Luis Ouriach - How are design systems changing?

00:00:00Basically overnight, LLMs made designsystems way more important. So, howshould we think about what a system evenis in today's world?>> By now the system becomes thecenterpiece [music] of all successfulengineering work. The more we rely on atool to create our code, the more werely on a system to tell that [music]tool what to code. How is this changingthe shape of design orgs in the way thatwe collaborate as [music] a team? Thatstarting from anywhere philosophy is thenew way of working.You might [music] start a document, youmight start in a canvas, you might startin the front end. You might start in anyof those places, but ultimately you'regoing to be bouncing [music] through thetools. Welcome to Dive Club. My name isRid, and this is where designers neverstop learning.>> [music]>> Today's episode is with Louie Orriach,whose role as a design advocate at Figmameans that he's constantly helping teamsnavigate today's changing landscape,especially when it comes to designsystems. So, we're going to do a littledeep dive into all of the trends thathe's noticing today and what it allmeans for designers. [music]But first, I asked Louie to give us arundown of everything that's changed inthe world of design systems since he waslast on the show. The last two and ahalf, three years since we spoke

00:01:15aboutFigma's variables launched, everything'son the table and everything has changed.Yeah. But you still have threads of thesame. So, I'd say that there was a overindexing on the importance of tokens,getting that right, refactoring systems.That took an enormous amount of time.Some companies took a year, two years toto get to that point of refactoring.Once that felt stable, I think theindustry threw us a curveball andintroduced artificial intelligence LLMbased code generation, where peoplestarted to rethink what systems actuallywere and what their purpose was.At least those at the forefront. Andthat made me a bit scared. I think itmade a lot of people a bit scared aboutwhat that meant for their roles and whattheir daily output would like be like.Then,let's say that was a year ago, plus. Nowwe're in a position where there's alittle bit more stable in what thatworkflow looks like, and we're startingto see, probably thankfully, theimportance of documentation inconsistent andpredictable output. So, I'd say there'sbeenswings back and forth between everythingbeing important, nothing beingimportant, and now we're landing at aposition where all those things you havebeen doing are exceptionally importantfor the new workflows that we're tryingto establish.>> Yeah. Yeah, I'm smiling because I've

00:02:30also had my own little vantage pointjust interviewing people. You know, I'veinterviewed like 200 people since welast talked, something in the ballparkof that. Andthere was definitely this sentimentaround how, you know, design systems aredead, yada yada yada. And it's like,everything's about speed, design systemsjust get in the way. And the pendulumhas swung to the complete opposite end,where everyone's doubling down on designsystems because it's so clearly a driverof speed. And all the documentation theywrote is actually really valuable, it'sjust not for people anymore. Yeah, Iwould say, to be totally honest, January2025,I thought we are all toasted. We arein trouble here because our jobs are nolonger relevant. And that was just athankfully temporary feeling, well,couple of months.And like like you said, the importanceof those docs, whatever form that theywere in before, is going to be thefoundation of everyone's workflow goingforward. They might just be writtenslightly differently and for a differentformat, for a different output, butguidelines are becoming paramount foreverything we need to do. Yeah, maybe wecan just go a little bit deeper there.You said the purpose of systems has kindof shifted as a result of AI. What'sbehind that?>> A lot of people looked at systems as ateam kind of potentially in the corner,

00:03:45doing their thing, raising the bar ofquality, and making sure that thebusinesses can scale efficiently and atspeed. But now the system becomes thecenterpiece of all successfulengineering work becausethe more we rely on a tool to create ourcode, the more we rely on a system totell that tool what to code.And the less we invest in a system, themore we're going to generate justhonestly rubbish into our systems thatwill just pile up like a landfill overthe next couple of years, and somebody'sgoing to have to fix it. So, unless wenow treat those systems as a centerpiecefor change, then we're going to end upin that situation a few years where aconsultant's going to have to come inand fix all the mess. You recentlypublished an article called a generativedesign system. Can you talk a little bitabout your thinking there? This isn't anew phrase at all,but people want to automate a lot ofthings, and we're in a process nowfiguring out what can be automated, whatshould be automated. And I speak toteams who do want to automate thegeneration of their components,and speak to teams who absolutely do notwant to do that, and want to automatesome other aspects of it. It might betoken pipelines, it might be someaspects of documentation. But we'removing to a world where a lot of thiswork can

00:05:01be done via a tool or an LLM,whatever framing you want to put on it.And we have to decide as an individual,as a company, or even as a team within acompany, what are our roles actuallynow, and what are the tools taking over,and decide that as soon as we possiblycan before that decision is made for us.Real quick message and then we can jumpback into it. An even smarter version ofLovable just released. It is 71% betterat solving complex tasks, [music] whichmeans it can do more work moreautonomously. There's more intelligentplanning, there's prompt queuing, so youcan stack requests while Lovable works.And my favorite part, there's automatedtesting, which means Lovable now testsyour apps like a real user. It opens itsown browser, navigates flows, [music]investigates edge cases, and catchesbugs for you. And the best part is whenit finds a problem, it fixes it right onthe spot. [music]I genuinely believe Lovable is theeasiest way to build software today, sohead to dive.club/lovableto try the new release.You know what I'm excited about? TheGranola MCP. It allows you to connectyour [music] AI tools directly to yourGranola meeting notes. And as somebodywho's been vibe coding a lot of toolsrecently, this release is a

00:06:16pretty bigdeal, because my meeting notes are someof the most valuable context that Ihave. And now I can find specific[music] topics, pull out action items,and get any question answered based onmy meeting history. This is allavailable today, and I'm alreadybuilding with it like crazy. So, if youwant to join me, head todive.club/granola-mcp.

00:06:37That's granola-mcp.Now, onto the episode.I want to tap into your perspective assomebody who does work with a lot ofthese teams, given your position. Andyou see a lot, you see what's working,you see pitfalls to avoid. What are someof the most innovative techniques orprocesses that you're seeing teams adoptas a result of everything that'schanging with AI? What I'm seeing is alot of people feeling empowered. Andthat it could be on one side atechnically curious designer.Or on the other side, it could be adesign curious engineer.Or maybe in the other angle of thistriangle is a product manager thatdoesn't feel like they can ship. And allof these people are feeling like withinreach, they can get closer to productionor maybe just a higher fidelity. I'mvery nervous about saying everything canbe production now, because maybe that'snot the point. Maybe we can get higherfidelity prototypes that we can test andvalidate a lot faster, and still throwthem away, because that is the point ofprototype a lot of the time, is to justtest an idea.But what is stopping anybody in thisprocess from contributing to that? And Isee this internally at Figma orexternally with other teams, wherepeople are, want of a better phrase,bleeding into another person's jobdescription. And

00:07:52I think that is great.Because we are teams.And historically, whether we like it ornot,people have worked in their own silos.And you pass an engineer a file tobuild, or a product manager passes you aPRD to investigate. The question is, whyaren't we all contributing to all ofthese phaseswith expertise?It's not saying that expertise is notrequired. Absolutely, we now get anopportunity to become even biggerexperts or larger experts in our field,but we all collaborate tighter as anindividual unit, and not relying onthose silos. I know a lot of the answersfor this episode are going to depend onthe type of team. So, maybe we'll hangout in bigger company land for a littlebit first. Like if somebody's listeningand they haven't actually taken manysteps in order to empower non-engineersto reach this higher fidelity level ofprototyping, or maybe have more accessto what exists in code, like what aresome of the baseline things that teamsare doing that are taking advantage ofthe moment, that you know, just shouldbe top of mind for pretty much everybodyin today's era? This this is anon-zeitgeistyresponse, but hanging out together

00:09:07as ateam more and building rituals thatallow you to collaborate synchronously,asynchronously. That is a Slack channel,it's a document, it's a Figma file orother file that you prefer tocollaborate in, but doing that as aunit, and not saying, "I'm the productmanager, therefore I own the PRD. I'm adesigner, therefore I own the design.I'm an engineer, therefore I own thehigh fidelity version of this idea."It's a designer having access to thecode base. It's a engineer contributingto the PRD. And it's a designer but justflexing in and out of these phases ofthe process.I'm seeing so many, speaking of largecompany land, so many large enterprisesgiving themselves a very aggressive Q1target of basically reinventing theirprocess.And that is exciting, it's a littlescary. That's what we're doing now isthat design process that we thought wastrue is up for grabs because things canbe done faster as a cross-functionalteam that we which we always wanted. Wealways wanted this. This is not asurprise or anything that is kind of outof ordinary.Speaking to CTOs who are telling me thatthis is not a design leader, it's atechnical leader.

00:10:23CTOs telling me thatthey're able to work a lot moreefficiently now. And my question back iswhat are you doing with the spare time?And the response is building moreproducts. People are seeing that theiropportunity for growth is wider now andthey can make more products based onuser feedback which they can also getfaster and build better businesses. Whenprocess is up for grabs, I do think alot of teams maybe get overzealous. LikeI've even seen it from time to time. I'msure that you have as somebody who is inmore of that role. For that scenario,are there like pitfalls to avoid or orsigns or signals that would point to thefact that like hey, maybe we are youknow chasing the wrong thing? Yeah, forsure. And this is linking back to thesystems centerpiece is I've had aconcern for a while that design systemsbeing a siloed team means that peoplecan just say it's someone else's job.And to me a design system is a bar ofquality in an organization.And that's everyone's job. So the thingabout reinventing process means thatthere's an opportunity for you to sayit's somebody else's job or not. And westill have to have clear lines ofresponsibility and output expectationson each individual. The designer isstill

00:11:38doing something specific to adesigner's role. It's not oh, the PM cancontribute to prototypes now, thereforeI'm not going to make one. So we have tohave this level of acceptance of whatour rolecan do at a bare minimumplus.And that plus is defined by yourindustry, by your company, by your team,by your leadership, by your experience,and your level within the organization.The more optimistic personis looking to widen their skill set tobecome more generalist. The pessimisticperson is is freaking out in the cornerand thinking I don't know what I can donow because I can do everything. It's upto an organization to decide the bareminimum of what the expected output canbe or should be.And the designers stretch into differentangles that they feel like they'd liketo.For me, that looks like asthe the closest opportunity is to becomea better writer. As a designer, writemore documents.Because that is in my opinion the bestway to communicate an idea.You could use audio, you could recordyour voice, regardless. Writing downbefore you commit to a pixelis the best way to get buy-in for anidea. And you're not going to get theability to

00:12:53ship an idea unless you getbuy-in. So a document is still centralto that for me. Kind of a spicy take intoday's world actually.>> [laughter]>> Like the prototype is the new PRD is thejargon now.>> Well, the the thing is that as soon asyou commit to a pixel, someone's goingto have an opinion on the visual design,but you need to get it signed off thatin a in aforward-thinking organization, you haveto have a design leader, a productleader, whatever the title is, sign offon your idea before we commit to it.And the best way I've seen to be able todo that is to write a document of yourfindings and your plan. And we huddlearound it as a cross-functional team,add comments, ask for other ideas, maybeadd aa loose mock-up of your idea.But before we go and build thathigh-fidelity prototype, let's make surewe're building the right thing. And thatgoes against my instinct by the way. Myinstinct is to commit to pixels as soonas possible. I don't know. I I I doboth. I I see value in both for sure. Soif anything, I think I just appreciatethe more nuanced take cuz it's very easyto come on here and say prototypesinstead of PRDs. I I want to double downon the team structure piece for a secondand maybe I'll just toss a hypotheticalyour way. So let's say that you'rebrought in as a consultant for a let'ssay like a series B org that's growing

00:14:08and they're feeling the pain of havingbasically made no investment into thedesign system up until this point. Theydon't have any dedicated people. There'sno like dedicated squad that's workingon the design system. So given almostlike a blank canvas, what would be someof the opportunities that you wouldpursue and what types of systems orpeople would you put in place to stewardit moving forward? Firstly, I would sayback to my point about quality, qualityis everyone's job.And quality is everyone's business. Andif we don't accept or pursue that, weshould be treating ourselves as failing.And that is something I just feeldeeply. We have to have a high bar andthat is com- company dependent what thatlooks like.But if we're not pursuing that, thereshould be consequencesuh because we're going to attract lesscustomers, we're going to sell lessbusiness, and we're going to produceless revenue.So that should just be sewn intoeverything we do is what is ourdefinition of qualityand how do we get there? That isregardless of the system because I thinkthat the pursuance of quality gets youto a systembefore you've even know it.But tactically, I guess you need atechnically

00:15:24curious designer and adesign-curious engineer. And for thesystem not to be an afterthought or aside project or something you're doingon top of extra work, it should just bea thread through everything that you'reproducing as an organization. So are webuilding components? That's a system. Wedon't need to put the label on itbecause I I worry that putting a labelon something like design system canintroduce conversations we don't want tohave about budget.>> And hiring and specialism versusa front-end engineer doing a really goodjob. And a designer doing a really goodjob to me produces a system as abyproduct. When we start to scale it up,so your scenario,you're building another product. Theeasiest one is introducing dark mode.That's when you have to startsystematizing properlythings like tokens. That is aconversation we can have at that time.But by that time, because we've alreadydone a good job,we have CSS variables and we can extendthem. If by that time we have raw valuesacross our front end,we haven't done a good job and we wouldknow that as we're building it. Thatlike Slack is just not going to flyanymore. I'm feeling that even in my ownwork. Like the difference of even justputting checks in like a claw.md file

00:16:39where it's like hey, we don't use rawvalues in these places. Here's how wereach for certain types of componentsand it just is automated in like aGitHub review, you know? It's amazingactually. Like you can feel how a verysmall amount of investment into whatthis system should be or in maybe yourlanguage like where the quality barshould be and then how much the LLMshelp you continuously hit it with everysingle PR from that point. It's amazing.It really is. Let me present the otherside side of the scenario though is thatbecause we can generate anything we wantat speedand so much of the success of anearly-stage company is how much it makesthe end user feel. How much do you wantto return to you?That is incredibly important. Whicharguablymeans deviation from the system is thewin or the keybecause we need to be able to attractpeople differently every single time wetalk to them.So what does what does an infinitelyflexible brand look like? Because we'reboiling it down to the brand at thispoint.And that might mean the raw values arefine because for this permutation of thelanding page, we're experimenting.And the sooner we get into like asystematization,we're committing to something. So

00:17:54thereis definitely an argument of do tokensmatter anymore or absolutely they haveto be there from day zero because we'rethinking about scale at every moment.And we have to dance between these twoworlds of experimentation and committingto quality.Experimentation can also mean quality,but an over-tokenization of somethingcan mean it's so robust we can't deviatefrom it.And every stage of your product ororganization has to figure out where youlie on that line.And maybe because everything can begenerated and can take you half a secondto scan a code base to see raw values,does the tokenization actually help youat that point? Or are you pushing thepixels around because you can?So the world we're entering now is is araw value fine? Or if it's not, then howdo youmanage that?If a raw value is fine, let's just limitthemto what the brand feels like. And thatcan be written in a document that theguidelines that you're talking about. Wecan express what the brand should makesomeone feel like and rely on the LLM toget us there. And this is where I likedance over my own words becausein some ways, deep down I'm thinkingsystem, system, system. In other ways,I'm

00:19:09thinking all the teams I talk toevery single day who struggle with thisword adoption that people are not usingthe right thing in the right place.And that world is just going to beflipped upside down.Now, it's already happening. It'salready happening. Yeah. Even the wordsystem is weird now because like thepoint of abstracting something was tominimize the amount of knobs that youhave to turn in order to make changesthat scale across the code base, but nowit's like the pain of that hasdiminished so significantly where it'sthe difference between a 10-second LLMcall and a 25-second LLM call. Like I Idon't care, you know? It's still goingto be able to find all of the things. Solike what is a system at that point? AndI do kind of like how you're bringing itback to brand actually and like thefeeling that you're trying to evokebecause the specific way that that isconstructed, I don't know, maybe it doesmean less. No, I'm not sure. It's weird.Like I doesn't fit into my box anymore.The the challenge we have, and thisisn't a new point at all, but if we relyjust on generation of code without avery strict indication of brand feeling,we end up with all the things lookingthe same.Yeah. And honestly, that is fine for themajority of cases. But if you reallywant to take your scenario that receiveda ton of funding,we need to accelerate, that is not good

00:20:24enough anymore.So, we have to figure out what thatlooks like, what that feels like, whatthe tone of voice is, what the imageryis. And it dials us back out of what isa component to what is the company'ssystem of approaching acceleration.So, taking the label of design systemoff what we're doing is going to help.It's going tobring you out of the corner and out ofthe silo and to the center of anorganization's change. It's almost likethe baseline quality line for a product.But like if you only stayed at thatbaseline at all times, then it'sprobably going to be a boring productthat someone was to vibe code in aweekend. So, it's like given thebaseline, now where do you strategicallydeviate from the system and do thingsthat simply don't exist in the set ofrules that we're giving to the models.And there'selevations there.The person spinning something up bythemselves at a weekend might help themsecure some funding todevelop it further.That doesn't mean that thing you builtin 12 hours is what you should ship.It's just a validation point forsomething.Then you accelerate all the way up tothe top of the line where you've gotthis fully fledged tokenized systemwhich you can expand andeven acquire and merge companies intoand build out a more robust

00:21:40brand in amore horizontal way.Butyou have to figure out where you are onthis line,which won't shift, to understand whatlevel of fidelity you need at everysingle point of a product, not just thesystem, because the system feeds aproduct, product feeds the system. Yeah,it's like a lot of the value propositionof a system was in making sure thatthings are consistent and standardized,but it does kind of feel like we'regoing to be getting a lot of that out ofthe box.>> [laughter]>> So, like that's not that's not enoughanymore. Like you you have to kind offind ways to make it so that there is alevel of attachment to a product. And Idon't know, even I think Shadcn played ahuge role in this. There's so muchnetwork effects around what the modernweb stack is, and the more that we kindof spin this flywheel of likeeverybody's building with Tailwind,everybody's building with Shadcn,everybody's building on top of Vercel,the baseline is rising.>> [laughter]>> Right? So, like a good clean consistentdesign istable stakes in many ways, you know? Idon't want people to spend their timereinventing something that alreadyexists.I think that those days are limited, butmaking it feel like your brand is theway you can put the energyon top of a system that is alreadyproven.We've spent so much time trying

00:22:55tofigure out how to make accessibilitysomeone's priority or a company'spriority.And if you can get that out of the box,take it.You don't need to go and custom spin upyour ownversion of a popover component. You'vealready been sold. And that is amazing.The the quality bar, like you said, isis higher.So, let's focus on whatwill help us sell more products, andthat might not be the most attractiveproposition for the pixel-pushingdesigner, but that's the job. And that'sexciting. You can get closer to revenueand showing impact and getting closer tothat promotion that you'redesperate for because you're able to getgo better access to stakeholders. I wantto dig into workflows and kind of likethe practice of a designer. Again, justkind of tapping into your vantage pointas somebody who's talking to a lot ofteams. And the first thing I'd love yourtake on is like how wide is the gapright now between enterprise andstartups in terms of the way thatdesigners are operating in thesesystems, but also just more broadly likeapproaching their day-to-day role as adesigner, whatever the heck that stillmeans now today. It feels like it'sgetting wider.And the reason it feels like it'sgetting wider is because the access totoolsis so vastly different based on yoursecurity standards as a company.

00:24:10If you're in a legacy enterpriseorganization, it might be really hard toget an AI tool signed off.If you're a high-flying startup, you'veprobably just got these things out ofthe box.You sign the contract and you just havethese tools. So, the have and have-notsor the digital divide, something that weused to call it,in this microcosm, it seems to begetting wider. It does give me a littlebit of concern because I don't wantpeople to feellike they could be stuck in a role andnot be able to get out of it into a moreprogressive organizationif they're in one that doesn't give themaccess to tools.So, that's something that the industryneeds to figure out prontobecause that could be a potentialproblem. But how people are working, thestartups are high-flying, they've alwaysbeen doing that, they're always tryingto push the boundaries. But what I havenoticed is that people want to get toproduction a lot faster.They want to almostskip design, and that isa potentially dramatic phrase, but it'show it comes acrossto I have an idea and I want to get itlive on a domain now. What's the easiestway to do that? And that was theconversation I was having last week.And on thefor an enterprise software where you'vegot millions of users, you have to be abit more considerate about how torelease something.So, that

00:25:25then looks like how can weshorten the time to production,what tools can we use toaccelerate feedback on this? Or how canwe increase the amount of users that getaccess to a beta? Or how do we usefeature flags to ensure that a subset ofthe audience can test this thing out forus?What's the community approach to oursoftware? How can we build a team ofpeople that we can rely on forconsistent feedback because they careabout us? And that is a playbook that Ithink everybody will use eventually asyou go up the stack from startup toenterprise, but the thankfully for meand for us is that community, people,users, passion is still incrediblyimportant for the success of an idea.It's just thatwe're doing it faster, but we can't skipthe qualityno matter how fast we want to move.One thing that Dive Club has madeabundantly clear to me over the lastyear is that the practice of design ischanging. And the old process [music] ofgetting feedback just doesn't quite cutit in today's world. That's why I'mexcited to announce that Inflight isofficially in open beta. It's thefeedback tool that I've always wanted,and it's built for a world that moves atthe speed of AI. So, I can share myprototypes, give context [music] invideo walkthroughs,

00:26:40and Inflight makesit easy to get the exact feedback that Ineed to move forward, whether it'svoting on directions or maybe evengetting the green light to ship a newidea. And all of this is available in asingle link that I can [music] drop intoSlack or maybe even share with powerusers to test out a new prototype. I useInflight every day, and it's totally[music] transformed the way that I sharework. So, I'm excited for you to try theproduct, and if you ever want to jamabout it, just email me at [music]rid@inflight.co.I can't talk workflows without askingyou about [music] theClaude MCP release that you justdropped. I think it was literallyyesterday. So, what are some of thethings that that unlocks in your mindthat you're excited about? It reallydoes get me pumped for the world that Ithought I was entering as a professionala long time agowhereI have an idea,I want to test it properly.I want to get it into a browser.I want to feel it. I want to knowinteractions. I want to knowtransitions, timing. And being able togo from the canvas to the browser,test it, and be like, that's actuallynot not quite right. Let's go back.Let's riff on it. That world for anon-technical designer is going to openso many doorsbecause

00:27:56you can communicate what yourintent is so much fasterbecause you can push it further. So, alittle phrase that we kind of throwaround internally is not faster, butfurther. And that enables people whoaren't technical jump in the browser,test something out, open the inspect,say, "Oh, that spacing should actuallybe 12, it shouldn't be 10 or eight orwhatever it is." Let's jump back intoFigma and have 10 versions of this thingspinning around. Push that one to thebrowser. Push that one to the browser.Ah, still not quite right. Let's comeback via MCP or whatever we're using andhave this feedback loop for ourselveswithout involving an engineer that mighttake days or weeks to come back throughno fault of their own, just throughprocess.So, that's part of the process feelslike we could speed it up massively. Thekey to unlock even further is thesystem.If you can push from your canvas to thebrowser and back all connected up to asystem that lives in a repo,then we're going to iterate so muchfaster. That is to me the gold mine ofwhat we've been trying to do.>> same. I close my eyes and I picture likejust havingthose two windows side by side at alltimes. Like that's how I want to work.There's there's an element of yes, it isthe system and they're pointing at thesame thing, but sometimes I also justwant it to like be a whiteboard, youknow? I just want to be able to likequick scratch and like,

00:29:11"No, no, no,like kind of like this instead." Andjust immediately have that go to Claudeor Codex or wherever I'm working andfluidly moving in and out. Like it feelslike we'll look back on the last, Idon't know what it is at this time. Thetimelines are blurring in my head, butit's like the last like three to sixmonths, right? As this blip on the radarwhere we had this false choice asdesigners where you had to kind of pickthis path and it was very difficult totraverse to the other one. And as aresult, I wake up every day and I'mlike, "Am I even a designer anymore?"Because a lot of times in the startupenvironment like you're describing, likeI just open up a localhostbring an idea to life and hit hit send,you know, andall I'm doing is writing code. Am Istill a designer then? I don't know.This is the beauty that the role isgetting wider and you tap in and out ofwhatever makes sense to you. When Ifirst started and I was contributing tofront end in professionally, I stoppedbecause things got so complex to set upan environment. I rememberDocker andOh god.>> NPM install andReact and bundling and I just thoughtthat is not where I am right now.Firstly, I've got no idea what you'retalking about when you say these termsand if it takes a couple of hours for anengineer to help me set something up,we've kind of gone down

00:30:26the wrong path.So we we reverse out of that into adesigner being able to do this withoutthinking about what any of those thingsmean, the idea wins.It's not a technical [clears throat]barrier that you have to jump over andpushes people further away, which whathappened to me. Push me further awayfrom production and more into thepolishing pixels on the canvas kind ofpersona and now we can reverse out ofthat one and maybe meet somewhere in themiddlewhere the canvas is very important forideation, collaboration and generaldirection exploration, but the browseris where we commit to it.We bounce back and forth between thesedifferent mediums to get the right idea.Which I think is just exactly what wewant. I love the idea wins. Itdefinitely feels like that's where we'reheaded in in that world. I'm prettybullish on designers. Anybody with anidea is the beauty of this. I like theword maker because it kind of doesn'tdefine specifically your role in anorganization. If an engineer wants toplay around with interaction details,absolutely go for it. If a designerwants to tweak the border color on acard because they've spotted thatthere's a better contrast against thebackground in this particular screen,please go and change that token value.

00:31:42Don't feel restrained bythe process that was before. And likehow multiple times you brought it backto collaboration too because somethingthat I'm feeling a lot is like yeah, I'mempowered for sure. I'm empowered. I cando a lot, but sometimes it kind of feelslike especially in a remote company,it's just me and my agents.>> [laughter]>> No, like all I am doing is just wieldingmy agents all day and I end upcollaborating with Claude so much morethan other people.And I don't know. It's like it's one ofthose tiny little signals where I'm likehuh, this feels weird. I I I don't know.I don't know what to do with it yet. Somuch is changing. But do you thinkyou're learning? Oh, I'm learning likecrazy. I think that bouncing back andforth between you and a tool to get moresure of your idea before you commit toeverybody, what do you think about this?I don't see a problem with that.Outsourcing everything you do tosomething is where I get a little bitmore concerned.Because Yeah. we are the experts in seatto bring something to marketand if we're outsourcing our ideas, ouroutput, our definitions of anything to atool, then we just need to ask ourselveswhat we really want to be doing and Iwould feel deflatedas a maker, as a creatorif I was just pressing like go go go

00:32:57allday [laughter]on a tool. Sometimes I do. If I'm makingFigma plugin and I'm coding it, Iabsolutely don't look at that code, butif it worksI'm happy.But I still have the idea. I'm laughingbecauseyesterday basically all I did was workon front end. I use conductor now and Ibasically it's Pavlov's dog a little bitcuz you just live from one of the tootof the conductor horn to the next.>> [laughter]>> It feels like I'm playing thiswhack-a-mole game where it's like allright, that agent needs me. Okay, nowthat agent needs me. Now this agentneeds me. I feel this trap sometimeswherelike going back to what you were talkingabout earlier, it's so important to getbuy-in and to be sharing early and oftenbecause when you can play ping pong withan agent all day on your idea and go wayfurther than previously possible as adesigner where I'm sweating the detailson every single part of this front endarchitecture and all of the hover statesand interactions and everything, it'stempting to just keep refining and keeprefining versus popping out and be likehey hey, I'm look I'm working on this.Like these are the couple directionsthat I'm thinking where that came veryvery naturally when my process wasslightly more defined and always startedwith me, you know, hitting F on a canvasand drawing the rectangle, whereas now Ican jump much further as

00:34:12step one. A lotof pros and cons, I guess. I'll put itthat way. Don't get me wrong. I wasstarting code too.And I think that starting from anywherephilosophy is the new way of working.You might start a document. You mightstart in a canvas. You might start in afront end. You might start in any ofthose places, but ultimately you'regoing to be bouncing through the toolsat some point. You might start in aSlack thread to be honest.Yeah. And that will always come back toa process that goes linearly throughthese phases if you want to keep riffingand get the best possible output. Butyeah, I don't have a problem withstarting in code because sometimes youneed to feel something to know what'swrong and it can take a lot longer oftento draw out on the canvasthan to say hey generate me a web pagefor this thing. I'd be like no, that'snot right. I would say though thesetools are not for hey, make me an app orhey, make me a website. That's just nothowthey're designed and not necessarily howI think we should be using them.But boiling it down to real specificeven like component level creationversus broad please solve the world forme. Don't you think using these tools? Igot a YouTube comment today and it wassome skeptic saying like I'm yet to seean AI design

00:35:28that solved a real problem.And I'm like hold on a second here. Whatdoes an AI design mean to you, you know?Because to me that's like assuming allwe're doing is just saying make me awebsite. I was like I'm using AI to geta level of control that I've never hadbefore, you know? And bouncing in andout of tools constantly. But don't getme wrong. Thethe average person who starts arestaurant they need a website and ifthey can get that done in 15 minutes,they're going to take it and that istotally fine by me because if I can goand read the menu of the website and itlooks kind of nice, I can get thecontact details and know where they'rebased, know what the prices are. I as aconsumer also wouldn't mind that. Whenwe start to talk about software, we'reentering a much different game.And that's where make me a softwaredoesn't fly. I'm helping my buddy rightnow who's starting a local contractingbusiness and he really wanted a website.And I asked him questions like you know,what do you want on your website? He hasno answers.>> [laughter]>> He has no idea, you know? So make me awebsite is not only valuable becausemaybe the skill is not there, but alsothe ability to articulate what you wantis not there for the vast majority ofpeople who do not do design or do notoperate in software. And so being ableto just spit something out on the pageis probably the easiest

00:36:43way to evenarrive at what you actually want becauseyou can see and be like ah, I don't wantthat. I want this instead, you know? SoI do see a ton of value in that.Additionally, this isn't a new thingeither because we've had the ability forpeople to go and spend $5 on someonemaking them a website for a while andthat made more makers or more people whocared about websites in this case. Andthe way that that's the same way I lookat this now is more people are going tobe able to create their ideas.We're not in a position to tell peoplewhat they can or can't make.But as soon as that restaurant websiteneeds to open another chain or anotherstore, then they might need a betterwebsite. That's where they probablyengage in a professional to build abetter system. All right, so it's onething to talk theoretically about howthe world is changing. I want to talk alittle bit about your story thoughbecause you drew us this picture ofJanuary 2025 and this little bit, youknow, there's that nervousness aboutwhere is all of this going? What doesthis mean for my role? Obviouslyeverything has changed from that point.So can you talk a little bit about yourpersonal journey with AI and where theseinflection points are? Yeah, I I took awhile of self-reflection and uhbrewing the thought of what was about tohappen

00:37:59and came out stronger. I can't even tellyou the amount of React courses I boughtover the past 5 years.>> [laughter]>> I can remember at least three or fourtweets from you about it. You've beencalling me out on that.I haven't finished any of them because Ijustdon't have the interest or time, but Q2last year, 2025, I thought hang on asecond.I couldmake plugins as a very easy entry pointinto what I've been trying to learn.Pull up an AI tool in that early daysearly days, that's funny to say that. Inthe early days of this, it was anenterprise ChatGPT license and I wasasking it to help me make a plugin,copying code from the browser into myMhm. IDE. Like a caveman. Like a cavemanand it worked. And I thought all right,okay, we're we've got something reallypowerfulhere where I can jump over those coursesthat I've wasted money on and shipped aplugin very quickly. A week, two weeks,something like that. Then it got faster.Couple days.Then it's a couple of hours.And I think I've published maybe five orsix Figma plugins or widgets over thepast 12 months, which is kind of nutsbecause they work.

00:39:14They solve problems.There's thousands of users. People aregiving me feedback. It's open source.People can look at the GitHub repos.They can contribute it to if they want.I received a GitHub star on one of therepositories. Look at you go. What thehell's happening here? That's amazing.And I think that is what we have on thetable.Yes. The ability to make utility toolsfor our use cases at speed using thesetools. So I've gone from the ChatGPT inthe browser copying code back and forth,to a more integrated workflow where I'musing a a tool that allows you to haveLLMs based in the IDE. I'm justchatting, bouncing back and forth,rewriting, rewriting, rewriting, gettingback into Git workflows and branches andall that sort of stuff. And hosting mycode and GitHub pages for landing pages.This just opened doors that I hadpreviously closed. And I'm able to workfaster on these ideas, get them outfaster, get validation faster, and solveproblems. So, all of these plugins orwidgets have come from justconversations with the community aboutwhat they're struggling with, andspotting an opportunity to help. So, I'min effect able to do my job better aswell. Inspiring because I'm also thekind of person who has bought morecoding courses than I care to

00:40:30admit. Iwould do a whole boot camp. I don'tactually think I ever technicallygraduated as from it. But, I spent allthe money. They just took all my money.I just didn't get the the gold star atthe end.I think I actually had some giving up ofhope. There was like a couple years. Iwas like, "I'm never going to be able todo this. Like, I'm never going to beable to do this." And it going back toyour word, empowering, I guess. Like, Idon't know. I'm at the point right nowwhere, if I have any free time at all, Ijust want to build things. Like, I Ijust feel like I can just shootlightning out of my fingertips, youknow? It's amazing. So, I've beenkeeping aGoogle sheet of startup websites for acouple of years. There's probably 8 or900 websites in this Google sheet. I'vehad a goal of just putting it somewhere.And two nights ago, I opened my laptopon the couch and just chatting away andsaying, "I want to make this website.Here's my Google sheet. What can we do?"And they planned it out. We builtsomething. And then I closed the laptopand haven't done anything with it since.I can make a bunch of new stuff, and Istill don't finish them. But, I'm stillable to feel like that hour or so in theevening was really usefulto just explore. And I can explore ahigher fidelity. I used to work with aguy over a decade ago, an engineer. AndI would just come to him on a dailybasis like, "I've got a new startupidea. Let's build the Twitter for this."Or,

00:41:45"Let's build a new propertymanagement website." Whatever it was.And he'd say, "Yeah, I've got somedevelopers I've got in my contact book.Probably going to cost us 50 to 100,000to to make it." And I think, "Okay, endof the idea." And now I'm thinking,spend a couple days prompting away, andprobably realize the idea is not good.But, at least I've been able to flexthat and test it out. And maybe one ideaone day will mean something to me. I cancommit to it. But, instead of committingfinancially now, it's committing withtime. A little bit of money, like amonthly subscription. As a designer, Ithink I'm not alone in this. Sometimesyou just want to throw things at a walland feel what sticks. And probably mostof the stuff doesn't.And that's okay. But, we can do itfaster now. It also changes what thedefinition of success is, too, becausewhen software is expensive, in order tobe successful, it has to have somesemblance of scale, you know? Itwouldn't make sense for you to makesomething for your immediate group offriends. In a world where you are havingto spend the amount of time and moneyrequired to do it. But now, like, I madea Super Bowl app just for my family,just to play games together whilewatching the game. And 15 people usedit, and it was a massive success. And Iwill totally do it next year. And itagain, yeah, it took took me about 2hours. The

00:43:00key thing I think though is,just because we can make somethingdoesn't mean it has to be widespread orit has to have mass appeal, because thereality is, if you wanted to do that asa business, that's the wrong way toplace a start. Yeah.>> You need to go and market the idea nowrather than build it. And I've alsoworked on sort of startups as a sideproject where that hasn't been thefocus, where a couple of years have beensunk into building something with no onewho knows about it. And that market hasshifted. And that won't change. Justbecause we have access to the tools, wecan make things faster, we can get itinto production, doesn't mean there'sany eyeballs at all on it. And no one'sgoing to sign up to this thing. So, keepit tighter, keep it the scope smaller.And maybe something for your family isthe best place to get that itchscratched. There's a phrase that I'm ona mission to use as often as possible onthis podcast, which is the nicheeconomy. Cuz I totally think that'swhere this is headed. You know, you cancreate something that solves a problemfora thousand people. Kind of what you'redoing with your family plugins, youknow? People wouldn't look at that andbe like, "That's your startup." Or,"That's your business," you know? But,all of a sudden, like, well, what if youcould do that 20 times as a designer?You just solve any problem that you canfind, you know? So, given given all ofyour explorations, given this windyjourney of the last 18

00:44:15months for youthat probably feels like a lifetime, Iknow it does for me.How the heck are you thinking about yourcareer moving forward given where you'reat? Honestly, don't know, because I'venever known. I've never planned mycareer. I've just taken the shots thathave come up. Or, put it a differentway, I've taken a lot of shots. And someof them make the basket, and some ofthem don't. But, I would not stopshooting. And I don't know where thatends. The industry shift will settleinevitably.And things will boom up and down. AndI'm kind of just holding on and focusingon what I like to do, which is help thecommunity at scale build bettersoftware. Where does that go? Hopefully,just better software. What does my rolein that look like?Using my platform to enable people to dothat. Do I have a one-year career plan?No. Two, five, 10? Absolutely not. Sixmonths? No.>> [laughter]>> Three months? No. The amount of thingsthat are changing on a weekly basismeans that I think that it's a fruitlesstask for someone like me to try and plantheir career, because I can't evenrely on a company's road map at themoment>> [snorts]>> to know where I fitin that.And that's for everyone.So, I'd say I'm going to continue doingwhat I enjoy, pushing that as far as

00:45:30possible, putting as much pressure onthat as possible. If I'm not enjoyingit, then what's the point? Do you feelpressure though as somebody who is inmore of like an education role to keepup with everything, especially given howquickly it's moving? That is difficult,I'll be honest. It's being resigned tothe fact that I will never knoweverything. I didn't, and I will neverwill. The market game much wider is kindof crazy. The tools, the expectation ondesigners to in some places become a lotmore technical.I'm never going to be able to talk atthe level of a full-time Reactengineers. This isn't just not going tohappen. But, hopefully I can helpfacilitate a conversation between thedesigner and that engineer to build abetter workflow for them in thatcompany. So, yeah, I do feel a personalpressure, not professional pressure, tojust know what is required at whatelevation to get the work done. And I'llfigure that out. It gives me theopportunity now to lean into differentparts of the role, different parts ofthe industry, and to find newenjoyments.That to me over the last year has beengetting back into GitHub contributions.That me in the next 12 months, whoknows? I'd hate to lose touch with theimportance of writing. I'd hate to losetouch with the importance ofquality bars. Pixels are stillincredibly important. And

00:46:45personally,I've been focusing visual design quite alot. So, that's always going to bethere. This it's open. Like,I'll take my shots. I think I'm someonewho has always just been the ultra earlyadopter everywhere. Like, I was obsessedwith scrolling on Product Hunt everysingle day for years. Try every singletool. Like, it just came so naturally tome. And I think I've wrestled with thisreality that I actually can't be thatperson anymore. Like, it's impossible.There are too many tools. Too much ischanging every single week. And so, it'salmost like a failure state to try. Butthen, you have to develop new razors touse internally to figure out where do Iinvest my time? What is worth pursuing?Why is 80% of my Twitter feed open callhacks? You know, I don't even I don'teven know, you know? It's a little bitintense as somebody who is just tryingto kind of hang on for the ride, youknow? And then I'm sure everybody feelsthat, which is why I was asking thequestion because I I know you do feelsome pressure to to stay up to beat thestuff. I do events every week.And people always put their hand up andbe very honest and say, "I don't get it.I don't know what to do." And that istough to hear. But, we have theplatforms to enable people to understandwhat is possible. And I think

00:48:01that youcan easily paralyzed by limitlesspossibilities. Even what you talkedabout then about opening Twitter andseeing a specific type of content, youcan just be flooded by people seeminglydoing very similar things. I thinkthat's probably one of the reasons,other [clears throat] than the obvious,why Twitter community felt like itfailed a little bit, because peopledidn't know where to fit in or didn'tknow what to talk about, becauseeverything just felt like everyone wassaying the same thing. So, communitiesare still very important, but peopledon't know where to go to find theirlike-minded people, they're just goingto be feeling like they're alone. So,community is still incredibly importantfor people's careers satisfaction and toknow what to do. I just don't know whereit is all the time, because it feels abit fragmented. I want to try a littlebit of an out there question before Ilet you go, because I think there are alot of narratives out there right now.And some of them are probably quite safeto ignore, and they're just overblown.Some of them, there is some truth, anddesigners should kind of take itseriously. I'm wondering if anythingcomes to mind for you in terms of anarrative that there's not that muchtruth to it right now, you know? Like,you you can you can safely ignore this,whereas, you know, actually this iswhere the world is heading, and we can'tjust

00:49:16bury our heads in the sand asdesigners. I went to my parents' houseat the weekend, and I felt very strongtonal shift in dinner conversation. Myparents didn't go to university, but AIwas raised as a topic of dinnerconversation, or after-dinnerconversation, after a few glasses ofwine.And the prevailing thought, at least inmy circles,is probably negative firstwhen it comes to this stuff, but peoplewere just telling me their use cases.My brother older brother has a financebackground. His wife is a psychologist.My sister works in admin tasks that sellfor a company that sells fireextinguishers.These people are not who we are, butthey were talking about how their jobsand their lives have been made easierwith these tools.So, my prevailing narrative I think wetell ourselvesis not the mass marketednarrative. And I think that as the dustsettles, people will find real use casesthat help them every day. And that'shard for me as somebody so into it tosee because I get hit on Twitter withnegative thoughts all day.But, I can see through it to theeveryday consumer has a use case forthis.

00:50:31My friend who works in marketing,he has a personal open AI subscription.With projects, he's got tons and tons ofdocuments running all the time helpinghim immensely with his everyday life.He's got ADHD. Helps him a lot. So, Itake myself out of the negative firstmindset, which is as a natural pessimistquite difficult. My personal one of anarrative that I have but don't projectnecessarily is whether we like it or notas designers, makers, people,most software sucks.And people don't really care about thepixels. And that has been a challengefor designers, me included in previousroles, to convince people to care aboutit. But, the role now isbecause there's so much software on themarket, because there's so manywebsites, so many businesses,we have to focus on sales, we have tofocus on getting revenue, and gettingusers, and sign-ups, and acquisitions.So, the pixels are secondary to thesuccess of the business.And we can come to the pixels at a pointwhere it makes senseor not at all. As we generate morethings, more software's going to becreated, more websites going to becreated, more apps by people who don'thave a design background. We've just gotto accept thatand not try to fight it. There will beroles that you go

00:51:47for whereyour manager or manager's managerdoesn't really care about the pixels.And that is a fight not worth fighting,I don't think, at the moment. I wasthinking about my own answers. I thinkthere's a narrative right now wherepeople who maybe are slightly moreskeptical of AI saythe blurring of roles is just becausebusinesses want to pay less, and ofcourse they want you to do more, andthat's what's really happening, buteverything will settle. I don't reallybelieve that's true.I think that we havedrawn an artificial line betweena picture of the front end and theactual front end just based off of thetechnical capabilities that wereavailable to us. AndI think that the actual front end willfall under design.And that it actually is just a wholecollection of things that totally fallunder the definition of UX, but as anindustry, we've refused to admit that.And so, we've just punted it off tofront end engineers who don't reallycare about the details as much as us.But, that at some point we're just goingto realize that's just UX design, andthat's the new role. There are morebusinesses being created. The hopefulargument of smaller teams means smallerteams in more places on more problemsbeing solved. As

00:53:02a person who enjoysthat startupaccelerated pace kind of atmosphere,[snorts] that's exciting.But, for somebody who enjoys workingvery large organizations, that might bethe opposite feeling for you. But, I dosee more startups, more opportunities toto raise money and build your own ideas.Now. Yeah, I totally agree. It's likethere was a line, this is my economicbrain thinking, but you know, there wasa literal line on a graph where if youcould not generate this much profitbased off of this problem,it's not worth pursuing. It doesn't makefinancial sense to do this. And thatline has just fallen. So, there's awhole subset, there's this wholecategory of problems that previouslynobody had the financial incentive tosolve that are just wide open.Look around you, like local businesses.Oh my goodness.There's so many problems that anybodylistening to this is totally capable ofsolving on their own, fully. Take ashot.>> Take a shot. Yes, yes, I like it. Well,I I can't think of a better line to endon than take your shot. And uh Louie, Iappreciate you coming on and giving us alittle bit of like the insider scoop,but also just telling it how it isbecause I think you're feeling this likeeverybody else, you know, it's a crazytime. And sometimes it's good to just

00:54:17shine a light on the reality, andI appreciate you coming [music] on todayand doing that with us. It's always fun.Thank you very much. Pleasure as always.Before I let you go, I want to take just1 minute to run you through my favoriteproducts because I'm constantly askedwhat's in [music] my stack. Framer ishow I build websites. Genway is how I doresearch. Granola is how I take notesduring crit. [music] Jitter is how Ianimate my designs. Lovable is how Ibuild my ideas in code. Mabin is how Ifind design inspiration. Paper is how Idesign like a creative. And Raycast ismy shortcut every step of the way. Now,I've hand-selected these companies sothat I can do these episodes full-time.So, by far, the number one way tosupport the show is to check them out.You can find the full list atdive.club/partners.