Ron Goldin - Building your ideas as a design leader

00:00:00And this is literally like right afterthis meeting where I was told by like VPpluses on the design side, this can't bedone. I was like, "Hey, check it out.Like I'm a non-engineer, but I justbuilt the ability to create an SVG thatcan be colored to any color you want to,can be transparent, can have all thesebrand colors, and a merchant never willhave to lift a finger." And we have anew online checkout that's like justpresent to the merchant be like, "Do youlike this? If not, here's some othercolors you can tweak." So like that'sagain that world of like I don't knowwhy I'm so excited right now. Like Ithink designers are literally have thissuperpower. Welcome to Dive Club. Myname is Rid and this is where designersnever stop learning. This week's episodeis with Ron Golden who's led design atGoogle, Shopify, Uber Eats, and in myopinion, he's one of the perfectexamples of what it looks like to thriveas a design leader in this new AI world.So, we're going to do a deep dive intohis journey as a builder, including howhe built a new product called Matchmakerfrom scratch. But before we get into allof that, I asked Ron to share a story ofwhat it was like as an early designerworking on Uber Eats. I remember like myfirst week on the job living in New YorkCity, had never

00:01:15done a delivery beforein my life. um wanted to do dog with theproduct and so I downloaded the the Uberdriver app and just went through theprocess myself. In New York City, you'vegot these city bikes everywhere. So, youjust like basically can check out a abike and you know, you get it for 45minutes or whatever. And so, I remembertaking uh a ride for myself and I mean,first of all, you know, we're bothdesigners. Like, the empathy buildingwas extreme. The onboarding was awful.Uh, I finally ended up getting thedriver app to work. Uh, I push a buttonthat says go online and I'm just likesitting there and like nothing ishappening and I I could not get a jobfor the life of me because Uber Eats wasstill pretty small. And so I basicallyjust quit. And that's kind of thebeginning of I think the first fewmonths of my Uber Eatats tenure like wasreally like more almost like of aresearch role than a design role becauseUber at the time like really had thisstrategy of like even though it was likea pretty large company like Uber rideswas in hundreds of cities it still actedvery much like a startup. So like youknow the driver app they literally justlike added a couple of buttons so youcan pick up a delivery but there wasnothing bespoke about it that reallycatered to the couriers. And so I foundthat to be true and so I started doingmore of these deliveries um

00:02:30first in NewYork City. Then I started bringing someof my colleagues along with me to do thedeliveries. We ended up getting you knowactual gigs and dropping them off topeople and didn't get paid a whole lotwhich you know is part of a whole otherthread of conversation I had with myPMs. But um it actually turned into aninternational research uh endeavor aswell. A big part of our challenge waslike trying to scale to non US cities.And one way to do it obviously is justlike flip a switch, but the other one islike actually see what it's like todeliver in Mexico City, in Tokyo, inBangkok. And these are places weactually travel to and of course foundlike, oh my god, this is like a wholedifferent sport when you're doing adelivery in the sort of skyscraperneighborhood of Tokyo where there'sliterally like your front door to theperson's actual apartment is like a15-minute journey through like multiplegates of security. Like there's nolandmarks. Um, and so like, you know, wewould literally ride around these citieswith translators trying to figure out,you know, which way to go, like dodgingtraffic, almost dying, like all of thatstuff. One of the things I really wantedto do was like figure out ways to sortof bring that feeling of like the fear,the anxiety, the pain of like being onthese bikes and literally almost dyinguh to home base, so to speak. And Idon't know if you remember GoogleCardboard

00:03:45at the time, but uh>> Oh, wow. Yeah.>> Yeah. Yeah. So, um, I had recorded allof these trips using a very earlygeneration of the Insta 360 camera. Iwould mount it to the bike and I'dliterally record these trips to likeliterally capture what it's like to beeverything from dodging traffic tofiguring out like which way to go andmaps is telling me the wrong directionsto like even having security guards tellme like, "No, you can't come in withoutcertain credentials to like deliverfood." And so, I was just capturing allof this stuff. And so I I I sent thisGoogle Cardboard along with these Insta360 videos that I uploaded to YouTube toour head of Uber everything it wascalled. So this like Uber Eats, UberRush, a bunch of other products. ThisJason Droy, he watched them. He likeliterally watched the videos of, youknow, me sitting in a car in Los Angelesoutside the biggest sandwich restaurantin all of LA for 3 hours in a car with adriver and not getting a single gig. Andthis is actually I mean this is aperfect example of this. Like that wasthe one I sent to him because ourassumption up until this point was likeservice the big restaurants, right? Likethey're getting all the traffic, right?And so I was trying to like replicatethat and like this guy literally wassuffering in his car. He had hadn'teaten. He's like had to go to therestroom and like couldn't couldn't doanything. But

00:05:00wait. So that was thebeginning of pretty heavy product anddesign investigation into like why isthis happening? And what we realized,which now is kind of common sense, islike if all of the cers are being toldto go to the hot restaurants, thenthere's too many cers and not enoughmeals, right? And in fact, we didn'treally have a really good sort of likevisualization at the time on the corerapp to even tell people where theopportunities were. And so peopleweren't even like using the app to likeunderstand demand. They were justliterally going to Google Maps or Yelpor whatever to find popular restaurantsand like they would just park there. Andyou might even see this in your in yourtown like down the street from me.People like park outside of Shake Shackjust like waiting hoping, right? So itwas the beginning of like a long projectthat ended up with uh a like a reallyrobust way to kind of visualize what wecalled hotspots. So places that likearen't like the tried andrue places thatyou might imagine have a lot of demand,but like places that might even be likea small mom and pop Chinese restaurant,but that like all of a sudden happens tohave a little pop of demand, right? Andso a it was a way for people to likestart to sort of spread out theirefforts across a city. The other one waswhich is really interesting was like

00:06:15alifestyle thing which is like whatpeople used to do which is crazy is thisguy in LA used to drive one and a halfhours to get to that sandwich shop thathad no business because he figured inhis own neighborhood in like Ventura,California, there wasn't enough demand.Well, guess what? There was. We justdidn't have the tools to say it. So nowhe could stay like in his neighborhood,save on like an hour and a half of gaseach way and actually support localbusinesses and like have less of acommute so to speak to get home. Realquick message and then we can jump backinto it. Jitter just released animatedcomponents so you can design once reuseacross your file and update everyinstance with a single edit. I'mpersonally very excited about thisbecause I use Jitter for everything fromanimating the in-flight logo toanimating the buttons in the divenewsletter even. So, this is going tomake it a breeze to keep everything insync. And it's available today. Justhead to dive.club/jitterto get started. I've anticipated DES'srecent release for a full year. It'scalled surfaces and it enables you todesign and prototype directly on top ofan existing production interface. All ofthe pages and flows that you need arepreloaded inside of Destin as

00:07:30startingpoints. So you can easily make changes,explore ideas, and then when you'reready, you can share your prototype in asingle click. Destin is the only designtool that allows you to design onexisting pages and flows pulled directlyfrom prod which is a pretty big deal forteams and you can get started today.Just head to dive.com/desen.That's dsn.Now on to the episode. You can tell thatyou're a researcher at heart just howquickly you're just like I'm just goingto do the whole flow end to end. Butthere's definitely a version of thisstory where you put things into a nicememo and you sent it off to leadershipand they read it and they kind of nodalong and then it gets tossed into aspreadsheet somewhere and maybe youthink about it again like two to threequarters later, you know? And so thepart that I think is so interesting ishow you you captured all of the videolike you're like I'm going to think veryintentionally about how to transfer thisempathy and ultimately it seems likethat played a big role in in motivatingpeople to actually take action. Jason uhDroy and my manager at the time AndyZabolski, he supported me to actuallyget that video in front of the entireUber everything organization. Like Jasonwas like, "People need to see this."Like our PMs, our engineers, our opspeople need to

00:08:46really understand likefirsthand like what that pain is. And sowhat we did is we cut like a version ofthat video down from obviously like twoand a half hours to like 3 minutes. Buteven that 3 minutes, we showed it in allhands and we just like left a minute. Weleft a minute in that presentation ofjust like the feeling of sitting in acar waiting for nothing to happen,right? And like you know, you'veprobably been in an all hands of acompany or maybe you maybe you haven't,I'm not sure, but you know, they tend tobe packed with content and so like thepower of having like hundreds of peoplein a meeting where we're all just likesitting there awkwardly waiting lookingat the app, nothing going on. Like itwas such a powerful moment and itactually spearheaded what became I thinkthe kind of Uber dog fooding program. Solike now at the time like I said it waslike pretty scrappy like I was sort ofdoing this with a bunch of people. I gotmore engineers to do it with me and soon but it actually became a programacross all of Ubres where people wereencouraged to basically do it. We hadresearchers real researchers actuallyplanning the operations of all of thisso it scaled and capturing the insightsand all that stuff. So like the appetiteI think for especially like a humanservice like Uber Eats where there'slike real human beings like goingthrough the motions of doing something.There's a real appetite I think foranyone building products

00:10:01to likeactually put themselves in the shoes ofpeople and like you know it's>> it's so hard to do I feel like if youdon't at least have that taste of likewhat research feels re at least a littlebit of the skill set and maybe some ofthe skill set the courage right to justgo out and do it right yourself and knowthat you're going to fail and that'sokay. listening to these stories,there's a line on your website that Isaved that makes a little bit more sensenow where you talk about how you like tojoin companies at inflection points andyou know there's this need to to dothings differently or to take a bold betor to change the way that you operateand so I'm curious given your backgroundyou've you've spanned so many differentcompanies so many high impact roles isthere another example that we could talkabout just to give people an idea ofwhat this moment looks like in practiceand then how you show up as a designer.>> Sure. Yeah. Yeah. Yeah. I mean, maybeI'll talk about Shopify. I wasresponsible for leading online checkout.And if you go to I forget it's likeshopify.com/checkout or something likethat. They have this marketing page andin big bold you know large text it'slike the highest conversion checkout inthe industry. And so imagine taking ajob where your job is to basicallyimprove what is supposed to be theindustry standard is already great.Okay. So that's challenge number one.And then imagine that

00:11:16the CEO ofShopify, Toby Luca, is telling yourentire team, "We need to redesigncheckout, make it better.">> No pressure.>> No pressure. And also no direction. Likeit just has to be better because that islike his and I I love this about him bythe way. Um that is his DNA. Like he'salways looking to like, you know, tearthings down and build them up again,like question things, make thingsbetter. I would say like that was asurprise to me when I joined thecompany. I joined at maybe by accidentthe perfect In the time I was there, welaunched the partnership with OpenAI.So, we were like one of the firstcompanies in like the world to basicallylike have that aentic e-commerceexperience within the chat interface. Welaunched cryptocurrency with Coinbase.So, the first time Shopify had cryptoglobally and the checkout redesign. Andthis was like my first six months of thecompany. I don't think I've ever shippedthat much in such a short time during aramp up period of unemployment uh in myin my world. And honestly, I loved it.again kind of brought that same energythat I had in those neat days of like acompany that like yeah technically it'spublic technically it's on the stockmarket but it really operated like ascrappy startup you know this is where Iwould say I had worked in like kind ofAI adjacent you know before we alltalked about openi you know

00:12:31MLbasedproduct work all that but I think thatShopify was really my first foray intolike cursor and you know like that sortof vibe coding world like at the timeand again just to frame this is like ayear and a half ago like this is ancienthistory, almost nobody at Shopify wasusing AI, period. Like nobody like I waspart of a small cohort of designers thatwere like early testers of like FigmaMCP connecting to cursor and it was likemagic for all of us and it was alsoinsanely painful. Like I remember likeit took me at the time and a bunch ofpeople like we had the Slack going likeit would take people hours to get MCP toconnect because that's where the techwas a year and a half ago. It was justlike clunky>> and just like not very like yeah there'sall these things you have to know andlike you couldn't just ask Claudebecause it didn't have the answers tothe stuff yet. So, we're just likefiguring it out. And it was just reallyfun to be part of this like little crewof folks who were just like before theKool-Aid like what is this thing and isit useful or is this just a fad and canwe use this and is it going to beinteresting and then you know maybe I'lljust share what my light bulb momentwith this um which was so Shopify hasthis big meetup once a year. I forgettechnically what it's called, butessentially like a big part of it, likemaybe 80% of the time is like a massivecompanywide

00:13:46hackathon. Everybody, notjust design, not just engineering, likeeverybody, whatever your role is. I wasin the online space. I love physicalstuff. They had this like really coollike little garage, like literallygarage with like point of sale hardware,um little NFC stickers and just like ifyou want a ticker, go for it. And um Igrabbed, you know, one of the designersthat uh reported to me said, "Hey, youwant to work on this project? I havethis idea for a POS thing that can talkto an NFC sticker." And in I'm going tosay 8 hours tops, we went from an ideato like a fully working thing. And thething just to kind of give you a pictureis like imagine like a sticker that hasNFC chip on it and um you're a merchantand instead of having price tags likestatic ones or like little signs, youbasically just have these um NFCstickers that you can put all over yoursweaters and pants or whatever it is yousell and then a customer can take theshop app and just hover, they don't haveto unlock their phone, just hover nearthat sticker and you have like a littlenotification that's like productdetected or whatever. Tap on that.Basically, it says, "Oh, this is like ablah blah sweater, you know, made of,you know, 90% wool like

00:15:02and price andreviews, all that stuff." And you canjust tap to buy. It sends the purchaseover to the POSOS in store so that theperson basically automatically paid forit. It checked out. Don't have to get inline, don't have to talk to anybody. Thecash register person knows you didn'tsteal because it's in the POS and youcould just walk out the store. And solike literally we basically just builtthis prototype, right? Which is like theexperience I just showed, right? And theNFC connection, you know, was again wehad an we borrowed an engineer for like30 minutes to say like, "Hey, can youhelp us with that thing that we werestuck on?" Cuz again, probably todayCloud Code would do this like heartbeat.But we needed help.>> You kind of had to earn it back then.>> Yeah.>> I say back then it's only a year and ahalf ago, which is is insane. is youhave this really storied career arc andyet so much has changed in such a tinywindow here toward the tail end. Somaybe we could even use that as anopportunity to reflect a little bit likewhen you think about>> your own early philosophies arounddesign leadership and how these UberEatats and like even Google experienceswere shaping it and then what you likethink and how you operate today. I'mkind of curious>> where do the biggest deltas exist? Ithink you're definitely seeing a lot ofdifferent ways people have framed it,but

00:16:17essentially like a flattening of alot of uh orgs, right? Like less layersof management. That in of itself means,you know, if you are a manager at acompany, that can mean a few things.Like one, you probably have a lot morereports maybe than you're used to. Likeso like we call that wingspan, right? Anincreased wingspan for each manager,senior manager, director, right? Theother thing I think that has changed iskind of along with that, right? Like ifyou are I am seeing more of a playercoach need um folks that have reachedout to me, folks that I coach personallywho are looking for their next jobs uhat times. Um I've done a little bit ofthat on the side as well. I'm seeingmore employers, especially like preIPOcompanies that are looking for this likeplayer coach type, which is, you know, Imean, there's probably a lot of ways youdefine it, but for me, it it does oftenlook like maybe managing a smaller team,but like in my mind that is like aShopify size thing of like>> two, three, five, six, something in thatrange. You're managing the team, butyou're also with the team. Like you'rebuilding things together. you know,you're kind of not in this ivory towerjust doing performance reviews andalignment meetings while the IC's aredoing the things the IC's do like youare all you are all part of the sameteam and you kind of do have this dualresponsibility

00:17:33of like making sure theteam is happy and healthy and like doingtheir best work and like for particularthings that maybe are you know mostsuited to your strengths you're alsolike diving in like for me as an examplelike there's a few things that I'mpassionate about and these days and likewhere I find I'm either asked to orspend my own time one which you talkedabout is like zero to one stuff. Like ifI'm on a product team and everyone'sthinking really incrementally, likeespecially these days, I'm not going tolike write a Slack or an email and say,"Hey guys, I have an idea. I'm going togo to Cloud Code. I'm going to spend 20minutes. I'm going to v code somethingthat people come back to and see."Because in every company I've everworked. Like prototypes win arguments,right? Prototypes get people psyched.like you know moving away from slidedecks, moving away from PRDs, movinginto like something that people canactually feel and react to where likeyour emotion is literally>> I want that you know like I want to shipthat to our customers like>> that's the area I spend a lot of timeand so any minutes right now that I haveI am like at my at my full-time job I amtrying to find ways to push on my own AIskills um finding ways to apply them tothe work at hand for payments forexample like I will vibe code like anidea for like a a new digital wallet orsomething like that just to kind of showteams like ways of thinking about thingsdifferently. And I think the access tothese tools and the

00:18:48power of these toolsand the speed at which people can pickthem up actually means that it nolonger, you know, for a manager to likemake something on the side is no longersetting aside a couple of days to likebuild out your components in Figma andcreate something in auto layout and thenput string up a prototype. Like you cando that in like 100th of the speed. Andso I think for managers who are curiousand crafty, they can do a lot more withtheir day with the 40 hours that theyhave. And so I think that's also helpedmaybe for those managers like myself whoalso just don't want to sit in 9 hoursof meetings and just like manage peoplelike there's absolutely a place for thatin the world and it's super important.But like um I love problem solving. Ilove being close to the problems andlike seeing the users and talking tothem. And so that's that's been reallyexciting for me to like know that that'sactually becoming more of a viablecareer path again for people that arelike kind of on that fence between thosetwo worlds.>> I hadn't thought about it that way. It'svery obvious in retrospect, but like youhear the whole prototypes of the newPRDs or some version of that statement,but you're right. It's not just that wehave the ability to communicate moreeffectively, but the opportunity cost ofexpressing your idea in a greaterfidelity

00:20:03is driven down so significantlywhere maybe as a leader you don't feellike you're being pulled from the largerneedles that you're responsible formoving.>> I've done prototypes in like Figma makeor cloud code in like 10 minutes betweenmeetings while grabbing a cup of coffee.Uh there's an example actually and it'son my website too of like a moment thatI had at Shopify where again we wereexploring this online checkout redesign.The actual design is still confidentialbecause it's still a work in progressbut there's kind of like a tease to sortof the technology behind it, you know,which is like we're trying to make amore expressive version of onlinecheckout, right? And there are certainassumptions for what that would looklike. Those assumptions were thatmerchants would actually have to like dosomething to like make their checkoutmore beautiful, more branded. Like theymight have to upload some highfidelitySVGs of some logo or something or, youknow, customize some colors. And likeone thing we knew from research atShopify is that merchants, especiallySMBs, small smaller merchants, smallmedium business merchants, either don'thave the time or are not super techsavvyand like they just might not go throughthe motions of like doing that stuff ontheir own. And so, you know, in 2025,26, the AI light bulb has to go on inthose moments of like, wait, there'ssomething that people aren't willing todo because they don't have time or knowhow to like, could AI do this? And sothat was kind of my pitch was like, hey,

00:21:18um, you know, could AI basically look atan online store for brand X andbasically like pull even like a low resJPEG of a company's logo and turn itinto a vector so that we can like swapout background colors.>> Like could we like look at their websiteCSS and pick out possible primary andsecondary accent colors to use in theircheckout? And my design leader at thetime was like, "It doesn't seem likemerchants are going to do that. I don'tknow if we can do that technically." AndI, you know, again, pre-2025, what woulda designer do? They would ask anengineer like, "Can this be done?"Totally. Which, which by the way, I did.I did. And engineers are like not sure/don't have time to answer your question,>> which I could have told you that wastheir answer even.>> So I'm like, "Okay, this is crazy. Thisis probably going to be like fruitless,but like what the heck, right?" Like Iso I went into cursor at the time andlike vibe coded an internal tool to tryto do what I just described which wasliterally it's the ugliest thing everbut it was literally just like a formwhere you can type in outdoorvoices.comright a brand and it would basicallyjust like suck all of the possiblecolors variants from the site that couldpotentially be the

00:22:33brand colorsincluding like it's intelligent enoughso it'll be like, "Oh, like this wasused for all the buttons. This isprobably a primary color candidate. Thiswas used in, you know, some of the otherstuff. Maybe it's like a spot orsecondary color.">> Um, and then the logo thing, and we lovethis about AI, right? Like, I don't haveto look for an API. I'm just like,here's what I want to do. I want to findthe logo, you know, look in the header,look wherever we can, like try to find alogo that maybe has the brand name orsomething, download the logo, and turnit into SVG. And then I prayed. I'mlike, maybe this is the thing. And so, Iwas like, great. Found an API that doesthat. that's open source like and itjust sort of built the whole thing inand then while like this is literallylike right after this meeting where Iwas told by like VP pluses on the designside this can't be done I was like heycheck it out like I'm a non-engineer butI just built the ability to create anSVG that can be color to any color youwant to can be transparent can have allthese brand colors and a merchant neverwill have to lift a finger and we canhave a new online checkout that's likewe could just present to the merchant belike do you like this if not here's someother colors you can tweak. So likethat's again that world of like I don'tknow why I'm so excited right now. LikeI think designers are literally havethis superpower like unknownsuperpowers, right? Like I just

00:23:48didn'teven know that Claude would be able orKer would be able to help me there andit did. I remember sharing this exampleuh to another design leader and he'slike how did you know it was going towork? I'm like I actually didn't. LikeI'm literally just like throwing stuffinto, you know, the AI black box. AndI'll tell you, like having been vibecoding now for two years, like the stuffthat I'm doing now was so much harderthan it was a year or two ago slash feltimpossible. And that's the other thing,too. Like that's such a fun and weirdthing about this moment is like youcan't even claim to be an expert on anyof this stuff because literally likenext month it's going to be>> totally different.>> 20 20% smarter. It's going to be able todo things you couldn't do today. So,>> there's a Cleo quote that I keep comingback to where he talked about the auraof inevitability at early Facebook andhow when things are real and alreadykind of close to having the technologythere, it's just so much harder to sayno, we shouldn't do that.>> And all of a sudden, designers have theability to bring the aura ofability tothe table. And so, when you gave that,it's like, all right, shoot, we can'tignore this anymore. You know, it works.Like, he made this.>> Yeah. Yeah. I'll add one thing to that.I I I'm remembering this part of uh kindof my experience was that again at thistime like I was one of the early andcurious in

00:25:03the AI space. Most of mydesign team that reported to me had notdabbled in like vibe coding at all. Wehad one designer that was using thejourney. Like that was the extent ofkind of AI practice on my team. Iremember showing that example, thatstory to like the designer that I seethat was working on online checkout withme and she's just like holy expletivelike I cannot believe that you know youdesigner/ you manager like were able tolike make that let alone in like 15minutes and also by the way having theaudacity to like having a VP say no andthen being like actually yes like it'spossible right so I think like againmaybe going back to the role of amanager thing it's like that courage oflike being able to sort of push on thesetools A and B dog fooding tools insteadof just like sending them over to yourteam and hoping that they work.>> I want to look at the state of the worldtoday from a slightly different anglebecause you recently navigated the jobmarket as a leader and I'm wondering ifthere was anything that felt clearlydifferent this time around or just othersignals that you're seeing that kind ofpoint to some of these ways that thelandscape is shifting.>> Totally. I mean, we talked a little bitabout this sort of player coach beingmaybe a more invogue right now. I can'tthink of a single role for leadershipthat I've

00:26:18applied to. And, you know, I'mI've I've been a director for a numberof roles. These are very seniorpositions. I can't think of a single onewhere they weren't expecting you to behands-on. Period. Full stop. What theywhat that actually means is not clear injob descriptions. It sort of varies caseby case. But I'll give you an example.When I was talking to an earlier stagestartup, they literally put me on a callwith an engineer in Eastern Europe andlike watched me use Figma to make surethat I knew how to use auto layout.>> No way.>> Yeah. Yeah. Yeah. Like literally theygave me a design exercise reallyliterally was a test to see if I knowhow to use Figma, which by the way Idid. Yay. Like I got, you know, like Ipassed that. But like yeah, that careerfirst, you know. Likewise, I don't thinkthe AI tools thing is necessarilysomething that people are putting on jobdescriptions, but I think like peoplehave definitely noticed my passion forbuilding. They've noticed that I betweenjobs was not only job searching, butalso built an entire company fromscratch called matchmaker with my wife.I think seeing that somebody is not onlykind of like a tinkerer of AI butliterally can do endtoend productdevelopment wearing the PM hat, theengineering hat, the design hat, andthen also the like entrepreneur hat

00:27:33oflike you know this perfectly well likeall the other stuff like incorporatingthe business and Delaware and likepaying the fees and like figuring outyour contract with your partners likeall of that stuff. And I think that issomething I've been saying for a minute,but like I do think some of the bestdesign leaders are people that havespent some amount of time beingentrepreneurs or working in early stagecompanies again because of thatownership thing. Um, but I do think thatI'm seeing more people that are findingwhat previously might have been calledquote unquote personal projects asvaluable, right? Like I call this apersonal project because nobody hired meto build matchmaker but it was somethingthat you know me and two co-founderscreated from scratch. It shows the kindof initiative and the resiliency ofiterating through it. It shows productthinking, right? It shows technicalsavviness. It shows all of this stuff.And so I'm definitely seeing morecompanies that are looking at thingsthat you built outside of work, not asextra credit, but as possibly reasons toget jobs in addition to or even overtraditional work experience.>> You know, it's funny like I have a lotof design leaders in emails and TwitterDMs of people who are asking to come onthe show. you know, it's a differentlens,

00:28:49but what I thought was sointeresting about you was matchmaker,too, for the exact same reason whereit's like, oh man, not only do you havethe pedigree, but like talk about handsin the clay. There's there's no level ofdepth that you can get beyond building atotally new product and company going 0to1 on the whole thing, right? Like thatsays so much about who you are. So, I'mwondering if we can maybe even just geta little bit of behind the scenes. I'mcurious what that, you know, processlooked like. Maybe just were therespecific I don't know lessons learned orhurdles that you had to overcome thatyou think people listening who areinspired to do a similar journey or tojust start building more would benefitfrom seeing.>> I think many of us probably have a lotof personal project ideas in our head oflike problems that we have in our ownlives that we want to solve. This wasdifferent. This was my wife Melanie is ashe's in marketing. She's been marketingfor her whole career. And then she gotthe entrepreneurship bug and started aan agency specifically on what's calledbrand partnerships. You know, I am a anoutdoor goods brand. I'm a luxury brand,right? I want to get more people tonotice me and to build up a certainbrand equity and to get people to buy myproducts, right? Partnerships is like,okay, well, like let's find otherlike-minded organizations or brands orcreators, right? and like find ways towork together so that we can have awin-win situation, leverage each other'saudiences, so on

00:30:04so forth. And so whatMelanie and her partner noticed was thatthey were in all these Slack communitiesthroughout the world, you know, that areall about marketing. And people werelike, "Hey, we're doing the Venicebianale, you know, in a couple ofmonths. We're looking for whatever likeinfluencers that are, you know,interested in, you know, the outdoors orwhatever, right?" So like there'sconstantly these like little miniSlackbased RFPs that are just fallinginto the ether and so matchmaker wasreally an attempt to solve that problemin software. Now in terms of the processyou know Ihad my in-house sort of you know PMslash re researcherfounder which is mywife she was my source of wisdom on likewho the users are you know how theymight behave things they're looking forbut to be honest I think we are now inour third build of the product>> and and just to kind of like paint thepicture like we started thinking aboutthis again probably about a year and ahalf ago when I feel like the AI toolswere they were okay they were fine M umI didn't know I had to necessarily builda whole custom thing using a cursor orcloud code. Um there were a bunch ofoff-the-shelf products. The one wascalled software at the time with no e.It's almost like frame with like alittle bit more built-in functionality.I was like oh like

00:31:19this could work foran MVP like let's try it.>> And then one it was ugly as hell by theend and I was just like as a designerlike I can't like I can't. After a fewmonths of that, like we already saw thatyou know the capabilities of cursorlovable started becoming a thing andlovable actually really caught my eyeand I'll say two things like one assomebody who studied computer scienceand have has learned six differentprogramming languages at some point inmy life I was tempted to be like allgeek in on like let me just code this byscript from scratch but I knew thatthere was a good chance that I was goingto go back to a full-time job and I hadthis like dream that I could build aproduct and hand it to my wife and herpartner who are not necessarilytechnical and they could run thebusiness in lovable without me actuallybeing involved. That was the dream. Andso I was like, I'm going to chooseLovable even though it doesn't scratchthat geek kind of part of me to justlike see the first time I builtsomething 0ero to one. I very muchreplicated my my my own process as adesigner. And what I mean by that isliterally conversations like lots of youknow discussions with my co-foundersaround like who is our user what are theproblems we're trying to solve withliterally we're writing a strategydocument writing a PRD

00:32:34like we we haveartifacts for all this stuff using thatto basically generate a series ofprompts which we then fed into I thenfed into lovable to create the firstversion. the first version as you canimagine absolutely terrible and verypurple but like just just to kind oflike start to build the foundationsright and so I almost thought about thatfirst iteration if you will ofmatchmaker as like what I almost calllike kind of like a build wireframeright it's like it allows you to do thecore functionality and engineers havebeen doing this for a long time by theway right this is not new but you'rebasically building the skeleton of theflow you have a homepage the homepage isgoing to have recent listings on itright it's going to have maybe anavigation that has a notification areaif you have like a chat from anotherbrand, right? Like this. We had all theelements. It was ugly. It was notintuitive. Half of it didn't work, butlike it was there. And that's I guesswhere the word vibe comes from and vibecoding. Like I I slowly started to builda design system in Figma at the time. Iknow some folks aren't using Figmaanymore, but like I found it very usefulto build, let's call it 5 to eighthighfidelity mocks of what I wanted theexperience to look like. I created avery very humble design system isprobably

00:33:49an exaggeration. A stickersheet of components of like, you know,button button hover state, right? Likemy type scale, color palette, few otherthings, right? And I literally just hadlovable ingest that ingest the screensand say update across the siteeverything to these new high fidelityvisual designs that I made. And by theway, the reason I did it this way againgoing back to my, you know, old schooldesign process is that we didn'tnecessarily know what the flow shouldbe.>> Yeah.>> Like we're almost like prototyping it tosort of feel what it looks like in thatkind of like lowfidelity wireframestage. We tore some stuff down. We had afew iterations in that kind of like uglystage to just make sure that the it feltright in terms of the AR architectureand then then we kind of went into likeokay now let's make this look awesome.now that it looks awesome. I mean, we weare living in the world of high fidelityfrom now on, right? Like we have adesign system, we have a type scale. Mywife is like launching features withoutme all the time. Like she like literallycan like create entire new pages and itlooks on brand. It uses all the rightthings and even today like lovable somuch better than a year ago that likealso makes like really good designdecisions too.>> That's good.>> You know, so I've been just reallyimpressed even seeing how it's evolvedum over the

00:35:05past. But I'll just say wekind of got to a certain point in termsof where we are where it is now fullyfunctional, not to mention anincorporated business. Um, it takespayments, we have real users, we havereal listings, but we're still very muchin like the early stages. And what Imean by that is, and people are talkingabout this, I think a lot on social now,because anyone can build something fromscratch, vibe code something fromscratch. We are all finding thatbuilding software isn't as hard as itused to be, but growing an audience isas hard as it always has been.>> 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. Oh, I think I want to tapinto your experience now having quite afew reps under your belt in terms ofactually vibe coding building software

00:36:20with AI. Not only the little one-offlowfidelity prototypes internally, butlike literally going 0 to one on aproduct and building that throughdifferent phases and rebuilding it. AndI would imagine that you've picked upsome lessons and experiences along theway that have kind of shaped the waythat you think about interfacing with AIand getting the most out of thistechnology. I'm curious if you're ableto see clear ways that your own processas a builder has evolved or maybe justdifferent ways that you approach thetools or anything that all that peopleagain who maybe have been kind of on thesidelines a little bit could learn from.>> Yeah, I think a lot to unpack there. Ithink one, you know, I I did, as youmentioned, like I had built a bunch ofstuff since then. Everything fromanother like SAS tool that I'm workingon that may or may not launch. It's justsomething I'm experimenting on that'slike almost like helping people lookingfor jobs and coaches in the designindustry. So, just for comparison, likethe first version of Levelable probablytook me a few months. Getting to adecent version of this took me probablya day.>> Whoa.>> Right. Like some of it has to do withthe power of these tools, but some of itto your point is about like just being alittle smarter now that I know how theAI thinks and works. I've also likelodged a bunch of like experimental artstuff. Again, like

00:37:35I think there's a fewthings I'd call out. I think one is, youknow, thinking about how you plan for,you know, a project. And I think this isa skill set that designers and PMsshould already have to some extent,especially if you're on the more seniorside, which is like all the usualexercises still help of like reallyclearly defining like what is the goalof this product? Who is it for? What arethe critical user journeys that thisproduct is supposed to handle? Like evenwalking through, you know, and this isall in words, right? Like was writingwriting this stuff down in like a doc.I've even seen some people like dictateusing a microphone if that works foryou. But just just getting it down inlike raw ugly ways. And what I'veactually been doing is like taking thatlike initial like call it a strategy/pdkind of hybrid thing that I've written.I will drop that into Claude andbasically say, "Hey Claude, I ambuilding something that's blah blahblah, right? I'll literally dump thatugly typoridden sort of PRD into it andsay like make me a series of prompts anda plan for how to most effectively buildthis in and then I'll say the platformin lovable in cloud code in whateverbecause I mean these tools now are soself-aware of themselves and of theseplatforms that they can actually

00:38:51understand like how lovable builds stuffand so they will Claude will literallyspit out like an entire plan. It willask you questions about your plan andidentify gaps in your thinking, which isamazing. Like you've probably seen thisas well, but like the Claude interfacehas really evolved to like literallyhave inline interactions of like, hey, Inoticed you wanted uh a notion of abrand, like what what constitutes abrand, you know, like you literally oror do you want me to make this brandpublic or private or both? like it'slike>> so there's there's just like it's a lotsmarter and it can almost like help youa ton with the pre-planning. So I thinkthat's one area that's been helpful forme to kind of get that first build outthe door. The second thing that so farhas proven true to me I'm curious if youhave a different opinion on this is thatI've generally found that when you tryto build everything at once with anyplatform it's kind of a disaster in away. I prefer to kind of build things inpieces and so sometimes Claude willrecommend that. Sometimes I will askClaude about that. Like let's actuallythank you Claude for your plan. Now walkme through a plan on how to basicallycreate a road map or stagger this thisbuild out>> right and so you know as an example

00:40:06youknow with matchmaker like maybe we'llstart with a few of the key screens.We're not going to mess with settingsyet. We're not going to mess with chatyet. We're right. So we're likeliterally building out things in phaseslike you would a sprint. Each sprintmight be 10 minutes instead of twoweeks, but you're still doing things insessions, right? Where you're kind ofbuilding a thing, checking it, makingsure you're not burning through athousand credits doing it. And likeyou're really like like kind of doingthis in pieces. And as you're doingthat, you and the agent are building amutual understanding of the product andof each other, which is reallyinteresting. It starts to understandyou. By the way, something I forgot tomention, like all that initial context,that PRD and all that stuff, I thinksome platforms like Claude are muchbetter at like long-term context.lovable hasn't been, but there areplaces within lovable when you can savecontext. And so like that is something Ialso have done is like make sure there'sa place that's sort of like the projectmemory, right? Like kind of your storedcontext for like what this thing is thatit can keep referencing to understandlike, oh wait, I'm building amatchmaking service for brands, notwhatever some email platform, right?Like it's kind of like>> that context seems to really help uhthese platforms quite a bit. I actuallythink I have a Substack post on this. Idon't even remember. I think I wrotelike 12 points on like how to savetokens, but like asking it to be lesswordy and more concise, right? Likeasking it to sort of like

00:41:22check itswork. Uh visuals, I mean, this is whereI feel like there's a big divide indesign right now. I mean, you'll you'llhave to pry Figma from like my coldhands. It might happen eventually, butlike I just love working visually. I'm aI'm a I'm a show versus tell person, andthat's even the way I lead teams. LikeI'll sketch out an idea I have insteadof just like using words. And so likewhen something is off in Claude, I couldback and forth with it in words for like10 minutes or I could go to Figma andsketch it out in 5 minutes and it'sprobably going to be closer to like thepixel perfect version I have in my head.And so I will very often kind of likedual task between AI building and Figmatools like at the same time, you know,like it'll build something and I'll belike, "Oh, like that spacing is like alittle bit off and I might guess move itup six pixels." Or I might like if it'ssomething bigger than that, I mightliterally just design that littlecomponent or a part of a component andbe like, "No, this is how I want theavatar to look next to the name of thebrand and the favorite button." Right?And so I think that's just somethingthat everyone's kind of got their ownworkflows, but like for me the designtools have still been such a helpfulkind of stop gap because AI does seem todo a lot of stuff on its own inspiredby, you know, like the millions

00:42:37of otherdesigners it's being trained off of. Andso I guess that's the last part of thistoo is that like, you know, I think wejoke about the purpleness of AI, but Ido think that you have a lot ofnon-designers right now. They're lookingat the outputs of AI tools and they'relike, "Oh, that's pretty good." But likea lot of them, as we both know, lookvery similar. And that is becausethey're literally all being trained onthe same world of stuff. And so I thinkthe last thing I just really try toencourage people is like don't wait tilllike the very end to think aboutbranding and differentiation. Likereally try to like I mean in generallike I think once you know what theproduct is and what it's for. I feellike that is a conversation and somedefinition you should have as early aspossible to start to infuse some of thatsoul and some of that like kind of likebrand and point of view into theproduct. That's been my experience toowhere if you have the nugget of what youwant the visual language nailed down inthe beginning, Claude's really good atextrapolating it and like yeah, maybesometimes it'll go a little too far andyou're like, "Okay, no, no, no, I don'twant that." But and maybe in that sense,I will design it in the canvas, butevery once in a while it'll do somethingbased off of my initial brand languagewhere I'm like, "Oh, actually, I kind oflike that. Maybe I would do that overhere, you know, and it almost helps mescale my own thinking." And I was kindof chuckling with you talking about the

00:43:52the canvas tools too because same thingwas like I'm only in a canvas and thenall of a sudden I'm like oh my gosh Ican code now and then the pendulum swungall the way and now I'm definitelysettling into the middle ground where Ifeel like a lot of the strategy andskill around being a designer today isknowing when to reach for differenttypes of tools and I'm having a lot offun being creative in that process.them. Before I let you go, I I have two,you know, total out in left fieldhypotheticals that I want to toss yourway just to make sure that we've coveredeverything. And the first goes back tomore like team building stuff. And I Ikind of want you to imagine that you aremaybe it's even in your current rolelike you get to build a new 0ero to onedesign org within the broader team andyou get to do it from scratch. newheadcount, everything.>> How has AI impacted the way that youthink about team building roledefinition or or maybe even like thetypes of skills or traits that you wantto prioritize on that new work.>> I I really do believe that, you know,now because every function has access tothe same tools, everyone can buildsomething that looks like a product. Youknow, I think still the things that

00:45:07wecan do better than AI for now is one ourbread and butter as designers, which islike how do you think about what problemam I trying to solve, right? Like whatuser problem am I trying to solve? Whatbusiness am I trying to solve? And look,what is the experience that best suitsthat problem? Some people just wouldcall that more generally productthinking. But I think that has alwaysbeen a very strong criteria for me forsenior and above designers. I think thatis something that would be anon-negotiable for designers in 2026.You know, you can't just be execution.You can't just be order takers. Not thatI think designers ever should be, butlike that is no longer valuable becausea product manager can doesn't need youanymore to do that one. Two, which wetalked about as well a little bit. So,so I think that's one kind of skill set.And I think like scrappy research isdefinitely something that I think isstill a valuable trait for that. Likeevery design, every firm I've everworked for, every company I've everworked for has always be under sourceand research. So I still think thatdesigners that like have the theboldness and the courage to like findways to like do scrappy cafe testing orlike user testing.com or whatever it isand know what that looks like, know whenit's useful, and like just do it. Like Ithink that's that's huge. Um

00:46:22the secondone, you know, we talked about this alittle bit is like that point of view,right? Like some people call it taste. Ifind that a little pretentious, but Iwill I will call it brand voice, right?Brand positioning. Like, how do youdifferentiate in a sea of tools that arealready coming out and they look exactlyidentical because they're all beinggenerated by the same agents basicallyor or code? Like, how do youdifferentiate especially kind of in theentrepreneurial world of startups or01s? it's most relevant. But thatability to like swing between productdesign and a bit of like branding>> and to some extent visual design as wellis I think is is very important. Youknow, it's those moments of delight.This is something I really tried to dowith matchmaker. You know, like youbuild products. I know you can feel thisempathy like when you're buildingproducts. I have found myself much moreforgiving of visual design because I'mlike it just has to work. It's been sohard to just get to work like well I'mnot going to worry about every littlepixel, right? But actually like once Iget it to work I really do find well youknow what like there what's stopping acompetitor from like literally v codingthe same exact thing tomorrow.>> Yeah>> it'll take them a day or two right it'snot that. So like finding the customersnumber one like like real growth stuffand then two like having a reallydifferentiating product right that feelsmore premium has an emotional

00:47:37connectionwith its users like those are thingsthat fall squarely in the design world.So I think those are two skills sort ofnon-obvious skill sets maybe three thereproduct thinking research and sort ofbranding that I think are extremelypowerful. The last one I'll say ispeople that have a somewhat technicalbackground are not afraid to look at aterminal and be like what in god is thisyou know and like be and I I even saylike read reading the output of cloudcode and what it's doing andunderstanding it actually makes you amore successful builder. And so I thinkthat's another sort of like persona thatI'm looking for a little bit or peoplethat maybe studied a little bit ofcomputer science>> or people that people have beendesigners that have coded a little bitor have been design engineers.>> I I think that those skills even thoughtechnically yeah AI does it all for us.I think those skills are really helpfulbecause I have certainly found myselfwhere I'm like, "Okay, lovable, I knowwhat you're doing." Like you're stuck inlike a for loop or something. Likeyou're stuck in sort of this world whereit's just like recursing over some weirdthing and I know how to help it debugitself. And I really do think that islike a muscle I've built as like a minorin CS back in the day that like I can'texplain how my brain has somehow likeapplied it to this world, but it stilldoes. And so, you

00:48:52know, I would say likemy dream design team right now is, andagain, I think this is why hiring is sohard. It's like we are asking a lot ofpeople, but it's like that T-shapedgeneralist that maybe has a few of thoseareas. Now, as a leader, I've alwaysbelieved T-shaped means that it is yourresponsibility as a leader at a companyto understand what people's strengthsare, what their passions are, and tolike know how to grow people in theareas to make them well-rounded. So likeI certainly am not that person that willonly hire people that are checking offevery box on my hiring list, right?Depending on the role, I might have acertain non-negotiable like consumermobile if that's like a big part of myworld, right? Like if you've just donelike enterprise SAS your whole career,maybe a little trickier depending. But,you know, if someone's never doneresearch before, I will train you orI'll connect you with a researcher andlike pair you up and like give youmoments to safely start to feelcomfortable in those moments. Right? Ifyou're not, you know, as much of aproduct thinker, visual designer likethis gets a little harder like, youknow, this I think just kind of comesthrough visual design to some extent Ithink is maybe education orself-education. And then um probablythinking honestly I just think isexperience to a lot of extent. A lot ofwhat you're saying even I'm

00:50:08kind ofnoticing in playing this role connectingI guess my own sort of matchmaker youknow kind of thing a lot of brands anddesigners and it's super lifegiving forme but something that I'm noticing isthere's a real big uptick in demand formore visual brand designers. A lot ofpeople are using the phrase brandengineers because they want you to bemore>> like technical. They want you to adoptthe builder mindset but be the personthat can really push what the visualidentity of this product is forward. AndI think my my understanding is that itis coming from this>> need to>> elevate beyond the baseline becauseeverybody's getting to the baselinereally really really quickly right now.>> And I think specifically um all that yesand I think motion design is one ofthose things that is so huge. I mean Ihave said this for a long time beforeAI. I think, you know, in a world wherea lot of product design is very similarand flat for practical reasons, like howmany times can you design a long feedwith a carousel cards, right? Likethere's a bunch of ways you can stylethat and do that, but like I reallythink a lot of brand expression todaycan come with motion design that is andI think it's a very underutilized skillby product designers. And as we allknow,>> it can be so bad when it's bad and sogood when it's

00:51:23good, right? Like I meanyou know one of my I have a very closecolleague that is uh you know real likelegit motion designer and he's just likeless is more obviously like he's like bevery subtle you know like nothing worsethan like a dancing hot dog type ofmoment on your uh>> on your UI. So I think that is that isone of those areas that again the AItools right now are not going tosurprise you with motion with an amazingmotion thing like that is just not dotoday. And so like again being able tocreative direct, art direct, createartifacts in, you know, After Effects,Rive, whatever it is to sort of likeshow and tell or to create those likebody animations like th that skill set Ithink is part of the visual set issomething that I find very hard to find.And so if you are that person and haveeverything else we described, just knowyou are in the top 1% right now.>> I've totally seen that too. Like theportfolios that resonate the most have alittle bit of motion. Like if you reallywant to separate, if you can make yourstuff move or just bring in one reallyseamless transition or even just the waythat you think about the compositions ofyour work, if those move, like even thatis a big plus one. AI is obviouslypushing a lot of this forward, but someof it too is just new tools are loweringthe floor like so much, right? Like I Iwas always so intimidated by Lah. I usedto mess around with it with AfterEffects. That's a hard pipeline, man. Tocreate something in After Effects

00:52:38andthen get it to work with L. That sucked.>> Yeah. All of a sudden now I've beenusing jitter. It's like, "Oh, this isreally easy. I can do this in 10 minutesand it's like quite awesome." Or anotherthing that is kind of been blowing mymind recently is>> given an SVG Claude's pretty freakinggood at just animating the SVG beyondwhat I thought was possible. Like theceiling for SVG animations is higherthan I thought it was. And I just didn'thave the skills in order to do it. Andall of a sudden I'm like,>> totally>> kind of back to your blackbox analogyfrom the prototyping the just checkoutexperience from like I have no conceptof this if this is possible. I'm like,is this possible? Can you even dosomething like this? I'm like, can youjust try it? And then it tries it. I'mlike, oh my god. Okay, cool. What if wedid XYZ, you know?>> Totally. And and the flip side by theway, the motion side is that everysingle place I've ever worked, theengineering skill set tends to be weakeron the motion stuff. Like you're firstof all, we're all very lucky to find afirm engineer that like loves design andcan do pixel level work. But even thoseengineers at top fang level companiesI've worked at, many of them are justlike I don't know how to work with, youknow, motion and have it be performant.Uh so again, maybe an opportunity forlike a design engineer type to like fillthat gap potentially and really kind ofmaster motion

00:53:53and code together.>> Well, it's a fun time to be a designerand everything about this conversation,I think, just points to that. It's beena real joy to hear about your journeyand then all the little experiments andthings that you're building now and justhow you're thinking about the differentroles and landscape. Wealth of knowledgeand experience. I really appreciate youcoming on and sharing it with us today,Ron. Oh,>> such a joy and such a such a great hangto to spend time with you for sure, man.