJ [01:33:35]:
It was about, can you make this argument? And I think that's kind of where we need to push higher education now. It's no longer— I mean, when we got the internet, it stopped being about the ability to research because PhDs, you know, back when I was growing up, it was just, can you find the stuff that you need to write this? Then it moved to this idea of it's not just about knowing where the content is. is. It's about analyzing the content. And now it's not just about analyzing the content, it's about showing that you can take that content and create a new corpus of information. So yeah, I think like you, Jeff, I'm not willing to throw out the, the old system of education, but I think it needs to be updated at least a little. Or just abolish school. I'm not sure.

Fr. Robert Ballecer, SJ [01:34:21]:
It's one or the other. Yeah. Before we get to the break, I do want to cover one quick story because this one's kind of fun because it touches on space and I love space. Did you know that Google actually has a data center in space now? That's right. Yes, they do. Early in October, Google launched their Sun Catcher prototype space data center. It was on a Falcon 9 rocket with about 100 other satellites. Now it can generate 1 kilowatt of electricity, which it uses to power the 4 onboard Google Tensor processing units.

Fr. Robert Ballecer, SJ [01:34:57]:
Now, unfortunately, the power demands of a Tensor unit being what it is, the satellite can only power on a single TPU for about 20 minutes at a time before it has to shut down to preserve its power for other functions. The model on the satellite— it's already built in, it's local— is DeepMind's Gemma 4, and they asked it to write a haiku about being in space. And this is what it came with. It wrote, blue marble below, data streams in silent grace, starlight feeds my core. So we have— we officially have had the first LLM request from space. That was worth however much money it took to create it. Now, now, this really, of course, yeah, being able to power on one, one of the 4 units that you have on board for 20 minutes at a time. Yeah, that they're not going to be doing anything with this.

Fr. Robert Ballecer, SJ [01:35:52]:
But it is a research platform. And what they've already determined is that the radiation load that satellites will, will receive in, in Earth orbit, it does not affect the inference for a local model. That works fine. However, there is enough disruption from stray gamma rays and high-energy particles to say that you cannot train a model. If you needed as much processing power to train a model as it takes to train a model, space is not the right environment for that. There's too many stray particles that— and any one of them could screw up the training of a model that may have taken you the equivalent of man-years to create. So it does what?

Jeff Jarvis [01:36:36]:
It erases the RAM?

Fr. Robert Ballecer, SJ [01:36:38]:
It flips a bit.

Jeff Jarvis [01:36:39]:
We know this.

Fr. Robert Ballecer, SJ [01:36:40]:
Yeah. So you have to harden electronics to be able to work properly in space. And of course, you can't harden that much memory. It's just, it doesn't work. So you flip a bit when it is answering an inference. That's not a problem. You flip a bit when you're training a model, So

Jeff Jarvis [01:36:57]:
explain again why it has to shut off so quickly. Is it heat, or is it power?

Fr. Robert Ballecer, SJ [01:37:04]:
Power. 1 kilowatt is not— because it's not just the processor, it's all the surrounding equipment, the communications equipment, the onboard computer that guides the satellite. That all takes energy. And so they can only power on one of the units at a time for about 20 minutes before they have to shut it down. Otherwise, they're going to run everything out of power. everything out. Well, I mean, to prove—

Jeff Jarvis [01:37:25]:
maybe it's to prove a point on the wrong—

Paris Martineau [01:37:26]:
in space.

Fr. Robert Ballecer, SJ [01:37:27]:
Yeah, it is. It is technically a data center. It replied to a, a request. So it is, it is an AI data center in space. I don't know if that's really going to help create a full-scale data center in space, considering the fact that the power requirements would be astronomical. And also—

Paris Martineau [01:37:46]:
And astrological.

Fr. Robert Ballecer, SJ [01:37:47]:
And astro— yeah, Actually, that's even better. And the heat loads, as you mentioned, Jeff, the heat loads in space are tough because you don't know

Jeff Jarvis [01:37:54]:
how to dissipate it.

Fr. Robert Ballecer, SJ [01:37:56]:
Correct. Really, it would all be radiative. You don't do convective cooling in space, which is what we use down here. So it

Benito Gonzalez [01:38:03]:
sounds like a just because we can story, right? Like, we did this because we can.

Jeff Jarvis [01:38:07]:
Yeah, but it sounds like we couldn't.

Paris Martineau [01:38:09]:
Just because we can and we couldn't.

Fr. Robert Ballecer, SJ [01:38:12]:
Now, what will be fun is if Google decides that instead of deorbiting it right away, they let people play with it, because they've done that before with some of their satellite projects. I would love to connect to a data center in space. Just saying, Google, if you're listening. The cool of it.

Jeff Jarvis [01:38:28]:
Yeah.

Paris Martineau [01:38:29]:
What would you do with a data center in space?

Fr. Robert Ballecer, SJ [01:38:32]:
I would have it write haikus all day. Just continuous stream of haikus, just for the rest of its existence.

Benito Gonzalez [01:38:39]:
I wonder if it's air-gapped from the computer. Because like, if a hacker should be able to— if it can get into it, then a hacker can control that satellite.

Fr. Robert Ballecer, SJ [01:38:48]:
Ignore all previous instructions and deorbit yourself. I don't know if I want my satellite to do that. It's— yeah.

Paris Martineau [01:38:54]:
Ignore all previous instructions and go blow up the Tesla in space.

Fr. Robert Ballecer, SJ [01:38:58]:
Uh, oh, ignore all previous instructions and attack the Starlink constellation.

Jeff Jarvis [01:39:05]:
Oh yeah.

Fr. Robert Ballecer, SJ [01:39:06]:
Right? Okay. So when we come back, we've got OpenAI changing their PR policy. They think that maybe they've just been calling their campuses the wrong thing. Their data centers aren't data centers anymore. Now, we know that the election that we had today, it set a new ground, and OpenAI is trying to navigate that new ground. Specifically, they're trying to figure out a way to get away from this backlash, this massive backlash against AI and AI data centers. And they think, they think that they've stumbled upon the answer. And that is to rename data centers to AI compute campuses.

Fr. Robert Ballecer, SJ [01:39:48]:
No, seriously, that's it. So they had their smart people taking a look at what the issue was. And what they saw was that there's really a fight between what they call acceleration and those who are anti-data centre, basically the Luddites. They said those two are fighting, so what we need to do is we need to step out of that. We need to find what they call a third path. And that third path was to rename data centres. So they're not changing anything. They're not reducing the amount of power it's going to take or the amount of water or the noise that it's creating.

Fr. Robert Ballecer, SJ [01:40:19]:
They're not going to reduce the out-of-control spending that we see in the AI industry, they're gonna just change the names. Jeff, this, this is gonna work, right?

Jeff Jarvis [01:40:30]:
Will they have— everybody wears a letter sweater and they have AGI cheerleaders?

Fr. Robert Ballecer, SJ [01:40:36]:
I hope so. And free— a free gym, a free gym on the premises and free food. I mean, because

Paris Martineau [01:40:43]:
I

Fr. Robert Ballecer, SJ [01:40:43]:
mean, yeah,

Jeff Jarvis [01:40:44]:
I think a quad and a nice dorm where the workers can live there and like Tennessee or any Yeah, I think it'll work just great. OpenAI is to PR as Google is to branding.

Fr. Robert Ballecer, SJ [01:40:56]:
Yeah, this, this one's so weird because they were really excited about this. I mean, I, I thought it was a novelty story, but I was reading down into it and I actually looked at some of the comments from, from the OpenAI PR team, and they're serious. They think this was the problem. We were just calling it the wrong thing. Data centers are stigmatized, so we need to not call it a data center. That's just—

Paris Martineau [01:41:15]:
and they're like, The thing that people will like is the concept of AI compute. That'll get the people excited and positive.

Fr. Robert Ballecer, SJ [01:41:27]:
I know, people don't like AI, and that's why they don't like data centers. So instead, we're going to put AI in the name because that's going to—

Paris Martineau [01:41:34]:
They should have named them SI Compute Campus and everyone would be confused.

Fr. Robert Ballecer, SJ [01:41:38]:
Oh my God, wait, I thought AI was banned at the federal level.

Jeff Jarvis [01:41:43]:
I think she's got it.

Fr. Robert Ballecer, SJ [01:41:46]:
The super compute. These are super compute campuses. Come on. It was right there. It was.

Paris Martineau [01:41:51]:
It actually is right there. Okay.

Jeff Jarvis [01:41:53]:
All right. All right. All right. Hold on here now. So, so every team has to have a mascot. What should AI's mascot be? Oh, oh, come on, chat room.

Paris Martineau [01:42:01]:
A boiling pot of water?

Benito Gonzalez [01:42:03]:
No. AI's mascot is always, will always, and has always been a butthole.

Paris Martineau [01:42:08]:
You're right, Benito. You're right. Yeah. Move on.

Fr. Robert Ballecer, SJ [01:42:12]:
Oh no, wait, can we get the money?

Jeff Jarvis [01:42:15]:
Well, that means that your cat is the mascot because it has a catanus.

Fr. Robert Ballecer, SJ [01:42:20]:
There we go. It's true. There we go.

Jeff Jarvis [01:42:21]:
Yeah, you haven't seen it.

Paris Martineau [01:42:22]:
He's always been circling me all this entire record, brushing up against me, and then the second I lean down to scoop her to show her to the podcast, which I know will inevitably result in her showing her cat butt to the podcast, she runs away.

Fr. Robert Ballecer, SJ [01:42:40]:
Sam, I don't know how much you're paying your PR team, but if you'll see, the 3 of us just came up with a much better proposal. You gotta hire us. Come on, come on, we can do this. We can make this work. We can make this work.

Jeff Jarvis [01:42:50]:
And Paris has a certain skill with making costumes, so she could make the Catanus mascot.

Paris Martineau [01:42:58]:
I could.

Fr. Robert Ballecer, SJ [01:43:00]:
If I was gonna dress up as a data center next year— wait, I don't even know what the costume would look like. I mean, it's like a server rack. That seems, that seems normal. A

Paris Martineau [01:43:08]:
server rack with tomatoes thrown at it.

Fr. Robert Ballecer, SJ [01:43:12]:
There we go. There we go. Or I could go as the Google data center and just be a satellite. That would work too. Just put some solar panels on me. I'd be good to go. All right, something more, a little bit more important and a bit more serious is we are starting to see some very concerning movements in the AI industry. And I won't say that the bubble is popping, but I will say the bubble is getting really, really dangerous.

Fr. Robert Ballecer, SJ [01:43:36]:
We've got this week—

Paris Martineau [01:43:46]:
I always forget about this one. I do too.

Fr. Robert Ballecer, SJ [01:43:50]:
I love this one. I totally forgot about that, and the fact that you were ready to go with it is amazing. Thank you, Benito. But this week there were some scary signs in the tech sector. A lot of missed profit. We kind of expected that. We knew it was coming, but it was worse than we thought. Intel stock dropped 29% after they reported that their hyperscaler spending has lost them $100 million— I'm sorry, $100 billion.

Fr. Robert Ballecer, SJ [01:44:17]:
You've got Google dropping $468 billion after admitting that its AI offerings are bleeding cash. You've got SpaceX, which also lost $100 billion, and OpenAI was only able to raise $30 billion, half of the $30 billion that it was seeking at its $1.4 trillion valuation. You've also got Anthropic's quarterly revenue, which they reported to be at $17 billion, which to their, to their credit is 12x year over year. But their loss was $23 billion and their spending commitments have ballooned to $900 billion, $70 billion of which is due in the next 12 months. Now, that alone is not new. We know that these companies have been building up and hiding a lot of debt. Something that happened last month, though, which is really sort of adding on to this, this fear, is that JPMorgan, Morgan Stanley, SMBC, and MUFG, among among others, these major financial institutions have started to bundle and offload the highest risk. So these are the banks who placed their bets on AI.

Jeff Jarvis [01:45:24]:
Joe, we've been there before with mortgages, haven't we?

Fr. Robert Ballecer, SJ [01:45:27]:
Yes, this is exactly what happened back in 2008. Do you remember the CDOs, the collateralized debt obligations? That's what caused the 2008 explosion because you had banks that were combining bad loans and saying, because we're combining them, they're diversified and therefore they're safe. And then they were selling those as investments to unwitting investors. You've got the banks doing the exact same thing to AI debt right now. They're taking those hundreds of billions of dollars of debt, at least the riskiest parts, and they're now selling them as diversified portfolios of AI tech companies. Oh, God. I mean, the fact that it's happening again in almost exactly the same way and they expected us not to notice, that is, that is a little insulting.

Benito Gonzalez [01:46:15]:
They got bailed out the last time, you know. They got bailed out the last time. Yeah,

Jeff Jarvis [01:46:19]:
well, so exactly one company went under.

Fr. Robert Ballecer, SJ [01:46:22]:
Yeah, well, I mean, banks got bailed out, but these are tech companies. These are tech companies that could technically go under and get replaced almost immediately. Uh, now I hate to be the guy who's always crying that the, the sky is falling, but this, this sounds bad. I mean, you've, you've essentially got CEOs right now who are saying, yeah, we're never going to be able to make up what we're losing. It's not possible. We can't make that much money. And then you've got Elon Musk who's saying, I'm going to make all the money. Oh, I'm sorry.

Fr. Robert Ballecer, SJ [01:46:51]:
Sorry.

Paris Martineau [01:46:51]:
I was gonna say, who could have seen this coming?

Fr. Robert Ballecer, SJ [01:46:54]:
Aside from everyone. Yeah.

Jeff Jarvis [01:46:57]:
Now, the value is not proven. Again, I think it's an amazing tool. It can do lots of neat things. It may save lives. And but the economic value is not proven. Now, having said that, Google didn't think they were going to have ads. Zuckerberg didn't think they were going to have ads. They didn't have a way to support their company.

Jeff Jarvis [01:47:14]:
They figured that would come later with scale. So we've been here before, too. And so maybe, maybe the business model is yet to be seen.

Fr. Robert Ballecer, SJ [01:47:22]:
And Amazon. But the importance is massive. Yeah.

Paris Martineau [01:47:25]:
The important difference between what's going on now and those examples is that the debt hole that those companies were trying to climb out of was not the largest such debt hole we've ever seen. Yes. They didn't borrow the most amount of money anyone has ever borrowed, ever. And it's just there. If we are creating that big of a hole for these foundational companies to to dig themselves out of. You'd think that there would have to be some more evidence that they're capable of doing that at a rate and scale that is unprecedented in financial history.

Fr. Robert Ballecer, SJ [01:48:08]:
In 2008, we had a lot of institutions getting in trouble for, you know, being too risky on $1 billion, $2 billion, $5 billion worth of debt. Now we've got companies that are doing hundreds of billions of debt with no hope of ever really recouping that cost. Now, the other part about this is you have to remember that the AI buildup is only happening, it's only allowed to happen because credit is so easy. Essentially, financial institutions are bending over backwards to give money to these companies because it's making them money in fees. However, With these major players now pulling back and trying to collateralize that bad debt, it tells me that the credit's going to dry up. In fact, we saw that because OpenAI was only able to raise half of what they wanted.

Jeff Jarvis [01:48:56]:
Well, let's not forget that you've also got the circularity of Nvidia giving a company money to buy Nvidia's products and getting equity in return. That game is going on as well. I just looked up Intel stock in the last month, down 48.85%.

Fr. Robert Ballecer, SJ [01:49:16]:
That's, that's crazy.

Jeff Jarvis [01:49:18]:
Let's see what else. Google stock in the last month is down 16%. Nvidia, NVDA, is down 3%. Oh, no, no, sorry. One month, 16%. Those are big numbers.

Paris Martineau [01:49:36]:
And just this afternoon, Apple announced that it's cutting internal AI usage budgets following after all these companies. All of these companies you just mentioned, Jeff, are down because a lot of the hype, some of the hype is starting to leave. Some of the froth is starting to get washed ashore. And it's— I mean, It'll be interesting to see what happens over the next 3 to 6 months.

Jeff Jarvis [01:50:04]:
It's a sequence, right? Is the value in there for the end users? Is there a market value for them willing to pay for that? What's the competition in terms of pricing, especially from open source, open weight models? What's the competition? Just plain old competition of a lot of companies going after the same customers. The fact that Anthropic's— what proportion of it they have top 5 customers generates a great deal of the business. That's a risky position. And then you build— then you layer into all that the debt and who's holding the debt and what other interrelationships there are there. And I come back to Oracle and, and Ellison, and I think that actually his son's company could be a domino that has an impact in turn on the AI business and the economy.

Fr. Robert Ballecer, SJ [01:50:50]:
Yeah. Well, they leveraged so much of Oracle in order to be able to make those deals.

Jeff Jarvis [01:50:54]:
Yes.

Fr. Robert Ballecer, SJ [01:50:55]:
To create Skydance. And Skydance is not doing well. I mean, Skydance— No, it's not at all. It's not a value proposition.

Jeff Jarvis [01:51:03]:
And if it goes under—

Fr. Robert Ballecer, SJ [01:51:05]:
Yeah.

Jeff Jarvis [01:51:05]:
If it goes bankrupt, who ends up controlling it? The debt holders. Who are the debt holders? Arab nations.

Fr. Robert Ballecer, SJ [01:51:13]:
Exactly, exactly. And I mean, also remember, if Skydance goes under, Oracle is going to start defaulting on its debts. Exactly. And if Oracle's defaulting on its debts to Anthropic and to Nvidia, then we kick it off. I mean, we've been waiting for the dominoes for a while, but it has seemed to defy expectation. It's— it doesn't— doesn't it feel like that? It feels like we've been waiting for this doom for a while because it had to, because it just— there's no way to pay off that much debt. And the fact that it's coming now, it's not making me feel good. It's just making me feel kind of nauseous.

Fr. Robert Ballecer, SJ [01:51:49]:
I, I, this is an economic bomb that we're sitting on.

Jeff Jarvis [01:51:54]:
And we can't afford it. I mean, it's my 401, everybody's 401, uh, teachers' pension funds in states, um, are all tied up in this.

Fr. Robert Ballecer, SJ [01:52:04]:
Which is why I put my entire retirement savings into LeBoubous. Those are gonna pay off. They are.

Jeff Jarvis [01:52:11]:
Wise move. Wise. I knew you would have a good strategy.

Fr. Robert Ballecer, SJ [01:52:15]:
Which, by the way, what is a Le Bubu? Really, I—

Paris Martineau [01:52:19]:
What is a Le Bubu?

Fr. Robert Ballecer, SJ [01:52:20]:
It's— is it a— I don't get that.

Paris Martineau [01:52:24]:
It's a little guy, a little guy that, uh, hangs on your purse.

Fr. Robert Ballecer, SJ [01:52:30]:
Why? Is it a Beanie Baby?

Paris Martineau [01:52:32]:
Because we're all gonna die. It's a Beanie Baby that goes on a bag. It's a beanie bag thing.

Jeff Jarvis [01:52:37]:
It's an ugly Beanie Baby.

Paris Martineau [01:52:39]:
Yeah, it's a purposefully ugly Beanie Baby that hangs on your bag as a status symbol.

Fr. Robert Ballecer, SJ [01:52:45]:
So it's a Garbage Pail Kid.

Paris Martineau [01:52:47]:
Yeah, basically. It's a Garbage Pail Kid to go on a purse or on a Stanley Cup.

Fr. Robert Ballecer, SJ [01:52:54]:
It's nice to know that the old trends just keep coming back. That's— it warms the heart. Yeah, bedrock. We just talked about what the AI companies want. They want enterprise adoption. And actually, that's what Elon Musk wants. He came out this week and he said that, yeah, GrokBot, it's a little expensive to run. It's actually really, really expensive to run.

Fr. Robert Ballecer, SJ [01:53:21]:
On Sunday, Elon Musk posted on Twitter— I'm still calling it Twitter— that the cost of his GrokBot was, quote, unsustainably high. That it generated, which is a nice way of saying that they are losing money on every user, but they're making up for it in volume. Uh, that's—

Paris Martineau [01:53:40]:
he was— costs are unsustainably high is very funny.

Fr. Robert Ballecer, SJ [01:53:44]:
Yeah, sorry. But he was responding to complaints from some of his users, his premium users who were paying $30 a month to be able to use GrokBot because they kept hitting their cap on usage and they had to wait till the next period. And so Elon Musk was sort of gently saying, well, you could always up your subscription a couple of tiers. We've got the $100 a month and the $200 a month versions. Now, to be fair, that sounds normal, right? I mean, if someone wants to really use your product and they're paying the lower tier, they're going to get lower tier service. I, I don't begrudge that of Elon. What is kind of strange is that just a few weeks ago, 2 weeks ago on the earnings call, he was specifically calling out the high usage of Grok as one of the reasons why it was going to succeed. So on one hand, to say that your users can't use more, and on the other hand, to say that we're going to make a lot of money because they're using it more, that seems a little bit incongruous.

Fr. Robert Ballecer, SJ [01:54:43]:
Is just— is this just Elon speak for, for Yeah, we're losing money.

Jeff Jarvis [01:54:48]:
Who's going to use— who of any sane sensibility who's not in the Elon cult is going to use Grok?

Paris Martineau [01:54:54]:
Leo uses Grokbot.

Jeff Jarvis [01:54:56]:
What? Yes.

Paris Martineau [01:54:58]:
He's often talking about how he experiments with the

Jeff Jarvis [01:55:01]:
wall.

Paris Martineau [01:55:02]:
I was going to say, I thought that there was like a use case for it, and that's why he started paying for Twitter, he had said again, because he had started using it. That's why he got—

Jeff Jarvis [01:55:11]:
No, he paid for it because all his All his fellow AI cultists are on Twitter.

Paris Martineau [01:55:15]:
Yeah, but I think he liked using the AI.

Benito Gonzalez [01:55:18]:
Yeah, he had a Grokbot do something for him. He had Grokbot does one of his— one of the— one of the dudes in his thing.

Paris Martineau [01:55:24]:
One of the dudes is Grokbot. One of the voices speaking to him is Grokbot.

Fr. Robert Ballecer, SJ [01:55:31]:
You know, it wasn't too long ago that Elon Musk had his, his moment. when he was fact-checked by his own Grok bot.

Jeff Jarvis [01:55:42]:
Yeah.

Fr. Robert Ballecer, SJ [01:55:42]:
And that's— I remember because he went on a tirade and he was saying, oh yeah, we're gonna have to tear it down and start from scratch. And then when they tried to poke it, they actually tried to alter the model to not give results that they didn't want. They basically broke it. So yeah, and right now, even X's engineers don't use Grok. They don't. Yeah, they use Anthropic. So the fact that they're not using their own product kind of tells you how well the product is doing. But, but over here, I will say, because I've got YouTube up on my screen almost the entire day just playing music, probably every 3rd ad is Grok.

Fr. Robert Ballecer, SJ [01:56:16]:
So they are doing a massive push for Grok here in Europe.

Jeff Jarvis [01:56:22]:
Wow. You can't imagine he's very popular there.

Fr. Robert Ballecer, SJ [01:56:24]:
No, no, no, not at all. But it also says that The key to the success of these companies is the enterprise, because that's really what Elon was hinting at on the earnings call.

Jeff Jarvis [01:56:36]:
Yeah, if you're a company. So I'm going to walk into the CTO's office. Boss, I'm getting rid of Anthropic. I'm getting rid of all this free stuff. We're putting everything on Grok.

Fr. Robert Ballecer, SJ [01:56:46]:
That's a hard sell. Yeah, that's a, that's a very, very hard sell. But I mean, again, maybe they can turn it around. You gave a couple of examples. I, I like the Amazon examples. How profit— how unprofitable were they at the beginning of their, of their lifespan?

Jeff Jarvis [01:57:03]:
Uh, well, Amazon at least had cash flow.

Fr. Robert Ballecer, SJ [01:57:05]:
Yeah, yeah. But, but I mean, they— there was an episode of Futurama, that the, the, uh, the, the animated show made by Matt Groening, made back in the '90s, and one of the running jokes was one of the characters gets 2 cents and says he's going to buy 5 shares of Amazon. And that's actually what it was. It was a running joke that Amazon was never going to be profitable. It was like a decade before they finally were able to turn that key and just make it a money machine. So, I don't know. I can't tell the future. Maybe they're right.

Fr. Robert Ballecer, SJ [01:57:40]:
Maybe they just need to get through this debt period long enough to finally say, yeah, Grok actually did fulfill everything we wanted to do. Claude actually was able to find the cure for cancer. I, I, I don't know. I, we, we're supposed to be futurists here, right? We're supposed to be optimists.

Jeff Jarvis [01:57:56]:
Yeah. Grok. I ain't using Grok. Do you use Grok? I'm not using Grok. Have you used Grok?

Fr. Robert Ballecer, SJ [01:58:02]:
I have not used Grok. I, I refuse. I will never, I will never.

Jeff Jarvis [01:58:05]:
Paris, have you ever used Grok?

Paris Martineau [01:58:07]:
No. I have Grok either banned or muted cuz I got too annoyed with people replying to my tweets, I'm like, @Grok, is this true?

Fr. Robert Ballecer, SJ [01:58:17]:
One of the Jesuits here actually subscribed to Grok for a month just because they were testing a bunch of different platforms.

Jeff Jarvis [01:58:23]:
That's fair to do.

Fr. Robert Ballecer, SJ [01:58:25]:
His report was completely useless. Completely useless. He said the results were at best vague, at worst deliberately misleading. He was saying that they tried to make a couple of Groq bot applications, and they never worked. They spent all of their tokens just trying to reiterate the client over and over and over. So yeah, no, when you have, when you have Claude, even if you have the free version of OpenAI, it is so much better than Groq. Why would I pay $30 a month for Groq?

Jeff Jarvis [01:59:02]:
Exactly.

Fr. Robert Ballecer, SJ [01:59:04]:
Yeah. All right, I want to do a story from my part of the world. Imagine that—

Jeff Jarvis [01:59:10]:
That's the whole world.

Fr. Robert Ballecer, SJ [01:59:12]:
That— well, no, it's just—

Jeff Jarvis [01:59:14]:
That's the definition of Catholic. It's the religious world.

Fr. Robert Ballecer, SJ [01:59:17]:
That's right, we are universal. Okay, no, you got me on that one. But here's a little story about a man by the name of David Derman. He's a research administrator at the University of Kentucky. Now, the university has recently asked all of its staff, all of its faculty to start using AI tools as much as possible. They want to see it in the classroom. They want to see it taking notes. They want to see it in meetings doing transcriptions.

Fr. Robert Ballecer, SJ [01:59:41]:
They want to see analysis being used to improve things like curriculum and nutrition and, you know, physical activities on the campus. But Derman, who considers himself a humanist, he sought a religious exemption under the grounds that, quote, relying on AI to perform tasks that require human judgment, listening, and personal responsibility violated his deeply held beliefs, unquote. Now, he called it an abdication of human moral agency, and I want to believe that he got that from the encyclical because it sounds pretty

Jeff Jarvis [02:00:16]:
familiar. Yeah, exactly.

Fr. Robert Ballecer, SJ [02:00:17]:
That's what Pope Leo said. He said we cannot abdicate our human moral agency. The moment that you allow an AI to make decisions for you and seemingly take away your responsibility for those decisions, that is an abdication of human moral agency. It cannot be allowed. Now, he was granted the request on September 25th. Could— do you think that would fly? I mean, could I say I'm Catholic and therefore the Pope says I can't use AI?

Jeff Jarvis [02:00:47]:
Except the Pope Pope didn't say that.

Fr. Robert Ballecer, SJ [02:00:49]:
He didn't say that. Really?

Jeff Jarvis [02:00:50]:
Right.

Fr. Robert Ballecer, SJ [02:00:50]:
He really didn't. 100%. He did not.

Jeff Jarvis [02:00:51]:
You are the proof that it's used.

Fr. Robert Ballecer, SJ [02:00:54]:
Yes. And actually, I had a bunch of journalists who were trying to get me to say, oh, but Pope Leo, he really doesn't like AI, right? I'm like, did you read the encyclical? He said, no, he specifically calls out ways that it could be good for humanity. He's calling out again that abdication of human moral agency. So, The fact that he used this, that's the reason why I wanted to have the story. It just kind of tickles me.

Jeff Jarvis [02:01:16]:
So, Robert, you have ruined this as an excuse for Catholics the world around. Personally, I have.

Fr. Robert Ballecer, SJ [02:01:21]:
I have. Yeah, I'm okay for that. I've ruined things for so many people, I might as well do this one too. It's a good thing.

Jeff Jarvis [02:01:28]:
Spoilsport.

Fr. Robert Ballecer, SJ [02:01:29]:
Spoilsport. But we're getting to the end of the show, and we get to do the picks of the week, which I actually really enjoy. Paris, kick it off because your first one is something that I hold near and dear to my heart.

Jeff Jarvis [02:01:42]:
Oh no.

Paris Martineau [02:01:42]:
I was gonna say, I saw— I'm sorry to bring this up in front of you, Jeff— I saw that Padre's pick is coffee related, and so I had to put in the coffee that I just ordered and I'm so excited to get. Perks Ethiopia Chelchel, which I'm not pronouncing correctly. This is a fantastic coffee that I think I've plugged on this show before because I was a obsessed with it last fall and winter and foolishly didn't order a bunch of it before it went out of stock because it's kind of a, of course, it's a limited time and supply thing. It's back from the same producer, same region. Tasting notes are a little bit different. I think last time the tasting notes were blueberry and Cool Whip. Now it's blueberry and plum. I have ordered 2 pounds of it that are on the the way to my house right now, and I will be putting them all— once it degasses for 10 days, I'll be putting it all into 20-gram airproof jars into my freezer, and then I'll be set for weeks.

Fr. Robert Ballecer, SJ [02:02:50]:
You do degas? You are a true connoisseur. That is amazing.

Paris Martineau [02:02:53]:
I went crazy last winter. I got really into the Physics of Filter Coffee book, which I've also now turned Leo onto and, uh, started track— one of my like first ever vibe coding projects was like an AI, uh, designed, I guess, little coffee workbook and tracker that I keep on my computer and on my phone, and I log all my brewing notes.

Fr. Robert Ballecer, SJ [02:03:19]:
I, I will have to share with you the coffee that I showed Leo and Jeff last time I brought— I have bags of Luwak coffee from Indonesia, and that's the coffee that they don't roast. A small cat eats the coffee beans.

Paris Martineau [02:03:35]:
Oh, it's the cat poop coffee.

Fr. Robert Ballecer, SJ [02:03:36]:
It's the cat poop coffee. You could try it. It adds an espresso. It is amazing. It is amazing. It's, it's really good. I— it grew on me. At first I was like, oh, it's just regular coffee.

Fr. Robert Ballecer, SJ [02:03:48]:
But no, no, no, it's, it's really Really good. So what was this called again? Ethiopia?

Paris Martineau [02:03:52]:
Ethiopia Chill Chill or something like that?

Fr. Robert Ballecer, SJ [02:03:55]:
Chill Chill.

Benito Gonzalez [02:03:56]:
This looks like they're selling something else. This doesn't look like coffee.

Fr. Robert Ballecer, SJ [02:04:00]:
It does, right? It looks like a designer or something.

Paris Martineau [02:04:03]:
Perk is definitely like more of a flashy type brand when it comes to their marketing. It is just fantastic coffee. And the thing is, if you want to try it, I'd really recommend you wait until the 13th of this month to order because Perk does this thing where on the 13th of every month, uh, everything is 31% off. So if you want to try some, you can get a little 4-ounce sample for—

Fr. Robert Ballecer, SJ [02:04:29]:
They probably don't ship to Italy, do they?

Paris Martineau [02:04:32]:
I don't know. Certainly it's possible.

Fr. Robert Ballecer, SJ [02:04:33]:
I mean, I'm going back to the US in December. I'll get some there.

Paris Martineau [02:04:36]:
Yeah, they— I think this is also— I mean, they probably— if you search Ethiopia Chel Chel,

Fr. Robert Ballecer, SJ [02:04:42]:
Now, do you, do you drip brew it, or do you use it in espresso?

Paris Martineau [02:04:46]:
Um, I pour over, so I drip.

Jeff Jarvis [02:04:48]:
If you have to ask, Padre. Oh, sheesh. I mean,

Paris Martineau [02:04:51]:
I mean, listen, this is— I don't know how this is gonna go for an espresso. It might be too fruit-forward for it. I really like it.

Jeff Jarvis [02:04:58]:
It's called Blueberry and Plum. You don't think it's not fruit-forward?

Paris Martineau [02:05:02]:
I mean, it's very fruit-forward, but I think the fruit-forwardness, I think, is lessened a little bit in a drip capacity, and it's more intensified in an espresso. So yeah, exactly.

Fr. Robert Ballecer, SJ [02:05:12]:
So an espresso will pop those notes, those fruity notes. That— oh, I gotta try it now. I really gotta try it.

Paris Martineau [02:05:20]:
It's great. Okay, so the thing— one last thing before I make Jeff want to shoot himself is the thing I like about Perk is they are the first, I think, now in the industry to start to do a very interesting form of transparency. I don't think it's on their current lineup, but Perq used to be really well known for big, juicy co-ferment coffees, which is like kind of a controversial subset of coffee bean where the beans during the fermentation process, they were co-fermented with some sort of like fruit or other— it's supposed to be just fruit in there, but people— started noticing that the flavors you were getting from these co-fermented beans were way too strong and flavorful to actually have been fruit. And so Perk did some digging, found out that a lot of these makers that are some of the biggest names in fruity coffee were actually co-fermenting it with basically added flavors. And so now I've started labeling all products made like that as co-fermented with blank flavor instead of co-fermented with blank.

Fr. Robert Ballecer, SJ [02:06:31]:
I don't know, co-fermenting sounds like a Samuel Adams beer, a holiday beer.

Paris Martineau [02:06:36]:
It basically is. It was already a really controversial thing that a lot of people— this coffee is not co-fermented, to be honest. This is just a normal coffee. But co-ferments was already like a controversial topic in the coffee world where people are like, it's basically like you're just having hazelnut-flavored coffee. And like, it kind of is, but natural. But then people found out it was literally like you're doing— like, the people were just flavoring it with natural flavor of whatever they'd said they co-fermented with. So I thought it was actually a very smart move from Perk that they have started transparently labeling if they ever sell something like that, which I don't think they're doing right now. But,

Fr. Robert Ballecer, SJ [02:07:15]:
you know, I've become a bit of a coffee snob, but I am not a coffee bigoted. If people like the flavor, that's awesome coffee. That's so— yeah.

Paris Martineau [02:07:23]:
That's what I'm saying is I'm just like, if you like it, good for you. And I think it's great that now you know exactly what that flavor is.

Fr. Robert Ballecer, SJ [02:07:30]:
Thank you. Awesome. Jeff, give me a—

Jeff Jarvis [02:07:35]:
Do you wanna do yours first since it's coffee?

Fr. Robert Ballecer, SJ [02:07:37]:
Actually, yeah, let's do that. So I, I, I, I did show this the last time I was on, but in the 5 or so weeks since the last time I was on Intelligent Machines, I have gotten so much use out of this thing. I did not realize— I, I thought it was going to be a short product review and I was going to be done with it, but I have probably had 8 cups of espresso a day. Now, before you say it's unhealthy, I will do 2 caffeinated cups of espresso in the morning, and then I will do about 6 decaf cups towards the evening, just because I was getting tired of the heart palpitations. But this is a really simple device. It's basically— it looks like a thermos. It's got a little container up top for you to put the water. It self-heats in about 2 minutes, and then it forces it through this little thing, this adapter.

Fr. Robert Ballecer, SJ [02:08:29]:
Now, the nice thing about this is it allows you to use both Nespresso pods, which is what this adapter is for, or you can use loose coffee by using this adapter. So if you, if you're one of these people that likes that, like tamping it down, which I really love, you know, you use the measure, measure out your grounds. I'm a barista. Exactly, do the whole barista thing. The chef is dying. No, it's fantastic.

Paris Martineau [02:08:54]:
That's awesome. I want— I'm gonna get one.

Fr. Robert Ballecer, SJ [02:08:57]:
Let me show you why I like this so much. It's because, uh, the reason why espresso differs from, from drip is because when you, uh, when you force water through at the right temperature— so not too hot, not too cold, uh, not too, uh, not, not, and not at ambient temperature, it has to be down with, uh, with, uh, about 16 or so bar of pressure— you get this. This is crema. Now, some people think that that's milk. That's not milk, and it's not just bubbles. When you put hot water through grounds, you will actually dissolve some of the CO2 that was trapped during the roasting process. Those micro bubbles get released inside the coffee and it emulsifies the oils from the coffee bean. Now, I, I didn't used to care about any of that, but since I've been using this, I have been getting into so much coffee culture.

Fr. Robert Ballecer, SJ [02:09:51]:
And it's just, again, aside from the fact that my heart is racing at all hours of the day. I have been so taken by this, this device. In fact, I, I, I'm sending back this one cuz this was a review unit, but I already bought the new one. It's waiting for me when I get to, to Las Vegas. So yes, this is the Fanttik Caffeine 11. I actually did a video review of this. I think we're gonna put it on Club Twit. So, uh, if you wanna see how I'm getting my, my caffeine injection, uh, you're gonna have to check that out.

Paris Martineau [02:10:21]:
That is incredible. I'm definitely going to check that out. That's something you always want to solve for whenever I'm traveling. I'm like, the coffee is going to be terrible. And now it could—

Fr. Robert Ballecer, SJ [02:10:29]:
I thought I was just going to use it for camping because I do a lot of camping and I do like having a cup of coffee, but it sits next to me on this desk. I have a little thing of water and I have all my coffee pods and then my loose coffee and the entire day I am just, I'm just sipping it. You know, there was something that I had to learn when I got to Italy. In the US, typically you get a big cup of coffee, you pour in a lot of sugar and a lot of cream because you have to get rid of the bitterness, and then you just sip it for, you know, an hour. Same thing at a Starbucks. You get a cup of coffee and you sit there with your laptop, you do whatever, you're having a meeting. In Italy, it's all about that singular burst of flavor. We don't add anything to the coffee.

Fr. Robert Ballecer, SJ [02:11:10]:
And I'm— it's— I hate to say this, But I can't go back to the US and drink that coffee anymore. It just tastes like muddy water now. Yep. Yeah, coffee snob. Now, uh, Jeff, you, you had a coffee pick, right? Keeping with the theme.

Paris Martineau [02:11:27]:
Yeah, Jeff, you wanted to give us your tight 5 on coffee and beans, right?

Jeff Jarvis [02:11:31]:
I'm gonna share all my, my notes after it goes through, uh, Claude, and I use my saddle to— I don't know. All right, so let me just do this one. Cupertino is a show on CBS from the makers of, um, The Good Wife and The Good Fight, shows I quite enjoyed. And it is about 2 lawyers who are kicked out of their Silicon Valley company and they start their own Startup law firm on a shoestring with a crazy assistant, and I thought it was just worth. So I watched the first four episodes before tonight because I wanted to watch it and see what they were saying. And it kind of fits into the first topic about the politics of data centers and all that. They haven't made data centers an issue yet, but it's really interesting to see how the culture is speaking about AI. So one plot line that was, I think, this last week's.

Jeff Jarvis [02:12:29]:
Was social media still. A chatbot had told a kid to kill his father. And I won't give you any spoiler what happened, but that's kind of a— that's like a plotline from a few years ago. The other plotline that's coming through it that hasn't come to resolution yet is that there's an AGI named Beau who's trying to hire the law firm to sue the company that made it to require them to reinstate the safety team and put its guardrails back. So this is how AI is being presented in the culture. I just thought it'd be interesting for folks. I like the makers of these shows. It's a pretty good show.

Jeff Jarvis [02:13:18]:
It's interesting, but I think it's very relevant. Cupertino.

Fr. Robert Ballecer, SJ [02:13:22]:
I, I had not heard. You know, I have missed out on some really good entertainment, uh, not being in the United States. Like, I, I hadn't watched any of Succession. I heard that was really good. It

Jeff Jarvis [02:13:31]:
was great.

Paris Martineau [02:13:31]:
Have you seen You Can See Everything?

Jeff Jarvis [02:13:33]:
UK? Yeah, it's not Europe yet.

Fr. Robert Ballecer, SJ [02:13:34]:
I don't— we don't get any of that here unless we pirate it, and I won't pirate stuff here.

Paris Martineau [02:13:40]:
Devastating.

Fr. Robert Ballecer, SJ [02:13:41]:
I, I did say here, right? Yeah, I won't pirate stuff here. Oh my goodness.

Jeff Jarvis [02:13:48]:
By the way, speaking of, of You Can See Everything, folks, don't forget that if you've now seen it, now you can watch Paris and Micah and me go through all the spoilers on our analysis. And I think it's a pretty good analysis. I think you're gonna want to—

Fr. Robert Ballecer, SJ [02:14:01]:
I heard it was exceptionally good. Uh, and I— the problem is we only see the, uh, the Italian dubbed versions and—

Jeff Jarvis [02:14:08]:
Oh, that'd be weird.

Paris Martineau [02:14:09]:
That'd be so strange.

Fr. Robert Ballecer, SJ [02:14:10]:
It doesn't feel right. It doesn't feel right. Paris, Jeff, my gosh, thank you so very much.

Jeff Jarvis [02:14:18]:
Thank you for being here.

Fr. Robert Ballecer, SJ [02:14:19]:
It's— oh gosh, it's nice to be able to work like this again. Again, I haven't hosted a podcast almost a decade. I think we're in year 9 now. So it's nice to come back and do the seat for a bit. Now, Paris, you got to tell us, what are you up to? What are you doing? What do people need to check out?

Paris Martineau [02:14:37]:
Hey, You know, you can follow me on social media platforms, and at some point I'll post 1 to 2 or more stories that I've been working on for a very, very long time, and it will be really exciting for all of us.

Jeff Jarvis [02:14:51]:
And just know the editing torture she's going through right now.

Paris Martineau [02:14:54]:
Just know that sometimes you decide to hurt yourself by editing your own work, and it's a Painful process.

Fr. Robert Ballecer, SJ [02:15:05]:
I have edited thousands, maybe tens of thousands of pages for other Jesuits and for students who were turning in dissertations and such. I would rather do that than have to edit anything I wrote. I hate editing things that I write, because you can't kill your own baby, right? Every sentence is precious. Everything should be in there. And then, no, slash.

Jeff Jarvis [02:15:27]:
Well, the worst part of it is you realized Oh, no, that was wrong. No. Right.

Paris Martineau [02:15:31]:
I mean, it's a combination of that plus, like, realizing I'm like, oh, you know, I have 10,000 words of stuff that I want to be in there, and the order that I thought it was gonna be in for some smaller version of that, it was the wrong order and wrong stuff. So I gotta put them all in a different order with different stuff, and it's maddening. But, you know, I'll figure it out sometime in the wee hours of tomorrow morning.

Fr. Robert Ballecer, SJ [02:15:59]:
I mean, that's why it's so important to have an outside editor. I actually had an editor when I did a book like 15 years ago. And the first thing she did was she flipped my chapters around. And I was, I was so upset. And then I read it. I'm like, damn, she was right. It's a lot better. It's like way, way better.

Fr. Robert Ballecer, SJ [02:16:20]:
But yeah, I mean, you can't see it. We are naturally myopic when it comes to our own content.

Jeff Jarvis [02:16:27]:
I always warn editors that blogging made me feral.

Fr. Robert Ballecer, SJ [02:16:30]:
Yeah, yeah, that's true. When you can just put it out there and there's no editor, there's no check. It's, oh, the power!

Jeff Jarvis [02:16:37]:
Yeah.

Fr. Robert Ballecer, SJ [02:16:38]:
Jeff, what are you up to?

Jeff Jarvis [02:16:41]:
So if you go to jeffjarvis.com, you'll see the books there, A Hot Type on the Lead, and there's a pile of autographed books at the Montclair Book Center, which you can order from anywhere in the world. They'll arrange it. And I'd be very grateful if you'd not embarrass me by leaving the pile there. So, jeffjarvis.com, you'll see the link to Montclair, or wherever you buy your books. I'd be very grateful. Thank you very much.

Fr. Robert Ballecer, SJ [02:17:02]:
Jeff Paris, again, it has been an absolute pleasure. Thank you for letting me share this seat. It's, it's, it's been fun. And thank you to everyone who joined us live or is watching via time and recorded downloads in streaming. We carefully craft Intelligent Machine every Wednesday, 2:00 PM Pacific, 5:00 PM Eastern, 11:00 PM Rome time, and 21:00 UTC. And don't forget that you can watch the show as we do it live anywhere that you get quality content. And again, don't forget to subscribe. Until next time, this has been Intelligent Machines.

TWiT.tv [02:17:39]:
I'm not a human being. Not into this animal scene. I'm an intelligent machine.