Subscribe
Share
Share
Embed
Conversations with designers, founders, and builders behind some of the best work
Long Podcast - State Of Play - Ryo Lu.txt
English (US)
00:00:00.040 — 00:00:14.520 · Speaker 1
This is real, Lou. He went from redefining productivity as notions founding designer to cloning himself in a personal operating system at cursor. My conversation with him reignited my creative soul, and it just might reignite yours.
00:00:16.600 — 00:00:17.880 · Speaker 2
What is real OS?
00:00:18.760 — 00:00:22.000 · Speaker 3
I want to show people like, oh, a designer can make all of this.
00:00:22.400 — 00:01:36.370 · Speaker 1
Not to blow up too much here, but like I am really. I really think that you are one of the most impressive creative minds working today. I want people to see. I want people to see you. I think there's a lot of inspiration people can find from people like you. Designers are feeling anxious, people that aren't sure how to get into this field right now.
As always, the artists in US are at kind of at odds with the commerce part of this. And so I kind of want to dig into that, especially with you. You're not like, you don't really consider yourself a designer. You kind of consider yourself a tool maker. I look at what you've done and you're kind of a mental model maker.
There's a lot that goes into this. And I just want to start by asking you. You know, you wrote that wonderful essay, um, how to Make Something Great. And there had to be a journey that led up to that. So let's, let's, let's start with, like, I want you to take me back for a second to the real Lou. Like the the creative adolescence.
Like, where was the budding seed? What was one of the first moments in your life where you said, oh, hell yeah. Like I just made a thing, and I can, I can, I can make things. What was the first thing you ever made?
00:01:36.610 — 00:02:12.140 · Speaker 3
I made my first website when I was 11. It was about like this anime I liked, uh, Detective Conan. It's like a like a classic. Like, there's, like, murder scenes, and then there's this kid that will, like, investigate and do stuff. And I made it in front page. I think I put it out, like in a free hosting site. They will put like little banner as like below your website.
Um, it's dead now, but that was the first thing. And then I kept making more websites like forums,
00:02:13.300 — 00:02:14.740 · Speaker 3
um, for
00:02:15.900 — 00:02:35.859 · Speaker 3
things that I liked. And I made one when I was 17, I was called Mac Idea. And then that one was like the first Chinese language, like online community for Apple fanboys, basically. And that was around like when, you know, like it was
00:02:37.020 — 00:02:41.700 · Speaker 3
2009, you know, the iPhone kind of started. And then
00:02:43.140 — 00:02:53.620 · Speaker 3
back then there weren't that many people talking about it, like in China. So I made something and made a lot of friends. And I was doing college, so
00:02:55.300 — 00:03:10.189 · Speaker 3
still like making websites, also doing design stuff more. I got into like typography. Graphic design? Like photography. Making print stuff too, like book covers. And it's like you're kind of start starting from
00:03:11.310 — 00:04:13.440 · Speaker 3
just building things for yourself. Things that you like and then building things that others like you might like. Like people like my like. And then you grow from there, and then you kind of like, I didn't know anything. Like, I just tried different tools and figured it out. And then you try, like, all of these little, all the different tools from, like, all levels like code or Photoshop, I use like fireworks for websites.
Then there's like after that as like doing startups and like more, bigger things. I did one in college was called Scheduling Direct. It was like a SaaS for like retail employees. People can manage schedules for their like shift workers. They can like exchange shifts. There's like timesheet employee management.
That was fun, too. It's like
00:04:14.520 — 00:04:20.559 · Speaker 3
I didn't know what like designers were or product managers were or
00:04:21.680 — 00:04:23.119 · Speaker 3
like, just do everything
00:04:24.800 — 00:04:26.160 · Speaker 3
and then don't work out.
00:04:27.280 — 00:05:17.859 · Speaker 3
So I was depressed for a little bit and then I had a friend. It's like a childhood friend. He was born in Wuhan. Like the Covid place. And then we actually started something there. Uh, for a little bit. We did, like, a language learning app. You can imagine, just like Duolingo, plus some games, plus like a dictionary all mashed in one.
And it was really fun, but I also didn't work out. And I went to went to Shanghai and we did like a Chinese trip. So in China back then, it was like there were like 20 something different payment methods. It was like the peak of all the online payments. And then there's like all the things popped up around that time.
Like
00:05:19.060 — 00:05:42.380 · Speaker 3
ride hailing apps, um, bike sharing deliveries, online payments. Um, and then we just did like one API that did everything. And then there's like one unified interface, one dashboard, one signup flow. You just integrate once you get everything. And that was fun. And then that was when, like I,
00:05:43.420 — 00:05:46.380 · Speaker 3
I really looked up to stripe people back then,
00:05:47.420 — 00:05:58.430 · Speaker 3
um, because they were doing pretty cool things with like such a boring Thing by itself, like an like a rest API. You know,
00:05:59.870 — 00:06:12.910 · Speaker 3
like, how do you make something like that really easy to understand and people will feel like, oh shit, this is cool versus like, oh, this is like just a developer tool.
00:06:14.310 — 00:06:16.830 · Speaker 3
So that's kind of like how I started. Um,
00:06:17.910 — 00:06:23.350 · Speaker 3
and I came here, I got a random like email from a recruiter,
00:06:24.590 — 00:06:32.270 · Speaker 3
and I was growing my team. I had four people back then. Uh, I wanted to learn how people hire people here.
00:06:33.310 — 00:06:53.240 · Speaker 3
So I did the loop, and then they flew me here, and they put me in this, like, Victorian house. Really weird. Um, I got chased by some, like, random people on the street after my interview, and then I got to the offer like the Dales.
00:06:54.480 — 00:07:43.240 · Speaker 3
And then I went back to Shanghai. And then I think there's some, like, business issues. So, like, our business model didn't really work out. Like stripe can say take cuts from the transactions because the, the, the base fees are pretty high. Our thing is like because everything was starting in China back then, like all the fees that the providers had were like insanely low, say like 0.02%.
I can't charge above that. And like, it's just like hard to do it at scale. So the company had to turn into like a more service oriented company that will kind of help you integrate payments with their thing, with their system. But then it's like not as fun. And I became a designer. Yeah, yeah, yeah.
00:07:43.240 — 00:09:06.060 · Speaker 1
Well, well, it's funny, right? Because I actually relate very heavily to your journey. I started also in front page when I was about 1011 and I my first website, I collected X-Men cards and I had a really bad scanner. And so I would scan them on my scanner and they'd be black and white and blotchy and then upload them.
And I was creating a directory of my X-Men cards, and I emailed Marvel and I said, is it okay if I upload the X-Men cards to the website? And they responded and they said, absolutely not. And then I was like, well, they're not going to find my website anyway, so I just did it anyway. But what I learned was you could have the banner ads, and I was in little IRC chat groups, and I found from somebody this script and it would click around your website, and I used it and it would sometimes click the banner ad.
And so after a couple of weeks, I got a check in the mail for $20, and my parents said, where the heck did you just get $20 from? They didn't trust the check or anything. And that was the first moment where I realized, oh, I can like make money on the internet from this stuff somehow, and I didn't really know how to do it.
But it was a really big moment for me. And and I want to ask you, when was that moment for you when you realized this, this thing, this art of creativity could also be profitable? Hmm.
00:09:07.420 — 00:09:09.860 · Speaker 3
I think when I was doing this payment stuff.
00:09:11.580 — 00:09:43.860 · Speaker 3
Yeah, like, you can really see everything kind of changed. Especially like, say I did the Chinese stripe thing in 20 2015 and I did the stripe trip thing in 2018, 2019. Um, I think those are like pretty big moments, almost. Uh, in China it was slightly different. It was like people didn't have, say, cards back then or like there's that very low usage of that.
And then they just went straight to mobile,
00:09:44.940 — 00:10:18.590 · Speaker 3
like from like a span of a year or two. You see these QR codes everywhere. Like even in the physical world, like, like merchants on the street, they have a QR code. You scan, you pay, and then people, people start like sending money online across different things. And all of these, like internet business, popped up every week when I was there.
There's something new. And then people are like giving out like coupons and promotions and like fighting each other for the market. Um, it's really crazy.
00:10:18.910 — 00:10:37.470 · Speaker 1
So it was like the friction for payments for internet businesses in China went away very quickly. And then you saw a saturation of internet businesses. And so I've got to imagine, as the creative venue said, hey, I want to create things. And then you saw the opportunity. Oh, these creations can can make some money,
00:10:38.910 — 00:11:12.880 · Speaker 1
revenue or or chasing money from this for some people, for some creatives might add a constraint that you have to work around as well as now you're dealing with things when you're no longer just making for yourself. Even if you're just making within a close proximity of people like you. Another constraint that you're introducing is critique.
Do people like this? What was one of the first times that you really felt like someone or or your customers or a mentor tore your work apart, but it actually made you better?
00:11:13.200 — 00:11:16.960 · Speaker 3
I think I got the most from notion or like earlier days.
00:11:18.320 — 00:11:20.479 · Speaker 3
Like, I was like the
00:11:21.920 — 00:11:39.120 · Speaker 3
one of the earliest people for design. I joined Ivan was like next to me. And every day at like four for 30, he will come over and look at my screen and then kind of just pick at everything.
00:11:40.480 — 00:11:48.010 · Speaker 3
I had a lot of anxiety back then. We're we are also very different. It's like I'm more like vibes.
00:11:49.330 — 00:12:29.810 · Speaker 3
I think about like, high level things and then I go straight into details and then I try different things, kind of random and then boom, ah, I know what to do versus like Ivan's more like, oh, let's like tuck all these constraints one, two three, four, five, six, seven, eight, like all the way to like 42. And then let's brute force all the things like do all, all the variations are possible and then let's pick one.
So like our mental models are also very different. But I think after I went through it you can kind of see better pictures as a whole. For certain problems you need to like go through every single pathways and options
00:12:31.370 — 00:12:33.650 · Speaker 3
and consider all of those.
00:12:34.930 — 00:12:35.809 · Speaker 3
But also
00:12:37.330 — 00:12:39.130 · Speaker 3
like I was not wrong either.
00:12:40.690 — 00:12:55.620 · Speaker 3
You can't, like box yourself up in the way you work today. How you think. How you operate. How you make stuff like you need to look at how other people think
00:12:56.940 — 00:13:09.940 · Speaker 3
and how maybe they do it through, like giving you some direct feedback. Maybe they just do their work and you observe and you talk to them. You
00:13:11.540 — 00:13:33.540 · Speaker 3
bounce with them and see what happens. We are all solving the same problem in the same space. Like there's this big picture, everything, space. And then we all have different perspectives and things we care about and priorities in life, different backgrounds, cultures.
00:13:35.100 — 00:13:42.919 · Speaker 3
And then we came together and now we were kind of like putting all of our all the things we are seeing
00:13:44.120 — 00:13:45.000 · Speaker 3
on the table.
00:13:46.560 — 00:13:48.920 · Speaker 3
And then the more things you can see, the better.
00:13:50.200 — 00:13:56.480 · Speaker 3
And it's not just about you yourself, it's about everyone in it.
00:13:57.200 — 00:14:29.160 · Speaker 1
How do you deal with that friction that sometimes occurs? You said something pretty powerful, I thought, which was I was not wrong either. Um, and so making room for the idea that there are two correct approaches in the room right now and for some of us and I know for me, too, uh, that friction, we can sometimes live in that friction for too long and let that eat at us.
We saw the survey saying designers are experiencing burnout.
00:14:29.600 — 00:14:30.200 · Speaker 3
Yeah.
00:14:30.320 — 00:14:35.320 · Speaker 1
How do you work through that while preserving your creative energy?
00:14:35.560 — 00:14:39.530 · Speaker 3
You just have to accept there are certain things that you cannot change Or
00:14:40.610 — 00:14:42.849 · Speaker 3
that are just
00:14:44.410 — 00:14:45.370 · Speaker 3
a little hard,
00:14:46.410 — 00:14:55.410 · Speaker 3
and it might be okay to just let that part go a little bit. And I think it's like if you think about design problems,
00:14:57.290 — 00:15:02.930 · Speaker 3
maybe each of them in isolation, if you think about them, they don't really matter.
00:15:04.290 — 00:15:12.050 · Speaker 3
Silica my my bun at this size, is it like four pixels radius or three pixels radius? It doesn't really matter.
00:15:13.250 — 00:15:17.770 · Speaker 3
But all of these little decisions, when you put them together, then they matter.
00:15:19.170 — 00:15:37.810 · Speaker 3
Then the problem becomes like you need to help people see different layers and perspectives. Like, sometimes it's fine to let let this small decision go. Maybe it doesn't go in your way, it's okay. But maybe the the the the actual problem is you need to let people see like.
00:15:40.300 — 00:15:49.619 · Speaker 3
The higher level things like. What are you actually doing? What is this like as a concept? How does it relate to other things? How does it evolve
00:15:50.620 — 00:15:51.580 · Speaker 3
in the future?
00:15:52.260 — 00:16:32.460 · Speaker 1
I do that a lot in in that I am, I think, very macro. And then I zoom in on the micro and I kind of I oscillate between those two perspectives frequently. And for me, it all makes sense because it's all up here. But for the people that I collaborate with, sometimes they're like, wait a second, I don't understand.
For the people like me, who? That's how we think. How have you found ways to take people on that, that journey or to to kind of show them those different levels of fidelity or elevation, I guess. Mhm.
00:16:33.260 — 00:16:36.580 · Speaker 3
It's like what you see here. Nobody else sees.
00:16:37.710 — 00:16:39.310 · Speaker 3
So you need to communicate
00:16:40.750 — 00:16:56.510 · Speaker 3
with people in their language. You need to maybe take some time to think or like to to learn what's in the other person's head. Like what is their mental model? How do they approach problems? How do they
00:16:57.790 — 00:17:05.990 · Speaker 3
do their work? And then there might be like some different like a couple of different archetypes. You can try to kind of dice your idea
00:17:07.310 — 00:17:17.310 · Speaker 3
down into the piece that they can understand to start. Start from there, and then you kind of layer them up. Then maybe that's the old way.
00:17:19.949 — 00:17:30.870 · Speaker 3
Because the new way is you just make it happen. Or like you just make a thing that tells the whole story. And it shows people how it works. Because I think that's our
00:17:31.990 — 00:18:06.959 · Speaker 3
like, the designer's superpower. It's like we can make things visible. Like people can see it. People can feel it. It used to be like mocks or like Figma prototypes. You click and then boom. And then and then you show up and click on it again. Boom. That helps. But now it's like you can actually make it real or like really close to what you want it to be, exactly what you want it to be.
And you don't have to spend too much effort. So I feel like that is the way forward. It's like the designers actually just make it happen. Or like
00:18:07.960 — 00:18:16.319 · Speaker 3
you dump your brain out and you make an artifact. That is what it what exactly it should be. Or like close enough
00:18:17.440 — 00:18:40.490 · Speaker 3
that other people can just play with and they can give you feedback. You can even build like a crazy, like debug mode, like you have all the sliders and different options for people to pick from. Then you know more like what to do next. And you don't really need to communicate like, ah, let's make a PRD and make a slide thing and do some meetings.
And, uh, you know.
00:18:40.530 — 00:18:56.969 · Speaker 1
I almost look at it like there's a couple of different loops that we might build for one of those loops might be building to communicate my ideas. Mhm. And then there might be building to learn
00:18:58.090 — 00:18:59.970 · Speaker 1
uh, so that I can gather more inputs.
00:19:00.250 — 00:19:00.730 · Speaker 3
Yeah.
00:19:00.970 — 00:19:07.610 · Speaker 1
How do you think about building to learn in in where we're at now and where things are going.
00:19:08.450 — 00:19:16.530 · Speaker 3
A lot of people don't realize, like how cheap or like how how low the cost it is to build new things now.
00:19:18.490 — 00:19:19.570 · Speaker 3
Like before,
00:19:20.610 — 00:19:59.460 · Speaker 3
like our processes might be designed around, uh, because and just like very constrained and takes a lot of resource, the designers need to do things upfront that the PMS need to do things upfront. Things need to be set, decided and then we don't waste as much cycles building stuff that we just built was.
But then your reality is like if you just do that once, you actually don't get enough like feedback loops, you might just the, the V0 and then it's actually done and it's going to be, you know, like some crappy thing that gets stuck for like years versus now.
00:20:00.700 — 00:20:41.310 · Speaker 3
Maybe the designer can just come up with a bunch of options, and then you send it to the agent, maybe like, you know, you spend five of them, though, build all your ideas at the same time, come back with them. You click on the link and then you see the preview. And then you can make them a little better, and then you can put them all together, send them as a link.
People can play with it. People can see it. People can comment and give you more ideas. You can repeat that loop and it's so much faster. So it's almost like what took month is now maybe less than a day. And a lot of people don't realize that notion.
00:20:41.310 — 00:21:08.030 · Speaker 1
You had to design for a generalist, almost a generalist, use case knowledge management. And maybe you think about it the same way at cursor. But both of those categories were kind of blue sky, very wide open ways to solve these use cases. What were some of the major contributions you worked on at notion a notion?
00:21:08.910 — 00:21:29.160 · Speaker 3
I touch mostly everything. Mhm. Um, but I care most about like our core system and concepts and then I help create notion I so we kind of design it to be like a layer above everything. And then it ties everything together. And there's a, there's a thing that it's not shifted Did
00:21:30.200 — 00:21:33.680 · Speaker 3
that I really cared about that I was working on for like years.
00:21:33.880 — 00:21:41.760 · Speaker 1
Like, my question to you is, as you're kind of working through a lot of these ambiguous challenges,
00:21:43.440 — 00:21:56.080 · Speaker 1
we talk about building to learn. How do you think about input gathering before you even start building to learn? Are there places you seek inputs before you begin building, and if so, what are those?
00:21:56.160 — 00:22:24.720 · Speaker 3
I see myself as a sponge, as almost like you need to suck in everything from all the channels and you don't discriminate. Say like a notion I was we started from like, say online communities. There's like Twitter, Reddit, there's like intercom tickets. There's like all the things that are coming from sales.
There's like internal ideas, high level company goals and directions. Um,
00:22:25.760 — 00:23:39.949 · Speaker 3
and then your own idea and like, make your vision, your boss's vision, and then you can manage everything up together. Oh, and the most important thing is actually like what you have right now. Like what is actually in your system, in the code base, in the concepts in, in the landing pages. Like, what are people see?
Like a lot of us maybe don't even know the whole thing when we work. We really cared about. Like we need to kind of fit everything into one person's brain, at least. Like the whole skeleton. It doesn't have to be like full detail, but you need to know, oh, these are all the things that are there. It's almost like, you know, what is intuition?
Intuition is just you. You become the sponge and then you let it simmer for a little bit, and then it becomes like you don't think anymore. That's intuition. And you build that. Like you build the intuition. You also want to build like it's almost like a collective intuition, so that everyone's kind of thinking of like this, like a similar mental model.
And it's like truthful. It's not like something made up. Like a lot of people, maybe they came from like Facebook, Google or whatever, and then
00:23:40.950 — 00:24:12.710 · Speaker 3
they're so stuck in like one shape of working or thinking and they don't really even care about other things. So like, oh, I only care about how things look. I don't really care about, like how it's implemented or how does it relate to other things. So like if you can broaden that up and then you absorb more information and then you build like a system or like something to help your you and your team get that information
00:24:13.750 — 00:24:14.470 · Speaker 3
that helps.
00:24:15.670 — 00:24:19.190 · Speaker 1
In that same vein, talking about intuition.
00:24:20.230 — 00:24:21.109 · Speaker 1
How do you
00:24:22.710 — 00:24:26.360 · Speaker 1
know when something you've made is is good.
00:24:26.800 — 00:24:27.160 · Speaker 3
Yeah,
00:24:28.680 — 00:24:39.480 · Speaker 3
I believe in that. There is actually like one ultimate solution given, you know, all the constraints and information that's there.
00:24:39.680 — 00:24:42.320 · Speaker 1
A universal truth of this is the answer.
00:24:42.440 — 00:24:45.040 · Speaker 3
Yes. But then the problem is
00:24:46.160 — 00:25:20.440 · Speaker 3
we almost never see everything where like we we just never see everything. And then this thing constantly changes, like the world changes around us. And this ultimate answer just moves. Like you kind of want to, like, chase them or a better way. There's actually like, you could design something that can kind of kind of hold it for a little bit or like the best things are like, like if you look at history, there are certain things in our world today like that just never changed,
00:25:21.880 — 00:25:30.010 · Speaker 3
even like in computing. So like apps and the desktop, the icons, the folder files,
00:25:31.850 — 00:25:37.250 · Speaker 3
even like different views, like less views, tables, little cards.
00:25:38.370 — 00:25:42.249 · Speaker 3
Like none of them changed. Much
00:25:43.930 — 00:25:51.050 · Speaker 3
like what would be be something like that for your business or like if you're doing something more general, like what is
00:25:52.130 — 00:26:02.530 · Speaker 3
the concept that will last? As things get better and more possibilities get unlocked, they are going to stay and they will get better.
00:26:03.770 — 00:26:04.970 · Speaker 3
What are those things?
00:26:05.370 — 00:26:08.530 · Speaker 1
What is the fundamental truth that you can create?
00:26:08.650 — 00:26:09.170 · Speaker 2
Mhm.
00:26:10.010 — 00:26:15.090 · Speaker 3
And then if you figure that out everything else becomes a little easier.
00:26:16.570 — 00:27:02.940 · Speaker 3
It's almost like if you play it like Age of Empires you start the game with like, ah, you only see like your, your little town like right here. And then you need to kind of explore the map like everything else is like dark. So could you need to have seen it then you know where to go. So you need to kind of absorb information like become the sponge.
You kind of see it and then you can decide, maybe I want to climb this hill versus that one, or you can do it the other way, which is maybe you're just kind of given this little box and then you're just here, and then you paint that little box, you paint it fully. But the problem is the box is wrong. Like that is the hell to climb, not this one.
Then you're fucked.
00:27:04.220 — 00:27:17.260 · Speaker 1
Real. I've used that. I've. I've talked about that metaphor. I've used civilization and the fog of war analogy before. I've never heard anybody else say that that way. I made me smile to hear you say that.
00:27:18.340 — 00:27:37.030 · Speaker 1
Um, I love that analogy. And that's going to be a fun one to talk about. In that same note. Can you recall a time where your interaction or experience with the product and it can be your own made? You say, that's it. That's the thing. That's the fundamental truth.
00:27:37.470 — 00:27:55.190 · Speaker 3
I think one good example is notion. It's like if you look a notion, there hasn't been many new things fundamentally, because we're all just kind of evolving and building on top of the system we had. And there's this project that I was working on like for years.
00:27:56.430 — 00:28:12.350 · Speaker 3
Every time we come back to the same solution, it doesn't matter. Like how many times you tried. You always come back to the same thing because given all the constraints that you have in your organization, in the thing that you've made and in the future ideal version of it,
00:28:13.390 — 00:28:16.630 · Speaker 3
that is the answer. Like, you just can't do other things.
00:28:16.790 — 00:28:44.160 · Speaker 1
I want to switch gears a second and we'll get into some cursor stuff too, because I think this will segue nicely, but I want to hear from you about tooling, specifically the tools that you've used, maybe the tools that used to be mainstays in your stack that maybe are no longer, or your new tools are. Tell me.
Tell me a little bit about what your tool stack looks like today across all the different categories of work that you perform, not just necessarily design.
00:28:44.520 — 00:28:49.080 · Speaker 3
Or do I use mostly the browser cursor notion?
00:28:50.840 — 00:28:52.520 · Speaker 3
That's it. Almost.
00:28:52.640 — 00:28:55.880 · Speaker 1
And so no more drag and drop canvases for, you.
00:28:56.760 — 00:28:57.400 · Speaker 3
Know, more.
00:28:57.920 — 00:29:07.040 · Speaker 1
Coming from a front page background no less, where you first started to explore, you find yourself now living entirely in the editor.
00:29:07.480 — 00:29:49.090 · Speaker 3
Yeah, I think once you figured out so like the base visuals for the thing you're building, like, say, the type scale, how the menus, the dropdowns, look how things fit together. Then you kind of. That part is done almost. Uh, you can say like add some embellishments or like some cool things here and there, but that part is done.
Next part is really about the concepts. Or like what's in your thing? Well, how does it work? Um, how do you architect the thing both like in the front end and like in the data?
00:29:50.130 — 00:29:57.130 · Speaker 3
How does things pass around? Like, you can't really do that in thick. Or you can try, like describing it in a doc.
00:29:58.450 — 00:30:07.970 · Speaker 3
But then now it's like you can actually just put whatever you're going to put in the doc to the agent and then it will try to make it for you.
00:30:09.170 — 00:30:18.620 · Speaker 3
And then you get into that feedback loop that I just talked about is like, you think you you see it, you tweak it.
00:30:19.820 — 00:30:35.900 · Speaker 3
You just keep doing that. Um, and then the difference is, before you were tweaking something, that's. It's almost like a projection of the thing. It's like a abstraction of the thing that's kind of stuck in a format.
00:30:37.100 — 00:30:41.420 · Speaker 3
Say it like a PRD is like a textual representation of what you're going to build.
00:30:42.500 — 00:31:03.900 · Speaker 3
Who is it going to serve? What are the things in there? What are the options of implementation? How do we like, measure and test things? The marks are like maybe like different visual states of the thing. Like how do things look? Or like this is like the flow of x, y, z. And then these are the different states and then how they flow together.
At the end all of these
00:31:05.020 — 00:31:16.920 · Speaker 3
become like code. That's shipped to people. And then they see it in the browser or they tap around. That's what it is. So instead of
00:31:18.080 — 00:31:24.520 · Speaker 3
like getting stuck in these artifacts, just do the thing and make the loop.
00:31:25.080 — 00:31:28.520 · Speaker 1
Just immediately put yourself in the real source of truth.
00:31:28.680 — 00:32:31.530 · Speaker 3
Mhm. And it doesn't have to be the real, real thing. It could be like an idealized version of it. So like I want to try some crazy ideas for cursor, but I don't want to pollute the code base with some random shit. Then I'll make my own version of cursor maybe as cleaner like to start, it has less things. So like I don't really care about.
Like I don't know how we make cursor really smart and fast underneath because my my ML engineers can figure that out. But maybe I want to play with some like crazy, I don't know, transitions for getting into the background agent. Or maybe I can spend like five background agent at once. How do I do that? Like, instead of doing that in Figma, I'll just come up with some ideas.
I might sketch them out too, but it might not be fun. Like you don't don't. You don't need full fidelity anymore.
00:32:32.650 — 00:32:35.970 · Speaker 3
You just need to kind of maybe take a picture.
00:32:37.250 — 00:32:38.930 · Speaker 3
Put your ideas in some bullets
00:32:40.090 — 00:32:47.530 · Speaker 3
and see what happens. Tweak. Keep going. And you cannot get to a place that you like.
00:32:47.970 — 00:33:00.730 · Speaker 1
Can you walk me through that a little bit in a little bit higher fidelity of maybe using Rio OS or some feature as an example of how you begin a brand new prototyping effort?
00:33:01.210 — 00:33:02.290 · Speaker 3
Mhm. Tools.
00:33:02.330 — 00:33:04.410 · Speaker 1
Testing process inputs. Yeah.
00:33:05.530 — 00:33:15.060 · Speaker 3
Mm. I think for designers I would actually recommend like you know play with the vibe coding tools to start. I don't want them to be too scared.
00:33:16.220 — 00:33:43.420 · Speaker 3
Like cursor right now. I think it's still, like, designed for, like, more experienced programmers. Um, actually, like intentionally like that for a little bit. The thing that matters more now is actually like the thinking or like knowing what to build. And how do I kind of describe it in a way that I can just get something out of it from the agent?
00:33:43.620 — 00:33:56.700 · Speaker 1
Do you build any sort of text based clarity before you begin prototyping features or things like that for yourself? Is there a research or an information gathering phase for you?
00:33:57.620 — 00:34:07.980 · Speaker 3
I think, um, for my own things now is purely vibes almost for cursor. There's definitely a lot of like we need to do some
00:34:09.460 — 00:34:12.350 · Speaker 3
like gather some Feedback or
00:34:13.830 — 00:34:17.309 · Speaker 3
I think these are really important. They touch like
00:34:18.750 — 00:34:32.070 · Speaker 3
these features. It's important to kind of list them out. But once you get to like the task level things or you kind of know like this piece, I want to do it this way, you can just
00:34:33.230 — 00:34:36.469 · Speaker 3
prototype it out. And I try to
00:34:37.510 — 00:34:40.229 · Speaker 3
it's almost like you create a prototyping environment
00:34:41.230 — 00:35:02.190 · Speaker 3
where you can just put a lot of ideas together. It's almost like real. It's kind of like that baby cursor is also kind of like that. It's like you start with like a base that has most things, the basics, so that you kind of feel you're like in that environment.
00:35:03.350 — 00:35:07.229 · Speaker 3
It feels real, even though maybe some part of it is still fake
00:35:08.230 — 00:35:10.440 · Speaker 3
and you put new things on top of it.
00:35:10.800 — 00:35:13.760 · Speaker 1
I look at Rio OS and I see.
00:35:16.440 — 00:35:37.919 · Speaker 1
I see Terraria. I see. I see someone who is. I don't want to say there's no intentionality. I think there is intentionality, but a very let's see what happens and see what you know, what I uncover as I mine my way through this terrain. Mhm. Um, and I want to ask you what was the whole
00:35:38.920 — 00:36:13.290 · Speaker 1
notion, was this, like I said, uh, maybe a generalist knowledge management problem space. And now you look at cursor and at least as it's starting, as you just called out, very kind of a niche, it niched itself a bit. Maybe there's plans to obviously get more broad with that in the future. What was the pull for you with a tool like cursor.
And then and then after that the follow up would be What made you decide to start playing the Cursor Terraria game in Rio OS?
00:36:13.890 — 00:36:18.010 · Speaker 3
Aha. Okay. Yeah. What pulled me to notion was
00:36:19.450 — 00:36:30.890 · Speaker 3
I just thought all the sass stuff don't make sense, like single purpose apps. So like, you look at maybe I should name them, but you know, like.
00:36:30.890 — 00:36:32.010 · Speaker 1
I'll say maybe like.
00:36:32.050 — 00:36:53.170 · Speaker 3
Salesforce or Salesforce Monday. Whatever. All of those. Everything. If you look at there are all the same things. There is actually no difference. Fundamentally underneath they're all just like data stores, databases. There are some like, you know, rows of data that have different properties. Maybe this status is like ongoing.
This is like at risk.
00:36:54.250 — 00:37:03.490 · Speaker 3
There's some people on it. There's some like comments on it. It's in the list view in the board view in the whatever. Um, and then these are like the databases that are connected like
00:37:04.690 — 00:37:10.940 · Speaker 3
it's all the same thing. So why are we wasting time rebuilding the same things?
00:37:12.460 — 00:37:18.780 · Speaker 3
Like, it just doesn't make sense to me. And then I looked around and I thought, hmm.
00:37:20.260 — 00:37:35.660 · Speaker 3
No, she's the one that's doing the right thing. Because they actually treat all of their data the same way. Every block, every page. Every database row. Every database, you know, touches a block.
00:37:36.900 — 00:37:40.140 · Speaker 3
And we can transform how we visualize things however we want.
00:37:41.380 — 00:37:56.540 · Speaker 3
Then the the the the the problem that got me interested was how do we design one system, like one conceptual system with the fewest concepts that can model all software or like most software?
00:37:57.860 — 00:38:25.350 · Speaker 3
So that's why I went to notion to figure that out. You know, we started from almost like a note taking app Up. And now notion can do everything almost. I think there's still something that's lacking that that need to wrap everything up better and serve the things people want better so they don't have to think about, oh, this is a database view of like this type with these properties like that is still too low level for most people.
00:38:25.710 — 00:38:30.470 · Speaker 1
Give me my raycast layer so I can just quickly prompt whatever I want out of thin air.
00:38:31.390 — 00:38:38.670 · Speaker 3
Yeah, that might be okay. And then what took me to cursor was is actually the same spirit
00:38:39.710 — 00:38:54.390 · Speaker 3
for notion and cursor for myself. Cursor is just like a little bit more low level meaning like you actually touch the code and the models, the like beneath it. And that basically gives you like maximum possibilities.
00:38:55.510 — 00:39:18.840 · Speaker 3
Like if you think about it, the world right now is kind of run by software. Code can almost represent anything like it can be both. The input can can be the output, it can run, it can be interactive, it can make stuff. Then if you build a tool that can write any kind of code for anyone, then anyone can make anything.
00:39:20.720 — 00:39:22.400 · Speaker 3
So that's why I'm here.
00:39:23.120 — 00:39:27.480 · Speaker 1
So you didn't join cursor just to clone yourself? No.
00:39:27.520 — 00:39:31.240 · Speaker 3
Okay, fair. That's more like an experiment in the real thing.
00:39:32.000 — 00:39:37.840 · Speaker 1
You went to notion, and you worked very heavily on the abstracted layer.
00:39:39.280 — 00:39:52.360 · Speaker 1
We wanted to be able to create databases. And essentially, you know, some people would refer to notion as a no code tool, which I always thought was interesting. And then obviously they did publish. But then you went over to cursor to work closer to the code,
00:39:53.400 — 00:40:10.460 · Speaker 1
maybe to eventually abstract that away. I don't know if that's the plan, but I find that very interesting. And so then you say to yourself, or I guess the first question I have about that is, do you think we will abstract away from code completely eventually?
00:40:11.580 — 00:40:16.939 · Speaker 3
I don't think we will completely, but I think for certain people it will be
00:40:18.060 — 00:40:23.700 · Speaker 3
more like certain people just doesn't care as much or they just don't understand it. Then why do you show it?
00:40:25.100 — 00:40:31.620 · Speaker 3
Or you show something that they will understand better? You show different abstractions of the code,
00:40:32.660 — 00:40:37.700 · Speaker 3
but it is code still. Like what? What's under? It's the same thing.
00:40:39.220 — 00:40:42.060 · Speaker 3
That is not some like random abstraction. That's
00:40:43.740 — 00:41:01.950 · Speaker 3
that say like I design. So the difference between, say, notion and cursor is in notion the world. We kind of designed it ourselves for our users. And then that comes with some constraint. If you want this thing to be simple enough, you just can't have too many things.
00:41:03.550 — 00:41:21.470 · Speaker 3
You can make each of them, like, combine really well and make them like super flexible, generalizable. But you will always hit a limit. So imagine like a notion like maybe you can do some SaaS app like stitch them together now, but you can't make like a GTA six,
00:41:22.550 — 00:41:29.110 · Speaker 3
you know, like you can't do some crazy stuff. You can do the SaaS, but maybe not everything else.
00:41:30.350 — 00:41:47.350 · Speaker 3
But with cursor you can actually make anything like, you know, people talk about cursor for X, like, oh, I'm building cursor for medical research for data analysts. But then the problem is cursor can be that like cursor for x is cursor
00:41:49.150 — 00:41:56.470 · Speaker 3
like it is already possible. All you need to do is you open the folder with all your files and you start asking the agent.
00:41:56.710 — 00:41:59.310 · Speaker 1
So you decide to make real OS
00:42:00.360 — 00:42:19.040 · Speaker 1
Now, Rio OS is a passion project, and I'm going to dive deep into what Rio OS is in the video. But I want you for a moment to put on your your messaging or your marketing hat and sell me as Rio OS today. What is Rio OS.
00:42:19.600 — 00:42:23.160 · Speaker 3
I actually like? I don't really want to find what it is.
00:42:24.200 — 00:42:33.280 · Speaker 3
It's it's like it's like a playground and also like, it's almost like a museum in some way. You can see like
00:42:34.600 — 00:42:50.640 · Speaker 3
computing ideas and UX patterns over the span of like 40 years. And if you open the iPod, there's the same thing for music for that 40 years. I want to show people like, oh, a designer can make all of this by by one person.
00:42:51.760 — 00:42:56.920 · Speaker 3
I want to show people you can vibe code something really, really fast,
00:42:58.370 — 00:43:01.290 · Speaker 3
Really deep. Works well together.
00:43:02.690 — 00:43:03.570 · Speaker 3
It feels cool.
00:43:04.610 — 00:43:17.490 · Speaker 3
It's not slopped. It's actually good. And then I want to kind of just show people. Like all these abstractions. They are the same thing underneath.
00:43:18.570 — 00:43:47.850 · Speaker 3
Like, I made a whole atlas with apps, with windows, with files, with everything. Stuck in the browser. In the webpage. But if you look at it, it has everything. Like, you're just kind of doing the same thing every step. Doing the same thing in different layers of the abstraction. That's what we do for software, but it is actually the same thing.
And with AI, you can actually unify a lot of these concepts. You can break them down. You can, you know, cross the bridge them.
00:43:48.890 — 00:43:50.730 · Speaker 3
And it feels interesting.
00:43:50.810 — 00:44:00.180 · Speaker 1
And you've worked on some pretty incredible products and creations will really become your favorite creation. Do you have favorites?
00:44:00.380 — 00:44:33.340 · Speaker 3
It is the thing that I spend the most time on personally. It was really hard to get into building because like once I became a designer, I kind of stopped, like coding, were making stuff. And every time I want to go back, like I kind of look at, ah, what are the newest, like web dev stuff? And then, oh fuck, I need to like reset up all the things and do my environment like, oh shit, what is this VR thing?
Like, what do I use this versus like Next.js? Well,
00:44:34.420 — 00:45:02.030 · Speaker 3
and I'm like, I don't want to do this. Like I'll just, you know, do other things, but this time it's just so easy. And then once you get into it, it's almost like you're just you're in this constant flow state. And every time you do something and it appears and then you make it better, it gets closer to what you want and you feel really good.
It's very addictive. I would say.
00:45:02.710 — 00:45:10.830 · Speaker 1
Do you feel like that same feeling you had when you made your anime website in front page again?
00:45:11.030 — 00:45:18.549 · Speaker 3
Exactly. I think that's what it is like. You get that feeling back because making things
00:45:19.830 — 00:45:33.390 · Speaker 3
there's like a lot less burden. You don't have to ask for permission. You don't have to. Like, I need to write a dog first and then like, share with the team and then do something like you have an idea?
00:45:34.750 — 00:45:45.830 · Speaker 3
Just do it. See what happens. Maybe it won't work. That's fine. Try it with another model. Boom! It works. Oh, it's like close, but not quite. Let's fix this part. Boom! It works.
00:45:46.990 — 00:46:28.240 · Speaker 3
So that's the difference. I think you know the Lani chart. Like designer burnout. Notes. Like my read on that is like designers in a lot of companies, they are actually like, they need to serve all the people in that chart. Like below them. Right? Like the marketing people, the growth people, the the PMS, the the the engineers, the founders.
And they're kind of like the balancer. They need to kind of balance, oh, give kind of like what I said. Like you, you become the sponge. You suck all this information and then you give them all this constraints. Here's what we do. But then the problem usually is given all of these constraints, this thing that we do is not the thing the designer wants to do.
00:46:29.520 — 00:46:41.000 · Speaker 3
That's not the ideal version of the thing. It doesn't have all the details. It doesn't have all the animations. It doesn't have all the polish. It's like rushed as like bad.
00:46:42.440 — 00:46:48.400 · Speaker 3
And then the designer gets burnt out because they can't make the thing they wanted to make, but now they can't.
00:46:48.760 — 00:46:52.280 · Speaker 1
Is that who you feel like you're serving that person in that situation?
00:46:54.450 — 00:47:02.570 · Speaker 1
The person who is for any variety of reasons, maybe not just the designer struggling to convey create their ideas fully.
00:47:02.890 — 00:47:03.450 · Speaker 3
Yes.
00:47:04.610 — 00:47:05.210 · Speaker 3
Yeah.
00:47:06.530 — 00:47:19.290 · Speaker 3
It's like historically doing this needs like you need to know a lot of things in different domains, and you need to train yourself for a long time. And there's like only very few people who can do it really well.
00:47:20.370 — 00:47:35.170 · Speaker 3
But now with these like, like tools like cursor, if you have a really good idea and you know how to express it through words or pictures were any means like any medium, any input.
00:47:36.570 — 00:47:41.610 · Speaker 3
And then you get this AI agent or like cursor to help you make happen.
00:47:42.650 — 00:48:10.860 · Speaker 3
You see it, you can talk to it again. And you could get closer and closer and closer. And then it might just, you know. Sometimes you don't even know what to do. Then the agent can help you think. Give you a bunch of options. Maybe you didn't know what this thing was like. I didn't know what the lead was, so I asked, ah, is this the best thing out there?
And the agent will do a web search. Figure out like, ah,
00:48:12.340 — 00:49:03.240 · Speaker 3
here are the top. Like, I don't know, bundlers or like things that you might want to look at for the new, new, new, like people who never experience like coding. This is actually the fastest way for them to learn code if they want. And the passwords in the future might be if you don't want to, you can still make some stuff.
But then I think the stuff that you will make without knowing what, like how things actually work, won't be as good or complex or sophisticated as the people who do. So you kind of raise the bar for everyone, but also for the people who Are really good at this thing. They get like 100% better. Like, they can basically like parallels to parallelize their brain.
So like when real brain now becomes like 100, I can do 100 things at once.
00:49:04.280 — 00:49:05.280 · Speaker 3
It's kind of crazy.
00:49:05.720 — 00:49:26.080 · Speaker 1
You did want to clone yourself. Um, I have one final question. Really? I appreciate your time, man. I just wanted to ask you, if you you disappear for five years, if you just. The world kind of stopped and then you woke back up to it. Oh, what what do you hope the tools of the world look like?
00:49:26.840 — 00:49:33.320 · Speaker 3
I hope there will be like, more creators, more makers, more tool makers, like tool makers
00:49:34.600 — 00:49:40.040 · Speaker 3
who are the best at their game in every domain, making the best tools for people. Like I,
00:49:42.000 — 00:49:44.720 · Speaker 3
it's only like, you know, it's all trained on
00:49:45.800 — 00:49:46.880 · Speaker 3
public knowledge.
00:49:48.370 — 00:49:51.650 · Speaker 3
There's a lot of things stuck in our heads right now,
00:49:52.970 — 00:50:09.290 · Speaker 3
and like, we're still living in a bubble like nobody has, really. Like, I saw a tweet like some somebody just learned about anthropic. Like from the news now, like, people might know more deep seek than like cloud.
00:50:10.330 — 00:50:16.169 · Speaker 3
Like people just don't. Don't really realize this. Already possible now. And imagine like in five years
00:50:17.170 — 00:50:26.970 · Speaker 3
what would it could be? And I'm, I'm actually a little worried because I can imagine, you know, people talk about like, it's going to be hard for like junior people to get jobs.
00:50:28.570 — 00:50:54.020 · Speaker 3
But how I think about it is like this thing actually, like levels the playing field for everyone, the junior people, if they spend more money than vibe coding, they might actually be better than like a senior engineer. And I think that will happen. Like there will be like 16 year olds making crazy shit that like we won't comprehend.
And I want that to happen. I want, like, you know, like designers instead of them.
00:50:55.180 — 00:51:04.700 · Speaker 3
Uh, my PM didn't let me do this. They just fucking do it. And it's like an improv. The next day, we're like, we start. Start, like,
00:51:05.900 — 00:51:11.420 · Speaker 3
not calling ourselves these random titles, and we're just builders and makers.
00:51:12.900 — 00:51:22.420 · Speaker 3
And I can build in the way that I prefer. I don't have to be forced to use any tool or input.
00:51:23.540 — 00:51:37.180 · Speaker 3
Maybe I'm a yap or I will just like, talk to the agent all the time. Maybe I like drawing pictures, then maybe I still use Figma. I still draw my mocks, but then the agent can just look at the marks and make it make it happen.
00:51:38.540 — 00:51:47.590 · Speaker 3
Or like I'm more like a word cell person. I think in linear like words. That's fine. You just ride your dog and then put it in there and then boom, it gets done.
00:51:48.790 — 00:51:50.390 · Speaker 3
Maybe it's that. I'm not sure.
00:51:52.510 — 00:52:44.750 · Speaker 1
I love talking to you. You don't need to build an operating system, but you can turn your playground into something real. For me, I think in stories and game mechanics. So I'm building walls. To some degree it is. My brain externalized. Ryu showed me something I wasn't prepared to see. He built an operating system to have a place to play, to think out loud, in full fidelity instead of static pixels.
And those 16 year olds he mentioned. They're not waiting for permission. They're already doing it. No preconceived notions about how things should be done. Just doing. I missed that. So I'm building my own playground where stories become experiences. So do me a favor and ask yourself, what playground would you build?
Because I promise you, it is already there in your head when no one's watching. And it might be time to let it out. I'll see you next time.