Loading summary
Transcript365 lines
- [00:00]
Peter
From like somebody who want to build product. I think there almost can't be a better time, right? We had this a few times before when mobile came around. New thing, new paradigm, Lots of stuff need to be figured out.
- [00:11]
Thomas
The biggest question is AI and how it all fits with us. It is a huge change in the way productivity software works.
- [00:20]
Rid
What actually happened that was a breakthrough for Raycast.
- [00:24]
Inga
There's this culture of designers shipping. Basically everyone has access to the code base and with new AI tools, it's so much easier to actually make a contribution.
- [00:34]
Rid
Are there any guiding principles or things that you've even picked up? Just working with AI as a material, designing an AI product that maybe somebody who's not deep down this rabbit hole wouldn't be thinking of.
- [00:44]
Jordan
When I shared it, I realized that wasn't normal at all and then it just blew up. And I think it does give people a realization like, yeah, there is a lot more effort put into these things.
- [00:55]
Pedro
That's like the most nervous I've been about anything I've done.
- [00:59]
Rid
Welcome to Dive. My name is Rid and this is where designers never stop learning. This week's episode is a special one because we're going to do the ultimate deep dive into one of my favorite products, Raycast. You can think of this almost like five different mini episodes that were all filmed in person. While at Config London, we're going to hear from the founders, the product design team, go deep into their AI strategy, get a behind the scenes of the new mobile app and even hear from the hype and brand team. So let's kick it off by hearing the Raycast origin story because it is way funnier than I imagined in 2017.
- [01:39]
Peter
17.
- [01:40]
Thomas
Yeah. In 2017 both of us moved to London to join Facebook one month apart and joined the same team. After some time we just started having coffees and that's how we became friends. And then we started building few tools on the side. MacOS tools.
- [01:56]
Rid
What year did you make the decision to like. Okay, let's actually start something.
- [01:59]
Thomas
19.
- [01:59]
Peter
Yeah, 19. End of 19. And then we got lucky and got into y beginning of 2020 with kind of the idea of Raycast and joined.
- [02:09]
Rid
What'd you pitch them on?
- [02:10]
Peter
Oh, God.
- [02:11]
Bruno
Oh, no.
- [02:12]
Thomas
Yeah.
- [02:12]
Peter
Are we saying what we pitched for real?
- [02:14]
Rid
Oh, let's go.
- [02:16]
Peter
That's embarrassing. So basically Peter and I, we, we played squash in the morning. That derailed really quickly, right?
- [02:24]
Thomas
Yeah. This is.
- [02:25]
Rid
It was not right.
- [02:27]
Thomas
Yeah, no, no, no, no. Raycast in this idea.
- [02:30]
Peter
So we basically play squash in the morning and then we were quite annoyed of like taking our gy the to the. To work and then afterwards maybe to cinema or pop or whatever. And he had this chimp bag all the time. And we thought, there must be a better way. Right? And y always tells you, like, solve your own problems. This was a huge problem for us, obviously. So what we thought of, like, wouldn't it be cool? You can go to the gym or to your squash place, just regular clothes, no bag whatsoever. You come there, your sports kids wait, there's fresh, you're using it, then you're dropping it there again and you leave without a sports bag. And a service behind the scenes would basically do all the cleaning and stuff for you. So I'm trying to keep it straight.
- [03:15]
Thomas
It's a laundry business.
- [03:16]
Peter
Yeah, let's cut it short. It was a laundry business. So that was what we pitched my fee.
- [03:23]
Rid
I am so happy that we have a record.
- [03:26]
Peter
For whatever reason, they invited us for like an interview in Paris. And that was also quite hilarious when you got there, really.
- [03:33]
Thomas
Yeah. We almost gave up on this idea between we applied for interview and they replied that they were accepted for the interview and there was some backup ideas as well. But then they invited us for interview and we thought, huh, they probably really like this idea about the laundry.
- [03:49]
Peter
This is it.
- [03:49]
Thomas
We should really invest in it. So we actually found a gym that, you know, we should show some progress there. We found a gym that's ready to do the pilot and, you know, we actually prepared the box and like, nice branding and so on.
- [04:02]
Peter
On.
- [04:02]
Thomas
It was quite nice.
- [04:03]
Peter
Even built a mobile app.
- [04:05]
Thomas
Yeah, yeah, there was mobile app.
- [04:06]
Peter
More of a prototype probably. Yeah.
- [04:09]
Thomas
And when we came there, YC interviewed, like is 10 minutes of just drilling and you know, like, they can interrupt you and you do a lot of mock interviews before that and before it always starts. What do you build? And even before I finished the pitch, we were interrupted with the question why Facebook engineers build that. That is where we realized they maybe didn't like how I did it and they just liked the idea of two Facebook engineers building something.
- [04:40]
Peter
And the rest of the interview is very blurry for me. And then we left basically the interview. Right. And sat downstairs.
- [04:46]
Thomas
Yeah, they asked some backup idea and I urgently started searching what was kind of the backup, which was Raycast. And I had something in Sketch, some like, search bar where you can search internal tools and stuff.
- [04:59]
Peter
Yeah.
- [04:59]
Thomas
And I just couldn't find it. And I think like on the phone, I. They like, okay, you know, the interview is finished. You know, go wait for Our answer. And then Aaron, so surreal. Our future partner is. He came down and said like, you know, we not gonna, we like you guys, we're not going to invest in this. Pointing to the box to literally anything else.
- [05:21]
Peter
Basically.
- [05:22]
Thomas
Like, you know, we can, how about we do the follow up and you know, you can maybe kind of developed this other idea which was Raycast. It wasn't named Raycast back then, I think. I don't remember what his name.
- [05:35]
Peter
Yeah, something like that.
- [05:36]
Thomas
And then just going to do a follow up interview which was fairly rare. Like YC usually don't do follow up interviews.
- [05:41]
Peter
Yeah, never heard of this before. Super weird.
- [05:44]
Thomas
We were in this weird state where we like is it good, Is it bad? In a month's time we built a prototype of what kind of Raycast became a search bar with access to other tools. Yeah, we jumped on a call with Aaron and he accepted us to IC after that.
- [06:03]
Peter
Yeah. And I think a month later, one and a half months later, we've been basically in Sfden for yc. So there was like really a rapid change essentially from one day to another which was quite fun.
- [06:14]
Thomas
But I think to me it was important message to others. Like we also want, to be honest, we wanted to start a company. I think it's okay to not have one idea. In fact, we've seen many people change and pivot through ic. They're actually looking just for founders that work together. They have some coherent story for what they're going to build. And that's why they asked what's going to be your edge? Building Laundry, high operational business. Are you good at logistics? No, obviously not. But are you going to build some tool that makes developers more productive? Like that sounds like a good story. And that's I think why they were happy to accept us after the kind of second iteration.
- [07:00]
Peter
Yeah, yeah. I think looking back that probably really shaped like the first year of Raycast. The thesis back then was that we experience ourselves like you work at a big company. The tooling and the apps you use there aren't great, but you got to use them. Right. There's no way around that. Like the company will not change just because you don't like one specific app. Right. And so is there a way that we can make it better to use those tools inside of companies? And that's sort of we felt is a pain that many people like back in the days we thought really strongly about engineers. Nowadays, I mean we own a design podcast, many other people use Raycast, but it really felt like it was sort of the coin, like the first 12 years was really where we dove into that and really focused on that part and yeah, that was sort of what started a lot of what you see nowadays in Raycast and Yeah.
- [07:54]
Rid
So you're both first time founders.
- [07:55]
Peter
Yes.
- [07:56]
Rid
The starting point was a lot lower than I imagined.
- [08:00]
Peter
It could only go up from there.
- [08:03]
Rid
So I kind of want to study the bookends of your journey a little bit and maybe we can look at some of the ways that either you've grown as founders, things you've learned, the ways that you've evolved, your thinking and a theme that I've kind of taken away from all these conversations I've been just talking with the team the last few days is this top down investment into design. And you know, as two engineering founders, I think it would be very logical to assume the opposite. In many ways a very builder first culture, you know. And so where does that come from? Like the people that you've assembled are deeply artistic and expressive and creative. So what does that say about your journey as first time founders?
- [08:52]
Thomas
We both actually coming also from Apple platform engineering, so iOS, macrosos, which I think kind of biased towards appreciating design more and you know, seeing journey of Steve Jobs and how he approached and journey how they approach the design in general. And I've built consumer apps before where it's very important to have high performance like nailing down like small interactions. I'm also fairly critical person to apps. Like I probably dislike majority of apps I use. Not like a designer, I think it's just the desire to use better products. I think for both of us we just wanted to use better software. Like I think this frustration with the status quo is kind of what makes you drive more of this. You know, we need better designers. We, we need to make sure that the features are polished.
- [09:50]
Peter
For three months when you were basing in YC, it was just the two of us, right? And 247 we just built and there was mainly experimenting and figuring things out and try what works. And then afterwards I think none of us is a designer, right? So for us this is a, is a skill set that we appreciate but we, we can't do it ourselves. I think in retrospect we really had to quickly learn to get our vision going right, like have somebody else making it happen, like where we still evolve because we probably have oftentimes quite strong opinions, especially at the beginning when you still shape up a product and you don't know what it is. But we also limit it in expressing ourselves right? So we need to find really quickly a balance of we really value design and we want to make this like front and center and feel like that's a differentiator in many ways, but we can't do it ourselves. Right. And so that's, I think, something that we really quickly had to get into and now assemble like a design team, which is for many early stage startups. I mean we have five designers out of 30 something people.
- [10:52]
Rid
And you're known for design in many ways.
- [10:53]
Peter
Yeah, exactly.
- [10:54]
Rid
You're kind of part of this group of companies that designers look at. It's like, yeah, that's the bar.
- [10:59]
Peter
Yeah, we knew we needed to design it pretty much as the first hire. We can code. Right, we can do all of that. But we also did this very radical thing basically off the yc. We then raised money and the first thing we did, we rebuilt the app completely new design, everything. Because we weren't super happy where it was. And so yeah, we hired Niklas, which was quite funny because he was. When you put out an early product, you do a wait list and he was one of the first people like randomly slid into our wait list. And obviously if there is a wait list, everybody signs up. You look those people up online and it's like, oh, that's a nice triple. That long ago, right? Triple account. And then we felt like, oh, cool, that's quite nice. And when we wanted to hire a designer, we reached out to him and basically got him convinced. Joining us where we were just to.
- [11:41]
Thomas
Do a lot, it's like to be honest, there was actually interviews, I think.
- [11:45]
Rid
We only interviewed, literally just grabbed the.
- [11:48]
Peter
Very first one, basically. Basically worked out, I guess. Yeah, that's fun. Gotta have some luck sometimes.
- [11:56]
Rid
So you start off with a vision just to build better tools. Didn't seem, at least from the way you're talking, that you had this master plan of like the 10 year vision of what this could be. But where have the inflection points come that have added, added clarity to the type of company, the type of product that you want to be? Even where the ceiling for Raycast can come from?
- [12:15]
Thomas
I think product wise we kind of knew it's going to be like a platform. And API was always in the initial discussion. So we certainly seen it as expandable tool because we knew we cannot build everything ourselves. But company wise, I don't think we ever discussed, hey, what's going to be our company.
- [12:32]
Peter
I feel like we have maybe like a good grasp for like a certain period of time and then it gets blurry. Again and so then you move this path along and then you get at some point need to clarify the blurry lines again and look a bit further. So it's very incremental for us. I think we don't have this necessary. Like this is what Raycos is. If you look back in the last five years, we'd be doing it. Initially it started as sort of this drop in replacement for Spotlight because we knew there was going to be just one of those interfaces that you're going to have. So we knew we're going to need to be better than Apple Spotlight. So that was sort of the initial milestone. Then we knew we want to build in a lot of powerful functionality. So we added like extensions so you can do literally a thousand things and made it a community thing, which was really important for us. We wanted to build a community around Raycors. We felt that something that can really work where people fall not just in love with the tour, but coming along on the journey and helping us building that. And then as we went we felt like, okay, we have this launcher where you can do this Spotlight replacement, but we felt there are many other things which we feel like missing a bit on Mac and where we want to bring. So we have the notes. Now we have an AI chat which almost like are standalone applications. So nowadays how we think about it is much more like a suite of productivity tools. And the launcher is the thing that ties it all together. Like oftentimes it's as cheesy as it sometimes sounds. Becomes a bit of this operating system, right? Like it sits on top of that. It's probably the closest to an operating system that we can think of right now. I think the most likely won't be any new operating systems in the near future, but you can still make those systems better. And that's what I really inspire too.
- [14:11]
Rid
Can you take us to the murky point where it's a little bit unclear what to do next? And I'm really interested to hear how product strategy works, how you know what to prioritize, especially as given your entry point, you can almost go in any direction.
- [14:27]
Peter
Oh man, that's often.
- [14:28]
Rid
Tell yourself a story for why that's the right thing to do. So how do you know what to do and how does that decision making process, process happen?
- [14:34]
Peter
So we had a few like sort of situations where it wasn't super clear what to do next. Like one was we launched Raycost for teams and we thought like, oh so we're going to be smart. So what we're going to do. We see every company, what they do, they're building a. Usually let's take notion, they build a product for individuals and then they make a paid version for individuals and then they're going to make a team version, right? And we say, if we're going to end up at teams, let's just skip this other thing and go straight away to teams, right? And so we did that didn't work that well. Let's cut it short, basically. And we felt then in a bit of a void, you come out and launch and like, so that was the first time when we launched something where it didn't blow us away because usually we had super good reaction all the time. And so then we were walking back and it's like, okay, let's look into it, right? And there were sort of two ways we could operate there. Like one, you iterate out of it, right, and say, okay, cool, that works. We iterate and iterate until we find it. Or you take a step back and take a different action direction, right? We decided for the latter. We basically said, hey, look, actually what we see here, Raycost is really a tool that spreads really organically well with individuals and they fall in love with it and then they're advocating for it inside of companies and a company might be, or a team might be not ready to adopt fully Raycost yet, but individuals are. So let's lean heavily on that and follow that thread and see where it leads us, right? And so then we basically said, okay, we're going to do Raycos Pro, which is an additional set of features, they're going to be paid. Which then also collided with basically AI coming around and it becomes something like a really attractive package. That maybe was a bit of luck, to be honest, but when we launched that, it was like, that was again a moment where like, wow, okay, there's real appetite for that now we're basically coming back. It's like, okay, there is appetite for that. And now we're seeing teams coming to us because individuals starting paying for it. And so yeah, that was sort of one this blurry moment with literally where they're like, oh, what are you doing next? You don't really see that. And then you oftentimes have to come back to the principles. And for us it's always like, we have a strong team, as you got to know, basically during this afternoon. So we feel we can sort of build a lot of things and pretty much everything we want as we are this entry point. But then it also needs to be a Cohesive story. Right. It needs to land with the user very well and they need to follow that and understand that as well. I think that was sort of one of this, I wouldn't say a make or break moment, but it was one of those moments where we, for like several weeks or months we were in this limbo state. It's like, ah, where is this going and are we going to do this pass or that pass? And I mean so far it seemed to be that was a good direction to like taking a step back and going back to like what are we really good at and like doubling down on that and following that threat.
- [17:17]
Thomas
Yeah. But before that, first two years, a lot of this just user driven. We are very close in the community. Feedback just comes. We don't need to ask people for that.
- [17:27]
Peter
Sure.
- [17:27]
Rid
You have no shortage of feedback.
- [17:29]
Thomas
Yeah. It got to a point where we don't know how to reply to every piece of feedback. Our support team is struggling. So there wasn't much of the road mapping. It was literally like what's the next thing we built? Okay. And people ask for that. I think this would be cool. Like quick link snippets. We build a lot of productivity tools within one. There's a lot of known things we can do, do. And I think first two years was pretty much that. And there was one big thing is the extensions and store, which was just a big project that took us a few iterations. Now the kind of the, the biggest question is AI and how it all fits with us. I think the future is uncertain for a lot of products and companies and what does it mean, you know, AI AGI. So we're trying to find our way there as well because it is a huge change in the way productivity software works.
- [18:27]
Rid
Uncertain for a lot of companies, but I'm sure a lot of people listening are also jealous of where you fit into the stack.
- [18:32]
Thomas
Yeah. For us we couldn't ignore it. It's the obvious productivity booster. The way we use computers now is vastly different even from three years ago.
- [18:44]
Peter
Yeah. I hope it's going to shift when we talk more about features and stuff and less about AI. Right. AI becomes more. I mean we don't talk about databases and things like that. Right. We talk about the features we build on top of or the apps or the products. And I feel like we haven't reached that point yet.
- [18:59]
Rid
No.
- [19:00]
Peter
We still hear way too much about LLM capabilities and way too little about use cases and problems you actually solve. But yeah, there's true utility there. Right. It's Hard to ignore that as well.
- [19:14]
Rid
Someone described it to me as like a residual AI tourism that still exists in the industry, where it's like, if you're silly to not label things as AI because it actually does have a material impact, and so you almost like, it's not even the tools that can make that change. I think it's actually like a societal level. Like, when there's no longer a benefit to tools, then we'll stop labeling everything as AI.
- [19:34]
Peter
Yeah.
- [19:35]
Rid
Now the fact is there is.
- [19:36]
Peter
Yeah, there's this nice saying. It's like, first it's machine learning, then it's AI, then it's just an algorithm. Right. And so I think we still had AI space, and even when we shipped it, we thought about how. How do we name it? Right? And we thought about, oh, we could use it. We could name it Smart Answers. And then we felt like, that's like, kind of silly, right? Is it smart? Maybe yes, maybe no. But it's obviously going to be called AI, Right? And I think in the Changelog, it was even like ChatGPT for Mac, because ChatGPT already was the verb or the term that you would use to describe AI. But, yeah, I think over time this will. Will go away and maybe there will be less sparkles, icons everywhere.
- [20:12]
Thomas
Eureka's intelligence.
- [20:14]
Peter
Oh, yeah, exactly.
- [20:18]
Rid
Real quick message. And then we can jump back into it. Look, I'm not a motion designer, but I'm working on a new product right now, and I've been using Jitter for everything, and it's honestly a cheat code. I've been creating Lottie files, micro animation showreels for Twitter, and I didn't want to use After Effects because it would have taken me all day to figure that stuff out. But. But there's no steep learning curve with Jitter. The whole product is super intuitive. They even have incredible community templates that I've been using to jumpstart my project. So even though I wouldn't consider myself a motion designer, I feel like Jitter gives me superpowers, which is why I can't recommend this product enough. If you had to dive dot club Jitter, you can spin up something stunning in a matter of minutes. I'm all in on paper as the next great design tool. And this is coming from someone who's taught Figma to more people than just about anyone. But I'm ready. It's time for a tool that puts creativity at the heart of everything, not systems and processes. And the way that we design and create software is changing faster than ever. And the tools we're using were made for yesterday's tasks. Instead, paper is putting designers first and foremost, building all kinds of features that I've always wanted. So if you want to be one of the very first people to use the next generation of design tooling, then head to Dive Club Slash Paper to get on that list. Early Access launches in May, so that's Dive Dot Club Slash Paper. Okay, now on to the episode. Next up, we're going to hear from the Raycast product design team. We're going to go into how they collaborate all the things that make their weekly rituals so unique and a lot more so. So let's start by meeting the team.
- [21:59]
Roy
Yeah. So I'm Roy and I focus on anything that has to do with AI within Raycast and I'm based in Berlin, Germany.
- [22:04]
Inga
I'm Alex, product designer based in Sofia, Bulgaria. Right now I mostly focus on our design system and our Windows experience.
- [22:14]
Bruno
Hey, I'm Jordan. I work on the iOS app as well as some Mac features and I'm based in South Wales.
- [22:20]
Rid
Now that we've met the gang, I wanted to start by diving right in to how they're able to consistently hit this high bar for craft.
- [22:28]
Inga
I think the important thing really is that it's not just design that's obsessed with craft. I would say the whole team at Raycast has a very high bar to what we ship. So when I joined Raycast I was really impressed at how high the level of understanding for design is, even throughout the whole team, engineering, leadership, everyone. So I think it's just something that's in the culture of the whole company.
- [22:54]
Roy
Yeah, yeah, that's a great point. I feel like it's a very design minded company and everybody's very good at what they do. We all have very different expertises. I would say we have the trust and the freedom to just explore our own sort of like niche. Like I'll go very deep on like technical stuff and like how things work. And obviously Inga is super visual and yeah, in a way that's all gets brought together and shines through the product and through like marketing and hypotherrial.
- [23:21]
Rid
So obviously like a design driven company. A lot of that's top down kind of baked into the culture from the beginning. But how does that translate into like a normal week? You know, like what does it actually look like in practice for Raycast to care about design?
- [23:34]
Bruno
I think it starts on a Monday with a team meeting. I think we all share our work, whether it's from a design perspective, an Engineer's perspective product founders. I think we're all sharing often and frequently with each other and that opens up the feedback loop early on in the week. And then as we progress, we have design digests and these weekly crits that help us sort of build on top of that and then go deeper in some of those details. And as Roy said, we all have different skill sets and expertise and I think that allows us to always cover everyone's work with the different sort of eyes that we need.
- [24:09]
Peter
Yeah.
- [24:09]
Roy
One interesting bit for the team meeting, like, it's quite low key already, but to just share your screen, it's always a little bit too much effort if it's just a tiny thing that you want to share. So every roughly once a month or so, we as a design team open up this Figma file and dump any visual work that we've done. It could be a silly exploration, could be a new icon, could be a new wallpaper, anything, basically. Then we present that as like a design digest. So that's ultra low barrier to like share something with the team. And then you don't have to say like, oh, I worked on a wallpaper. Share your screen for like five seconds. Someone from the team then just shares the design digest, which is like slides in Figma, basically. Right. We just have a wallpaper slide that we just dump our things on and mark it with our, with our name and then, yeah, walk through it. And usually most of the things don't need an explanation. So you don't need to go through the hassle of sharing your screen.
- [24:58]
Rid
I didn't realize that was happening in a Figma file. That makes so much sense because you're right, like, if you're going to share your screen, then you feel this pressure of like, well, the thing I'm about to share has to be worthy of the screen share and the gap and the logistics of doing the zoom settings. So that's really cool.
- [25:13]
Bruno
I think what's also tricky is, like, our timelines from design to dev are usually pretty tight, like the way we build and iterate. So sometimes you're like, do I share this design? Even though in like, the engineer might already have built this and want to share in production. So sometimes it helps just to go straight in with the production version. Even if it's slightly rough. I think it still opens up that discussion of like, design quality and craft and where we can sort of quickly address that to get it to just an even better version within a day or two.
- [25:40]
Inga
Yeah. And I think it's very key Also that this meeting is on Monday because it's a really nice start of the week to see what everyone has built recently. It's just inspiring to start your week with, you know, seeing everything that's been going on at Raycast and throughout the week. One of the key things which I think helps with the craft is people have a lot of time meeting free to actually focus. One of Raycast's core aspects of the Raycast mission is to help people focus more, to find their flow. And I think that really shows through in the way we work because Raycast is probably the company I've been in with the least meetings. You have so much time to focus on design, to focus on your craft, to iterate, to try a lot of different versions.
- [26:22]
Roy
Yeah, but I think it's not the case for everyone in the company. But definitely design team. We work on a couple of projects most of the time at the same time. But yeah, meeting wise, my entire Tuesday, my entire Wednesday, often half of my Thursday are completely free.
- [26:37]
Rid
It's the dream.
- [26:38]
Roy
Yeah, exactly. And the Thursday morning is focused on this thing that we call the Pixel Rumble, which is this session where every designer plus the two founders basically join and we just. It could be anything and explorations very early, something more polished. Like you're close to releasing something and it's about like, like, like an icon for the release media or whatever, anything you could bring there. And it's just crits, basically, but with the founders and usually that's a very easy way to like unblock yourself or just get some input and feedback.
- [27:05]
Rid
Let's go deeper into Thursday real quick before we move on. So, like, you have the Pixel Rumble, which also is amazing branding. So talk to me a little bit more about how that works, how you as a designer prepare, what is the ideal outcome coming from that meeting. Let's just go a little bit deeper in that part.
- [27:21]
Inga
What we have is we just have a notion doc for every Pixel Rumble where any designer can sign up to share a piece of work, whether it's work in progress, whether it's something they've been thinking about or an idea they have, or an active project. And then during the meeting, we would take turns to basically give everyone the context of this design problem that we're working on. We share our thoughts and usually we would come in with sort of a point of view of where we want to take it. But also it's really good to have both the design team and the founders really close to the work and involved in that discussion because you get really actionable feedback.
- [28:03]
Roy
Yeah, it's a very low key way of signing up for it. Like signing up sounds quite heavy, but it's just a bullet list for, for every meeting and you just put a topic and your name there and that's all anyone can bring anything. Basically, there's no set format. Usually depending on what you're showing, you could provide some context of like what you want to get out of it. Right. Like maybe you're stuck or maybe you just want to choose between two versions of something and get that type of feedback.
- [28:25]
Bruno
Usually for me it's like I've not prepped anything and I've signed up 15 minutes before and it's a, it's a mess of my FIGMA file. But I think that that shows the like low effort, low level of what you need. There's no barrier to present slides or have rationale behind that. It's almost like, let's just jump into figma and usually five different versions that are moving around.
- [28:48]
Roy
And it shouldn't turn into that thing where you're going to prepare for Pixel Rumble. Right. We don't want presentations there. We just want the raw stuff on the ideas and unblock people or help out.
- [28:58]
Bruno
I think we're also all guilty of maybe overthinking what we're doing. It's like, oh, it's too early, oh, it's too late. Oh, I'm not sure about this. Or actually, sometimes halfway through a session, I'm like, actually this is still early enough that I can share it. Like there's no need to hold this back because I know I'm going to get something valuable out of it.
- [29:15]
Rid
I like that a lot. Like, I think there's this malady in these types of meetings, especially as a company is scaling. And so it does seem like even from the early onset, you all are really intentionally making sure that this is a safe space to share. Whatever. Like, whatever, whatever you're thinking, whatever you're working on at any, at any stage too. Which is definitely. I'm kind of underlining that for myself.
- [29:34]
Bruno
Sometimes like the smallest thing becomes like a 45 minute discussion too that you didn't expect. You're like, oh, this will be quick. And then suddenly like everyone's got an opinion. You're back and forth bouncing ideas and then those are usually the best kinds. Right. Because it's like this started off as like a small, small idea and suddenly you've like solved it all or got a much better picture.
- [29:50]
Roy
Yeah. Which is maybe also related to the fact that we're now multi platform. Right. With iOS and with the desktop app and at some point soon also Windows. You know, have you considered how this would work if we, I don't know, need to sync the chat from desktop to Mac? There's so many complexity in there. So much complexity. Yeah. Sometimes in a very early exploration you sort of forget about that a little bit, which is also fine. I think that shouldn't restrict us of exploring things, but this could come up in like a pixel rumble where like a very small thing turns into like a bigger discussion. But that's like amazingly helpful.
- [30:20]
Rid
All right, so let's keep going through the weekly pattern here. We're on Thursday. Where do we go from here?
- [30:25]
Bruno
Friday. I mean, Fridays for us are Ray days, which is a free day to almost break apart from your weekly work and to pick up anything you want to explore. Sometimes that's like a new feature idea, sometimes it's bug fixing, sometimes it's just enhancing another feature. But I think we all have the freedom as designers, engineers, everyone in the company to just take time to pivot and think about something else. And usually some of the features that we ship come out of those, like raycast Focus was a Ray day, and yeah, quite a few others too. So they then feed back into the Monday and then you share that and that sort of inspires something else.
- [31:01]
Rid
I like that. Do you have to like declare what your Ray day is going to be?
- [31:05]
Bruno
Yeah, there's a Slack message in the morning. Right. And we just, yeah, add to the thread about what we're working on. Sometimes it's hard to like pinpoint what you're going to work on. Sometimes it'll take a couple of hours going through notes like Slack channels, the community. Maybe you find like a bug in there and you're like, actually, I want to fix that because I've had that issue. But yeah, we just post what we're doing and then sometimes we link up and it's like two or three people, like an engineer designer also working on something that day.
- [31:29]
Inga
Group ready. Sometimes you would come into Friday with a clear idea of what you want to design. Maybe you have an idea that you wanted to explore for a while or sometimes it's more spontaneous. You know, you wake up on Friday, you start your day and you do you have an idea that you want to explore.
- [31:45]
Roy
So yeah, yeah, just something. You realize something is super broken. Like recently there was the Pro walkthrough. So you sign up for Pro and then you get this little sort of extra walkthrough. Of like, hey, congratulations, super cool. This is the new stuff you have access to now. And I realized that we just never created assets for the light mode and it looked terribly broken. You can imagine, right?
- [32:07]
Rid
Yeah.
- [32:08]
Roy
And for me that was the first moment to kind of just set up the whole dev environment in XCODE and just fix it myself. Like it's too much work to go back and forth with an engineer providing assets and stuff. So it's on the one hand it was creating all those assets and at the same time fixing it yourself, but that was like a perfect grade A example because could go from start to finish in a day and just have a very demoable thing on a Monday and we ship it the next week. Basically.
- [32:33]
Rid
Quick context. How much of a technical background do you all have? How often are you working in code? Are we like pretty much just in Figma, how does that work? And maybe we can even use that as a entry point to learn a little bit more about how you all collaborate. Collaborate with engineers too.
- [32:47]
Inga
When I started my career I was both a designer and a front end engineer. But during my career I focused a lot more on design. So the past few years before joining Raycast I wasn't writing code actively except for maybe a personal project, my portfolio, website, et cetera. But when I joined Raycast there's this culture of designers shipping. So basically everyone has access to the code base and with new AI tools it's so much easier to actually make a contribution. So for me, I mean right now I would say I spend a lot of time in cursor apart from Figma. And it's really great to be able to onboard yourself to the code base of the product. So personally I would use it even in the beginning of kind of understanding how everything is structured, what lives in what file and understanding what everything is. It also allows me to make small visual improvements in the product without even bothering one of the engineers.
- [33:44]
Roy
Yeah, my background was a bit of a mix of design and software engineering and communication studies. Back in that day it was much, much easier to create a website, just code it up. At some point I lost that a bit like just front end engineering just became way too complex. But always interested in just prototyping like Framer Classic was my corner. Right. I love that and just the interactivity of it and all. At Raycast, indeed it is encouraged to just contribute where you can, where you want and just go in it's accessible to everyone.
- [34:15]
Rid
Something I kind of look for in teams is just creating space to almost sand down the ui like I really like that analogy of, like, just always having space to go back and kind of, like get rid of all the rough edges, everything that's, like, sticking out. And so that was kind of why I was asking about the technical background, because it was like feeling like that was the way that you all worked.
- [34:32]
Bruno
Those things are so painful to do as, like, linear issues, screenshots, annotates, wait for an engineer to pick it up and then you still kind of need to review it. Maybe it's still slightly off. So I think, yeah, like I said, if it's a bit of CSS or if it's a bit of, like, Swift UI styling, like, those things are pretty easy to pull off. I think the biggest hurdle is getting set up, getting environments working.
- [34:54]
Roy
Right.
- [34:55]
Bruno
That's the hurdle. And, like, committing to the main branch and, like, making sure you don't mess it up is, like, always my worry. Like, two years in, I accidentally did.
- [35:04]
Rid
That on the plane ride here. I, like, was just up late, like, making a bunch of changes in cursor, and I'm like, ah, this looks good. Commit. And I'm like, oh, no.
- [35:13]
Bruno
Yeah. And then if, like a bug report comes in, like, you start swearing and you're like, wait, did I introduce this bug?
- [35:19]
Roy
Yeah, that's good.
- [35:20]
Rid
Yeah, yeah, yeah. There's almost like a social cost, you know, like that back and forth. Like, if the back and forth is happening in GitHub. Comments, it doesn't feel good as a designer to be like, nah, I know you just responded to my feedback, but like this border radius, like, come on, like, it's like, you know, it's like off by two pixels.
- [35:36]
Bruno
Yeah, it's much faster. I think you cut timelines so much.
- [35:39]
Roy
Yeah. And in the end, I think I've always believed this, but I think in the age of AI, it becomes more and more relevant and accessible. So I think it's the responsibility of the designer, right, to own everything visual and how it feels and how it works. So the fact that you can jump in and just fix a couple of things here and there, I mean, that's kind of the case for quite a bunch of years by now, but I think it gets more accessible now with AI to be able to, yeah. Code up stuff. What are the vive codings that we would call it?
- [36:05]
Rid
Anything else in the collaboration space that we haven't talked about? I think it is interesting specifically because you are this, like, remote team where everyone kind of has clear areas of ownership to an extent. So, like, how much are you actually parrying on the same body of work or how does, like, feedback culture happen, even amongst the design team?
- [36:26]
Roy
Yeah. So even though we don't have that many scheduled meetings, I quite often just jump in a huddle with someone, like, especially if it's touching multiple platforms. So with Jordan, for example, we just sometimes huddle and we don't schedule that. Right. It was like, okay, it's Thursday. I'm working on this thing. I'm maybe a bit stuck. Or I don't know how iOS would solve this. Like, let's jump in a huddle and figure it out together.
- [36:45]
Bruno
We've also done some design only projects. Like myself and Roy worked on a bunch of, you know, AI vision work. Like, some of those things aren't like the now projects, but more looking at the future. So those were super fun to just bounce ideas off and almost felt like you're back in, like an agency or like a larger team, because we do work in quite small squads. So that's super fun as well, just to step back from production and almost just like, dive deeper from a design perspective.
- [37:12]
Inga
Yeah. Sometimes it's fun to collaborate with another designer on a problem you've been thinking about and to get a unique perspective from them, you know, and sometimes it could be an engineer because it's also really good to, I think, to discuss ideas early with engineering to make sure what you're designing is possible, not feeling awkward and like kind of messaging someone jumping on a huddle spontaneously. I think it's really important. And the fact we mentioned in the beginning that we don't have that many scheduled meetings, I think opens it up to spontaneous collaborations.
- [37:46]
Bruno
It's also, again, leaning into people's specialties. Right. Like we said, like, if it's an illustration, like, I'm not going to attempt this, like, what's the point? Or like, if you're working on a small icon, like, you might ping it to someone who's got a good eye for icon design and can tweak it and help you with clothes out. And those are. They're not planned.
- [38:02]
Roy
Right.
- [38:02]
Bruno
It's just you ping a slack message, you get it fixed, and then you move on. So I think, yeah, understanding and getting used to just being able to leverage people like, when you need them.
- [38:11]
Roy
That's usually when I huddle, Jordan. Like, hey, look at my ugly icons.
- [38:16]
Rid
What about the engineering piece? Like, Jordan, you said something interesting where maybe prod would be almost exactly, like, in tandem with design. It definitely doesn't feel waterfall y, but maybe we could talk a little bit about, like, how do you all collaborate with engineers? What are some of the patterns or ways that you've had success so far?
- [38:34]
Bruno
I think the way we've worked a lot on iOS and sometimes on like Raycast Focus and stuff like that is like you try and think about the smallest, roughest entry into the code base and you add that as soon as possible. That's deployed the next day and everyone's using it in internal and then you just build upon that. And I've not that often had designs specked out to the full feature set because you almost don't know where the feature set is that early on, especially if it's on array day or a couple of days after that. So I think that's where it's so in tandem and you're really heavily in huddles or on Slack messaging back and forth, discussing product and features and then trying to implement them as soon as possible. Because I think a lot of productivity is like, you have to feel it and use it to understand this whole flat design is really hard. I almost have to pull up figma mirror if it's on iOS, or even convert it to SwiftUI so I can feel it. I think it's that tangible stuff. And you can see Figma doing that with Figma make right. It's pivotal to making it feel right. So for me, as much as possible, getting it into the code base without doing too many polished designs early on.
- [39:40]
Roy
This is maybe the most frequent type of feedback we get in like a pixel rumble. Right. Like, it's both looks cool. We don't know. Let's get it internal as fast as possible and we can see how it feels. Like we can see how it feels is so important because Raycast is. So it's there and it's out of the way again. And that's the value.
- [39:56]
Rid
It has to be snappy. It's like an extension of my fingertips.
- [39:58]
Bruno
Exactly.
- [39:59]
Roy
Yeah. So we all know that we don't want to have slow animations in there and a couple of other things, but sometimes you don't know what's going to work. Like, do we want this?
- [40:06]
Peter
It's.
- [40:06]
Roy
Yeah, two options. Let's pick the one we feel the best about and if it doesn't work, we fall back to the other one and try that.
- [40:11]
Bruno
Yeah. I think one example for Focus was like, where does the bar go by default? Is it the bottom? Is it the top?
- [40:16]
Rid
Like, where does it go by default?
- [40:18]
Bruno
Bottom is. It is above the dock. They're like, focus, I don't know. By default, I Hope so. But that's the thing. We just like we made a decision and that's not locked in and you add in and until you've used it and you don't know, is it going to interrupt the bottom of my window, the top of my window notifications like. Like you're on a platform with so many other elements. It's not a website, it's not, you know, a contained app. So you have to figure those things out.
- [40:42]
Roy
Yeah. And also things like, for example, sound is a good one with focus. Right. Like at some point we realize, like, it's not complete when it doesn't have this dings everywhere at the very end. Right. Like that completes it. Like it's like the perfect sort of checkbox. Like, you're now out of your focus session. Go do whatever you want to do.
- [40:58]
Rid
Well, I want to go deep on Focus for a second. Who brought that to the table? Like, how did that originate and who worked on it? Because there's a lot of different pieces. Like, it's so simple and yet it kind of touches everything. But then you also have like these crazy animations. Like, what's the best?
- [41:11]
Bruno
It started with either Thomas or Peter had the idea. And then one Ray day, Peter messaged me. He was like, do you want to work on this idea for Focus? Let's jump on a call. I built out like a super rough UI of in Swift, of just like how the form would work. We were like, let's pick a timer, a title, like really rough. And then we demoed that on the Monday and then people were excited and we decided just to push it forward.
- [41:33]
Roy
And I think Focus is an interesting one because I think almost nobody initially felt like this is an amazing product.
- [41:39]
Bruno
Oh my God. I thought like five people would use it. Yeah, every single day. Yeah. But so unlike the stuff that we typically do, right? And you go, oh, this is so niche. It's like getting things done, you know, those sorts of like methodologies or principles. And you're like, oh, you try it and you give it.
- [41:55]
Inga
Yeah, well, that's an example of something that, you know, you ship it into internal. And because everyone at Raycast is a Raycast user, you get all of this immediate feedback throughout the company and over a few weeks you iterate on it, you make it better and by the time it ships, you've already solved a lot of these paper cuts or small problems and you have a feature that's actually really good, but you're kind of used to it, whereas users are experiencing it for the very first time. So for them very often we're very often surprised by the reaction.
- [42:29]
Roy
Yeah, we get.
- [42:30]
Rid
It's a good example too of like the functionality at its core is so simple. There's a lot of the little subtle execution details like hovering over to have the three second timer, the way that the success state operates. Like, I think the reason you probably feel comfortable getting something to code as fast as possible is because you all are empowered to kind of make it better and have all those like little micro improvements where I don't know, maybe you feel this way but like traditionally years ago if I was a designer I would be so worried to get something into code because then it's like, oh my God, I have to go through someone every single time. It's going to be so much more work to change things. And so having that kind of fluid.
- [43:09]
Bruno
Back and forth, I think it's key to the culture of Raycast is that like nothing's locked in until, you know, we all agree that and it feels right. And I think Focus actually had more features to begin with that we pulled out.
- [43:21]
Rid
Interesting.
- [43:22]
Bruno
And then we wait to see. It's like, okay, well we don't need that, let's pull it out, let's hear feedback and see how people use it and then we might bring it back. Like we had a list of things that we were like, oh, these probably will get requested and it's nice to validate when users use it and they go, oh yeah, they asked for that, that and that. And then we know that we're ready then to almost bring that back to the table.
- [43:41]
Roy
And maybe that's also interesting with the. I don't know if this is super common everywhere. I don't think so actually. But like all the projects obviously get a Slack channel, but the Slack channels are open, right. So everyone's encouraged to join that project and just provide feedback. And. And I would say almost everyone does this, right? Start using the feature and then maybe it's not your thing but you realize like ah, notice this little like visual bug or notice that doesn't really work on my machine. Right. Or doesn't we all have different browsers? Browsers, different languages set up like older machines, older versions of macOS, like all these things just bubble up. And I think that that's one of the reasons we get rid of the rough edges before it's out.
- [44:19]
Rid
In a culture of feedback though, a lot of times you get feedback that you just fundamentally disagree with as the designer. The way it's like there's a little Bit more of an open ended taste there. Is that something that like you as the designer, feel empowered to be like, you know, actually like I'm kind of owning this project, I think we should do this way or how do you even walk that tightrope? In a culture where there's open slack channels, a bunch of people weighing in, everyone feels comfortable expressing their opinion. What's that like?
- [44:44]
Inga
I think that. But the unique thing about Raycast is that everyone naturally uses the product. In many companies you have to actively make the effort, you know, to use the product for testing. But at Raycast, everyone uses it for their day to day work and life. So everyone's feedback I think is valid because they're a user.
- [45:05]
Bruno
I think it comes down to intuition, right? As designers we have that gut feeling and I think it's like, how strong is that? That like, if I'm like, ah, yeah, well maybe you're right. Like productivity is so like unique and everyone has a different perspective or like goal to achieve. So like, just because I can't see their point of view doesn't mean it's right or wrong. But I think it's, if it's like feedback towards something that's existing that you've added and you strongly believe in, then it's a different point of view and you, maybe you want to back that up or sort of argue your case. But yeah, I think that's the, the benefit is that like, yeah, if it feels right, it feels right and you almost don't have to like argue it too much unless the other person strongly disagrees or you figure it out. But like Roy said, I think sometimes you have to just try it to see and believe it or then pull it back out.
- [45:51]
Rid
I can't remember who was telling me, I'm like racking my brain trying to remember. But it was an async culture where the like format, the taxonomy of how you gave feedback was you had to put some kind of emoji in front of it. Like there was a little emoji system that was indicative of how strongly do I care about this, like what is my conviction level and what I'm about to say? Because that so often gets lost and it works both ways, you know. So I think that's a really interesting point that I'm almost thinking even for myself, like how do I want to capture that sentiment, you know?
- [46:22]
Roy
Yeah, you kind of want to attach a certain number or value to your feedback as well. Right. Sometimes you can. I think this needs to be different. But if, if you rate that on a scale, somebody coined this, like, many, many years ago. A designer. I forgot who this was as well. But, like, sort of, how many fucks do you give? Right? 0 to 10? Like, yeah, I don't want to die on this hill. Like, I think it should be different, but two? Or if it's an eight, then you better discuss that.
- [46:46]
Thomas
Yeah.
- [46:46]
Bruno
Save your energy for another day.
- [46:48]
Peter
Right.
- [46:48]
Bruno
There's gonna be a bigger discussion that you may be more passionate about. So.
- [46:52]
Roy
Yeah. And then, you know, it launches and then indeed, it wasn't a thing at all. Right. Nobody notices.
- [46:57]
Rid
I want to make sure we're covering everything. So as a way to either, like, summarize or cover things that we haven't talked about, I have a question for each of you, which is, let's say hypothetically, you have a disagreement with Pedro tomorrow. It is so extreme that you cannot stay at this company and you are leaving and you are going to go elsewhere. What is a element of the way that design works or the design culture at Raycast that you would want to make sure to take with you to the next company that you join?
- [47:25]
Inga
I would say maybe the open culture of sharing and this low barrier to sharing work. I think it's really important and it's missing in so many, so many organizations. I think just the. The focus we have, like, the time we have to actually explore ideas.
- [47:41]
Roy
I mean, you still have to fight for it, but I would really miss Ray Days. Everything else is very personal. Right. I cannot make the founders of the next company care about design as much as ours do. Right. That needs to be there or not. I cannot bring that. I think Pixel Rumble very much depends on that, whether that's the right people together in the room. But I think Ray Day, it just proved that so many ideas were born there and were actually shipped. They were actually meaningful. That is, I think if I would start a company, that would definitely be part of it.
- [48:11]
Rid
Yeah, it's really cool.
- [48:12]
Bruno
Yeah. I can't disagree with the Ray Days, but I think also for me, something else would be just like diversifying the skill set of your designers. I've seen so many companies who try to hire designers with similar molds. They focus on research, they focus on UX design. And I think the moat and the strength this team has over other teams I've worked on is we all sort of COVID these different areas and, like, quite deeply, I think, you know, I could pick a different designer to pick a different task. And I think if you're building a team or looking for a team, is finding a team with a diverse skill set so you can kind of COVID everything in house.
- [48:48]
Rid
Next up, we're going to dive deep into Raycast's AI strategy, all the way down into the details of interaction design, up to the big picture of product positioning. So let's bring back Roy, who is leading the design, and also Thomas, the CEO. And I want to start this chapter by going back before ChatGPT launched.
- [49:06]
Peter
When we started thinking about Raycast Pro, AI wasn't in that list. Right?
- [49:10]
Rid
How convenient.
- [49:11]
Peter
Yeah, how convenient. Right. So that was like very different. Like back in the days, like before 2022, we haven't really thought about AI. And so when this came around and then I think in January, we realized, yeah, we got to hop on that train because it became very apparent that this is hugely beneficial for productivity. And so really quickly we started with what is now quick AI, which is you put in a prompt and then basically get an answer. Initially this looked like a chat inside of Raycast. And then when we had that, people were like, oh, this is cool, but sometimes it gets into the way, so maybe want to detach the window. And then AI chat sort of was born, which lives separate, almost like a separate app by now, and went through many evolutions that we went there. And so, yeah, this really quickly became something we gotta have to have but are still exploring two years later or like three years later. Almost like we're still exploring what Ray Cost AI is.
- [50:09]
Rid
Let's go into the many evolutions. What were some of the things that you were exploring? What were like, the decision points early on from a design standpoint, we're like, okay, how do we think about which direction we want to take this? How do we want to implement AI?
- [50:22]
Peter
We're a fully remote company, right? So people are spread around Europe. But when you have these moments where you don't know what you're building, we feel it's really hard doing that remote because you kind of want to bounce ideas of each other and get really into this rhythm of figuring things out. And so we felt like we at a stage where, sure, the AI stuff works inside of Raycoast, but we felt not happy with the user experience. But we also felt we need to experiment and prototype. And so we said, okay, let's take a week where we get a bunch of people together who worked on AI. And so the core thing which we wanted to concentrate on, Raycours can do a lot of things. And the problem with that, there's a lot of muscle memory. People search for stuff. So they expect that the search stays like this. They have all these routines building up so we can't break any of this previous behavior. So the really tricky part is finding a pattern that fits into this without breaking your typing in, like launching an app and hit return. And so we experiment. There was a few different patterns. Like one was you would open Raycast and you press space and then you go into AI mode, basically, which was actually we had this internally and felt quite nice. What we didn't really like about it, it becomes two separate modes where you have to think about am I in the regular mode or in the AI mode? And so we decided against that. The other path that we had was basically what we more or less shipped now was, okay, what if we can do something with tab, which is a very regular key that you oftentimes use for auto completions and things to accept. And so the other thing was like, okay, you type in your question and then you hit tab and then you basically get an AI answering your style. And so that it feels like now very easy, as in hindsight it always is. Right. But getting to that point took a lot of time. We had a few other things that didn't really materialize. Like one was, can we sort of transform the UI a bit more so it becomes more obvious when you're going into this AI mode? But it was quite challenging because in Raycross everything is very snappy. We don't have animations, everything. Like it's really simple design to not get into the way. And eventually we basically ended up on the tap to AI is kind of how we're calling it.
- [52:38]
Roy
And it was very, very limited in the beginning when I joined. Right. So I've seen mocks of like all the explorations where I wanted to go. Because it's kind of funny how most of the AI in the chat, like you have the input at the bottom and the answer streaming in at the top. Right. But with Raycast, obviously you have your search bar at the top and then all your content is below that. So this switch to quick AI where we suddenly changed that around and like how do we deal with follow up questions? And the very, very first version, and it was there for quite a long time, it didn't have follow up questions quicker. That was your dead end. You would type something again and it's a new thread, right. So what you ended up doing was you had a prompt and it didn't do the thing that you expected it to do, especially with those early models, like, like 3.5 et cetera. It didn't really get what you wanted. So it came up with an answer for something slightly different and you're like, okay, I have to redo my previous prompt and then edit the prompt a little bit and fire off a new question. But that's in the beginning how it worked. And it was great for the time, but at some point we realized like, yeah, we need to have follow ups in there. And that was a whole exploration in itself as well. But that's when I had joined, so that's one evolution that I contributed.
- [53:51]
Peter
Follow ups was interesting. It sounds quite simple, but the UI we ended up on was actually quite hard to get to. Right. We had various explorations to make it kind of work with various different visual styles.
- [54:04]
Roy
Yeah. Because we're constrained to the small window. Right. And there's also a bit of a system prompt going on to keep the answer short for quick AI. So if you're in a chat it's slightly different but like basically if you're asking something, you don't want it to just have a very lengthy answer and then have the user interface where you have to like scroll through, you don't want want that. So we intentionally prompt the AI to give very short answers. But with follow ups it's a similar thing. Right. You're going to have a scroll list, but at the same time every answer in a way, the way Raycast works is kind of a selectable thing and you want to do something with it, you want to copy it, you want to paste it to your frontmost app or do something with a follow up on it. So there was these two things where it was a very long scroll view. It could be right, because you just keep on following up, which you have to be able to just scroll through, select through. But there's, there's on the one hand scroll and on the other hand selection. So UX wise, it was quite tricky to nail this.
- [54:55]
Rid
I love like those types of micro decisions, like even just a theme of like all these episodes, like I love getting into that level of detail because what we see, it does feel obvious. It's like, well of course they designed that, you know, like, how else would you design it? Like they probably just sat down and made that. They might not even had to design it. You just build it because it's obvious, you know, but it's like, well, when you start really teasing it apart, it's like there's a lot of different paths you can go and trade offs, you can take. And so maybe I want to take that idea Zoom all the way out. Because now, you know, back in whatever it was, end of 2022, you didn't feel like Raycast was an AI company at this point. It's very obviously like this is AI is at the heart of a lot of the vision moving forward, I would imagine. And you're playing this game that a lot of younger companies are where it's like, like, okay, we're building for that future. The future's moving so quickly, we maybe have an idea of where it's going, but also like, who the heck knows, six months is eternity right now. And so how do you make sense of this delta where you understand that AI is going to be a part of everything, it's going to expand in every direction, but you also have to think about like these individual building blocks in the short run where AI already is touching almost everything. So like what is the system, systems design mindset that you have to have when designing these features to make sure that you're not creating this series of one way doors for yourself?
- [56:24]
Peter
The struggle that we always felt, and we had a bit of like a slow summer last year where we were a bit in a limbo was the industry moves incredibly fast, right? It's like every week there's a new model, every month there is like a new trend, whatever, right? So it's really hard to know where you're going for. We felt like we were sort of in a trap where we tried to do something and then we hit sort of a roadblock where we felt, okay, this is maybe no longer the most important stuff we can do and the most unique we could solve it. And so we switched then over, which ultimately became AI extensions. But it's because everything is moving so fast, you oftentimes need to evaluate are you spending a lot of time on this one individual thing? Because by the moment you may be finishing, your solution is outdated or maybe not even relevant anymore. And so I think it's really hard to say, say this is the glory future you want to be. You maybe have a bit of a direction, but you probably need to live much more in the here and now and sort of do the obvious things that you feel like progressing forward and then kind of hope that this like compounds over time where you do like a lot of those. And then with those things you're basically building up momentum essentially. And that was one of the things which we, since the beginning of the year essentially focused on the first half of 2024. If you look at the changelog, we shipped a lot of stuff and then the second half, we suddenly felt in AI space in Raycoast, we just slowed down. We couldn't pin it down like one specific thing. And so when we came out this year, we said like, hey, we want to basically concentrate shipping more things more quickly and the more obvious stuff. And I think like since the beginning of the year, we shipped faster than we sometimes wanted. And it became a bit overwhelming almost because there was so much stuff that we put out where we now again at sort of recalibrating and saying, okay, now we went through like half a year of like shipping really rapidly. Where are we now? And we're sort of lifting our head and looking around and thinking, okay, is this the right direction? Do we want to do more of that? Which we're now doing a bit more. Yeah. Taking a step back and thinking about, okay, where are we in the moment of time now? Where should we go next?
- [58:27]
Roy
Yeah, and I feel a constant balance that we need to strike because as you say, and as everyone says, like, it's moving at such a fast pace, right? Like new models every week, new capabilities every week. And to some extent when something new comes out, it's very obvious that this, that's something we need to support. Right. Like early on, at some point these models got vision capabilities. Obviously you want to be able to upload an image, right? Like that's something we need to support. But with some other things, it's not as obvious and it's just also so much. And it's like we shouldn't just do all these things. Like if we do all the things that are hot in the moment, then we have no time left to just focus on what we're good at and how we can incorporate AI in raykast in the way is unique and is actually benefiting the users with the way we integrate with macOS.
- [59:14]
Thomas
Right.
- [59:14]
Roy
We allow existing extensions to create tools for AI within Raycast. So LINEAR was a good example. Like basically all the commands that you have in Linear that you can sort of manually trigger from Raycast AI, then also got access to all the tools that the linear extension got. And MCP is somewhat of a. Yeah, it was released and we were a bit confused. Like, is this now a new sort of standard that we need to also suddenly do, or are we going to continue with AI extensions? And we continue with AI extensions And I think it's quite a core building block of something we can incorporate later on with. Don't want to tease too much, but like all the smart stuff that we can do just the fact that there's now a building block that is creating so many interesting opportunities, basically.
- [59:57]
Rid
I might make you tease a little bit, but let's go back because it does feel like, yeah, AI extensions was a pretty significant breakthrough. You talked about this change in Velocity. Was it an unlock in just packaging strategy or were there specific things that you were pivoting away from? Like what? I kind of just want to take a magnifying glass on that moment. Like, what actually happened that was a breakthrough for Raycast.
- [60:20]
Peter
Usually when you think about things, you attach. You attach a file or like you attach an image, right? That's sort of the two or based things. But we felt like in. And nowadays you're not just working with files anymore, you're working with notion pages or with browser tabs or whatever, right? So we felt like, okay, that can be like a noise continuation. And so we then basically want to go down this path of allowing people to attach pretty much everything to a chat. But then we felt at some point like, hey, look, actually what probably is a bit more future looking is instead of like you attaching everything yourself, maybe AI can figure certain things out, right. And can read and more of those agendic systems that became popular in tools like Cursor and so on. And because we're like a quite small team, we sort of had to make a decision, are we going to go down letting it attach everything or are we going to go down and hopping on this enchanting path? And so we chose the ladder where we felt like there is for us with our audience, probably more to gain, and basically jumped in on that and said, okay, let's go with AI extensions, where we feel like that can be quite interesting, where suddenly a computer starts doing things for you. You're kind of like telling what do, and then the computer figures it out how to do, which is, I feel like always what everybody dreams of, right? In a way, we translating. We do, yeah. The translation in our brains of like, this is the intent I have. And then you kind of need to put it into clicks and keyboard strokes to achieve it. But then we can integrate with a lot of things ourselves and can we basically use that, like to give AI superpowers? And it really came back from. We sat down and felt like, okay, there are a few paths we can take here. One, we can always black catch up to all the other apps that are out there and maybe setting certain things up, like, oh, if a new model comes out, you need to support image uploads and stuff, right? But you're in this constant catch up mode which we quite frankly got a bit tired of because it's not fun, right? You're basically always behind, essentially that's the whole way is the thing. And we saw like how can we turn that around and like basically lead. And we oftentimes said innovate, right? How can we be ahead of others? And then we just looked at like what is the unique stuff that we have that others don't have? And we always came back to extensions. We have like over 2000 extensions in our store, many more private ones, but they were very sort of Raycast 1.0, Raycast classic Raycast, the old version, a human uses it, right? They were not ready for AI, but we felt like that's where the puck is going eventually it should be where things go. And we felt like that's sort of our uniqueness. We have this extension platform that's what tens of thousands, hundreds of thousands of people use every day to get their stuff done. Right. And so how can we leverage that with AI and combine that together? And that was the initial idea. How we got then where we are at the end of the day is a whole different story.
- [63:09]
Rid
It makes sense. I think that's something that a lot of people are wrestling with. And so, so much of product strategy when I talk to these AI startups is like what's the foundational system that we can lay in place that allows us to kind of be nimble and be able to move quickly in a world where we don't actually really know like where the breakthroughs are going to be coming. You just have to effectively position yourself to take advantage of those. And so working backwards from like where do we have a competitive advantage here? And I love the idea of flipping where like where can we be the leader versus constantly playing catch up. Because I do think we're seeing more and more like what felt like products are actually features and everybody's racing towards this commodification of the base layer of functionality.
- [63:52]
Peter
It's like, ah, yeah, that's definitely across the industry I feel. I mean one thing which we all need to realize, we all cook with the same stuff, right? Like if you look at the big model apps, yes, they do the models but then when you have access to the models then it's equal playing field, right. So there is no like massive advantage that they have there, right. That's why some companies can go really fast there and can build really cool products. And so it's really, if you think about it, it's a great way at the moment to be a product builder because it's such an open space, it's a level playing field, I would say. And so it really comes down, what can you do that is more unique to other things? Right. And that's exciting for somebody who want to build product. I think there almost can't be a better time. We had this a few times before when mobile came around. New thing, new paradigm, lots of stuff need to be figured out out. And then I think nowadays it's probably much faster the progress that is there on the things. But I find it fascinating because you need to find your own strengths and execute on those instead of like following others.
- [64:51]
Rid
It's a great day to be a product builder. It's an even better day to be a product builder at the pseudo OS level.
- [64:57]
Peter
Yeah, that's for sure.
- [64:58]
Rid
Yeah, you can kind of see where it's like, all right, AI, extensions, everything, right now, you know, you're the launcher, right. Like everything's user initiated and maybe this does create the pathway for things to truly happen a little bit more agentically, hopefully. I'm sure that's some of the things you're wrestling with right now.
- [65:15]
Peter
Yeah, I think the stuff that we're wrestling right now with is making it very intuitive to use because let's face it, like AI is here. Right. Tons of people using it, but it's still early days and it's very much early adopters still to step take the most advantage out of. Right. And so it's a new tool, but you need to learn how to use this tool.
- [65:39]
Roy
It's a lot of fiddling and trial and error and yeah, the fact that it's non deterministic. Right. And if you add tools on top of that, with the way we have AI extensions, it's like really, I wouldn't say super small niche, but you have to like get into it. Whereas turning it into a really valuable, impactful product, it just needs to be so much more accessible and, and in a way reliable, which is something we're working towards.
- [66:03]
Rid
Yeah, let's drill into that piece then. Obviously there's so much strategy, top level thinking, but at the end of the day, you do have to boil it down into this intuitive interface. What are some of even the patterns that you're thinking about or the challenges that you're having to work through as a designer to take this top level directive into something that people can use very fluidly every day.
- [66:24]
Roy
Yeah. So maybe AX Engines I think is a good example because always as a designer you work with the material that you have. Right. Like how does this, how do I work with AI? Well, you know, it's non deterministic. Right. And you know that it's. There's a certain lag I guess, or it takes a while before you get your answer back. Like all these things change a little bit. But those are kind of the key things that you need to keep in mind to design with AI. But very early on while we were exploring this, it was quite interesting because we like lots of sort of little privacy and control questions in a way came up. Like, do you really want to give AI access to your finder? Like is that something you want? Right? Like should we explore sort of different levels of confirmation? Like should you always. Like that was one of the key outcomes, I guess. Like, so the user wants to stay in control for any of the things that are destructive. Right. You don't want like playing a song and it's not the right song, it's not terrible. But deleting a file maybe instantly off your computer, like not moving it to trash, but deleting it instantly or renaming it, batch renaming hundreds of files in one go flow and instantly letting you lose trust in AI and using this tool. So that's something we definitely don't want. So we want to let the user stay in control over whatever the AI is going to do with the AI extensions. And it's quite intricate because we have access to quite a lot and you can give access to quite a lot of parts of your operating system. I really like the word pseudo asset you just mentioned. That's the perfect way for me to describe Raycast, to be honest. And never before.
- [67:52]
Peter
So I love that new tech loan.
- [67:55]
Roy
I'll take my 0.1% but like, you know, we have access to your find or we could have access to your finder if you give it. And the same with the calendar, same with a couple of other things that are quite, quite intricate in a way. So we decided on basically two levels in the end. Like either we just do it like it's an output, you get an answer from AI and it used some data from somewhere. But it's fine, it could trigger something. Like it could play a song in Spotify or create a playlist. In a way it's not terrible if it's going wrong. So we don't want to bother you all the time of like, are you really sure that you want to do this? This is what I'm planning to do and then giving permission. But for some other things, like the destructive Actions in Finder or moving things around, or maybe even compressing a file or something like that. It's like you get a command enter first to confirm that that's what you want to do. So you get a bit of a preview, which is a pattern that we introduced in the same constrained quick AI window. Right. It needed to work there, so it needed to be very straight to the point. Just a key value pair of like a title. And this is what I'm going to perform. Are you sure? Yes, let's go. And then it's going to do it and come back with an answer for you. But that was like one of the panels that we had to introduce to basically let the user stay in control and keep the trust in this system that somehow can come up with weird things that it's going to do.
- [69:18]
Rid
You talked about users staying in control, the separation of destructive assets. Are there any other guiding principles or things that you've even picked up? Just working with AI as a material, designing an AI product that maybe somebody who's not deep down this rabbit hole wouldn't be thinking of.
- [69:36]
Peter
If I think about something like Instagram, usually you maybe initially load something, but then it's kind of fine, and then you can doom scroll, but you never see a loading indicator again. Right? AI, because the latencies are so different, you often have to wait 1, 2, 3 seconds, or if it goes into syncing mode, it can go off for a minute. And so that's very different because we got used to basically instant networking calls and everything is super fast with AI, we're like, oh, this is slow. And so what we thought about a lot is perceived speed. You can't really make certain things faster, right? That's just what it is. But in design, you can make things faster. Right? Back in the days, I think there's always what Kevin Systrom of Instagram told a lot is like they made Instagram feel very fast because when you snap the photo and the moment you typed in your description, they already uploaded the photo. So when you then go to publish, it's already uploaded and it's like, wait, what? This was instant, right? So just by shifting a few things around, what we try to do is making this a very fluid experience. So it feels there's always something happening. We now have like a loading indicator. It's like a noise animation, which is, okay, cool. There's something happening when thinking happens, Like a model starts to reason for a bit longer. We show you, like, okay, there's thinking happening. You can follow that if you Want to read through it? Or you can keep it collapsed and see there is a spinner continuing. When an AI extensions perform something, we try to show you that the UI builds itself almost like there's a new list element coming in and then it goes back into a syncing mode and then there's another list element coming in. So it's sort of progressively builds up the UI so that you're not sitting there for, let's say, even if it's just two seconds, but up to like a few seconds and think, is this thing working? Did it fail? Did it become stale? So I think that was like one of the things. And then we have all the streaming text that before AI, like you never seen stuff that got streamed in, right? Like text, word after word. It wasn't really a thing. So I think there was a lot of that, which is quite a new.
- [71:39]
Roy
Pattern, actually, that hasn't exactly different from how Raygast usually functions. Right. Raygast is insert everywhere. Like we don't put any superfluous like animations anywhere because it just annoying. Like that's what we definitely shouldn't do. Like it should be snappy and instant. And then you have this thing that takes a couple seconds. Like, how do we deal with this? And what was interesting in me, this rotating sort of cursor indicator that is like a little diamond that sort of looks like our logo, but it's a bit more minimalist. That was one of the things we're quite late in the process came up with, because we used to just reuse the sort of the carrots that we had before that. But it was quite tricky to get it right because first of all, the small window. But we didn't want content to jump as much because it just started to distract. Right. So we have this in a way quite simple, a visual design system for this where each bar is the exact same height and the text that flows in. If you just look at it, some of these designs in isolation, you might feel like there's quite a bit of spacing between the prompt and the other thing. But if you're using it, you don't notice this because sometimes it's going to do a web search and that's going to just sort of flow in, in the same space as where the loading indicator was happening. And these are all things that we had to do because of. Of the nature of how AI works, right. It's going to think maybe after a little bit of maybe not even thinking, but it's going to take a little bit before it shows that it's going to do a web search, it could also instantly just spit out the text. It could also decide to indeed do a couple of steps before it's coming back with an answer and looking back at it. It looks so simple. But that took a lot of work to get right.
- [73:12]
Rid
I'm nodding along. I literally just experienced that same exact thing where I was like, oh man, this is jumping. Sometimes it doesn't jump, it just needs to be a fixed height and I would look at the sign, I'm like, I don't like that. I'm like, it' and there is like, there's like more intentionality that's having to exist in these like 2 to 3 second loading states where it was super easy to overlook them. Yeah, you can't really just dump a spinner out everywhere anymore.
- [73:34]
Roy
Going back to what we said about the Pixel Rumble situation that we had before, right. We discuss and discuss things in a very isolated way. Sometimes if you see a screenshot or snapshot of like a quick AI window with an answer where it's did a web search and something else and you prompted to give us feedback, that looks so overwhelming in the moment. But if you go through it gradually just by prompting and getting an answer, it's not at all. It feels very natural. But if you just look at that snapshot, you feel like, oh, designed this. Oh, I did.
- [74:05]
Rid
And again, it just goes back to like, you basically have to evaluate AI products in code. As simple as that.
- [74:10]
Peter
There's no other way because it's so unpredictable. Right. You never know what text you get back and so on. You can mock stuff up and get quite far. But then oftentimes, I guess you need to build it to really see when.
- [74:22]
Rid
It comes to context. Are you still having the user initiate that or are you doing anything behind the scenes to feed additional context to the memories?
- [74:31]
Peter
You're reading our mind. We were chatting literally about it a few hours ago.
- [74:36]
Rid
Okay. That was the question that we were talking. I was like, report history.
- [74:39]
Roy
And for the past months.
- [74:40]
Peter
For the past months. I mean, we definitely kept thinking about it. I think at the moment you need to be basically a wizard using all of this. Right? We talked about the building blocks. We have those. Right. But we feel like it could be much more intuitive and magical. If I would ask you, you probably can tell what you're going to do next in a way, right? Or even like if somebody sits next to you, you can probably tell that. And so there is a way where we think about, can we get to this stage where you can do much more things automatically, or at least not necessarily automatically, but suggesting. Right. There is something really interesting when you get the right suggestion at a time, all the time, and then becomes just that's your normal way to use it. Where with a lot of those AI features I think if you can make them really easy to discover, basically telling you, oh, this is the thing you want to do and you just press tab for example and it does it for you. That's really cool, right? I think that's sort of the UX that you see in some of the code editors that really became like the first thing that really made those code editors enchantic was the auto completion, right? It was writing more text for you. And it's like that's exactly what I wanted to write. And so this feels very magical. And at some point you've just tap, tap, tap and have the code edited. And can you do that on an operating system level? Right. That's kind of where our minds oftentimes goes. But there is a lot to explore to get there.
- [76:01]
Rid
You have endless opportunities and challenges.
- [76:06]
Roy
You need a super high signal, right? To be sure that the suggestions you going to show in this very limited space are going to be spot on. Like if you're going to show suggestions and like maybe half of the time, maybe even twice out of the 10 times it's going to be wrong, it just gets in the way then it just gets annoying. So there's a lot of context that we can work with. I think context gets very granular. Like if you look at context on just the shared context all of us have by spending some time together, that's one thing that you kind of want your AI assistant to know of, right? Like you're working on some project and you don't want to tell all the time that you work at some company and that this is your team and this is your code base and it should sort of know all of this. Right? But then you get much more granular in making it actually useful on the thing you're working on. You need to sort of bring in pieces of that context. And I think we're quite uniquely positioned to be able to do that. Like we established or build on this ention way of including AI extensions and more recently MCP servers as well. And that's kind of a pattern that we are thinking about. Like how could we use this more to allow you to bring in the right context at the right moment in addition to sort of the bulk context is what I Refer to it now as like everything it already knows. Right. And the combination of those two is, I think, quite key. But we need to explore a lot more in this direction.
- [77:26]
Peter
So, yeah, if you want to boil down Raykos, it's all about muscle memory, right? You have a Raycast user. Some of our users, when we sometimes have unfortunately a crash, they feel like they lost control of their Mac, right? Like, it's really this muscle memory. Like not having that, you feel like, oh, I lost it. And you don't even realize sometimes that you use it. It's so natural.
- [77:46]
Rid
Totally. That has happened to me once and I was like, I just, I was like, what's. Nothing's happening on my feet. What's happening?
- [77:52]
Roy
My keyboard's broken.
- [77:52]
Rid
Yeah, my keyboard is broken.
- [77:55]
Peter
Oftentimes where we can go through, can we distill it down to something so simple it becomes really quickly muscle memory, so you don't even think about it anymore. And breaking it down, down, like, what's this core thing that builds the muscle memory? Distilling something down to the essence is the hardest stuff, right? You need to like move stuff away, away, away until you find that thing and then you can like build it up again, essentially. And that takes time and effort.
- [78:20]
Rid
I'm a big believer in the power of video to explain my thinking as a designer. So when it's time to get feedback, I'll drop a loom link in Slack and another link to a Figma prototype and feedback will be scattered everywhere. And I mean, it's a mess. So I'm building the product that I've always wanted to exist, and it's called Inflight. You can kind of think of it like an Async crit. It's an easy way to share a video walkthrough along with an interactive prototype or whatever you're designing. And then AI interviews the people on your team to get you the feedback that you need and organizes everything for you in a beautiful insights page. So right now, now I'm only giving access to Dive Club listeners. So if you want to be one of the first to use Inflight, head to Dive Club slash Inflight to claim your spot. Next up, we're going to go behind the scenes to learn what it was like designing the new Raycast mobile app. So let's bring back Peter, one of the co founders from earlier, and also Jordan, who is leading the design.
- [79:17]
Thomas
To be honest, we never really thought too much about the before we released AI features. Like, it's just we were focused on Launcher and Launchers. Apart from Android don't exist much on mobile. But then suddenly people started asking for it. After we launched Raycast Pro, mostly for AI, but then once the notion of iOS kind of came around that there could be Raycast on mobile, people started also thinking, I would like to also have quick links and notes. But it was a big struggle to understand what would be the product because one worry we had is that if it is just AI, this alienates a big part of our audience. But we knew that it's also gonna be kind of one of the main features. So the first document was just a bunch of ideas, but nothing concrete. And it was for a long time after we had the team that we were just exploring. So when the team was there, we didn't really know what exactly it's going to be.
- [80:23]
Rid
So this is the point when you came in, right? So you're hired for mobile, you don't really know what this could be. It in theory, could be anything.
- [80:30]
Thomas
He had some ideas, like, I think he posted actually, like, what could be raycast on iOS, right? There was some. Some tweet, I think there was.
- [80:38]
Bruno
Yeah, I think I tweeted some stuff about like integrating with widgets and some other ideas. Like before I even interviewed. I think it was just before the. That as soon as I saw the job post in for like, we're hiring for iOS, I was like, oh, shit, here we go. This is a rabbit hole that I'm going to go down. But yeah, there was this one document that, you know, I join on day one and I read this, like, Raycast for iOS and I remember the one thing was like, it is not a launcher, there's no API, we can't do this. Everything else is kind of like an open book of like all these features that it could potentially do. And I think the first thing that I did was look at like, like, what are the limitations? What can we do at an OS level on iOS? Because it is completely different, it's pretty locked and usually you're contained to your app boundary itself. And I knew that like, that wouldn't cut it. Like having to always open an app from your current context just wouldn't work. So, yeah, I started and I created these. I remember there's like 12 different like touch points, so things like widgets, shortcuts, intents, keyboard activity, view. And I did all of that before we even thought about features. Because it's like no matter what we do, we have to be in all of these places. So if you just pick up your phone and it's locked. You should be able to access Raycast, or if you're in the Messages app, you should be able to access Raycast. Like, we had to be at that level for it to click and to be productive. So that was the start. And then from there, I think we pretty quickly started building and designing AI. Like, we need. Knew that was the one thing that I had to do. So I think about nine months of it was just looking at AI and starting to build around that and get some quick prototypes working.
- [82:14]
Thomas
Yeah. And while we were doing that, AI wasn't standing on the same at the same place. It was just moving really fast. So while we're trying to kind of catch up with what we thought previously was important, there's just more and more. And this kind of made it a challenge because we also didn't want to hire multiple people. It was a small team, one engineer, one designer. That it. Because we believe strongly in small teams. But getting this out took longer than maybe, you know, we would want to.
- [82:44]
Rid
Yeah, you're chasing a race.
- [82:45]
Thomas
Yeah, we're just ready. And also there is some sort of. We're not like, completely unknown company by this moment or so There is some hype, thanks to Pedro, there is more hype maybe than we needed. So we need to live up to expectations of what people have. And it's funny enough, people were asking for iOS without even understanding what exactly it is. Some people thought we're going to have a launcher, like replacing Spotlight, which is impossible. Some people wanted just AI, some people wanted something more.
- [83:14]
Rid
I can think of very specific Pedro tweets where he's like, what would you want the mobile app to be? I remember thinking like, man, they're having fun behind the scenes. There's so much on the table.
- [83:25]
Bruno
I would just like, search Twitter for, like, Raycast iOS and reply to people. I'd be like, what do you want? And a lot of people didn't know.
- [83:30]
Rid
They don't know.
- [83:31]
Bruno
Yeah, I scrolled one finger.
- [83:32]
Jordan
Yeah.
- [83:33]
Rid
They had no idea.
- [83:34]
Bruno
Exactly. That makes it harder, you know, to like, pinpoint because, I mean, Raycast is so broad. Right. And, you know, four years and of work and then you've got to try and catch up with that. It's. It's impossible. So trying to pinpoint those initial features and package it together took quite a while.
- [83:51]
Rid
I think that's why I'm so intrigued by your answer. Where, like the starting point for any mobile app people like. Yeah, well, I mean, I start thinking about the container of what a mobile app is, but you were very much so thinking holistically about all these different entry points and like having that be the start starting point. I've never heard someone give that type of answer before, but I guess it makes sense. When Raycast previously we were talking about it almost as like a pseudo OS level product and so maybe having that as the starting point for mobile. It's fascinating.
- [84:19]
Bruno
Yeah, yeah, exactly.
- [84:20]
Rid
Let's talk about the AI piece for a second. You're trying to catch up with the race car and nine months is an eternity in AI land right now. So were there specific points where you had to maybe, maybe pivot or change plans a little bit based off of how you were seeing the technology evolving?
- [84:39]
Thomas
Yeah, multiple times actually. That's why it took so much time. We started with simple, quick AI. We thought maybe we don't need chats, we don't need like chat history and so on. It felt like, oh, maybe this concept will go away soon. Maybe it's on mobile. It's not. You don't need to be in the long chats and have like long prompts months. We were wrong. And at some point we thought actually we maybe need to focus more on assistance part rather than just porting AI chat as this. And this was a fairly big detour and we want to still go in this direction. We built actually a lot of tool calling, like creating your reminders, working with your calendar and potentially even third party extensions. But this was basically cut away from the release two months before because it's just very hard problem. You don't want to do it in a kind of half working way. If it works with calendar, it should work pretty well. So we still want to do this but it took us a lot of time and it didn't even make the release.
- [85:50]
Rid
With other design challenges where you're thinking about what should even the interface of the app app be. You know, when everything's kind of on the table.
- [85:58]
Bruno
Yeah, yeah, definitely. I think visually that you think about like translucent windows and stuff like that. That doesn't translate as much. Right. Like you have to have a base, there's no wallpaper, there's so there's small details like that that we had to overcome. But it's also just like ux, it's keyboard first and stuff like that or like search first. And for a long time it was like people don't search as much on iOS at least from, you know, to launch something, you want a button on iOS right? To touch something. So yeah, there was a bunch of back and forth for a While we didn't have this like global universal search, it was only in the last two months, maybe two, three months, that we added that. For a while they were just buttons and yeah, that was a big challenge. Also navigation was huge. Because you think about them, they're individual apps, right? You have a notes app, you have an AI app and then quick link. So how can we make it possible to quickly jump between those? Also make a pattern that's scalable. Like the first release has four, but eventually you're going to have have 10, 12, however many apps. So that was a big challenge that we iterated on quite a few times. We met up in person to build a bunch of prototypes and yeah, eventually landed on where we are at today.
- [87:03]
Thomas
Then yeah, another big challenge was also context. It's much easier to grab context and work with it on macOS. You know, you can select text, open Raycast and do something with it. On iOS you don't have this concept.
- [87:15]
Rid
I would imagine, even like the homepage, like what's the thing that people see has all kinds of challenges too. When the web experience or the Mac experience is like deeply configurable. Right. Like, I enjoy Raycast because it's very easy to make it my own and surface the things that matter. And so you're designing a product that like I would imagine also has to be very moldable. Right. And it's like fighting all of these system level constraints. I was personally very curious where you were going to land for months, you know, because I understood like, I mean, there's a lot of challenges there from a design. Design standpoint.
- [87:47]
Bruno
Yeah, yeah, it was, it took a while to get to that home, that finished home screen. And I don't know how we like quantified that that was the right one, but it just like held it in my hand and it was just like that's, that's Raycast. Like it just felt right. And I think if you have that feeling, you know, that that's the option to go with. We tried a bunch of other stuff with like multitasking, so switching around or maybe a tab bar and stuff like that. But yeah, they were quite limited and restricted and not scalable. And I think, think be able to have this home screen that you can configure quite easily to suit your needs, but also be able to very quickly access, you know, your whole array of commands through the search. Yeah, it seemed to work quite well.
- [88:26]
Rid
Let's talk about the prototyping piece. Like the more that I talk to you all, the more that it feels very important that you have the real thing that you can kind of play with. So as a designer, how do you approach the practice of prototyping?
- [88:40]
Bruno
Yeah, I think we were moving at quite a pace where, like, in the end we weren't designing so much in figma, we were designing in code and just trying those things, especially those things like search or how you access and navigate. It just wasn't possible to just feel that in figma. So, yeah, in the end, I think in the second iteration we leaned way more heavily into SwiftUI, which allowed us to move at a much quicker pace and also allow myself to like jump in and also like craft some of those experiences. So.
- [89:12]
Thomas
Oh yeah, Jordan did insane amount of SwiftUI. I mean, he's coding now, basically. It's not even prototypes. He just go and changes things in the code base, which is actually huge. So if there are designers working on mobile apps that use SwiftUI, go learn it.
- [89:27]
Rid
Were you doing that in previous roles too?
- [89:29]
Bruno
No, this is. I hadn't touched SwiftUI before Raycast, really. So it's. Yeah, it was a complete learning curve and I think using AI really unblocked that because I could get responses a lot quicker instead of searching stack overflow and things like that. But stuff like the voice recording, be able to slide, like all of those gestures we built, I built straight in production without designing because it's already there and you know, there's no point adding additional steps when we can just, yeah, prompt and then tweak those things and then ship it straight away.
- [90:00]
Rid
Okay, then let's speak to someone very specific listening, who is inspired, has not coded ever. What's something that you've learned about using AI? To learn how to code with AI.
- [90:14]
Bruno
I think you have to be like, sort of specific about what you're working on. I don't think you can start with like, build me raycast on iOS or you know, build me this entire view. I think you have to break it down into like, chunks. So for me it's like describing like, okay, I actually want this to be a rubber band effect where like, it stretches as I'm pulling it and sort of slows down the momentum. And I think if you can describe those words and articulate it in a better way, it gets it sort of the AI and the model understands that and gives you a much better response.
- [90:45]
Thomas
Did you do any learning, like reading books or anything like that?
- [90:48]
Bruno
No, no, it was all done through AI and some like, websites like SwiftUI. There's a blog like the coding with Swiftui like, for some references, I actually wanted to understand a bit more, but most of it was, yeah, it was quicker to just ask AI And I think having a purpose and a reason, like, I wouldn't have the time to sit down and just build an app myself.
- [91:08]
Rid
Yeah.
- [91:08]
Bruno
So having the purpose of, okay, I want to convey this idea of how sliding, you know, a button works and how the gradient animates, but doing that in Figma and then trying to convert that over and. And all of those things, it was easier just to understand the limitations too. We are moving at a pace where everyone would be like, oh, well, you can just build down a metal shader, right, Johan? Like, go do that. It's like, okay, if I just know the constraints that I'm working with and I can sort of mess around with AI for a couple of hours, I know that that's a production code that we can ship and then we can move on because we had a million other things to work on at some point.
- [91:44]
Thomas
I got to a state when I was coding some features and I was like, okay, build the basics, Jordan. Pick it up and finish all the cool details so I can move on to the next kind of core thing.
- [91:56]
Bruno
Yeah, cool.
- [91:56]
Rid
A lot of work.
- [91:57]
Bruno
Yeah. And I think, you know, Johan and Peter were helping with, like, the backend stuff or, you know, getting those connections in and those functions in, and then building the UI on top of that is a much easier job because you're just piecing and tweaking, like, visual stuff there. No interactions. And. And yeah, it was a lot of fun.
- [92:15]
Rid
Let's talk about the tweaking visuals piece then, because you mentioned Pedro Hype Team high expectations. There's an element of, like, the broader high expectations, but then also, very specifically, the expectation for the craft of the product was very, very high because there's this baseline of what Raycast stands for in terms of quality that we've all kind of come to expect. So not only are you learning to code, you also know that you're going to ship to people who are going to inspect it with a very fine lens. So are there specific details or interactions or something that you were playing with where you really, really went deep that you're particularly proud of that we could talk about?
- [92:51]
Bruno
I think interactions that I'm proud of is like, that composer area. Like, yeah, the. The recording, the gradient, the, like, pull to create a new chat for me. Those are the ones that I worked on and I was super proud of. Did I go super deep on them? We made, like, a few different Recording gradients. But yeah, I guess if I look back a lot of it was you know, a day or two on each sort of area rather than, you know, finally, finally refining those. I guess it just like progressively happened over time over the weeks and months we were just, you know, maybe on a ray day I was just spending time tidying up the modals that we have. Cause we had like four different modals or things like that would just like you wouldn't micro tweak those things. I think the other gesture I quite liked is the pull to search that like Peter worked on. And it's think just getting the, the way the blur works and the curve that it comes in I think took.
- [93:40]
Thomas
A bit of time to be honest. Like in, in this last four months before the release we didn't have much time to go deep into some micro interactions. But quite a lot of was done in basically last one two weeks before the release as polishing. But we could have gone more if you know, we had more time.
- [93:59]
Rid
Your guys bar is high I'm just gonna say right now.
- [94:01]
Thomas
But like even now like I'm not happy with a lot of polish spots.
- [94:06]
Rid
But it's, that's called being a founder.
- [94:08]
Thomas
It's, it's. Yeah, but it's, you know, there are many ways to make it better. There's also was fair amount of frustration with some parts of SwiftUI. So there are some good sides of the fast prototyping. But Also this last 10% can cause a lot of pain to get right, especially if you have high standards or if you want to go a little bit more custom and some smaller bugs just was driving me mad. But you know, in, in a way it's just the first version, it's fairly good, but it's only going to get better from there.
- [94:42]
Rid
All right, well then before I let you go, I want to kind of get a little peek into the future because the more that I talk to you all the more I'm understanding the importance of systems thinking. Allowing you to scale, be flexible, move in different directions. So when you kind of imagine where this might go, what are some of the things that you're excited about about?
- [95:01]
Bruno
I think the next thing for me is like the custom keyboard. I think that's the unlock at a system level that gets us into, you know, WhatsApp, iMessage, Twitter. Like right now you have to use like the share sheet. It's a couple of taps, there's a slight friction there. And I think the custom keyboard feels like the closest like distance we can get from being at an os, like complete OS level.
- [95:24]
Thomas
And I think worth mentioning what custom keyboard in this sense is. Right? Like it's not actually the keyboard where you type. If you ever seen giphy custom keyboard, you just switch to it and you see a grid of GIFs. You can quickly attach what we can do. We can select the text and you run AI command with it, like fix spelling. So it's right in place. You can insert your snippet or share your quick link and bring more. You know, people keep asking for clipboard history. You can access clipboard history right from the.
- [95:59]
Rid
Oh, that's cool. I wasn't even thinking about instant keyboard as an unlock. That's huge.
- [96:04]
Thomas
And that happens right in the place where you communicating with other people or, you know, typing in browser or whatever. So this can bring us a little bit closer to what Raycast is on desktop, which is being omnipresent.
- [96:17]
Rid
Last but not least, we're going to talk about how the Hype team works at Raycast because they're doing things in a really unique way and it is definitely working. So let's take a second to meet the team.
- [96:28]
Pedro
So, hey, I'm Pedro and I'm basically just hyping up Raycast through videos and education stuff. And I'm living in Barcelona at the moment.
- [96:37]
Jordan
I am Inga and I just create like over complicated stuff on Figma and I am part of the brand and Hype team and a tiny bit of the product team. I'm just kind of like everywhere. Yeah. At Recast and I live in Northern Ireland.
- [96:52]
Rid
Let's jump right into the. The backstory of the mobile launch video. And, you know, it had a heck of a viral moment on Twitter, a lot of positive feedback. And I think it. It's so uniquely Ray cast and I don't even know if I'm able to articulate why. But when you guys go to market with things, there's just a freshness. It's a little bit different. And so, like, how does something like that come to be? Like, what's the backstory that led up to that? Really?
- [97:23]
Pedro
Yeah, man. I'm glad you start with that because that's the last projects that we worked on and that's like the most nervous of being about anything I've done. Yeah, it may seem from the outside that's like, oh, we know exactly what we're doing.
- [97:36]
Rid
It did come across that way.
- [97:37]
Pedro
You know, we're also just figuring stuff out. And so the way it came about is that maybe four weeks before we launched we got together in Paris. Inge, myself, and Bruno, who is also part of the hive team. He's not here with us at the moment, so. Shout out, Bruno. We were together in Paris just catching up, and we knew that the iOS launch was coming up, so we wanted to do something a little bit special for it, rather than just a video in my office, which is like, if you follow Raycast and our videos, then most of them is just me, my office, right? And we're like, you know, this is a new platform that we support, and it's the first time the Raycast is on the go. You know, let's try and do something a little bit special. And. And I don't want to make a video About Raycast on iOS in my office when clearly I've got my laptop and my desktop here, you know? So, anyway, we decided to make an ad for it. So we spent a day or two in Paris just brainstorming what this ad is about. Right. And Inga did some storyboarding. We came up with a bunch of ideas. Right.
- [98:42]
Rid
How wide you go? How wide was that?
- [98:44]
Pedro
We went wide.
- [98:44]
Jordan
Yeah, we have very wide. Like this. There is, like, throwing things outside, you know, like the idea, which I think.
- [98:51]
Pedro
It'S the most important thing at that stage of brainstorming. Right. Before that, though, we had already done a little discovery phase with the iOS team. So the discovery phase, like, think about the Hype team as like, an internal mini ad agency. I've worked in advertising for a few years. Like, you know, the process is somewhat similar in the sense that you get a brief, you do discovery, exploration, execution. Right. So we did that internally. We talked to the team, and we identified some of the takeaways from this launch. Things like, who is it for? This is not a release for, like, someone who's never heard of Raycast.
- [99:32]
Jordan
Yeah.
- [99:32]
Pedro
This is a release for people who are either on the iOS wait list, and we had, like, thousands of people there, so they're already waiting for it, or people who are already using Raycast on Mac and also want to use it on iOS. Right. So it's not like we're not Apple. We're not doing a release for, like, billions of people. It's like, it's kind of a niche thing. We're basically just communicating to people. Raycast is now on iOS. Right.
- [100:01]
Rid
So slowly, too, you're communicating to people like me who are willing to sit through and watch, like, a minute of you just going back and forth with nothing.
- [100:09]
Pedro
It was a minute and forty seconds.
- [100:11]
Roy
Right.
- [100:12]
Pedro
And so, given that we were like, okay, let's be creative, right? People want this. We don't have to sell anything here. This is not an ad where we need to go deep into the product values. This is just like Raycast is now on iOS. But we want to do this in the most beautiful way. Like, beauty is the thing in this ads, Right? That's what we came up with.
- [100:40]
Jordan
Yeah, because, like, it kind of like, betrays quality. Like, if you think about it, like, there's time and effort being put into this thing. How we kind of treat like our video stuff, because we're kind of just like, starting and, like, figuring it out is like, we want to almost treat like our product like a production studio. We're not just. Every video we release isn't the same, because why should it be? Like, every feature is different, so why can't our style just change up Every.
- [101:04]
Pedro
Night, Inga, myself, Bruno, we brainstorm a bunch of ideas. We went like, we had some crazy ideas, dude. They're like, you know, that's. That's what allows you to then come back down to reality and, like, stay within budget and time and stuff. You know, it has to be aesthetically pleasing. It has to be beautiful. But the message, the underlying message is, pack your Mac. Raycast is now on the go, right? That's the message. I think we can do something that talks about this and tells that story. The story being, here's a beautiful setup with, like, the best things you can own, like the XDR display, you know, and we're removing it. You know, like, who in their right mind would remove an XDR Pro display from their desk? Nobody. Everybody wants that display, right? So we're removing it, and we're going to remove item by item, and we're going to take it away. And people don't know where that item is going. It's just going out of the frame. And then eventually we review that it's going in a box. And then the last shot, which the pack shot is just me reaching out for the phone when the app shows up. And then we leave, you know, so it's like, it's very, very simple, you know, but what makes it beautiful in. In our opinion, obviously we're biased, is that it's very slow. It tells a kind of a little story about productive on your phone to, you know, you can carry on from your desktop onto your phone. At the moment, the things that you can carry on is AI chats, snippets, quick links and notes. There will be more. And so we started working on It Inga was doing some storyboarding for us on Figma. Like, you should show the illustrations you did was insane. Like the most beautiful storyboard ever.
- [102:50]
Jordan
Yeah, it was when I had access to Figma draw as well. So I was using all like the blur. Yeah, I was using and like the progressive blur. Whenever, like, you didn't even know I was.
- [102:59]
Pedro
No, I didn't know that.
- [103:00]
Jordan
So I was like testing that.
- [103:01]
Pedro
Oh, wow. You got early access. Nice. And then Bruno was heavily involved in, you know, the script and he flew over from Brazil to Barcelona where we shot it. I hired a production agency there. We got an amazing director and DP to like work on it. We had the music composed for it.
- [103:22]
Rid
Really?
- [103:23]
Pedro
Yeah, yeah, yeah.
- [103:24]
Rid
That's so cool.
- [103:24]
Pedro
I started saying like, that was the most nervous I've been because it gets bigger every time. Like. Yeah, it gets bigger every time. And everything that could have gone wrong went wrong. Like, we worked like non stop for the last maybe 10 days on this.
- [103:39]
Jordan
You got sick.
- [103:40]
Pedro
Like, I got super sick. There was a blackout in Barcelona. We had like three days to do the whole post because we decided to shoot this on film. 35 mil film. And so to shoot an add on film is a whole different experience. You've got limited amount of time takes. You've then got to fly over to London. We had a career, like a physical person coming to Barcelona to take the film and take it to London to. I think it was called Cine Lab. To get it developed and scanned and then it's sent to Barcelona. The scan so then we can edit it was like. It was crazy, man.
- [104:14]
Rid
And it's so beautifully over the top. Like even just getting to know you over the last like 24 hours, I'm like, this doesn't surprise me at all.
- [104:20]
Pedro
Yeah, it was over the top. But I like, I feel like the takeaway also is that if something's so beautiful, it's so high quality. What we're also saying without saying it, is that Raycast is high quality, you.
- [104:35]
Rid
Know, and it goes back to the thing you were saying earlier, like you didn't want to be selling or like promoting features, which is totally. It was nothing but quality. And it spoke directly to the people who had that level of context where like I hit play on the video and like every 10 seconds my smile was just growing a little bit more. I was like, okay, this is pretty awesome.
- [104:54]
Pedro
Yeah, yeah. And it's funny because, like when we got the first cut, the video was like 35 seconds. And then it was like, it was so Fast. And I was like, no, man, that's like.
- [105:05]
Jordan
That was our original idea.
- [105:06]
Rid
Yeah. The slow burn is what does it. Yeah, it's like that feeling, you get it 50 seconds where you're like, oh, my gosh, this is actually what the.
- [105:12]
Bruno
Whole ad is going to be.
- [105:14]
Pedro
But then, like, turns out, like this, editing companies and stuff, the production, they're very used to, like, ads having to be within 45 seconds. You know, usually, like, you got to keep it short. We're like, no, no, this is just going on YouTube, you know, like, we can do whatever we want.
- [105:28]
Rid
All right, so let's talk about Hype Team. Now, you've used this word a few times. I'm sure that somebody is listening. And they're like, okay, well, they've just rebranded marketing as hype, you know, so, like, what is a Hype team? What is, what are you trying to accomplish? What are you doing that's unique in this role? When you think about how to grow and also, like, build a community around. Right.
- [105:48]
Pedro
Podcast, it's not truly just marketing. Right. Hype Team incorporates a few different disciplines. Let's say a little bit of marketing. It's a little bit of branding, brand awareness. It's a little bit of, like, education, like, you know, teaching people about the product. It's a little bit of content creation, like taking photos, making videos on YouTube. Like, some videos are not educational. They're just, I don't know, social and entertaining type of videos. It's also about community. Like, last night we hosted the Figma configs after party. Like, 700 people signed up. There was a turnout of like, at some point, like over 250 people, you know.
- [106:25]
Jordan
Yeah, it's not like a harsh label. Like, I feel like whenever you're like in a marketing team, it's just such like, you're conformed into that marketing team. We're like everything. And we kind of just make it up as we go. Like, if we want to do more content, we're going to focus more on that, or if we want to do more events, we're going to focus more on that. But we're very fluid. Like, right now we're doing like, a lot of events, like, more events than normal. And like, we're loving that and we're really enjoying that. So we're going to focus more time on that.
- [106:51]
Rid
It makes sense even just given the nature of the product too, you know, like, yeah, there are a series of user journeys, but you also have this really kind of long user life cycle in many ways. Where it's like there's so much to learn, you know, like you're, you can't actually use Raycast in your first session, you know, like you kind of have to tease people along and give them little snippets and teach them. So it's interesting to hear like the role that education even plays in, in the idea of hype and what that entails.
- [107:19]
Pedro
Yeah, I think everything's hype. And I'm not saying like we coined the term and like we invented it, but like, I've not really seen a lot of companies doing that the way we are. I think it's very easy to overhype something. You know, I think that's like actually not going to say any names, but like that's what we normally see and that's a problem. And that gives the word hype a bit, a little bit of a negative connotation conversation because it's like, oh, you're just overhyping it. Right. But a lot of what we do is like figuring out when to under hype something. A lot of our videos, we're actually under hyping it because we want to manage expectations. We want people to feel excited, but when they finally get this thing, it meets their expectations. This is so closely related to like brand awareness and how your brand is perceived. Like, that's why I said this is also a bit of a branding team. Like Inga is obviously our brand designer, but brand is not just about design, it's about everything else. Right. We use Raycast as an asset to promote Raycast. It's not an exciting looking tool. It's very functional, it's productive. We need to build a brand that is so powerful that people want to be part of it, you know, and that comes from the human emotional side of things. And then we hope, and this is proving to work anyway, that then people want to use the product too. So it's like lifestyle first emotion. We try to tap into the emotional side of things. Like the ad we made for raycast Focus was 100 psychological and emotional, right? That feeling of chaos, stress, distraction that comes from everywhere right now in our lives. Like my phone is constantly on silent and focus mode because otherwise it's impossible. Right? And then you've got all of these apps and social media, all these algorithms fighting for your attention and you're just overwhelmed. And every time you want to get into focus mode, something pulls you out of it, you know, like this is the feeling we're evoking. We're trying to like get people to relate to. And then at some point in the ad, everything changes. The light changes, the music changes and you just see a typewriter again. It's like, it's just a metaphor. Right? You're not going to switch to a typewriter. You know what I mean? But the typewriter is a tool that can only do one, one thing. And that's why we replace the laptop with it, is. That's what I think. It's part of the brand. You're planting these little seeds, you're telling them about these messages in ways it's not so in their face and so salesy, you know.
- [109:52]
Jordan
Yeah, but like we're building it like this thing is like, I don't know, it's like growing with us. Like if you look at our brand like a year ago, it's completely different. Well, to us it's completely different. It's because, like, I don't know, I always think whenever I see like the brand and it's. Or like the visual identity of Raycast, it's not just like one voice, it's multiple. It's like, like a large part of it is you, a large part of it is Nicholas. A large part of is Thomas and Peter. There's elements of me in there and this kind of adds like a little bit more soul to it. I feel like whenever people are like, okay, we're going to make an ad and this ad has to get this many sales or this many whatever. We don't think really about that as much. We're like, what do, what do we want? What's our tone? What do we want people to feel? We've spent months crafting this, this thing. We're excited for people to use it. We want people to feel that same feeling that we're feeling.
- [110:41]
Rid
As you were talking, it was really dawning on me. Like, you're right. Like it's. There's not a lot of UI assets that you can use to then express the brand visually. Which gives me an even deeper appreciation for some of the things that you've been able to do because you have these time lapse videos that, I mean, even for people that are listening that still somehow don't use Raycast, like, you've probably seen these time lapse videos, you know, and there's like an affinity that's established with the pure visual design and to be able to do that without real UI is. I now have an even deeper appreciation for some of the things that you're doing on Twitter.
- [111:19]
Jordan
Yeah, but people are like, why are you doing that, like, Just, like. Just, like, blander.
- [111:24]
Rid
That's, like, some of the beauty of it, though, too, because it's like, it's not about a checkbox. It's like, this is me.
- [111:31]
Jordan
There is a lot of joy. And that, like, actually, I. To. I've been illustrating like, that for years. I never really thought much about it, but, like, at Raycast, like, what is it? The Windows wait list? Media. Like, we were doing, like, these posters and stuff, and I thought, okay, I'm going to share internally, like, a visual, like, or a screen recording just to show the guys what I'm actually doing. Because I don't think anyone knew what I was doing. You just thought, ingo, why are you taking so long to, like. Like, draw up this thing when I show it? I realized that wasn't normal at all, and then it just blew up. And I think it does give people, like, a realization. Like, yeah, there is a lot more effort, like, put into these things. Like, a lot of people, like, oh, I thought it was a 3D render. I thought it was. Because if it was a 3D render, yeah, it probably only take 20 minutes to put something together. But, like, there's thought behind it. Like, while I'm drawing these, like, illustrations, I'm thinking about the rest of the design. Fine. Like, that gives me kind of space to, like, think about how everything's gonna go. Like, so I'm not just completely zoned out. Like, tapping about my trackpad. Like, I don't know, it gives you a little bit of a breather.
- [112:35]
Pedro
She does that on the trackpad on a laptop. I'm like, there's no stylus. There's no, like, 27 display.
- [112:41]
Jordan
And I don't. I don't use. I. I'm just Mac. I. I can't. I don't even work at a desk. I can't do it. Like, like, what is it?
- [112:49]
Rid
I believe you.
- [112:50]
Jordan
We're doing. I know we're doing, like, design. We're, like, sitting at a desk just talking about, like, cool things we're going to do. And I'm just like, I need a. So I need my legs crossed. I need to, like, zone in, lock in, and not get distracted. Like, if there's, like, anything there, I'm going to fiddle with it.
- [113:06]
Rid
Yeah. On this spectrum of, like, this is the Raycast visual identity. And I know you don't have, like, these rigid brand guidelines, but if that's one end and the other end is like, I'm literally just going to sit down, open up, start fiddling with my trackpad and see what comes out. Like, how do you think about the ways that you want to express what Raycast takes on in terms of like a visual identity? How do you even figure out what that can be given, like that next opportunity?
- [113:32]
Jordan
We are just figuring out. I don't even think we know at this minute. Like, we have an idea. Like, especially me and Pedro are very aligned on like, where we want to go with like our visual identity. Yes. I have all these illustrations that I'm doing and I'm getting more known for them now, but I don't think that is. Is like just us. It's just an element. This is something we're exploring. This is something that's like really cool for us to do. It's like these long crafted things. It shows again, it's all reinforcing that our product is a high quality product without saying it because, like, we don't want to ever be like, quality, nice. I don't know what the words are.
- [114:07]
Pedro
Like, nice.
- [114:08]
Rid
There's not a lot of thought leadership coming from.
- [114:12]
Jordan
Like, it's just more. More like this is a really cool way, an interesting way of doing things.
- [114:16]
Peter
Yeah.
- [114:16]
Jordan
And this person has put a lot of effort and thought into this and that's cool.
- [114:21]
Rid
It's so clearly an expression of the people too.
- [114:24]
Jordan
Yeah.
- [114:24]
Rid
And I, I'm even going to say that because I was able to, you know, have lunch and hang out with you all leading up to this and you all pulled out. Everyone's carrying their own camera. Right. Like, everyone has this aesthetic angle and this like, artsiness almost in like different ways, you know. And so. So I'm now able to see kind of behind the scenes where it's like, wow. Everything that Raycast puts out there is really coming from like, this is genuinely who you are as people.
- [114:54]
Jordan
We all have similar ideas of things and we're very aligned, but we're all very different people. Like, I have a farm. Well, attempting like, I have donkeys and like, I'm chaotic and I don't have a desk. Well, you have like this beautiful setup and you're like a lot more like cleaner and like. And like. And then Bruno is like, he's more into like artsy things but like more organized as well. I don't know. We're all just very different. So.
- [115:21]
Pedro
But you know, it's funny that, like, your illustrations are very analog in a way.
- [115:27]
Jordan
Yeah.
- [115:28]
Pedro
Right. Like, you are replicating color and reflection and lighting and shadows.
- [115:33]
Jordan
Yeah.
- [115:34]
Pedro
By drawing. You're not relying on like 3D modeling, like any of these things. It's a very slow way to do it, although you're very fast. But like the process of drawing something from scratch is like slow. That's something we use a lot in our brand. These illustrations from Inga. What we also do a lot is photography. We start using a lot of photography in our branding because we feel like photography is a great way to like get closer to that emotional and human side of things.
- [116:02]
Rid
Things.
- [116:03]
Pedro
All the photography that we use pretty much is also analog photography. Right. We shot the ad in analog, but the photos are also analog events. It's also the slowest way to take photos. Like you take a photo now, you got to wait a week to get them.
- [116:16]
Rid
You know, it's a fun juxtaposition for a company that's very AI.
- [116:19]
Pedro
That's exactly it. Like Raycast is about being fast. It's about productivity, it's about taking shortcuts. You know, that's again the 180. Right. But everything else we do to promote that is in like the whole opposite.
- [116:31]
Jordan
Yeah, but at the same time, like because you're taking it on this, you're not super focused on this. At the same time, like, like if we're at like an off site or whatever and you're like taking pictures of people or whatever, you're not focused looking at the images in the bus or on the. But you're talking to people. Like you're. I don't know, it's kind of like there's a separation. You're still like there is productivity in a way.
- [116:52]
Pedro
It's not taking over your whole life without saying anything. We are saying things, right. It's the same when we did a giveaway and like we decided to hand print some photos. Photos for, for this giveaway, right? And we had this idea, I think we're talking about it in one of our one to ones. We have this photo, this one we did break as focus, right. We had this photo that I took of a girl lying on the bench in front of the beach and she was just sort of lying down and she was like looking at her phone, you know, like. But super focused and stuff and there was like nothing else around. The photo was a little bit like skewed, was not super straight. You know, like normally in photography, you know, you, you hear that people should sit, say like your horizons need to be super straight. That was like super like diagonal and, and I, I always love that photo. And you too. I love the photo. We're like, oh, why don't we do that? As a giveaway for the focus thing, we do the cap and then we do the photo. I was like, yeah, that sounds good. I'm gonna hand print them, you know, and then I just went to the dark room and I hand printed the negative. You know, like, zero reasons to do that. Could I just printed on an inkjet printer, you know, if you get that postcards that was hand printed, Every single one is unique. Yeah, every single one is a little bit different.
- [118:05]
Jordan
Not mass produced. It's not like, oh, you've just sent it to like a printing company and it's done. And then, yeah, like, there's thought behind every single one. Right when I tell every.
- [118:14]
Pedro
And then the postcards had a little text on it that said, like, less chaos, more.
- [118:18]
Jordan
Yeah.
- [118:19]
Pedro
And even that was like analogically done, you know, using like a film. Like, it is not, like, done in Photoshop. So this is like some of the behind the scenes of, like, what we do here. This is the first time I'm working, working way more on the marketing, brand awareness, creative side. I've always been a developer. That's what most people know me for. Building Radix and Stitches, working at modules, you know. And now I can use all of that knowledge that I've accumulated over the years of being a developer, of talking to developers, of like, breaking complex topics down into simple ways, which is what I do with Raycast, but now use more my creative side of things and try to get the things that I like in my real life and use that to promote Raycast.
- [119:00]
Jordan
It's like having all those understandings. It's like I've been a product designer for years, yet I never say I'm really a product designer because I'm also an illustrator and I'm into brand. I'm into, like, everything. And no matter what company I worked at, I always kind of floated or I was this, like, bridge across different teams and I was like, oh, there's something wrong with me. I'm not a good enough product designer, or I'm not a good enough illustrator because I'm not focusing all my time in illustration, or I'm not good enough this. And I. And it's like, but that's the perfect person to be. Because you're multiple things, you're multifaceted so you can have more understanding. It's like whenever I'm designing, like a visual whatever, I have the understanding of the UI while I'm also creating something fun.
- [119:39]
Rid
And I'm sure it's like kind of how this little team of three has been able to have a very significant impact. You are kind of able to straddle a lot of different disciplines. So, yeah, it's been very interesting to see like the behind the scenes how you all work and even just the core, like, you, you've spoken not just to the identity. I knew we were going to talk about, like visual ident identity stuff, but this really actually is like the Ray cast identity. So I appreciate you pulling back the curtain a little bit.
- [120:04]
Pedro
Thank you, man. And just one, one thing to add is, like, this only works for anybody listening that they might be interested in, like doing something similar. I get that asked all the time. This needs to come top down, you know, like, the only reason this can happen because Peter and Thomas just trust. They just trust so much trust. They just like, just do whatever you want.
- [120:26]
Jordan
Yeah. Like, honest to God, like whenever. Like, even with brand stuff, I'm like, are you sure like this?
- [120:33]
Pedro
Because even I think it's crazy.
- [120:35]
Jordan
Like, and then I'm like, oh. And they're like, yeah, you go for it. But it's so rare like that. That's not the norm in a lot of companies. And like, like, I was even chatting to like, Peter and Thomas. I was just like, this is such. And this is like, you might hear such a good company, but like, it actually is because you can actually just do whatever the hell you want. You could come to me and be like, oh, I'm going to do this. Or I could come to you. I'm like, oh, I' to make like a coffee mug. Oh, I want to make this, I want to do this. And you're like, yeah, you go for it. You, you lead, you, you do whatever you want, like, as long as it makes kind of sense.
- [121:07]
Pedro
Yeah, that's good actually because, like, as you, you, you know, have talked to a lot of other product people here, right. And I don't know if you've heard it yet, but, like, there's a lot of autonomy in Raycast. People work on their stuff and they own it.
- [121:19]
Jordan
Yeah.
- [121:20]
Pedro
And within the Hype team is kind of similar.
- [121:21]
Peter
Right.
- [121:22]
Pedro
Like, we follow that because that's what culture is about. The culture needs to be defined top down and people going to look up and follow that. So I've worked in so many companies, man. Like, the idea that you feel trusted and you're given enough freedom to explore things helps us make things that maybe slightly unconventional, you know. And I'm not saying like, oh, wow, Raycast is like the coolest company. But no, there's a lot of Companies that have done similar things that also make beautiful stuff. But, but we are trying to just take what we like and promote the brand. Because if I didn't use Raycast, what would I like to see in order for me to use it? Yeah, you know, wow, you're a company that doing analog photography and you do this fucking. I want to, I want to check out your product, you know, So I feel like it's important that people at the top, they believe in that and oftentimes they don't. Brand is very unquantifiable. Ethics, you know, it's not like, all right, so you spend, you know, 30k on this ad. How much are we gonna get back from? It's like, I don't know, mate. Like, I don't even know if people are gonna like it, you know, so you can't really quantify a lot of these things. You just have to believe that it's a compound effect.
- [122:38]
Rid
Before I let you go, I want to take just one minute to run you through my favorite products because I'm constantly asked what's in my stack. Framer is how how I build websites. Genway is how I do research. Granola is how I take notes during crit. Jitter is how I animate my designs. Lovable is how I build my ideas in code. Mobin is how I find design inspiration. Paper is how I design like a creative. And Raycast is my shortcut every step of the way. Now I've hand selected these companies so that I can do these episodes full time. So by far the number one way to support the show is to check them out. You can find the full list at Dive Club Partners.