Sharp Tech with Ben Thompson - Voice and Video Demos with ChatGPT, How AI Could Redeem Meta’s Mixed Reality Bets, OpenAI Explores Hardware

Episode Date: September 29, 2023

Ben shares his experience with a ChatGPT demo that inverted his expectations, a reminder that reducing friction is essential for AI’s next steps, and reactions to Meta’s Smart Glasses and reports of Jonny Ive and Sam Altman brainstorming. At the end: Delta follow-up, Disney capacity, and the Swift-Kelce bundle.

Hello and welcome back to another episode of Sharp Tech. I'm Andrew Sharp and on the other line, Ben Thompson. Ben, sincerely tonight, physically, spiritually, emotionally. How you doing? Not doing well, Andrew. I'm going to be totally honest with you. I'm kind of delirious. This is going to be an absolute adventure. But we will do our best. I'm excited because I'm delirious also. I just finished a two-hour podcast, greatest of all talk. What a week it's been, you know? You took a sick day on Tuesday, like five massive text stories hit. And then just as the dust was settling, a story that's a hundred times more important hits and forces us to actually delay recording for a day. Damian Lillard is a member of the Milwaukee Bucks. Wait, let's be clear.

I did not delay the podcast. You delayed the podcast to do an emergency episode of your basketball podcast, which is fine. I approve of that. But the whole problem is that I took a sick day, which I hate doing. And don't send me emails about how I'm too hard of myself. That's part of the secret sauce. I also need to note that you took a sick day only a couple weeks after you proudly announced that everyone else in your house is sick, but you never get sick. So I think you might have done that to yourself this week, but it is what it is.

It's very possible. So I said, I'll write on Thursday instead, which is usually always a mistake. If I write out of day, I don't usually write. And so I'm on Thursday and then the Damien's Lord story breaks. Obviously it's fantastic. I appeared on the Bill Simmons podcast, which weirdly has not dropped. I think more Boston Celtics cope. But I'm working on this article.

I'm working on this article and it's not coming together. I scrap the whole thing at like 11 p.m. I'm like, I need to start over. This isn't good. I should have probably just bailed on it. But I'm like, you know what? I feel irresponsible if I not just take a sick day, but also on the same day that Damien Lurter gets traded in Milwaukee Bucke,

people are going to think that I am issuing my responsibilities because I am too happy about this outcome. So I had to power through and get it done. I have no idea if it's good. I have no idea if this is going to be good. But we are here and we are ready to podcast. There you go. Tech comes first. You know, basketball is fun.

But the main thing is the main thing here at Stratory. And yes, you can listen to Gras of All Talk by going to the show notes. You can also listen to Ben on Bill Simmons by going to the show notes. We received a bunch of requests. to talk about the Lillard trade on this show. We're not going to do that, Ben. We'll start with what you wrote on Stratecary Thursday, and we'll see whether any of this makes sense tonight.

Meta is releasing its own chatbot called Meta AI and a fleet of celebrity-based AI characters with a promise of more to come and a developer platform to boot. I haven't used any of these products, you write, which are, for now, limited to text interactions. What the releases point to, though, is the removal of another kind of human constraint.

In the very near future, literally billions of people can talk to Tom Brady or Snoop Dog all at the same time. So can you explain what's happening there? The Tom Brady or Snoop Dog detail kind of threw me for a loop. But what should we know about the text-based chatbot before we move to the AI-powered glasses that meta-announced this week? So if you go to chat GPT and talk to chat GPT, you're talking to chat GPT and there's like the chat

GPT voice, right? That's sort of sterile and very, very sort of clinical in a certain way. And you can ask it to adopt a sort of a different tone, but that's that's not the default. And you have to like sort of put it in the prompt. And I think that the approach of meta is taking here is these chatbots can be sort of anything you want them to be. And they can have personality. And also chatbots, don't need to just be tools. I, you know, to date, chat botts, you know, chat GPT, everyone talked about it. We talked about in the context of is this a competitor for Google for search. We've talked about co-pilots and in Microsoft and the enterprise sort of context and helping

you sort of find information and figure things out. And I think meta and sort of rightly so sees this, look, these are fun, right? It's, it's just interesting to sort of talk to this sort of entity. and I think it's sort of a proof of concept. The idea is to get a bunch of celebrities. They're not actually called Tom Brady or Snoop Dog. They actually gave them different names. I think they have sort of like the plausibility of like this isn't actually them talking

in case it says something terrible. Right. Yeah. And I think it makes sense. I mean like and you know the it's fun. Why not do it? And you and it's in the context of these are going to be sort of like profiles on Facebook. So you can go to a profile.

you could follow someone and you could initiate a chat, I get a bunch of like messages randomly to like the trajectory Facebook account, which apologies, they're buried and like you're not following this person and I look at it like sort of once a year sort of thing. Not the best way to contact me to say the least, but it is sort of possible. Of course I don't,

not because I don't enjoy talking to people or subscribers. It's not scalable, right? Like I'm one person, you know, you know, tens of thousands of readers or whatever, might be, it's not, it's not sort of scalable. Right.

But you could imagine a world where it, it would be scalable. And in that world, if there was, you know, tens of thousands of me, well, sure, I could do that. But there's not. There's a human constraint. And the reason to talk about this human constraint is sort of broadly, it ties into a lot of things we've talked about. Why is it, you know, what's the actual locus of competition? The, you know, if you're watching a YouTube video, you're not watching a Netflix stream, which means you're not watching cable TV.

It's all sort of like one, the attention is sort of the limiting resource. And that's been the defining nature of the internet. The way to understand competition on the internet. And I know where to talk about Amazon today. We're saving it for next week. I know you're doing some studying so you can take me down. It'll be sort of great sort of battle. Spending my weekend with Lena.

It's going to be terrific. I mean, you know, you're going to have to do some good digging deeper than the case, I think, to find some good arguments. Oh, boy, here we go. Tell me about competition. It's a preview. It's a preview of next week. Competition on the Internet. The big thing about competition Internet is user attention.

The aggregation theory is about sort of collating demand. And that demand is for what you want to read, what you want to see, all those sorts of things. And the reason why the way to think about AI and why it is, I hesitate to use the word disruptive because disruptive has a very specific sort of business context that we sort of talked about. But the reason why it presents the opportunity for upheaval is because it eliminates the human constraint in all kinds of ways. In this specific case, it eliminates the human constraint as far as harvesting attention. and any one entity can serve tons and tons of people.

Now, I'm not saying that these chatbots are going to be super compelling and that you and I are going to be sitting around talking to our chatbot, you know, Tom Brady and pseudo chatbot, whatever it might be. No, of course not. This is V1. It's super early. It's not even sort of, I haven't even used it yet. It's not even broadly available.

But a lot of things in tech are about seeing where something is going. and realizing that the sort of implementation details and technical details are going to fill themselves out over time. And we do see things like Character AI or Replit, these startups that offer chatbots that by all accounts, they kind of are staying under the surface by and large, but by all accounts are actually doing bonkers business where people are actually sitting around and talking to chatbots.

And that chatbot is over time understands them and is tuned to them and is their sort of personal companion. and again, separate all this from the sort of dystopian sort of questions of is this sort of a good thing. It is a thing. And Facebook bringing it to bear because you already have a messaging app. You could imagine this talking to your Tom Brady. It's just another chat in your WhatsApp. Right.

It's not downloading a new app. It's not meta. That's right. That's right. And so I think it's compelling in the direction that. that it's pointing, even if the product itself maybe makes you want to sort of roll your eyes a little bit. So, yeah, I mean, along those lines, that's one of the things that I'm curious about. It just feels like there's an avalanche of chatbots right now. Like every other day, a company is

announcing a new chat bot. And I don't see a lot of evidence that people love using chatbots. And I also am a little unclear about how they're monetized. And it seems like the technology remains incredibly expensive, the technology that's undergirding all this. So I just wonder, like, is there a point where we hit like saturation or am I going to just continue reading about new chatbots like every 48 hours indefinitely? What is the population of the world? Eight billion? I think that's your addressable market for everybody's getting an individual chat bot, huh? Is that the plan? This is what happened with the internet the first time around. You start out sort of V1, the internet is just sort of publications online, right? So you can go to your local newspaper or you can go to a newspaper

anywhere in the world, which actually was tough for them because of the locals, they were now competing with everyone, but everyone got the same sort of experience. It was sort of like a broadcast experience and you went there and consumed it. What changed was the feed, the newsfeed, where you had a perfectly customized stream of content that pulled in content from everywhere. It could have an article from a newspaper here, could have pictures of your cousins over there, and it actually wasn't over there. It was. all in the same same stream. And it was infinite. It could sort of keep going. And it was new every time you went there. And now you had this fully customized feed that was very compelling

because it was customized to you. And now you can have customized advertising. And then that's where the business model sort of came along. And so we've already gone through this. We've already gone from consumption that is broad-based to consumption that is sort of individualized. Right. Now, all that stuff in that feed still had to be generated by a human. It still was sort of a discrete moment in time where someone took a picture and posted online. Someone wrote an article and sort of put it, put it online. And there's so many people in the world generating so much content that you can get a pretty customized sort of feed that is attuned to your interest. But what AI promises to do is actually

generating the content specifically for you. And this content isn't just push out an essay and you read it, right? Like you had like these horror stories of like these, you know, news sites that are like, you saw that an obituary of like a few weeks ago. Like I can just use the wrong words. And it was like really brutal. That is, but that's like V1 internet. That's just taking what you did before and doing it with sort of the new tool.

No, the missing piece in the internet currently is interactivity. Now, we can interact, right? We can be in a group chat and interact. But ultimately you are gated by the availability and. and presence of sort of other people being there. And, you know, an AI, though, is something that is there all the time. It's not gated by time. It's not gated by place.

And it's not gated by the interest of anyone else. It can be fully and completely tuned to you on an individual basis and be available 24-7. It's the total removal of constraints as far as the generation of content and interactivity and customization. A huge thing for Facebook has been, a thing they're focusing on is monetizing their messaging services. And the big thing here is not putting ads in messaging services, which is a terrible platform for ads if you sort of think about it, but it's these sort of quick to add sort

of concepts where companies take out ads or they have opportunities to put a button and you can start talk to someone. And that's a great sales channel. And there's a whole culture around that sort of here in general, like selling stuff online and stuff sold through line groups and all this sort of thing. it's very compelling and you get like sort of talk to someone it doesn't scale do you mean Taiwan yeah here in Taiwan is right right yeah it doesn't scale it's sort of that is trying to do the same thing in India they just announced recently actually with WhatsApp right they're working

with businesses in that way that's right and but what works in some countries like potentially India I think it's done very well in Thailand in like Vietnam you have a relatively cheap workforce so you can actually scale up the customer service agents to make those chat experiences compelling and you can complete a sale and you can have a sufficient margin to sort of make it worthwhile. People, that doesn't work in most other countries, it doesn't work in the West because wages are too high. You can't actually scale up the people. In this context, though, is paying for GPU time to have a chatbot expensive relative to general compute? Sure, it definitely is more expensive. Is it?

expensive relative to hiring a human. No, it's drastically less expensive and the scalability is much, much, much higher. So to the extent there is value and interest in actually interacting, having this 24-7 ability to interact with an entity that is fully customized in tune to you and overtime web memory and be able to have ongoing conversations, the cost is actually very low if you put it in the context of what it's actually competing against. Which is other humans that earn 10 times as much in salary. Yeah, that makes sense. Yeah, and I don't ask the question to be kind of a snarky dick about it. It's just as I sort of take in all the AI hype, it seems like a lot of it is then actualized in the form of chatbots that humans may not love interacting with in every single sector of life. And part of that is why I thought what you wrote on Thursday was really interesting because, You were talking also about a meta announcement, about a new generation of their smart glasses, which is met as mixed reality sunglasses that they've been developing with Rayban.

And as Zuckerberg announced all this, he said, I think that smart glasses are going to be an important platform for the future, not only because they're the natural way to put holograms in the world so we can put digital objects in our physical space, but also, if you think about it, smart glasses are the ideal form. factor for you to let an AI assistant see what you're seeing and hear what you're hearing. And that is actually really interesting. And you wrote that these glasses could be the ideal form factor to integrate AI into our daily life. You also noted that it's possible that the full realization of AI's capabilities depends on removing human friction from the process. So can you elaborate on what you mean there? Yeah. So I did this. I wrote about

this also in the context of Open AI's announcements. And the two big announcements that they had, you know, number one, you can talk to chat GPT now. And it's kind of a weird experience when you're in the app. You have to click this headphones icon. And then you're just sort of on this black screen with like rippling bubbles or whatever that when it's like thinking or whatever. Wow. But then you talk and it talks back. And when you're actually conducting the conversation, you're not like pressing any buttons. It's just sort of happening. And then number two is this vision capability, which you've probably seen tweets online and people showing it like you take a picture of a math pages does all the problems or whatever it might be.

So I talked to opening I last week, sort of like a briefing about this and kind of just brushed over the talking. And then the vision demo was obviously super compelling and very interesting. And like you see the demo like, yeah, I can see how this is a big deal, right? Makes a lot of sense. Then I actually got access to the products. And it was the exact opposite. The talking was so compelling.

And it was compelling because of the absence of what was. there because I wasn't typing, because I wasn't having to read an answer. And there is an aspect of obviously reading and writing is a foundation of sort of, you know, human experiences. Well, yeah, it's, you know, there's a lot of interesting research that has actually changed the way people think, right? I don't know if you've read that book weird, but like basically about how, you know, aspects of Western civilization are grounded on text and read. and the rise of logic and all these sorts of things. And it actually has changed the way the brain function sort of over time,

as opposed to the default human experience, which is more of a sort of oral experience and stuff isn't written down. And it's more sort of like tradition-based and myths and all those sorts of things, which sort of the default state. The talking shifts the experience from the weird mindset to sort of the default state. And it, it, it, it just, it's, it's, it's, it's transformative. It really, it really is. Again, I'm saying this, this is a very early beta.

There's still, it's not a great experience yet. But it's something where you can see this is a step change in the experience of bots. And in this step change, chat GPT's attempt to not offend and to, like, I want to have a talking conversation with Sydney, not with chat TPT, right? Yeah. But it's because it's compelling how much different it feels, right? And this is the aspect of sort of product design that is very ineffable and sort of hard to put your finger on. It's not just about feeds and speeds.

It's about how it feels. This is sort of like a Steve Jobs sort of axiom, right? It's the experience and what it makes you feel when you're sort of using a product. And that's why I'm skeptical of chatbots, just so you know, in terms of typing in questions and then getting back answers. There's the feel of it is more. frictiony and it's not necessarily something that everybody is going to want to do all day long for the rest of their lives. That's right. And particularly, why would you want to sit around and type and read a response from a Tom Brady? Right. Like it's, it's all pretty cumbersome. That's right.

And that is why that's what it was compelling to me is the cumbersomeness. When the conversiveness disappears, it's not just easier and more interesting. It changes what you want to talk about. Like it was much more interesting to talk to chat GPT. Like I'm like, well, I'm writing about, you know, the nature of virtual reality. What is virtual reality? It's like, well, it's about headsets, blah, blah, blah. But it could be things like,

oh, that's interesting. That's actually the same, you know, what I'm sort of thinking about. Like, why is an audio experience not virtual reality? Oh, yeah, it could be virtual reality if you think about it because you're doing blah, blah, blah. It's like sort of go this sort of back and forth. And a sort of thing I would not do in the text chat GPT at all because it's like a drag.

It's like I'm typing, yeah, typing all this sort of stuff out, right? The longest conversations I've had has been with Citi. just because it was such a surprise all along and sort of like what's coming back. I have not had hour long conversations with Chap, GPT. I could envision having an hour long conversation with this, though, because it's easy. You're just talking, right? I could be doing something else and be sort of having a conversation. Again, once everyone gets access to it, you're going to see it's not there yet.

There's still things to do to get there. In what respects is it not there? Are there certain things that it just doesn't understand very well? Like, do you have to repeat yourself? Just as someone who's a total foreigner to the demo here, explain what it feels like to use it. Well, you have to open up your phone. They have to unlock it. They have to open up the chat GPT app.

Then you have to invoke the mode. They have to wait for it to connect. And then you have to talk and you have to, you know, not pause two months. It might kick over, over, and then it would come back. And yeah, sometimes it misses a couple words. And then it's a little too verbose. And then also it has that sort of like clinical feel. And if you try to, like, if you try to, like, direct it in a certain direction, sometimes it sort of clams up.

It's like, oh, sorry, I didn't mean to get that one wrong. It's like, no, I wasn't, I wasn't offensive. It's all good. We're all friends here. Yeah, no, exactly. So, but the fact you kind of feel a little bit of emotional twins, sorry, I didn't make you feel bad. It's like you're talking to a computer. Like, chill out, dude, right?

And so, so there's just a lot of friction in the process. There's only been one piece of friction removed, which is the typing and reading. It's gone to death. All that other friction remains. And so part of my point in. this article was to talk about it makes a massive difference to remove that bit of friction. And just the feel and experience of it is substantial and makes me optimistic in sort of an analyst sense about like what Met is doing with all these multiple chat bots, the idea of everyone

individual chatbots. Because once those become talking back and forth, then it's suddenly a lot more interesting. But it also highlights to me how essential the removal of any sort of friction is to this experience. You are at the end of the day talking to a computer. And there is a necessary illusion that needs to be maintained. And my bit was like, what is virtual reality? To me, it's the complete untethering from reality such that you can fool yourself into believing the new reality you're in is sort of real. And the only way to do that is to relentlessly remove friction. And the challenge for meta and the challenge for open AI is that as

As long as they don't own a hardware platform, they are going to be fundamentally constrained in removing friction. By who, though? Aren't they just constrained by like physics also? No, by Apple and Google. Like Apple, if chat GPT were Siri right now and I can say to my phone the magic word to wake it up and start talking. So Apple can sort of gate the applications. Yeah, that makes sense. Okay.

You know, Google is certainly going to do the same thing. particularly they see this as a threat, they're going to be just as locked down as sort of Apple is. I mean, I think the interesting for opening eye, I should have mentioned this in the article. They should be pursuing a partnership with Samsung. Like, Samsung tried to develop their own chatbot, Bixby or something like that.

Like, they should be going to Android OEMs that don't necessarily want to be under Google's thub and trying to deliver this sort of experience sort of in their phones. It's super compelling. But again, just the, I think your willingness to dip in and out, because there is a certain suspension of disbelief that happens when you go, like, that's what's the obstacle to a virtual reality experience? It's actually getting out the headset and putting on your head and then diving into it.

And I think one of the key things to, to virtual reality, like I talk about meetings are super compiling for virtual reality. You go into a meeting and if everyone is in virtual reality, it feels like you're there. It's hard to explain, but it's actually true. And you remember you have a memory of that meeting. That memory is in the virtual space. It's not in your sort of locale. But as long as you have to go and get a headset and go and put it on, that is going to stop you from getting into that experience. Once you have a headset that you're in all day, for example, because maybe it's just a much better experience to work in, like I think like the Apple Vision potentially is, then going into a meeting is clicking a button or walking in a door, like a virtual door and you're in the meeting.

And once that friction is removed, you're easily transported. into an experience that when you're in it is super compelling and when you're not in it, it's easy to, it's difficult to muster up the motivation to get into it. Right. Yeah. Well, and the meta smart classes were compelling to me for two reasons. Number one, I feel like while we're talking about cumbersomeness, which is a word that we've invented for the purposes of this podcast, the goggle experience is going to be

fairly immersive but also fairly clunky for the first. foreseeable future. But if you're telling me that you just have to put on a pair of glasses and then that will have a personal assistant right there, it removes like a ton of that human friction that you're describing. And the other reason meta smart glasses are compelling to me is that this technology is just frankly much further along than I would have guessed, given what I've read about it over the last couple of years. And it seems like Zuckerberg is confident that they're going to be able to introduce software updates that integrate some of the technology that Open AI was announcing this week. And it may be worse than Open AI.

But if you can do it with this like Rayban product, that gets really interesting. And then like where will we be in four years and five years? And as this evolves, it feels like it's accelerating faster than I would have guessed. That's exactly right. That's why I wanted to talk about these two announcements together. that vision bit in use was much less compelling to me than watching someone else do a demo. Again, I'd open up the app, then I had to click on the camera icon, then I'd take a picture of something that had to type in my question.

I couldn't get it to work with the sort of voice sort of bit, right? And so I'm looking around trying to figure out what's something I can test this on, right? And I can imagine being out about and think, oh, I should use opening eyes vision thing. That would actually help me right now. Or I'm helping myself with my homework and definitely see that being a situation. where it's useful, but that friction is a, it's a big thing. We, you know, we've talked about it a couple weeks ago. How can you break out of the, this is good for homework sort of area?

To me, and this is kind of a story of tech broadly. So I'm not crossing a huge bridge here. Removing friction relentlessly is how sort of tech products sort of end up winning, winning in the market. And so I just thought the inverse of when I watched the demo on a Zoom call, vision super compelling talk. Oh, that makes sense. That's interesting. Yeah, you can already use, you know, just talk to a little bit. I get it. Experiencing it, talk. Wow. This is really cool. It changes the way I feel as I'm using it. Vision. Oh, what should I do this for? Right. It's like, I'm like back in the initial chat GPT, what can I type in the prompt to impress myself, right? And so you do that. My home office isn't that interesting. That's right. And so you, and then in conjunction with meta and these glasses to your point, like, you know, Snap was the first with the glasses.

sort of years ago. And, you know, I don't want to mention how much I paid on eBay to get a copy of the first pair. It was sort of embarrassing, particularly when the price fell through the floor later on. But, you know, but it was just like taking video, taking pictures, right? And it was very interesting, Meta's presentation, because they lead with all this AI stuff. And the use case is so obvious. It's so clear. This idea, you look at something, you say, oh, what is that? Or like, what is that like translation, right? I imagine that being in Taiwan. What does that say? And it's just saying it in your ear and it's right there. You're not pulling out your phone. You're not trying to point at something. And, you know, like, that's a high cost. It's a cost that you don't think about, but it's, you're less likely to do something. What it's just there and you just have to say something, it's much more plausible.

And so that's super compelling. And then they, you know, and then Mark Zuckerberg's like, yeah, AI is going to make these actually super useful. And then, and then, which I want to return to it a second. And that's like, but before we realize to say anything's useful, we did this live streaming thing. It's like you can live stream from your glasses now. And they had, you know, Charles Leclair, you know, sort of driving go carts around wearing

these glasses, which was cool as an F1 fan. I appreciated it. I'm also not Charles O'Clair. I'm not a live streamer. And well, I love gocarts. Wouldn't be that interesting? Well, I mean, I mentioned by, is that the good thing? Yeah, I mentioned by high speed predilection previously.

So it's probably not the best idea for me, for me sort of generally. But it's funny because that was, it felt tacked on. And the bit of our felt tacked on was what was clearly what these were built for from the beginning. And there's a bit here where Facebook like fell into a pile of gold and they had no idea. It's like they were just they were like we're going to do the metaverse. And it's Mark Zuckerberg talked about we got to have like holograms out there. When we say tacked on what's tacked on to be explicit about it is okay. So this is how it connects to Instagram and Facebook and our social.

networks that have like undergirded this trillion dollar business right well no i mean tacked on in the context of the when he talked about the smart glasses he spent the first five minutes talking about this AI use case right smart glasses and then he's like oh we we do have to connect it to our social networks but right well no but the point is that was the feature that these glasses were built for they've added on at the last second this AI stuff and it's so interesting because you know they you know you look at this whole metaverse sort of investment in building these headsets. And it's like it feels like you're too early. Like what's, you know, how many people are going to actually use this? It's going to be a niche. All these sorts of things. And oh, by the way, given the technological capabilities, it's only going to be

VR or maybe Mix Star in the meantime. You have a big headset. These glasses, okay, I get in the future. You want people to have glasses on. They walk down the street and there's little things in their vision. We're nowhere close to that technologically. This feels like a waste of time and money. And turns out, actually, it's the glasses that in my estimate, are way more interesting. They're interesting not because of the audio sort of communication possibilities. And they're interesting because this AI moment. And I just love that Mark Zuckerberg basically said that on stage.

He's like, yeah, it turns out this is actually a much more interesting use case. We didn't realize that to like the last six months, but here we are. Totally. If you think about it, smart glasses are the ideal form factor for you to let an AI assistant see what you're seeing. And he threw out a couple use cases, grilling with your family and you want to know how long to cook the chicken for. You're playing pickleball and the ball hits the line and you want to know whether that's a fault. Don't be that guy that's playing pickleball and cites the AI glasses.

Yeah, seriously. If you want to be like that, then go play tennis. We've already established that. Or say you want to settle a debate and you're out in the world. Just ask your meta glasses and get your response. That is interesting to me because I, you know, I'm not sure. We were arguing about Amazon earlier this week. Oh, boy, I can't wait for Monday's episode.

It's going to be so much fun after my weekend of research. But here's my question. Is like, are the glasses really going to be that much more useful than the phone? Like how much more utility are you getting there and how much more friction is there? Yeah. I thought these examples were poor, right? Right. Well, he did two sets of examples.

This was the first set. And they were things that you can do now because you can talk. to it. It doesn't yet have the capability to see. The more interesting ones were the ones where you could see where you're looking at a monument or a building. If you want to know what a building is. Yeah, exactly. Or a translation sort of thing, right? And that is, I think, the interesting bit. At the end of the day, what he described in that first use case, you could get with like AirPods, for example. And I think AirPods are the first, you know, I've said this before. They're the first sort of AR experience. You have this sort of, you know, it's augmenting reality quite literally. Like it's

magnifying some people's voices. It's sort of muting other people's voices. You can talk to Siri when you're sort of out and about. It is very much a sort of AR, AR sort of experience, but without actual sort of AI sort of layered on top of that. And that was sort of this first use case. And I thought that wasn't that compelling. It's the bit about where you just say, what is that? And there's no pulling out your phone. There's no. You can use prepositions, you know, or pronouns, I should say. You don't need to actually like define it. And it's it's just naturally and seamlessly seeing what you're seeing and answering it.

And this is where that vision demo, which moves away from, oh, I can point it at my son's homework and figure out how to do it to why would I, I wear glasses. Why would I not want to have these glasses on? Yeah. sort of all the time. Yeah, well, I was blown away by how far along they are, assuming that all of this becomes viable next year, as Zuckerberg was talking about, in terms of issuing a free software update to make some of the vision stuff possible. And to be clear, right now, the glasses can hear your audio commands. And that's also pretty interesting. But man, it could get crazy from here. So one stupid question before we move forward.

If you were running meta, would you be using the meta name as you released all these products? Because for me, it still conjures associations with the Metaverse rebrand and the VR play and everything else. And I see these glasses and it just feels like a totally different class of product. And so I wonder whether meta smart glasses is the right way to try to try to. bring these to market. Well, this is part of the reason why I had to restart my article at 11, because I was going on in sort of a metaverse tangent. I realized, number one, it wasn't really very good. And number two, I'd already written it previously. Like, my take all along has been that the metaverse is the internet, right? And, you know, that bit that sort of like was preserved

from my old article at the beginning about what is the nature of media? It shifts time. What is the nature of communications? It shifts place. You know, this is an aspect of, though, watch a movie is not reality. Like the initial idea of virtual reality was, I think, first coined in the context of like theater and like the idea that it's presenting a reality that is not present to you and it's something that's sort of bringing you to a different place. And in that view, all of technology is virtual reality. It's, it's not real. It's not in the real world. That's why it's scalable. That's why this idea of zero marginal costs. But there's a bit where we don't call any of that virtual reality. And it's not a bit like, you know, I would say AI. Everything was AI until

it's figured out. Then it's machine learning. And then when it's productized, it's a spreadsheet. Right. You know, that, you know, we've, I think settled on we're actually with AI. We're not going to rename it sort of machine learning or whatever it might be. But what is the missing piece? What's different? And I, it's something about this removal of sort of the human sort of constraint, where there's something discreet about watching a movie or watching a TV show. You can imagine a future and wasn't there yet, but where the quest, you know, the sort of VR experience or the Apple Vision, you're going into an environment that is generated on the fly. That is actually completely unique. And you're interacting with Tom Brady as an avatar who is actually having a conversation with you and doing whatever it might be.

And it's fully generative AI. And you can do that sort of at any time. That's obviously virtual reality, right? But there's a bit where getting to this, getting to this space, it's not an all or nothing sort of thing. We're getting bits and pieces now. When you're talking to chat GPT, that is obviously sort of a virtual experience. It's a unique experience that you're having sort of in ones and zeros that no one else is sort of having at the same time. And it's interesting that AI to me, I wrote this when Dolly came out, that AI was going to be key to the metaverse in making it happen.

because you needed to have a scalable way to generate content. And I think I undersold it. And my point is AI is virtual reality. It is the suspension of disbelief where you're fully immersed in what feels like reality to you, but is not actually your physical reality, is, I think, inseparable from generative AI. It is the same thing. And the whole thing now is getting the tools and getting the hardware and getting the experience to a place where it's easy to dip in and dip out and go back and forth into that reality.

And it's so fascinating to think that a year ago, these glasses are like, yeah, they're super speculative in 10 years. Imagine you'll see holograms. What we're working on now is VR because it's more technologically possible to, in my estimation, the glasses are much closer to virtual reality because of the use case we just talked about. And the headset feels miles away because the capability to generate these immersive environments. Right. is much more difficult.

And to me that, that's super fascinating. We've been thinking about this all wrong. Yeah. Well, and I, in the same way that the way humans interact with a product is sort of an ineffable question. The way humans interact with a brand and like the associations that a circle of. Oh, sorry. I forgot to get to. I forgot to get you.

You never circled back. But this is what we promise people today. Both of us are operating out very little sleep. It was on the back of my. mind that there was a point I was supposed to get to and I forgot. I mean, as far as the meta brand, whatever, it, it is what it is. I mean, it probably is a whatever question. It just occurred to me because I was like, this is a really cool product and the metaverse stuff they've been pushing for the last couple years has just really fallen flat with the general public. And so I would consider,

if I were Mark Zuckerberg, I would go that direction. This whole thing. The metaverse is like walked into this perception publicly speaking of a 3D experience where you put on sort of like a helmet on your head. And Facebook has always pushed back against that. Mark was like a way to push back against that saying, no, you will have a metaverse experience on your phone and on other devices. But that's been presented as, oh yeah, don't worry, you can have an inferior experience on your sort of phone and have the full experience sort of with a headset. And I think that misses the point. A full experience is achievable now as long as you constrain yourself to what is technologically possible because the fullness of the experience

is a matter of human perception and feeling, not technology. The technology is a means to the end. And so if you can, like, I was in virtual reality when I was talking to chat GPT. There was a moment, even if it was relatively brief, where I felt like I was talking to someone. And there is a bit about the human experience where we always, our expectations are always changing. And we'll look back and say, oh, that was so rudimentary and poor and whoever thought that was interesting. but the technology will get better as well. And so I think the key thing for thinking about AI and virtual reality is how can you actually

appropriately constrain the experience such that it achieves this sensation of virtual reality. And I would say that is virtual reality. And then you expand the experience in line with the capability of hardware to deliver it. It's not going to be this sort of, you know, I think we're led to stray by science fiction. The whole idea of you sit down, you put your headset on, and you go. to a place, I think is actually backwards. I think we're going to sort of like slide into places and it's going to be gated by our capabilities, but it's going to be the seamlessness and frictionlessness that makes it

compelling, not the sort of like jarring shift in where we are. Yes. Well, speaking of science fiction, a little dystopian, but I think the reality is, is that as virtual reality takes hold, it's eventually just going to merge with reality. And nobody's going to say virtual reality out loud. It's just going to be how we interact with technology and each other. And that's just going to be. No, I mean, you sound like Mark Zuckerberg.

They sort of said the same thing yesterday. And look, you know me, Ben, big Zuck guy. He looked great on stage. A shirt a little too tight on Zuck would be my one note for him. Hey, man, he's been working out. I know. I was going to shape as he was. I'd probably be wearing tight t-shirts too.

The SEAL Team 6 workouts are doing great for him. I love the vibe of the presentation, by the way. It was so, it's like the total opposite of like an Apple super antiseptic, overproduced sort of thing. Which, by the way, that's not a critique. That is Apple's brand, right? Like Apple's brand is like these are the most, these are jewels of products. We somehow figured out to mass produce. You go into our white, gleaming white stores and you sort of buy, that is Apple's brand.

And so like there's a bit where Apple doing these fully produced commercial, hour and a half commercials is very on brand. There was a bit about like, like, but the Facebook presentation could not. I've been more different. He was, Zuckerer was only on stage for like 15 minutes or 20 minutes. It was pretty casual. And, and that was good for him because he's compelling when he actually shows like emotions and

excitement. And you, right. Like, when he's a canned sort of speaker, it's like, this is the worst canned speaker. It's kind of like jocular.

And he's still a little awkward, but it's endearing. And I came away feeling like, I like this guy. Yeah. Yeah. No, exactly.

Because he's, he's clearly enthusiastic. about this. Look, you don't stay at meta and go through all the controversies and all the criticism and all the blowback that he's gone through if you don't still sort of have the fire in the verb, right?

It would have been easy to leave at any time in the last six or seven years, hand it off to someone else. It would have potentially been better for the core Facebook business because just sort of taking him away as sort of, if we look at it like Google, like generally skates by because people can't even pronounce an underpricide's name, right? And

he could have done that. He stayed because he loves it. And how do I know that? You can see it on stage. You can see that he's juiced up about this. And that is it that that's, I think that that's how you redeem the meta name or the brand name is by making cool stuff and being excited about it. Interesting. Yes. Well, I will see whether they can successfully redeem the meta name. But one other note from the hardware world, Johnny I, the renowned designer of the iPhone and Open AI CEO Sam Altman have been discussing. building a new AI hardware device, according to two people familiar with the conversations. And then the Financial Times followed up confirming that reporting and added,

Sam Altman, OpenAI's chief, has tapped Ives company Love From, which the designer founded when he left Apple in 2019 to develop the chat GPTs creator's first consumer device, according to three people familiar with the plan. Altman and Ive have held brainstorming sessions at the designer's San Francisco studio about what a new consumer product centered on Open AIs technology would look like, the people said. They hope to create a more natural and intuitive user experience for interacting with AI in the way that the iPhones innovations in touchscreen computing unleash the mass market potential of the mobile internet.

So, Ben, I went to Ben Thompson University over the past few years, and I learned that software companies usually struggle when they try to make hardware. why might this be different? Is it ultimately, because I mean, meta is throwing like billions and billions and billions of dollars of resources. Is that going to be the open AI game plan if they move forward there? Yeah. Oh, they don't have billions and billions of dollars as much money as it might seem like they have.

Right. And your learnings have been well received because it's probably a terrible idea and is not going to work out. That is generally the history. There's basically the only three successful at-scale hardware start. in Silicon Valley history is like Apple, Tesla, and SpaceX. Like there's like, everyone else is like flamed out, right? Or they're like they sell to corporate enterprises. Like it's consumer hardware is unbelievably difficult.

It is sort of unbelievably challenging. And so this is maybe sort of a flawed pursuit. And I'm not sure that Johnny I frankly is the right person. He, what he did with Apple is incredible. But I think the older and more respected and renowned he became, the more sort of persnickety and deliberate and overly focused on form over functioning became as well. And there's a bit from being a startup like you need the scrappiness and hunger that I'm not sure there is there.

But that said, the potential here goes back to what I talked about sort of at the beginning. If it is the case that achieving virtual reality, which to my mind is the same thing as achieving mass market penetration and sort of dominance of AI is dependent on this removing friction of making this sort of seamless experience, then both Open AI and Facebook have a problem because of the phone platforms. Number one, they're limited. Number two. So in order to realize the potential, they need some sort of step change. That's right. Number two, is the phone actually the best form factor. I mean, you've, there's the movie her, which is it's sort of like an earpiece and you slip it in, you have the conversation. Now, the most unrealistic part of her is he never seems to

charge the earpiece and it seems to work all day long and it has like great reception and, you know, and all those sorts of things. So, uh, not clear how that happens. That's why we need the, the superconductor. That'll be the next step there. Right. No, exactly. So there's, you know, that's why the phone makes the most sense. You're already. carrying a large device with a lot of compute and connectivity and a big battery, that will probably be the key thing going forward. Now, maybe you can build some sort of augmentation like the Facebook classes that, you know, has a connection to the phone and works through an app or sort of whatever, whatever it might be. Maybe it is possible to build something sort of independent.

It's probably, again, for all the reason I just stated, it's probably crazy and won't work, But it might have to work for this to truly reach its potential. And there is a bit where there's an opportunity. This is a fundamentally new way of interacting with computing. And that was the key thing for Apple, why Windows could be dominant, but Windows afforded no advantages going to mobile. Everyone assumed Microsoft would do well in mobile because they were dominant PC. and it turned out it was a different paradigm

and it required a complete rebuilding of the software stack and that was the shift to Intel to arm happened and all sorts of things happened because you needed a fundamentally different paradigm for this sort of use case. The question is can AI be tacked on to that and must it be tacked onto that because of the physical constraints

or is it so different and the actual superior experience will pull out a new form factor or a new way of computing. And I don't know. Again, I think the safe bet is to be skeptical, but I can also completely understand why Open AI is exploring this. And I have much more sympathy for Facebook's massive investments in this area.

Yep. No, that makes sense. And I just wanted to make sure I wasn't crazy as I was reading those reports earlier this week and thinking, I think this is a major long shot if they do go this direction. And also there's a difference between meta doing it as meta makes like gobs and gobs of money every quarter and open AI still try to figure out how to make money and having to then throw like hundreds of billions of dollars at a new product, a new platform, creating a new piece of hardware. It would be an adventure, but also kind of exciting to think about where some of this stuff might go. So I enjoyed you kind of fusing all the different news late in the week there. on that note, do you have any final thoughts on AI and hardware or should we move to a couple

emails at the end? I mean, I just think the interesting thing for meta. I mean, I said this last year. I said it in the past. Like, it's not if this metaverse stuff works out and working out by changing our definition of the metaverse, but hardware still mattering counts as working out, right? It doesn't, like, like, there's lots of stuff like this in tech where you don't actually, And this is the problem with a lot of tech demos is you have to come up with a demo to show the capability.

But the entire point, by going back to the newsfeed bit before, is it's use cases that emerged that were not possible previously. By definition, you don't know what those are. There is a process of discovery and emergence that when you look back, it's like, oh, yeah, this is obviously sort of the use case, you know, for more. But it wasn't clear at the time. So it's dangerous to overindex on demos. and you could argue it was billions and billions of dollars wasteful for Facebook to over index on a 3D experience that you get from putting something on your head. Mark Zuckerberg said it himself. He's like, we thought smart glasses.

We would play around with it because we want to put holograms in front of you at some point in the distant future. And it turns out there's actually this super compelling use case that was sitting right there that we didn't even realize. But thank goodness we already were working on the product. And we could sort of plug it right in, right? And I, and you put yourself in open AI's shoes and it makes you appreciate where meta is in relative terms because they've been sort of laying the groundwork for what might be next year. Even if they go through with this, they're years away from Facebook's current capability. And, you know, that again, it might not work out for Facebook. We will see it might be, are they going to sort of earn a return?

But I would say that the, you know, the pendulum is swinging in Zuckerberg's. direction, not just in terms of the stock market and sort of public perception, I would say that the argument and for his entire metaverse bet is looking stronger today than it has in years. There you go. Summer of Zuck could be the year of Zuck before all is said and done there, Ben. So earlier in the week, you mentioned you wanted to do a quick follow up on Delta. The Points Guy reported that in his first public comments, since changes to his airline's frequent fire program, outraged scores of loyalists.

Delta Airlines CEO Ed Bastion admitted that the airline may have overdone the changes, alienating some customers. No question. We probably went too far, Bastion said. Our team wanted to kind of rip the Band-Aid off. And I enjoy how the Points guy is sort of like the Woodward and Bernstein of the Delta airlines. scandal here.

So what were your additional thoughts, your follow-up? No, it's actually connected to this. I mean, I actually think ripping the Band-Aid off is good. Like, what you don't want to do is have to change your frequent fire program for the next three years running and endure the pain like year after year after year. And develop like, like, so there's a bit, just make everyone mad once and sort of get through it and get over it. The problem, and this was sort of a friend of mine that works in, in businesses with

very high fixed cost and sort of things. He sort of raised this. He's like, the problem for Delta is it makes sense why they're doing it now, you know, my articulation of the fixed cost and, you know, at the end of the day, prioritizing people in the front of the plane and who frequent travelers, that's all true. But that other bit about frequent fire programs also exist to fill up the back of the plane, that doesn't go away. And at some point, capacity will catch up.

And the nature of high fixed cost businesses is capacity doesn't catch up perfectly. It always overshoots. And then you have too much capacity. And suddenly there's too many tickets available. And there's actually too much lounge space. And now you're looking, you're back trying to capture the sort of hoi-pooi to sort of like fill up your seats. But you've made them all so angry and they switch their frequent flyers to someone else. And it's a big lift to bring them back.

And both parts are important. And so that was just sort of, it was actually before Delta, this Delta CEO comments came out. But it's basically saying the same thing. There are sensitivities to make sure you're aware of on both sides of the spectrum. Right. The fact that frequent firemowers serve both sides isn't just a point of interest. It's a necessity because the nature of these high fixed costs sort of businesses is you have to strike the balance. the other nature of high fixed class businesses

is you always overshoot in both directions. You swing back and forth from too much capacity and prices are low to too little capacity and there's overcrowding and prices are high. And the danger in ripping the Band-Aid off is if you go too far, yes, you're going to have more lounge space next year. Yes, you're going to sort of make people happy.

There's that Wall Street Journal article about the people that were happy about the change. It's like, thank God if I get it allowed. Just like, I mean, who agrees to be? be interviewed for these things, right? I mean, I don't understand it. But, uh, but you are setting yourself up for more long-term pain because the pendulum always swings back. Right. If you bleed normies, eventually you start to feel it. So we'll see whether they calibrate it differently. Yeah. And my pushback to my friend that, that we sort of talked about is my concern from an airline perspective is there are

constraints that they don't control, which is specifically runway space, pilot availability, and lounge space. And all of those depend on. And all of those depend on. it depends on building in America, which is kind of a tough thing to get done. And you have, you know, the pilot thing is probably the most likely to change. The increase to 1,500 hours to even sort of be a first officer. To go up from 200, it was, it went too far.

And you're sort of cutting off. It's too hard to get people in the system. And that can be changed. Now, it's going to be politically difficult because who wants to have, you know, potentially butt on their hands, blah, blah, blah. And this all happened because of this crash in Buffalo. And the weird thing was is that actually had nothing to do with the hours thing.

It was just there was, it was such an awful crash that there was a demand to do something. And so this regulation was sort of passed. And we ended up in this huge pilot crunch that was exacerbated by COVID in the early retirements and all that sort of thing. The pilot thing will smooth out because, you know, they're raising salaries hugely. It's a pretty sort of lucrative career, you know, potentially sort of going forward. And that will help balance it out. And there might be regulatory changes once it actually gets through consciousness. head that, look, the reason why you're getting all these flight cancellations, the reason why

prices are high. They're not pilots. It's because there's no pilots, right? Yeah. And just to be explicit right now to become a pilot, you have to have, what, 1,500 hours? 1,500 flight hours before you can even sort of like be a first mate who's like sitting alongside a captain. Used to be 200.

The 200 to 1,500 is a pretty big job. It's expensive to put in 1,500 hours while not being paid to be a pilot. So you're just limiting, artificially limiting the pool of a very big job. available candidates. That's right. It's a bummer because, you know, I think it's a great, it's a potentially great job that you have to either be rich or go in the military to sort of, you know, have a chance at. And so, but that could change, right? So the pilot thing could change.

Building new lounges, building new runways, that is how long does that take? You see these projects that take years and years and years and that is going to be a constraint. I suspect that was part of Delta's thinking, which is, look, we understand we're in this. large fixed cost business. We swing back and forth. We don't want to risk alienating our best customers. But realistically, when is this problem going to actually be fixed? It might be a while. And we, you know, we can go buy our oil refinery like Delta did sort of like 10, 15 years ago. We can't build new runways. That's dependent on government. And the track record of that getting done, particularly in the last few years, is not particularly great. Not great. Yep.

Well, speaking of infrastructure, Matt says Andrew was incredulous about a $500 year subscription to Disney IP streaming services. I've heard plenty of more experienced parents, including Ben, just giggle at the base naivete of Andrew. But what I think is interesting about the $500 a year bundle idea that Disney could offer is the sweeteners Disney can offer that others can't. For instance, hotel and park discounts. year my family had a Disney World vacation booked and then a couple months before to boost post-COVID Disney World business. Disney offered Disney Plus subscribers a really nice discount on Walt Disney World hotels. It was well worth canceling our original trip and rebooking with the Disney Plus deal. That's just one lever available to Disney. In a $500 a year bundle deal, what if Disney also offered

a $100 food card for every day of your stay on a Disney property? It sounds like a great little sweetener, but in reality, a whole lot of folks who might be nudged into signing on for $500 a year wouldn't even get to use the card. Those who do will only be further entrenched as fans of Disney and their streaming bundle. Disney parks make up a massive portion of Disney's profits, so there's a lot of room for creativity and finding new levers. I just thought that was a really perceptive read on the landscape, and I think it tracks with where things will head. which is that ultimately some of these digital products will be blended with live experiences and bundled together. And we're already seeing that with certain sports experiences too.

And so I just appreciated the note from that. The problem, though, and the reason why I, you know, you really wanted to do the Delta thing a couple of weeks ago. And I'm like, I don't know if I want to wait into it. But I'm like, okay, we can because I think there's an interesting tie to Disney, which we've actually been talking a lot about, which is the problem with this. All this is true and right. And Disney, you know, should continue developing this and understanding their customers and giving them, you know, imagine Disney emails you, your customized trip to Disney World with a discount. We know, based on what you view on Disney Plus, you know, we think you'll enjoy this and

XYZ and it's all ready to go. And we have a $100 gift card, all those sorts of things. Number one, there's obviously the technological capability to do that. And Disney, you know, is working hard on that for sure. But the other limit. limitation is park capacity. They have to actually build parks. Capacity. And this is the whole point is that physical infrastructure, large CapEx businesses can be good businesses, but they face real constraints that digital businesses don't. Disney used to be a virtual reality company. They're in the framework that I presented in this article, this idea that you could watch a movie, you could watch a TV show, you can watch sports in a different.

time or place than when they occurred and they could monetize that sort of again and again and that scales perfectly. It scales to everyone anywhere, anytime. They are having to be a reality business, which does not scale. The whole point, the whole excitement about the AI bit is it is the scalability. This is not scalable. You have to actually build the parks. You have to actually spend 60 billion over 10 years. They just said they were going to do to expand your parks, expand your cruise lines, and you're gated by how much service you can do. And the challenge in the long run is this same thing as Delta. On one hand, if they build all this capacity, they need to fill it.

So they need to make sure that it's being utilized. On the other hand, if they fill it too much, it's not going to be a great experience, right? They're going to be constantly in this back and forth to sort of properly optimize delivering the best experience while also getting full utilization of their assets because it's no good to, go to Disney if number one. It's overcrowded. Yeah. Every ride is full lines and everything else. Yeah. And Disney

benefits from the anticipation of wow, we're going to go to Disney World this year, right? And it's, yeah, go there and end up spending ungodly amounts of money. But it's okay because it was like the experience wasn't just going to Disney. It was the year anticipation and the planning and sort of all that sort of thing.

If Disney, say Disney put a Disney world in every city and town in America, uh, suddenly it's not very special anymore. It's not it's not really interesting right and this is of course that they're a long ways away from that. It's just to point out the reason this is so I call it it Disney's Taylor Swift era. It's a good business. It's not nearly as great a business as sort of the digital era was because of the fundamental constraints of the physical world. Right. Well, and one of the constraints that some of these companies are encountering are

just apathy on the part of customers as they pitch more expensive. streaming services. And so the idea of bundling together real world services and sweeteners. When did Disney do this discount in the post-COVID Disney World business? Disney had an extra shock on their business. So they had a reason to drive business. If their parks are full, there's no reason to offer this. Now, they might want to offer it to increase the attractiveness of their streaming service. But if you offer something that actually sucks, that's not It's valuable that could get you in trouble also. Yeah. Well, okay, so Disney and Delta have even more in common than we initially thought when we were discussing this a week ago. With the reason why I mentioned in the article yesterday, the reason why tech telecoms and media are lumped together is because their nature is high fixed costs and zero marginal costs. They can scale infinitely. And this is sort of like when you understand the cost structure of a business, there's all sorts of things that are downstream from.

understanding that implementation or that implications or whatever might be. In this case, the reason why Disney and Delta are similar, it's immaterial of their business. It's the nature of their costs. It's that they have large fixed costs. And it's interesting because there is a zero marginal cost to an extent, right? A plane's going to fly regardless. Like adding one more person is good. But those costs sort of expire, right?

Once the plane flies, you can't sell the seat anymore. Once the day passes, you can't sell a ticket to Disney World for that day. And it's just a fundamentally different business. It's not necessarily a bad business, but it's not, it's not necessarily going to be as profitable as a digital business. There you go. Well, last question. And congrats to us for making it to the end here and remaining mostly coherent throughout. Ashton says, long time listener, first time emailer.

Speaking of bundles is the Taylor Swift, Travis Kelsey item, the epitome of celebrity bundling. What's the overlap of NFL fans and Swifties? Aside from Ben, maybe a few thousand other girl dads, are there other comparable celebrity bundles, the Kardashians, I guess, but they haven't strategically grown their brand in such a crossover way. And who out there could help grow this particular Swift Kelsey bundle grow? Who could be added to make it a thruple? I have no thruple nominations, Ben.

Ashton was throwing out a few foreign dignitaries that could join the Taylor-Travs fun to conquer other continents. And I just want to say, this was one of the more concerning emails we've ever received at Sharp Tech, where I started to worry about what we're doing to people's brains with this podcast. And then on the other hand, what is marriage, if not the world's oldest model for bundling? you know, so maybe that's something we should all consider as we head toward the weekend here. I know it's not the world's worst powerful bundling. I just, oh, look, I can't wait to see where you're going with this. There's a reason why I said the Taylor backlash is coming because, God, this has been the worst.

It's terrible. I want to be able to browse by no tech band sports feed in peace and not be viewing pictures of Taylor Swift in the box of a chief game over and over and over again. It's super annoying. And all along, my quote unquote fandom of Taylor Swift has been overstated. I listen to the music, women with my daughter. I admire her business sense. I actually think the bundle concept here is interesting because I don't think this probably about love and marriage.

It is about a total different fan bases that sort of like extend across each other. So, I mean, that's kind of a cute sort of analogy here. But let me be clear, this is super annoying. I am the person who has the cable bundle, and my price keeps going up, and I'm like, why am I paying for the history channel? Why am I paying for the news channels? This is annoying. All I want to do is sort of watch sports. Now, that's not a great analogy because the ones raising the prices are the sports channels.

I'm the problem. It's me, if I may say so myself. But the point remains. It's becoming pretty annoying. My best call of the year was saying that the backlash is coming, and I am living it and embracing it. Go away, Taylor. It's annoying. Well, listen, man, I believe you when you say that you're annoyed by this Kelsey, Taylor Swift story and the avalanche of tweets that you're encountering every time you open X. I'm just going to continue to use X and tweets and Twitter interchangeably. I don't believe you when you say you're not a Taylor Swift fan, no matter how many times you try to back away from the hysteria we all saw this summer. We all, we were here. We were listening and some of us were hosting the podcast with you.

I am not going to lie. I'm sorry for being interested in one of the most fascinating cultural experiences that speaks to the impact of the internet and the the nichification of society and what happens when you rides above that and all these sorts of things. Sorry I'm interested in the same things I've been interested in for years, Andrew. There you go, man. And look, I want to be annoyed by Taylor Swift and Travis Kelsey, but the algorithms on my various social media products, know that I'm not annoyed by this story and know that I will continue

to consume all the content in services. And so I can't wait. Taylor on Sunday night in New York at Sunday night football, it's going to be great. Still pretty convinced that this entire thing is like a CIA op or something, but I'm going to enjoy it while it last. Enjoy the social media content. Oh, how the pendulum swings. Yeah, that makes one of us. All right. Well, on that, No, we are coming back next week. And Ben, until then, get out there, touch some grass. Virtual reality is all around us. That's the lesson.

Happy Moon Festival. It's my favorite Taiwanese holiday. It's kind of, I frame it. I characterize it as the Taiwanese Thanksgiving because Entail is getting together with family and you barbecue and eat a lot. And that's kind of it, which to me is like the perfect holiday. It's hard to make it worse. Happy Moon Festival. I should say.

So, all right. question. Do you like mooncakes? Bill is pretty down on mooncakes. There is a mooncakes are mooncakes are like Drew Holiday. Oh boy. And what I mean with this is what the people came for. You know, he's a very streaky shooter, very high variance. And you know, when, when, you know, during the playoffs, you know, these crazy streets, do hit these stepbacks and it's all great. And the problems you'd get in the playoffs and it would sort of sort of not material. wise. A great mooncake is fantastic. It's really, really good. A poor moon cake is just crushingly bad and like just one of the worst things you've ever eaten. So it's high variance. And in a pressure situation where you're on an island and you're like the only thing I have to eat on

mooncakes, I'm going to be a little worried because the bad is just so bad. I don't know that it's worth sort of potentially getting the good. And so give me, you know, give me, you know, give me give me another delicacy if it's the only thing I can eat. But I did get some mooncakes this year, you know, yesterday. You got that 2021 ring also, you know. I got a lot. It's true. I mean, I always get a lot of mooncakes, but these ones, oh, best mooncakes of my life.

Incredible, a transcendent experience. I ate like three of them and I felt terrible afterwards. That's great to hear. Well, I look forward to hearing what food stuff Damien Lillard can be analogized to. I know. This is where the slowing mental gears were a struggle. I was trying to come up with analogy on the fly. But I burned out trying to remember that one Taylor Swift lyric and I'm kind of done now. It's been a great week for both of us. A lot of great content. Go to your show notes to hear Ben on Bill. And I've recorded like eight hours of greatest of all talk over the last couple days. So we'll come back next week.

But Ben, until then, have a great weekend. Happy Moon Festival. And I will talk to you soon. Talk to you later.