Did DeepSeek Just Break the AI industry?
01:35:34 Share

This episode is brought to you by State Farm.

You might say all kinds of stuff when things go wrong, but these are the words you really need to remember. Like a good neighbor, State Farm is there. They've got options to fit your unique insurance needs, meaning you can talk to your agent to choose the coverage you need, have coverage options to protect the things you value most, file a claim right on the State Farm mobile app, and even reach a real person when you need to talk to someone. Like a good neighbor, State Farm is there.

So the only things I've seen is, number one, it was an AI model that costs like a tenth of what OpenAI's models cost. Allegedly. And they're just as good. Allegedly. And then I heard that it was just trained on OpenAI stuff and it's not actually that impressive. Oh no, they stole our stolen content. What is up, people of the internet? Welcome back to another episode of the Waveform Podcast. We're your hosts. I'm Marques. And I'm David. And that means that...

Andrew's not here. Andrew's out this week. Ellis is also out. Adam's holding down the board. Zuri's sleeping in the corner like usual. It's going to be a good time. It's been a couple weeks, so we do have a lot to talk about. We got an exclusive look at Samsung's Project Wuhan headset. A lot of interesting thoughts there. We're going to touch on the new DeepSeek R1 model. That's kind of all of the rage in the AI industry right now, but also talk about Pebble. Sort of a...

Can I say nostalgia about Pebble? I think so. At this point, I feel like I can use that word. We'll talk Pebble for a bit, and also we're going to play another game of something or nothing at the end. Love it. But first of all, I'm going to give a shout out to a piece of content. So Nothing, that's a great headline. Nothing. Nothing, the company. Oh.

did a video on their youtube channel that was almost too smooth um but the video was so they basically emailed me a couple months ago and we're like hey we're gonna make a video on like what it would take to make your dream phone you know how you've had like a dream phone in the past where you just like frankenstein together a bunch of parts like just give us one of those and we'll make a video of it and i was like all right so i just threw together you know some specs and yeah it'd be cool if you can combine these sent it off and then two months later they're

They were like, hey, we made the video. So I watched the video, and it was pretty interesting. Basically, it's them breaking down roughly what the costs are of a smartphone. Most interestingly, the bill of materials itself in different components of a phone that go into actually building it. So I'm sure you've heard before that, like, yes, you buy a new phone, and it costs $800 or something, but that phone only costs $100.

$243 of materials. And then everyone goes, well, they could have charged less or they could have, you know, here's all these other thoughts on the difference between bill of materials and final retail price. But

I gave them some specs, and this is what they came up with. So go watch the video if you haven't already, but I'll just give you some top-line level spoiler alert stuff that I think is interesting, and maybe you have a reaction to this stuff. So I told them a huge battery. We rounded up to 6,000 milliamp hours, 65-watt fast charging, 15-watt wireless charging. They said $13 per battery. Your dream is 15-watt wireless charging? Well, I didn't specify, actually. I just gave them... I said fast charging. Oh, okay. My dream would be...

Give me 100 watts of both. 100. 100 watts of both, but $13 per battery. Okay. Which doesn't seem that terrible, actually. I wanted a 6.1-inch, 120 hertz LTPO AMOLED 1440p. 6.1-inch being on the small side, but everything else basically being like a high-end S24 ultra-level display. $35 per display. Okay. Okay.

That's more than I expected. Yeah. I mean, this, I guess, includes the glass and the panel itself. The digitizer and all that. Yeah. Yeah, that's fair. $35. Triple cameras. I just handed them S24 Ultra roughly, but also not really. I would like a further telephoto than the S24 Ultra has, but the idea is triple cameras on the back and Pixel 9 front.

Front-facing camera. And $80 total for all of the cameras. Okay. $80 for the cameras. That seems... We got a bunch of sensors. Yeah, a bunch of sensors and then all the glass and optics and whatever has OIS in there. Right? It seems crazy that it's that much higher than the display. I always figured the display would be...

Just as expensive as the cameras. Yeah, I imagine the plastic actual lenses are not that expensive, but you never know. The sensors, Sony has a monopoly on, so what are you going to do? Fair. Or does Samsung... Well, no, those are Samsung sensors. Samsung makes some of their own sensors. I forgot, they're Samsung ISOCELL sensors. Yeah, so I mean, I'm not tied to those particular sensors, but yeah, roughly $80. The storage...

Give me a terabyte of fast storage and 16 gigs of RAM. $90. Oh, for that. What? Yeah. Really? Yeah. Oh, storage and the RAM. Storage and RAM. Okay. $90. That still seems like a lot. It does. And, you know, whenever you go to upgrade, I guess it makes sense whenever you go to upgrade a phone and you try to double the storage, then you see, like, real price increments. It's not like you have to pay $200 more to upgrade.

in bill of materials to increase the storage, but it is a pretty significant cost in the phone. Yeah. So that's interesting to see. A 1TB SSD is much cheaper than that. It is true. But, you know, it is fast. Yeah. It's a tiny little UFS 4.0 SSD. We'll go with it.

And then a Snapdragon 8 Elite. So again, they said they couldn't reveal their exact supplier costs. Totally fair. I'm sure Qualcomm doesn't let them. But they basically referenced that a typical high-end flagship chip will cost about $190. Wow. For them and their volume levels. Dang. So pretty pricey. I actually have some more thoughts in my S24 Ultra review on the cost of the Snapdragon 8 Elite. We'll get there. Okay. But yeah, $190 for the chip.

And then some other small things, a motherboard, antenna, electronics, haptics, other small electronics, miscellaneous, $15. Okay. Packaging, cable, stickers, box, everything.

$30. Wow. That's a lot more than I thought. Design materials. So if there's fancy stuff you want to do with the back glass or any decoration or maybe some special paint or something like that, $8. That would be why we see so many insane weird designs on the back of phones. Yeah, not that expensive. You can do Dragon Ball Z for $8. Exactly. Structural parts. Buttons, mid-frame, screws, things like that, vapor chamber even, $10. Okay. Patents, licensing, software.

$29 per unit. And then rough total bill of materials. Again, this isn't the all-in cost to make the phone, but the bill of actual physical materials, $500. So that's for my crazy super high-end dream phone idea. $500 to build. Wow.

But then, of course, you have to add on R&D, all the teams, the employee costs and licensing and things that actually one-time engineering costs to make this bespoke motherboard, work with this RAM, work with this storage, all that stuff. And then, you know, it breaks it down in detail in the video, but it might cost $20 to $25 million to develop a phone. That was a very funny part of the video because they're like, all in all, it should be about $500. But R&D is about $20 million. Yeah. Okay.

Okay, so you got to sell a lot of these. You do have to sell a lot of phones to sort of break even, obviously. My favorite quote was they wrote, as one of the costs, purchasing products from competitors to understand the market better.

And they didn't give an exact cost for that, but I thought that was interesting. Priceless. It happens in a lot of industries, and I just kind of forgot that that's a thing you have to budget in. Like car companies just buy tons of other companies' cars. Yeah. And then drive them around and go, huh, I kind of like that. Definitely more expensive for car companies than smartphone manufacturers. Well, in total cost, but maybe per...

in development relative to the final cost of the product. Maybe it's reasonable. You buy five or six cars to make a car. You buy five or six phones to make a phone. So yeah, anyway, go check out that video. It's pretty interesting and insightful. I don't really see that sort of behind-the-curtain...

Number stuff very often from smartphone manufacturers, so we'll check check that link in the description, okay? But anyway, okay, so I think the big interesting thing we got to try this week with them could be HD video We put out project muhan Samsung's headset collaboration with Google and Android XR I thought it was very interesting yeah, I have so many questions about this I have all the answers okay Yeah, so I tried to put as much as I officially know in the video we don't know things like the exact release date or the exact price and

We don't know a couple final details on things like field of view and resolution. I can just tell you what I thought looking into the headset. But most of what this experience was was me getting to try it and use Android XR and Gemini on a premium-ish Samsung-built headset.

And yeah, it looks a lot like a Vision Pro. They clearly were inspired by a lot of parts of the Vision Pro, but they did improve on some stuff like the battery, even though it is a battery tied to the side and in your back pocket. It's a removable cable. It's USB type C. You can pop on another battery, which is pretty cool. Can you plug it into a wall socket? You can. Okay, cool. Well, the battery that comes with it had another USB port on it that you can

plug into the wall. So you can't put the cable directly into a... Oh. That's a good question that I didn't get an answer to. I would imagine you can, but that didn't get an official answer. Because it'd be cool if you could plug it into a laptop as well, and then you have lower latency of some sort of screen expansion. True. That's a great point. Yeah. The cable's long enough to plug it into a laptop for sure. I don't know about a wall, but it's a cool idea. Okay. Yeah. I mean, they made this headset...

A while ago, they've been finalizing it. It's near final, so what we saw and what I put in the video is probably basically what you're going to get when you eventually get to buy one of these things, and they'll name it, and it'll come out, and it'll be a

some unpacked event later this year, probably. I was just going to ask if this is the name they're going with. It's probably not going to be called Muhan. Or no, Android XR. Oh, Android XR, the platform name, yes. That is the name of the software, officially, yes. Muhan is a code name. It means infinity, and it's a code for this infinite canvas that you get to play with, which is kind of cool. But it's not the name of the product. Is reality really an infinite canvas? Technically, yeah.

I would say so. The universe is expanding, Marques. It's infinite and expanding, Candice. So questions. Yes. Okay. So something the Vision Pro does really well is that when you're looking through it, the frame rate is high enough that it doesn't really feel like you're looking through a screen. It kind of just feels like you're walking around and the quality is a little bit worse than real life. Right. How does that feel through this headset? Yeah. Okay. Couple notes. Okay.

One, the resolution of the display was pretty good. I would say a notch below Vision Pro, but it's pretty crisp. The overall pass-through from the cameras was good enough. It wasn't stunning or amazing. Vision Pro is definitely better. When you look through Vision Pro and pass-through, you feel better.

Very close to like just looking through like some blurry glass or something. This was, yeah, a notch below that for sure. But they did something interesting, which is they let you take off the bottom of the eye seal. And that gives you this peripheral vision where light is coming in on the sides and the bottom. And for some reason that made it feel better because your brain sort of –

fills in the bezels around the display with the rest. So like if I'm looking at a wall and I see through the middle, the pass-through, and then my peripheral vision sees the outsides of the wall, my brain kind of fills in the gaps. Kind of merges them in a way. Yeah, which is interesting. So I prefer using it without the light seal because of that. But generally decent...

B plus pass-through, I would say. Okay. Yeah. I have a question. Yeah. So we actually just dropped today an Apple Vision Pro bonus episode. Go watch that if you haven't. Go watch it. But...

One thing that the developers were telling us is that Apple did a good job at making Swift easy to code in for iOS apps. And when they released the new headset that you could just kind of like know what you're doing over there. Did they mention anything about that with the Android version? Like native coding? Yeah. Yeah. So the details that I got from the team are, number one, all Play Store apps are compatible.

Off the bat, no extra code. They will work. Just like as a little window pop-up? Yeah. So I went to the Play Store. I downloaded a regular phone app, and you could just pull it up and use it like a normal phone app, change the aspect ratio, scroll around.

it works. But there obviously would be a benefit to having optimized apps, so what they call spatial apps. And a bunch of the built-in apps are optimized, and apparently it's a few lines of code to update your app to be optimized for spatial. And then you can go crazy customizing and adding way more stuff. But like the YouTube app, for example, was a spatial app. It

chose to curve the window a little bit. It added these floating side panels for related videos and comments. And then you could add a background for an environment that you're in and then watch YouTube videos in this like floating background on a mountain. Not as high fidelity as Apples, just not as good looking, but cool feature. So my understanding is that it will be pretty easy to make spatial apps, but if you don't update your app at all, it will still work on the headset.

Do you think this is the reason that they haven't made a YouTube app for Vision Pro? Because they want it to be exclusive to an Android XR headset? I wouldn't be shocked. That's a pretty huge thing, especially since they shut down the third party Juno player. My question is, what's a better experience? Watching YouTube in the browser on Vision Pro?

Or a built-in YouTube app on this MUHAN headset. Right, because in Vision Pro Safari, they have a special window mode for a video player now. Yeah, so you can watch a video full screen with a background image

kind of exactly the same way you do on the muhan headset it just won't be curved right but you can change the size make it huge make it further away closer to you it's similar yeah i mean generally i feel like a player like an app is going to have better ux design as well than just the website in a window so the second you have to move around the actual browser it's better on the samsung headset

Better on Samsung headset. Yeah, because I don't want to have to move around YouTube.com. Right. Oh. With my pinching and scrolling. Yes, I hear you're saying that. So, yeah. Cool. That's an interesting thought. Okay. But I enjoyed that. Okay. Yeah. Let's see. Other questions. Do they have any sort of group...

you could do? No, none of that was demoed. No. I would assume no, honestly. I think it took Apple a while even to start building these shared experiences and SharePlay is their magic sauce that works kind of cool with certain apps. I had no shared space experiences with another person in a headset interacting with the same model as me. So no personas? No.

No persona. No crazy persona. That's a good question. I didn't make any video calls, so I don't know what that would look like to the person on the other side. Maybe they just haven't. They just see your eyes really up close. Yeah. Yeah, there was, I mean, I was in there playing with the headset for maybe two or three hours. I didn't get to do everything. Video calls is not something I got to, but I did play around with some stuff that's not in the video, like just web browsing and poking around the stock apps and things like that, throwing apps all everywhere. Honestly, my biggest takeaway was, damn,

Damn, this might be the most useful thing I've done with Gemini. Just because you can interact with it with your voice? It can actually poke into the UI of what you're doing. And it seems like it can actually move into spatial apps and use the UI for you inside of spatial apps. But I could ask Gemini to clean up my windows, and if I had six windows, it would line them all up next to each other. That was great.

I could ask it to open an app and open Google Maps and show me this thing. And it would do that for me. If I was looking at something in pass-through, like a painting on a wall, I could

I could say, who painted this? And it would just pull up a web browser and Google something like, who painted this? So it was pretty hands-free. Obviously, it still has eye tracking. It will still have controller support. But the complete lack of having to do anything with my hands and just talk to the thing was surprisingly nice. That was something I noticed for your video. So you said that you basically just have to put it on, do the eye alignment thing, which takes like a second. Yeah.

and it automatically does hand tracking. - Yes, hand tracking is automatic. - So you don't have to do the whole Vision Pro thing of like look that way, look that way, look up, look down, put your hands down. - For eye tracking, you do have to do the looking at the dots to dial in eye tracking. So by default, eye tracking was actually off and I was doing everything with my hands. And so I'd point my pinched fingers at the subject and select things, which is a lot of arm movement. Obviously in Vision Pro,

you straight away set up eye tracking, and you never really do that. You can do essentially the same exact setup. It's like looking at a bunch of different dots around the screen, and then it's like, all right, we can see your eyes now. And then you can do the same sort of stuff. MARK MANDEL: Does the eye tracking feel as precise as on Vision Pro? MARK MIRCHANDANI: Yes. It's the same magical look at something. It's highlighted instantly.

move around the UI with your eyes and you can see like things like highlighting as you look at them. Yeah. It's pretty good. Okay. So when Vision Pro came out, it definitely felt like a very magical type of technology. Do you think that this feels like it's exactly caught up? If the Vision Pro is an iPad Pro, the Project Muhana headset is a Samsung Galaxy Tab.

So if, yeah, yeah. So if you love the magic of this is a super thin piece of hardware that is actually a computer and is amazing that it can do all these things on a display, you will feel the same thing with the Galaxy Tab. There are certain things that Apple does in their own ecosystem that are another level of magic on top of that, whether it's share play or shared experiences or whatever apps that you're using that will not quite be the same on the Google One. But-

Google has Gemini. Google has YouTube, Google Apps built in. It has that extra functionality to make up for maybe not being as magical. So it is very functional. There's no way it costs $3,500. I think this thing is going to be 2K tops, probably less. And I think that's going to be a big reason people will consider it over a Vision Pro.

But yeah, not as magical. I think that's fair. Interesting. I guess the relevant comparison could be like chat GPT to Gemini where because Google has all of these Google apps that Gemini can tap into and it already has all your data, it's more useful than just like a platform. Yeah, in certain ways, yeah. And it's cool. And I still want to play around with it again and try things like

like searching inside of apps. Like I wonder if I could ask Gemini in the headset about an old email and I wonder if it would go into the email app and like search through that and find it. That would be amazing because essentially it's just pulling up an instance of Gemini Live, which is the conversational thing that's on your phone already. So I imagine whatever that can do

this can do in a headset. And since it's multimodal, I can take a picture on my phone and it can search that. Well, the headset, it's running past her all the time. Gemini Live can see that. And so I can just ask about something I see. So there's a lot of potential, I think. And when you were talking to Google, did they have like a specific positioning for why this is useful to use? Because Apple never really defined what it...

thought that Vision Pro was best suited for. It's true. And I feel like it tried to figure that out after it launched it. It was like, this is great for media consumption, but also kind of work. And as they have updated Vision Pro, they've updated the categories that they found people were using it for more. But it still isn't that defined. And I feel like that's part of the reason it is not doing well. Did Google have like a, we want people to use this for work. We want people to do this. Like, do they have a positioning for this thing?

It was mostly centered around Gemini. I think if you asked them, they would mostly tell you like, this is the best way to use Gemini. You could use it on your phone, you could use it on your tablet, or you could just use it on a pair of glasses or a headset that you're wearing all the time and it's always able to talk to you. So I think their pitch would be it's the best way to use Gemini. And that's

It reminds me of your AI pin part, like the part of the video where you ask the AI pin to name the car and then you just pull out your S24 Ultra and it does it faster. I think there's some instances where this is the best way to use Gemini. In a world where you're walking outside with this headset on and a car passes in front of you, you literally look at the headset and go, you just look at the car and go, what kind of car is that?

And Gemini can see the car you're looking at and can tell you right away. I had a clip in the video where I just held up a book. And the book had a picture, and on the bottom it said the name of the photo and where it was taken. And all I did was pull up the picture, look at it, and go, can you take me there? And it understood that context, understood that it was a location, opened Google Maps, searched that location, and then took me to that location in Google Maps. That's pretty cool.

So if it's able to do that sort of parsing and understanding and carry out actions into apps on my behalf, I think that's the pitch. That would be cool. I would like to see what the limits of that are outside of a contained space.

Totally. Also, do they have any sort of workspace expansion stuff where they can expand your laptop screen or anything like that? Did not hear about that. Okay. Didn't hear about that. Interesting. I would be curious what happens if there's any screen mirroring or if we're plugging into a laptop, what happens? Because I didn't get to do that either. I want to see their first ad for this device because I need to know what they're trying to tell people to buy it for. That's a great point. The ads say a lot about their...

What they imagine people use it for. I did get to connect a Bluetooth mouse and keyboard. Okay. And there wasn't like a computer also, so it was just me using the mouse and keyboard to move around Android XR apps, which is cool. But yeah, you know, quickly going from like pinching something and then typing on a keyboard and then a mouse, all of that was very fluid. Interesting. But yeah, I'm curious about that first ad too. Yeah. That'd be interesting. Yeah. Yeah.

I got two questions. One, were you able to look at a bright thing and then back at a not-so-bright thing? Yeah, good question. So like you said, this was a very controlled environment, and in the interest of getting as many shots for our video as possible, we basically stayed in this evenly well-lit room. There was a display in the room. I could look at it, and I could look around very quickly, and that seemed very fluid, but I didn't really—

I don't have any memory or footage of looking at a super bright light and then something dark. So I don't know if it would handle it the same way as Vision Pro. I kind of imagine it will, though. Just kind of trying to balance out exposure quickly and hopefully not peak. Yeah. Okay. Next question. Do you think that this is, like from what you experienced and going back to what their first ad is going to be, is this basically just a dev kit?

Again, just like how the Apple Vision Pro was. - Yeah, it's basically a dev kit. I think that's accurate. Honestly, right now they are using this headset as a dev kit, and that's where you'll see it right now before it's shipped. But yeah, it feels like, okay, Vision Pro is out, Quest 3 is out, there are headsets that are out in the world, and if we want any sort of advantage or if we want people to use our platform, we need people to develop for it. We need to get something out in their hands. This is the first headset that has Android XR.

Take it away, developers. Make awesome stuff. And then hopefully that leads them to making magical, amazing apps that make you want to use it. The Vision Pro problem still applies when it's Android. Yeah. So I just really need to know. But it applies at half the price.

Probably. Hopefully less. Maybe less. Because even at $1750, this thing is way overpriced. Yeah, but for a cheaper AR, VR experience, you get the Meta. The Meta Quest. Yeah, all the way down to $350, $300. Yeah. Plus, if you can't plug this into a computer, like, Android apps are great. I'm sure you'll be able to mirror your Chrome tab or something, you know?

Because right now the Quest can already have your computer display. I feel like you need to have that feature in this headset. But also, if you're going to charge more than the Quest, then you have to be able to do more than the Quest. And the Quest has a huge library of games. That's what they're really good at right now. Obviously, the work aspect is okay. You can computer mirror things.

I would imagine that this kind of gets positioned as, yes, you can work in it, but also you're just going to get so much done with Gemini all the time. It's going to be the super useful thing that helps you all. You're going to want to walk around the office and write emails while your hands are in your pockets. Yeah, but the problem is this thing is so big, just like the Apple Vision Pro. Like how long until this is in a pair of Ray-Bans?

Like if they're just going to put Gemini in glasses. Yes. Okay. Great point. And I think this will be probably the last big point for this. Android XR as a platform can be in a bunch of different form factors. So in this video, I said that Project Wuhan, think of it as like the pixel or the nexus or whatever of this platform.

meaning that it's an example for a version of the hardware you can put it on. But there will be, from other manufacturers, higher-end versions, lower-end versions, different form factors. We were, I think, maybe having a conversation about PCs. Maybe, was this last week on the podcast? But like, desktop computers are still a big thing that you put on your desk.

and people will buy higher-end components or lower-end components, but they'll just run Windows again at the end of the day. And so Android XR will run on smart glasses too, but if it's on smart glasses, you will have less total functionality. You can't watch a movie on them. You can't...

see the Gemini UI as easily. It'll just be text-based instead of like this whole animation thing. So there are pros and cons to different form factors, but the idea is they're introducing Android XR and then hopefully the hardware ecosystem blooms and the developer ecosystem...

Right. And everything's blossoming at the same time. So more of instead of trying to crash the VR and the smart glasses into each other, you're just saying, we're just going to run on all the types of platforms. Yeah. All the types of headsets. Exactly. Yeah. Android XR is the base software layer for all of the future AR, VR, XR experiences that will come in the form of headsets, glasses, et cetera. Okay. Interesting. Yeah.

Okay. I'm very interested in where this goes and how they position it and all this stuff. I did hear that they were planning to release it in 2025, though. Yep. They wouldn't stop telling us that. Okay. They couldn't stop. Every time we asked about it, we're like, so this year? And then Andrew would ask, so actually this year? And then they'd go, yeah, yeah, yeah. Yeah, a couple months. Yeah. So they're, I think, planning on it. You should have asked them where the Galaxy Home is. Okay.

And I'd feel more comfortable. I know, yeah. Every company's got an air power that makes us go. That's true. Literally every company has announced something where we go, can we trust them anymore? Are they going to make, even if they have a track record, you're like, but you didn't ship that one thing, so now I don't know. That's true. Interesting. Yeah, Samsung, we'll see. 2025 allegedly. Okay, and nothing about price? Nothing.

Nothing about price. All of my estimates would come from, like if I was betting, we could probably estimate in an educated way

Like I looked at that display in the field of view and the materials and the battery. And I don't know the bill of materials, but I imagine from Samsung they would position this in a way that they could undercut Vision Pro. Yeah. But still be looked at as premium. Let's take a market that has very few people in it and make it even smaller. Well, Quest is popular.

That's true. But for gaming. Right. Like, you know what that's for. Nobody knows what these other things are for. Yeah. Well, hopefully. So that's it. I did play one game.

briefly. I don't even remember. GTA 6? It was one of, it was, I'm gonna launch it like such an idiot. I forgot the name of this game. It was like a flat game with like a bunch, it was a house cross section with a bunch of flappy bird. Oh, was it Fallout Shelter? Yes, that. Oh, okay. And I played that for 30 seconds and I was like, wow, it works with my hand control. That's crazy. Or I could have a mouse and it would work with that. Did it feel intuitive at all? I don't know how to play that game so I couldn't answer that.

But yes, I mean, I zoomed in and I clicked things and I held things down and I moved them around. So yeah, I think it was pretty intuitive. Decided to play mobile games on a thing I have to wear on my face. I mean, there's going to be games that don't work at all. Yeah. I imagine like a need for speed game is not going to. It'd probably just be in a window, right? You got to have controllers at that point. Yeah. I think you need controllers. Right. So is this built on Daydream or is this just a brand new situation? We don't talk about that. I don't think Daydream, I don't think Google even remembers what that is.

Do you remember cardboard? Yeah. It was under my seat. Cardboard was a great idea. It really was. Just put your phone right on your face. I'm going to guess when this Samsung headset comes out, it's going to retail for... Nope.

More. No, I was going to say $15.99 or $16.99. I'm guessing like $6.99. All the way to $6.99. I think it'll be more than Quest, but it'll be way cheaper than Vision Pro. Wait, $699? $799. $799? Yeah. Not $1,799. I'm going with $15.99. Ooh, okay. Okay.

Think about an S24 Ultra is already $1299. Yeah. But everyone knows what to do with an S24 Ultra. That's my thing. In order to get people to buy a thing, people buy the Quest even though they're not totally sure what they're going to do with it because it's $500. And there's two games they know they want to play. And during Christmas, it's $350. They want to play Beat Saber. It's a splurge moment. The Vision Pro is almost never a splurge moment. Yeah. Yeah. Yeah.

I just, I think it's all about positioning from Samsung. Totally. Like in the way that they position their phones against other phones, they have to... Do they even position their phones? Well, they call this... We know you're going to buy it. Like they have the S25 and the Plus and then the Ultra, which has to be more because it's bigger, but it's not that much better, but it has to be more expensive. So it's $1299. Yeah. That's...

Because they know that there's that market of people that will just get the best thing no matter what. Right. And I think they'll see other headsets come out after this. And then this one is the premium one in that world. There will be $700 headsets that run Android XR. From Lenovo. But I think, yeah, I think they're going to look a little different from this one. And I think the one that looks like Vision Pro, I think they're aiming it at.

people who might have thought about buying a Vision Pro. Come on HTC. Somebody who has $3,000. You got to make your comeback. We're going to get those cheaper headsets, but I think they're going to go right up well past a thousand. So that's why I'm going $1599, $16. For some reason, $1647 is in my head, but that's not a real price. $1647. Great year. Let's go $1599 is my official guess. I guess I haven't touched it, so I don't know how the hardware quality

There's glass on the outside. The frame is metal. It's not plastic. Remember, Quest is just plastic. Yeah. And I think this has metal. It had fans. It had that Snapdragon chip inside. Fans? You could hear them? No, I didn't hear them. Okay. But it had fans. It had nice machined buttons. It had the braided cable for the battery. It was definitely supposed to be premium. A Vision Pro. Yeah.

Yeah. It did look like a Vision Pro. You had the top shot of it on your head but above, and it has the same freaking button as Vision Pro. I said in the video, and people got mad because I compared it to Vision Pro so much, but I was like, it's not a digital crown here. It's a button. And they were like,

Why do you keep using Vision Pro as a reference? Oh, I wonder. Because Samsung did. That's why. Because it looks like a Vision Pro. Yeah. And it's also the one that a lot of people have seen. That video has 30 million views or something. People have seen Vision Pro. Yeah. So this is the reference. This is the Android version of the thing you've already seen. This is we've got Vision Pro at home. There you go. For people that want to buy it for their kids. Do you want to make an official price guess? $999. Okay. $1,000 flat. $799. $999.

$9.99. Oh, I feel like it's going to be more now. It is. If it's really now. I'm telling you. I'm giving you a chance. $12.99. Okay, final answers. Lock them in. Now we should do trivia. Okay. Okay. $12.99. Trivia, dude. So, David Kaz and Damien Henry were two French Google engineers.

Very Francais. With their 20% innovation time off, they made something that was introduced at Google I.O. 2014 Developers Conference. 2014. In fact, everyone that attended walked away with it. What was it? I have an idea. I think I went to that. I think you did too.

What was the thing 11 years ago? I should know this. When we were not old. When we were wee lads. When we were wee lads. All right. Well, we'll think about this one. Answers will be at the end, like usual. We'll be right back. Support for Waveform comes from Life360. So we'll be the first to tell you that not everything goes to plan. Life comes with its own set of curveballs for you to handle, especially when it comes to protecting your family. But

But while you can't foolproof your life, with Life360, you can family proof your family. So Life360 is a location sharing app that makes family life easier. Knowing where everyone is at any given time, making coordinating daily routines and activities a breeze. And you can attach a Life360 tile tracker to all of your family's important stuff and then track that stuff down right within the app. So whether it's your keys, wallet, even the family pet, you can get on a collar.

If it tends to go missing, it's a great candidate for a towel tracker. I have many things that tend to go missing in my life. Generally, my cameras tend to go missing. Ubers famously tend to go missing with my cameras. There was this one time that I definitely left my camera in an Uber and it

It was driving around, but I knew that it was in the Uber because I had a tracker on my camera. So now all my most valuable cameras that care with me a lot, they have trackers on them. Perfect. Yeah. So this year you can stay connected with location sharing and stay coordinated with place alert notifications for when someone arrives or leaves a given location. So you can family proof your family with Life360. Visit Life360.com or download the app today and use code WAVE to get 15% off. That's Life360.com.

That's life360.com code WAVE. Support for Waveform comes from Hostinger. So making your presence felt on the internet can be a challenge. Sure, you can try to make all the viral TikTok videos you want, but if you want to have a place where people can interact with you and be able to check out your product directly, you need a home base. And not just one of those cookie cutter ones that we've all used. This is your brand. It's your aesthetic and your footprint on the internet. So

So that's why you might want to check out what you can make with Hostinger. So whether you're looking to create a website or an online store, you can do it using Hostinger AI Website Builder.

Just describe your site, name it, and get a personalized site in seconds. And we're not just talking about the design. You'll also have the ability to generate content that matches your style and tone and tailor it to the length and topic that you choose. And you'll also be able to launch your business with a free domain and professional email included with your plan for the price of a cup of coffee. Plus, Hostinger provides intuitive tools that anyone can use and 24-7 support if you get stuck. So take your big idea online today.

with Hostinger. Head to hostinger.com slash waveform and start your website today. Use code waveform for an extra 10% off.

Flex your business with an American Express Business Gold card. You'll earn four times membership rewards points on your top two eligible spending categories, like transit and electronics, each month on up to $150K in combined purchases per year. Plus, you can now earn three times membership rewards points on flights and prepaid hotels booked on AmexTravel.com. Terms and points cap apply. Learn more at AmericanExpress.com slash business dash gold. Amex Business Gold, built for business.

by American Express. All right, welcome back. We've got to talk about this deep seek thing. Yeah. Now, here's my relationship with the deep seek headlines. Okay. I was working on the Project Wuhan video. Okay. And trying to distill all of my thoughts on this headset that I just tried and like over a flight back and like writing it and carving out all the footage and color correcting and editing and animating and all that, blah, blah, blah. And when I uploaded it,

Deep Seek had already gone through an entire news cycle of it coming out, being the talk of the town, and then suddenly is maybe not so hot. It's still in the news cycle. It's definitely still a news cycle. A lot. I need you to explain what I missed. Okay. Yes. Okay. So it turns out we all missed this, actually, because the original big Deep Seek news, well, part of the big Deep Seek news, happened on Christmas Day. So clearly a lot of us were going to miss this. Oh, wow.

That's my fault for not really digging into the forums on Christmas Day for the groundbreaking AI reports. Actually, the only things I've seen is, number one, it was an AI model that costs like a tenth of what OpenAI's models cost and are just as good. Allegedly. And then I heard that it was just trained on OpenAI stuff.

and not actually that impressive. Oh, no. They stole our stolen content. So that's my entire familiarity. All right. Yeah. High-level overview. This is an AI language model similar to OpenAI's JetGPT. Language model. Yes. It is specifically a language model. They also introduced an image model, but the language model is the thing that people are mostly talking about. Yeah.

So on Christmas Day, this company called DeepSeek, that is this Chinese AI company, released this model called DeepSeek V3. V3 is a foundation model, came out on Christmas. No one really talked about it that much because it was Christmas. The thing that is causing a lot of stir in the community is that it performs really well. It performs very similarly to things like GPT-4 and like 1.0.

one of Claude's best models, stuff like that. But probably the biggest thing that caused the stock market to start tumbling and all these things to happen is that reportedly it only cost $5.6 million to train. I think the headline I'm remembering now is that Nvidia lost like $600 billion of market cap. They lost half a trillion of market cap in one day, which is a whole Stargate.

so because of these headlines about deep seek yes wow so if you sort of extrapolate the fact that this costs 5.6 million dollars most of these ai companies basically are trying to create these gates not gates they're trying to create these uh walls on who can compete with them by saying oh it costs a hundred million dollars to train a foundation model so only these giant companies like google and openai

and I guess Claude, can train their own models to be competitive in meta. So it's semi-open source. It's like open weights. So they released this really big paper about how they trained it. And the paper basically discloses all of these different methodologies that they use

that basically take similar methodologies to what the big companies use, but they take the most efficient approach possible, right? I think a lot of the companies like OpenAI and Meta and Google,

Their big thing has been, oh, we're just going to throw more compute at it because for a while, these AI models scaled with the amount of compute that they threw at it. Eventually, they started kind of slowing down, and now they're trying to figure out different ways to make them more effective. Like, for example, OpenAI's O1 model is like this reasoning model that takes multi-steps. It does multi-step reasoning to think through problems, which has a higher level of accuracy for actually solving complex problems.

But things like that O1 model are in OpenAI's $200 a month tier. So they have all these insane gates, right, that stop people from actually using them or competing. I think I saw them saying something about how the usage of that tier was so high that they were still losing money per user. Yeah, which is...

Crazy how expensive this stuff is supposed to be. Right.

You can basically say, okay, well, I'm going to take 37 billion of these parameters and specify it towards things like code, right? Because when you feed a model a ton of data, you've got a ton of coding data. You've got a ton of just natural language data. You've got a ton of data about different topics. So it's not actually efficient to do this next word processing thing if you have to run through the entire model and find the next token.

It's more efficient to say, oh, I've got an expert on coding in the model here so that let's only run a parameter of 37 billion instead of 671 billion. Okay, sure. And that's both more efficient and more effective. So it can be cheaper, it can be faster, and it can be just as maybe even more accurate. Right. And it's also, yeah, it could be more accurate.

And it can also be much cheaper to inference. So if you're actually the end user using it, it costs less per token because it doesn't have to go through the entire model to find what you're looking for. Okay. So this V3 model was already a big deal because of how cheap it was. But the thing that's causing the bigger stir is just last week, they announced a new model called R1, not of Rabbit R1 fame. Prove it. Okay.

Can't prove it cannot prove it okay, but our one is a reasoning model similar to open a eyes Oh one model, okay, and it does thing it does the same Multi-step reasoning that the Owen model does but it does it out in the open which is very interesting So the difference is when you ask Oh one a question It'll say thinking thinking thinking thinking and then just kind of gives you the answer. Yeah. Yeah, I

That's what I do usually. Whereas the interesting thing about this R1 model is when you ask it a question, say, like, you give it some complex question like, oh, there are four blocks on top of each other. One is green, and it's in the middle. One is red, and it's here. If I were to move the green block on top, can you tell me the order of the blocks? This is an example from the Computerphile YouTube channel that I just ripped off.

Now, normally it would get those things wrong because it's like very difficult to, you know, take all that information and understand the physics and all that stuff. But in a reasoning model, it's similar to a human being who writes down the problem and you have to work it out, right? Because you can't just instantaneously come up with that answer. And so it will step by step tell you, okay, so theoretically, if you were to move that block to the top, then now the blue block is below the red block, which means that blah, blah, blah, blah, blah, blah.

And OpenAI used to hide all of that compute. But R1 shows it step by step, which is very interesting for a lot of people because now it shows you how these models are actually working, which is really crazy. I kind of like that. Yes. Because it's not like sourcing it necessarily, but it's kind of interesting to see it and maybe catch a mistake in the process as it happens. That's cool. Right. The other thing that is interesting about the way that they trained this is

is that generally the way that you train a lot of these AI models with complex problems is you would tell it the question, you would tell it all of the steps in between working out the answer, and then you would tell it the answer. And that would be the data that you feed into the model. So you're just feeding like,

all of these worked out questions into a model, which is a lot more data. And it's also not really learning how to do learning in the broader sense of the word, how to actually work out the reasoning for these questions. It's just kind of mapping. Here's the question. Here's the reasoning. Here's the answer. And so it's less likely to get those things correct. Takes a lot more data, et cetera. This time, they're basically just giving it the question.

And they're giving it the answer. And they're basically saying, you get a cookie if you get it right. And you don't get a cookie if you get it wrong. And so now they're sort of training the model to, instead of just mapping the question to the workout to the answer, they're actually kind of training it to actually do that reasoning itself. And that is way cheaper to do as well. So yeah.

The thing that is crazy about all of this is because it is so much cheaper, and they did it also on a bunch of NVIDIA H800 GPUs, which are the older, less fast GPUs that they were able to buy before the export ban because currently under the Biden administration, NVIDIA cannot sell H100s to China.

Because we want to be the... In America, they want to be the fastest, most brawniest, whatever. So they're saying, oh, we did this on H800 GPUs. We did it on like...

to 50 times less GPUs, it only took two months to train instead of three months that it took to train the OpenAI model. And the OpenAI model also had way faster GPUs. So suddenly, an order of magnitude less GPUs, 50x less GPUs you can use, NVIDIA's like, ah, maybe,

People aren't going to buy as many of these anymore. I see. That's why it related to NVIDIA. So, okay, so DeepSeek, it sounds like it's a genuinely innovative and useful, much more efficient, much less expensive both to use and to build model. Sounds like this is like the Liato Mega of...

Language models. People are considering this a Sputnik kind of moment. Sure. If by people you mean Marc Andreessen. Well, no, more than just Marc Andreessen.

In that it's like a new innovation that's not from the U.S.? In that we kind of, like, yeah, a lot of U.S. companies have been like, we're the only ones in town. And then out of nowhere, someone releases something that's better in some circumstances, and so it freaks everybody out. But what happens when you take this efficient model and put it on Amazon?

Google servers that are running like 100,000 NVIDIA GPUs? Like, would it just be that much better? So there is this paradox that I cannot remember the name of that people bring up a lot in regards to this, where if you can run something more efficiently and cheaper, more people are going to want to use it. So it scales with the efficiency. So it's like, yeah, if you have more servers, you're still going to use those servers. So, you know.

It'll be cheaper to run. I've heard the same thing. But more people are going to use it. Yeah. Yeah, there's some paradox around it. In that you just like fill the demand with your savings. Yeah. Right. The other thing is that theoretically now you could run this on like a 4090 or something. You don't need like these supercomputers anymore to be able to run these models. And so if you're like a university and you've got a small cluster, but it's not like an open AI size cluster, you could run a local version of this that's tailored to specific tasks. Yeah.

locally on your servers and you don't have to pay opening high anymore, right? Yeah so this is the other thing because the the weights are open and people are trying to like re-engineer it and It's free to use and it's free to the research paper is out there It's it's kind of open sourcing this thing that these few companies have kept closed for a long time because they want to maintain this like Monopolistic kind of leadership stance Wow. Yeah, okay

So that's the first half of the news cycle. I can see the reason for the hype and all of the movement in stock prices and all the headlines and everything. But then there was another half of it where it sort of cooled off or at least started getting...

broken down and exposed more for what is actually happening? Is that also real? Yeah, I mean, people are finding technicalities. So they said that the last run that actually trained the model is the thing that cost $5.6 billion. Yeah, billion dollars. Million. Yeah, million. $5.6 million. Big difference. Instead of from zero. Yeah, instead of from scratch. Like all the R&D hours, all the other compute, all the data collection, everything.

obviously that costs a lot more money. So it's kind of like the same thing as your bill of materials thing, right? It's like, what did this actually cost full on? They haven't released that. So it's a little bit different. Has anyone asked Carl pay? Bring it down. Yeah. Other thing that I think is very funny is they're basically saying that they are pretty, there's this other thing that you can do when you're training models called distillation, where you basically take your model and you take,

you take outputs from a good model and you use that as training data for a new model and they're thinking a lot of people are thinking right now that they used outputs from like GPT-4.0 as data to train this model because you can get you can create a much smaller model with better data and if the good data is in the outputs from another AI model and

then you can do that. And so a lot of people are freaking out. But also a lot of data. Exactly. Nice. So a lot of people are freaking out because they're like, oh my God, they might have stolen OpenAI's outputs. But it's like, boo-hoo. Yeah.

Yeah. I guess that makes sense because at first when you were saying that, I was like, okay, so we're just making another GPT-4. It's like, what's the point? But it is still cheaper and more efficient to run and all those benefits. So even if it's giving you the same outputs as GPT-4, it could have advantages. So I guess I get it. Way cheaper to run, way cheaper to train. You could train like a...

A local version. You could use your own local version. Chain of thought reasoning is really cool. Yeah. Yeah. Just a lot of things that I think a lot of people were not expecting. And then it also made it seem like, okay, maybe we don't actually need these supercomputer clusters the size of Manhattan to be able to run these things. Also, the entire nuclear industry kind of crashed on this too. Yeah.

Because people were investing in a lot of nuclear companies because they were thinking, we're going to scale up AI clusters so much, we're going to need nuclear power to run this. And now they're like, actually, you need 50 times less GPUs than you thought. That's really funny. And so now the nuclear industry is like, uh-oh. That's really funny. I don't know what to do. I mean, I am not the expert on this, but in a far enough future that I'm very...

envisioning of people using AI all the time, you have to be able to do it more efficiently. Efficiency has to be one of the focuses. We can't just keep scaling up to just bajillions of GPs everywhere and just blasting through. I mean, that's what we've been doing. That's what NVIDIA wants. That's what we have been doing. And if you follow the graph of NVIDIA's stock price, you can see how we've done that. But it seems like, yeah, efficiency also has to eventually be a focus in some of these. So I'm glad that it's at least started to

I like that. I think in most industries and markets, when more people have access to innovating on a thing, you get so much innovation that the efficiency and the value proposition just go up exponentially. But this market is so closed. You only have OpenAI, Google, not even really Apple, Claude and Meta. That's sort of it. There's a bunch of small AI companies that have spun up, but they're all basically using OpenAI as their foundation. So...

Yeah, I think this just kind of shows that open source is going to cause a lot of commotion. Fun. Yeah, pretty big deal. I'm here for it. Pretty big deal. Also, apparently, if you use the... There is now a DeepSeek app.

that you can have that is now the number one app on the App Store. Sick, that was fast. It will apparently not talk about Tiananmen Square. Oh, because it's still a Chinese app. Yeah, so is it going to get banned? There's so many things that are unanswered about this. I would not doubt it if that got banned. So, yeah. If you run it locally, it will tell you whatever you want. If you run it through the app,

which is kind of going through the Chinese filtration systems. It will not tell you whatever you want. I just want to go sign up and it says, due to large-scale malicious attacks on our servers, registration may be busy. Please wait and try again. Yeah, they're basically saying they got DDoSed. Yeah, everybody's probably... They have kind of a magnifying glass on them right now, so that doesn't...

Really shocked me, but yeah. Okay. Yeah. Things are happening. So that was a big deal. NVIDIA stock price has since kind of gone a little bit back up. There have been many think pieces about, well, it costs 5.3 million to train, but what about all the other money? And what about this? And oh, if they're using distillation of open AI data, then the actual true cost is blah, blah, blah. And it's like, okay, I feel like that's a lot of hand-waving personally. I don't know. I think it just shows that if you make something open,

Way more innovation can happen and also we're not the only ones in the AI race Alright pebble. Do you know this name pebble the word pebble bro? And it's not not the social media now pebble brings me back cuz pebbles the first smartwatch I ever had yeah, technically. Oh, well it was before motor 360s I had like a red pebble. I definitely did a video on the thing to a red pebble with a black band and

And all it was was an e-ink display on my wrist with notifications. Yeah. I didn't really want it to do too much else. It was just so I could have less screen time. I could just check my wrist and be like, oh, I can see what time it is. And then I just got a text and I'm not even going to read it. And there it is. Yeah. That's my Pebble experience. Yeah. So Pebble was one of the most successful Kickstarters of all time. At the time, it was the most. At the time, I think it was the most.

But they eventually went out of business, right? Yeah. Okay. So they raised a ton of money. They raised $10 million at their first launch when they expected to raise $100,000. They became really popular with nerds. They released a bunch of new Pebbles throughout the years. They lasted about four years. And then eventually Fitbit bought Pebble. Fitbit bought Pebble. Fitbit bought Pebble. Okay. And they shut down and –

They shut down the development of the Pebble 2 and the Pebble Time 2, and then they refunded everybody. It was a whole thing.

And then, if you remember, Google bought Fitbit. Oh, right. So there's just been a bunch of distillation of the IP of Pebble. And not a lot of people have used e-paper displays since Pebble. So that's something we didn't mention. It is a smartwatch, but it is an e-paper display smartwatch. Part of the kind of sell of Pebble is that it lasted seven days. It did very basic stuff.

It was very nerdy. It was a very open ecosystem. You could develop for it. It was this awesome community. You could see it in sunlight. You could see it in sunlight. Tiny three-pin charger. It was pretty solid. I remember that now, the battery life being so long. Yeah. Now, something you might not know is the founder of Pebble is Eric Michikowski, which is the same guy that made Beeper that we talked about a number of months ago. Beeper was that app that basically...

hooked into iMessage and kind of temporarily disrupted Apple until they shut it down. And I was completely wrong about whether or not that would last. But you know, here we are. So Eric has been apparently rallying Google for a while now to open source Pebble because they have not been doing anything with it. Yeah, seems like Google forgot. Yeah.

I bet Google bought Fitbit and forgot that they had Pebble stuff in there. Yeah, I think they also forgot they bought Fitbit, but, you know. So, yeah, he basically was able to rally Google to open source all of the Pebble stuff, and because of this, he's bringing Pebble back. And e-ink slash e-paper has gotten a lot better since 2012, 2014. Much better.

Bringing Pebble back could be kind of neat. Like, I don't know. I'm still picturing the same general premise, which is a smart roch-type device on my wrist with an e-ink display, a week-plus battery life, and, like, a decently fast enough refresh that I can get notifications, I can swipe them away, maybe I can...

control media on my wrist or something like that, like basic stuff like that, I think that would still be pretty cool. Yeah. I think that smartwatches generally are very overpowered for what a lot of people use them for anyway. And just like on a lot of smartphones, people would trade battery life for functionality a lot of the time. I love that. So Eric is basically starting a new company that is going to use the Pebble OS.

But he cannot call the company Pebble because Google still owns the trademark. Of course. So he basically put out this website with a sign-up form where you can sign up to get updates on when new Pebble stuff is coming out. Okay. So they won't call it Pebble, but they'll probably have, I'm predicting now, some really punny ripoff name from Pebble that reminds everyone that Pebble existed. Yeah. I believe the website is called RePebble. There you go.

So there you go. You know. Okay. I got on a call with Eric a couple of days ago to talk about this just to kind of do some Q&A. You will be able to build on top of the new hardware software because Pebble is going to be open sourced and everything that the new company builds on top of, on top of the software, they're putting back into the open source project. Yeah. Which is cool. Which means you can build your own hardware basically and like flash it to it. That is pretty cool. Yeah. So this is kind of like an open source device.

They are going to build new hardware, so I would love to see what that looks like in 2025 because, like you said, the technology has gotten much better. I want a Pebble in the shape of a Moto 360. Yeah, without the flat tire. Without the flat tire. Well, they had the Pebble round. They had the Pebble round, and it was basically in the shape of a Moto 360. Bigger, please. Bigger.

Bigger screen. Bigger screen. Bigger screen. Okay. Other cool thing, because it is open source, it can run on basically anything that has a microcontroller.

So if you want to flash your pebble OS to your fridge, you could probably do that. Whoa. Someone will probably do that. Someone's going to do that. Someone's going to run doom on their fridge on pebble OS for sure. Um, they are planning on sticking to the core feature set of the original pebble. Eric says that not everyone wants all of the features that you're getting from the Apple watch and all the stuff like that. So that's exciting. He put out this blog that was about, uh,

The things that he learned from having Pebble fail in the first place. And someone is working on a port right now so that you can develop on top of it. So it's pretty exciting. The more you talk about this, the more I'm picturing like a pretty simple, clean smartwatch. Like simple watch face, step tracker, sure. Yeah. Maybe heart rate. I don't want too much more than that. Like keep it.

Keep it easy. Keep it simple. Keep it like a week plus battery life. I think a lot of people would be into that. And it would probably be cheaper. Yeah. Because you don't have this huge expensive LED display. Yeah. It has to get super bright. Okay. I'm in. Yeah. I'm in. Eric had told me he was going to CES to talk to vendors. And I was like, what for? Vendors for? Now I know why. Oh. Yeah. Exactly. I'm very curious to see what it ends up looking like.

for people to build on this thing because the GitHub right now is like all written in C, which is very, to me, difficult to understand. So I wonder if there's going to be like an easier way to write for this platform. Remember the play date? Yeah. That thing is like,

The platform to develop for it was a custom software that they released that is super easy to use. Like a game builder. Yeah. So I wonder if there's going to be something like that. That'd be nice. Right now he only has four part-time employees working on this. That's all you need. That's all Google has on it too. Yeah. And they forgot that those people were still working for them.

Yeah, so I'm very interested to see if the community that loved Pebble in the first place builds this back up, if we get some normies on this thing. I think a lot of people are starting to feel a little bit overwhelmed with how much their technology is tracking them and the overkill of everything right now. We've innovated past the amount that we need. I think the reason that Pebble initially took off and had one of the most successful Kickstarters ever is going to be even more true now.

I think people will still like this. What was the reason in your head? Well, it was a way to use your phone a little less. Like you would get the notification on your wrist so you didn't have to take your phone out of your pocket. And then the phone, the notification would say like it's a text from someone or it's an email or whatever it is and then you can just put it down. And then it would just go back to showing the time and if it's simple, beautiful, shows you the time and you're not like poking around and doing all the gestures and checking your heart rate and all that stuff. If you just want a simple smartwatch, I think that's, it doesn't have to be $700 smartwatch. It could be one.

149 smartwatch and it could do all the things you want. The animations were also very nice. Yeah. Also this came out when the iPhone 4S was the newest iPhone. So back. Yeah. Throwback. That nostalgia too. Yeah I think so. It's the square one before they got round and then eventually got square again. I think the 5S was the first one I ever reviewed.

Oh, wow. The 5S. Wow. To me, that's nostalgia. Yeah. Pre, yeah. Wow. Anyway. About all for that. We'll keep an eye on it. Okay. Cool. All right. Well, we should take another quick break. We got some something or nothing coming up. But before that, trivia. Trivia. We're out to prize.

I know. Oh, yeah. Slacking. So at CES 2018, Lenovo launched a standalone headset running on Google Daydream's platform. Do you remember what it was called?

what the Lenovo headset was called? No. Q87986X. Close. You have to put a ThinkPad in front of that. The ThinkPad face. The ThinkPad Elite Facebook. Face? Facebook? Damn, I have no idea. I'll be totally guessing at that one. Yeah. Answer's at the end, though. We'll be right back.

This episode is brought to you by Indeed. When your computer breaks, you don't wait for it to magically start working again. You fix the problem. So why wait to hire the people your company desperately needs? Use Indeed's sponsored jobs to hire top talent fast. And even better, you only pay for results. There's no need to wait. Speed up your hiring with a $75 sponsored job credit at indeed.com slash podcast. Terms and conditions apply.

Welcome back. So now we are going to do a game that we call something or nothing. Let's play a game. Let's play a game. So the rules are simple. I'm going to read a headline and you guys have to tell me, is this something that we should care about or is this nothing at all? And I got you. Yeah. Nothing at all. Nothing at all. All right. First up, iPhone SE4 leak shows single camera and no dynamic island.

Is this something or is it nothing? Nothing. I'm going with nothing. It's like the one thing that's interesting about it is it usually the SE is in the body of an old iPhone that's already happened. Yeah. Which is why they can make it so cheap. It's because of the part that they make already. I agree. This one...

Slightly air quotes new design which is like oh they're making like a new part for this thing Other than that single camera does not surprise me. No dynamic rain. No dynamic island does not surprise me. Yeah, so I'm going nothing Yeah, they're keeping the dynamic island for the nicer phones. The SC is the phone I always buy my mom every seven years So I don't think she's gonna worry about the dynamic island. You guys don't think it looks beautiful. I

The dynamic island? No. This is a, I like how it looks. Yeah, there's a leaked. It might seem like a Nexus 5 almost. I feel like this is going to be Ellis' next phone. Oh, 100%. Oh, because it's small? Because it's small. It's not even that small. Yeah, I was going to say. Oh, you're right. Isn't it like a six inch screen? It kind of looks like an XR. You mean XR? What? Tenor? Tenor? The tenor, what they call it in the UK.

Yeah, I'm sticking with nothing on this one. All right. Yeah, nothing. Nothing, so we shouldn't care about it. Rumored to be an OLED. That could be cool, but still probably not that crazy. Still nothing. All right. Next up, OpenAI launches GPT-4 Government Edition. Less than nothing. Is this just going to be a repackaged GPT that's more expensive because it's government? Yeah. Yeah.

Cool. I think it's also like specifically tuned to government agencies. I think it's just ways for open AI to milk the government. Yeah. So the top comment on the Verge article is why not deep seek? It's open source.

And then the top reply is, yeah, it's going to be banned. Yeah, it'll probably be banned within a week. What's the over-under that DeepSeek is banned in a week? No way. It doesn't have the clout yet for that guy in the Oval Office to do anything about it. It took like a trillion dollars off the stock market. That's actually very fair. Totally fair. I do think that, yeah, I think they need Congress to be able to actually ban it, but I feel like they're going to ban it. Wait, so what's the over-under?

Which will be banned first, Lemonade or Deep Seek? Nobody remembers Lemonade. I got 20,000 people over there on Lemonade. That's actually true. I got 1,700 on Pixel Fed. You know what? I have the 18th most popular photo ever on Pixel Fed. Wow. Is it a good photo? What is it of? I would say so. It's of Glacier National Park.

Okay. You can't even break top 10. I'm so disappointed. Wow. Ever, Adam. I made my account two weeks ago. No, that's a flex. That's a flex. Yeah, let me flex. I respect that. Anyway, GPT-4 Government Edition. I'm going nothing. Nothing? Nothing. All right. Next up. Threads is officially getting ads. Something. Everything. Everything? This is everything. I called it. Didn't I tell you guys, like,

Obviously, Threads had their moment of exploding onto the scene. 100 million users in five days, whatever it was. The headlines were all like, this is going to overtake Twitter. This is the thing. We finally have a competitor. And it's meta, so meta's sneakily in the background like, yes. Yes, come to Threads. Meta's an ads business. But eventually, yeah, it's meta. We knew that they were eventually going to do ads, so...

We knew they were building in the background, waiting to find a good time to turn it on, and they're going to turn it on. And it's going to be another meta property with ads and what else is new. Yeah. But I'm saying it's everything. It is everything. Because this is what they do. Yeah. And now they've acquired a lot of users. And then people will migrate to Blue Sky, right? That's the next move? Come on, baby. Because Blue Sky, now think about Blue Sky. 30 million yesterday, just saying. They passed 30 million users, Blue Sky. Will Blue Sky get ads? No.

I have no idea. 100%. Because they are run by a company. Obviously, they're run on that open protocol, but they're still run by a company. Correct. Servers cost money. Servers cost money. These things cost money. These employees cost money. So does Blue Sky have to do ads as well? You know, Blue Sky doesn't even use AWS.

They run their own servers. That's expensive. It's actually cheaper than AWS, which is surprising. With their current volume? Because AWS is expensive. I guess it scales up with volume. So if your 30 million users are active on this thing, it costs a lot. And they charge you for convenience. But you still have to buy those servers and run everything. It still costs money. It costs money, but it is technically cheaper. But yeah, I have been talking to the Blue Sky crew a lot recently, and they said they are...

Thinking about the right time to move the 80 protocol into the internet engineering task force Which if you watched our hey if you watched our special episode we all just glazed over come on Secret history of the internet the IETF. They're the same guys that do email Shout out to them

They have one of the nerdiest sounding names of all time, but I appreciate that. Daddy Vint. Yeah. Daddy Internet. I was emailing with Vint this weekend. It was crazy. It was crazy. I was like, I just don't have access to it. Yeah, Vint. You mean Big V? Internet Zaddy. Our father. Anyway, yeah. So BlueSky...

Public PLC wants to eventually move the AT protocol into a standards body because they want to instill that trust. Because currently Blue Sky PLC is the same company that runs Blue Sky the platform. And the whole idea is that, you know, they are not the only one on the platform, on the protocol. So you're saying next one. Everything. Everything. It is something for sure. It is something in the way Meta moves. Do you think this will matter?

What ads yeah when ads hit there will be a another moment another reckoning for everyone on threads who has been Preaching the word of threads and I'll have to decide if they're cool with it or not Yeah, but most people haven't been preaching the word of threads They've just been on Instagram and then they're like what's this click the button and now you have a thread to count Adam You're gonna start me right again. Well, there's that movement. There's also the I'm leaving Twitter movement Yeah, and then

When it gets ads, they'll have to be like, I'm cool with that. That's not what I hated about Twitter. That's exactly my point. Everyone is used to ads already. Yeah, exactly. Maybe it's nothing. Maybe this was going to happen the whole time and it's just going to be background noise. It's everything for meta because it'll probably boost their revenue by a lot. Will it? They have WhatsApp and Instagram with 2 billion users. Yeah, but a cool 100, 200 million more isn't that bad. Yeah. Yeah, I could use another 20 bucks, I guess. Yeah.

All you need to do, Blue Sky, if you want to make all that money back, make very highly customizable profiles where you can have a pink background and a song that auto plays when you go to my profile. Nice. I love it. Next on the list is OpenAI's new operator AI agent can do things on the web for you. So did you read these articles about this? No. I did not. Okay. I have an understanding of what it sounds like it's doing based on the headline, but yeah.

Maybe, is there something deeper? Well, okay. So, Operator is this new feature that they have with the expensive model that they have. And basically, the- The $200 one? Yeah. That's crazy. Basically, it's in research preview mode right now. The idea is that you can say like, I want to go on a vacation. Can you book it for me? And it'll go to like TripAdvisor and then it'll- I still think this would be a banger video. We should do that. How?

having AI book a full trip for you. Have AI book a trip to Boston or somewhere nearby and follow the itinerary. Oh, boy. So they have, like, it's very weird. They have specific partnerships right now where it'll use specific websites and services for all of these things that it wants to do for you, which feels very Rabbit R1-y to me, if I do say so myself. What do they call those? Agents? Agents, yeah. Oh, same thing. Yeah, exactly.

It's basically an agent that goes and does things for you, but you can see it clicking around. It was funny. I believe Casey Newton did this, and it booked it for the completely wrong date. Nice. Something like that. And everyone that has been using it, that has been talking about it on Reddit, says that it's really bad. It barely works. So is it something or nothing? If it works, it's something. But if it doesn't work, or if it works as well as rabbits...

which is not at all. Which I mean, okay, it's open AI, so maybe I have to give them the benefit of the doubt, but like, I don't know. Do you? I don't know. I don't know, man. It's in research review. I am heavy, me personally, in the camp of I want to overthink each individual detail of the things that I'm doing. So instead of asking an AI agent to buy something online for me, I want to just double check it's going to the right address and that I will have the right shipping speed and all that stuff. And if I'm booking a trip, that's even more. Yeah. No way. I don't know.

I tend to think it's already easy enough to do most of these things online. The capabilities and the trust level have to merge. And both of those things, the capabilities are moving at one mile an hour right now, and the trust levels are going the other direction. So I don't know. I don't know. I think currently it's kind of...

Kind of nothing. And I also don't really think that this is the way that we're going to be using AI agents in the future. So like Marquez said, like, I think people just...

want to make their own decisions. I can just book a flight. Let me look at the flights too. Like instead of me going, hey, book a flight to Boston for me for this weekend. I want to look at all the flights, pick a time. Yeah. Which, oh, the one at noon is actually cheaper than the one at 3 p.m. So let me do that one instead. Now I'm going to plan something to do when I land at this time because it's earlier. Like I want to

I want to do that. Yeah, if you can do things like check the available flights and give me options quicker than I could do it myself, then that's cool. You know, there's probably CEOs and stuff that are like looking at the $200 a month price and they're like, yeah, like book me a flight, just do it. I don't care which one. And they'll just do it.

And they don't think about it. And then I'll get to the airport and they're like, sir, you're going to Paraguay? And it's in three months. I need your passport. Yeah. Yeah. It'll hallucinate. Yeah. I'm going with nothing for now. I'm going with nothing for now. Interesting. Okay. Next on the list. iOS 18.3 officially launches, which makes notification summaries italic.

If this ain't something, I don't know what is. That's not the official headline. That's what I wrote. Basically, Apple was getting a lot of flack because the notification summaries were saying crazy stuff and just lying a lot of the time because it was incorrect about a lot of things and trying to, especially because it was trying to group multiple things into one sentence, like multiple different events into one sentence. And topics. And now...

when it has a notification summary, it is italic to signify to the user that it's in beta or whatever. Something. Really? I also think it's something. Shows that Apple's listening to something. They just made it italic. True. Do you think... Okay, this is my question. I have not talked to a normie and been like, hey, your notification summary, is it more obvious that it's an AI guess now because it's italic? Because...

I actually hate how it looks now. The italic look is not a good look. It's like it's... So it has to just look different from a regular notification. That's the point of being italic, right? Correct. Okay, Apple...

that people have seen these AI summaries and don't like them. And if we make them look visually distinct from a normal text or a normal notification, then that's better than not. So that's that.

That is something. That's a something. I feel like it's something because this is a red flag for just Apple intelligence in general, I think. Yeah. Summarizations are supposed to be the easiest thing that AIs can do. Right. Like, hey, take this thing and summarize it. Yeah. And this isn't even the full Apple intelligence yet. Like, it's still rolling out. Still need that Siri update. So if they're having problems with this...

How do I trust them to be able to actually do a thing in three months? I really want them to do that reach into apps thing. That's supposed to be March. Yeah, I don't even know if there's problems with this. It's just not a useful app, not a useful place to deploy summaries. Like when I get a notification of like one single email,

I don't need you to summarize the email. I'll just look at the subject line. Like that's already good enough. When I get three texts. - A subject is a summary. - Yeah, when I get three texts in a chat with someone, I don't know, it's three texts. Like maybe a summary is useful, but all I really need to know is I have three texts from this person. Context in my brain is like, oh, I remember what I was talking about with them.

That's the summary. I think there's some value to like the urgency of the text. You know what I mean? Like if it says like urgent request needed, that's...

then I'm more likely to stop what I'm doing and go deal with the text than I am if it's just like this person texted you three times and I'm like I'll deal with it later you know notifications are inherently useful though like yeah we get a lot of them and they're annoying but they're there for a reason so I feel like adding this layer of friction in front of the notification to like summarize it that may or may not be accurate is not really what people want

They are funny, though. Yeah, they are funny. If I could just go and open the app or swipe down a control center or something and then choose a summary notification, then sure. But to offer it right off the bat by default, it's just another step. Apple only really added AI to a piece of the shareholders anyway. Bar. Bar. This is nothing. This is just more of the nothing. Yeah. More of nothing. Apple intelligence is nothing. You know what happened? What?

I tried to use Apple intelligence on, on my Mac when it launched on Mac and I used it on a, on an Apple note that was a whole script of a video I had written and, uh, it summarized it. And then I hit control command Z because it was like, Oh, and I'm just testing the summarization feature. And then it crashed. Oh,

And then when I reopened it, I could no longer Command Z it. No. So I lost the entire script. Damn. So that was fun. Damn. Don't test in production, my friend. Oh, my God. It's a bad thing. So I should have copied that note, but that's on me. Anyway. Is that on you? Yeah, no, it's... No, it's not on you. I'm going to blame it on Apple. That's on them. A-I-N-G.

It's bad. Okay. Next up, David did not win a ticket to the Switch 2 showcase in April and neither did Adam. Everything. Yeah. This is everything. This is everything. I agree. Switch is so hype. Yep. Switch is so hype. So hype. Dude. You had to win a ticket to look at it? Yeah. Is that what's happening? Right. Correct. And so you applied for a chance. Three chances. To look at the Switch.

at the thing that you might give them money for later. No, no, I will give them money for it. This is crazy. Let's get back clear. The hype is out of control. Dude, it is out of control. Okay, so now that you're not able to look at it, you're still going to buy it anyway. Yeah, probably. So you just wanted to look at it just to look at it. Yeah, so they basically, they were going to have three different showcase events where you could like play with it early, but they had to ticket it and they had to have a raffle.

And so they had multiple sessions for New York, Austin, and Los Angeles. Wait, is this paid? No. No. Okay. It's a free event. So this is just a hype cycle. This isn't Nintendo making more money. This is them just driving more hype. Yeah. People will show up with a phone and make a 30-second video of it on their phone and put it on YouTube and Shorts and TikTok and be like, look at the Switch. And it'll be like, oh my God, I want the Switch. And that'll be it. Yeah. Yeah. That's like in my content strategy. No, that's smart. That's smart. Yeah. Nintendo's smart. Everything. Everything.

Nintendo Switch 2, Nintendo Switch 2, Nintendo Switch 2.

And the last one. Boom's XB-1 supersonic passenger jet goes Mach 1.1. I put this in here. Yeah, so I'm assuming you think it's something. I think it's nothing. Oh, boy, this is something. The return of the, what was it called? The Concorde. The Concorde. If you have time for a tiny story time, I can tell you the story. We have nothing but time. Okay, so passenger jets. And that's it. Thanks for watching.

No, the Concorde was like that super, super fast supersonic plane that did like cross country and transatlantic flights in like no time because it was super fast. It was twice as almost... I think it was twice as fast. It was much faster than... So it was basically twice as fast as a normal passenger jet, much faster than the Speed of Sound. The Speed of Sound is like this benchmark of the obviously supersonic travel, but it's a benchmark for speed for planes. Perfect. It's like 800 miles an hour, right? So...

The Concorde that crashed, you know, sort of retired that plane and we went back to not supersonic travel. But the thing that really hit me about this is, and I remember we probably talked about this months ago, but what did you used to do on a plane before the Concorde when you had a six hour cross country or transcontinental flight?

Nothing. You just sat there. There was no internet. There was no TV. There was nothing to do. You just sat there. And so you were highly incentivized to get the faster flights so you could save more time, get to your destination, and then start doing things again. Now, if you have a six, seven, eight-hour flight, you're on the internet. You're on your phone. You can do email. You can watch a movie. There's a whole bunch of stuff you can do. So people kind of stopped caring about –

Getting there twice as fast and they have slowly gradually ramped up this like prototype of okay We can go faster and faster than they got to like just under Mach speed and they just did a test run where they went Mach 1.1 so over the speed of sound yeah their eventual goal is something like Mach 1.7 They want to be able to go twice as fast as passenger jets basically you can go anywhere in the world from anywhere in the world and

Up to 5,000 miles in like four hours. Now, it's going to be more expensive. It's the first one. It's going to be pricey. I think United bought like 15 of these things and they're going to slowly start making them. Yeah. Yeah. They got to make them though. This is still in like the testing phase of like they're okay validating we can go Mach 1.1. We still got to go faster and make sure it's safe.

But we have purchase orders, and when we start making these bajillion dollar jets, people will be able to buy a ticket and go from New York to London in three and a half hours. That's sick. Okay, but one thing. We live, we record right next to an airport. Yeah. When you go faster than the speed of sound. Boom. Big, big time noises. Yeah. Yes. That's...

Not ideal. One of the downsides. Okay, but there are certain slots of airspace that are approved for supersonic travel. So over the ocean, for example, or in the corridor for going from New York to London are certain paths where you used to be able to go supersonic. They did this testing in one of those corridors.

So there are places where it is acceptable to go supersonic. And so I think the idea of being able to get anywhere in no time and probably also still be scrolling your phone on the way, I think is pretty sick. Yeah. That's why it's something. I would love to get to the West Coast in two hours. That'd be amazing. Think about that. That'd be crazy. Think about that. The flight to California is typically five and a half hours for us. Yeah. What if it was two?

two in 20 years well when they finish making the jets and they start shipping them and the tickets come down in price because they've got a bunch of jets out there so that's 60 years I'm not getting on that first flight is what I'm saying no no no nobody's on the first flight I don't want to be none of you are allowed to be on the first flight but the idea is it will be something in 10-20 years when we are back to these super sick like the jets we fly on now are so old

Yeah. They're not even necessarily old. They're just the same technology as forever ago. And we don't really, we're more concerned with like, I mean, we're cramped back there. The leg room's not that great. The internet's not that great. The TV is not that great. Yeah. Like it's, we would like to do some upgrading. So I think it's something I'm excited. I want to give it trains.

Because it would take a whale. I have to channel my train to California. Channel my inner Ellis here and say bullet trains exist. These are different. Bullet trains exist to certain land to land destinations. It is fair. Not in the US. But not to London. Not yet. Not with that attitude. Big tunnel. Big tunnel. An undersea tunnel to Paris. A really big tunnel.

And you can put your Tesla on a little track. Yes. And then it'll go, it'll, it'll. There can be, all I'm saying is there can be tiers of travel where like there will still be regular flights like this, but there will also be, this will be an expensive ticket. Yeah. And this will be like, okay, you splurge, you take the bullet train instead of the regular train and you get there in half the time.

And it's just an available option. I have a proposition for you. We put all the rich people on the Concord. They come to you and they say, hey,

You can be the first person on this flight and get exclusive video rights and everything. No, no, no, no, no, no. I don't want to be first. Nothing. I don't want to be first. No. So they're doing these test flights. It's like a Navy pilot, and they're going in this specific corridor and slowly ramping up in speed. That's all great. Don't put me on those. Then they're going to get to Mach 1.7, and they're going to be like, okay, we can do this at this speed, and it's safe.

Now we can, it's almost like car testing. Like when you see the top speed runs from like the Hennessey Venom F5 or some random car that's like, you have to do the top speed run in both directions with all the safety equipment and the professional driver. Once you get past that stage, then you can ship them to regular people and regular drivers. And so you can put real passengers on it and not worry too much. By the way,

I'm just Tangent City. These Boeings can go faster, just like our cars can go faster. They operate at a certain speed in certain jet ways because it's most efficient for fuel. And sometimes I get on a plane. This thing is just going to say screw it and just book it. No, it's going to be able to go even faster. Oh.

But it's going to go at Mach 1.7 because that's what's safe. Mach 4, let's go. Sometimes I get on a flight and we'll be like 30 minutes late from the gate and the pilot will go, we're going to try to make up some extra time in the air. And you're like, what do you mean? Really? And then I'll feel like we're going faster and you're like, yeah, we caught a jet stream from downwind and we're actually going to get you 15 minutes early. And I'm like, you could have done that the whole time? Yeah.

So I'm like, why don't they do it the whole time? Because it's like going faster in any vehicle takes more fuel. So they have to be more efficient. So this is a plane that's going to be capable of going like whatever Mach 2 or something and it's going to go Mach 1.7 and it's going to be great. Well, I for one am excited to be one of the first to get on one of the supersonic jets from Boom Technologies. Boom Technologies.

Coming to an airport near you. Hard pass. We'll see. We'll see. Okay. That's all I got. Is that it? Yep, that's it. That is it. That means that we should figure out those trivia questions. Figure out? I think you mean...

Where are you going with this? Tell Adam the answers I already know. I feel like I have one of them. I don't know. Question one. David Kaz and Damien Henry were two French Google engineers. With their 20% innovation time off project, they made something that was introduced at Google I.O. 2014 Developers Conference. In fact, all attendees walked away with one of these things. What was it? All attendees? Yes.

Well, I wasn't there, so I can't confirm that all of them walked out with it. Dude, they would give you freaking Nexus devices. Yeah, now you got like a water bottle and a t-shirt. Yeah, disgusting. Although, pretty great water bottle. Two and a half trillion dollar company and they gave me a water bottle? The same one as last year, too. Flip them and read. What do we got?

I wrote Google Cardboard, but I added a diagram to the bottom. I also added Google Cardboard and added a diagram to the bottom. So we should go by whose diagram is better. Well, first of all, are we right? I was going to say, you are both right. Okay, good.

I have three dimensions to my diagram. Oh, Marques wins because it is VR, 3D. You need three dimensions. Hold on here. Although cardboard is 2D. 3D now, sucker. It was literally just a cardboard box you put your phone in. But you have to fold it yourself. That was what was sick about it. Yeah, it was great. It was pretty dumb. It was pretty smart. And it wasn't...

I think it was a Verizon promo or something. Somebody did it with the packaging of the box that your new Nexus device came in. What was cardboard? That's interesting. That's hilarious. Oh, yeah, that's right. Yeah, it was awesome. All right, next question. Going along the same theme. At CES 2018, Lenovo launched a standalone headset running on Google Daydream's platform. What was it called?

I don't know if you can tell, but I have complete faith that Android XR is going to be a thing. That's why I keep bringing these review questions. I'll give you a hint. It was the Lenovo Mirage what? Mirage. Mirage what? There's another word there. Mirage face headset. I don't know. Flip him and read. What do we got? The Lenovo Mirage boom. Dang it. But air. Air.

Marraige Air. Marquez just wants to fly in that airplane. I can tell. That's what I'm hearing. The correct answer was the Lenovo Marraige Solo. Cool. Wow, that's depressing. Yeah, it's a little sad. Yep.

Because you're the only one who can see the thing. You want to be more lonely? They were the only ones that were realistic about it. That's true. I'm solo. Yeah, this is the solo. So you know what you're getting into. This is like the Asus Republic of Gamers. You have a lot of friends now. That's what it's called. Republic. Republic. Democratic Republic of Gamers.

Well, I'm glad I got a point out of that. Me too. Yeah, that was fun. Hey, let us know in the comments, would you get on a flight that goes faster than the speed of sound? If the company was called Boom. If they offered you a flight, if you're on the first one, it's free. No. It's free. You're on the first plane and they go, hey, pick a destination. We'll fly you there for free. Oh my gosh. What?

I just remembered a dream I had like a few weeks ago. Was it flight related? No. Well, kind of. I had a flight related dream recently. It was about going to the moon on like a Blue Origin thing, but it was all of us. But we had to do it like individually. And when it came up to my turn, I had like the worst nerves ever and something went wrong with the system. And I was like, I'm f***ing out of here. You are not putting me on the... They're like, yeah, we had some engineering errors. It's going to be an extra 10 minutes. I was like, nah.

That's a fair choice. But Brandon was already in space. Oh, no. Oh, God. All right. Well, all our witches in the comments, please let us know. Yeah, hit us up. Tell me what the dream means. That's what we really want to know. Thanks for watching. Catch you next week. See you later. We've heard produced by Adam Malia. We are part of the Vox Media Podcast Network and our intro and outro music is by Vain. So do it. And Ellis. And Ellis.

I can hear it. You're good. I need to also change. You're good. And perfect.