The Vergecast - Google I/O recap: AI is taking over Search

Episode Date: May 12, 2023

The Verge's Nilay Patel, David Pierce, and Alex Cranz discuss the big announcements and takeaways from Google I/O 2023. Further reading: The nine biggest announcements from Google I/O 2023  The AI takeover of Google Search starts now  Google Perspectives: the new search feature helps you find human information online Google rebrands AI tools for Docs and Gmail as Duet AI — its answer to Microsoft’s Copilot Google’s new Magic Editor uses AI to totally transform your photos Google drops waitlist for AI chatbot Bard and announces oodles of new features Google announces PaLM 2 AI language model, already powering 25 Google services  Google teases Project Tailwind — a prototype AI notebook that learns from your documents  Android’s new generative AI can reply to your texts and design its own wallpaper  Google’s Find My Device will soon use billions of Android devices to locate your stuff  Google is bringing YouTube, Waze, and Zoom to cars with native Android software The Pixel Fold is Google’s $1,800 entry into folding phones Google’s new Pixel Tablet is a $500 slate for the home Google Pixel 7A review: a better deal Google’s new Project Starline prototype isn’t a giant booth Disney is finally combining Hulu and Disney Plus into the same app  Apple launches Final Cut Pro and Logic Pro on iPad with new subscription pricing Learn more about your ad choices. Visit podcastchoices.com/adchoices

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,

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.

dropping May 14th. Tap in with us. Hello and welcome to our chest, the flagship podcast with the JBC. I'm not going to tell you what that is, because if you watch the Google Ioki note, you know you need a massive amount of processing power to understand this information. Yeah, we can't afford to tell you. Hi, I'm your friend, Eli. David Pierce is here. Hi, I'm in a hotel approximately two miles down the road from the hotel that you're currently in.

It's good stuff. Yeah. Well, explain what's going on in a second. Alex Cranz is here. Hi, Alex. I am not in a hotel, but I'm really excited to talk more about the JBC. So today was the Google I.O. keynote. It was a big one.

A lot of pressure on this one. A lot of AI-related pressure on this one. I believe we counted, and Sundar Pichai said AI 80 plus times on stage. Wow. Just a lot of pressure on Google for this one. And they delivered in pretty enormous ways. We should talk about all that. But it is true that David and I were there at I.O.

along with V-Song, Allison Johnson, Alex Heath from our team. We saw Deeter, which is really fun. And now we're in separate hotel rooms on the road. It's easier to do this podcast in separate hotel rooms. Yep, 100%. Just given how we produce it, than to actually be in the same room together. So we're all far apart from each other, but together. Do you know what I mean?

Yeah, in our hearts. Alone together, together, together alone. Yeah. Now, I would like to begin the podcast by formally apologizing to my friend Deeter. It's been a long friendship, and sometimes you go too far. and you scream for the Android police to arrest your friend at Google I know. I believe I heard that you said Wii, Wii, you. That's a fact.

That did in fact happen. This from a source. I'm not going to reveal my sources, but I did hear that Eli shouted, weo, weo, weeoo. Yeah, arrest that man. And then someone had to say, Neelai, don't scream, call the police in crowded suit. Anyway, it's fine.

It's fine. and everything's going to be fine. Anyway, so it's Google I.O. Very fun to see everybody in person. David, I don't know if you caught this. There was a lot of pressure on Google, but then there was like a lot of confidence and then a lot of curiosity from all the attendees about whether Google would meet the moment. So like the energy was pretty charged.

Yeah, it was, it was an interesting space because this was the first fully in-person I.O. in four years. Yep. Which is like an eternity. And folks were psyched to be back. It had that same kind of like first day of camp. energy that some of these events have had over the last year or so it's at the shoreline amphitheater which is just a straight up concert venue so it's like outside and cool and it's there's a big lawn and it like feels

like you're in a concert venue so it's like it had a lot of stuff going for it coming in and there was this question of like this matters a lot to google in a way that i think google i.o we have often accused to being sort of a grab bag of announcements it's like somebody just like walks around to every team at google and has like got anything and whatever they have they launch at i.o And that's what Google I.O. is. But this year, it's been very clear that, like, this is the time Google had to make the world believe that it has a plan for AI. And Sundar Pichai, in particular, had to convince the world that not only does Google have a plan for the future of AI, but that he is the right person to execute that plan for AI. There's been a lot of smoke around, like, is Sundar the guy, are the people inside of Google upset?

the information has done some really great reporting on all of the ways that kind of his authority and leadership were being tested. It's just this really interesting moment. And yet, I think you're exactly right that like as soon as the thing started, it felt confident. It was like a company on top of the world being like gaze upon our kingdom. So it was like that context switch at the very beginning of the keynote was really fascinating to me. You felt like he did it. He kind of showed it. I wouldn't go that far.

So here's what I would say. Yeah. Well, and there's a very specific reason I'm saying I wouldn't go that far. Nothing is shipping. I mean, Bard. Bard is sort of shipping. Even the new features of Bard that they're most excited about are behind labs and they're gating how many people can get access to it.

So they showed a lot of really cool stuff and then you got to go sign up. It's Google.com slash labs. And you can put yourself on a wait list and they'll let some people in. So it's just like they showed a lot of really cool stuff. Yeah. All of it is plausible, right? Nothing seemed too far afield, but it's not actually shipping. So I would just like stop right there like they did it.

Like they showed very much they can do it. They have not actually done it. The switch that David is calling out that I felt most keenly was sort of right at the beginning where Sundar was like, we have 15 products that have half a billion users. And of that six products that have more than two billion users. And now I'm just going to put the AI in front of them. And I was like, oh, right, they're Google. They can just do it.

Right. And like that, I think they had not yet made that case very convincingly. And to some extent, this still was kind of a grab bag I. It was just everyone was instructed to pull the same thing out of the bag, which was that AI to it. And so there was like thematic unity to the point where like the pixel conversation, which will come to at the end, like Rick Osterosso is like, it's an AI phone. And it always has been. And I was like, has it? Like, I don't know.

Yeah. That's weird, right? Like that, but like that little bit of sort of thematic unity that everyone has instructed to talk about AI and their products meant for huge sections of it, there was confidence. Like, Google is going to do this. It's going to add AI to all of its products. And everything they show us is kind of cool. Yeah.

Well, and at the very beginning, I mean, I think the way that these typically go is you get a few minutes of the CEO at the beginning who tells you sort of a big story about the future. And then you get a parade of executives who do the actual announcements, right? That's how Apple does it. That's how Google typically does it. But this time we got, I didn't count, but maybe like 20 or 30 minutes of Sundar actually going through some of the announcements and explaining what this is and explaining why this matters. And it's like he got up there and like explained Google in a way that you don't typically see at an event like this from a CEO like that. And it's like you and I were talking after it were it's like the people Sundar was talking to were like investors and analysts who have

spent the last eight six to eight months now wondering is Google going to lose this AI race that nobody really saw coming but is moving really really fast Google just has not seemed like it was ready for this and was it going to be able to catch up and this was very clearly Sundar standing up and saying we're ready we have a plan I've got this and we've got this and it was like Normally, these things are for developers and they, like, get really into the weeds about how all of this stuff actually works and goes into your Android apps. This was not that at all. Yeah, this felt like a road show. Like, we're going to just show off everything we can do.

And I think there were a bunch of, like, folks from the government there, too, right? Like regulators and stuff kind of also curious. Yeah, there were a lot of people there, like government people. There were people walking around. They had these signs called lollipops or like circles on sticks. And, like, one of them says press. And, like, that's how we know where to go. And this time, for the first time, in a long time, they had one that's a government affairs.

And there were just lots of government affairs people floating around. And it's because the AI risks are real. And the push to regulate AI is real. And just last week, Sundar and Saja Nadella and Sam Altman and some other folks were at the White House to have like an AI summit. And then they're going to go do that in Europe. And then they're going to go do that in the UK. And then they're, I don't know, the government of Belize is going to demand that Sundar Pratra shows up and talks about AI like existence. AI risk. So there are, they have to just be really proactive and say, and this was their catchphrase

over and over again, bold but responsible, like over and over again. And at one point, you know, they're like, there's a natural tension when we say that, we understand it. We want to take big, bold bets and what this thing can do for people, but be responsible in how we deploy it, which is fascinating because, you know, the open source models are not being deployed responsibly. People are just using them. And so Google is trying to position itself is like, you should regulate AI, but you should trust us because we're already responsible. And that was a real undercurrent of all of this. But we should start with the actual announcements. So the most important one, and David, as you said, Sooner talked for 20 minutes, and he kept coming back to this. We've pivoted

to being an AI first company like six or seven years ago. And Google's mission is to organize the world's information. And that mission is timeless. And he just said that straight out. Our mission is not changing because of AI. Our mission is timeless. We've been working on AI for seven years or however long it's been, and now it's time for the future of search. And then he showed off a new version of search, and you got to play with it a little bit. Yeah, I spent a bunch of time on Monday with some executives at Google walking me through some demo. I got to sit in a sort of half-staged red team process where they were trying to, like, test this thing. But basically, the shortest way I can think to describe it is essentially they're putting a new box at the top of search results that they call the AI snapshot.

And what it's doing is using all of these large language models that's using Palm 2, which is the big kind of general purpose model that Google announced at I.O. But also some of the other things that Google has been working on in search for a long time to take all of this information and give you essentially like on the one side an AI summary, which is the stuff you'd get from Bing or chat GPT or whatever you ask, what is cool in Peru? And it just like answers to the question. And then on the right, it gives you a few links from around the web that kind of substantiate that information on the left. If you do a shopping search, it'll also show you some products with AI generated summaries of them. If you do local stuff, it'll do the same with the local results. But now, instead of getting links at the top of search results, you're going to get this big, colorful box that is totally AI generated. And it won't show up all the time.

It won't look exactly the same everywhere. But it is like, it is a new thing at the top of search results, which I would argue strongly is, like the single most important piece of real estate on the internet. It's a complete redesign of it. And they kept saying over and over, like, the search results are still below. You can scroll down and get to web search. And that's true, but it's still like a fundamental change. And especially on mobile, where more and more people are, this just eats the top of the screen.

Like when you do a Google search now, instead of seeing web links, you're going to see this AI snapshot first. And that is a huge, huge, huge deal. I want to be precise with that. Most people will not see that. You have to opt into this experience and it's a wait list. This is what I mean. Like they haven't done it yet. Yeah.

Like the thing is still gated. Yes. Yeah, right. I think there's a difference between they haven't done it and they haven't shipped it to everybody, right? They've done it. Like, it works. I've used live demos. Like, it's there.

It's a thing. Yeah. But so they have they have this thing called, it's called the search generative experience. They call it SGE for short. And it's a feature of a thing called Search Labs, which also has a bunch of other little features in it that we can talk about it if we want to, but that's kind of the main thing. And so if you get into search labs and you turn on SGE, you'll start to see those snapshots. And their essential plan is to use this kind of opt-in group, which is going to be a self-selecting set of people,

to figure out what works and what doesn't, and then start doing the kind of, you know, 2% tests that lots of companies do where you just roll it out to a few users to see what they think, and then roll it out more broadly. So what eventually billions of people see might look very different, but this is, this is directionally where all of this is going for all of Google search. Because right now, when you, when you go and you Google something, that top thing is usually like, here's all the stuff we want you to buy from Google aggregated links. Does that go now below it or does it just disappear? And so you get like this search thing and then all the relevant links. Like are you, because right now I just, I always scroll down.

Like the top of the real estate that's supposed to be valuable, I usually just scroll past because it's like, here's the stuff Google wants to sell me. And here's the actual thing I was looking for. Here's the first organic result. Right. And so, like, does that make that experience better? In a way, yes. I think in the sense that, like, if I Google some, you know, stuff to do in Peru, right now,

the odds are if you do that Google search, the first few things are going to be ads, right? And this is like the fundamental problem with Google search is it has just been absolutely overrun by sponsored stuff. Yeah. And links to Google stuff. these sort of zero-click searches and like Google search as a way to explore the web is not nearly as good as it once was. With the AI stuff, what it's doing is combing through all of that information and pulling useful data up to the top. And it's like, we can talk about that what that means from a business perspective. We can talk about what that means for publishers. We can talk about

what that means for SEO. Like, it's going to screw up the web in all kinds of ways that we cannot even predict yet. But from a pure user experience where I'm like, I want relative relatively good, relatively substantiated information. I think in a lot of ways, this is actually going to be really great. It's going to have a lot of the issues that a lot of large language models have. It's going to hallucinate. It's going to make mistakes. Google, I think, has done a better job than most. One of the things they showed me, but didn't show in the demo is there's this menu in the AI snapshot that if you hit a button, it will break up the like three paragraph summary sentence by sentence and actually have links below each sentence with the information

that corroborates that sentence. Okay, that's cool. Yeah. It's like a constantly sourced thing and it becomes, I think is like amazingly useful and cool. And so at the limit where that really works, and again, I have not seen it enough to be able to say confidently, like, how broad that is. But if it really works, I think from a user experience, it's kind of great for lots of things. Yeah, what it will do to the web is a very different and messier thing. But as a, as a person who just like wants information from Google, this makes sense to me. So there are two gigantic ideas in there. And you hit on one of them like half. way, which is the very concept of what you ask Google for is now different, right?

Yes. You used to Google stuff to find things on the web and where Google is going, and they called this out in the keynote. They're like, I want to find clothes for a five-year-old that are good for a birthday party is not a typical Google search. And they're changing what the Google product is, what search delivers to people to be answers to general questions. instead of here's the web.

And that's just a massive shift. Like, it's worth sitting with it for a minute and thinking, okay, what if Google the product is no longer organized around I'm searching the web? And now it's just organized around, I'm telling you the answer, which is the road they've been on for a long time. And I feel like this is the last step. Can you do Boolean searches in this new version? No, I assume.

Well, it's prompt engineering. It's an LLM. Yeah, I mean, prompt engineering is essentially. just normal human bullion search, right? Like, it is functionally the same thing. But I think, Nilai, like, I think you're mostly right, but I think that's not sort of the whole story, right? So I think the way Google explains it, and I'm sort of boiling this all the way down, is

like, there are basically three things you can do on Google. One is navigational, right? It's the people who go to Google and type in Facebook, because they want to get to Facebook.com. And a gigantic percentage of people do that. Like, single word names of websites are the most popular thing searched on Google. It's insane, but it's true. It's all our parents.

Those people are super well served by search results, right? Like, if I search Facebook popping up an AI snapshot with a bunch of information about the history of Facebook is actually bad U.S. You should show me Facebook.com and send me to Facebook.com. Now I'm just doing this in a browser window. I'm like, this is actually great. This is like the best Google experience you can have. It kind of works. Yeah, and Google search is really good for that.

The second thing is basically like questions with answers, right? And so it's how tall is the Empire State Building? It's like every Google executive's favorite version of that. Or all the way down to like slightly more complicated but good questions, right? Like what's the best restaurant in my town is like a question with an answer? But then the third thing is questions that don't have a right answer and that don't have someone on the internet who has tried to answer that question. And that is questions like what you're talking about. like clothes for a five-year-old kid to wear to a birthday party with all of the other context

that potentially comes along with the things that come from that. Google has never been good at that. Like I spent some time with Prabhakar Raghavan who runs Google's search stuff. And he described this as like task completion, right? And it's like if I go to Google and I say, make me an itinerary for a weekend in Paris, it'll show me some information about things to do in Paris, but it can't pull together information about what's closed and when, what's busy, whether I have kids with me, how much things cost. They even said this about local search.

If I say, what's the best restaurant? It's like, okay, what can you afford? How long are you willing to wait for a reservation? What kind of food do you like? Where do you want to sit? What do you, like is best the most Michelin stars or the one that matches your taste? Like, these questions don't have right answers. And they're based so much on context.

And Google has never been good at that stuff. And increasingly what Google says, and I don't know that this is true, but this is what Google says, that is the stuff that people want from Google, and that's where they think AI can be really powerful. Like, they don't think, and I think this is an important distinction, they don't think a chatbot is the future of search. Like, Bard is not going to replace Google search, at least as Google sees it now.

AI is another tool on top of everything else that search already does, and I think this is a convenient thing for Google to say because it's Google and it makes a hell of a lot of money on the way that search already works. But that's the way that they see it. is like, this is another tool to do a kind of searching you've really never been able to do before on Google. Yeah, but that thing, that's a big change, right? It just goes from what Google was, which is we're going to help you navigate the world of information on the internet to we are now answering complicated questions with all this additional context.

Yes, absolutely. Again, they've been on this road for quite a while. Like, this is the promise of Google Assistant in many ways, which, by the way, came up zero times today. Yeah, they didn't say the word assistant, did they? Not once. They didn't. And I think it's because LLMs are still too slow. for like a voice interface, so they just didn't want to, like, get into it.

But they just didn't say to it. The second thing, though, that's, like, complicated here is Google has to go get all that information. It needs to know what times the restaurants are open and closed. It needs to know if critics have liked the restaurants. It needs to know if they still exist. And if they're just scraping the web and taking the information and never sending anybody to web pages, the businesses that do that work don't exist.

If Google never sends anybody to. to Yelp. And I'm picking Yelp very specifically because Yelp is very mad about this all the time. If Google never sends anybody to Yelp and scrapes all of Yelp's data, then Yelp is dead. And then Google has sort of killed itself along the way. And there's a tension there that I'm not sure Google has figured out or anyone has figured out. Just from the outside, because I didn't go to these briefings and stuff, but it sounds like they haven't. Because a lot of these things that they're doing that they're pivoting towards, this is stuff that publications already do, right? Like the best restaurants in your area, Eater, which is a Vox Media property, and a lot of other companies do that already.

Foursquare does that already. Oh, yeah. V, who was sitting next to me at one point, it was like, oh, it's what's making buying guides, which is the thing that she does. Right. Right. Like, it's already piggybacking off of that data. And I am really curious to see, like, okay, what do you do? Like, you're going to fundamentally kill those data sources if you're not sending the traffic back to them.

So how do you exist? There's absolutely no answer yet. There is one mention of like, we want to send traffic around the web and we'll work with publishers. And there is a lot of politicians there today. There's a big swirl of that in the background, but it was not really brought up on stage. It's just the question. Like, here is search. And, you know, I wrote about it this week and we're going to keep writing about it for this year.

Like, once you change this thing on the internet, the internet changes forever. And I think we saw it today. Google's intention is to change it. I don't say it as a positive or a negative. Yeah, that's just what they want to do. Yeah, and maybe it should. Like, you can't look at this product is in beta experiment, whatever it is. You can't look at it and say, that doesn't look cool.

Like, it looks very useful and very cool. And there's a feature of it where it is still kind of a chatbot, right? You ask you a question that shows you this like AI generated box, and then you can ask that box a follow-up question. That's cool, right? Like, it's just the whole thing is cool. But that means, like, all of search change, like the thing that Google, does, which is essentially direct people around the open web, is going to change.

Yeah. There's just going to be a huge set of like ripple effect changes that come from that. And I think it's just worth it to call it. Like, this looks cool. It's not a competitor at ChatGPT. Like Bard is their competitor, ChatGPT. And we can talk about all the things are doing with Bard. This is them saying, we know that everyone's using ChatGPT and saying this is a more useful way

to search when what they mean is I'm asking, random questions to a text box, and it's delivering some answers to me. And our product should already do that, and that's the road we're on, versus what they want to do with Bard, which is, like, write me an email sucking up to my boss about how good his presentation was, which Bard was like more than happy to do. Very good at sucking up to its boss. Yeah. Yeah, I think the thing that I kept getting hung up on during this presentation was when they were showing one of those buyer's guides that they created, which I think was for mountain bikes. And they had the ratings in there. And I was like, Where are those ratings from?

Right? Like, where's the sourcing of that? Because if it's Amazon, that's one sort of audience, right? Amazon's not letting Google scrape that too. Yeah, Amazon, and Amazon doesn't want to let them scrape that. And if it's like REI, that's a totally different one. And if it's like Google's random own algorithm, like, there's just that lack of clarity there that I think makes it not necessarily a good product for some of those use cases.

They were showing, like, the buyer's guides. But that might just be because I work the verge. Setting that aside, the disclosure, the verge rights reviews of tech products, not Mitenbikes yet, but we do a lot of e-bikes lately. It's strange. The Europeans work here. All that aside, the implicit assumption of this is that the economics of making the web will continue to produce the web for Google to draw upon to make its lucrative search product. And like, that's not true. This will change how the web gets made. And that's fine. Like, we should go into it. open. So that's search. It's rolling out. You can sign up for it. It's Google.com labs. It's neat. Play with it. Look at the demos. It's neat. It's definitely going to have, we're going to be talking about this for the next five years. I promise you. Oh, yeah. I'm going to ask it the worst questions as soon as I get access. It's bold but responsible, Alex. I'm going to have a bold but responsible series of questions for it. They also rolled out generative AI in a bunch of other products. Most notably, I think, in like the Google Doc suite, they call workspace. So Gmail. Google Docs, slides, the whole thing. And all of that looked sweet. Yeah. Didn't it just all, I was like, oh, that's all super useful.

Yeah, I continue to think that, like, we're in this phase of thinking about AI as this, like, huge thing that understands everything about everything and can answer all of my questions. And I think by far the more useful version of AI most of the time is going to be the one that you're just like, what is in this dumb document that I don't feel like reading? And, like, there was a demo they did in Gmail where you just say, I need to be. I cancel my flight is essentially the prompts you give to Gmail. And it goes through your email. It finds your confirmation numbers.

It finds the, you know, all the information that you need about your flight in order to be able to cancel it. And again, I have deep, deep doubts about its ability to do this successfully and well every time. But at least this is the idea. And it just writes that email for you. And it has access to all of that information.

It's just sitting there in your email. And it's able to comb through stuff that is yours, that you care about, that you have decided is relevant to you and make that stuff for you. And that's the stuff where I'm like, okay, this is real, this is something. And they showed off,

what was it, Project Tailwind, that, like, kind of AI notebook thing? That's like their notion killer, right? Yeah, it's really hard. They framed it as, like, an educational experience, but you look at it and it's just like, this is just a research tool

for anyone who stores and needs to make sense of information. And for me, it's just like, I can just, like, put a PDF of a lawsuit in there and just ask questions about that lawsuit to Google's AI. Like, that is massively powerful. This is where bold but responsible has to come into play. So they said at the end of that part of the presentation, they're like, we thought about this

for students, but like anybody, like, a lawyer could use this. And I'm like, if I'm a lawyer, and I'm like, here's all the cases I didn't read for this lawsuit I'm doing. Summurize them. And this thing hallucinates a little bit. it's like, oh, I'm so sorry you're going to jail. That's going to separate the good law students from the bad law students. Whether or not they pay for the version of Project Tailwind that hallucinates or not.

It's like the next tier up in pricing. They had another demo in Google Sheets, which I thought was really interesting, where they were talking about colleges and like, if you find me at college, it's like good if I want to be involved in video games and I'm interested in animation. And I made a list. and then the person demoing it was like, tell me if they're public or private at a table. And so it added another column to the list

and it said whether their college was a public or private. And I'm looking at this. I'm like, this is super cool. Also, all that stuff we talked about was search where it like has the button where it tells you all the sources. None of that's here. This is like the most dangerous place where you're now having it draft documents

and do stuff for you and you have no way of checking. And when you're like, make a list that I will depend on in my college search, And you're going to, like, if it gets it wrong that a university is public or private, like, I'm not sure the stakes are here. Maybe they are. Maybe they aren't. But there's a lot of trust that's downstream of that. I mean, the average tuition, you ask it for the average tuition. Then you get accepted. And you're like, oh, this is actually $100,000 more than I thought. I mean, if you get all the way to applying and you don't double check that, like, I think that's like.

Look, 18-year-olds are a different, different class of folks than us. One of the conspiracy theories that I've heard about Bard in particular is that it exists to make people do more Google. searches because you have to fact check every single thing that Bard says to you. So it pops up this thing over and over, right? That's like, Bard is an experiment. It's going to get lots of stuff wrong. And then there's that button at the bottom that says Google it. And so it's like every time I, you know, fill out this table and then I have to go Google

of these just to check to make sure it's a public or private university. And Google's like, sick. We just got 12 Google searches out of this. They announce these upgrades to Bard. Like, do you know, like, sometimes you have a friend who sucks and then like they get a little a bit better. I don't know how else to describe this. Like, what's a meaner way of saying, like, my friend got to glow up? Like, oh, shit, you're hot now. Like, you know, it's like that thing that's, like, very insulting. Bard used to be bad at math. And now it's good at math. Like,

they said this over and over again. Seriously, they're like, Bard is now better at math and better at coding. And it's like, because it used to suck before. So they've improved this. And the big change here, they upgraded the model to something called Palm 2. So the model they have been very proud of is called Palm. They've upgraded the model of Palm 2. They're very proud of how big it is. There are different sizes of model inside of Palm. And they have like adorable names. There's like Gecko and Bison and Unicorn. Unicorn's the biggest and Gecko is the littlest and Gecko can run locally on a phone. So they're very proud of this model and that's the thing that's now powering Bard. And that has led to a pretty fantastic like capability increase for

Bard, so maybe it won't be as confused and boring as before. And then it can now do really interesting things. Like, it can definitely write code. But then they showed, hey, write code, improve your own code, and then leave comments on it in Korean. And it just did it, which is pretty cool. Like, I don't know if the Korean comments were any good. Again, Bard, little shaky. But it was just cool that they could do that. And then they talked about a lot about multimodal interfaces so that you can talk to Bard with pictures or you can have barred generate pictures for you, which is really cool. And then they're doing plugins. So in addition to generating pictures for you, it can go talk to Adobe Firefly and have Adobe Firefly do generative AI images

for you. Like the flex for Google is like, oh, we're Google. And we can just call all of these companies and say, do you want to plug into our thing that's going to show up with a billion users on day one? And they all say yes. The plug-ins thing, I think, is like the most interesting story in chatbots for the next couple of years, because this is, now what everybody has decided is the next turn of this, is that the chatbot is the interface, whether it's chat bot, GPC or Bing or Bard, the chatbot is the interface, and you're going to use these plugins to pull in all kinds of other data. And that's either going to turn out, like, the App Store, which is like a huge, incredibly successful business that has birthed entire new

industries, or it's going to turn out like browser extensions, which are like neat and cool, but don't make anybody any money or solve any real problems for anybody. And it's going to be totally fascinating to see which of those it turns out to be. But I think you're right that the other advantage that Google has is it has most of those services already that it can plug in.

Like Google doesn't have to go out and talk to kayak to do a travel plugin for Bard. It just has Google travel. It can just do that. And it has maps that it can plug in. It has Google lens that it can plug in. And it's like all of this stuff, like AI is coming into search

at the same time that Google is coming into Bard in a really interesting way. And all of this stuff is first party right now, but it could get bigger very quickly. Doesn't that kind of set Google up for regulation troubles, though? Because that's essentially showing off just how big and huge and far-reaching Google is, which is the thing that has the FTC really concerned right now. I'm like, you know, rumbling about like, well, we want to write Google up because it's too big. And they're primarily focused on the advertising section of this. But like, if you can go do what all of these other apps and all of these other plugins and all these other companies can do in one place, that seems like a really powerful argument for, yeah, break that up.

Well, they got to actually do it. I think, like, yes, I think there's a reason that there are a lot of government affairs staffers walking around. Like, I don't want to, I don't think it's possible to overstate. Like, where we were, there was the press and then there was like large groups of people under signs that said government affairs. and then somewhere else there were developers, presumably. But it was very obvious Google was like courting this group of people. And you're right. Like they have to walk this line.

And when they did Google travel, they got into some trouble, right? Like they bought a company. They bought a travel agency, basically. Like all this stuff happened in the past, but it happened. And now Google owns it. And it's like pretty hard for them to get into trouble doing things that they already can do against deals that were already approved and went through. Especially when all they have to say is like, well, there's Open AI and Microsoft on the other side of the table.

Like, we're competing against the fastest growing consumer software product ever. And, like, no one thinks we're winning. Is there a Bing travel? Yeah, I mean, like, Microsoft operates all these services. And, like, they're Microsoft. If Microsoft wants to stand up a competitor to Google, like, it definitely can. And then there's all these other startups. So I think that moment for Google is they're about to say we're adding AI to all these products.

And they need to keep their regulators happy. But they haven't, like, like I keep saying, haven't actually done it. Like, they're just showing us these demos that are really, really cool, but they haven't actually shipped it yet. They're being very, very careful, and they're saying out loud, here's all these bold ideas we have. We're going to roll them out responsibly.

We know we've got these problems. To that end, one of their biggest announcements was every time you make an AI generated image with a Google tool, it's going to take, it's going to get a watermark and it's going to have metadata embedded into it that says it's AI generated and it's been synthesized media, and they want broadly the industry to adopt that. Is that going to happen? I don't know. I answer a question.

But that's where Google's head is at. We're not going to give you these tools until we've done what we can to make them safe and convince governments around the world that they should trust us that we can do these things. I think there's power to that approach for them. How much for you guys being there in the room with them, do this feel like kind of a response to a lot of those regulatory concerns? I'll give you one example.

Straight on. They showed a really cool demo where. Someone was speaking in English, and then they ran it through an AI. And I think it was in Spanish, right, David? I think that's right, yeah. And it came out, and the person's voice had been remapped with perfect intonation into Spanish. And in the video of that person speaking, their lips had been remapped from English into Spanish, too. Right.

Super cool. And they are staring at a group of government affairs people, and they're like, we know that there's a downside to this. Like, we know that this is, like, a very dangerous tool to unleash on the world. They actually said the words deep fake on stage. Yeah, we can now deep fake in every language. And they're like, and our solution to this is we're only going to give this tool to our trusted partners. Which is on the one hand, great, right? Google's being responsible.

On the other hand, comes with a huge set of questions about how they will decide which partners to trust and what, whether or not they will cave to bad faith criticisms. Right. Like, all these big tech companies have caved to bad faith criticisms about who they trust and why they trust. trust. Okay. And then it also is like kind of funny because it sort of implies that Google is the only company in the world. And it's like, y'all, like, people are doing this on their laptops today. Right. Like, deep fake revenge porn is a problem today. So you promising that you won't give this language translation tool to people you don't trust, like solves this problem 2% because the real problem is like arbitrary code execution on laptops, which no one. can stop. And maybe no one should stop. And there's, there's a tension in there that, you know, we could talk about for like the rest of our lives, probably. So like in the room, while you guys were in the room, did it feel like Google was doing a good job of saying that they were trustworthy?

Like, did you feel, did everybody seem to sense like, yeah, these are the trustworthy ones. These are the ones that are the trust for AI versus Open AI, Microsoft, and everybody else that's kind of in this space right now. I have spent a lot of time trying to decide how I feel about this, because all of my instincts say that no, everybody is moving too fast and nobody is being trustworthy. But also, like, in talking to these folks and even in watching the keynote, I kind of believe Google that it's trying to do this. I mean, one of the things that I think seems to be true is that these kinds of products have existed inside of Google for a long time, and it has been slow and careful in how it released them to the world. Right. And then I think when chat GPT and then

being happened, Google's hand was forced, right? And so it started having to play its cards much more quickly and much more aggressively. And I think whether that was the right call remains to be seen on a variety of different levels. But I do think Google is trying to do this the right way. And I think that's partly self-serving, right? Google has a gigantic search business to protect. Google has a huge number of regulatory issues that it's fighting against. So Google has a lot of really self-serving reasons to try and be the good guy here. But at the same time, I think Google is trying to be the good guy here. So, like, incentives aside, I do think the company is trying to be responsible.

And it's possible that I'm just, like, being taken in by the narrative they're trying to spin, because that's clearly the narrative they're trying to spin. But that is what it has felt like to me so far. There's a lot of people at Google who really truly believe that they're five years away or 10 years away from AGI, artificial general intelligence. And, like, that they should be the old. only ones to do it because they'll get it right and it won't kill everybody. And like, that's something. That's a certain amount of confidence. It's a take. Yeah. I would say more cynically,

if you're Google and Microsoft, what you want to be is like the regulated monopoly. And I think if they can convince a bunch of lawmakers to effectively, like, we only trust AI to these three companies that we will regulate directly. Like Google's like, yeah, we'll be one of those three companies. I just, it's just like the kids still have laptops, man. They're going to figure out. I don't do it, whether or not the government says, and like, there's, there's just something in there that's, like, I'll give you the example of AI Drake. My favorite, my favorite character in the entire game. You want to stop AI Drake, right? Okay, what are you going to do?

You can go to Apple and say, don't let AI Drake into logic. You can get a pro tool. None of that makes sense. That's not going to stop it. You can go to Google and Microsoft and say, don't make tools. Oh, okay. But that's not going to stop it because they don't. Open source tools just let you do it.

You're going to go to Dell and Intel and say, don't let this code run on your laptops. There's just, at some point, like, the kids are just going to run software and they're going to make AI Drake. And like Google is in that kind of position with all of this stuff. There's a really interesting leaked memo or allegedly leaked memo from inside of Google that is I think the headline is, we have no moat. And it basically argues that it's not actually competing with Open AI, that they're competing with open source models. Like the meta model got leaked and they're saying we're competing with open source. Like that community is going to go faster than we ever will because we're, you know, we're like wringing our hands about safety and worrying about open AI. And that,

that's just running at the speed of like community on the internet. And like, there's just something there that like, what's Google, what are a bunch of Google lawyers and a bunch of government lawyers going to do in a room to stop a bunch of teenagers making AI drink? Like that is just an unfair from the jump. And like I think that's going to be. one of the most interesting dynamics that we see play out, because at the end of that all, like a big decision is, will we let people run whatever software they want on their laptops? And I think the answer is almost always going to be yes. Yeah, well, it just makes me think of like that's one of the reasons Google is nervous, right?

Because it's going to be a lot easier for the government to regulate Google than it is for the government to regulate AI. And I think it may see it as a proxy for regulating AI to tell a couple of these companies what to do. But it's like shutting down Napster didn't end piracy, but it is one button you can press that does some of the work. And I think if I'm Google, I'm worried that I am the biggest most pressable button in that space right now. Yeah. One thing I definitely want to talk about in like the tweets came in instantly. Google demoed a new magic editor function for Google Photos, AI-based, genera-based, where they were like,

you know, now you just added a photo. And the demo was a child holding balloons on a park bench, and they just circled the child, and they scooted him over on the bench. And then the genre of AI just figured it. And like, my Twitter is pretty quiet. I haven't tweeted in a long time. But, you know, during events I keep it open just to see.

And the immediate instant, what is a photo replies? Like this I think I put it in live blog this is a singularity level event for what is a photo like it is right there with produce an image of my child as a text prompt right we're looking at a photo and we don't like its composition we don't like the way the sky looks we don't like this and with a tap with a circle we can just instantly change it yes there's some Photoshop conversation there but this is different like meaningfully different than that stuff yeah the demo the that really got me is there's this picture of a woman standing in front of a waterfall and doing the thing where you sort of stick your hand out as if you're like, you know, catching the waterfall or like the thing they do where you're holding up the leaning tower of Pisa or whatever. And she got the angle slightly wrong. But with Magic Editor, you literally just drag her over to the side of the photo so that she gets the angle right. And it's literally it's just that simple that just like tap on the thing and just slide the woman over in the photo and then put it back and it looks just as normal.

And you can change the sky, you can change the background. It's the kind of thing that it's like, this is like the natural finish of Instagram filters. If you just like take Instagram filters and play it all the way out to the end, you have this where you can just change every part of your photo however you feel like. And this is another one of those. Like on the one hand, like as a person who tends to take just slightly wrong photos all the time, this kicks ass. I'm going to use this all the time. And it's coming to Google Photos. So it's not just pixel specific.

like this is going to be applicable to like many photos that I take from now on. And yet I'm going to look at every one of these and be like, am I lying to the universe about this photo that I took every time? Right. No, it's like Photoshop exists. Right. And like if you just think about Instagram is a really good example here. Because if you think about the Instagram norms that have changed over time,

do you remember when using DSLR photos on Instagram was like kind of a faux pa? Yeah. Like that was a norm on that platform that you were only supposed to use your phone camera. It was like weird. And that's gone. Like obliterate. by the Kardashian machine. Like, DSLR photos on Instagram are there all the time.

Mirrorless photos and Instagram are there all the time. And then it was, you shouldn't Photoshop so much on Instagram. Like they have the filters, but like a straight Photoshopped, it still needs to be a little bit raw. And that is fully obliterated too. And now it's like, oh, now everybody has this like completely like hot shit version of Photoshop that's AI powered. There requires no technical skill.

And you're like, the example that showed was I think it was a gray sky. and they're like, do you want the sky to be as blue as you remember it? Which is an incredibly philosophically loaded question. Yeah. How blue do you remember the sky being on this day? Like, I don't remember how blue the sky was in the day I proposed to my wife. But if you asked me, it was the bluest it has ever been in world history. Should it be that blue?

Like, I don't know. And so, like, you just get to this place where it's like, you really could be like, as an influencer, like, you know, where's that place in Italy with it, like, everyone stands on, like, the hand and the cliff. Do you know what I'm talking about? you can just make a photo of me standing on this thing. Like, whatever. Yeah. Like, does it matter if I was actually there to take this photo if I'm going to have the AI totally correct it?

It does because you're still a liar. Like, when you lie and do a Photoshop, you're still like a liar. But I'm saying the norms on Instagram, right? And then they had another AI thing where like Bard was generating captions of cute puppies. And you could just see this thing being like, all right, I'm going to like AI generate the photo and then AI generate the caption and like automate. in Instagram account. I'm telling you, singularity event for what is a photo.

I want to be more angsty about it, but I just can't because I keep thinking about every time, every time we've seen these turns on Instagram, at some point everybody's like, hey, you know, this is all just lies, right? And then there's like a whole thing like Jezebel and all these other sites are like, these are full of lies. This is destructive to body images and stuff like that. I'm not saying Instagram is more important than it is. I'm saying it's a useful thing to examine how like norms and effects.

photography community change over time. Yeah. So you start with, you shouldn't even use a mirrorless camera on this platform. Like that was a thing. And you're like all the way up to Photoshop is fine. You're all the way up to like there's a lot of professional photography Instagram. And now you're all the way up to, yep, maybe there'll be some blogs talking about lies, but everyone's just going to use it and it's going to be fine.

And like there's just something about the norms of photography in there that is interesting to pull out of Instagram. But I think it's like broadly true of all the. platforms. And it's not going to be a huge leap until you push the shutter button and the camera start doing this stuff preemptively for you. It is already happening. That is already happening with Samsung in the moon. And like, hey, we see that you're trying to take a picture of the Eiffel Tower with your partner in it. We're just going to, when you hit the shutter button, we're just going to get to what you think you wanted. It's like, it's just like obviously the next thing that's going

happen. Yeah. I have deeply mixed feelings about this, but it was wild to watch the demo where they scooted the kid over on the bench. The utility of that is a thousand percent obvious, right? Oh yeah. I wish my kid was in the center of this photo. The magic eraser where they're constantly circling people and deleting them from photos. Like, bye, bye. A generation from now, some Gen Z boss is going to like fire someone by deleting them from a Zoom call. Like that is absolutely going to Just circling them on the call. Like, Tony, it's been good. Just circled out.

You're out. See you later. This is Steve Jobs, let Tony Fidel know that he was going to get fired by deleting him from the contacts in the first iPhone demo. This is a real thing that happened. But, like, that's all just wild. And it's like, we're very close to accepting magic eraser as being a normal thing that happens in photos. Take this person on the background. Take this distraction on the background.

It's like this close to being normal. The phone's going to do it for you before you decide to do it. Very close. to be normal. What is a photo? Singularity event for what is a photo. Okay, we have to stop because otherwise it will be the entire show. We're going to take a quick break. 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 Murrow to move faster. With real-time collaboration and a robust CMS, with everything you need for great SEO, 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, scale-ups, 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 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 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.

With hiring pro, you spend less time searching and more time connecting with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward. Join the 2.7 million small businesses using LinkedIn to hire. Get started by posting your job for free

at LinkedIn.com slash track. terms and conditions apply. All right, we're back. Okay, so I want to talk about the most, I cannot believe we forgot to talk about this. So Google organized I.O. into sections like you do. It's two-hour keynote.

So first, like, here's all the stuff we're doing. Here's all the AI we're adding to our products. And we just talked about that. Then they're like, and then there's Google Cloud. And here's all the ways you as a business can add AI to your products using our tools. And they talked about their partners. and like some of them are terrifying. Like, what do you think Deloitte's going to do with AI?

I don't want it. Is it going to fire a lot of people, companies around the country and world? Yeah, it certainly is. Yep. With the power of Google Cloud. Like, what's Accenture going to do with AI? Yeah, dude, they're going to downsize your shit. Like, it's just like very obvious what's going to happen there.

But then there is this moment. And I cannot stop thinking about it because it came up twice. First, the company said it. And then later they referred back to it. Google was like, and here's it in action. I just want you to listen to this. And then we can all just figure it out together. Folks come to a wandies, and a lot of times they use some of our acronyms.

The junior bacon cheeseburger, they'll come in and give me a JBC. You know, we need to understand what that really means. And voice AI can help make sure that order is accurate every single time. Okay. Now, I can be charitable here. We should we listen to it again, shouldn't we? So, the junior bacon cheeseburger, they'll come in and give me a JBC, you know, we need to understand what that really means, and voice AI can help make sure that order is accurate every single time.

I'm just imagining just hordes of teens running into Wendy's screaming, give me a JBC. So charitable what this means is they're going to build some sort of voice recognition order-taking system. Again, downsizing their workforce. And apparently the teens have started calling junior baking cheeseburgers, JBCs, and they need the voice AI to recognize this. Right. That's more or less what he's trying to say, if I understand it. He said that without any of this context. He just said it and Google ran it as though we would all understand that the problem of the robot order-taking system, not understanding what a JBC is, is so challenging that they need to use Google Cloud. I'm just going to say paying a human being like $15 an hour to understand what a JBC is is a better solution than building a voice AI.

I feel like they, maybe it's just because I'm about to move. And so I've been having to make a lot of calls to, like, utility companies in the last couple of days. And today I called, I've called a gas company to, like, get my gas turned on in my new place. And they wanted to know what my phone number was. So I said my phone number, and my phone number starts with an eight. And they went, okay, your phone number starts with a one. And I was like, no, eight and one sound nothing alike. Like, we need to understand what a JBC is.

Like, yeah, like, y'all can't even understand what a one and an eight are. And you're going to tell me that you can understand what a one. what JBC is. It's good. It was just, it was like a pin drop silence in the keynote today. In particular, the idea that Wendy's was going to invest massively in AI to take cheeseburger orders is,

it's very good. It's like, it's a perfect sign of where we are that Google was like, what's the best compelling customer case for Google Cloud? They couldn't get like Waffle House. You know, like I don't know what smothered means. Tell me what my smothered hash browns being.

Also, I just want to point out that understanding that JBC means junior baking cheeseburger, you don't need AI to. You can just tell the computer. That's like a super easy computer science problem. That was the thing that kept jumping out. Like, nothing that you just described requires Google Cloud. Like, you're good. But no, I agree.

I mean, I think, so I was reading Wendy's is obviously very proud of the fact that it's doing this. Like, this reminds me of all the stuff that Domino's always does, right? Where like, anytime somebody launches a new technology, Domino's is like, we're going to use it to deliver your pizza. There's like that thing where you can order a pizza from CarPlay, and they're really excited about that. And it's like Domino's, that's nothing. Like, you've done nothing. But anyway, so Wendy's is very proudly announcing this thing. And their CEO, Todd Penninger, said in an interview with the Wall Street Journal, it will be very conversational.

You won't know you're talking to any but an employee. One, don't buy that. Two, is that what we're going for? Like, this is back to the, like, everybody's just going to use A-I to Fire a bunch of people. Like, I feel like I'm pretty happy talking to like a stoned 17 year old about my JBC behind the counter at Wendy's. Like, I don't need a robot to solve this problem for me. Also, which side of the counter? Are you, is it drive through or is it just walking in?

Because walking in, don't they all have the screens now? What if you are stoned? I just like, let's just walk through the situation which you're yelling. Give me a JBC. This is very realistic. Let's go. All right, LaserBong. Walk me through the situations in which you are screaming.

Give me a JBC to a robot at Wendy's. I mean, yeah, that happens a lot of the times. And usually, like, I want that and some chicken nuggets, the spicy kind. That shouldn't be hard. Like, that's a very easy conversation to have because usually if you're stoned, you're saying it very slowly anyway. So you're enunciating. A voice ad can help us understand when people watch ABCs.

All right, let's talk about Android. There was a lot of Android news at I.O. as well. Sort of. There's a really fun demo of, like, new lock screens. They, like, talked about Android, but they didn't talk about Android, right? Okay, yeah, like, it was Android 14, but like there was a pixel fold, there was a pixel 7A, there were new lock screens. They did a wallpaper demo that was like 96% of the Android demo. It was just like Dave Burke, who runs all the Android stuff and is lovely and did a terrific job in his demo, just goofed around with his wallpapers.

He's like, what up? We have a new Android. Let me change my wallpaper. I really liked the wallpapers. Yeah, I thoroughly enjoyed that demo. But, yes. But, like, Apple, like, half the time is, like, look at my wallpaper thing. There was one part during the demo where he was showing off how they could make cinematic photos.

And he was like, I did this thing where, like, here's a photo of a kid in front of a lake. And, like, I pushed the button. And now I cut out the kid and generated a eyes filled out the background and look at this parallax. And he's like, it's beautiful. And I was like, this looks insane. Yeah. I don't want my wallpapers to move that much. Yeah.

Like, it was wild. Okay. But there was Android news as a whole, right? There are new pixel devices, the pixel tablet, the pixel fold. A new home app. At one point, they were like, RCS now has 800 million users and Android messages. And then very pointedly, we're like, we hope every phone maker adopts RCS. And there was a cheer from the crowd.

And no Dieter. I kept expecting him. Yeah. And all the spotlight's shown on Deeter. And he was like, yeah. But now it's like the yearly ritual of Google asking. asking Apple politely to support her CS continued. And it's just never going to happen.

If you recall, at the Code Conference last year, someone asked Tim Cook, a question, like my grandmother hates the green bubbles or I can't send photos. And he was like, buy your mom an iPhone. Like, he doesn't care. The man doesn't care. They're not going to do it. There was a cheer for matter, which is like, like, Google is just like a very different audience. Like when Apple's like, and this will support matter, like, first of all, you're like,

I'm in a parallel universe. Apple said the name of an open standard. but it's not going to happen. But there's not rarely cheers. This was like, and it supports matter. And there was like a cheer from the crowd. And they're like RCS. And there's like a cheer from the crowd.

So that was cool. They added a new Find My Devices update. Alex, you're in on this, right? Yeah, it's going to use like, it's going to be better at triangulating, right? Because right now Apple does this too, where it uses all the Apple phones to try it to better triangulate so you can find your thing. And now they're going to do that with Google and Android. but the difference is Android has way more users than Apple does.

No offense to all of the Apple fans listening. I'm sorry, it's just the fact. Android has way more users. But I'm curious how, like, there's some stuff we don't quite know, I think. Like, what version of Android do you need, right? Because there's a lot of Android users out there, but most of them aren't on the newest version. So if you need the newest version, this becomes a lot less useful than the Apple one. I feel like this is like a perfect microcosm of the like potential and peril of Android.

Yes. Because a lot of the stuff, especially in that demo, was stuff you can do with Apple devices for your other Apple devices. And what Google is saying is we're actually going to make this a standard that anyone can tap into. And, you know, tile devices are going to use it and other companies' devices are going to be able to be part of this. And actually, they're working with Apple to do some of the work on the unwanted tracking notifications. if there's a tracker that it doesn't recognize it's following you around, it'll alert you and try to help, and that's really great. Matter is the same kind of thing.

Google's doing the same kind of stuff with FastPair. It's like the best version of Android basically takes all of Apple's good ideas and makes it available to everybody. And that is amazing. And in theory is like that actually assumes that Apple goes first on a lot of things, which it doesn't. But that's the idea, right? It is an open, successful version of this that plays nicely with everybody,

except that it cannot get its shit together enough to put everybody on the same version of Android, such that if you use a phone from 15 minutes ago, you're on the wrong version of Android and none of this works anymore. And so Google has tried to solve this by selling a lot of pixels, but it can't figure out how to sell a lot of pixels. And so it's like what Google is trying to do is right and powerful and good and would be massively successful. And it just can't do it. And I don't know. It's possible that those two approaches are actually like, mutually exclusive, that if you're going to be the open one, you actually can't force everybody

to, like, play along and update their software. But it just, it just breaks my heart every time. So they gave some stats here that were really interesting, though, that kind of support your point, right? That it's like a long-tail play. And it was like, there are 300 plus pairs of headphones that support Android fast pair. And it's like, I don't know that number is that big. I didn't, I don't know that there are literally 300 different kinds of headphones that support Android fast pair now. And that rules. That's awesome. That's awesome, right? And it's like, oh, that just happened over time. They announced this like several years ago.

And then, you know, all the new phones came out with last year's version of Android. And over time, they all just inherited Android fast pair. And the head phone makers did it. And that's like pretty cool. But you can't announce it. Right? You can announce it at the beginning when no one can get it. And then you can announce it now when like it just sort of happened.

And in the middle, there's like this vacuum of news. Whereas Apple's like, I don't know, special audio. Everything has it now. And that's attention for Google. I think this goes back to the to why. Android wasn't a huge feature of I.O. this year, or at least the big keynote, right? Like, they're going to do a whole Android keynote, too. Most years, Google I.O. is like, Android 14, Android, Android, Android, whatever. And this time it was like Android. Yeah, that's here, too. And they didn't

really focus on the timeliness of it. I think it's because they can't, right? They just are too diversified. They're too many phones, too many devices out there. There are, but they did the thing that they did, which is they announced the Pixel 7A, the Pixel Fold, and Pixel Tablet. And so to some extent, like they solve their own problem. Yeah, but they're just like, they're no longer focused on what's new right now. What is the thing that you're going to go be able to download on your device tomorrow to have a lot of fun? And said it's like, this is this is stuff that's going to come down the pipe eventually. You're going to be really excited about it.

Well, you get emoji home screens. What they're focused on is you're going to spend $1,800 on a phone that turns into a tablet. And here's a bunch of tablet optimizations. Do you know how many times they mentioned that Google has updated 50 plus of its apps to run better in tablet mode. Yes. Oh, and they did the thing where they like nagged themselves again, where they introduced the tablets.

Like, they were like, so tablets kind of suck. They're in a weird middle place. And sometimes you buy them and you shove them in a drawer and you don't use them anymore. You should buy our tablet. It's like, oh, that's a weird cell. So the pixel tablet is cool. We saw it. I would not say it looks remarkable except for the stand built into the case, which is

really cool. It's like a metal ring. It's very clicky. It's cool. And it's got a charging dock that comes with it. It's $4.99. The charging dock by itself, I think it's $1.49, but you get it for free if you buy the tablet. It's cool.

It's an Android tablet. There's nothing about it that's not an Android tablet. Except when you put it in the dock, it does the cool photo frame thing that Nest hubs do. And it has a new smart home Google Home app. Other than that, it's an Android tablet. This is why we talked a little bit about this on the Wednesday show with Dan, who got an early look at some of the stuff. But I think the thing about these large screen Android devices is that apps are bad, and they have always been bad. And Google has begged and pleaded many times to make them not bad.

And all the developers say, we don't care. Nobody sells good Android tablets. We're over it. And there was even a great moment in the keynote where Rick Osterlo was like, we have a better lineup of Android tablets than ever. And he showed like three tablets no one's ever heard of. And that was it. He's like, would you like this Lenovo Android tablet? And everyone's like, no, this is nothing.

But the Google optimizations are all well and good, and I'm very excited about that. There was a suspicious lack of really great big screen third-party Android apps still, because Google has such a massive hill to climb to make anyone care enough to build these big screen devices. But I think the genius thing that Google did with the tablet was essentially treat it like a TV screen, right? Like the demo that we saw, Nelai, has a big Google TV widget. It has Chromecast on it.

So you can cast to it, which is really cool. So it's the first. Right. It's the first Android tablet that you can cast two. Right, which is a big deal. And it is like very clearly this is a thing that like it'll run the Google apps that you want, but mostly this is for you to watch things on a screen. And most apps that you can watch things on have a button that you can press to make them full screen and that does the job. So they've like so aggressively constrained what this tablet can do that you almost don't have to worry about the app problem.

I think the foldable phone is a very different thing that is going to have a much higher hill to climb. But the pixel tablet, I'm like, I don't know that I'm going to care all that much because like, Netflix is going to look fine and Peacock is going to look fine and it'll all move on with my life. Yeah. How many apps are people using tablets? Yeah, they're using consumption. Like, what is a modern TV? Like, go look at the TV in your living here right now.

It's an arm processor. It's a Linux-based OS and it's a bunch of video playing apps. Like, it's a giant Android tablet. That's all it is. And like, now they made a little one and they made a dock for it. And that's cool. Like, I can definitely see myself buying it. I wish that the camera.

could be electrically disconnected. I do not like having a camera on those kinds of on smart displays in the house. It's just like, that's the line for me. It's like, I'll put, I'll put microphones anywhere. But like, I wish there was a button or a shutter or something on that camera. By the way, they showed a video where a celebrity, I think she's the narrator of the circle. That's what the view was telling me.

Michelle Boutot was taking meetings. She's like, I take three to five Zoom meetings a day. And she was just sitting in front of a tablet on the top. We're like no keyboard. This didn't happen. This is not the way to do it. No, but I'm definitely going to buy one. I'm a second for this.

I'm also thinking about buying the fold because we played with it and it is exactly the right form factor. Like they got the size and shape and hinge of this thing right. The hinge is right off the surface duo. Like this is what the surface duo should have been. But it folds perfectly flat. The hinge is really tight. It's not tall like the galaxy fold. It's just the right size and shape.

I have some concerns about the screen, though. And also that it's $1,800. Well, that's like... Those two things are related for me, right? Because I think, like, to just hold it as an object, it's very impressive. Like, it's really well made. It's really solid. It feels really good.

It's basically an iPhone 10 that opens into an iPad mini, which is like the dream. That is literally, that is what I want size-wise. But it has that crease down the middle of the screen that is very much there. It's a little better than the Galaxy Fold, but... Oh, boy, do you still notice it? But the big thing that threw me was, and we were using it outdoors in a relatively sunny place in Mountain View. But, like, the glare on that thing was awful. Like, really, really, really bad.

There was a cheapness to it. I don't know how else to describe it. And there's a screen protector that you can see. So, like, there's the edge of the screen protector and then the display itself and then, like, the little lip of the bezel. Just learning nothing from Samsung on that one. No, right? Like, I want it. And it's cool.

And, like, the tricks they gave it where you're, like, watching a video on a thing and you open it. It pops up in the video there. And then, like, you're taking a selfie with the back camera. All that was neat. We all tried to open it with one hand, which was just a terrifying experience. Like, we're outside holding $1,800 phones. I'm like, like, like, whoa, trying to open it with one hand. Alex, I wish you could have seen the face that Nile made.

So it was Nelai V song, Allison Johnson, and me. And we're all standing around this wooden table. And we all decide we're going to try to open this phone with one hand. And I go first and sort of like push my thumb in and try to pop it open and just fully almost drop the thing. Like it, I had to like bobble it to catch it. And Nelai literally goes, whoa. Just like reflexively makes a noise. It was terrifying.

It was so good. And then we all did it. All of us in turn were like, what if I made my other friend have that experience that I just had? The phone cannot be open with one hand. hand. We did see Deeter. He did insist that he could do it. He was like, you just do tut-t-da. Like, he's Deeter. That's a very Deeter move. None of us have mastered it yet. V thinks she's mastered it. I will say that. She had a thumb thing that it was, she was better at it than the rest of us for sure. It was like she was pushing it straight up into the air. Like,

it was getting higher like from the ground, from disaster. I don't. No, no, no, no, no. This thing is not made to be open with one hand. Yeah, but there's like just a little bit of, like right in the middle where the creases and the hinge, like the bezel stops and there's like a little flat plastic piece. And it's just like, ugh. Is it smushy? No, it's not squishy. It's just, I don't know, there's something about it that just didn't give me $1,800. And that's weird because every, like the, it's nicer than the fold, right, David? Like, oh, much. Like, all pixel products are that way. They always are like the same price as Samsung, the same price as Apple, maybe $100 cheaper, just not quite as nice.

There's something about it where it's just a little cheaper, just a little more affordable feeling. I will say this is the right size and thickness and aspect ratio. They got all that right. It looks gorgeous. And I do love the idea of just opening it and doing a thing, you know. Yeah. No, like we've talked on this show about how 6.1 inches is the correct size for a smartphone screen. And this is not an invitation for debate.

Like, that's just a true fact. This is the official position of the Vergecast. Don't worry about it. Keep going, Eli. Keep going. Gold. But it is, it is, I would be curious to see, you know, slight variations on it. But this size immediately felt right to me.

And like Alice and Avee were able to close it and slip it into their pockets. You're able to hold and use the thing in one hand. You can open it and hold it in two hands and type on that split keyboard fairly comfortably. Like, I agree. I think this is really right, just sort of blueprint-wise for what one of these is supposed to feel. Like, you can use it closed, and it's just a regular phone. It looks and feels like a regular phone. But that protector on the display, I'm just looking at the photos of it.

Is it that? It's weird. Like, in dance photos, it's really big. Is it that big in real life? Yeah, like I said, there's just like a little touch of jank in there. Like, you know, just spice it up a little bit. I think David had like the best take on it. At one point, David, you were like, I can't wait for the next generation of this.

And it's like, yeah, that feels right right. Change nothing about it except make everything a tiny bit better. Luckily, that's kind of the whole pixels move. So I'm optimistic. Or they'll shut it down in 12 months. Or they're going to shut all that. On that note, we also saw the Pixel 7A, which is a very nice $5.00 fun. No, there's a review of it on the website now if you care to read it.

Nothing much to say. We also saw a very impressive demo of Google's project Starline, which is their video conferencing thing with stereo cameras that look like you were looking at somebody in 3D. It was awesome. BJ loved it. He saw it a couple, like a year ago, I think, and adored it. No, so we saw the new version. Yeah, this is the new.

This like uses regular cameras, right? Yeah, so it's just a 65 inch display in cameras on three sides of it. Yeah, because the original version, you had to like have a whole setup. You're like, I have to go down to the Starline office. Yeah, this still had a whole setup. Like, we were in like a special booth and like whatever. It was very controlled. It's a tech demo.

But it still felt closer to like, oh, this might be achievable for normal people one day. Like this, here's a piece of hardware you can put in a conference room and it will make sense. And it was very cool. And it's nowhere near shipping. But, you know, Google made a cool thing and we saw it. It is really cool, though. And the idea is basically it's just designed to be, it does a lot of work on like gaze correction so you can make something that feels like you're approximating eye contact. They've done a lot of work on latency.

see they've done a lot of work on the depth so that like, Neil and I both had the experience of like the guy we were talking to, Andrew Nucker from Google, uh, would hold out an apple. And you, you can actually see the apple in like real true high fidelity because like he's large and you're on this big screen.

And then, and he was like, you can look at it and you sort of instinctively lean towards it. And then he's like, you know, I'm, reach out and take it from me. It feels just like you're taking it from me.

And like, no, it didn't. It like super, super didn't, Andrew. It felt like I was reaching. through a ghost. Like,

that's what this was. Yeah. But, but still, the demo, as far as, like, virtual conversations

I have had with people, that was leaps and bounds better than anything I've ever tried. And the display tech is really cool, right? Because it's just a flat 65-inch screen, but then it's got a lens array mounted in front of it that sends different images to your right and left eyes.

That's so cool. And it's just like, that is cool. And, like, they're like, basically, they're interrupting, like the stereo sequence of images that you see. And so they're just sending a different picture to your left and right eye and doing it with head tracking.

So it's basically just a TV with like a lens array in front of it. That's nuts. And they're trying it with like a handful of customers. I think Salesforce has one and T-Mobile has one. But it's still very much tech time. But it's cool. It was fun to play with. When does like James Cameron get one?

Yes. The next avatar movie in order to save on costs is you alone will go into a small booth and talk to a V. just like straight eye contact with a Navi and he's like you're really really messing us up dude Disney's thinking about Disney's like working on this right now can you imagine making like dead on eye contact with a Navi person and they're like I'm going to be honest with you the way we have sex is by putting our hair into each other well now I can also it works with our animals but that's different I feel strongly that we should take a break Probably. We'll be back. The JBC's for all.

Support for this show comes from Whatnot. 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. 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. That's W-H-A-T-N-O-T dot com slash sell. Whatnot.com slash sell. Support for the show comes from MongoDB. If you're tired of database limitations and architectures that break when you scale,

it's time to think outside of rows and columns. Because let's be honest, you didn't get into tech to babysit a broken database. You got into it to actually build. something. MongoDB lets you do that. It's flexible, developer first, asset 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.

We're back. We've talked about Google enough. Let's do a lightning round. Alex, you want to start? Yeah. I'm going to fight David in a little bit. But we each chose ours before the podcast started, and David got here first.

So we're going to fight later. I chose mine first. So I went ahead and went with Apple is launching Final Cut Pro and Logic Pro on the iPad with the new subscription pricing, and mainly because I didn't know people were still using Final Cut Pro, and this, like, blew my mind this week. Like, there was, like, an hour of Dan Sefer trying to explain to me in Slack that,

no, everybody does this. This is very normal. But I didn't know. Yeah, like, MKBHD, a lot of creators use it. I'm very used to Premiere. Premier, like Final Cut Pro was huge. Then 2011 happened. There was a really bad version of Final Cut Pro called Final Cut Pro 10, and everybody

stopped using it. And then apparently everybody started using it again. And now it's going to be available on the iPad. It sounds like it's pretty, like, the feature. feature parity is good. It's going to have a lot of the same features that you've already find in the Mac version, which makes sense because they're using the same processors. So this is exciting if all you have is an iPad and you just really want to edit. There's still a lot of questions we're waiting to hear.

And if you were one of those people that uses FCP now, like, just message me. I just want to know, like, what's your life like? I know there's a lot of you, but this is still like blowing my mind. And I want to know more. I want to understand you. They're like straight up open letters to Apple being like make FCP great by doing these things so we can switch from Premiere and like from the Hollywood community. So like there's it's been around. It's it has a very loyal user base, especially among solo creators because it's cheap. It's really, really fast on Macs.

I remember Jonathan Morrison used to edit 4K footage on a MacBook air with Final Cut Pro because it's so fast on M-Series chips. That's wild. Yeah, I just want to know like when that turn happened because I remember me and like I had my copy. And I was like, oh, tin sucks. And I threw it out. Yeah. So that was 2011. I think I actually physically threw it out.

And that was 10 years, like 12 years ago. And I'm like, when did things change? The relentless passage of time, Alex comes to Russell. I know. But when was the moment? Call me. Let me know. No, I'm fascinated by the subscription pricing with this.

If they bring that to the Mac, I think it's going to get real weird. And I think like the Luma fusions of the world will have a big opportunity. Because that's one of the big reasons people hate premiere. And as I understand it, that's one of the reasons Final Cut Pro became so You can just buy it and be done. Whereas with Premiere, you're spending, I think, $50 a month minimum, maybe $20 if you do certain kind of gross versions of it. Yeah.

Don't do that, especially for an iPad app. This is a big deal, right? Apple's bringing pro apps to its supposedly pro iPad, which has the same chip in it as some of its max. Especially after Adobe spent years being like, we're doing it too. And it's always just not quite good enough. Yep. And Apple just went and did it.

Yeah. I mean, and there are a lot of, there are a lot of things about editing on an iPad that are better than editing on a Mac. That like direct input being able to move things around with the pencil and your fingers. And it's like, it's actually a really great, powerful editing machine once you figure it out. And I like, I know a lot of people for image editing in particular who will like use their Mac all day until they have to do sort of really detailed fine image editing. And then they will actually like punt over to their iPad and do that work there. And I think the video interface for that is hard work. Like, Imovie is not a great piece of software on the iPad.

Imovie has been totally replaced by Capcut. We could do, if anybody wants to write the story of how Capcut, which is basically the TikTok editor broken out, has just replaced Imovie and Apple missed it. I'll take that pitch 10 days out of 10. Like, it is a wild thing that has happened in the industry and it's like kind of unremarked upon. Yeah. No, I think that's right.

And I think there's the thing that Final Cut has had going for it is that it is most of the power of Premiere without most of the overhead of Premiere, both like financial and like performance. And so if they can figure out how to actually map that to the iPad in a way that really works, I think it's going to mean a lot to a lot of people. Yeah. To have the subscription money. All right, Dave, what's your landing around thing? I just want to say, just last thing on that, it is super funny that Apple announced that the same week that Google announced a pixel tablet that basically doesn't do anything but play movies. It's just like, his two corporations have very different ideas about what tablets are supposed to be, and I think that's delightful. My lightning round thing is Disney just had its earnings, and one of the things it announced is that it's putting all of Hulu content into the Disney Plus app.

It's also raising the price of its ad subscription and not killing Hulu or running away from Hulu. Because it can't. It can't, and also doesn't seem to want to, but it definitely doesn't seem to want to buy the rest of Hulu from. Comcast for what would cost at least $9 billion. So it's just doing this weird move where it's going to sort of quietly try to like obviate Hulu and just like let it die to make Disney Plus more successful. Because Disney Plus, they're now going into ads. And one of the things that we've seen is that actually if you can build a subscription

slash ad hybrid business, that can be super successful. It's working really well for Netflix. Disney like everybody else is losing an insane amount of money on streaming. and needs to find ways to make more money. As a longtime Hulu fan who has debated canceling Disney Plus because I don't watch anything on it anymore, this makes me sad. But this was always kind of inevitable, right? Alex, I feel like this is, this was always where this was going to go.

Yeah, this was always, because right now they have to, they have to decide whether is Comcast going to buy Hulu, is Disney going to buy Hulu? And Disney's been putting a lot of money into stuff to go on Hulu, but it still owns the right. So it can just pull all of that stuff. and put it on its own service and be like, all right, Comcast, enjoy your three shows that are still on the service. Suffer. And I mean, Hulu Live, I think, is still like an important product that bears some thinking about, right? That's their live TV function.

It's basically the biggest competitor, I think, to YouTube TV. So there's still something there. But it's very clear that like Apple, or it's very clear right now that Disney is cutting costs. They're also saying they're going to drop stuff from Disney. plus some of the shows and stuff. We don't know what, but they're going to be dropping content because they don't want to pay the licensing fees for it. This kind of idea that, oh, wow, you have a utopia, you go and you pay for a streaming service and you get all of the content ever is just rapidly disappearing. So everybody, go bust out your Blu-ray players, start collecting

your media again. This is the moment. We'll put up in that. All right, here's my lightning round one. Google announced it's bringing YouTube ways and Zoom to cars with native Android, which is very funny, as David pointed out to me, everyone seems to be very interested in making your cars more functional when they're in park. So now you can watch YouTube in park. I think this is because of charging, right? They just think people with EVs are going to be sitting in their cars charging for 20, 30 minutes at times. Well, you go and you buy your JBC and then you don't want to go home and eat it. You got to eat it fast. You're like, it's a JBC and the robot's like, what do you mean? There's a data center on fires trying to figure it out. You're just getting salad after salad.

You didn't want that. This is not actually my lightning round item. My lightning round item is that David came to pick me up from the hotel to go to dinner last night. Oh, no. And it's Kia Soul. And he looked at me with absolute swagger in his eyes and said, wireless car play. And then the wireless car play immediately didn't work. Like, immediately didn't work. We didn't even make it out of the driveway before it broke. We literally didn't even get out of the driveway. I have never been so disappointed in Apple than I was at that moment. This rental Kia Soul, It's a hot whip. But he was just like, why this carplay is great?

You know, like, Google Map the restaurant. And I was like, this thing is, this arrow is not moving. Like the app kept quitting. And then Nilai got out of the car and it worked fine for the rest of the night. And then again, it has worked fine today. There's just like, there is like a Nilai Patel force field that just destroys car play. I don't understand. Carplay's like, I don't want it.

I was just 100%. He was like, why it was car play? And then just, I. I hit go on the directions and just nothing happened. No, no, no. He had to pick up his phone and use Google Maps on his phone. It sucked so much, Alex. I can't even tell you the shame that I felt in that moment.

I'm going to pick you up one day, Eli, in my car with wired car play. It'll work. All right, that's it. We got to wrap this thing up. That's the Vergecast. I'm sorry that we're completely loopy. It has been a long day, but a fun day. It was fun to see everybody.

We saw listeners at I.O. It was great to see all of you. Really cool. be in person. We've got lots more from I.O. on the site. We have hands-ons with all the gadgets, all the stuff. We have much more coverage of Google coming this year, as we've already previewed. So look on the site for that. I'm really interested in your thoughts about Google. Search changing is a big deal. So let me know what you think, because I wanted to shape our coverage a little bit. And if you want more on the hardware, the Wednesday show this week was Allison and Dan talking about

all those hands-on. So there's lots more about the hardware if you want it. On the Redcast, Wednesday show this week. All right. That's it. That's a rest of it. And that's a wrap for Vergecast this week. We'd love to hear from you. Shoot us an email at Vergecast at theverge.com. The Vergecast is a production of The Verge and the Box Media Podcast Network. The show is produced by me, Liam James, and our senior audio director, Andrew Marino.

Our editorial director is Brooke Minters. That's it. We'll see you next week.