Apple Joins the A.I. Party + Elon's Wild Week + HatGPT
01:13:19 Share

This podcast is supported by KPMG. Your task as a visionary leader is simple. Harness the power of AI. Shape the future of business. Oh, and do it before anyone else does without leaving people behind or running into unforeseen risks. Simple, right? KPMG's got you. Helping you lead a people-powered transformation that accelerates AI's value with confidence. How's that for a vision? Learn more at www.kpmg.us.ai.

All right, Casey, where are we? All right, Kevin, we're here in the belly of the beast here at Apple Park in Cupertino, California. We are heading down to the keynote and the atmosphere is electrifying. What are we expecting to hear today? We are expecting to hear about the following things. Artificial intelligence, AI, generative AI, machine learning, large language models, on device, chip. Those are some of the topics we can expect today. That's the word cloud.

WWDC 2020. What's the vibe here? Just describe it for... The vibe here is that people were chanting about capitalism at 7.30 in the morning waiting to get in. Wow. And how was the breakfast food? How was the food? It was incredible. I don't think I've ever had better overnight oats than that in my entire life. Have you? Yeah, they actually leave them overnight for two nights. They're big innovation.

Is this fun for you? Yes. This is my Coachella. I like just being around a bunch of nerds that are excited about software. Like, these are my people. I'm Kevin Roos, a tech columnist at the New York Times. I'm Casey Newton from Platformer. And this is Hard Fork. This week, Apple makes its big move into AI. Kevin and I discuss our trip to WWDC with the Times' Trip Mikkel. Then, another tumultuous week for Elon Musk. And finally, let's play some Hatch GPT. So, Casey...

We've spent the first couple days of this week at Apple's headquarters in Cupertino for WWDC. This is their big annual developer conference, and this was a big one. Yeah, you know, we love a field trip on this show, and this was actually the first time that Hard Fork has been down to the Apple campus, I think, in our official capacity as podcasters. No, we went down for the Vision Pro demo. That's a great point, and that's why it's so important to have a fact checker on your show. LAUGHTER

Well, that was very memorable, and this might even be more memorable than that. Yes. So this was a big deal event. It's a big moment for Apple and their sort of official entrance into AI. And today we're going to talk about everything they announced, the event itself, what they talked about, what they noticeably didn't talk about.

and what it all means. And to do this, here with us in the studio, we've got my colleague Tripp Mikkel here. Tripp has been reporting on Apple for almost a decade. He wrote a book about the company, and he's just totally steeped in the politics and history of Apple in a way that neither of us are. Absolutely not. So, Tripp, very happy to have you here. Welcome to Hard Fork. Thanks so much. Yeah, thanks for having me. So I think we should just start by setting the scene a little bit. Yeah.

What is WWDC Trip and what typically happens there? It's the biggest gathering of developers that Apple has every single year. It's all the people who make the apps that make your iPhone powerful and worthwhile. And they're all the people who are going to sit on their edge of the seat while you're

while Craig Hare Federighi parades around on now a video screen. You're like there to watch a movie, you know? It's not like what it used to be where you were there for like a live action show and demo. Yeah. I mean, as someone who does not sort of cover Apple as a regular beat, I was just like sort of surprised by how much

kind of inside lore there is around this. Like the whole event started with this promotional video where a bunch of like Apple executives are, are sitting in a plane and then they go skydiving. And Craig Federighi was one of these Apple executives who's very well known for his like white hair has like a special helmet on and people are laughing about that. And then he does like a bunch of parkour moves. There's like this sort of stunt double sequence of,

The whole thing was very sort of bewildering to me as a person who does not, I mean, I know who these people are, but I do not know the whole sort of like cinematic lore of the Apple universe. Well, it was very over the top and it was ridiculous, but I have to say I did enjoy it. Like, it pains me to say that, but I was all the way in by the end of that opening video. Yeah.

You were giggling. They had you with the throwback iPod at the start. That was what Phil got you, the original click wheel right there. Yeah, they had the click wheel iPod that they were using, I think, to help navigate the plane. And then when they all jump out of the plane and they pull their parachutes, every parachute had the name of a different Apple operating system on it. And they sort of went off one by one. And it was a little introduction into all the stuff that they were about to tell us about.

Yeah. I would just say it's like, it's very interesting to be on Apple's campus. It is a very interesting and strange place. Like it's this giant circular building. It's like one big building on their campus. And it's kind of hard to know like at any point where you are in real

in relationship to where you came in. It's very easy to get lost. Exactly. And it sort of feels like being in kind of like a spaceship that's like very modern and slick and minimalist and futuristic, but it's also a little eerie. You know, people were chanting, Kevin. You know, when we got there, there was a sort of pen where...

where they had the developers and the developers were like chanting like, Dub Dub DC, Dub Dub DC. And then when they finally let people in, there was whooping, there was cheering, there were employees who were whooping and cheering. So they really sort of make you feel like it's Christmas and Coachella at the same time as you're coming into

this event. And it's also like in a giant ring. So it's like the Coliseum, right? Like there's like a sporting event or something, you know, where like you got a bunch of Arsenal fans chanting outside going into Emirates Stadium or something like that. Exactly. Okay. So that is the sort of general vibe of dub dub as they call it, which I'm probably never going to say again. It feels slightly embarrassing to call it dub dub, but that was sort of what it felt like to be there. Let's talk about what they announced. So you sit down, there's this sort of

keynote presentation, Tim Cook comes out. He starts listing off a bunch of kind of incremental updates to various Apple products. Casey, I'm wondering if anything stood out to you. Oh, I mean, sure. A bunch of stuff. I think in terms of like what is actually going to just be a useful everyday thing that people can do, if you have a Mac laptop or a desktop, when the new version of iOS comes out, you're going to be able to control your iPhone from your desktop, right? Have you seen this? I saw it, but like, I don't understand what the point is.

What's the point of that? Is it hard to get out of your pocket? Yeah, here's the thing. So I have my phone charging next to my big monitor every day, and there are just so many apps that are on my phone that are sending me notifications that are not on my desktop. So maybe I ordered something on DoorDash, or I'm having a delivery on Instacart, or there's some messaging app that is on my phone that is not on my desktop. Now, instead of picking it up 40 times a day, unlocking it, putting it back, recharging, now all of a sudden it's just going to...

beep and boop on my desktop and I can use it there. You have to understand that Casey is a screen maximalist. He doesn't leave his home. Everything comes to him, right? He wants so many screens that he even wants screens inside his screen. I want a screen on my screen that shows me a virtual screen that represents a physical screen. That's what I want. Okay. So there was this, this feature called iPhone mirroring. Yeah. Um, uh,

I was fascinated by the reception that the iPad calculator app got. I mean, this truly, sitting there in Apple Park, this was the loudest cheer of the day was when they said that the iPad was finally going to get its very own calculator app. I was afraid for my life the way people were cheering. I thought, are we safe here? There might be a stampede. Why were people so excited about the iPad calculator app? I have no idea why that was such an obsession for people for so long. I mean,

been times where I have wanted to use a calculator while using my iPad and like you the first time you realize there's not an app it's extremely confusing how many times are you on your iPad I'm always trying to add things up like how many times did Kevin interrupt me on the last podcast how much money does Kevin owe me there's all sorts of calculations I have to make and sometimes I

So I think part of the, this is an inside joke because people have been complaining about the iPad calculator app for years or the lack of one. And finally they made one, but I should say like the app itself is quite cool. They showed off this sort of like AI feature that will, you know, basically if you're writing equations on the iPad with the Apple pencil, it'll sort of like,

answer the equations for you in what appears to be a version of your own handwriting. So there's some cool stuff around that, but I think it was mostly a campy inside joke, like, ha-ha, we finally made a calculator app. Well, yeah, so that feature that you just described is called Math Notes, and it was one of the coolest things that they showed off. Now, obviously, I don't know anything about math, and they went through this elaborate demonstration where somebody was trying to calculate the height and the angle of a ping-pong ball, I think, for reasons that I've already forgotten. But

actually just sort of watching the answers update in real time as you're just kind of scratching around with a virtual pen. Like, you know, I, we feel a little bit jaded and cynical about this sort of thing today, but I truly think like 10 years ago, if someone had showed you that it really would have felt like science fiction. And now it's just, you know, going to be a free update to your iPad. Totally. So a bunch more, what I would characterize as small updates. Apple will now also let you schedule a, an iMessage. So you know, are you going to do this? I,

I am because I am an inveterate late night texter, as you know. But I don't know the etiquette about this anymore because I assume, and maybe this is a bad assumption, that people don't have...

have their phones turned on late at night when they're sleeping. This is what I assumed. Isn't everybody just in sleep mode? But are you not texting people on like the East Coast? Or you're like, isn't this going to be weird when this lands at like two in the morning for them or something like that? Well, I think it's a fair question. Here's my feeling. My feeling is if there's a time of day where you don't want to receive text messages anymore, you just tell your phone, go into sleep mode and don't send me notifications. Now, I hear you. You know, you can't rely on apps

absolutely every single person to do that, but if that's a big problem in your life, that's what I'd be doing. But if you've got a VIP person who's going to text you, it's still coming through while you have it in focus mode or do not disturb or whatever, right? Wait, who are these VIPs who are texting you, Tripp? Oh, God, I don't know. Dua Lipa, Barack Obama. Yeah.

They're constantly trying to text Tripp and he's like, it's two in the morning, Barack. I will say that, you know, there is this idea out there that like in a work setting, you know, if you were a manager, it actually is really bad form to text your employee at like 8 p.m. being like, you know, don't forget to bring the presentation tomorrow or whatever. So yeah, maybe scheduling something to send first thing in the morning. I'm not denying that. So you've got some platformer restraint now. Yes, exactly. Yeah. We have to, you know, keep the employees happy. But like,

I think there's a serious use for this, but I'm just not one of these people who's ever going to use this feature. So, Tripp, what stuck out to you as one of these features from the first half of this Apple keynote? I thought the messaging updates were the most important things, really, to be honest. I know you're like, the scheduling to send, but that's something that people have been asking for a long time. And then also, the notion that you're going to have more emojis that you can tap back

with. I've always felt like, why are there only five of these, right? Slack has like 500 and you've always been constrained in Apple. Every time I use the haha tap back, if you're not an iOS user, we should say, you know, you can respond to a message, you press and hold on the message, and then you get a very small number of choices. You know, you can do a question mark or an exclamation point. Thumbs up, thumbs down. Thumbs up, thumbs down. Heart, I think is probably the most used one. And then there's just this haha. But every

time I tap ha ha I'm like this this is not how I actually feel you know how I'm actually feeling is maybe lol lmao ha ha

Ha ha sounds like a fake laugh, you know? Yeah. So there's only been a fake laugh. Now you're going to be able to use any of the emojis or even create your own emoji using AI. Which we should talk about. Yeah. So that was sort of the first half of the keynote, which is what I would describe as basically these sort of like minor updates to existing Apple products. Then we got to the second half. And this is the big one because this is when Apple announced all of its AI-related projects, which it is sort of,

labeling under this big umbrella of Apple intelligence. Which is just so sneaky. It is so sneaky. So AI at Apple no longer stands for artificial intelligence. It now stands for Apple intelligence. And I want to talk about that. But

But before we talk about what they actually announced and all of this sort of strategy behind it, I'm wondering, Tripp, can you give us sort of a capsule summary of Apple's history with generative AI? Because my understanding is like up to this point, they have not been considered kind of a major player in generative AI. And I understand from your reporting that they've actually kind of struggled to figure out what to do with this new technology that everyone's been going crazy over since ChatGPT came out. So just tell us like what...

Apple's role or approach to generative AI has been? - Silence? Yeah, I mean, you know, that's the best way to explain what they've been doing. They just, they have not talked about it at all. And it got to the point where it just became uncomfortable when Tim Cook was doing briefings with Wall Street, because that's all Wall Street wanted to know about.

And eventually the pressure got so intense from Wall Street, which wanted to hear, what are you doing with generative AI? After Microsoft had released the co-pilot PC and Samsung had released a phone earlier this year and everything else, that Tim essentially broke with Apple protocol around secrecy and said, we will announce something generative AI related in a few weeks. This May, he said that.

And is your sense that Wall Street believes that there is just a ton of profit to be made with these features and that Apple needs to start doing that immediately? Or are they worried that Apple will just sort of be left behind if it doesn't have a response to what folks like Microsoft are doing? Probably twofold. One, as you all have talked about,

A ton. Generative AI is presumed to be this thing that's going to create trillions of dollars in new economic value, right? Two, like the iPhone's been a little stuck, a little stagnant. There's not been a major reason to upgrade in a long time. You know, it's like, do you need your camera to be slightly better? Well, like we did that last year, but we're doing it again this year, right?

So like bringing these tools to your iPhone has the potential, and this is what Wall Street is betting on, to make your iPhone interesting again and make it worth buying a new one. And why have they been so slow relative to other large tech companies? I mean, you know, Amazon, Microsoft, Google, these companies have been sort of turning themselves inside out to try to rush these generative AI products into their existing products and onto the market. Like,

What explains Apple's relatively sort of sluggish embrace of AI?

I think the best way to look at this is just look at what's happened with Siri at Apple, right? I mean, that's probably been at the forefront of what we all associate with AI. Their virtual assistant, which they released in 2011, has just languished ever since then. At Apple, things get attention where they don't. And Siri hasn't gotten attention in a long time. One other thing that you've reported on that I thought was interesting as sort of a way to explain Apple's position here is that

their secrecy, they're, you know, they're a very private company. They don't, they don't publish a lot of like research. They're not sort of part of the kind of academic AI community has hurt their ability to sort of attract top talent in AI because top researchers want to be able to publish their stuff and show what they're working on. And so it actually, in some sense has been hard for them to build up like a world-class AI team because of this sort of reputation for secrecy.

Absolutely. They've acquired a number of companies, brought in really top research talent, and those people have stayed about four years, vested, and then walked out the door. Because they want to be part of the AI community that is driving the conversation around where generative AI is going. And when you go inside the circle, as you guys experienced, you get lost. I mean, like, literally, you got lost there, right? And that's kind of what happens to people who join Apple. Yeah.

Okay, let's talk about what Apple actually announced this week. So Apple intelligence is sort of the umbrella term that Apple is giving to all of its generative AI products. One of the things that they announced that got a lot of attention was this kind of revamp of Siri. They have given it a basically a brain transplant using generative AI. They say that the new Siri will be more conversational. It'll be able to

answer more complex queries. You can now ask Siri, like, play the podcast my wife sent the other day. And it'll kind of go back in your text messages. It'll know, A, who your wife is. It'll be able to sort of like search through your text messages and find the podcast and then actually use a podcasting app to play that all-in-one sort of command. Yeah, if it turns out that your wife recommended the all-in podcast, it'll actually show you the number for a divorce lawyer.

I thought that was interesting. You can also now type to Siri as well as talking out loud to it if you're in a meeting or you don't want to talk out loud for some reason. Also great in a hostage situation. If you're captors, you're worried about them hearing you.

But if they've somehow given you access to an iPhone, you can now type to Siri. Siri got a literal glow up as well. It glows around the edges instead of shows up in the bottom of your screen. Yeah, I mean, it sort of looks like they tried to design what it would look like if Jesus were present on your phone, you know, where it just sort of like shimmers in various like neon gradients. Yes. So lots of interesting announcements related to this new

Siri, they said it can also answer a bunch of questions right there on the device. This is one thing that Apple has been working on that they hope will differentiate it from other AI providers is that unlike using ChatGPT, where every request has to go to OpenAI servers and you have to be connected to the internet, a lot of this new stuff that Apple is announcing under their AI umbrella will happen right on your device.

And if it doesn't happen on the device, it goes to this thing, private cloud compute. Which, Tripp, can you explain to us a little bit about what this thing is and why you think Apple designed it? Because they can't do everything on the device. Like, you know, our requests are going to be too complicated. They showed a great example of driving to the airport to pick up your mom, right? And you saying, hey, what time is my mom's flight landing? Siri would essentially go into your email, find the flight number, check with...

let's say, Delta and see what time that flight is going to land and then come back with a very current update, that's going to require some cloud inference work. You can't do that all on the device itself. And I thought that was interesting because I always just tell my mom to take an Uber. But,

Let's talk about the privacy feature here, right? Which is they've designed this in a way that they say is more private than the other systems. Why are they able to make that argument? What does it do? They have designed a cloud network that will be powered by their own semiconductors.

and it will send up that request like, hey, we need to double check what time is this Delta flight. Unfortunately, they didn't provide us a great example of what would be done on device and what would be done off. So I'm riffing a little bit here, but it'll check that, it'll send back the answer. And then what'll happen as a result of that is like all that information will vanish. They made that commitment that we're not gonna keep any of that, we're not gonna log it. It's not gonna be used to help improve our system

and or help what we're doing in any way, shape, or form. And I believe they said that outside experts are going to be able to come in and sort of verify that this is private for real. They really made a big deal out of privacy. This is clearly one of their core, you know, missions and priorities

Maybe part of the reason that they have been a little more tentative when it came to implementing AI is that it's been just hard to know how to do it without sending a lot of people's personal information somewhere they don't want it to go. Yeah, this is something they've talked about for a decade now. But now that the AI moment has arrived, it's become something that's a real asset for them. You know, it's a principle that they can really push forward and really persuade people around.

So Apple also announced a bunch of other tools outside of Siri that use generative AI. A lot of them are around kind of writing and communication. You can now use AI to modify text in pretty much any application on your iPhone. There will now be a tool called Rewrite.

Say you're writing an email or a cover letter. You can highlight it. You can go into this tool. You can say, rewrite this to be more professional or more friendly or more concise. And it will just do that right there in the message window.

But am I right that there is no actual just write command, that it will only rewrite text and maybe sort of exempt apples from some of these difficult conversations about what sort of things people might want to generate with just sort of a pure text generator? Yes, it's a much more limited sort of tool than you've seen in other chatbots, and I think that's probably on purpose.

They just want you to basically be able to take something you've already written and kind of zhuzh it up a little bit. So Apple's tools can also now give you high-level summaries of your messages. This is a feature that I actually thought was pretty cool. If you're like sometimes...

you know, you'll be in a meeting or you're driving or something and you get back to your phone and you have 72 unread messages on some like unhinged group text that you're on. And it can now use AI to basically just summarize what's happening so that you don't have to wade through all 72 messages that you missed. You know what I hope the summary is for the group chat that I'm in that's going off all day when I get this feature? I hope it just says, your friends need to get jobs. That's what I hope it says to them. Yeah.

Yes. So that's the text AI stuff. And there's a lot more there, but we don't have time to go through it all. But we should also talk about the image stuff that they announced because this was something that also got a lot of attention. So Apple has built its own sort of diffusion models like Dali or the ones that are in Gemini and Apple.

These are pretty limited, like you can't go in and create sort of like photo-realistic images, and we should talk about why we think that is. But what you can do now, or will be able to do as soon as these updates are released later this year, is to do things like clean up your photos. This is a version of a thing that Google has had called Magic Eraser, Apple's version. You can go in to a photo that you've taken if there's someone, you know, in the background that you don't want to be there. Like Kevin. Yeah.

Yes, you can basically just like use AI to erase them from the photo. You can also use these things called Genmoji, which are their sort of AI emoji that you can now make. You can create custom emoji using a prompt like you would on any other AI image tool. Is this like the death of emoji? I mean, you think about it, like the whole emoji process is once a year, the Unicode Consortium approves some handful of new emoji and it just sort of grows every year and year. If you can just make new ones on your phone from now on, what the heck do we need the Unicode Consortium for?

Yeah, I had the same question. I do think this is going to cause chaos in a lot of group chats because people are just going to make Genmojis like roasting each other and it's going to devolve. Let me just say, I'm so excited to test the limits of what Genmoji is willing to let me get away with. The things I want to do with this eggplant emoji, I literally can't even say on this podcast, but it's going to be a good time. So they also announced this thing called Image Playground, which is basically a dedicated app that you can use to kind of create

various images out of prompts using AI. It's very limited in terms of the styles. It'll let you do sort of like a sketch style and a kind of animated style, but it won't let you do sort of photorealistic images. Yeah, it's like the styles are just like things that will not confuse anybody as to whether that they are real images or not, right? So nothing that looks like a photo, it's truly just designed to be whimsical and delightful. Yeah. Tripp, do you think people are going to use this stuff at

Apparently Casey is to do strange things with eggplant, so we can't really even talk about on this. Do kids listen to this podcast? I have no idea. Sometimes, they email. Yeah, so there was a lot of chatter online about this one image that Apple showed off of a superhero mom. Basically, someone was texting with their mom, and they wanted to like,

I don't know, like gas their mom up a little bit or something. So they were like, basically they told the sort of image generator, like make an image of my mom as a superhero. And it sort of went into their photos, found an image, you know, a picture of the mom and then kind of made this cartoon version of her wearing a cape. Yeah.

And I would say I did not think it was a flattering image. I haven't seen what – assuming this was based on a real person, I have not seen that real person, but I'm very confident that that person is much more beautiful than what was represented in that image. Yeah, so this feature could tear apart families is what you're saying. Yeah, imagine you send that and people are like, this is what you think of me? Am I a joke to you? Yeah.

So a couple more things we should mention. These AI tools are not out yet. We were not able to actually test them ourselves. We did get a little demo, but when I asked, like, can I put in some prompts? They were like, sir, your demo time is over. Wait, tell them the prompts that we wanted to do.

So the one that I wanted to do was create an image of the founding fathers. Yes, exactly. Because this is the one that gave Google's Gemini so much trouble and caused that big scandal over there. And they said, basically, you know, you'll be able to try all the prompts you want when this is out later this year. That's right. The prompt I wanted to try was draw me a picture of the person who won the 2020 election. Yeah.

but didn't get a chance to do that either. So we have not actually gotten our hands on these tools, but Apple says they will be available, some of them in beta this fall with the release of iOS 18. Other stuff like kind of these multi-step actions will come over the course of the next year, they said. They will also not run on every Apple device because they are processing intensive. You will only be able to run these AI features on Apple,

the newest iPhones, the 15 Pro and Pro Max, and the newest iPads and MacBooks that have what they call these M-series chips in them. And this gets to what we were talking about earlier, which is why does Wall Street care about this? Well, you know, I do think that maybe we are going to see a new upgrade cycle of these phones just driven by AI. That might be wishful thinking, but if you want to know, well, what is something that is on the new phone that is not just a slightly better camera, AI gives us one answer to that. Yep.

One other thing we should talk about, because I was fascinated by this, was the partnership that Apple announced with OpenAI, the maker of ChatGPT. This had been rumored for a while, but they finally sort of gave more clarity about what exactly the partnership looks like. So now if you basically give your Apple device a query that is sort of too complex for it to handle right there on the device or in its private cloud,

It will give you the option. It'll pop up a little thing saying, do you want to send this request to chat GPT? And, you know, chat GPT can then answer that question and present it. And I would say this was interesting to me on a couple levels. One of which is what is Microsoft thinking right now? Because Microsoft has this huge investment in open AI. They've bet a lot of money on open AI and they've integrated open AI technology into a lot of their products.

And here comes OpenAI doing a deal with Microsoft's biggest rival. And I just thought someone in Redmond, Washington right now is really upset. They're clenching their fists while they're watching this keynote. I mean, maybe, but also as far as we know, OpenAI is not a profitable company. Microsoft will only realize the value of its investment if OpenAI does become profitable. And what better way for them to become profitable than getting their technology onto a billion devices?

Yeah. I don't know, Tripp, what's my take on the Microsoft situation? I mean, the Microsoft situation, it seems like this is upside for them, right? I mean, all of these queries are going to be going to Microsoft Azure. So that has a certain degree of benefit for Microsoft. For OpenAI, you know, I mean, it must have been deflating to sit there in the audience and know that you were getting primetime exposure here in front of,

the entire Apple universe and then have Apple say, and we're going to add Google and possibly Anthropic and everybody else. Because what Apple essentially announced was we just created a new app store for model makers. And OpenAI is just the first one through the door, but we're going to bring in and usher in others as well. Well, let's unpack that a little bit because I think it's important to understanding how Apple views these AI companies. What Apple said in its keynote and in its presentations about all this AI stuff is that

They are doing most of the AI requests from users on their own models, right? These small models that live on your device, these slightly larger models that live in these private cloud servers. They are really trying to only reserve the most complex queries and requests for these external models. And they made a point of saying it is not just going to be ChatGPT, right? We plan to add more AI models over time.

So that struck me. I was sort of thinking maybe this would be a much tighter-knit integration, and it sort of sounds like Apple is kind of treating OpenAI as one of many suppliers that it hopes to build relationships with to handle those, maybe call it the top 10% of complexity when it comes to what users are asking for. I mean, to me, this is the real acknowledgement of where Apple truly is behind. They have not trained one of these

absolutely massive large language models like a GPT-4. And so they do need to rely on external partners. And there's no real reason for them to rely on only one external partner, right? They control all of the demand for the iPhone. The best experience on the iPhone will just be whoever has the best model at the time. And so they want to give themselves some flexibility there. And I would argue, like, are they behind? Or do they just not want to be in the chatbot business?

The thing that I keep thinking of is just how often you see hallucinations in and among chatbots. And like in Apple's controlled universe, it's a degree of control that they're not willing to surrender. Well, it's a really great question. You know, the thing that we've observed about Apple over the past couple of decades is that the more important a technology becomes, the more pressure they put on themselves to bring it in-house, right? And so as these things get more and more important, you know, like maps becomes really important. All of a sudden, okay, you know what? Actually, we're not going to,

use Google Maps anymore. We're going to build our own version of that. If you think about the conversations that we have on this show about AI, the people that are building it are telling us this stuff is getting exponentially better. You know, we're only a few leaps away potentially from something approaching an artificial general intelligence, maybe something that's even superhuman. If you believe that,

At some point, you have to figure Apple gets involved, right? Apple has to have an answer to that. Otherwise, what, it's just going to be like, no, we'll just let Google do it? Well, there's a great quote from JG, John Gianandrea, who runs Siri for Apple, in Josh Trangle's column in the Washington Post, where he basically said...

the people who are saying AGI is around the corner are freaking bananas. And like, we just think they're insane. We don't, we don't see that coming. We're trying to make technology that's useful. So Apple has already said, like, we're not, we're not playing that game. We're not chasing the sentient beings of tomorrow. We are playing with people's lives today and trying to give them a utility that,

that they find valuable. And I'll just say, I'm very curious to see how that quote will age over the next five years. Oh, you think sentient beings are around the corner? I'm not saying sentient beings, but again, it's like, well, you just like look at the rate of progress over the past 10 years. You assume that you're going to see some similar rate of progress over the next five years. I just think we'll have like very, very powerful systems and that it's one thing for Apple to say now, oh, well, you know, we're, we're, you know, that's, that's for other people to worry about. We're just going to focus on like making Siri a little bit better. You know,

You know, I don't know. I don't know that's the best approach. I will say, though, I think at the broadest possible level, I think Apple's approach here makes a lot of sense to me, right? They do not need to be first to the frontier. They do not need to have the most powerful models. The place where they are going to differentiate themselves is that, you know, a billion people

people plus are already walking around with Apple devices. Many of those can now run this new AI stuff. They just have a huge advantage when it comes to distribution. And if any company is going to make, you know, your parents, you're sort of like people who are not sort of early adopters of technology, if any company is going to be able to bring this stuff to them, I think it's Apple. Yeah.

It's not only that, but I honestly think privacy is huge here. You know, when we're talking about people's personal information, I think this is really going to resonate with people. And I think that's one of the things, yes, like OpenAI's ChatGPT voice assistant was really impressive. Like the versatility and the conversational nature of it was really sophisticated. But...

Do you trust ChatGPT to know your calendar? Do you trust it to go in and manage your email? I'm not sure anybody else has done a good job

a good job of persuading customers that privacy is at the forefront of what they do in quite the same way that Apple has. Yeah, you know, Apple described this stuff as AI for the rest of us. And I think that is going to be the standard that we should try to hold them to as this stuff gets rolled out, right? Because to me, what that means is it's going to be really easy. It's going to be something that even people who are not particularly tech savvy can use.

And also, it does sort of carry this idea of like, we're going to bring a lot of security and privacy protections to it. You know, you're not going to be, you know, taking your life in your hands when you use this stuff. So that's the bar that they've set. I'm very curious to see if they can live up to it. Yeah. I'm really unsure about what

the world looks like a year or two from now when hundreds of millions of Apple device owners just have all this stuff running on their phones. I think it, maybe it doesn't change that much. Maybe it's so limited that it can't really change much more than, you know, how we write emails and make emoji and stuff. But I think there's a real possibility that this just

becomes a big cultural shift, like giving people access to this kind of technology at scale, um, through devices that they already own just seems like I, I just am genuinely unsure what changes that's going to bring. I agree with you. I, I am always, uh,

overestimating in my own mind the degree to which people have already picked up on a lot of the stuff that we talk about on this show. I had a chance this week to give a talk to a group of about 4,000 people who work for in government finance roles. So city treasurers, you know, sewer authorities, that sort of thing. And at the start of the talk, I said,

How many of you have used ChatGPT? And like barely half the hands in the room went up, right? So there are so many people out there who have maybe just heard about this stuff and they're mildly curious about it, but they're not gonna go out of their way to find it. But when it lands on their phone, it becomes a different story. And I agree with you, Kevin, I think we're about to see a big shift. - It moves from software programmers to like everyday people, right? And that's a big difference, yeah. Also you're on the sewage speaking circuit? I had no idea.

The loveliest people you've ever met in your life. Yeah. Lovely people. Trip, thank you for helping us navigate all this. We'll see you next time. Thank you. When we come back, it's been a momentous week in the life of Elon Musk. We'll talk about some of the biggest changes and challenges in his business empire.

This podcast is supported by KPMG. Your task as a visionary leader is simple. Harness the power of AI. Shape the future of business. Oh, and do it before anyone else does without leaving people behind or running into unforeseen risks. Simple, right? KPMG's got you. Helping you lead a people-powered transformation that accelerates AI's value with confidence. How's that for a vision? Learn more at www.kpmg.us.ai.

I'm Julian Barnes. I'm an intelligence reporter at The New York Times. I try to find out what the U.S. government is keeping secret. Governments keep secrets for all kinds of reasons. They might be embarrassed by the information. They might think the public can't understand it. But we at The New York Times think that democracy works best when the public is informed.

It takes a lot of time to find people willing to talk about those secrets. Many people with information have a certain agenda or have a certain angle, and that's why it requires talking to a lot of people to make sure that we're not misled and that we give a complete story to our readers. If The New York Times was not reporting these stories, some of them might never come to light. If you want to support this kind of work, you can do that by subscribing to The New York Times.

Well, Casey, it's been another big week for Elon Musk. This character just keeps getting himself into scrape after scrape, Kevin. It's true. He's the gift that keeps on giving at least to podcasters like us. So let's talk about some of what has been happening to Elon Musk over the past week or two, because I think it's safe to say

It's been a little wild. Yeah. You know, this first story that we want to talk about gets at one of the big questions about Elon Musk, which does he have enough money to get through the day? Yes. So Tesla shareholders this week voted to approve a $45 billion compensation package for Elon Musk. Back when Elon Musk's compensation package was first sort of

vetoed by this judge in Delaware. It was reportedly worth $56 billion. Currently, it's valued at around $45 billion, but it's stock options, so it's always going up and down. All of which is to say, this is a lot of money, no matter how you count it. Yeah. Casey, have you been following this compensation debate? I have been, because it's a very unusual case where a judge actually intervened to say that the company could not pay the CEO as much as he wanted to be paid.

Yeah. So let's talk about the history here. My favorite wrinkle of this entire story is that this entire thing started because of a heavy metal drummer. Did you know this? No, I did not. So back in 2018, when Tesla was a much smaller company, it was worth about $59 billion at the time, Tesla's board came up with this compensation package for Elon Musk. And

And basically, they set out a series of financial targets. And if Tesla hit those targets, which included boosting the company's market cap to about $650 billion over the next decade, he would be entitled to this massive payday. And at the time, people said, well, sure,

Promise him whatever you want, because there's no way Tesla is going to be worth that much. Yeah, that was a talk. I mean, you want to talk about a stretch goal, right? The idea that you could essentially what more than 10 X the market cap of a company. I can understand why there were at least some people who said like, yeah, sure. If you can do that, buddy, we'll give you a lot of money.

Yes, but then Tesla broke through all these thresholds, became one of the most valuable companies in the world. Its market cap rose actually above a trillion dollars at one point. It's now back below a trillion dollars. And Elon Musk was, he thought, entitled to this giant compensation package. And he would have gotten it if not for one man, Tesla shareholder Richard Tornetta.

And Richard Tornetta, we should say, he is not a major shareholder in Tesla. In fact, according to some reports, he only owned nine shares of the company's stock. He was also a heavy metal drummer who was in a band called Dawn of Correction. Yeah.

And he said, there's going to be a correction to Elon Musk's pay package. Exactly. And in 2018, he filed a lawsuit basically saying this is way too much money to pay Elon Musk. This pay package is unfair to investors because every dollar that goes to Elon Musk is a dollar that Tesla shareholders do not get.

And in 2022, a judge in Delaware voided this pay package, ruling that Elon Musk had basically stacked Tesla's board with his friends. They were not his brother, I believe. His brother was on the board at the time. And because of that, this was not a fair package for Tesla's shareholders.

Tesla's board then kind of responded by asking shareholders to vote. Basically, do you want Elon Musk to get this money that he feels he is owed, or do you not want him to get this money? And this week they voted that they think he's entitled to it. And as a result, Tesla hopes the Delaware court will reinstate this compensation package. So, Casey, what did you make of this?

Well, it's a really extraordinary thing in the world of tech to see any of the most famous and richest tech CEOs be held accountable for any reason. And so I, for one, was heartened by the fact that a judge came along and said,

Actually, it does seem weird that your brother is one of the people saying that you can take $45 billion out of this company, right? Maybe we ought to design a better process. So, you know, there are going to be a range of views on this. Other people would say, look, you know, this person created a lot of value at Tesla, and he should be entitled to sort of reap the upside of that. But it's $45 billion we're talking about. It seems to me that we could compensate somebody with maybe a little bit less than that.

Yeah, so Elon Musk obviously has been campaigning for this pay package for weeks now, trying to rally Tesla shareholders to vote for it. His position basically is, look, I took a big gamble. I poured everything I had into Tesla. It became one of the most valuable companies in the world, and I should be rewarded for that.

And there are also a bunch of other people, you know, Silicon Valley investors and entrepreneurs who basically say this was a dangerous precedent, this ruling in Delaware, because, you know, people deserve to be compensated when their companies become successful. And so a lot of people rushed to his defense.

But there were many Tesla shareholders who objected, including Norga's Bank Investment Management, which is the manager of the largest sovereign wealth fund in the world for Norway, who said that they had voted against the deal because basically they thought it was too big.

Well, and look, I just think it is transparently false that Elon Musk has poured everything that he had into Tesla when during that same period, he was starting company after company. Most recently, he started an AI company. We know that he's wasted ungodly sums of money and attention on the company that used to be named Twitter. So if you're a Tesla shareholder, the idea that this guy does nothing but stay inside Tesla all day and work is just obviously false. Yeah. I mean, it's...

It's unusual. These kinds of pay packages, you know, they are not uncommon in the tech industry for executives to have pay plans that are tied to specific financial goals. And this is sort of the most high-profile instance that we've seen of a judge actually stepping in and saying, that's too much. Casey, what do you think happens next now at Tesla with Elon Musk?

Well, Kevin, I think this is actually really exciting because we're finally going to get an answer to the question, how much ketamine can you buy for $45 billion? If I were a ketamine dealer somewhere in the greater Los Angeles or Austin regions, I would be thinking about...

Buying a house, putting my kids through college. Very exciting. No, look, Tesla doesn't have this industry locked up the way that it did maybe 10 years ago. And I would be surprised if these Tesla shareholders didn't have some buyer's remorse a few years hence when other electric vehicle makers are doing much better and Tesla isn't nearly back to those trillion dollar heights that you were talking about a minute ago. What do you think? Yeah.

I think at the same time, there was this sort of opportunity here that Tesla shareholders had to kind of impose some accountability on Elon Musk, who has not historically had a lot of that at any of his companies. There was a good opinion piece by Brad DeLong in The Times the other day, who is an economist at Berkeley. And basically, he was saying, like, Tesla has become kind of this meme stock. It has become this thing that is sort of disconnected.

from the actual market realities of their business in electric vehicles. It has essentially become kind of a vibes-based entity on the stock market that is related to people's feelings about Elon Musk and their future and AI and all this stuff.

but that they actually have not had a lot of discipline in the actual business. And this was a chance for shareholders to say to Elon Musk and to the board of Tesla, like, no, you actually do need to manage this company with more discipline and more rigor. And this was sort of one chance that they had to kind of impose some of that

management discipline on Tesla, which now, you know, now if you're Elon Musk, you think, okay, I got my pay package and I'm going to keep doing what I do. Yeah, and what are the odds that he goes and now spends more time focused on Tesla? No, he's going to say, I got my bag. Let's go see what's going on over at Grok. Yeah.

Right. All right. Next, Elon Musk story from this week. This one is a very different kind of story. This came out from the Wall Street Journal this week. On Tuesday night, the journal reported in a story that Elon Musk had had, quote, boundary blurring relationships with women at SpaceX.

The article, the journal says, is based on conversations with more than four dozen people, including former employees, people familiar with Musk's interactions with female subordinates, and friends and family of the women. Tell me about these women who talked. Well,

Well, first I should say, do you know when you sometimes you read an article and you can just like feel the lawyers behind it? Yes. Like this was one of these articles where I'm just like, okay, a committee of lawyers has been, you know, wrangling this article into a place where it could be published for weeks because obviously these are very, you know, sensitive allegations. And Elon Musk is, we know, is very litigious about

And so it's just a very careful article. It's not a bombastic, you know, it's not making sort of claims that it can't back up. And I would say it is diligently sourced. Yes. And we will try to be equally careful, I think, right now in how we talk about this article. The main thing I would tell people is go read this article.

And I want to say congratulations to the reporters at The Wall Street Journal who got this one over the finish line. Joe Palazzolo and Khadija Safdar. I don't know either of these people, but I became a huge fan reading this story. Yeah, so the sort of broad summary of this story is that Elon Musk, according to the people that the journal spoke to, has a pattern and a history of having these boundary-blurring supervisor-employee sexual relationships with female subordinates.

And they focus on four women in particular. Please tell me about them. So the first woman is a woman who the journal does not name, but that they describe as a former SpaceX intern who met Elon Musk in the early 2010s when she was doing a summer internship at SpaceX.

The journal reports that they bonded over Star Wars and eventually kissed and a year later took a trip to Sicily together. The journal also got its hands on a bunch of text messages between Elon Musk and this woman that she apparently showed to a friend and reports that they had this kind of romantic entanglement and ultimately that she was hired full-time at SpaceX to

in a role that some people at SpaceX thought was maybe too senior for someone with so little experience. She was hired as a member of his executive staff, and this struck people at the time as being strange. Now, we should say that the journal also reports that this woman has denied having a romantic relationship with Elon Musk while she was employed full-time at SpaceX, although she did sort of cop to having had a romantic relationship with him before that.

And there's some other weird stuff involving lawyers. She's now got the same lawyers as Elon Musk, and they sent the journal some signed affidavits basically disputing some details of the journal's reporting, saying that while she was employed at SpaceX, nothing Elon Musk did toward her was, quote, predatory or wrongful in any way.

I mean, look, setting the particulars of this woman aside, I would say that were I in a similar situation, I would feel a lot of incentive to deny a lot of the details in a story like this, right? It would be personally embarrassing to me. So the mere fact that some of this stuff is being denied doesn't necessarily convince me that what the journal published is not true. Yeah, and I think there's also just a lot of

contemporaneous evidence that the journal cites to support the fact that there was something, you know, more than a normal employee-employer relationship going on here. Apparently, Elon Musk texted her nighttime invitations, including, "'Are you coming over? If not, I will probably trank out. Too stressed to sleep naturally.'"

And they said that around this time, she and Elon Musk would hang out his house watching anime and talking about the technical future of humanity. She also eventually— You've just described the worst evening I can imagine having, by the way. But go on. And then eventually she moved off this team, started reporting to another engineer, and eventually left SpaceX.

So that's the first woman. There are three other women included or referenced in this article by the Wall Street Journal. One of them is a former SpaceX employee who left the company in 2013 and alleged that Elon Musk had asked her to have his babies on multiple occasions.

This woman, which you think you have a demanding boss. Imagine being this woman. The journal says that she declined to have his children, but that their relationship deteriorated after that and that Musk denied her a raise and complained about her performance. She did receive an exit package of money.

Cash and stock valued at more than a million dollars, according to a person familiar with the agreement. Imagine being this person, your friends are like, how's your work-life balance lately? Like, well, not great. My boss keeps trying to get me to have his children. Ha ha ha ha.

Well, we know that this is a thing that Elon Musk, you know, does. He's very concerned about babies. He's very concerned about sort of the decline in fertility and the birth rate. He wants to sort of populate the world. And we know that he has had children with women who worked for and with him before. Some people

have taken him up on this offer. Yes. Yeah. So that's woman number two. There's also the Journal reports an employee who had a month-long sexual relationship with him in 2014. This one is a little complicated. Yeah, you're going to want to go to your whiteboard and sort of follow along at home as Kevin describes this because this one has some twists and turns. Yeah, this one is giving telenovela.

So this is a woman who worked for SpaceX, who apparently reported directly to Elon Musk and

and had a very complicated relationship with not only Elon Musk, but with Gwynne Shotwell, SpaceX's president. According to the journal, apparently Gwynne Shotwell thought that this employee was having an affair with her husband and basically allegedly retaliated against her, according to some emails and accounts that she made to friends and family at the time. And then

After that, in the late fall of 2014, the journal reports that she started a sexual relationship with Elon Musk that included drinking and hanging out at his mansion in Bel Air. The journal reports that they had sex and that this all became very complicated for her. And then the fourth woman—

It's true. And then the fourth allegation in here is one that we have actually heard before. This is a SpaceX flight attendant who alleged that in 2016, Elon Musk exposed himself to her and offered to buy her a horse in exchange for sex acts.

This one was reported back in 2022. This woman's shifts were apparently cut back after she rejected Elon Musk's advances and the company eventually agreed to pay her $250,000. Musk has called this flight attendant's allegations utterly untrue. So that

is the article. It's very juicy. I'll say it. It's gossipy. It's very carefully sourced. It's very diligent. And I, yeah, I learned a lot. Well, I have some thoughts to share, but first I wonder if you could read us this statement that Gwynne Shotwell, the SpaceX president and COO, said about what the journal reported, Kevin. So Gwynne Shotwell, president of SpaceX, said,

told the journal, quote, "'The untruths, mischaracterizations, and revisionist history in your email paint a completely misleading narrative,' she said."

Right. So when she says all the forces acting against us, do you think she's referring to gravity? Because I have to imagine that if you're working at SpaceX, that's the force that just pisses you off every single day. Yeah, that's the one you can't lawyer up and defeat. Yeah.

Gravity is undefeated. So, okay, there are these- You know what this story's about, Kevin? What? This story's about a pattern of behavior. Yes. You know, there's obviously going to be a lot of, you know, going back and forth about this detail and that detail and how romantic was this relationship and, you know, did this exact thing happen? But you, this is the value of this story is you pull back and you say, wow, this is

For a decade now, the same thing keeps happening over and over, which is that people across this company, including direct subordinates and even an intern, are being put in positions where they're now having to say, do I want to have sex with this person in order potentially to preserve my status at this company? Or do I want to say no and potentially risk the wrath of one of the most powerful people in the world? So that is just something that is very much worth pointing out, I think.

Yeah, and the article says that there actually were some SpaceX executives and former employees who were so concerned about a culture of sexism and harassment at the company that in 2022, they went to the National Labor Relations Board with a complaint about a high-level group around Elon Musk who they say were failing to apply the company's own rules to him, which they say also created an environment where he could act with impunity.

So this is just, as we said with the pay package dispute, this is just a man who has not had much accountability at the companies that he runs. He has been able to kind of run them as he sees fit. And maybe, I don't know, maybe this is the start of him having a little more oversight. The, the,

The absolute moral rot at the center of these companies that this suggests is just truly appalling to me. And, you know, look, I know that in 2013, a lot of people still had their eyes on the stars and they said, isn't it so inspiring what this man is doing with renewable energy and getting us into space? And I understand why that was inspiring for a lot of people.

But man is the truth starting to catch up with this person. We now know what sort of culture has enabled him to keep at this for so long. And I really hope that more attention is being paid to what is actually like to work inside these companies. Because my gosh, if I were one of these women working inside one of these companies, this would be an absolute nightmare. Yeah.

Okay, so that is Elon Musk's alleged relationships with female subordinates. We now have the third story of the week, which is about changes that are coming to X, the social media platform formerly known as Twitter. And there are two...

that I'm not sure if they're technically related, but they sure seem, you know, to be spiritually related to each other. The first is that X has loosened its adult content policy. Basically, porn has been available on Twitter for years. The company has known that this goes on. It's not really...

taken many steps to rid not-safe-for-work content from the platform. But under Elon Musk, it has now been officially permitted. Their new policy says, quote, we believe in the autonomy of adults to engage with and create content that reflects their own beliefs, desires, and experiences, including those related to sexuality. The second change is that Elon Musk and X have now made laws

likes private by default. Now, what does this mean? Normally, when you like something on X or if you like something on Twitter, you could then go onto someone's profile and I could see here are all the posts that Casey has liked. Which has led to some great little mini scandals over the years as various politicians have been found to accidentally like something a little too spicy. Yes, there was a famous incident or infamous incident years ago where Senator Ted Cruz, um,

One of his Twitter likes was a porn video and people had a great time with that. So X will now hide likes. According to X's director of engineering, this change is meant to protect users because many people feel discouraged to like edgy content. So basically Elon Musk didn't want likes to be public to prevent this kind of embarrassing thing from happening. And now you're not gonna be able to see what people like. Why do you think he's doing this?

Well, I think that it seems like that they have realized something that is true, which is that porn consumption is a primary use case of X for many, many people. And I think that this particular use of X, formerly Twitter, accelerated after Tumblr banned porn many years ago. Right. Tumblr was sort of previously in this role of being a place where a lot of people would go to see porn.

Tumblr got rid of it. And then all of that sort of moved to X. And now, you know, X is really kind of at the center of the adult content ecosystem. You know, OnlyFans is basically a way that adult performers who are on X and grow their audiences there by sharing short clips are then able to go and monetize it. X has considered launching an OnlyFans competitor of its own over the years. But,

You know, I have to say, like, my feelings about this are very mixed because on the one hand, I do think that we should have an internet for adults where, like, adults can engage in sexual behavior and express that part of themselves and maybe even build a business around it. But I think X has been a really problematic actor in this space. Yeah. I'm not upset about the hiding the likes from public view. You know, a lot of other social media platforms don't, you know, you can't go on someone's Instagram profile and see all the posts that they've liked.

You used to be able to see when they liked other people's posts. I remember there was that whole feed. Yes. It was very controversial. Yes, it was very controversial. And look, I'm sympathetic to the idea that we should just have the ability to use our social media accounts to like whatever we want, and then that shouldn't be sort of available to the public.

I think a lot of people also didn't know that their likes were public, and so they would like something, thinking, you know, this is just for me to bookmark, maybe come back to later. And then all of a sudden, people are saying, hey, did you like that spicy tweet? Yeah, look, this is a privacy feature, and privacy features are good. Like, we should celebrate when platforms, however bad they might be, launch privacy features. Yes.

Okay, so the last story is that Elon Musk has dropped the lawsuit that he filed against OpenAI. You remember we talked about this on the show when it was filed. Elon Musk had accused OpenAI of sort of breaching its founding agreement by converting from a nonprofit research lab focused on creating AI for the good of humanity to a

basically for-profit company that has deals with Microsoft and all these reasons that he was upset with OpenAI. He has now dropped that lawsuit. What do you think of that? Well, this was the sort of famous case of Elon suing OpenAI for breach of contract and then not being able to identify any contract, right?

Right. Because whatever agreement he was talking about, like for all legal purposes, was intentionally scribbled on the back of a cocktail napkin and didn't even appear to really be in agreement about anything. So, you know, this is not surprising to me and is another case of Elon Musk sort of like, you know, legal saber rattling.

winding up being nothing. Yeah, I mean, every lawyer I talked to about this case, everyone who has sort of had familiarity with the case thought this was a total frivolous case that was going to go nowhere and was going to get thrown out by court. So I guess it is not surprising in that sense that Elon Musk has decided that he would rather not go through with it.

But I think his beef with OpenAI is sort of well-established at this point. It'll just be interesting to see where that goes from here. Yeah, also, like, I'm sure he's going to threaten a lot more lawsuits in the next year. And as, like, reporters in particular think about how they want to approach coverage of those, I think they just need to remember that sometimes he will threaten and sometimes he will even file these lawsuits. But then he'll walk away once, you know, the obvious becomes more apparent. Yeah, until he becomes the secretary of the Treasury in a second Trump administration and decides to shut down OpenAI. Yeah.

When we come back, I want it Hatway. It's time to play Hat GPT. This podcast is supported by KPMG. Your task as a visionary leader is simple. Harness the power of AI. Shape the future of business. Oh, and do it before anyone else does without leaving people behind or running into unforeseen risks.

Simple, right? KPMG's got you. Helping you lead a people-powered transformation that accelerates AI's value with confidence. How's that for a vision? Learn more at www.kpmg.us.ai. Casey, it's time for another game of Hat Cheap E.T. All right, Kevin, let me pass the hats. ♪

Hat GPT is, of course, the segment where a variety of news stories from the world of tech and the future are placed into a hat. And one of us will take a headline out of the hat and make the other person discuss it until we get bored. And at that point, we will say, stop generating. All right. So let's use our bucket hat today. Yep.

All right, you go first. All right, let's see, Kevin. The first story, Eric Schmidt is secretly testing AI military drones in a wealthy Silicon Valley suburb. This is from Forbes. Last year, Eric Schmidt, who of course you'll remember was the former CEO of Google, quietly founded a secretive military drone company called White Stork.

And apparently it has recently been spotted testing this drone in and around Menlo Park. And so, Kevin, how would you feel if you had bought a house in a wealthy suburb and then one of Eric Schmidt's murder drones started hovering nearby? I don't think I would feel great. And actually, I saw this story and I just thought to myself,

Thank God the effective altruists got into AI because this is my hottest take. If these people, these effective altruists, had not gotten interested in AI a decade ago, the people who would be at the forefront of the AI industry now would be these like weird military...

building drones to like use in wars and like testing them in suburbs. So like, I feel good on one hand that I've never heard of White Stork and that it does not appear to be like a major player in defense AI. But I'm also terrified because like this is not

good. This is not good. You know, Kevin, when I was growing up, a white stork was a symbol of someone bringing a baby to your front doorstep, and now it will probably be the last thing I see before I die. All right! Stop generating! All right.

Here's another one from Forbes. Why perplexity, cynical theft represents everything that could go wrong with AI. This story is from Forbes, and Forbes is taking perplexity, the AI-powered search engine whose CEO we've had on the show, to task over what it described as stealing its articles, using AI to summarize and repackage them, and doing so without prominent attribution of links. This week, Randall Lane, Forbes' chief content officer, wrote that perplexity repackaged one of its recent scoops,

published its own AI-generated article that summarized it, including several sentence fragments that appear to have been lifted word for word, sent a push notification to users pointing them to its summary, created an AI-generated podcast out of the summary, and made a YouTube video that now outranks Forbes' original article in search. In response, Aravind Srinivas, CEO of Perplexity, a former Hard Fork guest, said that Perplexity was the number two source of referral traffic for Forbes, acclaimed Forbes disputes. Kevin, what do you think?

Well, I mean, this is something that you were saying. You called this one because back when we had Aravind on the podcast, you said, look, this is what's going to happen. This company is going to hoover up the entire internet. It is going to summarize that. It is going to put these little tiny links and attributions in its summaries. No one is ever going to click on them and they will basically be repackaging, paraphrasing, regurgitating, whatever you want to say. They will be sort of extruding this sort of journalism product

and sort of claiming it as their own, and people will engage with that and not the original articles that people had to work so hard to produce. Yeah, and I'm really sad that I was right about this one because my hope would be that Arvind went back and tried to build a sort of better and more ethical framework for doing this sort of thing, but they didn't. And look, I'm really worried. Like, this stuff is a moral catastrophe, and the tech industry needs to do a lot better.

Maybe legally they'll find some way to get away with it, but it is absolutely wrong and it is self-defeating in the long run. If this truly becomes the way that most people access their news, this journalism is going to disappear. And perplexity is not going to have push notifications to send to anybody about anything because nobody's doing the dang work anymore. So what a disaster. Yeah. And I should say perplexity has said they just told –

another publication that they have been working on revenue sharing deals with high quality publishers. Basically, they're saying, you know, we're working on this. We're trying to make these citations more prominent, but I do not think publishers are going to be happy that there are now AI companies out there scraping their hard earned stories, repackaging them as news articles and podcasts and YouTube videos and boosting themselves above the originals in Google search. So, yeah,

Bad news for journalism. Bad news for perplexity. Stop generating. Okay. This one is from the New York Times. It's called a four-hour-long hotel review that is actually about so much more.

This is a story about a viral YouTube video called The Spectacular Failure of the Star Wars Hotel. Apparently in this video, YouTuber Jenny Nicholson talks in a four-hour video essay about why Disney's Star Wars Galactic Star Cruiser, a spaceship-themed hotel which opened in 2022 only to close a year later, was such a disappointment.

I have not watched this YouTube video. Did you watch this YouTube video? I did. And let me say, masterpiece. You watched a four-hour long YouTube video? I watched a four-hour video and I would have watched a fifth hour because that's how good Jenny Nicholson is as she dismantles this experience where she and one other person went and for $6,000 they spent two days inside this hotel and had a miserable experience. But it wasn't...

Maybe for the reasons that you would think, like it wasn't that, you know, the food was terrible and that, you know, the staff was mean. It was that Disney had come up with such a cynical product that was basically based on the idea that people love star Wars so much that they will accept absolutely anything and pay $6,000 for the privilege. And she breaks down why that's absolutely not the case. This is honestly just an amazing, uh,

commentary on this cultural moment where these mega corporations try to find these fans and then just serve them absolute slop and tell them to love it when in reality it actually sucks. Okay, so if I don't want to spend four hours watching this YouTube video, but I want to see what it says, is there a tool that I can use to have AI summarize it for me? Yeah, I recommend Perplexity AI. It's going to have a really nice summary for you, Kevin. Okay, stop generating. Go watch the video. All right, pass the hat. All right.

Well, Kevin, how excited are you for

this one. I actually am excited because I have a little bit of like an Instagram use problem and I'm excited for a feature that repels me every time I go on Instagram. So you think this solves the problem that you see the ad break and you think, oh, the heck with this. Yeah, and then I get time off my screen. I get to reconnect with my family. I think this could be a really good thing for my life. No, this is obviously annoying and bad. And it reminds me, do you remember a year or two when this viral patent made the news? Do you remember this? This was a Sony had patented a device and

for interactive advertising where basically you would get shown a commercial for something and in order to make the commercial stop, you would have to physically get up from your couch and yell something at your TV. So I think that the one in the patent was just yelling McDonald's and then the ad stops. And at the time, people thought this was like dystopian and weird. And I got to say, that's where we're going. It is basically here. Well, God help us all. All right, stop generating.

Okay, this one is from the New York Times. It's called Every Elephant Has Its Own Name, Study Suggests. Scientists say they have found evidence with the help of AI that elephants have names and that they call each other by their names. This was published Monday in the journal Nature Ecology and Evolution. According to the researchers, they used an AI tool to analyze the rumbles that certain elephants produce

And they found that individual elephants seem to respond to certain rumbles more than others. So far, the scientists aren't sure precisely which part of the rumble might be the elephant's quote unquote name, but they did find that their AI tools ability to identify basically which elephant the rumble was intended for far exceeded what random chance would seem.

Dictate. Well, actually, there I can help out the scientists here, Kevin, because my understanding is that your elephant name is the street where you grow up on and then the name of your first pet. And that's your elephant name. No, this is very exciting. We're using AI to sort of understand how animals communicate. What did you think? Yeah, I like this. I like the idea that I could go on safari and an elephant could come charging toward me and I could just be like, hey, Bob.

Cool it! You know, I'm actually working on a beautiful story about you on a safari, and it's called Call Me By Your Elephant Name. It reminds me of my favorite joke. What's that? Why do elephants paint their toenails red? Why is that? So they can hide in cherry trees. Why do elephants paint their toenails red so they can hide in cherry trees? Yeah. Have you ever seen an elephant in a cherry tree? No. Then I guess it works. We'll be right back. All right, pass the hat. All right. All right.

Microsoft recalls screenshot feature after outcry. So some listeners may remember when Microsoft launched this feature, which takes constant screenshots of every single thing that you're doing on your PC so that you can quickly search for them later. We said, this sounds like one of the greatest and most intrusive privacy violations of all time. Well, Microsoft is now making changes. And the most important thing to know is that this will now be opt

in. So if you would rather not have a record of every single horrible thing you've ever typed in your entire life, you can rest easy. What did you make of this, Kevin? Well, my first lesson is that you should never call a product something that will be extremely funny if it turns out to blow up and you have to dial it back.

A product called Recall being recalled? I just don't know how they didn't see this one coming. You know, I will say when we talked about this, I heard from somebody at Microsoft who basically just accused me of, like, you know, wetting the bed over something, you know, very inconsequential. Oh, I think I know that person at Microsoft. Yeah, Tom.

I was like, I don't know. It seems like this really sucks. And over the next couple of weeks, as security researchers looked into this, they said, Microsoft, what the heck are you doing? So I feel very validated about this one. All right, stop generating. All right. Last one. Yeah.

This is from The Verge. Spotify is increasing U.S. prices again. A year after Spotify first increased the price of Spotify Premium in the U.S., they are doing it again. Next month, subscribers to Spotify will start receiving an email detailing the price increases to their plans. One Spotify Premium package will now cost $11.99 a month, up from the $10.99 increase announced last year. I...

saw this one a couple days ago because I got an email from Spotify saying that my subscription plan, which is a premium family plan, was going to increase from $16.99 a month to $19.99 a month. Casey, what the hell is going on with all these subscription increases? Well, there was actually some interesting reporting recently that Spotify has the lowest subscription

churn of any major subscription media service. So what that means is once people start using Spotify, they're just sort of very unlikely to bail. And if you're a Spotify, that's just an opportunity because you know, you can just keep jacking that price up little by little every year. And most people aren't going to run fleeing for another service. So kind of a bummer. But I got to say, Kevin, as somebody grew up in the era where listening to one CD would cost me $18, the fact

that the price of a basic Spotify premium plan has gone up, what, now $2 since launch in the United States does not actually make me that concerned. Yeah, I mean, I can sort of justify it still. I'm not going to cancel my subscription. I use Spotify all the time. I like the family plan. You know, I'm not one of these churn risks, but man, you start adding up increases. I mean, YouTube premium increased its price this year. Netflix has been steadily increasing prices like

All these subscriptions are just getting more and more expensive, and I got to think there's some breaking point where people are going to say, enough, I've outstripped my budget here. Yeah, and if you're planning to start a revolution based on media prices, let us know. We'd love to talk to you. Yeah, is Platformer planning any price increases? You know, I believe that we're actually the last media property that has not increased its prices since launch. We're four years old. We're holding steady at $100 a year. You're like the Costco hot dog. You'll never change your price, no matter how high inflation gets. We're the Costco hot dog of the year.

media okay stop generating that's chat gbt that's at gbt that's at gbt chat gbt something else you can now find it on siri this podcast is supported by kpmg your task as a visionary leader is simple harness the power of ai shape the future of business oh and do it before anyone else does without leaving people behind or running into unforeseen risks

Simple, right? KPMG's got you. Helping you lead a people-powered transformation that accelerates AI's value with confidence. How's that for a vision? Learn more at www.kpmg.us.ai.

Hard Fork is produced by Rachel Cohn and Whitney Jones. We're edited by Jen Poyant. Today's show was engineered by Corey Schreppel. Original music by Alicia Baitube, Marion Lozano, Rowan Nemisto, and Dan Powell. Our audience editor is Nell Gologly. Video production by Ryan Manning and Dylan Bergeson. If you haven't already checked out our YouTube channel, you can find it at youtube.com slash hardfork.

Special thanks to Paula Schumann, Quee Wing Tam, Kate Lepresti, and Jeffrey Miranda. You can email us, as always, at hardforkatnytimes.com. Yeah, tell us your elephant name.