Flora Guo - How to accelerate your design career with AI

00:00:00How can we use AI to just reallysupercharge our own process as learnersand thinkers and designers? What AI isreally great at is giving us a map ofthe territory. It's like when you'replaying a video game and you see thatthere's like clouds surrounding areas ofthe map that you haven't looked at. AIis really good at blowing those cloudsaway. Basically now you have a veryclear idea of what is possible, but it'sstill up to us to walk the terrain.Welcome to Dive Club. My name is Rid andthis is where designers never stoplearning.Today's episode is with Flora Guo who'shad one of the more impressive careertrajectories of anybody that I'veinterviewed on this show. She'scurrently the founding design engineerat Paradigm and before that Guillermo,the Vercel CEO, personally DM'd her toget her to join the team. So we're goingto do a deep dive into career growth,design engineering, what it's like beingthe founding designer at an AI-nativestartup, and a lot more. But I wanted tostart this episode a little differently.So I asked Flora to share some of herfavorite takeaways from the recent TokyoDesign Conference. Brian Collinsbasically brought 50 of the mostinteresting

00:01:15designers in the space. Sopeople like RioRyan Lovin, Gabriel Valdivia, a lot ofyour guests. And it was just such anincredible opportunity to learn fromeveryone across those two days. I'msomeone who always always taking noteswhether it's online on my website orphysically in my notebook and I justreally wanted to create a place forthese notes to live. And recently stampshave been living in my head rent-freethanks to some incredible partners. AndI was also really inspired by one of myfriends Josephine Ong who just shippedsuch a cool portfolio with this infinitegrid interaction. Same one that you'dsee maybe on like It's Not About Cosmoslike public.work. And I just knew I hadto like sit down and build something outmyself just to scratch that creativeitch. So, this is ultimately what I'mgoing to do with it. I was totallysomeone who wanted to go, couldn't makeit, had incredible FOMO, and then youreleased all of your notes like this andI was like, "Oh, thank you." I madethese little stamps for each talk that Ihad notes for. And Three that reallyresonated with me are, if I go into likethis view where I can see them all aspages, Brian Lovin did this incrediblesession about building software with AI,and I think

00:02:30it was just so incredible tohear about how AI has been empoweringhis team at Notion with ProtoPiePlayground, and just the way that we'removing as an industry, it's going fromcreating pretty images and artifactsthat are just for show to designersbeing able to showcase interactions andreally get into the nitty-gritty of howthings will display and be experiencedby the end user. One that really stoodout was Rio's talk on building the nextcursor. You know, cursor is a tool thatwe all use, but just getting to see thephilosophies that underlie what he'sstriving with the team. So, this idea ofblack-box AI interfaces versus classinterfaces. How can we really surfacethe intent that the user has? Also, whenwe look back at the history of softwarewith folks like Bill Atkinson and AlanKay, um in the very early days, thepeople who were creating the designs andbringing it to life in code were one andthe same. And because of that, we justhad such a tight integration of theiroriginal intention and what peopleactually got to see and experience. Realquick message and then we can jump backinto it. One thing that Dive Club hasmade abundantly clear to me over thelast year is that the practice of designis changing, and the old process ofgetting feedback just doesn't quite cutit in today's world.

00:03:45That's why I'mexcited to announce that Inflight isofficially in open beta. It's thefeedback tool that I've always wantedand it's built for a world that moves atthe speed of AI. So, I can share myprototypes, give context in videowalkthroughs, and Inflight makes it easyto get the exact feedback that I need tomove forward, whether it's voting ondirections or maybe even getting thegreen light to ship a new idea. And allof this is available in a single linkthat I can drop into Slack or maybe evenshare with power users to test out a newprototype. I use Inflight every day andit's totally transformed the way that Ishare work. So, I'm excited for you totry the product and if you ever want tojam about it, just email me atrid@inflight.co.If you've been listening to this showfor a bit, then you've heard me professmy love for Raycast multiple times bynow. So, I want to highlight a couplepeople who started using Raycast afterhearing about it on Dive Club. The firstis Rex Harris. He says, "Raycast is thetool that most shapes how he uses hiscomputer and keeps him in flow." Thenext is Greg Hunton. He called it alife-changing product and says, "Everytime he pulls a thread on something inRaycast, it wildly deepens the power andexperience

00:05:01of using his computer." Icouldn't agree more and I can'trecommend Raycast enough. It's totallyfree to start. Just head todive.club/raycastto check it out. Now, onto the episode.So, there's one talk that really broughtthe FOMO for me, which is Saelios'Geometry of Luck and he was teasing onTwitter and I just wish I was in theroom. So, what was your experience withthat one and maybe could share a littlebit of your notes there? I think this isdefinitely a talk that really resonatedwith me because Saelio proposed thatluck isn't, you know, something thatyou're born with or just falls in yourlap, but it has a specific shape and ageometry. And what that means is that,like any other skill, it's somethingthat we can, you know, just learn aboutand invite more of into our lives. So,he uh us through a few differentexamples, people like John Chapman orJohnny Appleseed. So, someone who fromthe outside looking in, you're like,"Oh, wow, he had so much prescience inlike scattering these apple seeds andplanting these orchards before, youknow, settlers came into the states."But, he was actually someone who juststudied the maps of where, you know,where are the rivers going, what seasonswill people be coming in, and then justbeing so aware of like everything thatwas happening around him. Kind of likefor us, you know, we see this

00:06:16constantdeluge of contact coming in from X, justbeing dimly aware of what thepossibilities are, so that you're wellpositioned as someone to just reallyjump in and start exploring with thesetools. People who just perceivethemselves to be lucky are also able toidentify the opportunities as they cometo them. They're creating a greatersurface area of this luck, whether it'sbeing in rooms or creating those roomswhere serendipity can happen. I thinkit's often times just like one momentthat can completely change yourtrajectory. This project in itself isalmost an inception of that very ideabecause you've shared this thing thatyou enjoyed, you put it out on theinternet. I I don't remember how manyimpressions it got on Twitter, but itwas a lot. It's like it was like almostlike 100,000 or something. Like peopleobviously enjoyed your work. Thisresonated with them. And that makes mewonder how intentionally you arestewarding your own surface area forluck. And maybe you could talk a littlebit about like the role that sharingonline has played in your own careerjourney. There's this quote that I oftenreturn to by I think either Ray orCharles Eames, which is that the role ofa designer is that of a very good host.And I think showing up in today's world

00:07:31where, you know, like you said, it isultra-competitive and there is like alot of stuff going on all the time withall these new tools. I think just reallyasking like what can we give, whetherit's holding space for an event orcreating an artifact about somethingthat we're excited about and want toshare. Really flipping the script oflike, instead what can I ask of thisperson, but like what can I offer them?I think has been just something that Iwant to keep returning to in the waysthat we show up online or in person forthat matter, especially when we hostevents, you know, inviting people into aspace, thinking about like what can Igather them around? Like what is theintention that I can set around this?The ways that we host space online isvery similar. Whether it's gatheringpeople around a common interest of likewe're going to screen a design film orwhether we're inviting them to show offtheir own work. I think it's reallyfiguring out like what are the threadsthat connect all of us together? Likewhat is the reason for us to be togetherin this space? And then figuring outlike how can I pull this out of othersas well. I'm just incredibly grateful tothe people that I've gotten to meetonline. People who start off maybe youjust like seeing each other's projectsyou're like, "Wow, this is really cool."Then you send them a DM and it kicks offmaybe a conversation

00:08:46IRL over coffee ora phone call and then it sparks eithercollaboration over projects or evenroles. When I was working as anindependent designer for a couple ofyears, I would say the vast majority ofmy clients came in from Twitter. And assomeone who's just always lovedexperimenting, you know, trying thingsout for the heck of it, it's been such agreat way to meet people with similarinterests. And even my current role itjust felt very serendipitous with howAnna, my CEO, found me online onTwitter. And when I was at Bracel aswell, it was just through conversationsthat started off in the DMs. So, I thinkany designer who's building projects,it's just such an incredible time thatwe live in to get to share with theinternet online. Can you go a little bitdeeper into how intentionally are younurturing that community that you'vebuilt for yourself on Twitter. Cuz Ithink a lot of people look at whereyou're at, and it's really inspiring,but also a bit of a black box in termsof how to even get to that point whereall of a sudden Twitter really is a verybig deal for your career. I remember wayback in March of 2022, that's when Ifirst got onto Twitter, and it was justme sharing things

00:10:02that I was reallyexcited about, you know, going to eventsat IRL, being like, "Oh, wow, like Ijust learned about X thing in design.Like here are my takeaways." And justshowing up, being excited, and reallydown to have conversations, I think justled one thing to another. Like differentroles that I think started off from a DMwould evolve into in-person things, likehosting an event together, or eventuallyworking together. People can really feelthe excitement that you have, whetherit's through a project or through notesor something that you're hosting. Sojust creating things that are artifactsof what you love and what you'repassionate about, and putting it online.There's putting things out there on theinternet and being a magnet for peoplewho care. And then there's alsoproactively messaging others and beingthe one that initiates. How much are youdoing the latter? Cuz I think that'slike something that's so valuable andyet a little bit scary probably for alot of people listening. Wheneverthere's a designer whose work you reallyadmire, I love just like dropping aquick note and saying like, "Hey, I lovewhat you shipped. Here in particular iswhat really resonated with me." So Ithink whenever I get those in my inbox,it really makes my day. I would say thathosting events is also a really greatway to reach out to people and to havethem reach out to you as well.

00:11:17Whetherit's hosting like a design film night ora design engineering PowerPoint night orsomething about agentic interfaces, Ithink creating like very specific thingsthat you can offer people. So like, "Oh,come into my home or come into myoffice. I'd love to host you." Or, youknow, sending them something that youthink they'd enjoy. Just create somemoments where it feels so much morenatural. You're just talking to a newfriend on the internet. I think thereason why I had so many follow-upquestions on this category of topics isI look at your journey and I see anincredibly steep trajectory. Like howfast you've been able to really insertyourself and build momentum in thisspace. And you know, now you're atParadigm and working this high-profilerole. It's really cool, right? And solike a goal of mine in this conversationis to help people find little pieces ofthat equation that they can use toreverse engineer your trajectory. Somaybe we could even just talk a littlebit more about what that journey haslooked like starting with the designengineering internship at Vercel. Vercelis a company that I've looked up to formany, many years. Even before when Iknew what a deployment was. I rememberbeing on the homepage and

00:12:32thinking,"Wow, this is just so beautifully builtout." I actually started this designengineering journey with V0. I rememberback in 2024, I was looking around andjust seeing such incredible projectspopping up in my timeline and having noidea where to start. I think theincredible thing about these tools isthat you can go on with just an idea andstart noodling out what does the pathlook like to creating something that I'mreally proud of. So I started off withmaking shaders, you know, things thatlooked like sunrise and then moving onto personal projects in V0 as well. Onethat I really had a lot of fun makingwas this idea of a digital gift box. AndI feel like this was the first projectthat I just sat down and was like,"Okay, today I'm going to makesomething." I really love the idea ofsoftware as a gift. And I think theseare always just such an incredible wayfor me to translate that intention ofcreating like a space on the internetwhere I could package things up for myfriends, whether that's images, links,and then send it over to them. Andsharing it on like this basically foundits way to the Vercel team, and we wouldjust like have

00:13:47these one-offconversations. And eventually about amonth later, so in January, G DM'd me onTwitter and was like, "Haven't you everthought about working at Vercel?">> No way.And that was really crazy. I had, youknow, messaged him in the past, I think,just saying like, "Hey, reallyappreciate you. Thank you for buildingVercel. Like, I'm having so much funusing it." Okay, that's really cool. I'mjust going to underline that foreverybody listening. The fact that youDM'd the CEO of Vercel just to say thankyou says so much about you and how youlike this is kind of what I was gettingat earlier, right? Like that's a bigdeal. And then, you know what probablyhappened? I bet he forgot. There's noway that he actually remembered. There'sno way. But then he saw the project ormaybe somebody sent it to him. Heclicked in, said, "Okay, maybe I'llreach out to this person." And then atthe top of that page was your thank youmessage. And it was like, "Oh, wow. NowI really want to talk to that person."So, that's amazing. I just I'm shining alight on it. I think it's amazing justthe way that you are positioningyourself in this community. I think G isalso someone who's just incrediblyproactive with reaching out to peopleonline. So many of our hires at Vercelwere just people that he saw like on theinternet. And he was like, "We need totalk to this person."He has a mental called like "GBrain"

00:15:03where he just like puts all theselittle notes and then it's like, "Okay,now what's recruiting? Let's follow up."Just being the type of person that heis, he creates like so much serendipityboth at Vercel and in the world as well.I'm sure you're freaking out. You getthis DM. It's like unbelievable, right?You end up joining the team. You spendthis chunk of time earlier in in yourcareer surrounded by some of the bestmakers on the planet. And my assumptionis you're probably thinking a little bitabout who you even want to become inthis space. So, how did that experienceshape your mental model for the type ofdesigner that you wanted to grow into? Ithink it was such a surreal experienceto get to work alongside like peopleJames Clements, Matias Shaw, and Rono. Ithink just seeing the level of care thatthey put into their work, whether it'scrafting interactions for a marketingsite or really diving into theperformance of the app, it really showsyou that design engineering goes muchdeeper than the surface of how somethinglooks. It's really having the contextand the ability to take an intention andtranslate it into what the user isactually experiencing. At Paradigm aswell, it's something we're alwaysthinking about like how can we make surethat the vision that we have is whatpeople actually get to play with inprod. And I think there's just

00:16:18so muchto learn from the designers who arealways picking up new technologies andtools as well. People who are just doingit because they're really curious. Ithink something that really stood out tome about Rono is how he's always cookingon side projects, whether it'sdeveloping details or his personalwebsites. I think it's so inspiring tobe surrounded by people who make for thelove of making, not because you'relooking to wrap a specific project or,you know, just get something donebecause that's the finish line. It'sreally appreciating design as thisjourney of this practice that neverends. And I think there's a lot ofbeauty in just walking that path. I wasclicking through to Rono's websiterandomly the other day, and it wasdifferent again, you know, just acompletely different everything. And youknow, the joke as designers is we cannever find the time to update aportfolio. And this dude has a newpersonal website that is dripping incraft every 6 months. And you can tellit's like it's not even about the endoutput for him. Like it is that act ofcreating, which I don't know has alwaysbeen something that I aspire to growinto as a designer. It's just beingsomebody who creates because it islife-giving for my very soul, you know?I think especially when we're

00:17:34like justtesting things out, just having peopleonline that we can look towards that'slike, "Oh, wow, they built this reallycool project. They're talking about howthey built it here, the stack that theyused, the tools that they're playingwith." It just gives us that extra nudgeto be like, "Okay, today I'm going tosit down and figure out how exactly thispaper MCV works." Or like, "Today I'mgoing to sit down and set up my cloudcode files." Just having, you know, rolemodels like that I think has beenincredibly powerful for me. Before wemove on, just take us through a few ofyour favorite experiments as an exampleof what it looks like to just create forthe sake of creating and putting it outthere on the internet just because Ithink your Twitter feed is beautiful. SoI don't want to move on quite yet. Onethat I thought was really silly was myfriend's favorite movie is Wallace andGromit, and I wanted to make him just alittle artifact that he could play withon his home screen, which was, you know,his favorite character, which isFeathers McGraw, andyou know, having little things of theplot details. This is just somethingthat more of like a visual artifact thatyou can hook up, you know, create like a3D model, and then create like a 3Dmodel viewer, and just give something topeople as a gift. I think that'ssomething you guys can resonate with me.This is a project from way back whenwhere basically

00:18:50when I was starting offwith making projects, I wouldn't reallyknow how to approach it. So just a verysimple web app that you take any ideaand it breaks it down for you in termsof like here are things that you mightwant to think about on the front end.Like here are what what you can learnabout for the back end, just to serve asmore of a map maker for me. I think thiswas a project that I really enjoyed.Just returning again to this idea ofsoftware as a gift. Like what are waysthat we can create things that feeltactile and create things that you cansend to people and just say, "Hey, thismade me think of you." Recently, what Ireally love as well has been like whatare ways that I can play with thingslike animations, whether it's, you know,product or the Tokyo project. It wasreally figuring out like what are cooleffects I can play with. This one Ireally have to shout out like DanielPito, he put out this CSS like littlesnippet of how can we create these like3D rendering like a 3D infinite canvasto explore like things coming at you.And when I was playing with the 3Dcanvas as well, I think justexperimenting with like performance.Maxim Petkov is a friend who's justincredible design engineer. And, youknow, I sent it over to him. He's like,"Oh, by the way, if you change the CSSclass to transform,

00:20:05it'll be a lot moreperformant." So, I thinkjust always playing around with likewhat are new things that have beensitting in my cache and I now want totranslate into a project I'm actuallyworking on. It's clear that you'relearning a ton about even just frontend, you know? Like just the more youbuild your ideas, I'm seeing a lot ofexamples of like little technologies oryou had that 3D render for Wallace andGromit kind of thing. Like I have noidea how that works, you know? Butyou're figuring it out for eachindividual project and it's obvious thatthose are helping you reach that nextladder of learning, which then kind ofmakes me wonder like where was thestarting point for all of this? Like howmuch technical understanding did youhave when you first opened V0 and thenmaybe we can use that as a launchingpoint to understand what some of thosemile markers of understanding have beenfor you. I remember when I first startedto play around with V0, I had like avery basic grasp of HTML, CSS,JavaScript. I'm also someone who lovespicking up internet courses. One was UIengineering 101 by Mariana Castillo withGeneral Oh, cool. Yeah, I went throughthat. I loved how they walked me throughhere's a component, here's how youtranslate it with just HTML CSS, and alittle bit of JavaScript.

00:21:21And eventuallyalong the way, I was reading, you know,a lot of React documentation, trying tofigure out like, what are Reactframeworks? What's Next.js? And the waythat I think about it is that there'salways a minimum viable amount oflearning that you need to get you to thenext step, and then you backfill a lot.I think what AI is really good at doingis giving us artifacts, but they're notnecessarily grounded in understanding ofthe system, which I think can be reallygreat with making prototypes, littleone-off experiments, but the moment youneed to debug something, you're like,"Oh, man, I don't even know where tostart." Mhm.>> that's where, you know, it's reallyhelpful to push yourself to feel likeyou don't have a complete grasp withthese tools, and then step back, like,"Okay, what do I need to zoom out in? Orwhat do I need to zoom out on and reallydeepen my understanding of now?" For me,I know when I started off playing aroundwith these different tools, I justdidn't grasp how components should bestructured. Like, when should we breakthings down into modules or when can wepass down props versus like sharingcontext. And I think just learning bybuilding and like learning by asking hasbeen such an incredible way of likelooking

00:22:36at reference material online andjust, you know, asking your friends whomight be good engineers or people whohave done it before. There's a tensionthat I think a lot of people experiencethat hits on some of the things thatyou're talking about whereit's like, how much do I actually investin my own knowledge that like exists inmy brain? Maybe some of that is likesyntax, you know, like code levelknowledge versus just improving the waythat I work with AI and maybe I abstracta lot of the details, but I'm going tofocus all my attention on that front.And with everything changing, you know,that's like a interesting spectrum tothink about, like how you view learningand where you want to concentrate yourefforts. How do you think about whereyou want to fall on that spectrum? Forme, I think this is something that Iexperienced very acutely last yearbecause I was making a bunch of personalprojects. I'm like, oh, cool, like look,it's a demo. But when it actually camedown to debugging something, if I wasplaying like with an app API and all ofa sudden it wasn't loading well oreverything's crashing, I just felt likeI didn't have the necessary foundationalknowledge to like actually hone in anddive deep into the levels ofdocumentation. What AI is really greatat is giving us

00:23:52a map of the territory.It's like when you're playing a videogame and you see that there's likeclouds surrounding areas of the map thatyou haven't unlocked yet. AI is reallygood at blowing those clouds away.Basically, now you have a very clearidea of like what is possible, but it'sstill up to us to walk the terrain. It'sone thing to see what we can make andanother thing to understand it in itslike full entirety. And I think with AIwe're also moving up the layers ofabstraction. So we're not writing everyfunction by hand, like no one's doingthat. But the more context that you dohave about how it works under the hood,so how React works, you know, how ourprops being passed through, we're muchbetter equipped to actually debug andbuild things that are resilient. Itdepends on the level of abstraction thatwe're working at. So if we're saysomeone who's working with a very strongteam of engineers, what's most helpfulfor them is maybe like here is aprototype in code of like exactly how Iwant it and the engineer is, you know,hooking up all of the data layer foryou. Maybe they start off with uhPR that's just like the skeleton of likeall the API calls and how the back endtalks to the front end. And then as adesigner,

00:25:07you can go in and like polisheverything using Claude code and makesure that it's just exactly how you wantit in terms of the micro interactionsand the animations. Or if you're someonewho wants to be deeper down in terms ofimplementation, like understanding howshould this component be structured tobe the most performant, I think reallyjust going back to the basics andstudying has been like something thatI've returned to much more lately aswell.There's one question that I can't stopasking myself. What if companies appliedto talk to you rather than the other wayaround? And that question is thefoundation for the all-new Dive TalentNetwork. And it's working. Like rightnow, I'm helping many of the mostexciting startups that I know to hirethe designers and builders who listen tothis show. So, if you're curious whatmight be out there, and maybe you wantto get on my list, or maybe you're evenlooking for your next design hire, headto dive.club/talentto join today.You're someone that takes learningseriously. Just listening to your talk,that is something that's very clear tome. So, I want to ask a little bit of uhin the weeds question here, which is ifyou would throw up how you work with AI

00:26:22on a daily basis when coding, and youlook at the version of yourself sixmonths ago versus the version today,where do the biggest deltas exist? Ithink one of the biggest differences hasbeen creating explainers as I go withClaude. For example, in the past when Iwas learning about how React works withReact hooks, I would be building it toput in and be like, "Oh wait, like thisis totally something that can beabstracted out into its own customhook." So then, I would, you know, pullup some documentation, have aconversation with Claude, and then atthe end say like, "Okay, what are thekey points that we should remember fornext time or even just for my ownlearning?" Cuz with like skill files,you know, you can give Claude context,but I also think about how as we workwith AI, these mental models that we'rebuilding up like skill.md, fold.md, itreally encourages us to think aboutlike, "What are ways that we can buildup our own context?" Or "What are waysthat we can define our own invariance?"So, something I really love doing islike after I feel like I would trysomething new or I I learned somethingthat I didn't know before, I'll likequickly write up a markdown file andjust save that for next time for my ownreference. And like sometimes I'll postthese notes to my website, but justcreating artifacts that I cancontinually return to. I think that'sone of the biggest unlocks. Is there anexample that you posted

00:27:37to your websitethat we could look at? If you're goingto notes,umI remember like monorepos is somethingthat we use a lot at Vercel. Likeeverything that you see from the Vercelmarketing website to the app to allthose ship conference websites is hostedin this like massive repository calledfront. And I remember, you know, pokingin for the first time and I was like,"Wow, there's so many subdirectories."And for my own personal website, it'shosted on a Next.js app, but I wanted tocreate like a shared component systemthat I could use across projects.Figuring out like, "How do I do that?Where's turbo repo? Like how do I setthis up?" And then uh deep diving thingsand just writing down notes for nexttime as well. I think that's been reallyhelpful. I think the part of this thatactually is the coolest is just thetimestamps. Like how often you arecontributing notes from your ownlearnings to your own website.>> to think of websites as like livingartifacts. You know, being able to justjot things down. If there's a cool linkthat you saw, something that you want tosave for next time. I think justcreating things that you can return touh has been like such a theme umI love. Any other mental models thatinfluenced the way that you collaboratewith AI that you feel like you'verefined over the last year or so thatyou think might inspire

00:28:52other people? Ireally love thinking about AI as thisinfinitely patient tutor.I think it's one thing to ask for it todo something for me and not, you know,question anything about how it works orwhat assumptions it's building off of.But really adopting this posture of nowwe live in a time where I can ask anyquestion and I can ask it to break itdown into as fine a level of fidelity asI need. I would really encourage peopleto just think about ways that they canuse AI to scaffold their own learningand then to flush it out by buildingprojects, trying to understand like whatare ways that we can prompt ourselves tobecome better at learning. One of theways that that inspires me is justlooking at your own role right now. Likeyou are a design engineerand yet, correct me if I'm wrong,there's not like a traditional designerat Paradigm, right? Like you are kind ofthe designer, but you're also owning alot of the front end. It's a journeythat, in theory, a lot of people thatare listening to this could go on. Iactually think that I'm going on itmyself. Maybe not as quickly on some ofthe engineering concepts, but like I'mthe designer, but I'm just owning moreand more of the front end. It can bescary, you know, for people who have hada code of version historically and yetthe blueprint and and some of the waysthat you

00:30:07are leaning on the AI and andleaning into your own curiosity, I thinkis ultimately like like that's thescaffolding that allows people toprogress in this direction, which evenjust from, you know, the comments ofpeople listening to the show, I can tellis inspiring. Like it's a journey that alot of people want to go on, but ifthey're like trying to figure out likewhat are the right footholds to grab onmy own journey. At Paradigm, I'm very,very lucky to work alongside someextremely experienced and justknowledgeable engineers. Like we haveShirley front end engineer and Jerry ourCTO. I think I'm always learning fromthem and as someone who still feelsrelatively much earlier in myengineering journey, I think a lot of ithas been figuring out, you know, assomeone who's feeling out this designrole at Paradigm, what does it look likefor me to be able engineers who areworking across all their layers of thestack versus like what are PRs that makethe most sense for me to make myself. Ithink it's really figuring out like whatis the right level of abstraction andboundary to straddle. And for somedesigners, you know, they really justwant to tunnel deep into how thingswork. And I think some designers areincredibly skilled at just spitting outthis exact vision that they have andthen working closely with engineers orwhoever they're working

00:31:23with to bringthe exact vision to life. Can we talk alittle bit about your practice then?Maybe we could even look at, I don'tknow, a recent release or even just ahypothetically how you work on somethingthat feels a little bit, you know,meatier, more ambiguous. Like how muchare you exploring in code? How much areyou owning like all of the front endimplementation details? And you know,what are the steps in between? Like forsomebody that can reach for really anytype of tool, how do you think about theway that you work? And maybe if thereare certain ways that that has evolvedover the last 6 months even? I thinksomething that we as are in the processof figuring out are what are theintermediate artifacts that you cancollapse? And what are the things thatyou should spend your time like reallyfiguring out high fidelity versions of.Sometimes that doesn't even look like astraight. It's like sitting down withthe team and having a conversation oflike what is the right mental model? Sofor example at Paradigm, we recentlyshipped our workflows feature. And forus, you know, it's figuring out how dowe define the primitives of workflow?It's zooming all the way back out tolike asking what are the objects thatwe're concerned with here? Like whatdefines a workflow? What are theworkflows that users are building out?And then from there like what are theintegrations that we want to have? Likewhat are the actions that we want tohave? And

00:32:38I think it's so important toreally be aligned at a conceptual levelbefore we even start moving into likevisual experiments and prototypesbecause I think one trap that we canfall into as designers is coming up withlike a bunch of options and then pickingone too early. I think it's aboutfiguring out like what is the rightthing to create when through a paradigmonce we're aligned on like what exactlyto make, then we can hop into tools likeCloud Code or Figma. And I recently beenplaying around with Taper a little bitas well to just like put down a bunch ofscreens, sometimes in varying levels offidelity. Like, sometimes all you needfor a a meeting is to like draw thingseven as rectangles to be likeconceptually like here's the dialoguethat I want. And then uh progressivelygetting just like more alignment as ateam before you move into like acoded-up prototype. Something werecently played with was what are waysthat we can show this like one specificworkflow and then working with ourengineers to put up like prototype PRs,playing around, testing them, and theneventually uh converging into what wesee in prod today. For something likethe workflows,I'm just going to ask you to even gomore specific into what were

00:33:53the maintypes of artifacts that you explored?Like, once you were aligned with theteam, what were those first few visualartifacts that you made and why? For me,I still like to use Figma to, you know,visualize what are the different waysthat we could display this interaction.And then sometimes what I like to do aswell is like take a few of the versionsI'm working with and then put them intovariant UI and just see like oh, whatare some other ways that I could explorethis. And what I really love aboutVariant's product is that it works as ascrolling feed. So, once we've likedefined here are the constraints of whatwe need to see in this interaction. Likethere might be a pattern that like Icompletely do not have in my realm ofunderstanding and then just thinkinglike, oh, okay, like that's interesting.Like why what might be a fitting usecase for this? Because I think it'sunderstanding like what should the userbe doing and then like what are thedifferent form factors that we can servethis up in the interaction now. So afterwe play around with some kind of likestatic shots of like, okay, I want thisbutton to open up like this entry toyinto overflow dialogue, then it'splaying with a prototype that can

00:35:09eitherbe in cloud code directly in ourrepository or sometimes we have peopleon the team who are more product focusedand like to play around with other toolslike directly in cloud to just show theinteraction that they're thinking about.And we'll take like little screenrecordings and like little snippets andgive each other like GitHub PR links tojust test on staging and then see whatfeels right before I eventually we'relike, okay, like working backwards likethis is how we want it to feel in thefinal final version and and now we havelike this little graveyard of PR linksfor the test. Let's talk a little bitmore about your role as the designerstartup very AI native product. A lot ofunique challenges there. So talk to us alittle bit about how even the nature ofthe product has shaped your designpractice. So Paradigm, we work on a wayfor new to prompt basically a thousandagents at once. There is this interfaceof the spreadsheet which people are sofamiliar with as knowledge workers, butit's just a completely different modelthinking where in the past, you know,you define a formula and the spreadsheetfills it out for you. But now we see itas a way of interaction with this aswarm of agents,

00:36:24right? So like how canyou define a set of steps in yourworkflow once and then have thesecolumns each columns has a promptexecuted it out across like tens orhundreds or thousands of rows. So for ussomething that we're always thinkingabout is scale and how can we bring thisto an interaction that like feels goodwhen you have like 200 agents versusjust 20 running at once and I think areally great way that we've beendogfooding it as well as just thinkingabout like what are things toparallelize. So for me like I lovebuilding lists, I love collecting listswhether it's resources I love orlearnings that I have I'll put them inlike one column and then I'll createprompts to see like oh what are otherideas I can pull up related to this orwhat are other resources that you wouldrecommend. For example, if I'm trying tolearn animation like let's first buildup a list of like animation resourcesand then like what are threads I cankeep pulling and if really using the AItool as a way to help myself to figureout like okay where do I want to movenext. I want to double check on thescale piece for a second. Like what aresome of the design possibilities thatyou find yourself exploring or reachingfor to solve that type of problem? One

00:37:39thing that we've really honed on is whatare ways that you can represent thingsat scale in a very structured manner andwhen we returned to this idea of likegrids uh rows and columns it makes a lotof sense because this is a paradigm thatpeople are already familiar with. So forus it's figuring out like if I want tohave something at scale what does itmean to first like generate that list ofitems so like using an agent to scriptthe web and help me build this list andthen from there it's like what do I evenneed to extract or like what is the mostoptimal way of like interacting withthis like body of knowledge now. So forus like something we're referencing islooking at like other very data heavyproducts. So like what are ways thatwe've represented like vast swaths ofdata in the past, whether it's in aspreadsheet or you know, maybe at somepoint looking at graphs and documentsand things like that. I think it'sidentifying with the structured datalike what are the invariants? What areways that you can pull out somethingthat is constant across all of them,whether it's like a field that you wantto have or a shape of the data that candefine so that the AI can really helpyou build it out as you scaffold it.You're working on

00:38:55very cutting-edgetech, workflows, mental models, likejust the way that you are empoweringhumans to work is so different than whatwe're used to even as as a society,where my assumption is given your roleas a designer at a startup, you'rehaving to keep a bit of a pulse on likewhat else is happening around and how dowe think about the ways that this shapesour own product roadmap and patterns topull from. I know the answer is scrollTwitter, but like go a little bit deeperinto what you're doing as a designer tohealthfully stay up to date witheverything that's happening in the worldand hopefullybeing able to like identify the signalfrom the noise. You know, Twitter candefinitely get very overwhelming, but Ithink keeping a pulse check on here'slike what everyone's been talking about,putting it on my list of like things toexplore and then for me a really greatfilter is just chatting with my designerfriends or design engineer friends. LikeI have a backlog of products and tools Iwant to try and just comparing noteswith each other. I think makes it a much

00:40:10more fun experience to just do ittogether as a collective or as acommunity. I really love in-personevents for this, whether it's hostingthem myself or going to ones that myfriends put together. It just createsplaces for people to share littlesnippets of their curiosity and demo toeach other as well. Like, here'ssomething that makes me really excitedand that I've been diving down thisrabbit hole lately. I think sometimesit's a matter of picking one thing thatyou want to dive down into for a weekendor two and just trying to fog things outfor that period while you choose to godeep. So, staying broad, keeping allthis context and then when somethingreally grabs your attention, you'relike, okay, now I have to build it out.Now I have to try it out. And then it'sjust continuing to return to like gobroad and then go deep and repeatingthat process. I did that yesterday witha friend. Just kind of catching up. Wehadn't talked in I don't know, 6 12months or something. It's anotherdesigner andhe was sharing his workflow with penciland I was sharing my workflow with paperand it was so interesting to just seethe different things that he wastinkering with and experimenting withand both of us had arrived at thissimilar point where it's like, gosh, inthe last two monthshow I'm exploring on a canvas isfundamentally different than it's everbeen at any point in our career. And wewere both like so excited about it. And

00:41:25then he's like, have you tried this? I'mlike, no. I'm like, oh, have you triedconductor? He's like, no. And it's justso valuable to have those littlesessions with people because I feel likewe're kind of all just blazing our owntrail and figuring it out right now.Absolutely. Yeah. I think getting toshare your like personal workflows withfriends is just such a fun thing to dothat because you can dive into like,wow, like this project that you maderecently, like, how did you do that?Like, what are the tools that you'replaying with? And continuing to likehave this demo culture, I think it's sopowerful as well. Like in Canada, and Iguess well by now, we have somethingcalled Socratic Coffee where basicallypeople get together every Sunday. Theyco-work for 3 hours and then at the end,people demo what they've been workingon. Okay, cool. Being spaces where youcan foster that curiosity to make thingslike not necessarily for work or for aspecific project, but just havingdedicated space where you're carving outto be like, "Okay, today I'm going tolearn something new. Today I'm going tomake something new." And having thatjust because love the process, I think,is such a powerful thing to do. Well,that connects back to a lot of thethings that you were talking aboutearlier where like when you werescrolling your Twitter feed, it's somuch of you just saying, "I learnedsomething new. Here's what it is." youknow? And people really like that reallyresonates with people. And there'stotally this moment in time. Like gosh,if

00:42:40you're listening and you're inspiredbyyour journey on social media and puttingyourself out there in the community,this is the moment where you could dosomething similar where you just exploreand tinker and try different tools andtry different workflows and like what ifI did this instead? Or what if I usethis as a starting point and then justdumping your takeaways and yourexperience? Like people are hungry forthat right now. Even my own journey,like that's kind of how I started wasback in like 2019 when Figma was alittle bit newer and it was becoming amuch more technical product and a lot ofmore powerful like functionalitycomponent props. And I would justexplore. I would just get up and I wouldjust do weird things with components andthen make a, you know, tweet about it.And I would just do it again and againand again. It was fun. And I think thatsame opportunity is is definitelypresent in today's world just becauseeverything's changing so quickly. Doingthings for the fun of it is probably themost powerful thing and just alwayskeeping that true awfully alive andbuilding. Like how does the shape of ourtools shape us as the users of them?Like what are the interactions thatthey're affording but the ways ofthinking that are really becomingembedded in our psyche? Something youreally love is the cloud code like askuser tool where basically you can use AIas an interviewer to really help

00:43:55youcrystallize your own thoughts. So Ithink with AI there's like two paths youcan go down. One in which you're kind oflike, "Okay, just like do this for me.36 36 36." And you're caught in thislike infinite loop where you feel likeyou're not really breaking new ground oryou're hitting the same bugs again andagain. But, I really encourage us toflip it and be like, "How can we use AIto help us build the right mental modelsthat we need? Like to first figure outlike what is the level of the ofabstraction that I should exist at? Andwhat is the necessary context that Ineed to acquire in order to be effectivehere? Whether it's, you know, actuallygoing into React and understanding howdoes this component work? Like howshould components be structured? Howshould I be writing these prompts andthe way it's passing them down to, youknow, if I'm working with like a shaderor something, like how much math do Ireally need to know? I think with AI,we're getting much better at figuringout like what is the minimum viablelevel of knowledge I need to makesomething. But then like what is thelevel of knowledge that I need to becomesomeone who actually feels well versedin this specific technology or thisspecific tool. And I think for everyone,you know, it really depends on thecontext of like whether you just want tocreate a fun demo or prototype orwhether this is a product that you wantto sustain

00:45:10over time. Yeah, it goes backto what you were talking about in termsof viewing AI as the ultimatepersonalized tutor. Yeah, exactly.>> And actively finding ways to lean intothat rather than just to accomplish yournext goal and then your next goal.Absolutely. It's like how can we use AIto just really supercharge our ownprocess as learners and thinkers anddesigners. I feel confident in sayingthat you have supercharged a lot ofpeople's process listening because theway that you think about these tools,but also just the way that you approachyour own journey and your hunger forlearning and experimentation, it's superinspiring for us. So, I really, reallyappreciate you coming on and sharing itwith us today. Yeah, thank you so much,Red. It's been such a pleasure.Before I let you go, I want to take just1 minute to run you through my favoriteproducts because I'm constantly askedwhat's in my stack. Framer is how Ibuild websites. Genway is how I doresearch. Granola is how I take notesduring crit. Jitter is how I animate mydesigns. Lovable is how I build my ideasin code. Mobbin is how I find designinspiration. Paper is how I design likea creative. And Raycast is my shortcutevery step of the way. Now, I'vehand-selected these companies so that Ican do these episodes

00:46:26full-time. So, byfar, the number one way to support theshow is to check them out. You can findthe full list at dive.club/partners.