He tried 200 to-do apps so you don't have to

He may have tested every productivity tool in the world, but in the AI era, which ones actually solve your problems and which are a waste of your time? That's this week on Platformer. This episode is brought to you by Jira by 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. Last season, we talked about what AI means for jobs and the risk that huge numbers of jobs might soon go away. This season, we're turning our attention to what you can do to keep your job. What tools and strategies can you use to keep your advantage in a world where AI capabilities continue to advance. There's nobody better to help us map that landscape than my friend David Pierce. David is an editor-at-large of The Verge, co-host of The Vergecast, and the author of Installer, the weekly newsletter where he tells the world what to download, watch, and try. He spent his entire career test-driving the tools that the rest of us use to organize our lives. And today I'm going to ask him what's real and what is just productivity theater. First, though, as always, we begin by checking in on the state of AI and the economy. And that means it's time to bring in platformer fellow and Gen Z AI correspondent, Ella Marchianos. Ella, how are you this week? I'm wonderful. I just had a beautiful 4th of July weekend. I went to a nice cookout and also to a more heterodox celebration of Canada Day. Really? Which is apparently on July 1st. And how do Canadians celebrate Canada Day? There was only one Canadian at the gathering, interestingly. But I think a lot of kind of strangely flavored Canadian sodas were consumed. What is the strangest Canadian soda that you encountered? I really cannot remember the names of any of them. And they like flew through my brain. Well, perhaps something for our listeners to investigate on their own time and discover the dynamic world of Canadian soda. In the meantime, though, Ella, I'm wondering if you have any news to bring us from the realm of AI and jobs. Yeah. So, in fact, I have some news from both the U.S. and Canada about some people who really don't want to use AI. at their jobs. So the Digital Education Council, which is a consortium of like global university educators, did a survey that surveyed 18,000, over 18,000 faculty members across 35 countries. And one thing they found is the percentage of faculty members in the U.S. and Canada who plan to use AI in teaching has gone down 9% from 2025 to 2026. Wow. So even though the technology has been getting better, the number of Canadian teachers that plan to use it as part of their work is going down seemingly significantly. Yeah, American and Canadian. Oh, wow. OK. Yeah. And in the rest of the world, this actually like isn't what we're seeing. Like they have three other regional breakdowns. There's like Latin America. There's a group that's Europe, the Middle East and Africa. there's Asia, and all of those groups are hovering around 90% plan to use AI somehow in their teaching. And that's about the same as they were last year. But in the U.S. and Canada, it's only 67% of professors, and it's down from last year when it was still in the 70s. And what do we make of that discrepancy? Did they sort of follow up with the American and canadian teachers and ask them why they felt this way yeah unfortunately we do not have like direct responses from these teachers right now um so it's hard to tell i keep on like trying to think of wild theories and like none of them actually fit regional breakdowns i can think of like data centers wouldn't be that different in like or energies wouldn't be that different and how big of a deal it is um in terms of like actual data center build out between regions i don't know They're like, I this has like in some sense stumped me, but also on a vibe level, it makes a lot of sense. Like, yeah, just like when I talk to most mostly the population I talk to about AI is AI researchers and AI policy people. And like in general, when I talk to AI researchers and AI policy people in the US and Canada, they just are way more negative about AI, like across the board and like a really strong way where it like almost seems like this like very pervasive cultural effect. Right. And we know that when we look at the surveys of Americans of how they feel about AI, its popularity has been rapidly declining. And so it sort of makes sense that teachers are just sort of part of that American populace that doesn't like what they're seeing. At the same time, I can also imagine that many of these teachers have now just had a few years of experiences with AI in the classroom, and they're not liking what they're finding. They're finding students are cheating. They're finding students are not learning the material. They're finding that their place in the classroom feels threatened in a way that it didn't before. So part of me feels like it would be surprising if the adoption of AI were going up in American and Canadian in classrooms. But what makes you think that it ought to be going up? Yeah. So I don't know, like AI knows a lot of stuff. Like I was a university student, I don't know, like last year when the models weren't even that capable. And while I would not necessarily want to be getting lectures from an AI lecturer, like, you know, as a CS student, while I was not vibe coding and I was not allowed to vibe code. And I'm actually kind of grateful for that. I do think I learned stuff creating antiquated artisanal code in C. I essentially used it as a replacement for Stack Overflow. It told me a bunch of random stuff that I was forgetting. And that was a little bit faster, I think, for memorizing all of the syntax stuff I needed compared to like if I didn't have AI. That said, like I do think like most of the places where I've seen myself or students I know actually finding AI useful is like when they seek it out in contexts where they're allowed to like customize towards what they actually think they need as opposed to like teachers intentionally doing some sort of AI integration in the classroom. Right, that makes sense. I mean, I do see a world where a teacher could think of AI as a partner to them in the classroom. And maybe it's able to do some one-on-one coaching and tutoring with students that are falling a little bit behind. Maybe it can help with grading or evaluating papers, although, of course, you know, that could be very fraught. But it's easy to imagine a world where, like, if it worked, a teacher might like it. But again, I think all those other concerns that they have are real. Any other takeaways from this study? Yeah, I don't know. It's just like there's a really big difference between these groups of people, like a roughly 20 percent difference in like how much they want to use AI in the classroom. And like people in these different regions have access to the same tools. They're teaching different curricula to a certain extent, but not completely. And so I think that this really does show us that the extent to which people are willing to adopt AI is just not some sort of standardized objective sense of how useful the tool is that everybody will agree on. It's very contingent. It's very culturally mediated. And I think we should all keep that in mind whenever we feel super confident about some assessment we have of AI in a specific context. I think it is a great point that the opinions do seem to vary widely around the world. Well, it's an interesting study and it speaks to a question that we're going to be asking a lot this season, which is what kind of tools do you want to use in the AI era and which tools do you not want to use? And so I feel like today we learned at least what American teachers seem to not want to use. Speaking of that, I think it is now time, Ella, to bring in someone who can tell us a lot more about tools. Ones both that he likes and does not like. After the break, my conversation with David Pierce. Jira by Atlassian is where agent speed meets team intelligence. Assign any work item to your favorite coding agent and see exactly what it's doing without leaving your flow. When something stalls, you're not digging through logs. You can see what's running, unblock what's stuck, and stay in control at scale. Don't let your agent start cold. The teamwork graph feeds them context from across your entire stack, delivering 44% more accurate results with 48% less token usage. So the work they pick up is the right work, done right the first time. Try it free at Jira.dev. That's J-I-R-A dot D-E-V. My guest today is David Pierce, editor-at-large at The Verge and co-host of The Vergecast and Version History. I've known David for about as long as I've been a tech reporter. He worked with David Pogue at The New York Times, was a senior editor at Wired, was the personal technology columnist at the Wall Street Journal and editorial director at Protocol before returning to The Verge. And since 2023, he has written my literal favorite newsletter, Installer, which is The Verge's weekly guide to what to download, watch, read, and explore, where among other things, he tries every productivity tool so the rest of us don't have to, although of course I almost always want to. Last October, he also launched Version History, which is a great show about defining products from tech's past, like the Hoverboard, Guitar Hero, and Vine, and what we can learn from them. So why David? And why now? Because before we spend another season asking whether AI tools are making anyone more productive, we need a map of the territory. And David has walks every inch of it. He wrote the definitive pan of the humane AI pin. He's covered AI's pivot to the enterprise. And this year, he's been testing the new wave of agentic assistance from Google and Microsoft. Most importantly, he is a productivity obsessive himself. He's a time blocker who assigned every minute of his day a job on his calendar and cycled through more notes apps than most people have apps, period, before landing for now on a combination he'll talk about in this conversation. David and I are friends, and in January, he had me on the Verge cast to talk about my own productivity system and adventures in Cloud Code. So today, I'm returning the favor, and I think you're really going to enjoy this conversation. So with that, let's bring in David Pierce. David Pierce, welcome to Platformer. Hello, sir. I'm very happy to be here. I am thrilled for you to join us for the first episode of our second season. We're talking all about productivity and what tools people can use to maybe get ahead, stay ahead at work at a time when some folks at least are worried that AI could take their jobs. And nobody has a better view of the landscape than you. Nobody is trying more tools. And with that in mind, I have to ask, I know you recently tried the Trump phone. Is the Trump phone the advantage people need in the AI era? It really is. If you've been waiting for the thing that will make you want to use your phone less and go do something more productive, the answer is the Trump phone, which I cannot recommend enough as a way to just get you off your phone and back out into the real world. It's a huge victory. This is the screen time solution if your kid is looking at their phone too much. Just give them the Trump phone. Next thing you know, they're on a swing set playing. Yeah, they're going to throw it in the ocean and go back to being bored the way we're supposed to be. I just want to say, by the way, that I find it deeply upsetting that you figured out how to get a productivity podcast before I did. This has been a thing that both of us have been running towards for most of our adult lives and you got here before I did, and I'm very upset about that. Well, you know, as you know, I've been very inspired by you, and I don't know if I would, you know, even care half as much about all this stuff if you didn't do all the great work that you do with The Verge and Installer. I'm so sorry. And you should be sorry because I've wasted a lot of time, thanks to you, but I've also really enjoyed myself, and that's the tension that we're in. So, David, I think it's fair to say most people try to minimize the number of new apps that they install. You go out of your way to find new ones. So give us your sicko credentials up front. How many to-do apps do you think you personally installed, used, and abandoned? Oh, wow. Conservatively, 200? Like, I think if you made me pick a number, it's probably somewhere in that range. And that's limited by the number of them that exist, not by the number of them that I'm willing to try. Right. If there were 400, you would have installed 400. Without question. Yeah, I think really, I mean, all the way back to like my first experiences with smartphones, you know, 20 years ago, the very first thing I downloaded in the App Store was Evernote, which is just it tells you something about who I am specifically as a person. and this has just been an ongoing journey and then especially as my professional life has gotten busier and as my personal life has gotten busier next to my professional life I think it has gone from a thing I do because I think it's interesting to a thing I do both because I think it's interesting and out of like deep necessity to remain a person and in that sense actually the last like 18 months has been totally fascinating for all the reasons you said up top right that we're in a moment where everyone has decided that everything about how we do everything is about to change, but actually none of it has changed yet. And so we're in this complete chaos phase of just trying to figure out what the future of any of this stuff is before any of it has settled. And it's just delightful. It's so fun. It is. Well, let's dive into the chaos a bit. You're infamous for a kind of perpetual quest to find and use the best Notes app. You've just mentioned installing 200 apps. After all your years of tinkering with your setup, what has actually survived and how solid does your current setup feel? Nothing and incredibly fragile is the honest answer. I'm down to the thing that I have learned pretty explicitly is what I need. I think, and we should talk about this more broadly too, but I think I spent a long time building really elaborate systems that would make sure that sort of everything was in its right place. And I knew exactly how everything was going to be organized and everything was tagged and everything was in projects. And I could tell you all of the lexicon of every single app that existed and discovered that actually the problem was I would make these really beautiful to-do lists and never look at them, just never. And then I'd be like, well, that's weird. I don't do anything on my to-do list. Maybe I'll get a new to-do list. and so the biggest change for me has been figuring out sort of what I need to be successful and the problem is I've become very clear on that and now there's basically this incredibly annoying thing happening where there are a handful of apps that all have every feature that I want minus one and then they keep adding things that are like the thing that I want so I go back to that try that app and then realize it still doesn't have the actual feature that is very important to me. So I leave for one of the other ones that has it. And I'm just, I'm in this cycle basically between five it two to list apps I basically am switching between to list and reminders at all times for a variety of reasons And then I bouncing between four or five Notes apps depending on which particular pain point I am willing to put up with it at a given time. I switched today, Casey. Like this is not a bit. I switched Notes apps today. What did you switch from and to and why? I'm just an insane person. No one should ever hear me say this out loud. So I switched from an app called NotePlan, which I really like. Obsidian has become this very popular app by basically just building an app on top of a set of Markdown files. I think that idea is very important, right? That fundamentally the thing at the base of it should be a bunch of things that I own that do not exist in some proprietary format that are not hard to get out, that are not hard to read on some other device. I have a folder of files that I can upload to Google Drive. I can put on a thumb drive. I can print out and hand to my wife. Like that's very important. NotePlan takes that idea and basically spins a bunch of really interesting like task related stuff on top of it. Obsidian is very good for writing notes. NotePlan is much more geared towards being like a sort of daily notes and tasks app and does it very well. So I used it for a long time. And then Craft, which is the white whale of my productivity applications because it is the one that is the closest to doing everything correctly. It's almost a really great calendar app. It's almost a really great notes app. And it's almost a really great tasks app. And every time they make a tiny little bit of progress in one of those directions, I throw my entire life back into it, run into the edges again, and leave. So the update literally today was that they've done a much better job of integrating tasks around the app. So now you can see all of the relevant tasks everywhere you're supposed to, which sounds like a small thing, but it's just not how it worked before. And so now, like, it integrates with Apple Reminders. So I can say to my phone, like, remind me to water the grass tomorrow morning. And now that is in my daily note in craft, which is a very important little bit of synergy that makes this stuff go a long way. So for now, I'm in craft. By the time anyone hears this, there's a strong chance I will no longer be in craft. And when you say you're in craft, you are also in reminders. And does that mean you are also in Todoist as well? So Todoist and reminders are, again, two apps with opposite strengths for me. My sort of running theory of all of this stuff is that capture is the most important thing by a mile and none of the rest of it matters. And I think AI is actually increasing that fact that you don't have to build a system. You don't have to care about where any of it is because we actually have technology that is making it easy to find and organize these things after the fact. The whole idea is just being able to get it out of your brain and into something. It's useless in your brain and it's actually incredibly useful as long as it is somewhere. So reminders I really like because it has – it's unparalleled in its Siri integration, right? You just say, remind me to water the grass at 7 a.m. tomorrow and it just does it. I don't have to like remember the weird syntax to say the name of the app before. Todoist sounds like to-do list. So that trips up Siri like every three times. It's just one of those things that like reminders is the cleanest way I have ever found to get something out of my head and into a system. And so I always end up coming back to it because it's just so fast. Todoist is a vastly better app for like managing products and to-dos. but most of the time I come back to what I actually just need is a list of all of my tasks. I don't need it. I don't need it any more complicated than that. I just need them all somewhere. And reminders, because it does a good job of sort of percolating around other apps too, is mostly the one I've gravitated to. Got it. Okay. So it sort of cartwheeling between this like native system feature of the Apple ecosystem, a really good to-do app, and then a like pretty good at everything app. That is the current landscape of how, David, you are trying to stay on top of your very messy personal and professional lives. There's more, there's slightly more to it than that, but that's the basis of it. Everything else sort of flows around those things. Yeah. When you interviewed your colleagues at The Verge about their own productivity systems, My takeaway was that the lesson you learned was that simplicity and a single source of truth beat the perfect system. Why are those two elements so essential for getting work done? I think it's the – how should I say this? I think having a single source of truth solves two problems at once. And we don't actually think enough about either of these problems. It's knowing where to put it and then it's knowing where it is. and I think that is like that that's the whole thing right like I have something in my brain and I see this with people all the time because I have I have made myself public as a productivity nerd so people love to talk to me about this stuff there's like that you can have this sort of best fit solution for every individual thing you're like well if I have this link for this thing I'm going to put it over here and if I have this thing that I need to do for this thing I'm going to put it over here but then if I have this other google doc that I need to put and what actually happens is all of the systems just break down because it's there's too much friction to put stuff anywhere. So the idea of just saying this is the place I put things is actually the single most important thing you can do. And then the reverse of that is I know where it is now. Right. And there's you don't have to build some elaborate system that will reveal it to you at the right time. You don't have to go and like jump from app to app to app to try to figure out where this thing is that you were doing the other day. I think people have this experience all the time. Right. You're like, is this in my email? Is this in my Slack? Is this in Google Docs? Is it in my to-do list app? Is it in my notes app? Like the fewer number of places that can possibly be, the more likely you are to actually go interact with that thing. And then this is the newest part of my theory. And I'm curious how you feel about this. I've become a big believer in mess in those spaces. I used to think I don't want to see any tasks that I don't need to worry about right now. I only want to have the most important things in front of me every minute. And then I started talking to authors and very productive people, people who are working on lots of creative projects. And what they discovered is the idea of actually seeing the things over and over and over and over and over again is really important. That I'm forced to wade through my to-do list every day makes me more likely to accomplish the things on my to-do list because it actually plants it back in my brain. It helps create connections between things that actually you should be confronted by all of your stuff as often as possible. And that's another reason that I think having fewer places for that stuff is just really valuable. Right, and this is the logic of like seeing your tasks in an app like Craft because you're going there in the morning, you're starting a daily note, maybe you're writing a little bit in your journal, you're looking at your calendar, but all the to-do lists are right there and you're not just sort of hunting all over. So if folks take nothing else away from this conversation, it is that simplicity is your friend, a single source of truth is your friend. And now David and I will spend the rest of the conversation talking about how we've worked to needlessly complicate our lives and spread all of our writing and content across as many apps as possible. That's correct. Can I give you one really stupid example before we move on? So there's this guy named Jeff Wong who wrote a blog post in, I think, 2022 called My Productivity App is a NeverEnding.txt File. This is like a formative blog post in the David Pierce theory of technology. And basically what this guy did is he started an incredibly long text file, and he would have the date and then kind of a record of all the things he had to do and all the stuff that he did that day. And every morning he would open the thing up, scroll all the way to the bottom, and add the new day. And I hated the system for that thing alone. I have to open it and scroll all the way to the bottom. It's just like what an annoying little tiny bit of friction every time I want to go into my system. But what I started to realize as I started to do stuff like this is actually the fact that you stop accidentally like two days ago, and then you have to scroll a little bit further down to get to today. That's the whole point, right? It's this is the thing that makes me go back to what I was thinking about yesterday and the stuff that I didn't finish doing yesterday, that like putting this in front of your face in this very simple way is actually vastly more powerful than having some elaborate set of Todoist queries that is supposed to tell me exactly what I'm supposed to do today. Because that doesn't work anyway. Right. And I do think that, like, particularly if you work in a creative field, having some sort of mechanism for just revisiting things that you have added to your system is a huge part of it. And it feels like one of the ways where this ecosystem has evolved in the past 10 years. Totally. You built yourself a thing like that, right? Yeah. Yeah. So I built a system called Blips, which was built, which was based on this blog post by a guy named Andy Matushak, Matushak, who's very foundational in the my personal productivity universe. And his basic idea was like, when you start to have a thought like, you know, there's a backlash against data centers, you want to create a note where you just sort of like aggregate every little link and every little thought that you have about that. And that at some point in time, at least for me, that turns into a column that turns into a podcast segment. But maybe I haven't thought to write it down yet. Maybe I haven't thought to do that podcast segment yet. So I have a system in capacities, which is how I, uh, it's the, the app that I use that you're using craft for. They're just kind of collects all my notes and everything. And then when I open up my daily note in the morning, do a little bit of journaling, I see five random blips and I'll think, aha, you know what? I have been meaning to write something about that. So that's good. Does that work a lot? Like, do you actually have the experience of being like, oh, I should do something with that today? Yes, absolutely. And you know, the other experience that it gives me is I'll think, oh, I know five things that happened about that that I haven't put in that note yet. And so it draws me back into that note. So it's like kind of helping me keep them a little bit fresh. I love that. Now, interestingly, most of the stuff that we've talked about so far, I would say it's not really very high tech, right? Most of the tools that you and I have talked about, you could have built them in 2010. We've seen a lot of productivity fads over the past 15 years. David, Inbox Zero, getting things done, second brains. Now I think we're in some sort of like Claude Code, Obsidian, Open Claw moment. Every week, you're asking your readers, hey, what are you using this week? What do you think AI has genuinely changed about the way that normal people work in last year? not as much as i would have thought to be honest i think there there is a certain kind of annoying like maintenance task that i think people are really quickly starting to learn to do with ai which i think is really cool like i think claude co-work was a really big moment in that that it's like i have i have a folder of stuff and i need that folder of stuff to be different in some material way, right? I need all of these PNGs to be JPEGs. I need all of these files to be in this consistent format. I need all of this stuff to just be different in some way. And it's the kind of thing that you could do manually that is actually not that hard to do manually, but it's super annoying. And it's the kind of thing you just don't do. Like that is like chef's kiss, perfect stuff to give to AI. And I'm starting to see people do that more and more. And I think that's very cool. What I have not seen is any enthusiasm for the idea of like letting AI rule your life. I think there are things that AI is very good at. I think I've heard from a lot of people that one of the things that they do that they like is say, you know, what, show me all the stuff in Slack and my email and my to-do list that I missed yesterday. Just like show me some high priority stuff. And getting kind of mixed results, like those are great when they hit, and they're sort of useless when they don't. And that's fine. It's a relatively like low stress activity. So people get something out of that. But there are so many tools out there now that are like, we are going to, you know, perfectly prioritize your time for you, we're going to base it around your energy levels. All you do is plug in the API keys for five apps that you use, and we'll take care of everything for you. And there is just absolutely no evidence in my life or anyone else's that I talk to that any of that stuff works at all. You know, I don't know how much time you may still be spending on X.com. I feel like this phenomenon only happens on X.com. And it is Silicon Valley folks who do these sort of very performative discussions of how they're using AI. I love to pick on Gary Tan, who runs Y Combinator. Gary's a very nice guy, but he's always tweeting about his G stack and how he's used AI to write a hundred thousand lines of code to sort of, you know, run his entire life. Again, to me, this stuff seems performative. What do you make of the folks that say like, no, no, no, no, no. I have the Mac mini in the closet. It is running my life. And like, this is the way of the future. It's, it's mostly productivity porn, right? Like I remember a few years ago when the notion community really took off and it was notion went from being like this sort of niche little productivity app to like genuinely huge. Like it became kind of a mainstream thing that lots of people knew. The Reddit for Notion just became completely overrun with people showing off these like gorgeous, incredibly beautifully designed, bespoke dashboards that I hate to break it to you, accomplish nothing. They just accomplish nothing. And it's so fun to make that stuff. And so I totally get it. Like, in the same way that I really like building myself elaborate new systems that don't make me any more productive, setting up a bunch of agents that feel like they're doing something feels great. And it's fun to do. It does not help you get anything else done. I do think a thing I have come to believe relatively recently is that there is a lot of like basic day to day life task stuff that is very helpful to offload to AI. And it's the same as like the sort of Claude Cowork level of things to do that. It's like I need to find a plumber, just a real thing that happens all the time. Right. Like I can I can go on to Nextdoor and I can go on to Yelp and I can like cross research all of these different platforms to find a plumber. or I can just go to Claude or Gemini or Chad Chibiti and be like, can you find me a plumber that kind of generally has pretty good reviews? And they do a good job. And so I think there are a lot of tasks like that in our lives that are simple, but time consuming. And I think there's a lot of that that is useful to do as a way to like run through your day-to-day life. But like people are telling you stories about all of the like the agents that they have managing their finances. Like don't do that. Just don't do that. And if you're offloading your work priorities to AI and trusting that it will do a good job, it is going to fail you over and over and over again. It just will. You may get yourself into trouble. Well, so let's talk a little bit more about work. One of the topics I've been fixated this year is the way that many CEOs believe that AI is making them massively more productive, but their workforces tend to be much more skeptical. goal. At The Verge, I feel like you sort of straddle kind of the line of like you're a worker, but you're also like in leadership there. What's the dynamic at The Verge right now? Are y'all AI curious, enthusiasts, skeptics? We kind of run the gamut, I would say. I think a kind of foundational truth of The Verge is that we evaluate the products, right? And I think it is like And Neil, our editor-in-chief, always likes to say that's the source of our power. It's like, you can give us all the bullshit you want, but if the thing doesn't work, I'm going to find out. And then I don't care what you say anymore because the thing doesn't work. And I think we've just run into this weird thing with AI over the last three and a half years now where the gap between what people say and the ground truth of life as a user of these products was so enormous for so long that everybody just kind of lost the benefit of the doubt. And now it is genuinely smaller. Like there are a lot of things that we have been promised for a long time that actually now kind of work. A lot of them still don't. And a lot of them don't work all the time. But some of them do. Some of them really earnestly do the things we were promised back in 2022. But I think- Is there any- Go ahead. Go ahead. Well, I was going to ask if there is anything that AI does for you well enough at work now that you would hate to live without it. Oh, that's interesting. Um no No I don think so I think at this point the number of things oh no that not true I have a good one for you. It is very good at drawing connections between a bunch of sources, which is actually a thing I have come to appreciate. I use Notebook LM from Google for this a lot in a way that I really like. And to just basically sit down and read a bunch of stuff and it's really helpful for finding people to talk to about certain things. Like I'll read a bunch of white papers and be like, is there anyone who was quoted in all of these? And like often there is. And so it's like, oh, I should go talk to that person. And so as a way to sort of take a bunch of stuff and reduce it to something much smaller, that's very helpful. And I think especially in the stuff that we covered, the part of the job, frankly, is to be kind of a mile wide and an inch deep. And so it's like to be able to be that wide very quickly and then be more selective about like, okay, here's where I'm going to go really spend my time and try to find the smartest people the fastest and try to go do the best work. That doesn't feel irreplaceable. It just would take longer. And there would be a certain kind of thing that I don't go read as a result. You know what I mean? That makes sense. I think for me, a big one is like transcription, right? Which like, I think maybe most people don't even think of as AI really because maybe it doesn't feel quite the same as like notebook LM, which feels way fancier. but my god the hours that i used to spend transcribing interviews that i just like don't have to do anymore um you're completely you know that's also like that actually is the only one i can't live without like yeah i because it used i mean i used to spend an hour interviewing somebody and then four hours transcribing it and now shout out to turbo scribe ai i just upload it and it spits out a not quite perfect but like functionally useful transcript in 10 minutes and I don't, I, God help me if I ever have to go manually transcribe anything ever again. That's, that's wild. I don't have to listen to my voice anymore. It's so great. And that's the dream to not listen to David's voice. No, we love David's voice. We love David's voice. Um, I will also say, and this, here's where I feel like this gets a little edgier and like some journalists, like I think would say like, I would absolutely not do this, but I've been using it for podcast prep, you know, like when I have a pretty good, I did it for this episode. I had a pretty good idea of what I want to talk to you about. I wrote a long prompt and I walked away and I came back and I had a shell. Some of the questions were stupid and I deleted them. And a lot of them were like directionally right. It's like, I do want to ask about this. I do want to ask about that. And I feel like it saved me time in a meaningful way. And then you know what else I did? I, in the same conversation, I said, summarize these topics so I can send an email to David to kind of let him know what I want to talk about so he can prepare for this podcast. and I'm telling you, this stuff took like three minutes and that would have used to have been like an hour and a half or two hour process for me. So I'm always trying to find like, where am I actually getting more productive as opposed to feeling more productive? And that is a space where I just actually feel more productive. Well, okay, talk this one out with me because I've been thinking a lot about this recently. So I also host this podcast called Version History. You've been on it. And it's great. Go check out Version History. And the show is... it's part sort of litigation of interesting products over time. And it is partly just like straightforwardly a history show, right? So it is like, it has to be deeply researched. It has to be full of like, interesting reporting and lots of stuff. And I spent an enormous amount of time researching those shows. And like, on the one hand, it's very fun. And I sort of set it up as a show because I like doing that work. But I've also found myself wondering like, okay, is there a beginning of this process that I could offload? Yes, absolutely. But every time I've done that, I think the show is worse because I think the less I know the material, and I think I know the material less when I'm not doing as much work. I really do believe the idea that there is a linear relationship between how much of the work that I'm doing and how well I know the thing at the end. Reasonable people could disagree on that, but I firmly believe that. there's a great like andre carpathy quote that's just like you can't outsource your understanding you know like there's a lot that you can hand over to ai but like if the more that you hand over the less that you are going to understand and that does really matter when you're hosting a podcast i think for me like if yeah go ahead so i think one thing i've been thinking a lot about is like i will i will go and i will read you know for hours and hours and hours and i will talk to people and i will end up this giant giant massive notes and one easy thing to do with ai is to basically say, hey, I have all of this, put this in chronological order. Another thing AI is fabulously good at, right? Like, here's a bunch of dates. Here's a bunch of things that happened. Can you put them in chronological order? And I'll go through and check at the end. But like, I wrote all of this. Can you put it in the correct order? I don't think I have an intellectual problem with that being an AI thing, but like, and I don't think I also have an intellectual problem with being like, hey, can you recommend, you know, the 10 most cited things about XYZ? Like if I were to start researching Nest, where would you think I should start? I don't think that is outsourcing your understanding. But it falls down such a slippery slope into I've actually done none of the work and now I have an outline for a show that I've never read. And I'm just I'm so terrified of getting to that point that I don't quite know where to cut it off. yeah and i think um it probably is worth a little bit of exploration i do think that yeah there is a version where you just sort of snap your fingers and it's podcast prep and there are podcasters that just read things that their producers hand to them and they basically are seeing it for the first time when they sit down in front of the microphone i think those podcasts are mostly not great um i don't want to make that kind of podcast myself i know that you don't either um and yeah i I can also imagine a world where like, you know, version history follows a certain format. You could take the first 20, 40, however many episodes you've done, sort of feed that into an LLM, say like, this is kind of the format that our show follows. Our next episode is about Nest. Like, go out and do some research and like create a prep document that is kind of in this format. And then, you know, maybe that does save you a couple hours. I think the tension is the LLM is now going to choose things for you that it thinks are the most important moments. And you're sort of outsourcing, you know, to the LLM, it's judgment of what is important. And you might have made different choices had you not done that. And so now you maybe are doing a different show than you would have done had you not made that choice. So I think that's kind of the tradeoff. And and I don't I don't know what the right answer is there. Yeah, I don't either. You probably want to maintain your own stamp on it. Yes, I think that's right. And I think I'm also more and more suspicious of whatever LLMs think is important. I've been going through this thing recently where you mentioned inbox zero at the top. I was for a long time an inbox zero person. I was very on top of my email. And holy God, has that fallen apart on me. I have my email is in the worst state it has ever been in my entire adult life. And I went in and I was like, OK, let me actually this. This will be a useful experiment and also actually helpful in my life. Let me see how far I can get having email clean up my life or having AI clean up my email rather. And it was the most fascinating experience because it turns out there's a bunch of stuff in your inbox that is obviously not useful. Right. Like, yeah. Land's End sends me an email every eight minutes and I don't need those. And Claude very quickly was like, David, you don't need those. And then so it was able to actually go through and cut out a lot of the stuff in my email very quickly. And that's the kind of like co-worky busy work that I think is incredibly useful. I have a lot of marketing emails in my inbox. I would like them gone. It's a task that would have taken somewhere between a few minutes and a few hours and it took me 45 seconds. Like, fabulous. Huge, huge, huge win. The minute it turned to my primary inbox, it started getting almost everything wrong. Because what it does is it looks at something, and it's like, well, this person seems to think that what they're emailing you about is very important. So it's probably very important. And it's like, it just made me think of SEO, right? That it's like, you can game your way to the top of the search results by making everything seem like a five alarm fire all the time. And people are just going to learn to write emails that sound super personal, that use my name a lot, that make it that, you know, start with re-colon something so that it seems like we've been talking before. Like it's a completely gameable system. And the LLM has no ability to look at this and be like, here's what's important to me, David, right now or even ever. And so I just extrapolate that out to like, yes, can you do a reasonable job of telling me the facts in order about the history of Nest? Absolutely. Probably very well and increasingly well over time. Can you actually point out the things that matter in this story? No, I just don't see it. Yeah, yeah, I think that's fair. Looking at the AI landscape we have right now, there are chatbots, there are co-pilots inside of existing apps, there are agents that will sometimes promise to do the whole task for you. Which of those three is delivering the most value for you right now? So this is not the answer I would have expected to give you, but it's the AI apps themselves. For whatever reason, my own behavior is less open the app in which I want to do things and then sort of invoke the LLM into it. I actually think long term that probably is how it ought to work in the same way that like if I'm going to search inside my notes app, I should talk to the LLM about my notes inside my notes app. But for some reason, the behavior for me has just kind of hardened where I just open up the Claude app to do whenever I need to do things. I own up the Claude app. I give it the folder and we're off and running. I don't think that's the long term answer. I also just don't think anybody has figured out the UI of this. Right. Because everywhere else is just like, here's an overlaid kind of slow chat window in whatever other app you were using. And that hasn't, that's not it. I'd rather go to Claude, which is at least like a better made version of that thing. But I, yeah, for me, it's mostly been those apps. What's, what do you use? I use the apps as well. I'm not a big agent guy. I don't feel like the agents are working for me yet with maybe the exception of Claude Code, which can obviously get a lot done. But I don't really have a lot that is being done for me, like on an automated basis where like I wake up every day and like a bunch of things have been done for me. I've experimented with that stuff. Some of it's sort of fun, but like it hasn't really stuck. Um, I think I'm basically with you. Like I do just mostly use the apps themselves. Um, and I'm waiting for the day when like, I can basically just install one of them, like at the root level of my Mac book and just like, let it run my computer. You know what I mean? Like, I think that feels like the next step is when can one of these intelligences just like start to kind of replace the operating system of the hardware? Because I think that's going to happen faster than people think. And I think once it clicks into place and does work, like it might actually be kind of thrilling. I think you're probably right about that. Have you ever used this app Poppy? No, tell me about Poppy. You should. It's really fascinating. It's just a relatively small startup created this app. and you install it on your phone, and you give it just a terrifying amount of information about you, right? Like, I mean, this is the trade you have to make with AI, right? The more it knows, the more it can do, but in order to do all that stuff, it has to know everything about you. So you give it access to your calendar, and your email, and your text messages, and your location, and everything else going on. And then it starts to just sort of proactively text you things that might matter. Like one, one little thing is you get these long, uh, when you're doing two factor authentications, you get the long messages that are like, we will never send this to you. We'll never ask you for this. Here is your Amex. Uh, Poppy then looks at that text message, pulls out the six digit code and texts you the six digit code. So you immediately get a notification. That's like two, six, two, three, one, seven is the code. And it's like, thank you, Poppy that you have just done me a service, like good AI. Um, and it's, it's good at seeing your calendar and being like, hey, I know where you are. I know where you need to go. There's gonna be some traffic. You should leave now. And it is that kind of lightly proactive thing that actually when it works feels really powerful. I wound up turning it off because I had like three days in a row with no tasks and it would send me a thing being like, oh, I see that you still have do laundry today. And I'm like, buddy, I haven't had do laundry in three days. Like I did laundry. We're good. This is, I understand that you don't know about this, but it's annoying to me that you keep telling me to do my laundry. And there's just all these like little things that aren't quite there. But I had a bunch of those moments where it just figured out something that I needed to care about and just sort of put it in front of me that I was like, oh, this is something. And I think we're headed towards more of that. But who knows what the interface of that is? I don't think it's chatbots and I don't think it's text messages. It's going to be something else, but I don't know exactly what yet. I think that Siri wants, I think that is what Siri wants to be. and like the Siri equivalent for other operating systems is like there will be the like Siri panel and you'll click on it and there will just be a bunch of like ambient stuff that is that. And, you know, once they figure it out, it'll probably work well. What is an AI workflow that you suspect takes more time to maintain than it actually saves you? Everything agentic. Yeah, yeah. I think realistically any kind of go out and, you know, book something on my behalf or buy my groceries or whatever, even these sort of very repetitive tasks, there are easy ways to do this stuff that exists already. Like just one dumb example I'll give you. I wrote a while ago about the new Starbucks ChatGPT integration, which sucks, right? It's like, it makes it very hard to figure out what you want to get. I ran out of tokens trying to order coffee, which is the most insane experience I've ever had. And you go through and it's like menus exist. Do you know what I mean? Like the experience I had with Chat Shui BT was as if you're there were no menus anywhere in Starbucks and you walked up to the barista and they spoke to you a list of every available drink at Starbucks. Like we've figured out better ways to do it than that. And and then when you say, you know, I want I want an iced coffee. She goes, okay, well, let me read you the possible milk options that you can add. And then you can, and it's like, none of this works. And then I heard from a bunch of people in and around Starbucks who were like, oh, I don't know if you know this. There's one button in the Starbucks app that just says reorder. It's like, we have solved this problem. We don't, we do not need to AI our way through this problem. We have pressed the reorder button through this problem. And I think there's so much of that stuff right now that everybody is, trying desperately to like reinvent the wheels. And what we're actually doing is just coming up with needlessly elaborate, needlessly brittle ways to do everything. Like I'm not here to pretend that Airbnb is the greatest interface in the world, but it is so much better to use it yourself than to watch ChatGPT try to use it on your behalf. Absolutely. My answer to this question for what it's worth and somewhat to my surprise is deep research, which seems so cool. when I first started using it, but then I just realized, like, all I'm doing is generating 10,000 Word documents that I am not reading. And whenever I zoom in on anything, it either feels like filler or something that has a questionable source. You know what I mean? So you sort of are just better off doing the actual research than the allegedly deep research. Yes, I think one of the things that I've thought about AI since the very beginning is that it is a very useful, way to do things that don't have right answers and that are pretty low stakes, right? Like my first favorite thing to use LLMs for was movie recommendations. You just input like, I like this movie, this movie, and this movie. What movies should I watch? It's actually pretty well suited to put those things together in a useful and hard to replicate way to be like, here is a movie you will probably like based on what I know about those things. Does it need to be the absolute exact platonically perfect answer to that question for me to get value out of it? No. Does it need to, does it need to even like get the movie right every time? Like, no, if it hallucinates a movie, who cares Right Like I just don watch that movie and then we I go find something else So that the sort of thing that it like it back to the sort of plumber example right Like I don I do not actually require the single greatest plumber in the history of Alexandria, Virginia. I just need one who's going to be good because I'm going to get three quotes and I'm going to go with the one I like best and is the cheapest. Like I need you to start this process for me in a way that does not require you to be completely correct about everything all the time. And the problem you run into with deep research is we're operating under the assumption that you are giving me lots of correct information. And if I have to go redo all of your work, what the hell was the point of any of this? So you're sort of, you're like, you're just giving the thing a task you can't actually expect it to correctly complete. And that mismatch just still feels so real to me. Right. There's a version of the future where AI doesn't reduce the amount of work we have to do at all. It just raises expectations for everyone. Everyone writes more emails, generates more slide decks, produces more slop for other people's AI to summarize. The workday stays the same length. The volume of content triples. Does that feel like a realistic scenario? And are there any tools or practices you've found that sincerely give time back to you? Not really so far, to be totally honest. I think the thing that is true and the thing that I find hopeful is that in the same way that technology has always taken busy work away, this could happen again, right? Like I think the history of spreadsheets is really instructive in the history of AI, that it didn't obviate accountants. It didn't change the way that we think about business even. It just changed the tools with which you could do the job. And in fact, what it lets you do is do the work much faster. And it let you forecast more into the future because you could operate with more numbers, changing them more quickly. And you could see the outcome of those numbers instead of doing it by hand. There's some very good, Stephen Levy wrote a great thing many, many years ago about spreadsheets that I encourage everybody to go read because it's a really interesting cultural study of what happens when suddenly a thing can do the work for you. And everybody thought we would never have accountants again. And then what it turns out is Excel is really hard to use. and one of the skills of being an accountant is now proficiency in Excel. I think there's gonna be a lot of that with AI. And I think the thing that I am hopeful for and frankly already grateful for in the technology is that it can do a lot of the busy work I don't wanna do. It can go in and clear out the thousand emails from land's end for me every morning. It can relatively successfully surface the tasks that I have today. It can collate a bunch of information from a bunch of different places and put it all into a single text message that I get every morning. Like, that's the kind of thing that is useful. It's not creative. It's not even thoughtful. It's just administrative. And I think that kind of task, there are a limitless supply of administrative tasks in our lives. And so I'm not worried about people who do administrative tasks running out of them to do. But I think the hope is, I don't think, I think anyone who tells you that we're all going to stop having to do work because all of all this and we're going to like live on universal basic income and just live lives of leisure is ridiculous. But I think there is some possibility that we're going to get to sort of cut off the like bottom pyramid of what we do at work and spend more time doing more interesting work, which is why everybody should stop firing their workers. Actually, you should hire more smart people to go do smart things because that's going to be the thing that like raises the ceiling of all of this because sitting around turning a bunch of paragraphs into a PowerPoint deck is not useful work. Like no one needs to be doing that job. We can all find better things to do than turning some Excel files into paragraphs in a PowerPoint. But there are people who do that for a living because we have to. And so I think if we can automate that out of existence, I am actually extremely hopeful for what we can all go find to do instead. You wrote the review that doubled as an obituary for the humane AI pin, and you were similarly unmoved by the Rabbit R1. That's right. Two years later, we're seeing a similar promise minus the hardware agents that do things for you while you do something else. How close do you think we are to living in the future promised by devices like the Rabbit? And what do you think the role of AI hardware might be in the next couple of years? Input. It's the whole thing. I think the idea of having a device that is easier to get some kind of information into than my phone is really powerful. I'm kind of out on an island on this one. Like there are a lot of people I work with and in the world who are like, you're an idiot. Your phone is fine. I actually don't think that. I think the friction in I have to get out my phone from wherever it is, unlock it, open an app, tap on a thing and type is actually quite a lot of friction. And if we can get it down to I talk into my ring, like, you know, Eric Mijakovsky a little bit. He has this thing called the Pebble Index that is literally just a button on a ring. You press it, you talk to it, you put it away. That is vastly less friction. And if they can build that into a thing that's like, oh, you just set a task. Let me put that into your task manager. Or, oh, you just set a note. Let me put it into your notes app. That is like a powerful new productivity tool. And you know, the other thing that has going for it is when you get out your phone, you also have to navigate through a thicket of notifications and distractions. A hundred percent. Yeah. Right. How many times have you opened up your phone to do something and you end up on TikTok? And you're like, wasn't I going to write something down? Right. Or even worse. Yeah. You open up your phone and then 10 seconds later, you have no idea why you open up your phone. It's like, there's never a moment where I feel worse about myself than that moment. Completely. And so I think where I'm at is trying to figure out how all of that is going to actually practically work. Right. Because what we have right now is this run of devices that are like, we'll record your meetings and get out some of your useful information and action items. I think that'll eventually work fine. I think that has a bunch of weird societal problems. Like if you and I sit down to have coffee and I'm like, I'm just going to turn on my AI pin so that we can hang out more efficiently. Like that feels bad and it will change the way that we are together. But that kind of thing, I think used deliberately is going to be really powerful. I'm not super bullish on the idea that I need another thing with a screen in my pocket. But some sort of discrete, useful input tool, I think is going to be the trick for AI. Hmm. And then it's like, does it have a camera? Like, where do you fall on smart glasses? Which are objectively useful input system. They can see a lot of things. They can hear a lot of things. They make people feel bad. And I don't know that they're ever gonna get past that. Yeah, I mean, this is one where like, I think the brand is so tainted by the meta association, you know, like there was just that whole recent controversy over the fact that they had installed a facial recognition software, like the precursors to it in the glasses and then said, well, you know, we're not going to use it and it's irresponsible to suggest we would ever use it. I was like, yeah, that's going to be hard for me to trust. This is one where like, I can imagine Apple doing a version of it and I'll be like, okay, Like now I'm ready to kind of give it a look like, you know, maybe Google, I would take a flyer on it. But I do think that all of the privacy and surveillance concerns that people have are super legitimate. I don't love the stories I read about how much law enforcement has embraced the meta glasses. Like in general, I'm always trying to figure out how can we live in less of a panopticon rather than like more of a panopticon. Yeah. And I think 10 years ago, I would have told you that the Panopticon is going to win because everybody will always trade convenience for anything. I don't think I believe that anymore. I think that actually it has gotten to a point now where I think we are starting to see people in real numbers push back and make deliberate choices against that stuff and take the hard way to not pick convenience. and to say that actually I am going to make intentional choices about my life, which you, dear big tech, have spent two decades systematically taking away from me. So I think that is a trend I actually didn't see coming and am really heartened by. Yeah. Yeah. Yeah, me too. You spend your other podcast version history taking a look back at popular tech from years past. Your thesis is that there's an awful lot to learn from old tech. if you did an episode about chat gbt in 2036 how consequential does it turn out to be is it the smartphone or is it the hoverboard interesting it's much closer to the smartphone than the hoverboard yeah um it's to me the question is is it the smartphone or is it crypto um because crypto captured everybody's imagination and kind of amounted to nothing right like a bunch of people made a bunch of money, the world didn't change, and we all went about our business. There's a possibility that that's what ChatGPT did, right? Like, did ChatGPT kind of all on its own completely turn the tech industry, the U.S. economy, and like, in very real ways, the world on its head all at once? Absolutely, yes, right? Like, there is a world before and after ChatGPT. uh will there be another world after that probably but i think in that sense its influence has been just gigantic uh right or wrong it did that and so i think there's no question that it is it is going to be massively influential i think the question of is it going to completely remake society in the way that the smartphone did like i honestly think we're going to spend what did you say 2036 yeah i think by 2036 we are still going to be unraveling the ways in which the smartphone changed us as people um i think we're only at the very beginning of understanding that and like you start to i mean there are then these studies that are connecting like falling birth rates to the rise of the smartphone i mean it's nuts man like the world everything about what it meant to be a person in the world changed because of smartphones and and that is good and bad in thousands of ways on both sides. I think it is too early to say if AI is going to be that consequential. I think there are a lot of people who stand to make a lot of money on AI being that consequential. And so they have a lot of incentive to tell you that it's going to be. I actually don't see a ton of evidence in the world yet that it is going to change who we are as people on that level. I think there's certainly a world in which we get there, but I'm skeptical of it being the next smartphone, at least for now. I don't know. You talk to these kids who are on character AI from the time that they're 12 and having all their most important early relationships with a version of Harry Potter that will have cyber sex. And I feel like, yeah, we're probably going to change a lot as a result of this. It's very possible. We are running in that direction in a lot of ways. So let me sort of bring back to the kind of central question that I want to try to answer in this series and just get your perspective on it. I don't think there is any right answer to it. But for the people who are having the anxieties about AI and their job, how would you encourage them to think about that? And how would you encourage them to think about how technology could be helpful to them as they try to navigate that? Does this feel like a case where staying open and curious and willing to try new things is going to help you stay ahead? Or is it something else? I think staying ahead is wildly overrated, first of all. I really do. Like, I think this idea that everyone is selling that you have to either get on the AI train or get left behind is kind of pernicious and just frankly wrong. But I do think you're doing yourself a disservice not to at least goof around with these tools. Yeah. If for no other reason than because it is genuinely helpful to offload some really boring tasks to them. Like, you have a bunch of photos that need to be organized in some way and you'd like to rename them all. or like the thing that happens to me all the time is I go out with a new SD card and I take a bunch of pictures and it's like the date on all of these is September 13th, 1982. And it's like, do you know what Claude is really good at is fixing those things, right? Yeah, yeah. So take all the dumb tasks and offload them. And I actually think a thing that has started to happen in my brain and I don't know if it's happening in yours is I'm starting to notice those things more in order to give them away. Like the kind of thing that I used to struggle to think about what are useful tasks for me to offload because I just don't really notice those little bits of busy work that I do all day or the like little tiny bits of repetitive actions that I take over and over that are kind of annoying. The more you see those, the more you can just punt them to something else that can do them. So like don't give it creative tasks because that's good for you and bad for AI. Don't rely on it to do things for you because it can't and it won't and you should learn how to do things for yourself. But like, can it be a helpful tool just like any other tool? Absolutely. Like think of AI the way you think about your laptop, right? Which is like a thing that I can do things with, not a replacement for my brain or my family. And I think you're going to be fine. If you take nothing else away from this, don't make AI replace your family. Listen, man, you'd be surprised. No, it's a legitimate thing to say because some people are trying. Now, it's also true that some people have terrible families. So I'm, you know, I'm sympathetic to the, you know, the folks who are looking for support elsewhere. But what I don't say a lot is if you just if you read a thing about AI and you just find and replace the word AI with the word software, everything gets a little more understandable and a lot less scary. And I think we should all maybe do it that way. Like it's just like that software. And that's OK. It's sometimes it's very good software, but you should treat it like software. I like that. Well, speaking of software, David, last question. I can't let you get out of here without giving us a recommendation. If a listener is in the mind to turn over their life to a new piece of software tomorrow that does anything in the realm of making them more productive, what might you tell them to check out? There's an app called MyMind that I really love. It's spelled like it sounds, M-Y-M-I-N-D. and they developed it as a really beautiful sort of visual reference guide. So the idea is like, I'm a designer. I don't know, I guess designers like read magazines all day. I don't know. I like wander through beautiful museums and I take pictures of them and I upload them and it does a really interesting job. This actually I think is a fascinating and great use of AI. It automatically categorizes everything that you put in. It sorts it by color, it sorts it by type, It sorts it by like all the pictures you take of cars. It understands are of cars and it will find all the cars for you. It is just this sort of endlessly reorganizable set of stuff. And again, when I come back to like, you just need to know where everything is. My mind is a really great place to put stuff. So the way I've been using it is, is just an endless compendium of things that I like. If I read an article that I like, I save it into my mind. If I watch a movie that I like, it goes into my mind. If I read a quote in a story that I like, it goes into my mind. if I take a picture of my kid that I like, it goes into my mind. And so now I can go through and be like, oh, this is, I have a, just a running set of my favorite pictures of my children. And I, did I read a great article in the last three months? All of them are in my mind. Like this is, there's a very old idea of a commonplace book that people have been using for forever, which is essentially this thing. You write down little snippets of things that are important or relevant to you. This is that, but in a way that is sort of searchable and remixable and you can find lots of stuff. And it's also just beautiful and lovely and available on every platform. It's expensive. I think it's like $15 a month, but it is maybe my favorite piece of software that I use every day. So there's my recommendation. I love that. That's a wonderful recommendation. I've also tried it and can confirm it is very fun and cool app. David, Thank you for the recommendation. And thank you for joining us today on the platform podcast. It's been a blast. Platformer is produced by Lindsay Chu and edited by Fitz Harris at Story & Sound. You can watch this whole episode on YouTube at youtube.com slash Casey Newton. My email is Casey at platformer.news. And we'll see you next week. Jira by Atlassian is where your team and your agents work from the same context. Try it free at jira.dev. That's J-I-R-A dot D-E-V.