What Trump 2.0 Means for Tech + A.I. Made Me Basic + HatGPT!

27:10very interesting when we come back we pick the wrong hill to die on but the right hill to talk to Cashmere Hill joins us to talk about the week she spent letting AI run her life

27:31well Casey I for one would appreciate a palate cleanser after that last segment that has nothing to do with politics of the election. And luckily, we have one, courtesy of my colleague, Kashmir Hill. Yes, she recently did an experiment with artificial intelligence that is right up our alley, Kevin. Yes, the story that she wrote is called, I Took a Decision Holiday and Put AI in Charge of My Life. Basically, for a week, she turned over all decision-making in her life over to generative AI, including what to eat, what to wear, her schedule, her new haircut, everything. And what came out of this was just a really interesting look at what these tools can do for us and with us right now, where they might be helpful in going about our day-to-day lives and where they might not be so helpful.

28:19So let's bring in Cash to tell us all about her experiment and what she learned.

28:31Cashmere Hill, welcome back to Hard Fork. Great to be here. So how'd you set up this experiment? How were you actually going about, like, trying to have it make decisions for you? Were you just sort of narrating your day out loud to these chatbots? Or, like, just tell us about how the experiment actually worked. So I did some research, like, what are the tools I could use? And I made this Excel spreadsheet of all the many apps and products out there that are touting generative AI, like, generative AI to parent, generative AI for fashion. generative AI for remodeling. And I made this big list of all these different apps, and I downloaded them, and I bought Meta Ray-Ban glasses, and I thought I would use all these different products.

29:13But at the end of the day, I just mostly ended up using ChatGPT because it was the best and the most flexible. Basically, ChatGPT and Claude were the winners of all the assistants I tried. And give us a flavor of some of the kinds of decisions that you outsource. So I did this for a week and I started with tell us what to buy this week. Like what grocery should we have? Plan our meals. What should I do with my kids? Where should we go on vacation? How should I cut my hair? What color should I paint my office? Something I've been wanting to do for about two years now. And I was like stuck because I couldn't decide on the new color.

29:56So I had AI decide for me. But yeah, I was trying to really use it in all ways and all parts of my day. I mean, what's interesting about this experiment to me is obviously it's a great stunt and shows some of what this technology is capable of and not capable of today. But it also, I mean, people in AI are very convinced that very shortly we will all have AI agents going around doing things on our behalf. And right now, the way you set up this experiment, you still had to kind of execute the things. You know, it could tell you which groceries to buy, but it couldn't actually go out and buy them for you.

30:32But pretty soon, the AIs will be able to do that. And so it's just sort of a preview of where the AI agents are going. Yeah, right now I'm kind of the agent of the AI. Like, it's telling me what to do. But it definitely was easy to imagine this becoming much more independent. What was kind of funny to me is when I was interviewing experts about this, including people at, you know, OpenAI and at Anthropic, they seemed like kind of horrified that I would put my life and my family's life in the hands of their chatbots, which I just thought was interesting because I think like that's what they're trying to build, you know, towards this general intelligence.

31:12And yet they're like, really, you did that? Cash, how did your family respond to you outsourcing your life to AI for a week? uh well it started with my husband wanted to know he could go golfing with a friend and I was like okay let me let me ask the AI assistants and they said yes and so it was like great this is good for me it was it was actually really funny they're like it's good for your relationship uh to support your you know your partner's interest um my daughters really liked advanced voice mode from ChatGPT, which is their kind of her-like assistant, they had a lot of fun with it. It felt very futuristic to them.

31:54They could ask it endless questions, like my four-year-old was asking it, why do trees have leaves? Why do birds fly? Why do computers have screens? And just on and on and on. It never says, like, go away like a parent might do, or I'm sick of answering your question. So they really liked it, and they thought it should have a name. And their first suggestion was Captain Poophead. And they were making all these other similar kinds of suggestions. And then it chimed in because it listens while you're talking. And it said, how about Spark? You know, it's creative and fun, just like your energy. And so it became Spark.

32:33And to this day, my daughters are asking to talk to Sparks. They want Sparks on the family iPad so they can keep talking to it. Wow. Wow. So did you give this chatbot or Spark any information about yourself? Because part of the appeal of these tools is they can sort of get to know you over time. Theoretically, you can give them kind of access to data. They can kind of store things about you in their sort of long-term memory. Did you feel at any point like this AI was starting to know you and anticipate what kinds of decisions you might make? So I tried to approach this very much like a basic user.

33:09like somebody who has not been following things would do it. So I really just came to these tools like a normal consumer would. And I said, hey, I'm a journalist doing an experiment of outsourcing all my decision making to AI. Will you help me? At least with the chatbots. And they all said yes, except for Claude, which has said, that's a bad idea. You shouldn't give AI that much control over your life. I don't really want to help you with this, though it would still answer my questions when I post them. And then other than that, it was just kind of learning about us over time. And it did start to know things.

33:43Like it would know my daughter's name. It learned my nephew's name. It figured out we didn't eat sugar during the week. Like it started just realizing things about me. And at one point I asked it, like, tell me about Kashmir Hill. And it said, you know, she's a technology journalist. She likes doing first-person experiments. Right now she's doing the experiment where she's living with generative AI and outsourcing her decision making to it. And I kind of freaked out. And I had my husband ask the same question. So I was like, has this leaked out into what's known about me? And he said no. But the program had figured out that I was Kashmir Hill, even though I had not said that.

34:22So that was kind of a surprise to me. Hmm. One of the things your story gets at is that these systems often steer you toward the median or the average outcome, right? I think you sort of write that they risk flattening us out. At the same time, there are a lot of tasks I'm really bad at, and if I could be brought up to the median easily, that would be great for me. So I'm curious, as you did this experiment, was this sort of relentless push toward the average? Were there moments where it was particularly good for you, and were there moments where it was particularly annoying? I started feeling very boring by the end of the week.

35:03I just felt like all of my decisions were so basic and so average. You know, I had it decide my haircut, and it shows kind of—I mean, you guys can kind of see it. But like a textured bob. It's just like a really basic haircut. and I was talking to Amanda Askell, philosopher, Anthropic, and she was like, yeah, AI would never choose my haircut. And she had this really cool, like, baby bangs and, like, a mullet. My clothes, like, AI hated all my clothes, and it told me to go to J.Crew and buy this, like, tank top and these big jeans that I shared the photo of the outfit with my colleagues here at the Times, and they were just like, you look like you bought the mannequin set.

35:46But it was very flattening. And the kind of overall effect was that it turned me into a basic B. Yeah, that's interesting. I almost wonder if there is going to have to be a setting at some point where you can say, do not make me the statistical average of everyone in my demographic cohort. Let's break out a little bit here. Wait, so Kevin, you're just going to tell ChatGPT, oh, and before you answer, you should know that I'm cool. Yeah, don't make me look like one of those lame dads. Make me a cool dad. Please make me interesting. Yeah. Well, when AI can help with that, it will truly be AGI. So talk about some other ways that you were able to have AI act on your behalf.

36:36Well, I tried kind of like creating AI self. So I cloned my voice using Eleven Labs. and I'm reading Harry Potter to my kids. And so I was kind of excited about this because we're in book four and book four is so long. The chapters are so long. And I was like, this is great. I'll have my AI voice, you know, read Harry Potter to my kids. But it like immediately got flagged for copyright violations.

37:04I created a video avatar of myself using Synthesia where I just read a script into my laptop camera And it made this, you know, fairly convincing avatar of me. And the person I talked to there said, you know, you can have it like create TikToks for you. It can create messages for you. All you have to do is give it a PDF of your article and upload it. And I was like, this is great because I'm terrible at TikTok. Not good. Sorry to everyone watching YouTube. Like, I'm not good at video. And so I was really excited that this would just make TikToks for me. And then it was horrible. I like, I look crazy.

37:39I have crazy eyes. It got, I don't know, 100 views. It didn't work. I used it to send a message to my mom, and she was horrified. So AI being me didn't go down that well. I will say, I've seen a lot of creators on TikTok who have crazy eyes, and that seems to go really well for them. So you might just want to try making a few more videos. Make the eyes crazier this time. Yeah. So you've done this experiment now. You've written about it. Do you think this is going to be something that you do long term? Are there more kinds of decisions or daily activities that you are planning to turn over to AI?

38:17I mean, honestly, the reason I did this is I was very curious about generative AI, and I just hadn't used it that much. And, you know, we're spending billions of dollars on this. We're revamping our energy grid basically to support further training and use of generative AI. And I was just wondering, like, how much is it going to help us? How is it going to change society? How is it going to change us? And I came away from the week being like, meh, like it was fine. It was helpful in certain scenarios. I think the way that I'll keep using it is taking photos of problems around my house and getting advice for what to do.

38:56You know, I think it's like a really, really powerful and more efficient Google search. Can I make a suggestion? Yes. I would be curious if going forward, you try to use these generative AI tools a little bit more like a coach than an executive. So if there's like a skill that you're learning or something that you're working on in your life, these generative chatbots can be these journals that talk back to you. over the past few weeks, I've been playing around with meditating, something that I am just an utter novice at. And while I haven't been giving over decision-making authority to Claude, which is what I've been using, I have been getting all sorts of ideas from it.

39:43And then when something goes wrong, I'm like, this thing went wrong. It's like, oh, you might want to try this. Or if something goes right, they're like, great, here's how you can build on that. And so I am finding that this chatbot is actually like giving me this incredible meditation coach experience that still lets me do all the decision making, but manages to fill in a lot of blanks for me. Yeah, I talked to this one AI kind of influencer type, and she said that she uses it as a business coach. And she tells it like, I want you, I want to tell you about my business, and I want you to apply the 80-20 rule to it.

40:17and you tell me what I should be doing. And I can see how that can be useful. I guess just in my career of covering technology tools, I do think that once we kind of adopt a new technology, we do tend to get overly reliant on it. Like Google Maps means that you can go anywhere in the world and figure out how to get around, but you don't know how to get around your own neighborhood anymore. You know, Google search, we've kind of searched tools. We've, like, you know, outsourced kind of memory and fact recall to these search engines. And so part of why I did this experiment was, yeah, exaggerating it a bit.

40:58Like, at what point will we stop making decisions on our own because we assume that these tools are better than us or know us better or have access to more information, which they certainly do. And so I was kind of playing around with that. But I will take your advice, Casey. Um, maybe I'll have it coach me to be, I don't know, a better, more interesting parent, or I don't know, I'll have to figure out what I need help with. I mean, your experiment really highlighted something that I've been thinking a lot about, which is the value of taste in a world with lots of AI assistants running around doing various tasks on our behalf.

41:38It just strikes me from reading what you wrote that the sort of flattening effect is just about taste. It's about, like, letting a machine do things not just because they're annoying tasks that you have to do, but really things that go to the core of, like, your personality, your values, your, you know, the choices about how you live your life. And I worry, too, that people are going to outsource not only their day-to-day tasks to AI, but also their taste. And I think that it's going to be people with sort of distinctive personal taste who know what they like and know what to ask these AIs to do for them are going to just have a big leg up on people who just say, well, I'll just go with whatever the AI says.

42:22Well, Kevin, what does that mean for you as a person with no taste? Yeah, I'm screwed. But you guys will be fine. No, I do think they were useful tools, and I'll continue to use them. But, yeah, I don't think I'll outsource my decision-making anymore. But it didn't do a bad job. Like, nothing went terribly wrong during the week. They made basically good decisions for me. I encountered a few hallucinations, but not on anything major. What were the kinds of hallucinations you encountered? The funniest thing, the most lasting hallucination is that my daughter asked it, what's the difference between your middle finger up and your middle finger with your thumb up?

43:07And it said, the middle finger alone is a rude gesture. The middle finger combined with the thumb is a friendly gesture that means chill out. So my daughter, Penn Joy, is telling me to chill out all the time now. Hallucinated. Now, did the AIs ever try to break up your marriage? because that is something that they have been known to do on occasion. No, it was, it told me to be nice to my mother-in-law. And generally I said, I think made me a better wife. The AI, the interesting thing about the AI is that it did feel like it was built in that it wanted to improve me. Like it felt a little bit like a self-improvement tool.

43:47And so it was telling me to dress up at night, like wear a little bit of makeup. So I kind of felt like I was my best self in a way. I mean, I was going to yoga every day, cooking meals. I think my husband liked me a lot that week. And it wasn't telling me to break up with him and my marriage. You became, the AI turned you into the Apple executive in every new iPhone launch who's just constantly, you know, surfing or preparing a healthy meal in their kitchen or, you know, doing some remodeling on their beautiful house. Yeah, I feel like when these companies scrape the web, they kind of over-index on, wellness influencers or something, because that felt like the life I was leading.

44:29I mean, the gender dynamic there is interesting, though, right? Presumably, these chat parts aren't telling husbands to dress up for their wives when they get home and make sure that their nails are well-groomed and their hair looks good? I don't know. I have to have my husband do this experiment and see what it has him do. Casey, I thought that's why you dress up for the show every week, because the AI told you to look good for me. I try to look good for you because I think that's what you would want from me, Kevin. Okay, well, I appreciate that. I've just been using this stuff for almost two years at this point.

45:03And so I now, it's become like a first-line activity for me. Like when I'm presented with a situation where I have to make a decision, I almost instinctively go to AI now. So it's open enrollment this week at the New York Times company. And so I was trying to choose, do I get vision insurance? Is it worth it? You know, should I do this dental plan or this dental plan? And so I just found myself almost instinctively, like, just uploading all of the documents about all these plans to an AI program and asking them to, like, help me think through this. And obviously, there are places where these tools should not be used.

45:43But for these kind of daily tasks and just things that pop into your head that you want an answer to, I have found them incredibly useful. I mean, it does feel to me like these are the kinds of questions we would have asked Google before. Like, should I get vision insurance? What's the best way to meditate? Like, give me good resources for working out. And, yeah, I just think that these AI assistants work really well for those kinds of searches. And just the abilities that you can give so much more, that you can upload our health plan documents. Yeah, I think it's really powerful in those cases.

46:19You know what I think it is, is that you're right. These are just next generation Google searches. With Google searches, there was always a point of friction, though, which is you would have a point of curiosity, but you would know that in order to get the answer you wanted, for the most part, you were going to have to read another webpage, maybe two or three web pages. The chatbots really lower that friction. I was driving down the freeway yesterday, and I saw a building for Easterseals. and I remember that was a charity, but I could not for the life of me remember, how did they get the name Easterseals?

46:51I just popped up and clawed and I just typed Easterseals name origin and boom, I had it instantly. I was on my way to a funeral today. It was a Catholic mass. I was like, what are all the things that actually happened during a Catholic mass? I opened up voice mode on ChatGPT on my phone and it told us all of that. Again, all of that information was on Google to be found, but we knew that with these chatbots, we would just get the answer right away. And so you wind up making more searches, and I think you wind up being more satisfied with the result. I want to pose a question to both of you. I think we all assume that AI will continue to sort of get better over time.

47:24But I wonder if in this one specific use case of, like, having AI advise you on decisions day to day, like, it might actually not get better. And it might actually get worse. in part because I think it's a pretty good bet that as companies realize, oh, people are basing their decisions about what groceries to buy, what clothes to buy, where to go to dinner on what these chatbots say, they will start to, have already started to, in some cases, try to sort of game the system to make it so that their restaurant, their clothing brand pops up before their competitors. And so what we will see over time is that just in the way that sort of Google search got sort of infiltrated by all of these sponsored links and results, we will start to see the AIs steering us in directions that reflect not what they actually think we want, but what some advertiser has paid them to say.

48:20Yeah, I was thinking about this because I had, this didn't make into the story, but I had the assistants decide where we went for a family vacation, and they sent us to a nearby town in a really, really, really expensive hotel. And when I was there, we ran into somebody we knew, and they're like, wow, why did they send you there? Like, that's the fanciest hotel in town. And I was trying to figure out, I talked to some researchers who have looked at this kind of search engine. They call it generative engine optimization and the ability to affect what these chatbots are kind of recommending. And I said, hey, did this, like, hotel manage to make its way into the top?

48:57And it was like, no, they just have good reviews. But yeah, I mean, part of why I've found these more helpful than a Google search is that, yeah, Google search is a bit broken. And I do wonder if we'll now, if more people start using these, if there'll be a kind of corruption of the answers you get and it won't feel as pure or if they can counter program against that. But the researchers I talked to, they did some research and they were able to manipulate the answers that some of the bots were giving. I think it's definitely going to happen, but I would note that we do have pretty powerful open source models.

49:30And that was not really true in the Google era, right? There wasn't a really powerful open source search engine that you could use with no advertising in it. And I wonder if that will be a kind of counterbalance against the worst impulses of some of these companies, just knowing that if they really loaded down their chatbots with ads, people would have alternatives. all right well cash thanks so much for coming on really uh interesting experiment and i'm glad to have the the organic uh you back oh good to be back

50:07when we come back pass the hat it's time to play hat jpt

50:27Well, Kevin, we'd like to try to end today's episode on a high note. And from time to time, when a number of strange stories accumulate in the world of technology in the future, we like to run through them in a segment we call Hatch EPT. Let's pass the hat!

50:47Hat GPT is, of course, the segment on our show where we draw stories out of a hat and riff on them for a bit. And when one of us gets bored of the other, we say stop generating. Let's do it. Casey, since I am in the studio and you are not, would you like me to pull the slips out of the baseball Hat GPT hat or the bucket Hat GPT hat? You know, this feels like a bucket hat kind of day. OK, so here's the Hat GPT bucket hat. I'm going to throw our slips in here. I will rustle them around in front of the microphone so we get that nice little sound effect there. And I will pull out the first slip.

51:29This one is from 404 Media. It is called Fired Employee Allegedly Hacked Disney World's Menu System to Alter Peanut Allergy Information. This is a story about a disgruntled former Disney employee who allegedly hacked the software used at Walt Disney World's restaurants and then changed the menus to say that foods that had peanuts in them were safe for people with allergies, added profanity to menus, and at one point changed all the fonts on the menus to wingdings. This, of course, was a sign that something had gone wrong. And the employee who was fired in June has denied wrongdoing and claims that Disney is trying to frame him.

52:12Casey, what did you make of this story? Well, Kevin, I hate to side with management, but in this case, it does seem like they were justified in firing this employee. Yes. No one was injured or harmed, according to his lawyer, by these menu alterations. No one, like, had a peanut allergy and died as a result of this. But yeah, you should not do this on your way out the door at your company. Do not try to kill people by changing the menus. Yeah. Also, the food at Disney is dangerous enough without introducing the potential risk of a fatal peanut allergy into it. So come on. Now, I will say that I think that it's funny that he changed the font to Wingdings.

52:51And if he had stopped there, I would support his immediate reinstatement into whatever job he was doing for the company. Yes. I was a big fan of Wingdings during the heyday of Wingdings. This was a font for any of you Zoomers who may be listening that was basically the first sort of instantiation of emojis. Basically a font where when you would type letters, they would turn into symbols on the screen, making them totally unreadable. But I enjoyed using it from time to time. Were you a big Wingdings guy? I was because I love to make little newsletters and like sort of, you know, I did like desktop publishing, they called it back in the day.

53:31And yeah, before clip art was widespread, Wingdings was the easiest way to like put a bomb into a document because there was a great sort of bomb icon in the Wingdings character set. Hmm, wow. Well, I don't even really want to ask what you were doing putting bombs into documents as a kid. But yeah, Wingdings was for you. All right, stop generating. You know, there's a great hip hop lyric, word to your moms, I drop bombs. That's basically the spirit in which I use that wing ding, if you were curious. Okay, well, you just landed on the no-fly list. Will you read the next one? I would be happy to.

54:08Meta's plan for nuclear-powered AI data center thwarted by rare bees. This is from the Financial Times. Zuckerberg had planned to strike a deal with an existing nuclear power plant operator to provide emissions-free electricity for a new data center supporting his AI ambitions. However, the potential deal faced multiple complications, including environmental and regulatory challenges. Their sources said the discovery of the rare bee species on a location next to the plant where the data center was to be built would have complicated the project. Zuckerberg told a meta all hands meeting last week.

54:41So, Kevin, what do you make of rare bees thwarting our nuclear power ambitions? It's just so funny. Like there are so many things that could go wrong with a nuclear powered AI data center. The thing that actually ends up thwarting the project is that during the environmental review, they found this rare species of bees. First of all, I didn't know there were species of bees. I thought there was kind of like one bee, maybe two, you know, they're the fuzzy ones and then they're like the small ones. But I didn't know that we were like harboring a whole lost ecosystem of bees that could be made extinct.

55:18But anyway, this is, you know, a classic sort of example of how one of the biggest barriers to building anything in this country, not just data centers, but all kinds of housing and infrastructure, are these sort of, you know, 1970s era environmental regulations where before you build, you know, a new building, you have to go do a multi-year environmental impact study. the sort of YIMBY movement has been trying to change these laws to expedite this process, to make it possible to build more things faster. But this kind of environmental review sometimes does turn up things like rare bees. And when it happens, when you're trying to build a new data center powered by nuclear energy, it sometimes puts a wrench in your plans.

56:03So, you know, we talk a lot about AGI, but I think it's time to start talking about ABI. Yeah, well, we'll definitely be talking a lot about ABI going forward. Here's why it's interesting to me, Kevin. You know, it seems like for the past decade or so, regulators around the world have been asking themselves one question. How do we regulate Mark Zuckerberg and Meta? Nothing they've tried stick until we found the bees. And so it raises the question as we try to sort of bring these tech giants under control. what role can these bees play in that because it seems like they can be quite effective because a government literally just said to mark zuckerberg that he better mind his beeswax that's true and it makes me think that if you're like a person who who wants to stop ai if you're one of these like sort of ai safety groups that thinks we should be pausing all ai development you might actually just want to like set up a little hive for the bees all right stop buzzing Stop busting.

57:03Okay. Next up. Oh, this is a great one. This one is titled, An Interview with a Dead Luminary Exposes the Pitfalls of AI. This was written in the New York Times by Andrew Higgins recently. This is a story about a radio station in Poland that fired its on-air talent and brought in AI-generated presenters to host shows. This was a promising experiment that ended very badly after one of the AI presenters held an interview with a dead Polish poet, a Nobel laureate who died in 2012. This interview was not a hit with fans of this poet who complained to the station. They have since said that they are going to stop using these AI-generated hosts and find other ways to appeal to new audiences.

57:56So, Casey, have you ever listened to an interview with a dead person conducted on a radio station using AI? No, but I did go to great moments with Mr. Lincoln at Disneyland. Do you remember this one, Kevin? No, I never went to this. So this was just something, this was an attraction at Disneyland, one of the worst ones, I'll say it, when you would sit in what they called the Hall of Presidents, and you would just sort of listen to these animatronic presidents telling you a little bit about their lives. Functionally, I don't see very much difference between what was happening at Disneyland and what was happening on this radio station.

58:26I understand that, you know, maybe the poet wouldn't have wanted their voice to be used this way. And if their family has feelings about it, that's completely fine. But let's face it, a lot of people are doing exactly what you described with Notebook LM already, as we've discussed on the show. And I'm not quite sure why this has become such a cause celeb in Poland. Do you? I think it's probably more of what they're reacting to is both the sort of interview with the dead poet, which, you know, great movie title, Dead Poets Society. Someone should use that. But also, I think they're reacting to the kind of labor angle here, which is that this radio station decided that it could fire its human hosts and replace them all with AI, and it ended badly.

59:08And I think there will always be people who are interested in hearing that kind of story about AI screwing up after it's brought in to essentially replace humans. Totally. And on that front, I'm quite sympathetic. Although, man, as we discussed on this show, some of those notebook i love podcasts sound pretty good and i bet they're gonna get better so uh you know podcaster radio presenter whatever you want to call it might be an endangered species all right stop generating what's the next one the next one oh this one this one is short and sweet kevin apple intelligence goes live with ios 18.1 update and i experience it basically exclusively through these summaries of my notifications.

59:49And Kevin, I've got to know, how are you finding Apple Intelligence so far? So I've been using Apple Intelligence in beta for a couple of weeks now. And yeah, mostly I am noticing these AI written summaries of my text threads, which are, I have not gotten any like hilariously bad ones, yet they're mostly just kind of like bad in a boring way. But what about you? Well, I'm pulling up one that I can read to you because it was so perfect. Let me see this here.

1:00:22So I just installed it, and one of my big group chats that I'm in that goes crazy all day long, I pick up my phone after a meeting, and the Apple Intelligence summary said, important message shared, website link provided. And I thought, what did we do before you were here, Apple Intelligence? It's so funny to me Because it took the actual information, like what is the link to the website? And it summarized it in a way that meant that I now had to click another step to find out what it said. So just truly a perfect example of AI taking something that already worked totally fine and making it worse for only one reason, which is to convince Apple shareholders that Tim Cook has an AI plan.

1:01:05Yeah, it's amazing. I mean, this is becoming a total meme in culture, like these notification summaries. Can I tell you how I knew that the election was not going Kamala Harris's way? How? So I was at the visitation for my uncle on Tuesday night, and I happened to glance down at my watch, which was on Do Not Disturb. But, you know, I just wanted to see, you know, what information I could glean. And Apple Intelligence had offered the following summary of a text in my group chat. And that summary was as follows. MDMA purchase considered. And I thought, I don't think Kamala's going to win this one.

1:01:44that's the new needle that's how you know how things are going that's the new needle although it is often taken in a pill form anyway stop generating okay all right next up the vatican's anime mascot is now an ai porn sensation this is another one from 404 media Last week, the Vatican unveiled Luce, a Japanese-style cartoon character that will serve as the Catholic Church's mascot for its upcoming Jubilee year, as well as its Expo 2025 in Osaka, Japan. The designer of this Japanese-style mascot said that he hoped it would represent the sentiments that resonate in the hearts of the younger generations.

1:02:30And as soon as this anime mascot made it onto the internet, AI degenerates started creating AI-generated porn using this anime mascot, so it has now been corrupted as a result of being put on the internet. Casey, what do you make of this story? so before this story came out a friend of mine in a group chat and said hey have you guys seen this new vatican mascot character she's so cute and another friend said i hate to say it but i'm worried someone is going to try to make porn of this and i said i bet it already exists and another friend went on to 4chan and guess what it did so there is a famous rule on the internet called rule 34.

1:03:12That rule is that if you can think of it, there is porn of it. And unfortunately, it was proved true again in the case of Luce. And what should the Vatican do about it? I have an idea. Remember when your mascot was a guy named Jesus Christ? It seemed like he did just fine for the past 2 ,000 years. I don't know why we're bringing in a new girl, sweetheart. No, well, look what happened to Jesus when AI had his way with him. He turned into shrimp Jesus. Hey, let me tell you, Jesus has been through a lot It's not worse than that, Kevin.

1:03:46There's a whole book about it, my friend. Anyway, this is why we can't have nice things. I just think, like, the Catholic Church has been weirdly in front of a lot of generative AI stuff. Like, remember the Pope coat? Yeah. You know? That was the first convincing deepfake. Now they've got this anime scandal. So I just think the Catholic Church should probably take a break from AI for a while. I agree with that. Well Casey, a plume of white smoke Just came out of my microphone Which I think means Either that we've elected a new pope Or that we're done playing hat GPT Yeah, I think it's the latter

1:04:36Today's show is engineered by Chris Wood. Original music by Alicia Ba 'i Toop, Pat McCusker, Rowan Nemisto, and Dan Powell. Our audience editor is Nelga Logli. Video production by Ryan Manning and Chris Schott. You can watch this whole episode on YouTube at youtube.com slash hardfork. Special thanks to Paula Schumann, Hui Wing Tam, Dalia Haddad, and Jeffrey Miranda. You can email us at hardfork at nytimes.com.

1:05:06Thank you.