The case for making your own apps

What if you could make a Mac app just by snapping your fingers? That is the promise behind Glaze, a new app from the makers of Raycast. Founder Thomas Paul Mann is here. 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 jira.dev. Welcome to Platformer. I'm Casey Newton. This week, we're asking whether the best productivity tool might just be the one you make for yourself. Our guest is Thomas Paul Mann, the co-founder and CEO of Raycast and the maker of Glaze, a new tool that builds custom Mac apps from a plain-English description using AI. I reviewed it in Platformer earlier this month. I've been having a blast with it. And today I get to ask Thomas all about what he is building and what are the implications of anyone being able to make their own software. Are they going to do it? What are they going to do with it? What should you do with it? First, though, as always, we begin by checking in on the state of AI and jobs. So let's bring in Platformer Fellow and Gen Z AI correspondent Ella Marquianis. Ella, how are you this week? I'm doing well. How are you? I am doing great. I know you got to enjoy a little time off. Was that restful and restorative for you? It was like 95% restful. I would say I went to Yosemite. It was beautiful. I had a great absurdly offline time, which was very nice. But I was also cooking with my mom the whole time. And we both have very alpha management styles when it comes to cooking. You know, it wasn't like my normal workflow where a lot of my conversation happens with Claude, who does what I say all the time. Right. You're finding that mom is a little bit less sycophantic than Claude, at least on the subject of cooking. Yeah. I am also now remembering my mom watches all of these. So shout out. I love you. It was very fun. We're going to go deeper with Ella about her relationship with her mom on the next episode. So that's something to look forward to. But in the meantime, we should talk about AI and jobs. And I believe that we have a story about maybe someone else that's trying to delegate work and having a challenge with it this week. Is that right, Ella? Basically, this new story, we're seeing for the first time a big union of workers strike about potentially being replaced with robots. Tell us about this one. Yeah. So Hyundai has been developing this new humanoid robot with Boston Dynamics, who many of us know for, like, the what we thought were the cute little robot dogs from 10 years ago. I never thought they were cute for what it's worth. I always thought that they were going to possibly kill us. But yes, I do definitely know the name. I see we have different intuitions here, but maybe you were right. So basically, in January at a trade show, Hyundai and Boston Dynamics unveiled this new robot, which is a version of their humanoid created for manufacturing and specifically manufacturing of their cars. And since then, Hyundai's union has demanded as part of negotiations, protections for potentially being replaced by these robots, which is kind of a first. Yeah. I mean, so this I feel like is one of those stories that we've just been waiting to see a big manufacturer brings in a robot. The robot starts to be able to do what the people can do. Tell us a little bit about this robot first. Like, what is it? What can it do? And how big of a threat might it be to the Hyundai workers? Yeah. So it's called Atlas. It's six foot three, 198 pounds. Is it single? It's unfortunately, it's maximum lift is only 110 pounds instantly. You know, I can lift more than that. So but, you know, there are a lot of auto parts within that range. And basically, demos of it have shown it, for example, autonomously, like picking up and stacking auto parts, like taking it from like one area of a floor to another. It's starting to be able to not just be teleported, which is where a human controls it, but also like do some kind of like stacking fulfillment, et cetera. Tasks automatically. It's not yet deployed. The CEO has in the past said that their plan is by 2028, they actually want them deployed in a plant in Georgia that's not unionized. But when you look at it, it is kind of like able to do kind of multi-step manufacturing tasks and it does move in a pretty human-like way. And it isn't just like a tiny lightweight thing that like dances, which is like a lot of what you see in robot world. Also, I can imagine by like 2028, like assuming that they can continue to update the firmware and the software, like these robots might be much more capable. So what does the union want and what has Hyundai said about it? Yeah. So, I mean, part of what the union wants is pretty standard. Like the union wants bonuses. The union wants a raise. But the more robot specific things they're saying is they want to shift from hourly pay to a fixed salary because they're worried that even if the workers continue being employed, these robots will mean that their hours drop. And then they're also demanding that this company set aside about 30% of its net income for bonuses and initiatives that guarantee job security, which is a lot. And that's partially because Hyundai, like a bunch of South Korean firms, is making a lot more money because of AI. And its stock price is rising because of AI. And, you know, like workers want some share in that increase in profit. Yeah. Seems reasonable. So the union has made its demands. And what has Hyundai said? Yeah. So basically the union had its first three days of strike last week and they since then haven't been able to reach an agreement. And the second week of the strike is continuing this week. And Hyundai has, like, yet to address publicly most of the workers' demands. Hmm. Is Hyundai alone here or are there other companies that are starting to bring in these manufacturing robots in a way that we think might affect jobs? Yeah. So a bunch of different manufacturers are investing in robots and interested in robots. I probably you've seen videos of Tesla's, like, Optimus robot, like, serving coffee, which so far is mostly all it's doing. But it's they're trying to start production by the end of the year. BMW and GM are also investing significantly in robot tech. GM also earlier this month kind of outrageously added dozens of robots that they called co-bots, which is means, you know, it's a co-worker. It's not going to replace you. And it added a few dozen of those to its factory at the same time that it laid off a thousand workers from that factory. And this was this was like a robot arm rather than the humanoid, which is kind of like a new wave. But, you know, we're already seeing this happen. Basically. Yeah. Calling it like a co-bot, which makes me think of something that is working alongside me while simultaneously, like, firing a thousand of my coworkers feels pretty grim. So it sounds like maybe the Hyundai workers are on to something here. Like, this feels like a pretty realistic threat to their jobs. Based on what you've read, like, where do you think this one goes? Like, do you think they have the leverage to get Hyundai to bend or is this just going to be one of the first cases of unions just having to, you know, I don't know, maybe have less power than they want? Yeah. I mean, we like a few weeks ago we covered a different labor negotiation with a chip manufacturer where the union did win out, which like kind of increases my local prior. And that was in South Korea also. And that kind of increases my local prior that, you know, workers have some leverage there. I mean, partially their demands clearly are reasonable because there is a threat here. I don't really know what's going to happen, you know. We will just have to wait and see. And while we're waiting, maybe we can turn our conversation to somebody who is helping people build their own tools. So while you wait for the robot to be able to do your own job, maybe there's something you can make by yourself that will make your day a little bit easier or at least more fun. That is the promise of Glaze. And so when we come back after the break, my conversation with Thomas Paul Mann. 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 jira.dev. My guest today is Thomas Paul Mann, co-founder and CEO of Raycast, the Mac launcher that replaced Spotlight for hundreds of thousands of daily users, myself included. And, yes, he is a German man named Thomas Mann. And like the 20th century novelist of the same name, he has also chosen a life of writing, although in his case, it is writing software. Thomas spent more than a decade as a software engineer at a German research institution called the Fraunhofer Society. Later, he worked at Audi and then Facebook, where he met his co-founder Peter on the same team. Their founding frustration was tool sp It's extremely powerful. And I think sort of the stepping stone that really enabled us to like grow and broaden like this amazing community was having extensions. So for a long time, like we built a majority of features. And at some point, we realized we can't build all the things ourselves. So how can we kind of twist it around and give people the opportunity to integrate with the tools they already use? Because sometimes we don't even have access to those tools or we don't know them, quite frankly. And so we built this extension ecosystem. So everybody can build an extension. They can publish it in the store. And that really took off. There are all the open source. There are like over 3,000, I think, by now, all on GitHub. And that was sort of this moment where we felt like, huh, interesting. So as soon as you give people a tool to create their own tools, it sort of changes the equation. And that was all pre-AI. So people had to actually write code. And we thought a lot about our API to make it like economic and easy to use for developers. And that was sort of like one of those inspirations for Glaze, where we said, why stop at like extensions for Raycast? Why not doing full apps? And so here we are now. Right. It's interesting. You know, I was reading some of, you know, your old writing, and I have seen you say at one point that one of your chief frustrations at Facebook was that every minute outside your code editor made your job worse, which sort of leads you to build Raycast, which is, you know, you can sort of command the whole computer from this one beautiful box. Walk me through that line from there to making Glaze. Because it sort of feels like you went from like, there are too many tools, so we're just going to shrink it down to one, to now, hey, let's make a product that creates infinite tools. Yeah, for sure. So yeah, beaming my time back to Facebook. So working, being like on a lot of different tools. And I always felt as an engineer myself by trade, like the most productive I can be is like writing code, solving a problem, putting something out, right? So that's like, by number of lines of code or whatever, how much you produce, right? You're being productive on a given day. So any moment I didn't spend in my code editor felt like I don't spend my time wisely, right? And so, quite frankly, the reality is like you have a lot more tools to deal with, but they're oftentimes aren't optimized for the people like every now and then go in a tool and need to do something. So we're talking about issue trackers as an engineer. Like I don't want to have all the fluff. I want to see what I have assigned. I want to close those things. That's how much I care about it. And then there is probably like a thousand features that I never, ever think of. And so that was sort of the inspiration for Raycast. We keep you in the flow in where you do your actual work. And we give you those little extensions that you can quickly do via keyboard shortcuts and like do those things very efficiently. So you don't need to open another app. And so that becomes like very muscle memory. And so you do those kind of things. And with Glaze, like we kind of in a way go the other direction. As you said, we let you create way more apps. But the difference is like those apps are hyper-personalized. So in an ideal world, like yourself, you build yourself the perfect to-do app for you. And this happened to be like, it has a comic in there. For other people, it's something else, right? But there are U apps. They can connect to services like to your issue tracker of choice if you want to, and you can update it there. But it is a U app. So you stay focused. You stay in your productive genius zone. And that's where we think like Glaze really shines because I think what we've seen in software over the years is like new features get added in every app. Everything gets bloated. Everything gets more and more generic because companies need to grow and attract new people. And what started as a little like tool becomes suddenly this glorious massive app and wants to do everything. But then the actual people using it, they have very dedicated things they want to do. And quite frankly, sometimes those apps get more in the way than they actually help. And so we thought like, can we flip the equation around and give people the opportunity to do whatever they want and basically the platform to do this. So that's sort of the broad connection between Raycast and Glaze solving a similar problem from very different angles. Do you feel like you have a guess at like what is the end state here? Like, are we going to have fewer apps that do more or are we going to have more apps that do less? I think that's the billion dollar question right now, right? So there are all the super apps that pop up left and right with codex and Claude. Then generally chatbots, do we even have apps or is everything just like, we yap into our microphones and it does everything. I'm personally a quite visceral like thinker. Like I want to see UI in front of me. I want to click buttons. And I think we developed so much of those patterns over the years that we sit in front of our computers that it's very hard for me to imagine that all of this goes away and we're just sending text messages back and forth, talk and listen to something. So like, even when you work on something, it is a bit deeper. You might want to whiteboard something. You want to really visually think about something. Even if you want to present certain ideas, you want to show a diagram which other people grasp and not want to have a wall of text. So I think that we're in this phase figuring it out. There is no right or wrong. Like our take is use intelligence to build apps. And then you can even have intelligence in those apps, right? So they connect there. I think apps are what we've seen. Like they stay around for quite a long time. I don't see them going away. But I think everything is going to change in a way. Like I think you mentioned like being interested in like an operating system where like it's built around an LLM, right? Like how do we leverage intelligence in the first place? And I think there is something quite interesting around that where like this gets pushed down basically to the operating system level and it becomes sort of a core part similar to how we think about files and other things. Like they're just there. So it's a tough one. I think we're all trying to figure it out. Yeah. I, like in my heart, I still believe that there is room for both like a very powerful operating system that is enhanced by AI and discrete apps that do specific tasks. I think that when you abstract everything away to the level of the operating system, if you're the average person, you kind of just lose track of what it can do. There's something very comforting about having an app where you sort of know exactly what it does. You know which button to press. And I have to believe that at least like through the medium term, that idea is going to retain some power. Does that track with you? Like I'm very similar in that regard. Yes. I think people get used to certain things, how they work. This will stick around for a long time. I think what is interesting if like we open up everything a bit more. And I think that's like a weird good side effect of AI because all the big companies, I think that DoorDash recently even announced yesterday a CLI to like order food, which is kind of like, I didn't see that coming. But I think this is a really nice side effect because it allows tinkerers and everybody to work a bit more open together. Because yes, you might want to have an app that does something, but if it's all proprietary data and never be accessible from somewhere else, it makes it really hard. So I think this is a really nice side effect and allows this dual world as well, where you have like a specific app to do something with the UI, but every now and then you maybe want to have the intelligence to do like some cleanup and just goes over a bunch of like your to-dos and like just groups them for you and you don't need to think about how to do this manually. Right, right. So I want to make this concrete for people, particularly for people who haven't, you know, built their own app yet. I think your launch line for Glaze was that most software is a compromise. Give us an example or two of that. What's an app on someone's Mac right now that they might be compromising with and how could they fix that by building their own? I mean, it starts with as simple as to-do list, right? David Pierce, our friend, like he tried out 200 different to-do apps, right? And I think he mentioned every to-do app has something good and something bad. And you flip-flopping around those different to-do apps. So just take the good parts and mix them together in your own app. That's like, we've seen those things happen very often. I did this myself where I have like now a little to-do app that builds actually on top of Apple Reminders and I can use that, which is a nice one. But you can also go much further. Our support team internally, they build the entire internal tooling on top of it. So they use APIs from our services that bring all the data together and they work basically day-to-day now with the Glaze app where they do actually get all the support requests. It's connected to our manual, to our internal knowledge. There is an AI built in, This is for open source developers. They put something out and somebody else comes randomly around and like maybe sometimes even improves your software or like says, hey, there is a bug, can you please fix? And like, oh, somebody cares about my stuff. And this has been like gatekeep to like just a fraction of people on the planet, right? Like the open source community is like relatively small. But that feeling, I think, is really good. And given we all live in a digital world now, everybody can create those things. So I think like the market will be dramatically bigger than what we think of as app stores right now. Because if you think about all this creation things, even take like YouTube and video, like that is a massive market now. At the beginning, it looked kind of silly and was obviously way smaller and like the ceiling was maybe like what you can do as movies, right? But then that basically didn't really encounter that everybody can create suddenly. And then so now we talk about so many million people on YouTube, right? So I think that we're gonna go through a phase of this with software. It's not really clear where we're gonna end up with, but it's exciting. Like it's I think the most exciting time in software in a long, long time and kind of need to see where we where we're ending up in a few years. Yeah. I mean, to me, it's easily the most exciting time in software ever because I can participate in it now, you know, it's like the previous era. It's like I could I could write about it, but like that was about it, you know, now I actually get to make my my silly apps. Yeah, and I think that's the big change, right? So everybody is a creator now. Yes, for sure. Another question that I have is sort of in this vein, is like, how much of the world wants to be apps versus how much of the world wants to be agents? Like, as you were, you know, describing the world that you're envisioning in that last answer, I started to imagine like, what would it look like for an app to just, or maybe it's Glaze, maybe it's the operating system, it just anticipates an app I might need and it builds it for me. And to use like the classic tech keynote example, I thought about travel, right? I went to Spain recently. When you leave the country, you're on the internet a lot because you're trying to figure out what, what sightseeing do I want to do? Where do I get tickets to that museum? What is the train schedule? How do I get from this city to that city, right? And you know, it's kind of fun to do if you like that sort of thing, but it is also the sort of thing I would happily hand off to a piece of software. And so I can imagine an app that just sort of spun itself up and was very visual and it walked you through all of the different places you might want to go and you just click a couple of buttons and it'll start to buy tickets for you. But I can also imagine a lot of builders out there saying, Hey, I'm just going to make the travel agent agent, and you're going to plug in your credentials and we'll sort of go off and do it. I don't know that there is a right answer there. Different people may have different preferences, but what is your sense of like when you want an app versus when you want an agent? Yeah, it's a good way. I think, yeah, the travel example is a classic one, right? It pops up in every keynote of every AI announcement. So we can stick with that one because I think it's everybody has this, right? But if you think about travel, like your use case, you travel to Europe, it's like a very timely thing. You have it for like a week and then it's like gone, right? And then maybe like somewhere later you have like something similar. And this is sort of this disposable software, which I think is like kind of really nice because that is like in an app factory is really nice. You could have this, you could look at it, you could click through it. It's really prepared for you. It maybe has your taste of like having some comic pictures generated of Spain. And so it can pick up all of those little traits that it knows about you already, because you maybe build a few apps. But you'll use an agent to produce this kind of stuff where the intelligence I think is good. I think the in the agent part, really what I think everybody kind of wants that it takes off a lot of the stuff off your plate, which we're coming into the other usual example of like, oh, let an agent clean up my inbox, right? And do work for me. So for me, I feel like an agent is almost like a person assistant, a co-worker, some sort of entity that I kind of communicate. I also need to teach that entity kind of what is important to me, what's my style, what do I care about? And then it can do a lot of things for me, right? So I sort of outsource it to this intelligence layer, which can do the stuff for me. And that maybe even like creates apps for me that I can then consume and do, right? So I think like the apps is for us like mainly a way to like visually communicate and get information across because even if you have the most intelligent thing, it just still needs to somehow tell you what the heck is going on here, right? And that's for me an app. What an app exactly is, I think we will see in the next two to three to four to five years, see that change, right? But that's like software, something that you have on a display and then you maybe click a button and some agent takes again off and do something for you. But we need something to communicate back between, well, it's the classic human interface, right? Design. And I think that's where we see Glaze slotting in is the best way right now to like get like a really proper Mac app up and running, which fits a new system. You can tweak it however you want. It's very powerful. And I think like this would come over time. But yeah, like I see the agent is more like unstructured, does things for you on the side and apps are the way to communicate between like your computer and the real world in a way, yourself. Yeah, that makes sense. You have uh said that some of the first maybe most enthusiastic users of Glaze have been companies like cursor, linear, Vercel. You mentioned your support team as well. They're replacing some of their internal tools. Does this personal software movement that we've been talking about actually look more like an enterprise story? Like, do you think that maybe that is where the actual biggest market here is, is in companies that want to build all of their internal tools really quickly? Yeah, honestly, we're still trying to figure that out. We're sort of like playing this dual game and let's say there's like a massive opportunity in the consumer space where basically everybody can try something, have this app store dynamic, have people like share and collaborate. And then on the other hand, like sort of the classic enterprise angle where like people use actually software to do their work. There is a clear ROI, all of the enterprise terms. But that's like where we see like sort of like two very different polls. On enterprise itself, I think there are two things which we've seen. Being on a desktop is actually quite nice because IT is in control. Like there is no weird server running when somebody leaves the company, they like return the laptop. OK, cool. Like the, the thing is in our garden basically, right? In our fences. So that's quite nice for them. Like it's a way easier adoption curve. Um, so they can get going with that. And so, yeah, we're trying to both. We haven't really like decided which one it is. Probably it's going to be ending up a bit of both. I think internal tooling, what we've seen is just like ripe for a change. And you see like more and more companies pushing into this, because like lots of this software gets built from the outside. They don't know the internal processes. At Facebook, we have the luxury that most software is actually internal written by a team. So it's highly specialized for internally. And we see like there's a lot of benefit to this because the people that work in companies know how the company works. They know what they care about. They can all connect it together. They don't need to do compromises. So there's a lot of appetite here to just like make it yourself and not relying on a third party supplier to like build software is maybe like not perfectly solving your problems and kind of approximating things that you know better. So I think like there is there is going to be a shift there, which we already kind of see with AI being adopted so crazy fast in companies generally. Right. You built Glaze on top of cloud code and and codex, or at least those are kind of two of the ingredients in it. Uh, those models belong to Anthropic and OpenAI, who I'm guessing, would love to own this layer themselves. What are the odds they don't ship a make me an app as a feature at some point? Maybe could we argue that to some extent they have already? Yeah, I mean, um, maybe they are already. Like, that's the thing. I think everybody does everything all at once. You see also every other week something gets shut down. I think big companies have also massive audiences to serve. So I've seen this firsthand when Facebook shipped products that got like millions of people using it, but it's too small that Facebook even cares about it, right? Um, I think you see the same with big AI companies that try everything. But then to do something very good, you kind of Very exciting to hear that this is coming. Yeah. So you also run an app store and some of the questions that people have around what you might call vibe-coded apps, I think particularly for companies that are thinking about playing around in this space, is safety and security. And so I think folks might want to know like what safeguards have you put in place to make sure people aren't uploading malicious apps into your store? And have you caught any bad actors so far? Yes. Yeah, that's like certainly something that is top of mind for us. We care a lot about that because it's also something that lots of people straightaway ask, right? So if you go through the publishing flow, we made this pretty nice. So it uploads your application and then we scan the application. If there are any things that pop up during that scan, you get like a report. And then if there are sometimes even things like you maybe committed or like added an API key yourself that you don't want that everybody has access to, we flag those things. And then you can ask with the press of a button, it's just like, Hey, please resolve this. It resolves it for you and then you can re-upload. So we basically made sure that when you publish, there is a review process, something comes back to you and then you can basically then decide yourself what you want to do with it. So far, we haven't had bad actors. There were like a few weird apps that got reported. So we have also the opportunity for people in the community to report apps and we take that very seriously. So we had some that we basically took off. It's the usual thing when you build platforms, there will be some weird content that you maybe disagree with. And so we like also trying to stay on top of those. The other thing, because it runs on your computer, which we are going to add soon is like what we call the sandbox. So it cannot like access everything on your computer. will be restricted. And then it asks you for permissions to access certain things. So think about certain locations on your disk. Think about making network requests to certain domains. Um, because if you download one app, you might not want to have that it like connects at all to the network. And those are sort of the measures that you know from MacOS and iOS already or from the modern operating systems. So we, on the one hand, piggyback on those. So there was a lot of that already built in. If your app wants to read your Apple reminders, it needs to ask for like a permission and you see the usual dialogue that Apple presents you. And we tightened that even further up to make sure that basically when you get those applications, they're in a good state and the security best practices are where they need to be because that's obviously one of those things when people no longer look at the code, there might be some things that also, um, yeah, can be dangerous if they're rolled out to others. So we like put measures into this that that doesn't happen. Let me ask about the SASpocalypse, Thomas, you may remember that in February something like $285 billion came off the valuations of SaaS companies in 48 hours. And in part, I think it was the fear of someone like you coming along and saying, Hey, just make your own version of it. So when you saw all of that happen, did you think, Yeah, this makes a lot of sense. Like there's a big opportunity for me here. Did you think it was sort of overstated or how are you thinking about the possibility that a tool like Glaze could, you know, create a lot of problems for a big software as a service company? Yeah, I'm in two minds here, to be honest, like on the one hand side. The flows, how we work internally have changed so much like from just even like this year is in the last six or seven months, it's like put our company upside down and so the tooling, like lots of tools that we used heavily before are not opened anymore. And a lot of those are SaaS subscriptions, right? I don't want to necessarily name all the companies to be still friends with them, but it's like, that's the reality. Like our flows have changed so hard that you question like, do we even need this tool now? And that doesn't apply to every tool because then you have other things and this is, I think oftentimes, I think the things where there's a bit of like, you hold your records in there, like your knowledge, those kind of things, I feel like they're still staying around for long. Like you're still gonna have somewhere your emails. You're still going to have somewhere all your documents and all of this kind of stuff. I think this will be around for a long time because you don't really want to rebuild everything yourself and in every company. This was like kind of like the opposite, right? The SaaS went over it and like generalize things. And so every company runs the same stuff. And now we like maybe in this other world. Um, I think for us, what really changed is how we consume those things. Lots more stuff goes through agents. Lots more stuff maybe ends up in a Glaze app for us internally. All our dashboards. We're not opening necessarily the websites anymore. We just have dashboards in our apps, like on a Glaze layout and can look at this. But we still use the service behind the scenes. So I think like in a weird way, a lot of those maybe become a bit more headless. And I think the ones that have the opportunity to be headless might be in a good position. The others that are in a state where this is not really an escape hatch, I feel like could struggle a bit and sort of this in between state where you not really like have this opportunity to remove your UI completely or your UI is not strong enough to like be a differentiator. And then it then it becomes the question, could this be just accessed via an agent? And then you think, yeah, it's actually not too bad. And then it's actually better. And lots of those things we do now, we like access a lot of tools, um, and sometimes multiple tools in one thing. You ask like questions. Oh, how could we improve this feature? And then it goes across our error reporting and our analytics and our internal knowledge and it ties things together. But I personally no longer open some of those things, right? I just like ask my agent to do it. And so it's this weird world, which I think we're going through, um, where we need to figure out how some of those things click together. And then quite often, like we say like, okay, if there's something which we reuse more often, we, we spin up a Glaze app because it then can consume it ourselves in the way we want it, right? And in the UI we want it. So I think it's like, it's a blurry line on the one hand side, I think it's maybe was a bit overdramatic to like, Oh my God, everything gets wiped out. Um, but I see day to day for us, like some tools became just quite frankly, less important, um, or less used and we could probably save some money. Yeah. Have you, you don't have to name them, but have you canceled any SaaS subscriptions this year because of this phenomenon that you've described? Not yet to be honest, um, but we were like, it becomes to this point where you've asked yourself, like, wow, we've been paying so much money and we don't use this anymore. And then everybody wants you to sell the AI integration. It's like, yeah, no, maybe not. Um, so we like, like everybody wants to upsell and you sit there like upsell. We haven't used this tool in a while. Um, so that's like, that's like a weird state to be in, in quite frankly. One tool which I will name, which, yeah. Oh yeah. Yeah, go ahead. One tool which I will name, which like, I mean, we're remote companies, so we're heavy on Slack, but weirdly, I think Slack is in a fantastic place again. Like all these agents in there, you trigger on Claude and whatever. Like it became such a weird, like centerpiece again of software and you see, like, even they realized it and shipped some stuff now, which hasn't happened for like ages. So I, I don't know. This one was a right time, right place. I don't know. Really fun. I went to uh a dinner that Slack hosted last week and uh they're, they're really feeling themselves uh lately. Uh I think they've, I think it's exactly what you said. They realized that they're in a good position to, um, to, to benefit from the AI boom. Um I'll name a name of a company that can go away. Um uh QuickBooks uh by Intuit. I pay them so much money to run my business and the moment I can vibe code that away, uh they're gone. So consider that a threat, QuickBooks. All right. On every episode of this podcast, I like to ask about jobs. Um you employ around 40 people, I believe. And when you last talked about it, or about three-quarters of them were engineers, you have built a product whose premise is that people like me are not going to need an engineer. So two years from now, does Raycast, uh have more engineers or fewer, do you think? So we've stayed for a while now flat and with the, that we're just doing more and more stuff. Every now and then I just look at the amount of code we produce and it's like, it's every engineer will tell you like, that's not really a productivity measure, but it's an approximation. And it's quite frankly insane, like how that went up. And we haven't increased numbers of people. Um, so for us, Pretty well by now and has all the memory and all of this kind of stuff. So just ask, like, whatever you use, ChatGPT, Claude, Raycast, to suggest you like three apps, and then you pick the one you like the most, and then you try it out, and then you'll see how it goes. I think that's a great idea. Thomas, thanks so much for joining us today. Thanks for having me. Always a pleasure. 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 CaseyNewton. My email is casey@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.