The Vergecast - AI is fixing — and ruining – our photos

Episode Date: October 8, 2024

For this full-on “what is a photo” episode, we start by chatting with Halide developers Ben Sandofsky and Sebastiaan De With about what it means to build a camera app in 2024 — and what it means to try and accurately capture a photo. Then The Verge’s Allison Johnson joins the show to talk about her experiment going all-in on AI-ifying her photos. Finally, we answer a hotline about which gadgets to attach to your head when you go for a run. Further reading: Halide Halide’s Process Zero feature captures photos with no AI processing Let’s compare Apple, Google, and Samsung’s definitions of ‘a photo’ No one’s ready for this Google’s AI tool helped us add disasters and corpses to our photos The AI photo editing era is hare, and it’s every person for themselves This system can sort real pictures from AI fakes — why aren’t platforms using it? Email us at [e-mail] or call us at 866-VERGE11, we love hearing from you. Learn more about your ad choices. Visit podcastchoices.com/adchoices

Welcome to the Virchcast, the flagship podcast of knowing that the key to the camera control on the iPhone 16 is that it does whatever it's going to do when you let go of the button, not when you press the button. Total game changer. I'm your friend David Pierce, and I am doing what I guess you would call a camera audit. So I've decided for a bunch of reasons that I want to take fewer photos and videos on my phone, especially of like life memories. I think if you're doing it for a purpose, it's fine. but I just want to have these photos and not take them in a way that sort of ruins the memory of the moment, if that makes sense. And I have definitely been guilty of taking out my phone, taking a picture, and then being like, oh, while I'm here, let me check my fantasy team for 45 minutes.

And that's bad. So I'm trying to get out of that. And I've discovered, I actually have a lot of options around here. I have old phones and really any phone from the last couple of years. I have old pixels, old iPhones. I have a nothing phone. Like there's a bunch of decent camera options here. Maybe I should use one of those. I have a DJI Osmo Pocket. This thing takes great videos. I do a lot of Vergecast stuff with it, actually, and I like it quite a bit. I have an old Sony Alpha 6,000 lying around here somewhere.

I have a lot of options, and I was thinking I need to buy a new camera, some fancy new, you know, the Fuji X100 or whatever. And I probably will, let's be honest. This is just the life I've chosen. I spend all of the money on gadgets that I don't. don't have. But I think there's something interesting about repurposing one of these gadgets. I mean, hell, I have the humane AI pin and the Rabbit R1 here. Both of those are cameras. Maybe there's something too thinking about the camera I have with me in a slightly different way. Maybe it's another phone. Maybe it's an old camera. Maybe I should go just like buy a bunch of disposables. I have a

friend who does that and it seems sort of delightful. Anyway, all we are here to talk about on the show today, almost all we're here to talk about is photos. We've been talking about. We've been talking about this idea of what is a photo for years now, as smartphone cameras have gotten more and more AI-based, and the whole idea of what happens when you hit the shutter button on your camera or on your phone or wherever has shifted. The question of what you get out of it, it feels harder to know than ever, and it's a thing we've talked about a lot on this show. And so for today's episode, we're just going to dig way into that for the whole show. We're going to talk to the folks behind the super popular, super cool camera app,

Halide, maybe it's Halide. I don't know, we're actually going to sort that out when we get them on here, to talk through how they think about photos and why their new feature called Process Zero, which just removes all of Apple's processing from the iPhone, has been such a hit. The photos are worse, people really like it.

It's really fascinating. So we're going to talk to them about all that stuff. And then Allison Johnson on our team has been doing kind of an AI photo experiment on herself to see how it changes the photos, you take the way that you take them, the way that you edit them, the way that you remember them when you know that you can change just about anything about them after the fact. Then we have a hotline question that has nothing to do with the photos, but it's still very fun.

All that is coming up in just a second, but first, I just realized that if I'm going to charge this Sony camera, I have to get a micro USB cable, and that means I have to go into the dreaded bucket of cables to see if I still have a micro USB cable. The good news is I'm who I am, so I'm sure that I do. the bad news is that bucket of cables is a deeply terrifying place. Wish me luck. This is the Vergecast. We'll be right back.

Support for the show comes from Retool. Too many companies run critical operations on duct taped spreadsheets, slack workflows, and whatever else they could cobble together. Not because they want to, but because building internal tools means weeks of waiting on someone else's backlog. That's where Retool comes in. Build custom internal tools just by describing what you need.

Prompt something like, build me a revenue dashboard on our Salesforce data. And Retool actually builds it on your company's data and your cloud with enterprise security built in. Go to Retool.com slash Verchcast. We all need to retool how we build software. What's up, y'all? I'm Skyler Diggins, seven-time WMBA All-Star, Olympic gold medalist, and mom. And I'm Cassidy Hubbard, host and reporter for nearly 20 years covering the biggest names and stories in sports. And mom.

And this is Am Mom, a community for athletes, game changers, and moms of all kinds. Dropping May 14th. Tap in with us. Do you ever wonder what's in your lotion? If you look at the back of the bottle, it could contain more than a dozen ingredients. And they may not all be regulated. The threshold is so high that only 11 cosmetic ingredients have been restricted by the FDA since 1938. This week on Explain It to Me, the chemicals lurking in your cosmetics.

New episodes, Sundays, wherever you get your podcasts. Welcome back. So good news. I found a micro-usb cable. I actually found like 12 micro-usb cables. So if you ever need a micro-usb cable, get at me. I apparently have lots of extras for some reason and we'll never ever get rid of them. Anyway, let's get into it. So this app, halide, or halide, again, we're going to figure that out in a minute, has been around for a bunch of years. and it has gone through a lot of changes as the iPhone has gone through a lot of changes. Not only, you know, the tech getting better and the things about the design changing

and the lenses moving around and all that stuff, but the whole idea of what the iPhone camera is supposed to do and be feels like it has changed. And the two guys, Ben and Sebastian. Hey, I'm Ben Sandofsky, and I'm co-founder of Lux Optics, and I'm the developer half that works on Halide and Kino and our other fun side projects.

I'm Sebastian DeWith. I'm the design half. with that. So the co-co founder and designer slash photographer part of it. Have been going through all of this and trying to figure out not only how to make a really great camera app, but what kind of camera app people need now, given what the iPhone and other smartphones are becoming? We had a really fun conversation. Nelai and I grabbed Ben and Sebastian hashed all this out for a long time. Let's hear it. Excellent. All right. And Nelai Patel is also here. Hi, Neelai. I think we should just get them to pronounce

It's halide over and over again. You just did the thing that I wanted to do first, which is how in, is it halide or is it haylide? Can we solve this for people in the world right now? You just said haylight, Ben. Is this the official correct answer? I alternate because it's based on halogens, right? Oh, God. And silver halides, halogens, as long as it's not halide, which is a Middle Eastern name.

No, that's what it is now. Well, yeah. Sorry, there was actually someone who worked at the verge. John Porter, I think he has moved on now, but he wants. pronounce it on a podcast holiday and we sent him a happy holidays Christmas card and we'll never let let him live right down. We continue to do this every year because it's so good. That's very good. Hallight or Hallide depending on how fancy you feel. No, no, you guys, this is the worst. We have to pick one. Let's do this right now here on the Verstcast. Can we name your app, please? What is your name?

On three. One, two. Hallied. Halide. Okay, there we go. All right. Good. I just want to point out that like when we release. I want to point out that you super cheated on Seb Cyan first. He went on two. You never go on two. We're a very good team, actually, when it comes down to it. I just want to close this discourse on names that we were like, when we're building keynote, no one can mess up that name. And then it was featured in the keynote from Apple. And they told us, we're having trouble because it sounds a lot like keynote.

And so I think every time we build a new app, it just needs to challenge some. someone. Well, the best part was that I was like, it's actually pronounced Canoe. And then they were like, really? And I was like, no. Happy Kino, everybody. That's what we're doing now. Okay, so I think we have a bunch of stuff to talk about. And we should just, we've been getting into our feelings on this podcast about what a photo is for years now. And we have invited you to to essentially do the same thing with us. But my first question as we get into this, I want to talk about process zero and a lot about what has changed with iPhone photography in particular and all this stuff. But I am desperately curious, like when you guys started building a camera app many years ago at this point, did you know at the time how big a like philosophical exercise it was to build a camera app? I think we think about it as like you add a bunch of filters and buttons. But it's really like you are making decisions about the universe when you build a camera app in a way that I think people are only coming around to. Did you know this all those years ago when you started working on on Halide?

So I guess like the origin story I think like the first the first code commit for Haleide was 2014 for about like one day I'm going to switch three times over Oh my God All over the place So for Halliday

In 2014 It was like three lines of code And then put away until like 2016 And so I'd say that in 2016 None of this computational photography was happening It was to put it simply, a lot of these apps were just wrappers around the APIs that Apple gave you, and it was just a more pretty veneer.

And so what we were doing at the time was, and then I connected with Sebastian 2016, like we were adding value like focus peaking so that you could actually tell what was in focus, you could dial in manual values. But for the most part, until Apple released Raw up thereabout, it was that you had no real say on how the final image would look. And what's interesting is, I think when it came out, so Hallite is now seven years old. It came out in 2017. That's exactly when the iPhone 10 launched that year.

And with the iPhone 10, I think Apple started using the word smart HDR for the first time. And they started using that very famous slide where to take one photo and they kind of onion slice it into like 12 of them. And to try to show you like, oh, look, we do a lot of stuff. And there's a lot of smart things happening here. And I think that was, you know, around the same band that the pixel started, you know, talking. The pixel team started talking about like the, the pixel team started talking about like the Italian Renaissance and how we can use science to influence things. It was a very early time where people trying to figure out like, how do we get better images out of these tiny sensors? And that's

when everything exploded with like night mode and all those kind of things when it basically turned photography from just taking a snap to like a data science. And it was like, okay, if we actually just use these processors and like do a crazy amount of computation on it, we can fundamentally alter the output. And we can just, we can beat physics. We can just cheat it. Like we can do Zeps of Field. We can all these things that big. cameras can because they have giant lenses and giant sensors. We can shoot at night and we can totally take photos with like boke. It's not real, but nobody cares because you're still taking a photo at night. You're still taking a photo with what looks like a shallow depth of field. And so we were kind of emerging at the same time as that all emerging. So and how much control over that stuff do you get? Like do you look at that in those early days? It's like, oh my God, look at this incredible toolkit we suddenly have to play with all of the light that this thing is capturing instead of just doing like a thing at the end of the pipeline. Or is it like, oh, this is barely photography anymore. How do we deal with that?

How do you look at that when those tools start to come available? I'd say like around 2020 is when things started to converge of where things were getting a lot smarter. Also, we had like this existential question of why should our app exist when the first-party camera is getting so much better? And fortunately, both Sebastian and I were already shooting raw photography, which in a nutshell means that you're capturing the sensor data as a file, and then you can bring it into a third-party app like Lightroom, and you can make different decisions on how you want to process

the final image. And so around 2020, it was like, we were starting to see, like, wait a minute, like what Apple's doing is great. You tap a button, and anyone can get a good photo. Like, my parents who don't know anything about photography, tap a button, and they got a photo of their grandchild, right? But if you looked at the underlying sensor, Apple is kind of underselling, like, how much better the hardware was getting. And then over time, the sensors were getting bigger and better, the optics were getting better. And so that's when we started being like, like, wait a minute, like, how can we build a feature around this where it stands on its own? One of the questions I have about raw is what it has come to signify.

What is a raw photo? Yeah, exactly. Not quite that. But people are like, I don't like my iPhone 15 camera or I don't like my iPhone 14 camera. And then you see in the comments on social media, people say, oh, you should shoot raw. And my reaction is like, like, that is, that's a horrible suggestion. question, but all it really, because you're going to get this giant file and you actually have to process it to a look. But what it really means is, oh, there's a way to turn off this processing, right? Even inside of Apple's like pro raw system, you're going to generate a file that lets you turn off this ultra-hDR look. And that just means something to people. And the fact that the word is raw, I think is deeply fascinating. Like the word and what it means and what you get are all kind of wrapped up in a, mess. But you guys are describing, like, we're going to take the sensor data and start making

different kinds of images with it. I think we make an effort to say native raw versus pro-Raw, because what Apple calls raw in pro-Raw is not technically raw. And if you read the comments on a photography website, like Petapixel, I'm sorry. But it'll be just constantly that guy being like, you know, it's not really raw. And like, yeah, yeah, cool. Awesome. I think the difference for people who aren't as deep in the weeds as we are. What is Apple's version of pro-Raw? What is your definition of native raw? So technically, raw is you have these sensors that have voltage values from light converted to, you know, these numbers, right?

And then that's it. And then you need to run an algorithm to take on the sensor. There's this Bayer filter, which are discrete red, green, and blue values. So it needs to kind of mush them up in a blender. And it's all like, it's like basically like if you were to buy the. ingredients of making a cake. And then the JPEG is the final cake. Well, I guess pro raw is kind of like a half-baked cake.

Like you get to pick the frosting. If you like sprinkles, whatever. Raisins, I hate you. But pro-Raw is Apple's solution to what you're just talking about, where it felt like you needed a Ph.D. In image processing to take one of these native raws and make them approximate anything that Apple has. And that's partly because Apple, when they take a photo, they're actually

taking a burst of nine photos and then merging them together. So the concept of having any of these bare underlying values just doesn't exist anymore because it's all going to be squashed together to create SmartHDR to reduce noise and do all these effects. I really wish Apple hadn't said pro-Raw because it's going to haunt me for the rest of my life explaining this to people. But it's like kind of halfway between the native raw and the final JPEG. And I think what is a really important thing to say there, because I see a lot of people saying something like,

oh, like, this is just the same. Like, if you just take a raw photo, that in itself is a really fascinating turn of phrase, because if you say a raw photo, a raw file technically is not a photo yet. So what Ben said, you know, it's like you've got this sensor data, red green, two green, two blue, one blue,

one red, sub pixel, basically to make up the pixels. All that data needs to be combined. You can think of it as kind of a giant spreadsheet that just like has all these values of what the sensor caught in that moment in time from one photo. one important one photo and that's not a picture like you have to process a native raw file to get the photo out of it so per definition the rafel has to be processed for it to be displayed on a screen it actually has too much information to even be displayed on a screen and so a pro rafile is essentially already done that step it has created an image because a lot of the processing that happens on iPhone images requires it to become an rgb image a bitmap essentially and so it's a very deep image with a separate information for stuff like what Neil I mentioned, like the tone mapping, which you now have the tone

control for iPhone 16, that's separately stored in the profile game table so you can now adjust just how much that HDR effect is. You can adjust the white balance completely. You can adjust the amount of sharpening. Noise reduction is something that is baked in per se, and not just because Apple things, noise shouldn't exist.

Notably, native raw files have a lot of noise because you have a tiny sensor. You can have a lot of noise. But if you combine images, you can think of it. If you perfectly combine the same image, noise is different for every frame. So you're going to average out the noise no matter what you do if you're combining images. Behind the scenes, like when we're talking about Raws, we've been trying to call them digital negatives, which I don't know if that's any better for casual users, but it's the same concept of when you had film photography,

you'd have a negative, but then a lot of the work of great photographers would be during the development, where you would take the negative and print it on paper, you would dodge and burn it to recover the dynamic range, which is now an algorithm that Apple does for you. But there's really cool videos of, like, Ansel Adams with some of his majestic landscapes. He's kind of doing, like, analog photoshopping where he's like, okay, I got to bring back some of the sky values and bump this part up.

So this isn't particularly new, but Apple gives you like, okay, you want it on or off or somewhere in between. That's all. You can have any colors long as it's black. Like, yeah. Well, I was just kind of say, it's just fascinating to me that a lot of people say,

I see all the processing that's happening, and maybe sometimes don't even have a word for it. That's kind of like what you're alluded to, Nila, is like, people are just like taking photos. And then they take a photo with this and it's like prefer it. Even though it lacks this absolute like, I mean, literally the best teams in the world right now in photography are working at Google or working at Samsung and working at Huawei at Apple to make this processing. So do all this processing engineering. And they work from the silicon up to do like insane innovations. And then people, because they're like, ha, ha, I'll download an app and skip all that and get what often sometimes is an objectively worse image if you were to object.

judge image quality. But we find that people, yeah, people just seem to have a current preference for authenticity through imperfection, I think. And that is really interesting. I just keep coming back to this idea that like, and we've talked a bunch about this on the show, that there is this increasingly pervasive feeling among people that their iPhone camera in particular is bad. And I think, like, to your point, on any objective measure of quality, it gets better and better and better. yet people seem to like it less and less at the end. And I, like, that disconnect is so fascinating to me because, and it goes back to what you're saying about, like, the iPhone has decided that noise is bad, period, right?

Like, noise sucks. Shadows suck. They should not be allowed, and we will get rid of them at all costs. And I think, like, to your point about imperfections, like, I think that's right. And I think everybody has sort of ideas about what older photos look like. And there is a thing where, like, photos can be too good. But I also think as the hardware has gotten better, the software has gotten vastly more opinionated about how a photo is supposed to look in a way that just feels sort of incongruous. And maybe what's happening now is we're getting some of that choice back where it's like, oh, the hardware's really good.

You can do a bunch of stuff with it. Here are some options. And it seems like maybe that's what you guys have even seen as developers is some of those tools are becoming available to you to make available to other people. And that actually that's what people want now, is some choice back with all this great hardware instead of just wild new ideas about shadows. So actually, can I just dial back? I just said that process zero is rebranded raw and I can already see a thousand comments on PetaPixel. And I thank you for pointing out. Haunted by the comment section on Peta Pixel.

That's what I'm getting into you. Yeah. No, it's, I think that it, you know, but that's between me and my therapist. Now, the thing is what the value add here is that, We do have opinions about, you know, there's a million different ways that we could develop the raw. And what I think with process zero is we decided, we like noise. I have a friend who works in visual effects, and he calls noise analog dithering. And you can't remove noise without removing these details because that's what was captured at the time.

We have opinions on sharpening. Our images are a little softer by design. And if you look at, you know, you look at analog photography. Actually, last summer before we really went all in on this. I was shooting with a Canon AE1 on 35 millimeter film. And I was like, oh, it just feels so warm. Like, it's so flawed, but I love it, the fact that you can't get this crazy HDR in there.

And I think the challenge is going to be as we're building out more features and we start, you know, adding the ability like in the future, if you want to recover dynamic range, okay, at what point do we allow users to make fake-looking images, right? And it's something we'll have to face in the future. That's a big question. And at what point does something become fake looking is back to big philosophical existential questions about photos. Like where to draw that line is super messy. I think one of the things that you're touching on that well, too, is like at some point, these cameras all started looking very similar, right?

So you get your photo out of Samsung phone, of the pixel, out of the iPhone, they all look like they boost the shadows way up. And there cannot be any reduction or clipping of dynamic range. That is just considered data loss. We cannot have any noise. We've all agreed on that. And I think that's simply the result of making a camera for the greatest common denominator. Like, this is a camera needs to work for everyone. I think at some point someone was mentioning, like, oh, I use my camera for like reading

a serial number on the bottom of like a car or something, right? Like at that point, like, if you don't, if you get an artistic photo that just has like really nice Chiroskura rendering, that's really cool for you. But like, this camera literally doesn't work for me anymore. And so that's a really, really difficult position to be in. I often tell people like, wow, I would, on the one side, it's a fascinating job to design the camera at Apple. but at the same time I have to make a camera that works for my one and a half year old daughter and like my 80 year old father and myself, like who is like an incredibly pedantic, annoying photographer that frequents the petapixel comment section.

Maybe it's simply the time that like cameras do allow, and this is what apps cater to do allow people to like find her own style and shoot in a different way. And you see that now with like the very unapple like thing of like adding so much control to like what kind of like final output you get in your image. And maybe that goes down to processing too because processing, I think people are becoming. aware of how many creative decisions are being made for you. And that's fascinating to me because there was one time, like in 2020, a year of unprecedented times, if I want to like pull an anecdote, where people ran into this because they were all taking photos with their Samsung or Pixel or iPhone. And they woke up in San Francisco one day and the entire city was orange, like completely orange because there was a huge wildfire near the place where I live now. And the sun was being

filtered through it. And it was all orange. And people took a photo with it. And the craziest thing happened, the camera recognized it as being like, oh, the white balance is way off. We're going to correct for that. This doesn't really happen because it's trained on everything that's happened before it. And it has, it's literally all the smart stuff is based on precedent. And we were in 2020 living in unprecedented times. And so people started downloading like a camera app, like ours to, you know, like try to correct that. It's at manual white balance. And that, I think, was like beginning to be the turning point around when people started like exploring options like that and being aware of something like that existing.

Like they were aware that as they took the photo, like, oh, wait, it's doing something for me. And I don't like that it's doing that. And that has become a greater awareness, even in like Gen Z, even in people who don't know what a file is. They know that processing is a thing and that that increasingly is making decisions for them

to go out a long rant. But yeah. One question I have, and I've talked to you about this ever so slightly said, Apple designs the camera as a system. Samsung is a system. Google is a system. You've got the hardware, the lens,

and then you've got a processing pipeline to lean on, and they obviously know exactly what's going to happen in their processing pipeline. In the case of Apple, they're designing their own chip as part of that processing pipeline. And it feels like they can make decisions about what they can handle in processing at the expense of what the hardware can do.

And if you all are making an app that just pulls the raw data, off the sensor, Apple gets to say, well, we know there's going to be some fringing. We know there's going to be some distortion in this lens, but we're going to optimize it to collect all the light we can, because our pipeline can fix almost anything except not having the light. Has that come up for you now in the newer versions of the phone where you can see that optimization starting to shift against quality out of the lens, for lack of a better term, in favor of just light collection? Yeah. That was actually one of the craziest things I saw

when we started shooting a lot of native raw again, because for a while I was just like pro raw maxing, like just shooting on pro rau all the time. And then I was like, okay, let's go back to native raw for a second. I noticed that the output was just getting a little, was just getting a little worse. Like if I were to objectively look at it, you know, I just take a, again, to clarify,

the difference between pro rata and native raw is you get raw sensor data, which you just, I just quickly threw in lightroom to develop it, I drew in person processes that sensor data and spits it out. There's actually some corrections already baked into that. So some of the pipeline does raw data on the output, Apple iPhone and some of it, then it starts converting it to images so that it can start merging. Geometry, geometry corrections, those kind of things.

For instance, your ultra-white camera will have a lot of corrections because otherwise it looks like a crazy fish eye lens, basically. And then, yeah, we noticed a few years ago, like the main camera especially, which is the one that needs to gather the most light, you know, the camera people use for 9% of the time, it has more color fringing than it has in the past. And we, I think we talked about it as in person briefly, Neelai. I was like, no, like the crazy thing that's happening here is they just, there's a realization there.

They're like, oh, wait, it might be color fringing, but we can just fix that in the pipeline. If we get more light in, we can just correct that out, and the smart photographer will make up for what would traditionally be considered a bad, the worst photographic lens, right? It just becomes a data problem, essentially. And that is coming up a little bit,

and I wrote a little bit, I just published a review of the iPhone 16 yesterday, and I wrote, like, look, you've got to get used to processing, because gadgets in our phones, I can see him going in one direction. David, I feel like you're the pixel fold kind of guy, but like you're going to get devices that are thinner and folding

and sometimes they fold three ways now. In China, I hear the devices are folding three ways. It's going that way. We're not going to opt for thicker phones and bigger cameras. And to make cameras work that are smaller and thinner and tinier and fit into like very small headsets and rayband glasses, we're going to need more processing. because the sensors, the images aren't going to get any better.

That's just physics. So I fully believe that, you know, as we go towards slimmer and foldable iPhones or whatever, our output on something like a process zero image is probably going to get objectively worse. That's just the way it is. But old-fashioned photography in that way, that's just, I don't know how long that's going to be around on these kind of devices as they get smarter and smarter because they'll just start relying more on the fact that they're magic, hyper-powerful computers, and they're not going to rely on the fact that they're good at collecting light,

because they're just not. Wait, so let me ask you about that. Let me send you all the way down the existential rabbit hole. This is where I live. Welcome to the basement with me. At the end of this, right, the hardware is just there to collect light, and everything else happens in the processing stack. Everything else is an average of values from the past, right?

is that, do you see that all the way at the end? That it, you just spin the dial all the way on the path that we're on, and we don't really care if the sensor is any good or the lens is any good, as long as we can collect as much massive light as possible. See, I think that's a great philosophical question, because if you come down to it, you can solve it the way the Samsung moon photo situation happens. Most of the world is completely been captured before and done before.

If you train a good enough model with a bad enough sensor, you can probably take photos of 99.95% of the things out there, and they'll be really good. So, like, I don't know, that seems kind of like a reality that we're heading towards, where the processing is just simply good enough, and the reality does it. The reality of capturing images just needs to correspond

to what the internal logic can cohere together. And I personally philosophically find that a really big problem. I want to make a camera that, like, if you go to it on the, what is a photo thing? I was talking to Ben about this. I was like, we should probably think about what we consider to be a photo. And to me, it is an image

that's generated largely majority, at least by photons that you have captured at a moment in space and time. But I'm not sure if that is going to be the case. I think in the future, some cameras, maybe not all of them, but some cameras are going to increasingly rely on datasets of what the world should look like and what images should look like

and just use data, use light as data, not as like visual aid as much. What is Hallide for all the way at the end of that road? What's your job when we get all the way to the end of that? Well, I mean, like, the invention of the automobile didn't put an end to horses. It's just horses were allowed to be this fun thing. You had to look at them and ride them and stuff, right? But as far as utility for every day, you know, people stopped keeping barns, I guess. And I mean, like, look, you know, photography didn't end painting. Painting, the reason people look at paintings is because the amount of skill and craftsmanship that goes into it and being able to work within these constraints. And again, it was really eye-opening last summer shooting analog because, there is just something about an analog photo

and actually the removal of details, right? And so I don't know if anyone tried to watch one of these high frame rate movies, like The Hobbit at 48 frames per second, or that HDR, Will Smith movie, the one with his clone. What was it called? Gemini Man, there we go.

And it was ultra-high-frame-rate, H-DR, and what ends up happening with movies is when you shoot high-frame-rate, ultra-realism, it breaks the magic, because your mind isn't able to fill in the detail, And it turns out 24 frames per second with a slow shutter, like it feels more dreamlike and you're able to suspend disbelief, right? And so with photography, there's just something about a photo that is less perfect that feels magical about it. So I think there's always going to be that as an art form.

But as far as something that you're going to just tap a button and get a nice photo, you know, I think that we are smart enough that we're never going to enter that kind of battle with a world's most valuable company. Let me see this. I just put your definition in the test. Photons you collected at a moment in space and time. Right. I'd buy that. That's a pretty good definition. Does that count if you're taking nine photos and merging them? That moment has extended to nine captures, right?

Yeah. So I think so. The thing is like, let's say, you know, and the same thing goes for cleaning up. Let's say use AI cleanup to erase a part of the photo. I still think it's a photo. But once you fill that in with the majority of a cleanup feature, so like 50% or more of the pixels,

become filled in by a model, I feel like it ceases to be a photo in a meaningful way, because it has now been filled in by a data set. You guys are making a tool. It sounds very much like you're focused on the creative and artistic aspects of photography. There's a whole other conversation about, can we trust these images that exists in the world? Have you looked at the various watermarking features from C2PA that say, actually this happened, right? This photo of Donald Trump, it occurred, we're going to mark it, we're going to send it to Getty. Getty is going to know this was real.

Those photos were marked at the camera level when they went to Getty. Have you looked at doing that? Is that something you're allowed to do on iOS? That's all bullshit. Am I allowed to swear? Like, if not, like, write down. Okay, good. You are allowed to swear.

So, all right, everything that Adobe is pushing as far as, like, content authenticity is, it's just, it's like, divix and anti-dvd-d-d-d-kind of stuff in the 2000s. It's, like, if you look at it, the ability to just strip the metadata, the fact that you can just point your camera at a screen and take a photo, like, it's not going to do anything other than convince people to sign up for Creative Cloud and use Adobe products. I don't know, I don't know how much deeper you want to get in that, but I don't think it's a useful use of time. Although, they do have a verification feature where I think it does like a signature of like you can, I think kind of like C-SAM detection, like it'll get a signature of a photo,

you can drag and drop it to UI and whoever first posted it, like it'll tell you who did that. That's pretty cool, but the rest of the stuff is just a waste of time. Yeah. I'm just curious. This is right next to the creative part of how much can I change this or how much data processing can I do is the, how much can I trust this? And it feels like both of those are unsolved. Yeah. When we launched Process Zero, like I thought, I was talking to Ben, I was like, can we use the fact that we shoot a JPEC that is the process zero component and like the raw, the sensor, again, the sensor, single shot raw, Bayer data as sort of a proof of provenance, like a proof like that.

There are not a lot of data sets of generative AI, if any, I haven't found any at least, that work on raw files, right? But conceivably, you could totally do it. That's the thing. So it's not a good, you know, it'd be a fig leaf of authenticity. It would be like, look, there is a raw file. You can see that it was actually taken. But, you know, like, it's still, it's just such a losing battle.

At some point, you philosophically, as a photographer or a camera app maker or both or esteemed member of photography common sections. You just go through the same existential crisis, and you just realize, okay, there's nothing stopping this. And can I just back up to what you're saying about editing photos and where's the line that you're going to draw? So in 2016, there was a famous National Geographic photographer, Steve McCurry. And so he got into hot water because people noticed one of his award-winning photos

was actually different on his website because there's like a dude in the background who wasn't in the final award-winning photo, and people started digging into it, and basically he would use what is iOS 18's object removal feature before it existed. He did that equivalent kind of editing, and this was a huge, like, you don't do that if you're actually taking photos.

So, and I guess this is also a question for you as journalists. Like, at what point when you're doing like an interview, or you're transcribing, you're pulling quotes, at what point is it okay to remove the likes and the, and the butts, you'll be doing a lot of that with me, and thank you. But at what point are you then editorializing what you're trying to capture and telling a story that actually didn't occur? And effectively, you know, everyone has to have that level of journalistic integrity

if they want to call it a photo. So David and I'm both going to answer on three, and we're going to see if we have the same answer. Okay. I'll let David answer. My quick version of that is that's norms, that's professional. norms. That's what you're describing. Right? This guy was a professional photographer. He entered a professional

photography contest. He broke the rules. He got kicked out. They got in trouble. Your journalist, you get a bunch of quotes. You go two sideways. You're over the line. Right? Like, the journalist club if it still exists, we'll get mad at you and your

reputation will suffer. If you're a Getty photographer, there's a list of rules of things you're allowed to edit. And most of them are like, you can change exposure. And that's about it. You can do some vignette. If you look at New York Times, photos, all they're allowed to do is vignette, so they're all just vignetted to hell and back, because that's the only button they get. So, like, that's professional norms.

I don't know what consumer norms are going to be, and what we're seeing is an explosion of consumer photography. And that, I think, is really challenging. I don't know, David, what's your answer? I think about the same. I mean, there's been a big discussion this year about how people transcribe Donald Trump. There's been all this stuff about sanewashing Donald Trump, where he'll go on like a nine-minute ramble of nonsense, and then a reporter will essentially try to make it make sense,

which on the one hand is like good, useful journalism. Like, here are the words that he was trying to say. On the other hand, it makes somebody who doesn't make any sense, makes some sense. And I think it's a complicated thing. Like, at what point is my job to make something make as much sense as possible so I can help my reader understand? And at what point is my job to represent something as it was? in all its messiness. And it's like, I think it is kind of a perfect analogy

for what we're trying to do with photography. And I don't have great answers. Because it's, especially when it's messy, how un messy you're supposed to make it is really complicated. And really the question is, can you make everybody believe in the professional more? The answer is like, obviously no. So I guess the closest that we have around ethics right now

is Apple's decision that they're not going to allow fully generative AI in iOS 18. And there was the interview with Gruber where they're talking like, okay, we're going to generate cartoons, but we're not going to generate photo realistic stuff. And so that's nice when you have someone who is, you know, not enabling outright bad actions, although then you have someone like Google who's also letting you like,

yeah, just change the background, type in what you want. That does like remind me of like that amazing quote that Nelai got from Apple's John McCormack that basically was like, okay, we believe a photo is a thing that really happened, which is still not going as far as. saying like a photo is a thing you capture with a camera or is a light that is captured. It is a thing that happened, which totally fits in the model of having a generated future where there is a super great Apple model that will just basically fill in the blanks if the image

data isn't good enough. It just needs to confirm that the thing has actually happened. So it's like, yeah, like surely we will continue to have journalistic standards and there will be a few of these things like C2PA to work in their niches. But, I mean, there's no stopping the train that's heading down a track. of like the absolute apocalypse of like what the meaning the meaningfulness of photos on the internet. I truly think if I look at my daughter,

she's one and a half years old now. When she's a teenager, I don't think she will believe any single image she sees on the internet. I think the piece of this that is lost in a lot of this conversation, I think is like, we talk a lot about the sort of really high stakes photos.

And we talk a lot about like the photos of a serial number. Right. But I think like, Sebastian, to your point, you have a young kid, I think I have a young kid. Eli has young kids. Ben, I don't know. I got one. He's got hand, foot, and mouth disease right now. So, yeah.

Yeah. Good times all around. But I think, like, Sebastian, to your point, your kid will probably grow up in a world where they don't believe photos they see on the internet. I think that's probably right. But the question of, like, will they believe photos about themselves from when they were kids? Is the other part of this that I think is, like, complicated? And is such a, like, normal everyday use case we don't talk about?

in this context. But the question of like, when I take pictures of my kid, what is, what responsibility in air quotes do I have to like truth and honesty versus like taking a fun picture of my child having fun?

I feel like is such a like mainstream every single person use case that we just don't talk about enough because it feels mealy. But I don't know. And as you guys think about how light, like people download this stuff in part because they want to make beautiful art that hangs in museums

and a part because they want to take better pictures that are kids. And like what is the job there in this world? All of my son's photos are in Native Raw, so I can show him the receipts afterwards. He can develop the negatives himself. Exactly. No, but that's a really good point.

On his 18th birthday, here, son, I've made you some negatives. You know, it's important that every dad has a talk with a son about noise reduction. It's a dangerous world out there and you know how to be prepared. Yeah. But, I mean, to that point, like, some part of me feels like, you know, we're going to go back and look at, we're going to recognize this era of mobile photography for looking a certain way.

That's just a fact. And we're recognizing the first generation of Instagram photos by the heavy filters, which also means people are going to look at the first generation of Instagrammed photos and be like, oh, these are real. There's no generative of, because you can see by how bad these filters are. This was before the end times of photographic integrity, which is fascinating to me about. the way, that those fake sepia filters are going to go down in time as like a symbol of authenticity, which at the time were derided as being universally. Deeply real photos. Deeply unreal. Yeah. Deeply real photos. Must be real. And I think we we'd love to kind of in a same way, in a way, provide that niche too, because we want to offer

that extra tool that just gives you a more authentic shot because it's a bit more true to like what an old-fashioned camera might have captured. It has a bit less of that process look going on. But two, I also believe those photos are more representative of your memories because they are more like the way you remember a moment. Like when I added my photos, actually, I tend to make the sunsets warmer and the cozy times nicer and the time when I walked in Tokyo alone 3 a.m. bluer and more cyberpunkky because that's just the way it felt. And I think those aren't perfect representations of reality. And that greatest common denominator photography thing is kind of what's making people pull towards that a bit more.

And maybe also, ironically, making the pull more towards generative things or more creative ways altering photos because they're deeply unsatisfied with how perfectly photos represent reality when they're sensing reality around them but not the world that they perceive per se emotionally. All right, we've got to take a break.

And when we come back, we're going to talk to Alison Johnson about how to take pictures of your kids with AI. We'll be right back. Support for this show comes from Shopify. Every thriving, successful business has to start somewhere. A good place to start is a relatively simple question. What if, given the right tools, I've really put my all into this.

One tool that can help grow your sprouting business to new heights is Shopify. Millions of businesses around the world rely on Shopify for e-commerce. They offer a host of helpful tools you can take advantage of, from payment processing to analytics to website design. Their design studio includes hundreds of templates to help you create the exact website you've been envisioning for your business. If you're wondering, what if I need help, then no worries, because you're never left to fend for yourself. Shopify's award-winning customer support is available 24-7. It's time to turn those what-ifs into a thriving business with Shopify today.

Sign up for your $1 per month trial today at Shopify.com slash vergecast. Go to Shopify.com slash vergecast. That's Shopify.com slash vergecast. Support for the show comes from Upwork. The days of doing it all, all by yourself, are over. There's no romance and burning out while you're trying to scale. Instead, you can check out Upwork. Upwork helps grow your business by giving you fast access

to specialize talent across more than 125 categories, so you can fill skill gaps, launch projects faster, and scale without committing to full-time headcount. And finding the right talent is easy. You can browse profiles, review past work, and get help scoping the role so you can get started quickly. Seriously, you could connect with the right freelancer in just a few hours, especially when you sign up with Business Plus. Their AI-powered shortlisting pairs you with the top 1% of talent in under six hours. No endless searcher required. You can visit upwork.com right now to post your job for free.

That's upwork.com to connect with top talent ready to help your business grow. That's upw-w-rk.com. Upwork.com. Support for the show comes from LinkedIn. If you're a small business owner, you know that every hire counts, but time and resources are limited. Finding, connecting with, and screening the right candidates takes up valuable time you could be giving to your customers.

That's where LinkedIn Hiring Pro comes in. It's built to be your hiring partner, helping you find the right candidates faster. That way you can hire with confidence without turning it into another full-time job. Hiring Pro streamlines the entire process from drafting your job to short-listing candidates and conducting AI-powered interviews for initial screenings.

It's updated conversational interface lets you describe what you need in plain language. Nearly 60% of hirers find a candidate to interview within a week. With Hiring Pro, you spend less time searching and more time connecting with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward. Join the 2.7 million small businesses using LinkedIn to hire. Get started by posting your job for free at LinkedIn.com slash track. Terms and conditions apply.

All right, we're back. So the Virges of Allison Johnson has been reviewing all of these AI phone cameras for us all summer. And she decided to run an experiment on herself that I find very fun. She basically wanted to go all the way in and figure out what life is like when you just lean into the AIness of your phone. All the things it can do for you, all the ways it can change your photos, all the upgrades and cool processing things it can do. just what if you didn't have any existential crises about AI photos and cameras and just bought in? What would happen? I have been excited for weeks to hear about the results of this experiment, and Allison is here now to tell me all about it.

Allison, hello. Hello. We're back to talk about phone cameras, because that's what you and I do on this show, is just talk about our feelings about phone cameras. I love it. I'm here for it. I think that might be the official mission statement of Vergecast at this point, which is weird. We make an audio product about cameras is a thing that I never really thought about.

Let's talk about photos for a while. So we've been looking for different ways to talk about this stuff. Sometimes it's just like Nilai having an existential crisis on the Vergecast. But you went out and actually tried to live out this experiment a little bit in what the future of photography looks like. Tell me a little bit about this experiment you've been running. Yeah. So the idea was, what if I put myself in the... this headspace of I am using this phone camera to take not photos, but memories, and just sort of,

like, ran with that for a little while. I carried around the pixel 9 pro, the memory making machine that it is, and just kind of set out to like, you know, I take a thousand pictures in my kid anyway, but I was like, especially on a mission for, you know, a, you know, a about a week. So we went to this science fair. It was kind of outdoors, indoors, very kind of steam punky, cool situation. We went to the park, rode in one of those little fake stationary jeeps. He loves that. I went to Starbucks. He's a good sport and goes with me to get my coffee. So all of the usual things. I just took photos for a week. and brought them into Google Photos afterwards

and push myself to use the tools that I probably wouldn't have done anyway, like the reimagine, you know, take things out of the frame, just kind of put myself in an uncomfortable position to like try and use as much of the AI as I could. Well, so it makes me think of a conversation you and I had, this is probably almost a year ago at this point,

where we were talking about the thing that happens when you go in and edit a photo and you remember you've edited the photo and that something is sort of inexorably changed when you edit the photo. That maybe it's good, maybe it's bad, maybe it doesn't really matter,

but like somewhere in your brain, the fact that you moved stuff around inside of the photo becomes attached to that photo. And I feel like a year later, we're even deeper into that theory being the way that this works. So I'm very curious to hear how all this went. But like, give me the setup how you were thinking about approaching this experiment.

Because I feel like I want to make memories, not photos is a very, like, beautiful, high-minded thing to say that we can interrogate forever. But I'm curious if you're like, I want to actually live out this definition. And I think it's Google in particular, right, that has said a camera is for memories, not for photos. So you're like, I want to live in the universe Google imagines. Like, how do you set yourself up to do that? Yeah. It was mostly in the back half, like editing the photos. So kind of just went into Google photos on the phone.

And that's where you get the like reimagine tool and all the AI stuff. So it was like a very uncomfortable territory for me because I really don't like edit phone photos at all. I just kind of like put all my face in the computational photography process. So this was an exercise and like sitting down. and kind of, I had to kind of clear my mind and be like, okay, what, like, what did this moment feel like versus what's here? And sometimes it was obvious. It was like, while this person was in the frame and, you know, my memory is of my child in this situation, not in this person's distracting. So I would take them out. But then I, like, pushed myself beyond that to kind of be like, is there something I would add to the scene that would make it feel more. kind of like expressed how like magical or whatever it felt at the time because I think that happens a lot too when you're taking pictures of your kid or like this is such a special moment and then you go back and there's just like snot all over their nose and they look kind of

annoyed you're like oh okay we missed the mark on that one but I can imagine to all of that stuff that it makes sense that that all comes up in the in the editing process right you sit there and you're like how did I feel and how does this photo feel and I want to I want to to interrogate that because I think it's really interesting. But I can also imagine that you would start thinking differently about the kinds of photos you're taking and even like the way you're setting up to take those pictures. Right. Like I'm thinking about I was out with my son yesterday. He loves trains. And all he wants to do is there's a there's a train track and he just wants to stand there and wait for trains. And a train goes by and he yells and dances and then it leaves and

then he just goes more choo-choo and then we wait until there's another train. And we do that for hours. And I was standing there, like, when he would get really excited about a train and sort of stop paying attention to me, I would, like, back up and take a picture of him or, and I was, like, running around trying to get the angle just right and try to get the train in the exact right spot. And I had this thought of like, oh, maybe I don't need to do any of this. Like, I'm just going to stand here and just sort of point my camera in, like, his vague direction and then worry about it later. But also maybe, like, do I even need this picture in an AI? Like, I don't know. It starts to mess with your head in terms of just like when you decide to hit the shutter button in the first place.

Did you have any of that experience? I did. And it was like more forcing myself to take a photo when I normally would have been like, ah, this is really going to work. Oh, interesting. Someone's in the way or, you know, I was like, no, I'll take the picture anyway and see what we can do with AI. So that was kind of more my experience. So you're kind of like maybe there's a good picture in here somewhere.

And all I have to do is take the bad one and see what we can do later. Right. It forced me to take more shots, I guess. Yeah, I'm like a classic parent photographer where my kid was like, he was on a bridge in this little park. And I'm like waiting for the perfect moment when he's going to like come running back. And he's in, you know, framed perfectly in the middle of the bridge and like just didn't co-operative. You know, like he didn't want to do that. But I'm like, oh, well, I'll take pictures anyway and see what I can get away with.

Okay. So, and then take me through the editing process a little bit. You know, you mentioned you're not really a phone editor. I think most people aren't. I'll do occasional, like, one edit, a little bit of cropping, a little bit of straightening, maybe like messed with the saturation every now and then. But I think most people just like you take 10, you pick the best one and you put it on Instagram. Like that is the pipeline we've all been taught at this point. Yeah.

But tell me about the process of going into these photos and you open one up in this headspace of what is the memory here. This feels like a very like philosophical thing to go through every time. Yeah. Yeah, it was odd and it was uncomfortable, honestly. But yeah, just kind of taking, taking myself back in the moment and like opening up those photos. And I really had to like sit with it for a minute because I was like, well, I'd say, this is what I remember. Like, it was just like this, you know, he was standing, my husband was standing here holding our child and we were looking at something. And then to kind of like push

myself a little bit more. Like the one I came up with is like, one of us is always carrying this massive diaper bag, like 100% of the time at like a pack mule. And my husband had it, you know, the strap around his shoulder in this photo. And I was like, well, you know, I don't remember the diaper bag. Like, it's just their, like, furniture all the time. So I took that out. You sent me a bunch of the photos that you took, and I'm looking at them now trying to, like, guess which ones you're talking about. And there's one, it's your husband.

He's holding your kid in his right arm, and they're looking at, like, machinery of some kind, it looks like. Yeah. This looks like a very old printing press is kind of what they're both, like, poking at. Yeah. And all I see on your husband is T-shirt. Yeah. But you're saying there was a diaper bag, a strap of a diaper bag. Am I thinking about the right picture?

Am I guessing correctly? Yeah. Like there's a big black strap. And it's, I don't know, it's not like that distracting or anything. You look at it, you're like, yeah, that's the kind of bag you carry around if you're hauling a kid everywhere. But yeah, it kind of amaze me just how easily you can take something like that out. And it's not a challenge for generative AI.

It's just like we got it. Yeah. I'm looking at this now and there's like, there's a little bit of a wrinkle in a black line kind of going across his chest that probably wouldn't be there. But if you had never said that, I looked at a bunch of these photos being like, okay, where did she use AI? What did she did? And this was one where I had nothing. No idea. Wouldn't have ever in a million years have noticed that this wasn't in here, especially at Instagram size, right? Like even blowing this thing up as big as it goes. Exactly. I never would have noticed. So what tools do you end up using in all

this. Or you're just like in Google photos playing with the AI stuff? Like what did you find yourself gravitating to messing with these photos? I stuck with Google photos. And I, some of these photos were taken on an iPhone in a different session. Sacralage. I know. I know. Still messed with them in Google photos. And I uploaded a bunch of them to Instagram, which I also felt weird about. I felt like I was tricking my friends. But Wait, why? Tell me more. Why did that feel weird? It feels strange because, you know, like, I've changed enough of a lot of these that they're materially, like, different. You know, I'm like, that is not how that scene looked. One of them is cheesy as how there's, like, birds in the sky that were there.

That was the one I was like, I really got to try this. Okay, wait, so let me, can I just explain two that I saw that really jumped out to me? There's one that's like the 10 out of 10 on this scale, I would say, is a photo that you took of your son in a park. He's driving like a little, it looks like one of those little like stationary cars that you can just sit in. You pretty obviously just put a dinosaur in the background. Yeah. There's just a dinosaur there now. I had to throw that one in to be a little like, okay, something's up here.

Please don't believe all of this. I will also say it's not a very good dinosaur. Mm-mm. The head is very good. Yeah, the teeth. It just sort of wisps as it turns into its dinosaur body. It has like just kind of too many limbs, I think. Yeah.

But the teeth are pretty scary. But it's still, and this is like the sixth photo in a gallery, and I just flipped through. And it's like, it works. It kind of, it's fun. Like, I don't think it's real, but it does give it a fun, like Jurassic Parky kind of vibe. I'm into it. Yeah. And then the other one, which is, I think the one you were just describing, he's running through, I want to say a field,

but it's like a field next to some kind of industrial machinery. I don't know where you were for all of this, but everybody looks like they're having fun. There's an interesting backstory, yeah. And so this is one that I actually lingered on for a minute because he's running and he's got his right arm up in the air, and the sky is beautiful and blue, and there is just this big flock of birds basically right above his head.

And if it's not AI, this is a lot. this is like a perfectly reasonable thing to have photographed, right? Like it would be a lovely shot. Like I would be very proud of you for getting this photo with this dimensions in these particular ways. But then I found myself wondering like, okay, are the birds just not there at all? Were the birds like off to the side

and she just sort of like grabbed them and moved them over so that they were directly on top of his head? Or is this just a perfectly timed photo? And I think my hunch is you put the birds there, but I couldn't say that with great confidence. Yeah. Yeah, I did add the birds wholesale. There were no birds there.

No birds. The funny thing is we were at this, they called it a science fair. And outside just to the left of where you see Lennox, there's like a giant iron kind of bird sculpture that flaps its wings when you like, there's a little pulley. It was so cool. And it was just kind of like, either like the sun was hitting just right, and it was like so beautiful. And Lennox is kind of running around after we played with this thing. And he just looked super happy.

I was like, what could I add that would like make it feel like that moment felt? And I landed on putting these birds in. And it is so cheesy. And like, I don't enjoy it. Like, I look at it. I'm like, you know what? The birds are convincing. And this is like a pretty photo, but am I going to like print it and put it on the wall?

Like, no, I can't. Well, and that, I mean, this is where you get back to the like, what is the, what is the memory piece of this? Because I think there is some part of this as a viewer that I'm immediately like, oh, what a fun, beautiful day outdoors. Right. And that's like you're communicating to me, I think, the correct thing. I don't know how much the birds really add to that. Like, it's, it's, it's, it's, it's, it's, smiling and running with his arms up that does most of the work in this photo of, like,

communicating the feeling. But putting birds there certainly helps, and it makes the photo cooler to look at. But it doesn't sound like it accomplishes anything for you. And I find that really fascinating. Like, as the person who is likely to look at this photo the most over time, it doesn't sound like this improved, like, you're not going to look at this with fonder memories because of the AI. birds you put there. Yeah. And it really, it really made me think about the whole exercise.

It's like, who's this for? Is this, would I make these edits for me? So it felt more like what it was in my mind or am I trying to portray something on social media that's like, looks more idyllic. And I kept coming back to like, we have kind of a set of expectations from like professional photography. Like if we had hired a professional photography, like, if we had hired a professional, professional portrait photographer to kind of follow us around for an afternoon. You would sort of expect, you like suspend your disbelief a little bit. Like, you know, they know how to get the right moment and the right light and like your kids smiling and not screaming and all that stuff. And I think this like one way that these AI tools are maybe intended to be used is to kind of

achieve more of that, like, without being a professional photographer. And that's where I was, like, iffy on it. I was like, I don't think this quite works because there's just too much that you can't, like, recreate afterward. It has to be there. And, like, the photos I like are, yeah, the things were just there. Like, Lennox is having a great time and the light was right. And that's what makes it a good photo, I think, or, like, a photo I want to keep looking at, not AI. birds. To keep on that photo, I think the thing about what you just said that I find very compelling is the idea that like a professional photographer would know how to get the moment when they're not screaming. And like your memory of this day is, it was a very fun day. We all had a really good time,

right? Like I bet Lennox at one point made a sad face or like, or like had a bunch of snot everywhere, right? Like these things happen. And so I look at this photo and I wonder, okay, what if instead of adding birds, you had just opened his eyes or you had just wiped away the snot on his face. Or like, I'm thinking about my own kid right now who has a, he's, he's an 18-month-old boy, so his face is a disaster all the time. And right now he just has this like little scratch right on the side of his face. And I'm like, what if all I did was just get rid of that? Just sort of yank that away.

And my memory of it is not the scratch on his face, right? My memory of it is we all had a really nice day, not that he was very snod. for two-thirds of it. And I wonder, like, to me, there's a very stark line between, like, I added a dinosaur and I wiped away the snot. But I guess somewhere between those things, it's very blurry. But did you do any of the, like, tiny AI cleanup rather than kind of using the tools to their full extent? Did that feel any different? Well, so the generative AI in Google Photos right now will not let you do anything to a person. Oh, that's right. And that's, that occurred to me a lot. I was like, well, you know, he looks kind of bummed in this picture.

Like, would I want to change that? And the closest you can get is with best take. So I did, like, take firsts of photos and then you can kind of pick, you know, the expression you want afterward. It didn't, it's not super effective with just one subject because you could just pick that photo rather than the other. It makes a lot more sense if you have multiple people in the scene. But yeah, that was the weird area. I was like, I don't know. I think I would, like, clean up some buggers and not feel too weird about it. In any of that, you get to, like, well, where do I stop doing this? You know, once you've, like, cleaned up buggers, you're like, well, what else can I do now? Right.

It's a slippery slope to adding dinosaurs. Like, it literally is. That's exactly where you end up. You end up in dinosaurs. Yeah. No, it was just a strange thing of like I would start taking something out of the back. Like there's a stroller in the background and I erased that. I'm like, well, what else can I take out here? And then the thing that is so wild to me, you just end up with this like bland picture. I'm like, I took out all of the clutter and all of the stuff in the background and it just looks, I'm like, well, where were we? You know? Some of that stuff adds the context and the kind of, of like, we were really here doing this thing kind of feeling. And that's what really struck me in this experiment is like, oh, you couldn't just AI your way into a picture that just could be anywhere.

Just looks lovely. Totally. Yeah. And there is a weird thing that you see in a lot of sort of overly doctored photos like this where everything kind of happens in a vacuum, right? Like there's so many photos that if you remove people from the photo, it's like a beach. is a perfect example, right? Like, if you're at the beach and you're the only person there, it's actually weird, right? Like a nice beachy day. So then you, but you go in and you do the magic

eraser and you get rid of all the stuff. And now it looks like the rapture has happened and you're the only person left on earth at the beach. And there's just, it's a strange thing to try to dial in. And there was a, there was a photo in the ones that you shared where Lennox is like, he's playing with a truck on a table and there's like a glass of milk in front of him and there's a blurry person in the background. And this was a, is another one that I struggled to figure out what, if anything, you had done with AI. And I'm curious what the answer is. But I was looking at this and I was like, okay, my immediate instinct would be to remove the glass of milk because it's only like two thirds in frame and it's not really

part of it. So it's just a little bit distracting. And then like, okay, maybe I'll get rid of the person in the background. And now I've turned this kind of nice photo of a person living in the world into like a kid at a table in an empty warehouse with nothing. And like, I think I've made the photo worse. Yeah, yeah. That's a conclusion I came to a lot. And actually, so the glass of milk is not real. Oh, really? Yeah, it is a I. Wow. That is a convincing glass of milk. Isn't it? We were at the, we were at a Starbucks, but a fancy Starbucks, you know, the ones where they would, like, roast the coffee in front of you. Oh, yeah.

He had a little orange juice bottle, which is, I don't know, how you feel about being a parent on Instagram is, like, loaded with sugar. And I'm like, well, what if I put something in front of him that wasn't so overtly unhealthy, quote unquote? I don't know. I don't care if my kid drinks orange juice. I think orange shoes is basically fine. It's so good. Anyway. Don't cancel us, internet.

I know. Bad opinions about orange juice. Yeah, I was like, well, I'll just put a glass of milk in front of him. Like, he ordered it at Starbucks. Yeah, and it's super convincing. I was like, well, that's wild. And yeah, you can go like one of the pictures where he's in that little Jeep in the park, the one that doesn't have any dinosaurs just has a park behind him. I actually took out like there's like this building that has the bathrooms.

It's like brick and it doesn't look like offensive. It was just kind of in the background. I was like, well, maybe I can just put some trees there. And it's a thousand percent convincing. But then it's also like, it just looks like some traits. Like it doesn't tell me about anything about like where we were. Yeah. It does.

The picture becomes very kind of generic in that way. Yeah. One thing I find myself wondering about this process is how you'll think about these photos differently over time. A, whether you'll go back to the photo of Lennox with the birds in six months a year, five years, or 50 years. and remember that those birds weren't there, or if you'll forget and think they're part of the photos and maybe that's success in a certain way. But also, whether changing the photos

changes how you'll remember them in a way that feels weird. It's like so much of what we talk about with photography is, it's a documentation of a thing. And if you change the thing, you're going to change your memory of the thing. And that all gets very hairy and philosophical. But even just like removing people from something is going to change how you remember. remember that experience when you go back to the photo.

I don't know. Are you thinking about that at all as you're going through being like, how do I make this photo what I remembered versus what I want to remember it as many, many years from now or how I will remember it? Yeah. I like personally, my Google Photos is my journal at this point. Like I haven't written in a journal in a long time. If I'm trying to remember, like, where did we go when we were in Portland 10 years ago?

I go to Google Photos, and that's where the answer is. So I did come across, you know, some examples of, like, I was taking a friend, like a friend of ours was kind of half in the frame, and I had a picture of Lennox and my husband, and I was like, well, I'll take him our friend out of the picture, and it just fills it in, you know, perfectly. I would never think there was another human there. I was like, well, am I going to remember that we were there with our friends on that day? Or am I going, is that like lost to time if I take him out of this photo? I started feeling really strange about it. And I undid it. Yeah, there's a little bit of like, are you lying to your journal in all of that, which is, which is very strange. Yeah. Oh,

which I never lied to my journal. I would never do that. No, who would? Yeah. And then figuring out where that scale begins and ends. It's kind of the same thing. It's that same slippery slope, right? Where you're saying, okay, is remembering that my friends were there an important part of this memory? Sometimes, sometimes probably not. But sometimes, yes, is remembering that, you know, whatever, he had a runny nose. Definitely not.

But, like, those are, it's just strange to think you're making those decisions now also about the future. Like, maybe, maybe you will want to remember. remember that you were there with your friends. Maybe it doesn't seem so significant now, but it will later. And those are just making that decision now is an interesting way to like, you're changing the way that future you remembers this without any real knowledge of how it might do it. Right. It's uncomfortable. It is. Uncomfortable is exactly the right word for it. Like even when it's useful, it still feels kind of uncomfortable. Yeah, exactly. Did you have any of these that you you came out of it and felt better about?

Like, I keep thinking about that thing you said at the beginning where you were like, I took photos. I wouldn't have taken otherwise because I'm like, I bet there is a good photo in here somewhere. I find that very compelling. Did you have any of those experiences? Anything where you came out of it and you were like, I have made this into something out of nothing? Yes and no.

I have only reinforced my belief that you cannot edit a bad photo into a good photo. Even with AI? Yeah. That's true of Photoshop. That's true. of putting birds and, you know, AI birds in your photo. Like, nothing's going to, I tried. I tried to make some just bad photos look better and they're still stinkers.

But the thing I am convinced now is I will take more photos where there's someone in the frame and I kind of wish they would move out of the way, but, you know, you can't ask them to move. Right. Or I'll take more of those photos. is now and just remove them from the background because I don't feel bad about that. And especially if it's like someone else's kid is in the background and it's a picture, you know, I can't get my kid to stay still forever. So I take the picture anyway and then I can remove the other kids so I'm not putting them on my Instagram. I'm like, I actually feel fine about that. Yeah. So that's where I land.

That feels like a good use case. And especially, like, the thing that immediately comes to mind there is, like, whenever, if you go to the zoo or whatever, you're inevitably going to take pictures of something with a thousand strangers right on all sides of it. Yeah. And you can make a case that actually not only you're making your photos better, you're kind of doing like a public service to the world by removing those people from the photos. Like, you're fighting the surveillance state with AI. Right. I love that. Not today. Exactly.

So how do you feel at the end of this experiment? Has it, you're, like you said, a person who takes a lot of photos anyway. Do you take anything away from this that makes you think you might act or operate differently with a camera from now on? I think the big one is I might rely on it more for those photos when I'm like, well, I couldn't reframe this. And there's someone here. and I just don't want them in my photo. It is very good at that. Sometimes it takes a couple of tries to get, like, the best version.

But, yeah, I was super surprised, even with, like, kind of a complicated background or things in the foreground. I think that's where Magic Eraser hasn't been good in the past and where generative AI actually is, like, makes it better. It isn't just, like, a gimmick. But there are also gimmicks. And like, I remain unconvinced that anybody was asking for the ability to, like, put dinosaurs in their photos. And, like, it was kind of funny for a minute. But, yeah, I'm going to, I'm going to stay away from the dinosaurs, I think. Yeah, I mean, and I think it goes back to that question of who is it for, right?

Because I think there's a world in which you can make the case that this stuff is, it's fun to share with people, right? like people who weren't there, people who care about you and your kid and like it's silly and fun in the way that, like, I think about the pictures you get at Disney World on a roller coaster, right, where they pick this one very specific moment. And it always has a silly background. And like, no one believes that that is like a representative photo of your life. Right. But it is, that to me is actually a perfect example of like, that is a photograph of a memory and not of a, it's not a photo.

Even though in a way it is a photo, it's so. exactly staged and put together and directed and surrounded with a frame that like they make a memory out of that in a way that I think is mostly cool and fine. Yeah. But there is still something different about it when you're doing it on your phone and your camera roll in these like very mundane things. Yeah. Like taking out the background of a shot of a park to remove a building is like not the same. same thing as the roller coaster shot at Disney World in a way that I'm still having trouble putting my finger on. It's not the same. Yeah. No, I agree. I think that's a good analogy.

So what has this given you in terms of new perspective on the what is a photo debate? Like, do you buy Google's memories, not photos thesis a little more, even if it doesn't quite pan out in reality? I, you know, I don't know. I don't want to give them too much credit. I do in a way, You know, the thing that struck me is really, it's such a different approach than the iPhone 16 camera, which is more about, you know, the new photographic styles and kind of their response to what is a photo is like, we'll give you the tools to dial in the color and the skin tones the way you see it, which is kind of, they're answering the same question. And it's not like, we're going to reproduce this thing the way it absolutely was. It's sort of putting it, giving you more input over like, no, it felt more like this. Or it looked more, it looked warmer or cooler. Adding dinosaurs to your photos is not a thing, you know, you can do with the iPhone camera presently.

But it kind of struck me how they're both kind of turning from this, like, you know, push to like, represent reality as faithful as they can and like do it with as many, you know, as much data as they can cram into a single image. And it's sort of turning to like, no, you make the call on this actually. Yeah. There's a real like theory of relativity thing happening in photography that is maybe true. Like it has, I think it has never been true that cameras perfectly captured exactly everything right in the real world. Like, Like, it's just never been true. It's just that now the tools with which you can make it more real are the same tools with which you can make it less real.

And I think you can't have one without the other. Right. Like if you give me the tools, I can inevitably use them in both directions. And we have to decide if we're okay with that, I think. Yeah. And like the magic eraser thing is really interesting because this tech is getting better like shockingly fast. And I think is probably going to keep getting better shockingly fast for a pretty long time. So, like, the capabilities are not getting worse anytime soon.

Yeah. All right. Before we go, give me the single worst example you had. Was there a photo that you made disastrously unusable, would never share with another soul in this project? Let's see. I tried to do a bunch of my vacate. We went to Iceland last year. And there's literally the waterfall that is in the magic eraser, like, promotion.

They're like, it's called like Scogafas or something. Like, I'm not pronouncing it right. But they're like, look at this beautiful waterfall. You can take all the people out. And all of a sudden, you had the waterfall yourself. I did that. And it's like the most boring friggin' photo of a waterfall. I'm like, what have I done here?

Like, I'm not, no. I put the people back in. This is why Samsung makes the moon beautiful. So you don't have to worry about it. There's just like, you want the moon? Here's the moon. Yeah. You can have it.

I put the people back in is a pretty good. That's like a memoir title for you when you write about what is a photo. I'm writing that down right now. All right. Allison, thank you as always. Thank you. All right, we've got to take one more break, and then we're going to come back, take a question from the Vergecast hotline. We'll be right back.

Support for the show comes from Anthropic. Not every question has an easy answer. And the ones that are really worth asking usually come with a healthy mix of inspiration and backpedal. AHA moments and quiet meditation. When you're working through one of those problems, you want a partner to bounce ideas off of and figure out where the deeper issue lies. That's where Claude can help.

Claude is the AI for minds that don't stop at good enough. It's the collaborator that actually understands your entire workflow and thinks with you, whether you're debugging code at midnight or strategizing your next business move. Claude extends your thinking to tackle the problems that matter. Plus, Claude's research capabilities go deeper than basic search.

It can have comprehensive, reliable analysis with proper citations, turning hours of research into minutes. Ready to tackle bigger problems? Get started with Claude today at cloud.aI slash vergecast. That's clod.aI slash vergecast and check out Claude Pro, which includes access to all the features mentioned in today's episode. Claude.a.ai slash Virchcast. Buzzwords like progressive and affordability

are thrown around all the time in politics. But what do they actually mean? For me, being a progressive means at least two things. One, being willing to unite lots and lots of people, all of the folks that are getting screwed over against the powers that be that are making your life worse. And then second, being progressive is essentially a hopeful enterprise,

that you think, I think that the world can be much better, that we don't have to settle for crumbs or settle for the status quo. And is there a difference between what it means to the elected officials and what it means to the people? So money is essentially the root of everything. I don't care if you're gay. I don't care if you have all that.

That's like secondary, third. Like that doesn't, that's not a priority. That's this week on America Actually. Let's begin. Complex and unprecedented. The Spanish authorities are calling it. Before the disembarko, asymptomatikas. Passengers who'd been stuck aboard the Hanta or maybe Hanta virus-stricken Dutch cruise ship

disembarked in the Canary Islands this weekend, prompting the highest stakes game of where are they now since maybe COVID? Some of the evacuees, American and French, have since tested positive for the virus. And yet public health officials seem remarkably calm. We do have one individual who was taken to the biocontainment unit early, early this morning. and we assess that individual. They are doing well.

Possibly because this is not the one to freak out over. Today, Explain drops every weekday afternoon. All right, we're back. Let's get to the hotline. As always, the number 866 Verge11, the email, vergecast at theverge.com. We love all your questions. We try to answer at least one on this show every single week.

Thank you, again, to everybody who reaches out. They all get funneled into a Slack room that more and more people keep requesting access to because it's so fun to hear all of your awesome questions. So thank you to everybody who reaches out. This week, we have a question. God help me about running. Hello, Bridgecast. This is Zach from Las Vegas. I have a few hundred dollars that I wanted to spend on a new gadget. And I've been looking at the PixelBuds Pro 2. I have the original pros, and I'm mostly happy with them. But when I take them running,

they often aren't very secure. And I thought the Pro 2s would fix that. However, I've recently been intrigued with the meta ray-band glasses. I think they'd work well on my runs, incorporating sunglasses and speakers for music. And it's a new fun class of tech that I don't have. However, I'm all in on the Google ecosystem and pixel products, and I didn't want to add a new assistant or AI to my life. If you were in my position, which would you purchase? Thank you.

All right, V Song is here to help me. Hi, V. Hi. I empathize so much. Right, you know why you're here for this, right? I know. I know. Yeah.

Okay, so I want to reframe this question very slightly because I think I don't want to spend a ton of time litigating the PixelBuds Pro specifically, in part just because everybody's ears are different, right? And the only actually good advice we can give is try them. Buy them both, try and see what happens. But I do think the question of, can the Rayban meta smart glasses be a kind of multi-purpose useful exercise thing is a question we've actually gotten a bunch of times. So I want to talk about that. And I also want to throw in, you and I have both tried and our fans of the Shocks Open Runs. So I think in those three categories,

you have the headphones that go in your ears, you have the bone conduction headphones, and you have smart glasses. You're literally in running gear right now about to go running. My first question is, what are you going to put in or on or near your ears when you run today?

Today when I run, I'm going to be wearing the Shocks Open Run Pro 2. I genuinely love them. I used to be a big Beets Fit Pro Gurley, and then I moved to the suburbs, and the range rovers here are unhinged. It's a safety thing.

I really need situational awareness, so I've been wearing these, and I just love them compared to the previous versions of Aftershocks because the base is decent. Like, if you really love bass, this ain't going to do it for you, but compared to other bone conduction headphones, the bass is quite good.

The sound quality for bone conduction is quite good, and it's because it's using, some like regular air conduction speakers on top of the bone conduction. So the base is the air conduction and the mids in trouble are bone conduction. Works great. Really love it. Pretty secure. It's not falling off. But to the point of the question, it's not great if you want to wear glasses on top of it. Like everyone's faces are different. Everyone's ear capacity is different. I have pretty narrow ear two head space, I guess what you would say. So when I don't feel like wearing contacts when I'm running and I put the open, uh, the shocks in and I have my glasses on or my sunglasses on,

is very crowded, uh, in the airspace. And that is like, I think my biggest complaint of those so far. So the meta ray bands are fucking great. If you just want to consolidate two things. And like, I'm, I did several miles this summer, um, wearing them. meta ray bands as as like a two-in-one. It's both my sunglasses and it's my form of audio. Like, it's actually pretty great for that. If you're running in a relatively quiet area or you're not like by like a highway, so you're by a highway, it's just not going to, it's not going to do that for you. Yeah. In terms of like running in loud environments, it's in your headphones of the move. And then honestly, I would say bone conduction is second.

and then like big gap meta rayband smart glasses. Like those things, if there's a truck idling that you run by, like your toast, you're not going to be able to hear anything. You're toast. And also, if you are someone who runs races, so I ran New York half marathon a couple years ago with the Bose tempo. Loved the Bose tempo. It was great during all throughout my training.

I was like, this is great. I'm going to do this race day because the general rule for runners is nothing new on race day. You want to have practiced everything beforehand. went, they are blasting music at these races. Critical failure on my part, because I have my very specific running playlist. It is highly curated. And I was just like, I do not jam with these jams. And it really kind of ruined my vibe.

So, like, in that sense, they're not great. So it really depends on where you're running. The other issue I have with the meta-ray bands is that they are heavier. That was my question. my only real worry in recommending these to a lot of people is just that they are large, like physically heavy sunglasses. And I am not a consistent enough runner to weigh in on this, which is the thing I was most curious about from you.

But like, you've complained about how heavy they are when you're not trying to run fast and long. How do they feel in a real like run scenario? I didn't say if you're doing like 5K or less, you're really not going to nervous too much. But like the more you run, the sweatier you are. especially if it's in the summer, especially if you're wearing a hat. So like, like I said, I wore them a lot from my mileage during the summer. And, you know, I'd have a little hat to protect me because the sun is deadly. And I have the ray bands. And like, I'm running in 80 plus degree weather, high humidity in Jersey. My face gets sweaty. sunscreen. It's just like melting

down my face. These things. And so, like, you know, these glasses are pretty good if you have a low nose bridge, I do. But when you have sweat and sunscreen, they can only slip a little bit. They're going to slip a little bit. That's just like inevitable because you are in an up and down running motion. If you're doing like really long mileage, I want to say like if you're training for a half or you're training for a marathon, you kind of got to deal with the fact that they are heavier battery life. Like it's hard to say because you could be a freaking cheetah and a half marathon is like an hour and a half for you. you so like whatever, it'll be fine for your long runs. Or you could be slow like me, where you're just like, you're like turtle steady does it and you're running like two and a half hour to three hour half marathons and you're just like, this is this battery life. It should, it should work if you're fully charged, but whom among us always has a fully charged headphone when we run out the door. So like it is you play a little bit with fire there. So yeah, like unfortunately you're going to have to try a couple of things.

But what it sounds like you're saying is that if you're just a person who just sort of casually runs for exercise, you just want to get out in the morning and run a couple miles to get the blood flowing, all those concerns you just described probably don't really apply. So you should mostly be fine with the raybans, right? Yeah, you should mostly be fine so long as you don't. Like you said, you're not running by highways and like a lot of construction because you ain't going to hear nothing.

Like I have done that and you're not going to hear anything. You know, obviously, fit is very personal. So before you buy, I would go to a store and see how they fit on your base. But I actually really enjoyed running with them the summers, particularly when it was really sunny. And I don't want to have multiple things sitting in my ear. It's actually very smart. It's a very smart option. But again, loud environments, races, you're probably not going to have a good time going, especially if it's a big race. It might not be a problem if it's a smaller race and they're not blasting like somebody else's playlist. But for me, like at a big race, when I wore the bow's tempo, I was like, I've made a huge mistake because I had to listen to like Pitbull and Pitbull doesn't get me going for a run. No one needs that for a half marathon. The thing I keep thinking about with all of this is just that we can't be that far away from like the wraparound Oakley's smart glasses. Like that is so clearly the next one of these you would do. if you're meta. Everybody's into health and fitness

is a wearable thing. It's exactly for the people you're describing, like, somebody is going to make those like rainbow-tinted wraparound Oakley's with the crokeys on the back that you can just secure

to your head and it's going to be Ray-Ban meta-glasses and they are going to sell so many of those. And that becomes the answer to this question because then it's like

you're going to look like a doofus but it's going to rule and you're going to win all your races. So congratulations. That was the Bose fit tempo. They already did it. That's true. Bose just didn't manage that side of the business well, but you can see some of my pictures from my first New York City half marathon.

I am wearing the Bose fit tempo. I look like a freaking jabroney. And, you know, if they could have beaten the pit bull and the pop music that I am not a fan of just blasting, it would have been perfect. As it was, I was just like, I can't hear my K-pop. I have no motivation. This is terrible. So those are all things to consider if you really want to go with the Meta Rade Dance for running. All right. What's on the running playlist right now? A lot of straight hits because as far as K-pop goes, they're not very bubble gum.

They're kind of like pots and pans crazy electronic music. And it's just like, I'm like, yeah, there's a beat. I can go to that. I hate everything right now, but I can go. I can go.

A lot of times I run to K-pop because I'm not super fluent in Korea. So I can just be like ignoring what they're saying. And it's just like, yeah, they're just saying you can do it. That's great. I love this for me. Yeah, I'm a very, I need a lot of bass and a lot of when I run. Otherwise, like, I don't know how people listen to podcasts or audiobooks because I'll just stop.

There's nothing like. It's so boring. Yeah. So it's like, I could just listen to this audio book on a couch. I don't need to be suffering while I run. So, yeah, a lot of straight kids, Shabuzi and Duckworth, they have the song called Start a Riot. And I'm like, they're like, holy here, going to start a riot. And I'll run.

I'll be like, me, me, I'm starting the riot up this hill. So, yeah. I love that. All right. Well, I think my recommendation for this question is start with the Raybans, A, because they're a really fun gadget. And if you just have a few hundred dollars you want to blow on a gadget, it's a really good one to mess with. They're super fun. You'll enjoy them.

take them on a run and a half, and you will immediately know whether they work for you or not, and then you can try the other things. But I think if the, if the Rayvans work for you, that feels like the most fun outcome here. So I would say start there.

That is the most fun. And if those don't work, highly recommend the shocks just because, like, if you look at any running influencer on TikTok right now, they all got shocks on their year. They all got that exact thing.

They all love it. And I'm like, yeah, I do too. I was pleasantly surprised by the latest version of the Open Run Pro. I've started using them for runs, for dog walks, for, like, pushing the stroller to go pick up my kid. And now it's like, you know the thing when you get a car and all of a sudden you start seeing the car that you have everywhere. That has been my experience with these. They are everywhere.

And everyone I see who is, like, way too good at running, those people who, like, you know, if you, like, met them, they would talk to you about how much they like running. They all wear them. And I feel like it makes me hate myself and them. But it's also, like, clearly there's something here. Yeah, they're like literally, I have been seeing every single running influencer just like upgrading to them and going like, oh my God. And they have the same opinions that I had. So I was like, oh, yeah. Like it's a real thing.

And they're sweat. They're very, I got caught in a rainy run the other day, four miles of just like rain. And they worked great. I would be less confident that the meta ray bands would survive a four mile run and like pretty, pretty decent. Hurricane Helene. leftover rain shower. So like you got to kind of like think about that as well. Fair enough.

All right. Well, I hope that helps. V. Thank you as always. Sure. I'm going to go suffer now and run. Godspeed. All right.

All right. That is it for the Vergecast today. Thank you to everyone who came on the show. And thank you as always for listening. There's lots more on everything we talked about, including Allison's great story about her memories, all of our coverage of howlied, all of our phone reviews, many, many thoughts about AI. Nilai did a really fun thing comparing the different cameras on the different smartphones and how they use AI to do stuff.

Lots of that. On the verge.com. I'll put a lot of links in the show notes, but, you know, read the website. It's good website. And as always, if you have thoughts, questions, feelings, or other cameras you think I should buy, you can always email us at vergecast at the verge.com or call the hotline. 866, Verge 1-1. We really, really, really love hearing from you.

This show is produced by Liam James, Willpore, and Eric Gomez. The Vergecast is Verge production and part of the Vox Media Podcast Network. Neil and I will be back on Friday to talk about all kinds of news going on this week. There's more AI news. There's some interesting gadgety stuff going on. We might get some Apple stuff soon. Lots to talk about. We'll see you then.

Rock and roll.