This AI notetaker won't sell surveillance to your boss

He built the AI tool that half of Silicon Valley is now using to run in the background of every meeting. So is that making us more productive or just making it easier to have more meetings? That's this week on Platformer. This episode is brought to you by Jira Byte and Atlassian, where teams and coding agents get the context they need to do the right work. Try it free at jira.dev. That's J-I-R-A dot D-E-V. Welcome to Platformer. I'm Casey Newton. This week, our guest is Chris Pedregal, co-founder and CEO of Granola, an AI note taker that runs in the background of every conversation you have at work, if you want to. It has quickly become the default for so many businesses. I use it myself and I wanted to talk to Chris about what does it mean that we have a note taker always hovering in the background, recording all of our conversations, learning things about us. I think you're really going to enjoy that conversation. First, though, as always, we begin by checking in on the state of AI and the economy with Platformer fellow and our Gen Z AI correspondent, Ella Marcianos. Ella, welcome back to the show. Thanks. It's wonderful to be here. Well, I'm glad you're having a good time and so am I. And you know, Ella, I thought this week we might turn the tables a little bit because I saw a story about AI and jobs that I wanted to get your take on. So are you comfortable with me flipping the script in that way? Yes. Thank you for always being so flexible and adaptable to my crazy schemes. So this is a story in the Wall Street Journal, had the headline, big companies are starting to hire again, defying predictions of an AI wipeout. Did you see this one? Yeah, I did see this one. I will say to not yes and you right off the bat, there is this take that econ blogger Noah Smith had like a year ago where he was like, no matter whether AI has any effect on jobs, we will see a mill of many stories asking, what is AI doing to jobs coming out like clockwork every month. And when I read this headline, I was like, it's happening. I'm inside the machine, the take machine. Yes, I think that's a fair point. Let me outline a couple of things that are in this story and then we can talk about whether this is a Noah Smith prophecy come true. So basically what the story says is that after 18 months of being very careful about adding headcount, big employers, including Alphabet, CSX, Booz Allen Hamilton and ServiceNow are telling investors that they plan to hire again modestly to hit their growth goals. And a reason that they are giving is AI. Executives are saying that AI needs people working alongside it, that even if you have agents working in your enterprise, you're going to need people who are shepherding them, managing them, fixing their mistakes, telling them what to do. The CEO of a company called Lattice, Sarah Franklin, said just because you have coding agents doesn't mean you're not hiring engineers. So that's kind of the case that maybe there is some reconsideration happening here. In the most recent week of jobless claims, we saw that they were the lowest since 1969, according to federal data. So that is the case that something is happening. The case against it is that these companies really aren't going on the sort of hiring sprees we saw around the pandemic. Even Booz Allen, which said it is going to hire some people, is still down, I think, seven and a half percent. The story quoted MIT's Paul Osterman saying the honest answer about AI and headcount is, quote, no one has any idea, which is the exact point of epistemic certainty that we try to begin every episode with. So those are the facts of the story, Ella. What do you make of it? Yeah, so, well, first of all, I do still feel like these firms need human employees. Like every month, I'm like, these large language models are getting better. Does my job need to exist anymore? And like every month, I do see them get less bad at writing. I do see them hallucinate less. Like I actually got my first hallucination in weeks while I was talking to Claude about distillation over the weekend. Although it was like I was on Opus 5 instead of Fable now that it's paid. So, you know. But it was still like, it was almost like hallucination used to be like a constant fear for me. And like, even though I fact check everything still, like it almost felt like a nice vacation to get a hallucination. I was like, there's a reason I still have a job. But like, despite that, it just like doesn't actually feel logical to have an AI agent take over my job. And I do think we're seeing this thing where the like, in some sense, AI as normal technology thesis is vindicated. Like even as AI capabilities get way better, there are like gaps we didn't know about that humans need to fill. Yep. I think that's right. So, you know, as you noted, we're sort of on this pendulum that swings back and forth between, you know, companies are laying people off due to AI and then they're rehiring them due to AI. And it's going to be a while before we see how all of it shakes out. But I did think that, at least in the short term, this seems like a somewhat positive story that people may want to know about. So thank you for indulging my reading of the Wall Street Journal this morning. What have you been reading about this week? Yeah, I've also been reading the Wall Street Journal. Worst stands. Honestly, I feel like they're my favorite tech vertical. I love the Wall Street Journal. All right. Anyway, although I do kind of have beef with this article. So there is a recent very excellent story interviewing a few Gen Z years about how they're using AI for social interaction. And as a fellow Zoomer, my mind was boggled by apparently what's going on. Like, so I do genuinely feel like it's possible. This excellent reporter found like two of the weirdest Zoomers I've encountered in a long time. Even as someone who, like, works in tech in the Bay Area. Because there was one person, for example, who uses ChatGPT to take note cards that she then brings with her on dates to make sure that she gets through all of the most important points in the date. And what are the most important points of a date that you would bring on a note card? I really want to know. They're not in the article. Like, do you, what, I, part of me, yeah, like, what would be on your note cards here? What would you be asking ChatGPT for? Well, I mean, the reason that I think this is so funny is, like, the reason that you generally would bring a note card to anything is that, like, you're worried you would forget. But it's like, what could you forget to ask on a date that would be, like, so consequential, you know, that, that it couldn't be solved just by going on a second date with the person and being like, I meant to ask you your net worth. That's true, that's true. Well, I, I'm glad to know what you're, you're, you've been asking. Okay, so the story, the story's headline is some Gen Zers are using AI for social interaction, maybe even too much. So one person is using note cards. Like, what else are people, what else are people doing and are we worried about it? Yeah, so, like, one person, like, told stories where she was like, some friends of mine, like, a friend of mine was, like, texting her new boyfriend or whatever. And he, his texts suddenly started having, like, extremely good grammar when previously he'd, like, misspelled stuff all the time. And nobody, nobody will suspect me of this because I still misspell stuff in my texts all the time. But he apparently had been just, like, using ChatGPT to, like, draft the texts with a girl he was dating so that he would sound more put together. Another person has used, has used it to, like, come up with pickup lines to say to people on dating apps. So basically, like, it's not like people have, like, scripted in-person social interactions, but there's, like, somewhere in between this where, like, they're actually using it to generate text messages. They're actually, like, using it to prepare for social interactions somehow. Right. And, you know, this is a sort of old dream, you know. You may be familiar with the story of Cyrano de Bergerac. It came out in 1897 when I was born. And it's basically about somebody who is bad at social interaction. And so he has, you know, a friend sort of supply him with lines to woo someone. And now, you know, thanks to the miracle of transformers, we can automate this process and distribute it to more people. So the desire is new, but I can imagine there would be some fears about what is lost if, you know, you're sort of checking with, you know, ChatGPT for everything before you deign to speak to someone. Yeah, for sure. And, you know, this even has come up in more recent science fiction, namely the AI 2027 scenario that, like, made some waves last year about AI doom, where they have a midpoint in AI development where every politician has a headset where, like, whatever ChatGPT-8 is, like, telling them what they should say in order to be successful. And that means that all of the conversations between our political leaders are ChatGPT talking to itself. I don't feel And so somehow, just within the past couple of years, Granola has become a company valued at $1.5 billion. It has customers including Vanta, Gusto, Asana, Cursor, and Mistral, and it also includes me. I signed up for Granola when it was still in early testing, and it remains my meeting note-taker of choice, even though now every time I have a meeting, Google and Notion and Zoom all wake up and ask me if they can help too. But something that makes Granola unusual, I think, is that it is the rare AI productivity tool that people will actually volunteer that they love. It makes people feel like they are being more productive. The question is, can it be more than that? Right now, the company is attempting a transformation from a very smart notepad to what Chris calls a context layer that other people's AI agents plug into. To put it lightly, it has a lot of competitors in that space, and it's had some controversies too, including ongoing questions about the privacy of the way that Granola works. So there's a lot to talk about here, and I find that Chris is one of the most interesting and thoughtful product thinkers around. We got into all of this, and I think you are going to enjoy the conversation. So with that, here is my talk with Chris Pedregal. Chris Pedregal, welcome to Platformer. Hey, Casey. So when you guys got started, I would not have said automated meeting summaries sounds like the seed of a big company. What led you to the ideas that would become Granola? So I guess we didn't start off saying, oh, you know, meeting notes is the thing. Basically what happened is we stumbled across LLMs before ChatGPT came out, and we were like, oh wow, these things are different, interesting, special. And we built a few prototypes with them, and very quickly, we felt like the tools we use for work and for thinking were going to be reinvented on the backs of LLMs. And that's what we set out to do. We kind of were like, okay, we're very inspired by like Engelbart in the 50s and Alan Kay and all those guys. It feels like another seminal moment in like the progression of technology. And so we had all these really big ideas of like, okay, we want to build this like contextually aware workspace. And that's still what we're doing, but it turns out, when we actually put stuff in front of users, what people got really excited about was the idea that they wouldn't have to take meeting notes and that a computer would do a good job at that. But not like a weird third-party crappy summary, but like something that actually felt like it was my own voice and cared about the right things and that I could control. And then, yeah, that's where we started. Honestly, I think we thought transcription was going to be a big thing. It turned out to be a way bigger thing than we expected it to be, right? I think we can kind of look around and see that change that's happening. But yeah, that's how we got started. Tell me about what you think it is about those transcriptions that proved to be so exciting for people. I have my own answer because I am somebody who used Granola one time and I was like, oh yeah, this is cool. I'm just going to do this for everything. But when you talk to users, what's the common thread running through all that? Yeah, so it's interesting. It's different things at different moments of usage. So when you first use Granola, I think what people were floored by, at least like two years ago when we launched, was once you've installed the app, it's basically no effort. Like you click a button at the start of a meeting, and then you get just these really high-quality notes, right? And you get them instantly. Like the moment the meeting ends, they generate. And I think that was a bit of a wow moment for folks. And then they were like, oh, I can still take my notes, but if I don't take notes, it's still have great notes. And then I think there's a second wow moment, and not all users have had that, but it's basically once you start using an agent, transcripts are this like goldmine of context, right? And out-of-the-box agents are really intelligent, but they know nothing about you. And turns out, we do tons of meetings and speaking is pretty effortless. So there's this like context, like exhaust that you're naturally putting out. And something like Granola just captures that. And we're still figuring out the best way to give that to the agents, right? But I think that's the second moment where like, oh my God, actually I get all this stuff for free. I think those are the main reasons people love it. Yeah. I will say that I find myself using Granola for everything but maybe searching through my notes less than I expected to. Like I get a lot of peace of mind from capturing it, but then, you know, I go about my day-to-day and it turns out I don't really need to like check back for, you know, what this one person said. How true does that prove across your user base? And is it maybe that there are just sort of like certain roles within an organization where they're searching constantly? Yeah. Yeah. It's a good point. I think the way... So I can talk about my personal usage and then like different salespeople are different, for example. But my personal usage is I go back less than you'd expect. I would go back and some companies would look at that and be like, oh, okay, well, how do we drive that number up, right? I actually think forgetting and like not going back and looking at stuff is a feature as much as a bug, if that makes sense. Like our brains are actually pretty decent at remembering the most important things. What's powerful is that moment when you're like, ah, crap. Like what did I say in that meeting or what was that important thing? And then just knowing that you have it is very, very valuable. So I think that's a lot, but I think that's a little bit about how the human brain works. Like I think it would be bad if we tried to force you to go back and look at everything. The thing that has changed, like if you look at our MCP usage is like the fastest growing chart that we have, for example. And there agents have a lot more time, right? They have more cycles. So agents do go back and look at stuff. And I think that trend will only increase. I think we don't do, as in the Granola product, but anybody doesn't do a great job of surfacing the most important things from your meetings, right? Like there's kind of been this mythical idea and it turns out to be really hard in practice of like, oh, there's that one thing I forgot about in this meeting and it's important to do. Like couldn't technology just surface it at the right moment at the right time? And you know, I think we'll get closer and closer to that, but we're definitely not fulfilling that promise yet. Right. Also just like the agent has to know an insane amount about your business to be able to, you know, flag like one sentence in a 90-minute meeting that is like the most important thing for you to do unless somebody is saying like, by the way, Chris, you better get this done by Friday, which just doesn't always happen. Well, I want to ask more about agents in a bit, but first I wanted to ask you about this design choice that you made that I think is like very powerful, but that, you know, people also have feelings about, which is that Granola is invisible, right? Like there had been these other AI note takers before. Often they would involve adding a bot to a meeting. How did you come to a place where you said, you know, we're just going to like record the audio off of a laptop? Yeah. So it's interesting. The reason we made that decision was not so that Granola was hidden. Like I really don't like bots and I can tell you all the reasons why I think bots are kind of awkward and not a great solution. But like that wasn't the original goal. Like the original goal was we, we thought it was really important for Granola to always work. And it didn't matter if you were in a Slack huddle or a Zoom meeting or an in-person meeting, we wanted Granola to be like, if you hit like start note, it just starts transcribing. And the only way to do that is by using the system audio on your computer. So that, that's where we came from, right? And we also, we didn't like bots just because we thought, we thought they're a really, really inelegant, dehumanizing force in a meeting. The downside of our approach though, is that when you use Granola out of box, like other people don't know you're using it, right? So we're placing the, the onus on the user to, to use Granola to, to say, to tell to the others that they're using Granola. And that's actually not, that's not our choice. Like if there was like an easy way that didn't involve the bot to make that visible, we, we would do it just because we think that's where social norms are going anyway. And it's just good for everybody to know if you're being transcribed or not. Like we, we have like a pretty, the problem here is like the Zooms and the meets of the world actually make this almost impossible, right? Like they, yeah, because they're pushing their own like note-taking systems and solutions, right? And there's no, there's no API call where you can be like, let everyone know, like ask everyone for consent that we could just call because we And so it's actually, like, internally, everyone's using the watch app. It's like a very surprising. I did, we didn't. We thought it'd be like a niche use case. It's like, oh yeah, we should have it because it's like there's that, whatever, 3% of meetings that are really important and you want to have the watch app. But it's actually, internally, it's cannibalized iOS usage almost entirely for the folks who have watches. And then I use it also oftentimes for like hybrid meetings just because like I, it's so easy to start and and it's just it feels like the most natural form factor, you know, I'd be curious to get a few more of your thoughts on the hardware use case. As you noted, I, you know, lots of these companies are trying to figure out whether hardware makes sense here. What were some of the things that made you say like, not like, like you you've sort of mentioned one, which is it's a high bar to wear something. Uh, I personally don't really want to wear, like, you know, a pendant or a brooch or something. Um, but like what else? Yeah, so I've, as you can imagine, like I play with everything, right? So I've played with um pendant. I, I don't like the pendant for all kinds of reasons. It's too, like, as you can, you, you know me a little bit, as you can imagine, it just feels awkward and it's hidden and um uh ring's interesting. There's a ring that's come out. Um, there's one that's come out. I've ordered it. I haven't played with it yet that apparently has enough battery life where you can do meeting recordings on it. Obviously the watch. Um, there's these like pens. They're like pens that can record you now. There's the plod, like the back of the phone format. Anyway, there's all these different ones. Um, my, my view is the following. Like we, we want to build a really, we want to build the best product experience. Um, and I just think the bar of people wearing a new piece of hardware is just really hard. And if you forget to charge it and you can't use it, and then you want to use granola and then you're really frustrated and that's a bad experience. So, um, I'm not against building hardware if we figure out the right form factor. I just, I think I have a lot of respect for hardware because hardware is capital H hard. Um, but also I think the new, uh, we'll see. I don't think anyone knows what the new like AI native like form factor is, if there is one. Like, it, it's possible. It's like, we all kind of want it to be, I don't know, there's all the rumors of like, you know, like the case or whatever. I think we're all excited about the, at least I am excited about the idea of a new piece of hardware. It's not obvious to me that, that there necessarily will be one. Um, you know. And, and if you have an Apple watch, it's actually, it's great. Like the, the one thing we have realized is like it, it does need a screen. So, um, one of the issues with the plod and all of those uh copycats is that, um, you can't, you can't associate it with a meeting. So you get all these like recordings and then you have to go back and post-associate them with the meeting. Um, or you could just assume it's the right meeting and then it gets it totally wrong. So with the Apple watch, it has a screen and, and when you start it, you're actually saying, oh, I'm actually talking to Casey right now, or whatever is is scheduled on your calendar. It just works. And we found that to be a, a, a really big factor. That makes sense. I, I wonder if these things are sort of like the Fitbits of the, of the 2020s, where like, it feels good to have one, but then you use it for like a month and just then just like put it in a drawer and it turns out that everything that you wanted it to do can actually be done by your phone or your watch. Yeah. So I am, I, I'm curious, I'm curious like what you've tried and what, what you think about this, but like, I actually don't like the Apple watch as a form factor very much. Like I, I think people say, it lets them keep their phone in their pocket. I find it a bit distracting. It's like now I'm getting these notifications on my wrist. So I am, I bought one a while ago and I didn't use it until we have like the granola app on it. And that, that put me over the edge. I, and then I'll, on for the week, on the weekends, I switch back to, I've had the same watch, you know, for like 20 years or not 20, like 12 years. And I switched back to that on the weekends, but um I don't know, what about you? I'm sure you've played with a bunch. Also, like, San Francisco, you guys are like in the future, right? So, like, is everyone walking around with glasses and a necklace and like a uh, like, what's it like over there? I mean, the glasses are becoming more popular. Like, I think that's undeniable. But those are also triggering privacy concerns, right? And I think that they're much more acute because the primary use case for those is not like, I'm in a meeting at work. It's I'm walking around in the world, like looking at strangers, right? And then maybe making weird content of them that I'm posting to Instagram. So that sort of feels way more acute. Um, but Apple watches, yeah. It's like I see a lot of Apple watches in San Francisco. Um, I, I do find that it helps me look at my phone less. Like I'm very judicious about the notifications that I'll allow, but man, that thing where like you're sitting at dinner and your phone is buzzed three times and your lizard brain is going crazy. Like who, like someone in my life is having a crisis. Like to just be able to look at my watch and be like, Oh, like, I got a new like on threads. Like that's, that turns out to be way better for me. Um, so, um, let's talk about agents. We were sort of, we were starting to talk about this earlier. And this, this seems like one where there's like a lot of opportunity, but that opportunity is still kind of hazy to me. So, you know, if you're like me, you've been using granola for, I don't know, well over a year now. granola knows a lot of things about me. What are some things that you're thinking about being able to do with that and how do agents get involved? Yeah, yeah, absolutely. I mean, I think it's a little bit harder today than it than it should be. So it's like, uh, there are the people, it's funny, you probably talked to folks who are all the like, um, whatever, like AI, AI-pilled or AI doctor spectrum. Like you probably talked to people across all parts of it. But um, uh, so right now, granola has, if you, if you use granola for a bunch of meetings, that's a ton of context for you, right? About you. And um, you can connect it to Claude or you can go into the chat in granola and then you can ask questions and they'll like reference that context. Um, the state we're in right now actually is that if you've used granola for any, any reasonable amount of time, there's just too much context to like search for like quickly. So like Claude will do like a relatively dumb job at like extracting information out. It'll look at like the 20 most relevant meetings. But really what it should do is like, or what I think we should do is we should be kind of compressing that down into, you know, like a apple juice, like from concentrate. Like we should, we should, we have that concentrate, like that compressed information about Casey like right now. And then that's what should be sent to all the other tools and systems and agents and whatnot. So like one is like um, I think we're in this like awkward phase right now where we have all the context, but it's not in the form that's yet like the most useful. Um, and then what do you use it for? I think there's like a, it's a black and white thing. It's basically, you've either, you're either a person who's realized that you can use AI to help you do a lot of like maybe most things or a lot of things. And if so, um, the critical piece is the AI needs to have the right context there. And, and, and it's like, so it's like, uh, when I talk to people, they either do that or they, or they don't. It's like, you're, you're in one of those two camps. And um, and that, that's kind of, it's, it's kind of everything actually. It's like if you think internally, um, it's like, uh, like any task that our team is doing basically today. And again, even inside of granola, there's a, a spectrum of how like AI-pilled or high, how AI native people are. Um, like any task you can, if the agent has the relevant context, I think that's the, that's the big question. Like that's, um, like it can be super helpful with. Sorry, that wasn't very specific. You're probably looking for something more specific than that. Well, I mean, I, I think what, what I take away from that answer is that this is just Like that is it. So by default, you're the only person who has access to it, which companies hate us for, right? Because they're like, hey, there's all this content, like, why am I paying you money? And we're like, no, no, no. We're capturing context, but context can be very private, can be very sensitive. It's like very, very nuanced. So even though your company is paying for it, those notes are private. Like legally, if, you know, the company gets sued, we have to hand those notes over. It's the same thing with like Gmail, right? But it's not like your manager, your manager's manager can look at those notes or run an agent on those notes. And it's actually something we, it's like a, there's some real trade-offs for us as a company, as you can imagine, to do that and something we like, we believe in. So basically, I don't think the line in the future will be drawn so much on whether something should be transcribed or not. I think the questions, the interesting questions will be, should this context be shared or what part of this context should be shared? So like let's say you have a meeting, you have like one-on-one. It's like, it can be very valuable, at least for me, if I have like one-on-one. It's like, what exactly did I talk about last time? Like, what did I say about what, how can I improve? Other people probably shouldn't have access to that, but like maybe, like, maybe you have a really open culture at your company. You want to do that and you proactively share it. Like that, that's cool, but there needs to be like a human judgment and philosophy in how you want to do that. And I think the bit that's going to get really interesting, so like a classic use case for granola is sales calls. Like if you're doing a sales call, having great notes, great transcripts is really valuable. Right now, you can either choose to like share the notes and the transcript or not with other people. Something we're working on is actually, like you can modify what version of that gets shared. So we'd like to strip out all PII out of that, like all personal information. We'd like to strip out any conversations about your health or your family, you know? And those are still valuable to you. Like, let's say we were friends, you know, and I might ask you about your partner or whatever, because like, I know them or I don't know if you have kids or whatnot. And I might still want to remember, like, crap, what's Casey's kid's name or whatever. You know what I mean? Like that kind of thing. But like, I don't need my work colleagues to know that, right? So like, and that's the kind of thing. I think that's where it gets really interesting. And I think that's like the nuances of human relationships and how we work, like really matter. And like the tools we design need to like fit into all these different situations. Did that make sense? That makes sense. The case that you have me wondering about is like, I think that, you know, you could probably use granola pretty easily to figure out, like, who is a bad manager at this company, you know, or like as an employer, as an employee, being able to ask granola, like, Hey, I'm being treated in this way. Like this doesn't really seem great. So does that feel like an opportunity to you, or is that something that stresses you out? Yeah. Well, I think, I think it's an opportunity, I think it's an opportunit- opportunity if it's bottoms up and I think it's um, it stresses me out if it's uh top down, right? So I think if you're like, hey, I don't think my manager is very good. Like help me and you're, you know, you're asking granola, like, help me like navigate how I should work with my manager. I think that's great. I think it's empowering. I think, like CEOs will ask, ask us all the time saying, I want, I want everyone at my computer at my company using granola and I want access to all the transcripts and all the notes to run the queries that you're talking about. And we say no, right? We're like, that's not, that's not what the tool is built for. You can set that up if you want, but then you need to basically, you have to create a company culture and set an expectation and everyone needs to share their notes and everyone needs to understand what they're getting into. But there's no backdoor way. Like if you, I think that that's a slippery slope. I think if you have a backdoor way for someone in power to use all this context to make judgment calls in that way, that doesn't feel right to me, and that's, and it's not a future we want to build towards. Yeah. Yeah. Me either. It's yeah. It's, you know, you, you sort of said earlier that, that this is probably a case where norms are going to evolve and that feels right to me. And I think, you know, the, the hard part is figuring out what are just, you know, good ways that we could improve this business, improve the culture of this business, like coach employees on their performance and not turn the company into um a panopticon. And um I don't know how you thread that needle. Like it actually seems quite tough. Um that makes for an interesting conversation. So granola was a classic case in my view of someone having a really good idea and the entire industry racing as fast as it could to copy it and deploy it. Um, and I'm now in this zone where it seems like every time I start a meeting, seven different apps are like, you know, I could also take notes for that. Sorry about that. Yeah. No, I mean, it's a, it's a problem of success. So what do you tell your venture capitalists to convince you that you can, like, build a moat here and make something really durable around this kind of core product? Well, I'll tell you what I believe and I can tell you what I tell the investors. Like that's a little bit different. Okay. And I hope they're completely different things. Well, so, and this, I'm always afraid this will sound like bullshit to people who don't know me. But basically, like, I think we are like 5% in this new technical revolution. It's like uh like the personal computer has just dropped, right? And no one really knows what to do with it. And the only thing you can do right now is is like a word processing, which is powerful. Right? But like, spreadsheets haven't come out yet. Like the GUI hasn't come out yet. Like, I feel like we're in that moment in history and um transcription and meeting notes, that feels like the first little thing that's useful. And I think the future, like what's ahead of us is the prize. And I think it's a lot more of what we were talking about before, which is how can AI, how can these new tools we build actually change the way you work and make you do better work faster, smarter, like augment you to do more than you can do today, right? Like I think that's the prize. And I think every big company is also chasing after that, right? So that's going to be super competitive. And we just happen to like, I think through luck, uh stumble upon transcription as being like one of the most valuable data sources for that future. And that's why everyone is going after it. But um I guess like the short answer there is uh if all we were going to be is a transcription and meeting note company, then I'd be really worried about it. But to me, the battle is much bigger and it's going to be just as competitive, but it's, it's ahead of us. So I'm worried about that future world, not today's existing world. Um Yeah. I guess that that's- What is that battle? What is that battle ahead that that you see or that you want to sort of be in the arena for? Yeah. I just don't think, so like, I think in short, it's basically, there will be a few apps that people use all day every day. And um for the last whatever, 50 years, those have probably been like docs, spreadsheets, email, then Slack came on the scene. You know, like it's been a very uh consistent set of apps. And all of that's about to change, right? And um, I don't think it'll be one. I might be wrong. Like maybe we'll all just be living inside of ChatGPT or Claude. I don't think so. But I don't think it's gonna be 50 either, right? I think there'll be like some, and it's, you might have 50 apps. I just don't think you'll spend significant time in all 50 every day. And then, and as AIs get smarter and agents get smarter, there's like basically a fight for what is the UI, what is the interface that people are using day in and day out to get work done? And I think that that's what we're, that's what we're fighting for. Right. All right. Well, that's definitely a big goal. Um, you know, you sort of said earlier that this is probably a case where norms are going to evolve and that feels right to me. And I think, you know, the, the hard part is figuring out what are just, you know, good ways that we could improve this business, improve the culture of this business, like coach employees on their performance and not turn the company into a panopticon. And um I, I don't know how you thread that needle. Like it actually seems quite tough. Um that makes for an interesting conversation. So granola was a classic case in my view of someone you can manage X agents, right? And if you hire another person, you could manage two X agents. And you make more money. Do you want to hire that person? Maybe. I don't know. It's like, I don't know. I think it's, we can sit around and kind of like navel gaze a bit, but I don't think we're going to get it right. I think that things are going to change, but ultimately I think humans are at the core of everything, and I don't think that's going to change. And it sounds like, at least at granola, you didn't have a moment where like, you know, at the end of last year, everybody starts going crazy with coding agents. You didn't say like, okay, we're changing our hiring plans for the year. Like you're on the steady course. We were just so, so deeply understaffed. Like our team was just so small for the product surface area that we needed to cover. It didn't, like that, that was the major gap. And then for us, I mean, if we were like a pure backend product, maybe I'd feel a little bit differently, but so much of what we do is still like in the pixels of, and like how it feels. And so far, like product and design has been the least affected by agents. I mean, it's coming, right? But so I think we're maybe not a good, a good case study here. We might be a little bit early in that, in that journey. I would also note, I've, I've never talked to somebody who left Google and said, Google turned me into a big company guy. I want to find another company of the exact same size to go work at because that's just who I am. Also true of most of the other big companies, I find. All right. You mentioned that it sort of seems like everybody is working as hard as ever. And maybe this is a good place to wrap up because while I love productivity tools, we're doing a whole mini series about them here right now. I do think there is a version of the future where none of this manages to give anyone any time back at all. We have AI writing more briefings and emails and generating slide decks and notes for other people's AI to summarize. And we don't seem any more productive than we were before any of this started. So how do we avoid that, Chris, and can granola help? I'll like avoid your question with something related, but slightly different. I think perfect. I think, I think the reason we work a lot is, is more systemic, right? It's like, I don't know. I mean, it's cultural. It's the pressures that are placed on us by the work environments that we're in. And now, this is who we are and blah, blah, blah, all that kind of stuff. I think what my hope is that better tools can elevate what we spend our time doing while we work. And I think, I think if humans are spending, let's say we work just as much, but we do less rote tasks and more like deep thinking about like, what is the right thing to do? Like, I actually think it's weird. It's like in a world where like agents are, we don't know how good agents are going to get, but let's say they get really good. I don't think meetings are going anywhere. Like I think we're kind of lucky or unlucky and maybe as humans, but like in some ways, meetings are actually powerful because it's like where humans come together and they're like, what do we actually care about? What do we believe? Right? And that can feel really mundane and dumb and like a Monday meeting where you're talking, but it can also be like pretty profound. Right. And, and so my, my hope is that our tools can get like the, if we can all design better tools, then humans can just elevate and think about like the more interesting, the more important questions. Uh, but I don't know if that means we're going to work less. Like I, I think history is against us here. Like, like, I, you know, it's like, again, technology that's supposed to save us time just ends up with people working more. And I don't know what's up with that. Um, I, I do think maybe one, one exception here is like, I think notifications, like, like notifications and apps and tools have really, have really fucked us in our ability to like focus and think and be present and all these things. And, and we become very reactive. And I, I just don't, I don't know, the, the incentives there are so strong on everyone's part where I don't think that's going to change, but maybe in theory with AI, if it like can understand things, like maybe that's the solution. Like I really hoped Apple would like, you know, we talked about this with the Apple Watch. I really hope Apple or like Google would have like a, I don't know, like the moral imperative to be like, oh yeah, we, we, we built this the wrong way. I'm sorry. Like let's, let's try this again, but it's, it's almost impossible to do that as a, as a big company. And maybe, maybe AI or new technology can help us with that or at least give individuals agency to better deal with the onslaught of notifications. Um, but I think if you took someone from like 50 years ago and you showed them how, how our brain flips from one thing to another because these little red circles, uh, show up on screens, I think they'd be horrified, right? Like I think they'd be absolutely horrified. So I don't know. That's probably not a, not a satisfying answer. What do you think? I mean, you talk, you ask a lot of people this question. What do you, what do you think? Yeah, I think that, you know, I have a slightly skewed sample because I talk to people who like their jobs and people who like their jobs work more and they actively seek out tools to help them find more things to do. That's kind of this mini series that I'm making, right? Yeah. Um, and I've, I've said for a while that I think the best like productivity tool is just liking your job. Um, it makes everything so much better. Um, so for folks who are, let's call them normal and just like kind of like their jobs sometimes, but mostly just want to not think about it too much or get home to their families. Um, I don't know. Like, I worry that AI mostly just kind of feels oppressive to them. It's a new system to learn. It's a new system of monitoring and control. And I think this sort of shows up in the survey data about how people are feeling about AI. So my hope is that AI toolmakers, like keep that person in mind and think like, what can I do for you? A person who has a normal relationship with your job that is maybe going to make you feel less overwhelmed. Yeah. I mean, I think, I think that's a great, I think that's a great point. It's, it's interesting. I think ultimately, um, and we, we have this tension at granola, but I think, does a tool feel like it's your, like, I think tools are very important to me. Like I think, I think, uh, I could talk about this for hours, but like tools are the reason why humans can do what they've achieved. And, and I think there's an amazing quote, which is, um, uh, we shape our tools and thereafter our tools shape us, right? And, um, and I think in the age of AI, it's, it's, it's really important. Does AI feel like it is my tool and does that mean that I, I control it and it makes me better? Or is it the company's tool or the government's tool or the system's tool? In which case it, it feels completely, completely different. Um, and like the, you know, the irony is not lost on me because we are a tool that individuals use, but for the most part, companies pay for, right? And then we end up in all these, like, well, you know, we make sure the notes are yours and private. And we want granola to feel like the most personal AI, but in the workplace. Um, so yeah, I, I do think, I don't know, I've been harping on this in, uh, since, since we launched, but I really, really do think, um, trying to do thoughtful design of the tools we create really matter because people are going to use it the same way. Like people, um, stop going to libraries and started Googling everything, right? In some ways that was great because information was accessible really quickly. The downside of that is like, you know, humanity is now learning a lot of stuff from like SEO optimized, like crappy content as opposed to like well-researched, you know, you know what I mean? And it's like, that's not what Larry and Sergey wanted to happen, you know? Like I, like their motivations are, I know very, very, like, let's make information accessible and useful, but that's like a weird, uh, a weird consequence of, of the tool they built. And I think, um, I think the unintended consequences of, of the design decisions we make around AI tools, uh, might be really, really big. Um, so yeah, good, good luck to all of us, I guess. Good luck to all of us. Um, well, Chris, this was a really fascinating discussion. I would even say it was a good meeting and I'm really looking forward to uh digging into the granola transcript of it. Awesome. Well, thanks so much, Casey. I really appreciate it. Have a great day. Thanks. You too. Platformer is produced by Lindsay Chu and edited