The Vergecast - AI assistants are so back

Starting point is 00:00:00 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. Prompts something like,

Starting point is 00:00:22 Build me a revenue dashboard on our Salesforce data. And Retool actually builds it on your company's data, in 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.

Starting point is 00:00:59 dropping May 14th. Tap in with us. Hello and welcome to the Virchcast, the flagship podcast, with multimodal search agents. They can kill you in a number of ways. Knives, guns, they're here. They're agents, and they can do anything they want. I don't know why the industry is picked toward agents.

Starting point is 00:01:20 I mean, I get it historically, but I don't know if they've ever seen The Matrix in which the agents relentlessly try to kill Keanu Reeves. Hi, I'm your friend, Nelyai. It's the Vergecast. Alex Tran is here. Hi, I'm the Avergast. I'm the AI of Alex Kranz. How are you today?

Starting point is 00:01:34 Not enough emotion, Kranz. More emotion. Hi. Oh, my God. Kranz is going to flirt with you this entire episode. David Pierce is here. Hi. I will not flirt with anyone this whole episode, I promise.

Starting point is 00:01:47 We should just turn up the knobs of how flirtier AIs are, like throughout the episode randomly. There's a lot going on this week. I don't know if you can tell for my voice. I'm very tired. I just came home on a red eye flight from Google I.O. in the West Coast. So Google had I.O. It was very exciting.

Starting point is 00:02:02 OpenAI had an event launching their own product. We should talk about those two things together because they're kind of in direct competition with each other in actually kind of hilarious ways. And a bunch of other stuff happened. David reviewed the iPads. A bunch of other emulators have launched for iOS continuing to reveal that Apple's application model for iOS. Maybe a little limiting. The CEO of AWS just don't expectably stepped out. There's just like a lot going on.

Starting point is 00:02:27 And then there's like a Windows event next week. in which Microsoft is basically being like, we're going to beat the M-Series processors. So a lot. Just a big week. And I just got off a plane. So, you know, who knows? Who knows where this episode of the Virchcast will take us?

Starting point is 00:02:43 It's either going to be the shortest or the longest Vergecast in history. And no one knows. You know, a thing that I have, I have a new theory about red-eye flights? So, you know, in your 20s, you just take them because they're cheap. Yeah. And your body can just take it. And then in your 30s, you get a little bit smarter. and you're like, I shouldn't do that anymore.

Starting point is 00:03:00 Like, I have a real job sometimes. Like, I want to buy, I'm going to buy a real plane ticket. And then a thing that happens, and I've only noticed this recently, I'm in my 40s now and I have a child and I take the red eyes because I just want to get home. Like, my goal is to just be back at home because I'm like, I don't, I've been places. San Francisco, it's Milwaukee with rich people. Like, get me out of here. Like, going back to New York. And then you look at the plane and the plane is like a gradient of like people in their 40s.

Starting point is 00:03:29 going home to people in their 20s who are hungover. And that's a red-eye flight. Yeah, that's about right. It's very good. One of my favorite early Neely stories that I still tell is when I think we were in Barcelona for MWC and your go-to red-eye advice was eat two McDonald's McGrittles right before the flight and you will instantly fall asleep because your body just shuts down and that's how you sleep the whole way home. It's the level of sodium. It's been like a decade since you told me that and I still think about it every single time I get on a plane. It's unbelievable advice. My friends, I used to call it

Starting point is 00:04:02 time travel. You just like eat two McDonald's sandwiches. You just wake up wherever you're going to be. It's like, what happened to me? You're like desperately thirsty and extremely puffy, but you are. Sweedy and yeah, it's bad. But you get there. Anyway, but I'm back. I'm here. It was a big, it was just a big week in AI. So let me set a little bit of the just competitive stage here. We have long known when Google I.O. is. It's around now every year. And they have it at the shoreline amphitheater, which is a big outdoor amphitheater at Google owns right next to Google Campus.

Starting point is 00:04:37 And they invite lots of people. Last year is a little bit smaller. They're coming up in the back. This year is big again. We just know when it is. Then there's this other company called OpenAI that likes to act like they're just a bunch of kids, even though they're some of the richest and most experienced people in all of tech. And they're like, oh, we should have an event.

Starting point is 00:04:55 We're not going to have an event. And they literally did this move. They tried to upstage Google I.O. But maybe leaking, maybe not leaking, maybe just like swirling some rumors. They were going to launch a search product last week. And then Sam Altlin was like, we're not going to launch a search product. And then they announced a demo. And they did a live stream called their Spring Event, which they announced, I don't know, 20 minutes before it started.

Starting point is 00:05:18 They just created some anticipation that they were going to have a big thing before I.O. Well, it was weird because there were rumors that it was going to be GPT5, which obviously would have been a huge deal. And then there were rumors that it wasn't going to be GPT5, but it was going to be a search engine, which would also be a huge deal. And then right before the event, Sam Altman, I believe, was the one who tweeted, like, not GPT5, not a search engine, but we have something to share. So, yeah, they are, they're doing this very weird, like, aw, shucks routine about, like,

Starting point is 00:05:49 it's not a big deal. We're just, and that's so on brand for OpenAI, where they're like, they're both convinced that they are doing the biggest thing that has ever happened in the history of humanity. And like, I literally mean that precisely as I said it. Like, they actually believe that. And also are just these like, aw shucks scientists who are like, oh, look, we made a thing that will flirt with you. Isn't that fun? Yeah. I would, again, I'm just pointing out, some of the most experienced, righteous people in tech. Yeah, they know exactly what they're doing. And they are, they are very competitive, I guess is the way to put it. Like, they're, they're, they're,

Starting point is 00:06:24 aggressively competitive with Google. I think they see that their opportunity is to displace Google, whether or not their opportunity to actually build AGI. We will come to that, right? Like, one way or the other, open AI is going to make AGI. But in the show, we will also come to that. And so there was just that little gamesmanship there of are they going to announce a big thing?

Starting point is 00:06:45 Are they an upstage Google? And then so they had their event. They announced their multimodal search GPT4O, O stands for Omni, that lets you search with video, with audio, with voice, it is natively multimodal. The, you know, the feeling at Google I.O. And Alex Heath reported this in command line is that OpenAI did the demos Google was going to do. Right.

Starting point is 00:07:07 So this was the big upstage. Like, Open AI saw what Google was going to do or somehow heard about it and rushed to demo those same capabilities first to make sure that Google was in response to them. I don't know. That's true. Right. That was just very much the feeling from Google folks. It's very much the feeling that Alex reported.

Starting point is 00:07:23 but there is some real gamesmanship there. And I think the important frame for all of the news is it feels like OpenAI and Google have identified some kind of end state of an AI chatbot, AI search product, and now they're in a furious race to get there first. I would actually frame it slightly differently. I think everyone has known this is the end state for decades. Like the idea that you should have a Star Trek computer as a thing is whatever. everyone has been talking about forever.

Starting point is 00:07:55 Like, you go back to when we, the early days of Siri and Alexa and Google Assistant, and they were talking about this stuff. But the sense that I've gotten is that I think the thing that has changed, and I think it's changed really recently, is that people think we're close. That there is a real sense that like, not only is this the dream, but like, oh my God, we can build it now. And I think that was definitely the sense I got talking to Demas Sassabas, who runs deep mind at Google and is in charge of all their AI stuff.

Starting point is 00:08:21 It was kind of the sense you got watching. the Open AI stuff. They think it's here now. And you get the sense that we've gone from like, oh, neat pie in the sky, a dream that someday we'll all get to, which everyone has always known is the goal. Whether or not it's a good goal, still debatable. We've tried it a bunch of times.

Starting point is 00:08:38 It hasn't really worked. Here we are. But there is this real sense that like, oh, it's happening now. And we have to go get it right now. And it just feels like everyone has gone from running as fast as they could to like running even faster because there is a sense. that like you can sort of see the finish line now, which is really interesting. Yeah, but they're running faster to a not particularly useful place.

Starting point is 00:09:01 Like, like, I just watching it, particularly the chat GPT one, was like going to do you, remember at CES, Intel used to have a big booth? And they'd always be like, look at the power of what we can do with Intel. We can read your emotions with the camera. And I was like, okay, so when ChatGPT did it, I was like, cool. I saw that demo from Intel four or five years ago. Oh, cool. You can like read things when you hold the camera down on it. Again, I've seen that demo before.

Starting point is 00:09:29 Like, it felt like a, it was like a nice demo. It was done very well. And like I can appreciate that there was a lot of new technology there, but it also didn't feel like it was doing anything new or fantastic. Alex, I think you're underestimating the level to which the billionaires of Silicon Valley believe that people want to bang an iPad. Oh, this is true. This is true. Yeah, like, where's your David's Sennel?

Starting point is 00:09:49 Zazlov, love. You've got to, like, love people, not just devices. Yeah, an FMK with David Zazlov, an iPad. Like, Sidney is like a real, it's like, I don't know, guys. Beautiful lineup. Nightmare rotation. It's a tough one. I think Sydney's going to lose. Like a lot of people are going to pick the iPad.

Starting point is 00:10:13 They are. So that's the frame. And I actually want to come back to that, Alex, because it feels like the core technology that has gotten everyone so excited has a fatal flaw with these hallucinations that everyone is ignoring. And we should talk about that a lot, but let's go through the news. So, you know, we covered the hell out of Open AIs event, Kylie Robinson, our new senior AI reporter, talked to Niro Moradi, their CTO about some of their technology. We sent an army to Google I.O like we always do. Soon our Pachai was on Decoder, which will give a little preview of.

Starting point is 00:10:47 That's going to come out Monday. So just a lot. We covered all of this stuff. Let's go through the news. Open AI's news pretty straightforward, right? The GPT40, that's the Omni model. It's faster. They change some of the pricing so you can just get it for free now. You don't have to pay for it like you did with GPT4. There's a Macap now, which is very funny because they have a gigantic Microsoft investment. By the way, their answer to why Macap is that's where the users are, which is both a statement, a true statement, and also the sickest possible burn that you can deliver to Microsoft. your major investor. They are letting people use the GPT store for free now. So, like, really, it's like a pricing change, this new model. And in classic sort of opening eye fashion, you know, they announced SORO, their video generator, no one can use it. They announced GPT40, it's out.

Starting point is 00:11:34 But the thing where you, like, wave your phone around and look at stuff, I don't think is hit yet. No, you can get at the model. And, like, every AI app on the planet has been integrating that model over the last, whatever, 96 hours at this point. It's out there. All accounts is very good. People like it.

Starting point is 00:11:50 It's really fast. But the, yeah, the, like, super agent stuff that everybody is talking about is, as far as I can tell, not anywhere, really, at this moment. Yeah. And so the idea here, and this is where the idea that you're going to bang an iPad comes from, is you've got a phone. It talks to you. The voice is very much like Scarlett Johansson in the movie Her. Sam Altman tweeted the word her. Mira denied that it was Scarlet Johansson's voice.

Starting point is 00:12:16 Boy, is it close. But the idea is you're just like looking at the movie. stuff with your phone, you're talking to an AI, and it's much more personable, even a little bit flirty. It's like telling jokes. It's saying, you're looking nice in that hoodie. And it's, like, more powerful because it's not going from your voice to text into the AI model, back out to text and back out to text to speech.

Starting point is 00:12:36 It's all just happening natively, which is a speed up. But that's, like, the thing, right? Like, the announcement here is very much this model can now talk, can, like, look at stuff with you and you can talk to. it in a much more emotive way. Okay. Right. Can we talk about that actually?

Starting point is 00:12:52 Because I was talking to somebody about this yesterday, and they just sort of casually made the point that this is a user behavior that does not exist and doesn't make any sense. And I have not been able to stop thinking about it. And like, I go back to the, when I was at the Rabbit R1 launch a couple of weeks ago, one of the demos he did was write a spreadsheet out by hand, just like a table of numbers, and then point his R1 at it. and say, I think it was like, can you invert the rows and the columns and then send this to me? And it's like, cool technology.

Starting point is 00:13:25 What the hell is that for? And is anyone in the world ever going to actually do this? And so many of these demos that we've seen both from Google and OpenAI this week are basically people walking around a room asking their phone what a thing is. And I just like, and this goes back to Alex, your point of like, what is all this technology actually for? That's either a user behavior that we just have never developed because we've never had the technology for it, or it's just not a thing people want to do. And I'm increasingly

Starting point is 00:13:52 leaning towards the second thing. I mean, if you can't see, you probably, it would be very useful to the point of phone in a general direction and be like, what's there? Yeah. That's a good one. It's a good accessibility feature. No question. A hundred percent. The demos are weird because they need to create them so that you can verify that the thing is happening in the way that you think it's happening? Like the dumbest one I've seen so far, someone holding up GPT4O at Buckingham Palace and being like, is the king here? And it's like, yes,

Starting point is 00:14:25 the flags are raised, so the king is here. And it's like, dude, like, you don't need all this technology to get that, like, they've been answering that question since the 1200s. You know, it's like, that's why that's the thing. That's the whole thing. And there's just a

Starting point is 00:14:41 piece of that where it's, you need to construct the demo, so it's obvious the computer is doing the thing. And the actual use cases are yet to be seen because real people have to get the stuff and try doing things that no one's ever thought of doing. This actually ties in with what Google demoed, right, which is you have some sort of problem, you point your phone at it, you're like, help me fix this problem. We will come to that because both of their demos were kind of weird for various reasons,

Starting point is 00:15:06 but it's like that thing where it's like, I don't know what this part is called, but it's rattling. Do you fix it? Right. And then like the computer knows some things and it like helps you get to that answer. Whereas all these demos, I think they're just trading on the fact that people will know the answers already. And what they're proving is that the various AIs can. And I think that's what you're getting at, David, is like these demos are not compelling.

Starting point is 00:15:31 Because at the end of the day, they're designed so that you can see that it's doing the thing that you already know. Right. Yeah, it was like the person, they were in the London Google office demoing, uh, Project Astro, which we'll get to. But running around with the same kind of multimodal assistant, and they point their phone out the window at Kings Cross Station and say, what is this? And it goes, Kings Cross Station.

Starting point is 00:15:54 It's like, who did we help here? You're in Google's London office and you're like, gosh, what enormous station is that? Just, I don't know. And I think, to your point, I think as a like search input, it's super interesting. But the idea of that as a sort of ongoing assistant that I'm just going to like pipe the video of my world into all the time and it's just going to chat to me about what I see. That I am suspicious of. I just don't see that yet.

Starting point is 00:16:21 The funniest thing about that demo in particular is I believe it also asked, in the demo, I believe they also asked Astra, what is King's Station known for? And Tom Warren and our team who is British said no Londoner would have given such a milk toast answer. They've been like, that's where you get hookers and cocaine. That's actually the answer. If Gemini had said that, I would be more impressed. Yeah, I'm like, straight up. You're just like, finally, billions of dollars of technology later,

Starting point is 00:16:48 we're heating the oceans with all this GPU usage. But I can finally ask our robot where you're at hookers? Yeah, like, hey, chat, GPT, who in this audience is most likely to sell me drugs? Like, now we're talking. Yeah. So that's like the open eye news. And the big piece of it is this emotion stuff. Right?

Starting point is 00:17:06 They made the thing multimodal, but I think the thing that was compelling to most, people in this demo was that it was flirting with the people who were giving the demos and in a voice that sounded very much like Scarlet Janson. Was it compelling though? I mean, it definitely was flirting, but like, was it compelling? So I'm just going to caveat this with two things. One, it's been a long time since I was called upon to flirt with anyone. And whenever I try to flirt with Becky, she's just like, I don't know what you're doing. Like, those days are a long gone.

Starting point is 00:17:38 And so, like, I read it as like, this is what they, this is what a bunch of nerds think flirting is. Yeah. Like, you want a flirty female voice to be like, you look good in that hoodie. But it's like, maybe you shouldn't be wearing a hoodie. You know what I mean? Like, I don't want a computer to just reinforce all the bad, stupid things I do all the time. I don't need that kind of hype person in my life because they're not productive. Yeah, I mean, there's just an element where you should listen to it. It's hard to talk about it just like away from the actual thing. But if you listen to it, what they've added is a bunch of stutters and pauses and like intonations to what would otherwise be a pretty dry text to speech output.

Starting point is 00:18:25 Yeah, it sounded the most human, I think, that we've heard one of these sound, right? And I know that like back to that project Astrodemmo too, they're actually saying don't talk to them like you talk to Alexa or. Siri or any of these other ones. Talk to it more like you would a human. And that's like kind of compelling. But at the same time, I go back to I don't want a yes man. Like the yes man is how I ended up with a leather newsboy hat in the 2000s. My friend was just like, she wanted to just get out of the store. She was like, yeah, get that hat, Alex. I worked for weeks before she admitted it. And that's what, like, I don't need that from a computer. Yeah. No, our nation's teens during that same period of time was like swing dancing. That's what we're going to do. Yeah. Yeah. And like,

Starting point is 00:19:08 like, that's just a collective delusion that honestly what you want is an automated system to put a stop to. Yeah. And we just didn't have the technology at the time. And now maybe that technology is pointed in the wrong way. I went through Brian Setser Faye. Like, I was a teen during that time. It happened to me. It's great.

Starting point is 00:19:25 If it's still happening to you, please. How many people do you fling over your shoulder and then around your waist? Again, when I say that he's no longer interested in these antics. The thing that's always in my mind is she is a divorce lawyer. Like, we just don't play. The thing about all that, right, is what people are reacting to is the personality. And then all the demos at the bottom, it said, check your facts. Yeah.

Starting point is 00:19:51 Right? The thing they have not improved is, is this thing getting more accurate and more useful or smarter? Which I want to put a pin in because, Alex, you wrote a great piece about that that also comes up in a Google example. but the part where everyone's excited about this enabling technology to build these experiences that everyone's always wanted the day we was talking about is like, is the thing good enough to do the thing that they want? Or is it just convincing enough? I actually don't know the answer to the GPT40 because I haven't had a chance to use it very much.

Starting point is 00:20:20 And the feature, the big feature isn't actually available to be used. I do think it's becoming increasingly clear that it's getting convincing faster than it's getting good, which I actually think is sort of scary. like just just looking at the reactions that people had to the open AI demos this week. Like it made a lot of mistakes. It got a lot of things wrong. The interaction was kind of weird, but it was mind-blowing to a lot of people. Just the fact that it sounded passably human and it made jokes and it was sort of silly and flirty.

Starting point is 00:20:50 And they could dial up and down the emotions. There's one point where it was two researchers who were doing this demo and one of them. He's sitting there and he asked a question and she responds. And the other guy goes, not enough emotion. Like give me more, more. drama and it like correctly amps up the drama and like to the point where it's funny like I laughed at the thing and that's remarkable but I don't know that there's any evidence that it is making fewer mistakes in those moments so far than these things have in the past or that it is like

Starting point is 00:21:21 meaningfully closer to getting to zero which is ultimately the only thing that actually matters yeah and zero is a hard place to get like people aren't at zero but at least I you understand the failure mode of another person. You know, like, you can evaluate whether a person is trustworthy or, like, you have some history, or you can find a number of TikToks going through someone's entire backstory to tell. Right. Like, there's a whole thing you can do to evaluate whether another human being is trustworthy. You're telling you the truth or whether they're experts or whatever in a way that, like,

Starting point is 00:21:53 AIs are just like a different kind of thing. Yeah. And we don't have the skills as a society or a culture to do it. And also, we're going to come to this. This is Alex's piece. Everyone seems to be just ignoring it because it's fun. Right? And it's like, that's weird.

Starting point is 00:22:06 If I'm going to briefly defend AI, I will say there are a lot of things about these tools that are very low stakes that are fine. Right. Like there's a lot of things for which there's not a right answer. And it doesn't matter if it gives you the slightly wrong answer. And again, I think the thing that we're learning is that more of the point is the interaction than we expected. Right. The more you hear about Kevin Ruse's girlfriends in The New York Times, I loved that story. Like, we've given Kevin a hard time.

Starting point is 00:22:35 It's an unbelievable story idea, and I'm very jealous of it. Yeah, I mean, his beat is, what if you banged an iPad? Yeah, it's an unbelievable idea. Incredible. And you hear about, like, the meta-celebrity AI things that they have. Like, half the point of these things is just to use them. And I think I really underestimated that for a long time. I was like, these things are tools, and the goal is to accomplish something.

Starting point is 00:22:58 as quickly and efficiently as possible. And having ums and ahs is actually just a waste of everybody's time. And for my own use, I actually still believe that. But I think in general, the fun of using it, it's a real, like, the real friends with the friends we made along the way is like, that's very much, that's very much where we're at with a lot of AI. And I think it forgives a lot of the wrong stuff in a way that is a little scary, but also if you're a company like OpenAI, really productive.

Starting point is 00:23:26 because as long as it's fun, people will forgive so much. And they're making it fun. I mean, but you're kind of getting the point where it's like, what if your computer was Calvin's dad from Calvin and Hobbs? Yeah, just like, I don't know. I'm just going to make some to go away, kid. Fine. You know, and that was life before computers, like before Google,

Starting point is 00:23:46 people would just make stuff up at parties. Like, that's what it feels. And fine. Like society made it through many moments without being able to Google stuff on their phone at a party. but it's just weird that we're at a place where we had that facility, and now we're like, it's less accurate than before, but it's flirtier, and the point is the interaction. That's a change.

Starting point is 00:24:07 It's like worth remarking on. That kind of mirrors, like, people's search behavior in general, right? Because everybody's like, I could use Google, which will take me probably to the correct answer. Or I could just go on TikTok and be like, does anybody know why this happens? And take the first answer I get from the most charming person. And the answer is your water doesn't have enough hydrogen in it. Yeah. Just packetful.

Starting point is 00:24:26 Open-in announced all this stuff, which is interesting, right? Again, the timing, then the day of I.O., they announced that former chief scientist now, Ilya Sutskova, is leaving the company after all of the drama about Sam Altman getting fired, Ilya was one of the people that voted to remove Sam, and then he said he made a mistake, and he went back, and there was the whole thing, and now Sam's back with a bigger board, whatever. It always seemed likely that Ili would leave, and then they announced it on the day of I.O. And I only bring this up, you know, it's like the sort of machinations of open eye are very interesting because of Sam getting fired.

Starting point is 00:25:01 But it's interesting timing and that they made all this noise about maybe we're doing search, maybe doing a thing. It's not, it's, but it's magic anyway. And also our chief scientist who led a coup will be leaving the company. Yep. Don't pay attention to that. Pay attention to Google's news. Like, weird. It's just a weird.

Starting point is 00:25:18 That I think Open Eye is in a weird spot where they are, it doesn't feel like this from the product. side. I think the product has captured more imagination, but they are as a company, like pretty reactive to Google. Like they're in they're weaving their way around Google in a fairly interesting way. Yeah, I think Google is

Starting point is 00:25:37 getting powerful in this space more quickly than anyone gave it credit for not very long ago. I don't think that. Do you remember 12 months ago when we were sitting around being like, Google blew it? They blew it. Where had they been? Google fucked it all up. We knew that Google was like slow, but like, I don't think it's

Starting point is 00:25:54 surprising that Google, one of the largest companies in the world, is going to run circles around this company that was originally created, to be clear, to protect people from AGI while also creating AGI. Like, there's still a startup. Yeah, they've got a ton of money for Microsoft, and they're hunting for more investment all the time. But Google has been investing in this space for a very long time, investing in those resources, investing in those people.

Starting point is 00:26:21 and I'm not surprised that it came up and said, yeah, we're going to do it and we're going to crush it. Like, of course, Open AI is going to be on their back foot the entire time. That's why they rolled out their store so quickly. That's why they've been moving so fast is because they're like, okay, we beat them to market. Now we have to keep beating them to market because that's all we have. We don't have all of the infrastructure. We don't have all the resources that Google has. So this is what we can do.

Starting point is 00:26:46 Yeah, no, I buy that a lot, actually. And I think that, you know, what Open AI is. good at is having a product and caring about a product and having the product be existential to the company and making the product good. And Google is not good at that. Real bad. Like, Google has a product that it cares about for a long time. Not in Google's core skill set.

Starting point is 00:27:10 It hasn't been for a while. And, you know, last year they announced a bunch of stuff. And I'm like, where is it? You know, again, they were at Bard last year. Like, all that's gone. They already wiped out one whole set of product. and replaced it with a long list of other products, all which are named Gemini, which we will get to.

Starting point is 00:27:27 But they had, you know, their announcements at I.O. We were very confident. We should go through at least the chatbot stuff here. I did talk to Sundar afterwards. We talked a lot about sort of how people feel about AI. You know, that's coming on Decoder next week. But it felt like he has a handle on it. You know, like he knows that he has to do this stuff.

Starting point is 00:27:48 And he knows it's like very much like can't make an omel. about breaking some eggs and he's like, I feel bad for the eggs. But he's doing it. Like, they're in an aggressive posture. And the thing that they are, I think, the most aggressive posture on is Project Astra is some of their search improvements. And it's the idea that they have the infrastructure to scale this stuff and make it cheap. Right. So Sundar said to Deirdrobosa on CNBC, our cost of queers are 85% cheaper than it was a year ago.

Starting point is 00:28:18 And that, if you're opening eye, is like, oh, shit. Like, Google can just deploy this stuff to everybody who has Google and it's cheap and they can like figure out their business on the back of search ads and all this stuff and we are at, well, 20 bucks a month was a lot of money and people weren't paying for it. And now we're giving away the best model we have for free. Right. Like, there's a real challenge in there. And you can see these companies kind of battling it out, each using their own advantage.

Starting point is 00:28:40 Open eye very much their advantage is they can tell a compelling story about a product that everyone believes will be here a year from now. And Google's advantage is that it's huge. Yep. And you also just described basically the whole reason that Microsoft and Open AI got so tied up because Microsoft promised AI that same level of scale and infrastructure. And I think if we've learned one thing, it's that it's very hard to do that in a partnership in a way that it's much easier to do it inside of your company. And I think if you look at what we've seen across Google the last year or so, that company has just like relentlessly consolidated itself to be able to move all of that stuff. faster.

Starting point is 00:29:20 Yes. And that's like, that was the Rick Osterlo story when they, they put him on top of hardware and on Android. It's to move AI into pixel and Android more quickly. And the infrastructure stuff, like one thing Demas Asa said to me is that the, the fact that Google's infrastructure exists is how you get Astra to be so fast.

Starting point is 00:29:39 Like, he just literally tied those two things together. And he's like, we spent six months taking the tech that we had and making it fast. And what that meant is putting it in the rest of Google's infrastructure. Like, that becomes the job. And you're right, cheap and fast are how you win. That is the next phase of this is we take all of this stuff and we figure out how to do it.

Starting point is 00:29:58 We figure out how to do it in a way that is like economically sustainable and doesn't have latency. Because latency kills all of these products just as a user. Like if it's slow, it sucks. And that's where we are. And Google has known that forever, right? I mean, Google has always known its speed is the thing. So we're talking about some of Google's news. The main one is Project Astra, which is their natively multimodal take a video.

Starting point is 00:30:19 of a thing, search for it, ask it questions. I did a demo of it with V-Song. It was weird. There were Google executives there doing demos on their phones, apparently. We didn't see that. We went into a booth where we put on like a gaming headset and there was a, you know, a top-mounted camera pointing down on a big TV screen. And we were asked to take objects out of like bins and like put them on the top, you know, so the top-time camera can see them and like ask questions about the objects. Which to David, point is a completely insane use case. Like, you are just never going to do this.

Starting point is 00:30:53 Yeah, what? But, like, the idea is that it could see the things that could just, like, instantly sort of latency-free chat with you about them. So, you know, V-picked three things and, like, said, tell me a story. Name this stuffed animal. Great. I broke it. Just instantly broke it.

Starting point is 00:31:09 I had a spaceship, a pleiosaur, like a plastic dinosaur. Did you know it was a pleiosaur? Or did it tell you it was a pleiosaur? I thought it was an ichthyosaurus, and it told me it. was a pleaser. By the way, I don't know who was right. I didn't have any dinosaur people there with me. You know, the one dude from Jurassic Park was not there with me.

Starting point is 00:31:31 But Alan, what's his name? Alan Grant. Yeah, Alan Grant was not there. Yeah. Good pull. There we go. What was the, who's the one who played Ellie? This is what I care about.

Starting point is 00:31:43 Laura Dern? Laura Dern. Yeah. Lord Dern. Oh, my God. I wish she'd been there with you. Yeah. I just wish she's around.

Starting point is 00:31:49 If Google's listening, the next time we do this demo, Laura Dern should be there. And I was like, tell me a story about these three objects where the Maraca is the winner, which, to be fair, is like an idea that can only come from my, like, booze-addled human brain. And it was like, get ready for a story. And I was like, I'm ready. And it was like, once upon a time. And it just stopped. It's just dead in its tracks.

Starting point is 00:32:15 Like, could not predict the next word in a sentence that added up. to the Morocco will beat this dinosaur in the spaceship. And there is nothing more awkward than that specific silence. Yeah. Where you're like, is it about to happen? Did it break? Is it doing the wake? Is there an hamster on a wheel somewhere?

Starting point is 00:32:32 Like there's a data center on fire. And the poor guy giving the demo just like hit the space bar. I was like, get out of here. Like he was like, we have a other demos to give. Like I don't trust you anymore. And it's fun. Like that is a weird demo. It's a weird use case.

Starting point is 00:32:44 It doesn't make any sense. The point is to show that it can like see. You can like use its camera. and do text to speech, like all at the same time in the same way, right, natively multimodal. But, right, like, you would never walk up. Like, if you ask a normal person, tell me a story where the Maraca is a winner, they would also be like, get ready. And walk away.

Starting point is 00:33:06 But I think there's just a part of this where we're asking these things to do, like, incredibly esoteric things just to see where their limits are, obviously. And you've got to bring that back to reality. So the one demo they gave, which is coming to Google Lens, it's like an extension of Google Lens, is a video search, which is sort of related to Astra. Like, Astra is a thing over here, but video search is a thing over there. And it's weird to pull them apart. So Astro, you're just like talking to a computer, you're like waving your phone around, looking at stuff and talking to it. Video search and lens is you like take a video.

Starting point is 00:33:39 Like you hold down the shutter button of a thing while you're talking and you say like, why is this record player broken, which is one of their demos? Or why doesn't the lever on this camera move, which is another one of their demos? And they feel to me like the very related ideas, but they are separate in classic Google fashion. Like these are not the same idea. They're different names. But the thing about it is the astrodemos, you know, they felt very esoteric. Ken and Mara rocket win. The video search demos were very tactile.

Starting point is 00:34:11 Like I could see how they would be useful. Yep. Right. This thing is broken. I'm just going to take a video of it being broken. I don't even know what the parts are called. Help me fix it. So the record player one, which they did a lot of on stage, it's pretty good, right?

Starting point is 00:34:21 They had a, people don't know about record players. There's a weight on the back of the tone arm. And if you don't have the weight balanced right, the tone arm just like skips around. Because you need some weight pushing down on the needle into the record to like hit the whole thing. If you're a record player nerd, this is like the most exciting demo that you've ever seen. If you're a normal person, you look at that and it just tells you this thing. Go watch the demo. I encourage you to watch the demo.

Starting point is 00:34:45 It's like that. It identifies the model, the make and model of the turntable. It's like, that's an audio technique of a turntable. And the reason that its tone arm might be moving this way is because it's not counterbalance correctly. And it has a whole list of like things you could do to counterbalance correctly. And it's like, oh, this actually isn't useful, which I thought was fascinating. It's the right answer. Straightforwardly the right answer.

Starting point is 00:35:05 Like, that's why it was doing that thing. But if you don't know what the tone arm is called, you've gone two steps too deep, too fast. Right. Right? You're like, the information you need is, like, record players work by putting a needle in a groove, and you need to put pressure on that needle. And this round weight at the back is called the counterbalance, and you need to spin it. So you put enough weight on the thing. And the amount of weight you put on it is actually dependent on your needle.

Starting point is 00:35:32 It's like, you fall down the rabbit hole and suddenly you realize, like, you own four record players. Yeah, then you're just an audio nerd. It's just a thing that happens to you. That's fine. But, like, that was really interesting to me. Right. It's the right answer, but sort of in a way that was like too far too fast. Oh, I think you're totally wrong. I think, I think you, what you just described is Google search, like, which is the way that it works, right? Which is that essentially...

Starting point is 00:35:56 But I'm saying, like, if you don't know what a term, I was called, you couldn't follow the directions that were given to you. Right. That's, I mean, that's fair. But I think maybe a better example of the opposite case is one that Liz Reed, the head of search talked about when she and I were talking, which is your dishwasher is broken and there's a light blinking. And you don't, don't know what brand your dishwasher is. You don't know what model it is. You don't know how to describe which. You're just like, my dishwasher's broken.

Starting point is 00:36:21 Like, off the top of your head, do you know that, do you know the model number of your dishwasher? I could make up anything right now. If you had confidently answered, I would have believed you. You're absolutely confident. It just makes up enough. But again, I don't, that's not information most people have. It's not information most people need.

Starting point is 00:36:36 It's not information most people should care about. It should just say, take out the filter in the bottom, rinse it out and put it back. You'll be fine. Like that, that is the answer. I don't need to learn about dishwashers. I just want mine to work.

Starting point is 00:36:47 And it's the like, I take my car to the mechanic. Some people are like, oh, I'd like to understand cars so that I can, I can make sure that my car runs better. Most people come in,

Starting point is 00:36:56 they go, my cars are going, Ravr, fix it. And that's the end of the interaction. That's great. But if you're asking, like,

Starting point is 00:37:03 what is this thing and why is it broken? And it's like, here's how to fix it, telling you what to do is actually important, right? And here it was just like, fix your counter balance. It just says,

Starting point is 00:37:12 pull the, Do you see the thing there? Pull it down. Fair. I just thought that was a particularly jargon. Maybe it's because I just, like, am a turntable person. But the answer there was, like, particularly jargony in a way that I thought was interesting, right? It didn't, like, it didn't help you understand the actual answer.

Starting point is 00:37:29 Well, I think you're also someone who thinks, like, you want to understand the thing that you're talking about and working with. Like, we're all gadget nerds. Well, I just want to lie to David confidently. Yeah. Yeah, that too. But, like, we're all nerds. We want to know how the thing works. We want to know all the things.

Starting point is 00:37:43 processes. And it's true. Like the reason, you know, when you first used Google search 20 years ago, you would have to be really think about your search prompt, right? Like, you'd have to think about what you type in. And now you just type in like, my dishwasher doesn't work. Yeah. And people just don't care as much. But, but to your point, like, yeah, if it said, if the dishwasher said, yeah, go replace the filter, I'd be like, where the hell is the filter? What the hell is that? Like, to your credit, like, I would still want to know all of that. And I don't think it would necessarily get me there yet. I'm just saying, I think that one's interesting because it was the right answer.

Starting point is 00:38:16 There's no question that was the right answer. Yeah. That's the reason the tone arm was flying back and forth. It was just, is it a useful answer? Right. And I think we can obviously disagree about that and we are. But that's the level that we're at. Like, is this actually useful for you to go sort of like spit out an answer like this

Starting point is 00:38:32 where it's fixed the counterbalance? Right. And you need to know some information to like take action on that instruction. Well, and there's a really interesting sort of AI all the way. down piece of that too, which is that like, what that indicates is that the, the AI for sort of understanding what's in a video that you're sending to it is capable, right? Like, it did the job, it figured out the question you had by you pointing a video. That's awesome, right? Like, I actually think the video search in lens is a very good idea. And I think pointing your phone at something being

Starting point is 00:39:03 like, what is this thing that's happening is an actually totally normal user behavior. So I'm like, all in on that. The output thing is actually much harder, right? Because it's like, okay, in this case is the right thing for Google to do to just sort of give me an answer about the tone arm, which is not a term that I know to your point. So if it says pull the thing on the tone arm, I'm like, what the hell is the tone arm? And you've lost me. Should it send me a YouTube video? Should it send me a compilation of stuff? Should it like superimpose AR on top? Like that is just such a different problem from the sort of search query recognition thing. And my sense is Google's getting really good at the search query recognition. Right. They did it with the voice.

Starting point is 00:39:41 They did it with video. They did it with images. Google is very good at understanding what you're asking almost no matter how you ask it. And then the question of how to answer is so different and so much harder and full of thorns and wildfires. So by the way, this is kind of exactly what I was getting at is what it presented was like a list of words, like vocab words. And then in the demo, she was like, and here's a link to Audio Technica's manufacturer page. which is more words. And what you actually want is like a YouTube short where someone just does it in five seconds.

Starting point is 00:40:17 Right. Somebody just reaches down and pokes the thing and it works. Right. And they're just like, here's this thing. Like do this thing until it floats and then like turn the number to whatever number is. Like that's what you, that's all you need to know. And like the actual answer to that is a video of someone doing it. And that's again, it just like you got it right in one specific way.

Starting point is 00:40:36 But it actually got it wrong like another way, which is like the form of the answer is not the most useful thing. Which is really interesting. Then there was the other demo, which it was just wrong. Like, wrong in a way where I'll just preview it. The first question I asked, you know, in the Dakota interview was, is language the same as intelligence? Because the question was that someone was pointing a camera at a broken film advance lever on a SLR, like a film, SLR camera. And so these cameras are really manual, right? So you take a picture, you know, move the lever to move the frame of film over in the camera.

Starting point is 00:41:10 If you've never seen a film camera, this is like a very satisfying mechanical thing to do. You like push the button, the big, you know, the shutter goes chagunk. And then you like move. You're literally pulling, you know, the notches in the side of film. You're literally pulling the film over and winding it around. Such a good feeling. So you can expose the next shot, the next frame when the next time we open the shutter. This is great.

Starting point is 00:41:33 I love, like, I'm a nerd for these things. And the answer, the question was, why can't I advance this thing? Why is this lever not moving all the way to advance the next thing? This is a problem with these cameras. They have them. There's a million reasons you might have this problem. One of them on old cannons is like there's a magnet at the bottom of the thing. And one of the answers is like get a magnet and like run it over the bottom of the camera.

Starting point is 00:41:55 Like there's all these. There's like folk tales and superstitions like, you know, it's like drink the blood of a goat and like wave at a crystal and like move your film in advance lever. It doesn't matter. There's just like a lot of information out on the web of how to do this. And the answer that Google delivered in its own video and highlighted is the most wrong answer. Like, it's right in one way, which is, as a last resort, just open the back of the camera. It will technically work. Yeah.

Starting point is 00:42:22 It will technically work. But if you open the back of a film camera, you expose all of the film. You ruin all of your photos. Right. So when people do this, when I ask Becca and Viren and a bunch of other photographers on our staff, like when you have a broken film event's lever, would you ever just ever just, you ever just, to open the back of your camera. And Beck was like, no, no, no, no.

Starting point is 00:42:40 You go into the darkest closet you can find and you pray. And then you open it and monkey with it because you don't want to expose your film. You don't want to lose your photography. And Viren was like, I've spent a lot of time in that closet, which is a very funny thing. The most Viren thing you can say. Right. Like these cameras break. They're finicky mechanical objects.

Starting point is 00:42:57 But there's not a place where anyone with even the slightest bit of intelligence about those kinds of cameras would recommend and highlight just open the door. and fix it because necessarily you're going to ruin all of your photos. So I actually asked you to know about this. I was like, this is not a thing. I don't see the intelligence going up. I see the capability, right? Like David's saying, it can see in the video what's wrong.

Starting point is 00:43:24 It can go search the web. It can synthesize some answer. But this answer is wrong. Like, this is not the thing you should advise people to do. And he's like, I went and talked to the team and they're like, well, in some cases, if you were willing to destroy all of your photos, it's the right answer. which is correct. Yeah.

Starting point is 00:43:40 That is the correct response. We should just run it. We should just run that clip. Let's run that clip now. Ironically, I was talking to the team as part of making the video. They consulted with a bunch of subject matter experts who all reviewed the answer and thought it was okay. I understand the nuance. I agree with you.

Starting point is 00:43:56 Obviously, you don't want to expose your film by taking it outside of a dark room. There are certain context in which it makes sense to do that. If you don't want to break the camera and a 40 of 10, is not that valuable. It makes sense to do that. You know, it's a good example of you're right. There is a lot of nuance in it. And part of what I hope search serves to do is to gives you a lot more context around that answer and allows people to explore it deeply. But I think, you know, these are the kind of things for us to keep getting better at. I do see the capability frontier continuing to move forward. I think we are a bit limited if we were just training on text data,

Starting point is 00:44:35 but I think we're all making it more multimodal, so I see more opportunities there. To me, that answer is like, yeah, we should have probably told you're going to destroy all your photos, right? Like, that's the answer. Yes. But it's also like, why are we rushing into a world where the products can't quite do the thing you need them to do?

Starting point is 00:44:55 And Alex, this was your piece, which is these things are hallucinating left and right, and everyone's kind of like over it. Yeah, it's, I just keep being baffled by it because people, it just screws up constantly. Like, if you give me one of these things, I can probably get it to hallucinate within four to five questions fairly easily. Like consistently, I could get it to screw up. And it's just because they're not good at it yet because they're computers.

Starting point is 00:45:21 They're not people. They're not actually all of the stuff that our brains do is really hard to do. And they're not prepared. But these companies are like, well, it's really cool. We're on our way. So like, enjoy. And this is now how we're going to have everything work. And it's like, well, no, stop that.

Starting point is 00:45:35 Like, I don't, I need my computer to know how to do my taxes. I don't need it to know, like, what a nice accountant should sound like. Have you seen the one where people keep asking it how to get a man and a goat across the river? And literally none of the models can, like, do it? Like, the question is, you have a man, a boat, and a goat, how do you get them all across the river? And it's like, first you take the goat across, then the man comes back and he takes himself across. And it's like, that's not the answer. This is a bad answer.

Starting point is 00:46:03 Yeah, they don't, like, the AIs don't think like humans do. They're not meant to. And that is remarkable. And it's like, it is cool that we're now effectively engaging with like these alien thoughts and things, right? Because it doesn't think the way a human does. But it also doesn't think the way a human does. And we keep asking it to. And we keep putting it in these situations. And then we're like, don't worry about it. So it hallucinated a little bit. So it like decided that Nelai wasn't the founder of the verge and that Alex is. a man, it's fine, keep going. And it's like, yeah, materially, those aren't harmful things to say, right? Like, I think we'll both be fine if somebody said that. But it still shouldn't say it, and that's what you want to now become the new form of search. Like, for me, it really feels like Google has lost the plot for what Google is supposed to do, and Google is supposed to get you to the information. It is not supposed to be the arbiter of the information. And increasingly, it feels that it needs to be the arbiter of the information. And that's like the whole point behind a lot of its new search products.

Starting point is 00:47:04 We're just going to tell you how to think. And it's like, well, no. One, you're a company, you shouldn't tell me how to think. Two, you don't actually know because you're relying on these AIs that don't know. Just stop this. Like, this is a really bad path for them to go down. I feel like it's going to eventually bite them in the ass. But they're so financially incentivized that they just keep doing it.

Starting point is 00:47:23 I mean, this is truly the competition between Open AI and meta and Google is ferocious. Microsoft, obviously, in the mix there. Apple wants to be. Well, Apple's just going to buy some model from the other, right? At least for this stuff that Siri goes outside your phone. That's the rumor. We'll see. WWC is coming up.

Starting point is 00:47:42 But it just feels like there's that ferocious race to get to this product that everyone has dreamed of for a long time, whether or not the core technology can do the task reliably. And, you know, just go back and think about that film camera example for a minute. But if you had a friend who was like, you were like, for some reason, you're like, my film camera is broken. I can't move this lover. And they're like, yeah, just open the back. And then you open the back and it destroyed all the photos you'd take in. By the way, film's not cheap, which is the thing the Becker reminded me of instantly.

Starting point is 00:48:10 Like, film is expensive. So you're going to ruin all these photos. You're going to ruin your expensive piece of film. You would be pissed at your friend. Like, just as a matter of course, you'd be like, what? Dude, like, you could have told me to go to a dark room. Like, I asked you because I assume you know there's other piece of information. And there's none of that accountability with these tools.

Starting point is 00:48:30 You can't just, like, be mad at it. Because, you know, in a market, you would leave Google and go to opening it. It's like, it's going to lie at you, but in a way that suggests it might also have sex with you. Like, yeah. Fine. But I just, that's the part where it's, it's like everyone's building a really fast car, but they're not telling you that the engine just doesn't work about 20% of the time. And it's like, okay, is there a path to stopping it? That would be a bad car.

Starting point is 00:48:56 I actually, I would put it slightly differently. I think it's, they're building a fast car and not telling you how to use it. Right. And I think like the camera thing to me is such an interesting example. Yeah. I would argue that's not a hallucination. That's a terrible UX mistake. Like it gave you a true answer without the actual bit of information you needed to know what to do with that true answer. You can open the back of your camera to get the film out.

Starting point is 00:49:21 That is a true thing. And so in a certain way, the model was successful. And like that's what I hear Sundar saying. It's like it didn't get it wrong. It just didn't tell you what you actually needed to. know. And so now we're in this place where like, I mean, that is the, you're dancing on the pins or the heads of wheels, whatever I think, I mean, there are, there are things that are wrong. Like, Alex Kranz does not have a beard. That is wrong, right? Like, got it. Nailed. Yeah.

Starting point is 00:49:53 But the rest of the answers on that list, that's the one I focus on because it was most destructive. The rest of the answers in the list that it generates are bananas. Like one of them is just like, nudge the shutter a little bit, which is like, no? Yeah. I've already done that. But I think that's not going to help you. One of the things I think a lot about with Google right now is Google used to make you do a lot of work, right? And that was the job, right?

Starting point is 00:50:15 Like when you did Google search, you did most of the work when you were interacting with Google. And now Google is trying to do most of the work for you. And it creates this incredibly different version of Google that needs to exist. and what Google thinks is that actually it can just put a paragraph summary at the top and then let you use Google the way you normally would because if you want to know more than just the paragraph, well, here's regular Google. And I think that is like woefully underestimating how differently people are going to start to use Google as a result.

Starting point is 00:50:47 And so to me it's like making these things in such a way that is like weird and different is one thing. But to just push them at people and be like, this is a familiar thing that you totally understand is a disaster. And that's so much of what we're getting. It's like, this thing sounds like a person. Talk to it like a person. It works like a person.

Starting point is 00:51:05 Like, it super doesn't. Not even close. And all it's going to do is run you into trouble. Yeah. To me, this is the question that like all of these products have to answer, right, is what are people's expectations and can you change them fast enough? I truly do not know the answer to that because I think what people are responding to is these computers are talking to them.

Starting point is 00:51:28 And it turns out people confidently, flirtatiously talking to you, charmingly talking to you, is a great way to get a lot of people to believe a lot of lies. That's how you get Alex Kranzh's in a newsboy cap. Yeah, exactly. This is how swing dance happened to America.

Starting point is 00:51:45 I don't know, that's true. But, like, Google knows this because it operates YouTube. It knows this problem intimately, right? There's a lot of charming liars. Here we are. We're on YouTube. But it's interesting that it's like running full force into making the problem. And like you said, Alex, owning those search results, like that's Google's fault that they told you that answer.

Starting point is 00:52:05 Whether or not you think that it's appropriate to pop open the back of the camera, that's Google's answers now. It's not some answer in a Reddit thread. It's not some answer from a friend that you're kind of mad at. It's just Google's problem now. It's going to be opening eyes problem. There's a long more to talk about. We should talk about search very briefly. We should talk about everything else Google now.

Starting point is 00:52:21 So we've got to take a break, though. We've been at this for a minute. We'll be right back. Support for the show comes from Framer. Framer is an enterprise-grade no-code website builder used by teams at companies like Perplexity and Muro to move faster. With real-time collaboration and a robust CMS, with everything you need for great SEO,

Starting point is 00:52:42 not to mention advanced analytics that include integrated A-B testing, your designers and marketers are empowered to build and maximize your dot-com from day one. So whether you want to launch a new site, Test a few landing pages or migrate your full.com. Framer has programs for startups, scaleups, and large enterprises to make going from idea to live site as easy and fast as possible. Learn how you can get more out of your dot com from a Framer specialist or get started building

Starting point is 00:53:13 for free today at framer.com slash verge for 30% off a Framer pro annual plan. That's Framer.com slash verge for 30% off. Framer.com slash verge. Rules and restrictions may apply. 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

Starting point is 00:53:45 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 shortlisting candidates and conducting AI-powered interviews for initial screenings. Its updated conversational interface lets you describe what you need in plain language. Nearly 60% of hirers find a candidate to interview within a week.

Starting point is 00:54:21 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. Support for this show comes from What Not. Whether you're selling online or out of a storefront, you already know the challenge. You're simply hoping for people to find your listing or waiting for them to walk in.

Starting point is 00:55:00 But What Not flips that. They say they're the live shopping marketplace where you can shop, sell, and connect around the things you love. On What Not, you go live and sell directly to people in real time. They see what you've got, ask questions, and buy. And they keep coming back. Whether it's beauty, collectibles, electronics, luxury fashion, and yes, even cookies, sellers are building real thriving businesses. And for a limited time, What Not says they'll match your first $150 sold in the first month. You can visit Whatnot.com slash sell to start selling.

Starting point is 00:55:43 That's W-H-A-T-N-O-T dot com slash sell to start selling. sell. Whatnot.com slash sell. All right, we're back. Existential crisis about banging iPad is on pause for the rest of this first podcast. I know you wanted more of it. I thought you were going to say over.

Starting point is 00:56:10 And I was like, well, that's a lie. But on pause, we can live with. No, Kevin Ruse still wanders these streets, David. By the way, we're good friends with both Kevin and Casey at Hartford. Oh, yeah. Out of love. Although I do worry. Out of love.

Starting point is 00:56:24 All right. Just not as much as we love our iPads. It's like my wife, my iPad, Kevin Ruse is like, I would say the big three. Man, the dog just got a job. A child. He's there. He's fine. He doesn't play me any YouTube videos.

Starting point is 00:56:46 Fair enough. Look, the open-AI stuff and the Google stuff, right, these like video multimodal search chat interfaces, those are the main events. Google announced 10 million other things at I. all of which were named Gemini, basically, or variations of it. It's very confusing. David, can you just run us down the stuff? Sure.

Starting point is 00:57:08 So basically, the theme of Google I.O. is just Gemini in everything. There was Gemini Google Photos, which powers a feature called Ask Photos that lets you ask questions of the photos in your Google Photos. The example they gave was you can just search what's my license plate number, and it'll find your license plate number in your license plate. a photo in your Google photos, which as somebody who only knows my driver's license number because it's in my Google photos, rules. All about that feature. Yes. By the way, that might be the single

Starting point is 00:57:38 most important thing that they announced was AI searching Google photos. Oh, it also, as far as I could tell, I wasn't there live, but it got a big reaction as far as I could tell, both live on the live stream and just in our office, people being like, yes, license plate search. Well, it turns out no one knows what their license plate is. You know what I think it is? It's Google Photos is so close to being able to do that now. So it's Apple Photos in its way that you start to ask it things that can't do. So if you ever search your Google Photos, you quickly start to ask for things that it can't do

Starting point is 00:58:10 because it's so close to being able to do that now, right? So you can just ask it for like, show me all the people wearing a blue shirt on this day in the past and it'll figure it out, right? Because it has some ability of that knowledge. I personally search for the word truck a lot. Just a gaze. My now departing pickup truck. I needed to know my VIN number for when I traded in the truck.

Starting point is 00:58:33 And I just searched for the words VIN number. And it just, like, found it, right? Because I'd taken a picture of it. The idea that you can extend that, I think people responded to it because they're already trying. Yeah. So now you're going to give the people what they want. It's like very good. Yeah.

Starting point is 00:58:47 Yeah. So that was one thing. That got a lot of excitement. Sort of in the same vein of like things you already kind of want to do. They announced a bunch of Gem and I stuff in workspace. There's a, there's a thing where you can like ask similar questions of your email.

Starting point is 00:59:01 The example they gave was if you have concert tickets or I think it was Nick's tickets was one of the examples they used. You can just say like, what time do the doors open at the next game tonight? And it'll actually look in your email for your ticket to find that information. That kind of stuff is neat. There was a bunch of stuff in just the kind of general model universe. There was a new version of the Imagine model, which is one of the ones that developers can use for creative tools.

Starting point is 00:59:29 There was VO, which is Google's answer to SORA, which is a text-to-video model. Which seemed really cool, by the way. Yeah, it seemed even better than Sora in a lot of ways. It also debatable whether either of them actually exist in the world. So we'll see. There was a new thing called Gemini Live, which was like an always-on voice chatty thing, basically very similar to the one that OpenAI announced. Project Astra is the big long-term vision, Gemini Live, is that just sort of specific voice

Starting point is 00:59:59 thing. There was an update to Synth ID, which is how Google is attempting to figure out a way to watermark AI-generated stuff. They're now doing it with text, which I thought was really interesting, but also with videos, which goes along with VO and all the stuff they're doing there. A bunch of stuff for Gemini Advanced and Gemini Nano and Genome and I, one of the 1.5 Flash, which is a new model just designed for all of it to be fast. Like, it's, the names are insane, and I'm sure I've gotten two-thirds of them wrong, and I don't

Starting point is 01:00:28 care because it's Google's fault. You've gotten them all right. The names are dumb. There was a new one. There was a new AI thing in YouTube called Music AI Sandbox, which I confess I don't totally understand because there's been a bunch of kind of overlapping AI music things inside of YouTube, and they glossed over it so quickly that I think it might just be kind of an amalgam of some of those things already.

Starting point is 01:00:49 but that's a thing that exists. They're working with musicians on that stuff. New Gemma models, which are Google's open-sourced AI models. Circle to Search is in Android doing more stuff. They had a bunch of little things in Android designed to make kind of Gemini all over the operating system in interesting ways.

Starting point is 01:01:11 So the Android piece is really interesting to me, like in a big way. I don't know if any of it will come true because Google but Dave Burke got on stage and started demoing AI features in Android and he's like this works as we control the operating system

Starting point is 01:01:27 and you put that sort of right next to we've combined the Android team and the Pixel team under Rick Austerlo and you're like oh like this is the like if you make this really good this is the thing that is actually compelling to switch from

Starting point is 01:01:41 Samsung phone or iPhone. This is how Google thinks it wins for sure both with Pixel specifically and with Android in general. Like, I think this is true or false the thing they have identified as their chance to leapfrog Apple in a really dead way. But they've always thought AI was their way. Yeah, but now it's like the user interface for the operating system is AI. Like, they've built like a super rabbit.

Starting point is 01:02:04 You know, it's like that's the thing that they're getting at. It's like, you can just be on your phone and like circle a thing and like the robot will like go do it for you. Or like, you can just talk to it and like use the apps for you. And unlike Rabbit, Google has some. AI technology. Yeah. Well, and there was one thing that they talked about where, uh, the sort of context awareness, where it'll actually understand what's on the screen as you're talking to your phone.

Starting point is 01:02:28 That's the kind of thing that you genuinely only can do if you control the operating system. Like no one else has that access, but also, uh, Google has been trying to do this for forever. Do you remember Google now, which was just the same damn thing? Like the, the promise of so much of the large language model stuff right now is all this stuff we've been trying to do for 15 years we can do now because the tech is better. Like these ideas are not new. It's so funny. They just think we have the tech to do it now.

Starting point is 01:02:56 Or the tech to do it convincingly. I just have to keep putting that asterisk there to do it convincingly. Maybe not to do it. Yeah. That's fair. To at least make you want to try it. Yeah. Is as far as they have to go.

Starting point is 01:03:09 There's more. There's a bunch of search stuff that we should talk about. But it was basically like they made all the models better and are starting to introduce some more specific models. Like I think Gemini 1.5 Pro Flash, in addition to being an unbelievably stupid name, is really interesting because it's just a different version of Gemini that is just designed for speed. Its whole job is we just paired everything down and made it as fast as we possibly could.

Starting point is 01:03:33 And I think you're going to start to see more little things like that as they tune to these kind of specific use cases over time. So it really top to bottom was just, here's the thing we geminied it. Is it better now? Do you like it? That was the whole vibe. So I will say, I think the most important one is Google Focus. It's the one that will hit the most people the fastest and change how people use that application, which I think is cool.

Starting point is 01:03:56 Like really cool. Then there's the rest of it, which, you know, I think the major criticism of Iowa is like, there's too much stuff. It was too unfocused. Like, they just said the word Gemini and Million. Like, if you just read that list of things named Gemini, Genini 1.5 Pro, Gemini 1.5 Flash, Gemini Live, Gemini, like, it's insane. Yeah. They announced the thing where they're like, here's a prototype and idea where you have an A. teammate who has their own

Starting point is 01:04:19 Gmail address, she just like yell at it. Yeah, his name was Chip. It's like, this is a ridiculous. Like, this is pure vapor. Yeah. Like, you are announcing vaporware. There's no reason to do this. Like, you're just, why, right?

Starting point is 01:04:31 You could pare it down to like, here's our product and here's all the things it can do. And that'd be more focused. But I think Google thinks it has to, like, cover the waterfront of every idea you might have with AI. I'm like, we had that idea too. And we have the infrastructure to pull it off. And we have Google workspace. So we can just deploy 100.

Starting point is 01:04:47 AI teammates to your Google workspace tomorrow. Startups, don't you dare. Yeah, I mean, I really, like, the way I've come to think about all of these things is that they used to be kind of half for developers and half for like the general public to sort of show the world what you've been working on. Now it's like a third for Wall Street. Yep. So that you can make sure investors trust that you have a real AI strategy.

Starting point is 01:05:14 Because fundamentally that's what a lot of this is about. Google is still fighting the idea that it got caught off guard by OpenAI and ChatGPT and is still trying to assert itself to investors specifically as a big player in the AI future, which means trillions of dollars if you can convince everybody that that's the case. So it's like it's a third Wall Street. It's like 50% developers and then it's just a little teeny tiny slice left for regular people. You almost did the rest of that math and you just walked away from her. 17% regular people. There you go. And it's just like, you could just feel Google getting up there over and over and being like,

Starting point is 01:05:56 we are good at AI. Yeah. That was so much of it. Yeah. Like, we did it. That's how I felt about Apple at the iPad event. They were just like the best AI piece. It's like, what are you talking about?

Starting point is 01:06:04 And it's like, this is made for analysts. So the one thing that's really interesting there, specifically in the context of Wall Street, is AI overviews and search, which they announced very quickly. So Google has had this thing called search regenerative experience, sort of in testing for a long time now, a year, where you search for something and there's like an AI answer. And it's like, here's this stuff. And they've been tweaking it and messing with it and like seeing how it works for a while.

Starting point is 01:06:30 And what they announced at I.O., which is a huge deal, is it's rolling out to everyone in the U.S. like this week. It's going to be called AI Overviews. And it's, that's it. This is search now. And on earnings calls last year, Sundar was saying, this is the future of search. Like, search will be this thing.

Starting point is 01:06:47 This is where we're going. And here we are where you search for something on Google, and depending on the query, Google just reads the web and delivers the answer for you. And next to that, they have this thing, AI powered search page, which is basically what ARC browser is doing, right, where you, like, go ask ARC a question and it, like, reads the web and it, like, makes a little web page for you with all the answers in it. Google's going to do that now as the search. engine results page, the SERP.

Starting point is 01:07:16 Instead of 10 blue links, it will just make a little custom web page of, like, best anniversary restaurants in Dallas and be like, we've concocted this webpage for you. That's wild, right? This is like, flip the table on how the internet works in a big way.

Starting point is 01:07:31 I would say the response from the media industry in particular is apocalyptic. Like, that's the word I would use. Fair? The CEO of the news media lines, which I think we have to disclose. Like, I think Fox Media is in it.

Starting point is 01:07:43 It's like not my thing. But someone's in it. Sure. We are in the media. We are in the news media. Yeah. We run, we operate in the media business and our executives are in various media business organizations.

Starting point is 01:07:55 That's their side of the house. I just complain. But there's a disclosure. But the CEO of the news media line said to CNN, this is catastrophic to our traffic. Like Google's going to keep all the traffic and they're not going to send us in traffic. We've been talking a lot about little sites that are already seeing Google traffic go to zero. I've been talking for years about this thing I call Google Zero. which if you've been listening to Verge Has for a long time,

Starting point is 01:08:16 you know that like our first big referer when we started was Yahoo. And we would like write stories about fish technology because the Yahoo algorithm loved fish. And we would just like do that on Fridays to fuck with Yahoo and it worked. I don't know. That's how I built the system. It's like I'm sorry. And then, you know, there's like Facebook traffic for a while and went away. And I'm always like the thing, it will go away.

Starting point is 01:08:37 That's like my inherent paranoia after all this time in the media. And I'm like, that number will go down. and that's Google. And that's the last referer of note across the entire industry. I've been calling this Google Zero for years. Like, what are you going to do if your Google traffic goes to zero?

Starting point is 01:08:51 On Decoder, I asked me to executives all this time. And this is the moment where I think it clicked into reality for a lot of those executives that Google is going to start answering the questions by reading their sites and delivering an AI summary of the answer. And Google's response to this.

Starting point is 01:09:05 And again, this is, we move through this quickly because I talked about this with Sundar and Decoder quite a bit. That's going on Monday. But Google's answer to the, this, Liz's answer to this, Liz Read who on search, is actually the links and the AI overviews get more clicks. That's their answer.

Starting point is 01:09:21 Like, straight up, they're like, actually, we're going to send more traffic this way. No one knows if that is true. The only people who know the data to evaluate that are Sunopor Chai and Liz Read. And it's like, well, if that's true, that's great. If it's not true or worse, if the halves get bigger, like you send more traffic to the big corporate sites and no traffic at all to anyone else, that's bad for the, like, that's bad, right? That will reshape the internet in particular ways. And maybe you think this is great. Like, maybe you think that indie media or whatever small sites should build their own network

Starting point is 01:09:58 of traffic and their own, this is our belief, right? Like, my belief very strongly is we should have our own audience away from platforms. But if you are like operating today and you're like, oh, my traffic is going to go to zero, my Google traffic is going to go to. zero. You don't have the runway to like do the thing. Right. They just turned it on. Like it's just on now. Yeah. Yeah. Like your business is going to go away. And that is the thing that I think the pressure from open AI, the competitive pressure from open AI and all these other companies, it pushed Google to do a thing very quickly that's going to reshape the internet in almost indescribable ways because the traffic flows of search traffic are about to like table flip.

Starting point is 01:10:39 Yeah. I do think the, the, the, the, example that most immediately keeps coming to mind for me for this is the what time does the Super Bowl start kind of thing. And that to some extent has already died thanks to the knowledge graph, which now is just like a, it just pops up a thing at the top. But what I talked to Liz read a few days before I.O. And one of the things that she said almost in as many words is that there is a type of information on the internet that Google just views as commodity information that I shouldn't need to go to another website to know the answer to, right? Like before, if you wanted to know the years that Abraham Lincoln was the president,

Starting point is 01:11:18 you would search that on Google and then you would click on a link. And Google at some point has decided that that is just commoditized information that belongs to no one, and so I should have to go nowhere to get it. And then if you go all the way down the spectrum, you get down to like human perspectives and original stuff and really beautiful art. Right. And that is the stuff that I think Google earnestly believes should still belong to those people and that Google's job is still to send you as the user to those places. And I think Google would like you to believe that it is very obvious where one of those things ends and the other begins. And in fact, the whole middle ground, which is most of the internet is somewhere in between and there is just no line that it is possible to draw between what is it just sort of out there in the ether that everyone can, know and the idea that you and I both used some of our time to write a story hoping we'd get some clicks and some ad sense, that that's actually like a unnecessary thing for the enjoyment

Starting point is 01:12:19 of the internet, and maybe that should go away. I'm somewhat receptive to that version of the conversation. Where that ends and good things that deserve to exist begins is now up to Google. And that is what scares me about this. Yeah, and they have not really articulated anything beyond like the cream will rise to the top. Yeah, they say high quality information and human perspectives and a bunch of just like vague buzzwords that don't give anybody anything. Isn't this like just what Ask Jeeves was doing for decades? Yeah. Like, it didn't we all to decide that that was a stupid way to interact with the internet.

Starting point is 01:12:57 And that's why Ask Jeeves still exists, but way fewer people use it now. I think it still exists. It's Ask.com now. Sorry. Jeeves was fired. Please. Keith's not canceled. Some old tweets surfaced and Jeeves is no longer with us.

Starting point is 01:13:13 But again, Alex, that's a perfect example of like none of these are new ideas, right? The only thing that's new is that everyone has decided it's now possible to do this well. Yes. S.com does TV show reviews now. Oh, my God. Sure. So that's pure just like Google Day. Perfect example.

Starting point is 01:13:32 Right. I mean, this is like why we spent a year last year covering the. the SEO industry. Because I think it's important to describe the thing as it was before you can understand how it changed. So we're like, let's just, these are the lot, these are the glory days of the SEO industry as it exists today, which is you buy ask.com and you search for which TV shows are trending and you shove a bunch of content onto a high ranking domain to collect some pennies.

Starting point is 01:14:00 This is happening by the way across the media industry. And that's happening at like AI scale. Like you buy Sports Illustrated and you show. a bunch of AI content on it is a real thing that has happened. Yep. Yep. Right? You buy these like zombie domain names.

Starting point is 01:14:13 What's the last one? The all. Gizmodo. The all. Gizmoda. Oh, no. Was it the hairpin? The hairpin.

Starting point is 01:14:19 It is a hairpin. That's what I'm thinking of. Like this beloved women's blog and you shove a bunch of AI content on it. Like that's already happening. This should wipe all that out. Right. I mean, that's, I think, what Google would like is to wipe all that noise out. That implies, by the way, that Google can make good deterrent.

Starting point is 01:14:36 about what's good and what's bad, which the current state of Google does not provide a huge amount of evidence for. It also implies that Google will prioritize the beautiful human content over the AI content when a huge part of Google's business is making the AI content. Weird. And monetizing the AI content. Yeah. That actually Google has an incentive for that content to exist at giant scale. It also kind of pulls the curtain back on like Google, probably more than Google should want it to. Right? Like we, we, especially in our industry,

Starting point is 01:15:11 we know that Google is an enormously powerful thing that really controls a lot of conversations and stuff simply by virtue of the fact that if you want to engage in these conversations, you first go search for them, and the first place you go search for them is Google. And so Google kind of dictates how you perceive these things. If Google just starts answering the questions, it is like the theater of, oh, I search Google to figure this out, and I have the agency of finding the answers is gone. It's just Google. Dude, this is the phrase.

Starting point is 01:15:38 Yeah. They use this on stage. Let Google do the Googling for you. This is the frame for how they talked about search. Horrible. A horrible way to do it. It's a lot there. We can talk about this forever.

Starting point is 01:15:47 Yeah. I mean, I personally can talk about this forever in tiresome ways. And I assure you that if you work at Vox Media, you have heard me talk about nothing else for four years. Like a lot of the verge is designed around Google Zero. And the point is, I don't think. anyone deserves traffic from Google, right? That expectation, I think, has actually been really unhealthy for the media industry, that, like, we're entitled to traffic from platforms,

Starting point is 01:16:15 especially for us because we cover the platforms. I would like some distance there in a real way. And also just, like, I'm a brat, and I don't want to be dependent on anyone ever for anything. But this is a pivotal moment. Like, Google waving through, we're going to do AI overview and search for everyone in the U.S. And then sort of immediately pivoting to, oh, and also you can search your Google photos. That was, they like lit a bomb and they're like, and you can find your license plates. And they spent no time on it.

Starting point is 01:16:43 It does not appear that anybody knew they were going to do this. We'll see how it shakes out. I think the next 12 to 18 months are going to be an absolutely bananas time in media. I think this is the thing that's going to, I would predict that there are some lawsuits filed over this from the various media lobbying groups, the same way that they've already started suing Open AI. there's going to be a big shakeout from AI overviews. And a huge part of it, and this is just what I'll focus on, is Google's claim is that it will send more traffic. You cannot measure that claim inside of Google's own tools today or with any data that Google is providing.

Starting point is 01:17:19 And if they want to close that loop, they have to make it testable that this is true. You have to be able to see the data for yourself and say, okay, like we were the one that showed up in the AI overview box and we got way to more traffic from it than we were getting, like, you cannot measure that today. So we'll see, but I, you listen to the decoder or Sundar because we talked about a lot. That's going on Monday, but we're going to be talking to this. I guarantee you for about 18 months because the internet is going to reshaped around this particular feature in search. And it's going to change a lot between now and then. Like there are, the weirdest thing about this is like SGE changed a ton in a year. And I think Google is maybe going to be surprised at how different.

Starting point is 01:18:01 it looks once you give it to everybody. Yeah. It's going to be weird. All right. We should take a break. We'll come back with the lightning round, which as of yet is unsponsored, but the bites, they're coming. It's going to happen one of these days. We'll be right back. Support for the show comes from MongoDB. If you're tired of database limitations and architectures that break when you scale,

Starting point is 01:18:27 it's time to think outside of rows and columns. Because let's be honest, you didn't get into tech to baby. set a broken database. You got into it to actually build something. MongoDB lets you do that. It's flexible, developer first, acid compliant, enterprise ready, and built for the AI era. Say goodbye to bottlenecks and legacy code. Start innovating with MongoDB. There's a reason it's trusted by so many of the Fortune 500. And that's because it's a platform built by developers for developers. MongoDB. It's a great freaking database. Start building at MongoDB.com slash build.

Starting point is 01:19:12 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.

Starting point is 01:19:40 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.

Starting point is 01:19:54 I don't care if you're gay. I don't care if you have all that. That's like secondary, third. Like, that's not a priority. That's this week on America Actual. Let's begin. Complex and unprecedented, the Spanish authorities are calling it. Before the disembarko, asymptomatikas.

Starting point is 01:20:17 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 president. 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.

Starting point is 01:21:10 All right, we're back. It's the lightning round. Sponsored by me, Neil Epitale. Thank you. The guy you coined the phrase. I didn't get any of that money. I don't know. It's like, you're so generous. Where is it? I'm just moving money around my own house. I was just doing tax fraud. There's kind of a lot. lighting around. David, do you have first? So there's a Microsoft Surface event on Monday, which there's been a lot of hype for it because Qualcomm and Microsoft and others have been kind of quietly intimating that they were about to get a new generation of Snapdragon chips that are going to be like as good as the M-Series Apple chips and the war between Mac and PC

Starting point is 01:21:48 is about to go on. And then we got a huge leak from Dell. I think it was a 311 page document detailing the next XPS13, which is Dell's best laptop. It's like the default which Windows laptop should most people buy. It's the XPS 13. And the things that they're saying on here are pretty unbelievable. 12 hours of battery life, 13 hours of battery life on the lower spec models, local video playback for 29 hours, tons of power that like, at least according to the marketing copy,

Starting point is 01:22:24 the things that we've been thinking might be true of this run of snapdragon processors might actually be true of these snapdragon processors. I just want to highlight that Apple last week, like, in kind of a surprising move, announced their M4 processor in an iPad. And I don't think that's an accident, right? Like, Neelai, you were talking earlier about how a lot of these things are for Wall Street. Apple did it for Wall Street because that is very aware that this is happening right now.

Starting point is 01:22:58 Yeah. We'll see. We'll see. Yeah. I would be remiss if I didn't note two things. One, boy, we've heard this story before. Oh, yeah. So, who knows?

Starting point is 01:23:08 But two, a lot of what we have, a lot of our coverage is in Tom Warren's new newsletter, Notepad by Tom Warren. Yes. Which is his new newsletter about all things Microsoft's, particularly AI, the future of PCs in this particular way, and whatever Microsoft is doing in gaming, which I won't even continue. continue to talk about because it will immediately be another hour of the first chest. You can sign up for a notepad now. You can get it in a bundle with command line. We have two paid newsletters. You see your empire is growing?

Starting point is 01:23:38 I hope there's some like arrows pointing down saying, buy now. Subscribe to Tom's newsletter so that he can buy one of these new XPS 13s. That's really what Eli is saying here. Tom needs the battery of life so he can write his newsletter. Please subscribe to a notepad. Yeah, he's got to write a lot of those every week now. Alex, what's first? Have you heard of P-S-S-S-S-S?

Starting point is 01:23:57 That's a great name. Otherwise, no one is, it's P-P-S-S-P-P. You could say it like that, which is how it's supposed to be, because it's P-S-P. That's how you lure cats, right? But it's P-S-S-P-S-P. You just keep saying it until somebody tells you to shut up, and then you're like, yeah, do you want to play it? It's a new emulation software. It's the PSP.

Starting point is 01:24:21 Do you guys remember the PSP? Oh, yeah, I had a PSP. They had one of the great cartridges of all time. I broke one of those cartridges, but I loved the game, and I was, like, poor, so I couldn't go get and do one. And I'd always have to be like flicking the plastic just to get it just right and then putting it back in. Only Sony invents minidisc and then invents another mini-disc for the PSP. It's somehow a little worse. But now you can play it on the iPhone, which is really exciting because it was like a tank game and I really want to play that tank game.

Starting point is 01:24:52 I'm going to be able to play it on my iPhone now. And also RetroArch, which is like the emulator software. So RetroArch lets you basically play just about any kind of. If somebody out there has created emulation software for a console you have never heard of, you can probably find it in Retroarch. Nice. Yeah, that's real. The software itself is a little, it is not the most user-friendly. It is not going to be as neat.

Starting point is 01:25:17 It's hideous. It is the ugliest app. It's the ugliest app. It's all right. Yeah. But it works. I just put it on my Apple TV. And almost as soon as we are done recording, I will be putting a bunch of games on my Apple TV and testing it for work and definitely just for work.

Starting point is 01:25:37 But it's cool. And it's just remarkable because, like, even a couple of years ago, you were, like, trying to figure out how to get all of these emulated games onto your TV. And it was really limited. And you could do it like a raspberry pie. you could mess around with Android, all of that. And now you can just do it on an Apple TV. And it's like, welcome, Apple. I've been here for 10 years, but I'm excited you're here because I kind of like my Apple TV more than my Shield TV.

Starting point is 01:26:03 I'm so sorry to all the Shield people. I'm sorry. Dude, end to the show. It's in the guest bedroom now. Wow. I have been stewing on the idea and being increasingly angry about it. The Apple TV is the only good set-top box left. they're all bad but it is the least bad

Starting point is 01:26:22 because it is the one least destroyed by ads and scams and all kinds of other nonsense I watched Palm Royale because it was like do you want to watch Palm Royale and I was like Apple has figured out that it can just do it can just be a Roku yeah it's headed there an extra it's infuriating but I will say retro arc in particular

Starting point is 01:26:41 being on the Apple TV it's like a native Apple TV app and it just native is strong but it exists on the Apple TV and some of the others work with like airplane and stuff and are starting to be pretty good. But the Apple TV as a like retro console thing is going to be a big deal. Yeah, you can just do it now. Retro, like there's going to be a lot more that will probably be much more user friendly,

Starting point is 01:27:03 but retroarch is the big one. You can just go play it now and that's dope. I keep thinking about something David said a couple weeks ago, which was all of this regulatory stuff was really like hard to understand. And then the emulators happen and now everyone gets it. like a little bit of regulatory pressure, opened up the application model on the, and it's like, oh, everyone wants this. Yeah. And it's like, particularly for the Apple TV, Apple has insisted that you, like, do its dumb stuff and it has not won.

Starting point is 01:27:33 Like, that product has not won. And it's by far the smallest install base. It has the least, like, the idea that everyone's going to be playing games in the Apple TV, just like did not come true. And now it's like, oh, shit, everyone's going to be playing games with their Apple TV. Yeah. I have never wanted to actually hook up a controller to an Apple TV before today. Like every other time, they'd be like, you want to play this game on your Apple TV? No.

Starting point is 01:27:55 Don't ask me that. I'm not going to lie to you. Like, no. And now I'm like, okay, I want to do that. I want to play Grotto Trigger on my TV. I honestly think the emulator stuff is just the biggest indictment of apples and, like, close-fistedness that you can get. Because it's like the second you let people do good stuff, they're like more excited about these products and they have been a long time. I have two in my lightning round.

Starting point is 01:28:19 One of them is just telling David to talk about his iPad reviews, which I know you already talked about earlier this week, but I wasn't here. So it feels like everyone landed in exactly the same place, which is, boy, I hope WWC involves an iPad of US up to it. So yes and no, the conversation around this has actually been really interesting because most of the reviews, including mine, basically say that. Like it is a, I actually worried I was going overboard

Starting point is 01:28:45 gushing about the hardware. There were others who went much further. Like, with the Ibed Pro in particular, like, it is a spectacular piece of engineering and design. But the question forever is like, what is this thing actually for? And I think what has come true this week is there's been a subset of people who are like, yeah, this thing needs to be more like a Mac. Let me run Mac apps.

Starting point is 01:29:09 Open it up. Let me do whatever I want. Like, give me a Mac in this body and I will be happy. And then there's a much. different set of people who are louder than I thought, who are like, no, this is, that is not the point. I, like, maybe I want a touchscreen Mac will have that conversation separately, but the iPad is supposed to be different. And one of the things I said in my review is that, like, after talking to people at Apple and about the iPad, and I've just been covering this a long time, like, Apple's running

Starting point is 01:29:35 theory is that it can do, like, whatever the opposite of death by a thousand cuts is, that, like, rather than have sort of one big thing that makes the iPad for everybody, it can have one that is for you, Neely, and one that is for you, Alex, and one that is for me, David, and we'll all buy iPads and we'll use them differently. And that that for Apple is, like, the ideal outcome. And there are a lot of people for whom that idea is, like, really compelling and really romantic. Those people tend to be people who have that specific thing in the iPad already. People who love the pencil or people who, like, do the insane architect thing of, like, holding up the iPad and walking up the iPad and walking around the table to show somebody your designs, which I'm not convinced is real.

Starting point is 01:30:16 We got an email from someone who, I don't remember the email specifically, but they're either, they are an interior designer or they're partners in interior designer. And they're like, we just run the iPad to the ground. Yeah, it's a thing. It's great. Great. Yeah. So I think the question then for Apple is like, okay, do you try to find the big mainstreamy thing,

Starting point is 01:30:36 which I think the obvious answer is make it more laptop-y or just like open, it up so that people can do more stuff on it? Or do you keep just chipping away at this thing, like one tiny use case at a time until you've built the magic? And I think like that second approach is way more fun and interesting and leads to something very cool in the long run. We're also 14 years into this and Apple has found two. Yeah. And to me, the thing that I got, I was, you know, like on planes all week basically just like watching people angrily posted each other about the iPad. And I was like,

Starting point is 01:31:11 these are the same iPad reviews we've been writing for years. Yeah. It's the same iPad. Like, I forgot that I wrote this iPad review in 2013 when the first iPad came out. Like, the same exact review

Starting point is 01:31:22 running iOS 7. Like, boy, it would be cool if I could do stuff with this. And then I wrote a version of that review in 2018. And then David, you wrote a version that review yesterday. And then I just saw Dieter at the Google event. He was like, I wrote these iPad reviews.

Starting point is 01:31:33 Like, Dieter is the one, I believe, who coined the phrase, it's an iPad. It's like, look directly to camera, remember like, it's an iPad and you know exactly what that means. That's us. We did that here. And the thing that gets me about it, the thing that I've, the reason I want to talk about it

Starting point is 01:31:46 right after emulators is it's Apple's business model that is holding this computer back. Yes. It is not the capabilities of the iPad or some idealism or blah, blah, blah, blah, blah, blah, blah, blah, blah. It is Apple's business model is preventing application developers from using the full power of the iPad. The fact that this thing cannot actually run a desktop web browser, they say it's S-Sive class, it absolutely is not. No.

Starting point is 01:32:11 Nope. It just isn't. Also, if you go to docksgut.gov.com in the browser on an iPad, it will punt you to the Google Docs app. It's the most miserable experience in the world. So what's really interesting is the Google Docs app. This is, you know, all our years are reviewing the iPad, we would complain about this. And Apple would always be like, who cares, only journalists use Google Docs. And then they realize the journalists who review the iPad use Google Docs.

Starting point is 01:32:33 And so they fixed it, like faking the user agent in mobile Safari to tell, Google Docs that it's desktop Safari. So you can run the good Google Docs in Safari and the iPad. What they did not do was make mobile Safari good. Right. Or fix it in Chrome or any of the other browsers that you can run on the iPad. Right.

Starting point is 01:32:51 And the reason for that is straightforwardly a business model reason. It's not that the thing isn't powerful enough to run the regular Safari that you can run on a laptop. Right? Yeah. I have an Intel Mac downstairs. I have a 2015 IMAX sitting right over there.

Starting point is 01:33:07 that is one-tenth as powerful as the iPad Pro. It runs regular Safari just fine. It runs Chrome in a way that suggests that I should buy a new computer, but it runs Chrome. Right. The reason Apple doesn't allow that to happen on the iPad is because the web is now the most powerful application distribution mechanism that has ever existed. It is by far the most popular in the world. And if you bring that to the iPad, the App Store goes away. Because now you just have a laptop.

Starting point is 01:33:37 You have a Chromebook. A really nice Chromebook. A really beautiful Chromebook. And Apple's business will not allow that to happen. And so this product is just absolutely held back by a business decision that maybe it's fair. I don't think, fine, make your business decisions. But we should be honest that this isn't a technical limitation or an idealistic limitation or blah, blah. Apple is making a hard business decision about the application models on its products.

Starting point is 01:34:02 Regulators are starting to crack that open and people are starting to see, oh, emulators are great. Like all this other stuff that we could do is great. And the last one to fall, I promise you, is they will put a desktop class browser on the iPad and it will suddenly become a hundred times more powerful as a computer. And more and more people will be like, I can just use this all the time. It's not the file system, although I love a file system.

Starting point is 01:34:20 It's literally the application environment is completely neutered. Yeah. No, I absolutely agree. And I think the browser solves a much bigger set of problems than like, you know, X-86 interoperability with the old apps that you want to use. because like, I'm fine. There's going to be a bunch of apps you can never use on the iPad. So be it.

Starting point is 01:34:40 The web, there's no good excuse for that to not be one of those things. That doesn't even work anymore because all the laptops are armed based. I mean, that's true. Like, including Apple's, yeah. Like the most powerful laptop processor line is now in the iPad first. Yeah. Yeah. I'm just like the thing about the iPad, the thing that it's been missing to me from the

Starting point is 01:35:03 discourse from the beginning is like Apple designs every inch of this product right that they are they make the screen and the input devices and the operating system and they're like there's not a thing in this that is not an Apple choice and this is the thing that they want yes fine but then you have to be like Apple chose to limit this this specific way and I think there's a version of that trade that is more palatable at you know $349 for the 10th gen iPad than it like yes I mean I saw people spending two and $3,000 on the, and I just like, I think the argument that this should be a different device, that it's meant for different things, it has, it does different things well. Like, no, at that price, it should do everything well that it is physically capable of doing.

Starting point is 01:35:49 And it just can't. And it is, it is like, you're not letting it. I do need to fight you on one thing, David, because I was one of those people that spent way too much money on an iPad pro this week. Yeah. Oh, yeah. I got the 11 inch. It's the first one I've, that first iPad Pro I've had since you.

Starting point is 01:36:03 2018 one, that magic keyboard case, even though it's the new one and it's all much better, is horrible. Really? It's so thick. It is so, like, as soon as you drop that iPad into it, all of the magic of it goes away. And I was like, okay, so it's the exact same as my 2018 iPad, but the screen's better. Like, that was the immediate feeling I had in a really, like, demoralizing way. There is something to that.

Starting point is 01:36:29 I mean, it doubles the size of the iPad. Like, the keyboard at this point, I think, is both. thicker and heavier than the actual iPad, which is bonkers and probably says more about how thin the iPad is than anything else. But also, like, if it were a little thinner, it would be floppier even than it is. It's just an unsolvable problem. Yeah, I mean, it was the same feeling you get when you use, like, those inexpensive Chrome books that are convertible Chromebooks or whatever, right?

Starting point is 01:36:55 And it's like, okay, what do you want? And, like, yes, realistically, having it be feel like a laptop is more satisfying, but also I hate it. Fix that. Yeah. I miss this. The surface does it better. I'm going to play with yours,

Starting point is 01:37:10 because I really thought. I'll bring it in for you. The surface, having a built-in kickstand, and then all you have to do is attach the keyboard, is the correct way. Yep. I believe that with every fiber of my being.

Starting point is 01:37:22 Man. Look, Mac versus PC is back on. Subscribe to Notepad by Tom Warren for all the latest on the newly resurgent Mac versus PC battle, everything old is new again. That was my lightning round. The other one I was going to point out as Andy Hawkins wrote a great piece about just like self-driving cars have hit a wall.

Starting point is 01:37:39 And I was going to make the same comparison to AI, which is everyone got really excited about that technology. And then it just didn't do the thing. And I would just caution everyone. Like, can they actually do the thing? But you should go read that piece by Andy because it's very good. I think we were like five hours over. Yeah, we should go. Should we talk about search for a little bit?

Starting point is 01:37:58 All right. There's much more to come on Monday. Again, we have the interview of Sundar. He and I got into it. I like talking to sooner. I've talked to him once a year for a long time now, basically. He was fire in this one. So stay tuned for that on Monday. And then it says here on Sunday, we're going to do the five senses of gaming smell and taste.

Starting point is 01:38:18 I don't know what that means. I'm very worried about what Liam and David are doing with the Rochast. That sounds great. This face on David is just... You officially know everything I need you to know. It's going to be great. All right. That's it. That's our chest. Rock and roll.

Starting point is 01:38:34 And that's it for the Vergecast. this week, hey, we'd love to hear from you. Give us a call at 866 Verge11. The Vergecast is the production of the Verge and Vox Media Podcast Network. Our show is produced by Andrew Marino and Liam James. That's it. We'll see you next week.

The Vergecast - AI assistants are so back

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.