Loredana Crisan - Figma's big bets for the future of AI design

00:00:00She said, "You can't blame the computerif the music has no soul. If the musichas no soul, it's because nobody put itthere." That is the case with AI aswell. AI is a tool, but yourresponsibility is to bring the heart toit. How will AI change what people areable to imagine, and how do we evolveFigma in [music] that direction? Theidea like putting the soul in the in themusic, putting yourself into yourdesign, is the message that I it giveeveryone right now.>> [music]>> Welcome to Dive Club. My name is Rid,and this is where designers never stoplearning. [music]This week's episode is with Lord DonaCreșan, who's the Chief Design Officerat Figma. And we're going to do a deepdive into everything that launched atConfig 2026, and what it all means fordesigners.Honestly, it paints a really excitingpicture about the role that AI plays inthe future of our craft. But before weget into all of the details, I wanted todig in to Lord Dona's unique background,and the ways that it's influenced howshe thinks about design and taste.>> I'm Romanian. I grew up there, and myparents were both engineers, [music]like material engineers, pretty nerdyboth. When I was six, my mother asked meif I wanted to play piano. I

00:01:15said,[music] "Yes." And then I did not standup from the in front of the piano forsomething like 20 years. This was mostlyclassical piano, but around kind of liketeenage years, I rebelled againstclassical music by starting to prototypeand my side project at that point likejust record and experiment with house,techno, rock, and music production ingeneral. So, my first job ever was to bea sound engineer at the recordingstudio. I did that for for quite a bitof time. And honestly, what musicteaches you that translates acrosshonestly any medium is what it feelslike to be a master of your craft, howmuch time you need to put in to developthat embodied intuition. We talk a lotabout expertise right now, aboutjudgement, about taste, but people don'trealize how much of that is justwhat you sense like in your body. It'snot your brain. It's how you see. It'swhat your eyes actually lock into. Andso that was the same for me with music.My My year learning how to hear soundsin a certain way. My fingers learned howto express sounds that was only in myhead. And like creating that system foryourself where you feel in

00:02:30completecontrol of your tool is what reallybrings out your heart, your soul intolike what you're producing. This is whyI love tools in general. Like my firsttool was a keyboard. It evolved tosomething a lot more complicated likethe SSL J console with like many manyfaders and buttons left and gear thatwas in a different room that you had tolike make sure that you get soundthrough. And then very interestingly Iwas in San Francisco at that time. And alot of the customers that were coming inthat could the clients of the studiowere working at startups. They wouldtalk about startups and technology andthe Silicon Valley bug sort of creepedin. And at some point I had to decidewhether I would go to LA and like reallyfocus on my music career as a musicproducer, you know, foreveror do this crazier thing which is join astartup and see where else it could leadme. And so my first role at a SanFrancisco startup was to help themprototype audio interfaces for anassistant-like experience. Thetechnology wasn't there at the time, butthat allowed me to design in my mediumin sound. It was something that was verynew and very comfortable at the sametime. Now I've repeated this story many

00:03:45times throughout my career and like Igenerally when somebody asked me for acareer advice around expansive thinkingand like taking on more, it's alwaysabout like have some something thatyou're really good at and then pick upthe adjacent area that could help youcraft something that's even moremeaningful. So, I remember being at thatstartup and we did a a user research umstudy with an external vendor and thatperson asked, "Who designed thisexperience?" And we all kind of lookedat each other because it was a littlebit unclear. And then I realized that Idesigned it and>> Cool.>> design was not in my vocabulary at thatpoint. It wasn't even a job that I wasfully aware of. I was just makingsomething and what's really cool aboutit is that we're still all makingsomething. We're talking a lot aboutlike what design is, how roles areblurring, etc. Just make something.>> Real quick message and then we can jumpback into it. If you still haven't triedpaper, even though it kind of feels likeeverybody on this podcast is talkingabout it lately, they just made it wayeasier to get started because you cannow copy and paste from Figma directlyinto Paper's canvas. All of the layersand their properties, including imagesand SVGs, are imported

00:05:00in an instant.It's pretty big deal and you can try itout today. Just head to dive.club/paperto get started. Framer 3.0 just launchedand it is exactly how I want to buildpro websites. You can get the power ofworking with AI agents, but directly ina canvas [music] that is tailor-made forshipping production sites. For example,I have like five different CMSs fordive, which isn't exactly my favoritething to deal with, but now I can havethe Framer agent do everything [music]related to CMS management directlyinside of Framer. And of course, you getSEO, staging links, localization, andeverything else that you'd expect. Buthonestly, at the end of the day, thething that I love most is just beingable to explore, design, and build withagents all inside [music] of Framer. Andif you head to dive.club/framer,you can get 30% off Framer Pro to start[music] building today. Now, on to theepisode. I think a lot about what thelittle interests and skills that I couldadd to my own repertoire to makesomething that is truly me, makes peoplefeel something. And for me, actually,the thing that almost is like my littlemini frontier that I want to explorenext is sound design. So, my secret hopeis that your DNA spills

00:06:15over to the restof the org, and that when we have theConfig 2027 release, we're talking aboutyour sound design panel that you getadded to the file.>> Well, I would love that so much.>> I want to talk a little bit about tastebecause I appreciate some of theparallels that you're drawing. And inorder to get at more of what youbelieve, I'm going to temporarily put onmy design Twitter skeptic hat a littlebit that sometimes I wear in theseepisodes. Andthe first is that well, okay, if we lookat your background as somebody, I mean,you 20 years of reps playing piano is iscrazy. Like, I took probably 10 or 11years of piano, and just it's anunimaginable amount of practice andscales and drills and everything. Andwith AI, it's almost like we'recompressing the speed at which we canget those reps by an order of magnitudeor more. And so, one sentiment might be,well, taste is actually not defensiblebecauseeverybody's getting these reps soquickly that your experience mattersless.>> Reallygoodperformance expertise requires time.Nobody, even somebody with likeincredible aptitude or, you know,biological,

00:07:31uh, you know, aptitude towards a taskstill needs to put in the work to get tothe NBA, to be a brilliant pianist, tolike be an incredible designer. Andtaste is a shorthand for like thisperson has it and they're miraculouslysomehow were born with it. It's veryvery much not true. And the questionbecomes to me is like how do you get toit if it requires a lot of time investedand often times it's just the love ofwhat you do.Are you drawn to it every single day? Isit when you have a free hour, what doyou do with it? Do you scroll some likesocial network or do you try to learnsomething else about your craft? Do youlove making more than you love likesitting and watching something? And sofor me taste equals care from thatperspective and the experiences that werecognize as tasteful, things that werecognize as good, they actually areexperiences that are just very veryintentional. I sometimes compare designto throwing a party. And if you thinkabout like the best party that you'veever showed up for,that person, whoever threw it, thoughtabout everything. You came into the doorand there was a place to put your coat.The music was just the right volume. Theflow of the house was like perfectly

00:08:47prepared. The food considered there aremany people and like they have likedifferent restrictions. All of thatstuff is care.It's taking the time to anticipate, tosimulate, to consider how other peoplemight feel, to put yourself in otherpeople's shoes. And yes, like we sayit's taste at the end, but it's justhard work.>> Okay, I'm going to keep pushing on thisa little bit because I love the partymetaphor.The where I think potentially it mightstart eroding in the long run isthere's no equivalent to copying andpasting a party, right? Like it's alwaysgoing to require a level of manualeffort, but this world that we're kindof sprinting at from all angles is youcan make something that's absolutelybeautiful. It's rooted in your personalpassion and experience and then somebodycan just replicate it and it's harder tostay unique when you have these trendcycles collapsing because of what ourtools are making possible. How do youreconcile that with your thoughts aroundlike the value of taste in the long run?>> I actually think that what happens wheneveryone has the power to build is isactually it elevates the need to standout. You know, another metaphor that Ioften think about because I believe inthis responsibility of the designer

00:10:02tocreate something that bakes in theirpoint of view and something that feelsauthentic to people. You could buy thosepretty funny or you know, soulfulcongratulatory cards, birthday cards,whatever cards and they have the messagewritten inside and it's a great message.Do you just give somebody that card? Youcould, but that's that's not what you doif you're a good friend. That's not howyou show up. And so I think that themore software is getting created, themore articles are being written.Basically, we're in a place whereeverything is getting produced at suchhigh high rate. And we we alreadyconsuming what was already there. And sowhat we're going to try to do is likelook for even more authenticity. I don'tknow if youif you've come across this, but whenpeople realize that something isdeveloped by AI only AI. So again, AI isa tool in the process to help you createwhat's in your head very much a fan of.AI is the outputvery much kind of down on. So whenpeople realize that what's being put infront of them is inauthentic and it'sjust created by AI, they disengage.Research says this very very clearly. Sothis is why I think it's not it's nottaste, it's like just your point ofview, your human intuition. Like whatyou want

00:11:17for that experience like needsto come through and then the tools thatwe build are very much like important toto help you drive that.>> I love that and it even ties back towhat you were starting with with themusic analogies because it's like thereThere this moment in time with thesynth, right? where people just wrote itoff. It's like, I you can't make thatwith your hands, therefore it is notart. And yet, it allowed people toexpress a sound or an idea that maybewas in their head that it was previouslyotherwise not possible. Like, when AI isused that way, it's beautiful. Then it'san instrument.>> There is a beautiful quote by Björk onon this because you're right, the theworld of music was like, absolutely not.These are not real likeabout them. She was saying somethingthat really stuck with me where shesaid, you can't blame the computer ifthe music has no soul. If the music hasno soul, it's because nobody put itthere. That is the case with AI as well.AI is a tool, but your responsibility isto bring the heart to it.>> Okay, let's talk a little bit about howyou guys brought the heart into what youshipped with config.>> I believe that technology changes theworld mostly by changing what we're ableto imagine. So, if you think aboutpatterns in time, like, you know,writing was initially

00:12:32created to helpcapture who owed money to who. But, inreality, where we took it as with ourhuman imagination was all the way toShakespeare. Or, if you considerphotography, it was invented to capturereality. And now, think about whatevercrazy output comes out of, you know,Nano Banana Pro because like we'vedecided that, you know, the world thatis already here is interesting, but theone that we imagine is even more so. Thequestion for us with Figma at thismoment was like, how will AI change whatpeople are able to imagine? And how dowe evolve Figma in that direction? So,it's really really important for us togive people the tools to dream andexpand what they're able to do. And so,um the set of releases that we we landedon are very AI forward in um in oneregard, but also very very muchexpressive on the other. and so this iswhy it probably feels like they're theyspan the gamut but in in reality youhave a design agent that is increasinglyable to hand off control to you. Youhave the ability to use code as acreative medium. Now that is anexpansion that allows designers toexpress themselves using like more

00:13:48interactivity, data as as behavior andship their vision all the way toproduction. That was always a dream wecould not do it and you have moreexpression with motion as an examplesomething that helps your product standout in a world that is increasinglyuniformitized by just like what themodels can produce.>> I want to zoom in on the word controlbecause that's something that I've beenthinking a lot about and I think it isthe answer to another chunk of audienceand design community that views AI asalmost the giving up of control in manyways because they're used to everysingle thing being directly manipulablein a canvas and anything else feels theopposite of that and what I reallyappreciate about this set of releases isthis idea of okay I can kind of build myown tools. I can achieve the outcomethat is in my head and actually givemyself the parameters to do so. So I'mwondering if maybe you can unpack alittle bit around like okay given thattechnology what are not only some of thethings that you are now capable ofimagining in terms of what's possiblebut what are the behaviors that you'reeven seeing internally at Figma in termsof how you're making the most of thisincreased level of control andcreativity.

00:15:03>> Basically I think the premise for forsome of our releases as an example thatthe shader effects that you could promptyourself and and the tools the premiseis that AI could get you to 70% this istrue with motion you could prompt likeyour your frame and then you will getthe first animation set up but then fromthere you need to make it yours. So,it's almost like AI sets up theworkspace. It gives you the parameters,but then you are the the one that likemakes that thing, right? And so, it'sreally important to me to get away moveaway from AI gives me something and thenI try to describe how it should bedifferent. That part is actually a veryslow. I talk about me, you know,dictating a painting over the phone.It's very, very hard. Like I haven'twritten down a shape that I I, you know,have in my head and I could easilymanipulate it myself to like get to whatI want. It's almost impossible todescribe. If there's a color that I wantto put in my design, the color wheelwill never be replaced by a prompt to,you know, so it there are moments whereyou just want to take take controlyourself. So, what we're doing withshader effects as an example, we'reallowing you to prompt the effect thatyou want, be very clear with whatparameters you might want. The AI willsurprise you by giving you, you know,extra, but you go back and forth

00:16:18justdeveloping kind of the tool and thelayer at the same time. And once youhave it, then you could actually tweakall of the variables in it. So, if it'smotion, you could like decide the speedof it or the responsiveness. If it's agradient, for example, that like createsthat shader, you could tweak all of itsvalues. And it's a it's a interestingway of working because increasingly asdesigners, I believe, we're we'recreating systems, not just just screens.Uh this is very true with with Weavus aswell and probably we'll get to it. Butat that point, that layer has the toolthat created it and you're able to shareit and tweak it with all of your all thepeople on the team. And this one of thereally important parts of design aswell. Design is never done alone. Designis evolved through critique. It's it's aplace where, you know, one of the toolsof design is other people. And so, it'sreally fun to be able to have AI haveagents have the direct manipulation anddirect control and your team in onespace.>> You said an interesting phrase, creatingsystems not screens. I'm wondering if wecan double click on that and I'mactually curious like given that shift,how does that change what signals youwould look for in

00:17:33new designers that youwant to hire?>> Okay, so the first part, systems notscreens. I'll give you an example withWeave. Weave is a design product that weacquired last year and what it allowsyou to do is on an infinite canvas use anode-based interface to string togetherinputs from you, prompts to the model,any visual model, whether it's likevideo or or image, and several variablesto create something that really feels asunique as you are. You could take atexture from one one image and, youknow, the object from another anddatabase of, you know, specifics thatyou'd want to implement those two andlike you create a system at that pointand what the output for that is isactually a mini app. We're actuallybringing these into Figma as well atConfig. And so these mini apps willallow you or anyone to just, you know,add the inputs that would generate theoutput, but the underneath of it, all ofthe special sauce, is in the workflow.And so that is a system no matter whatthe input is, the output will considerand match from a brand perspective, froman intentional perspective, what you'vedesigned in like in the guts of the ofthe workflow.>> And even thinking about it from thestandpoint of like somebody was leading

00:18:49the design of prompting tools andplugins that give you custom parameters,kind of thing. Yeah, like you thinkabout the components maybe that you'remaking available to the model, but likethat's a really loose grip on what thefinal output is. And so can we just go alittle bit deeper in terms of what doesit look like to thrive as a designerwhen that's the thing that you areresponsible for. It's not this finitelist of screens and flows.>> Yes, you have to zoom to the level of wewe actually talk about this whole MDconcept at at Figma a lot. And that isjust like, can you understand theintent, the idea behind what's beingcreated such that you could enable notjust more people on your team, butagents to actually build with with thatintent as well. And the reality, youasked me what's what types of designerswe're we're looking for. On the onehand, we want the people that really areexploring the tools right now andthey're not afraid uh to to trulyunderstand what's out there.Uh but I wouldn't say, you know, I'mAI-pilled and like I will look fornothing but how AI can do this for me,but actually the people that are able tofind those intersections because inreality, any tool is nuanced. If

00:20:04back topainting, you know, like if you're a5-year-old and I give you somewatercolors, you're going to get 10colors and like three paint brushes. Andthat is all you need, that's all youcould do. But if you're a painter, youryour set of tools will be very verydifferent and nobody could tell you thatthis big brush is so good like you coulduse it for anything you want. Inreality, you have to have range. So, welook at people that have range, butthey're willing to expand it becauseright now a lot more tools are cominginto our arsenal and we need to becurious and like understand how we usethem. Outside of tools, I think it'sthis system thinking and this idea oflike, can I distill what's importanthere? Is there a way to extract this sothat it becomes replicable and like thatis in the interest of making AI aprecision tool. Right now, it's verymuch not so. If you ask AI the samequestion or give it the same task, willimplement it in in very different wayslike each time you ask. And you canwield a tool that does something verydifferent each time you ask it to you.Generally, you'd want the hammer tobehave the same in the sameevery time you use it because you wantto understand and like learn how towield it yourself and all of thesethings that we're doing are in the inthe service of precision and wedesigners that understand that.>> I would assume there's probably a goodamount of

00:21:19research that you all areobviously doing at all times, butprobably leading up to some of thesestrategic bets you're you're trying tounderstand the shape and activitieswithin different teams and maybe toborrow your paintbrush analogy, whenyou're looking at these teams andtalking to different orgs and trying tounderstand how things are changing andwhat how people are working, what arethe behaviors or you know, what are thethe types of paintbrushes even thatyou're noticing in these teams thatyou're latching on to that kind ofgive you the ability to imagine, youknow, where things are headed and whatexactly you all should be building for.>> Yeah, there's so much. I would say thateveryone is experimenting and we get somany questions about like Figma, what isthe workflow that we should anchor on?Like how should this evolve? And youhave various organizations are atvarious levels adopting the newtechnologies and like really settling onon new workflows. A lot of what we'rehearing is on the one hand excitementabout what's possible now and then thatis almost always paired with the butXYZ and like the XYZ ends up being likewhat everyone can build. It's reallyhard to decide or

00:22:34even review all of thethings that are coming in. Review is abottleneck across design, acrossdevelopment right now and it's quiteinteresting because as humans weactually love making and we don't lovereviewing quite as much. But we're nowin a place where we have to becomeexpert reviewers and probably figure outhow to automate those reviews as well.But this is where I use Figma a lot as aprogrammatic canvas as an example likewhere I, you know, visualize differentthings in order to bring people in andlike that collaboration to speed up thereviews. We can go into some of thoseworkflows in a bit. So, that's one ofthe things that we're hearing. The Thesecond thing that we're hearing is fromdesigners sometimes, they're theythey'll say, "I feel like I have themandate to transform, to like, you know,use AI for for everything. And sometimesI don't feel like my leadership fullyunderstands that AI is not always thesolution for everything that they want.Even though I love AI and like, youknow, it's like it's part of my toolset. I want to expand and like be ableto to use everything that I want. So,everyone's very excited when we tellthem that AI is coming into Figmabecause it removes that, you know, ideathat I could I have to be in this tooland it like do certain things a certainway.

00:23:50Um so, it's that's that's anotherpart of like, you know, that nuance oflike how you use these systems togetheris starting to appear. Synthesis and anddata is really, really interesting.Like, the ability to unify what's beingsaid in Slack with, you know, your Asanatasks, with with your design workflows,with what you ship to prod. That levelof thinking really opens up theopportunity for designers in particular,like bring in Hey Marvin is an[clears throat] example and like, youknow, understand how your users use yoursystem what and what they're, you know,maybe want differently from you. That'sone area that that people are excited toexplore. If I was going topeel back a little bit the the thingsthat are really, really importantis can you enable more synthesis, betterdata sources, just like a lot morecontext to flow into your system. Thesecond thing is how much power do youhave to mold things in the way that youwant to and like with speed, but alsoprecision. And then, what do we do withall of this? Like, how do we buildtogether when everyone's kind of like ontheir own with, you know, building in ina certain direction, often alone? Likehow do we pull all of this material tomake something that feels considered andlike moves us forward not not justcreate more.>> That's

00:25:05kind of where I'm excited to seeyour continued take is it does feel likeokay, I've been given superpowers withAI specifically in like the code piecebut I can kind of make I feel like I canmake anything, you know, but everythinghappens locally in a silo on my machineand it's a very big departure as adesigner from previously where it feltlike I could just hop in to anybody'sfile. So I'm interested to see like inwhat ways these recent releases lay thefoundation for enhanced collaborationand in the feeling like okay, I'mworking together with people and notjust coding all day.>> And we really believe that you know,code is a medium and you should be ableto like being code and explore and likedata is really important as well andyour team should be there. So we want tomake all of the coding workflows reallyvisual. We want to give you all of thedesign tools that you need in the momentand we want to bring your team on thecanvas as well. If you think about it,you know, like the computer started withMS-DOS as well and like, you know, wedid not stop there and we didn't stopthere for good reason because as peoplewere visual we want to touch, we want tomanipulate and these systems are not aseasy to bring into alignment but it'spossible and we're we're very much onour way.>> Can we talk about the review bottleneckpiece? I'm sure a lot of people

00:26:20listening are experiencing that rightnow. How are you all thinking aboutsolving that right? Even just in the waythat you all like work internally likehow do you approach this problem that Ithink a lot of us are facing?>> This probably happens in every companyright now but you go in and buildsomething in code and you have a greatidea and like the way to get it out isoften, you know, record this Loom video.There are a lot of Loom videos and Iwould love to watch all of thembut they're when they'reit will take some time. So the questionbecomes what's a better way to actuallynot lose any of these ideas because theideas are awesome [clears throat] andideas are really awesome when somebodybuilds on them and you start like seeingthe the system and you bring themtogether and then ship something that'slike absolutely incredible. And so,visualizing them on the canvas whereeveryone is is really step one. And so,what we're able to do with code layerson the canvas is bring your prototypeonto the canvas and then from there youactually could simply tell your agent tovisualize all of the states of the app.You could ask it to visualize the flowof the app if you wanted to also see thethe diagram of it. You could run a skillthat maybe looks at

00:27:35it through the pointof view ofnew user. There's so many tools that youcould bring into that review processthat take you out of I just must consumethe 7-minute loom and like decide whatfeedback I might give or for the nextstep, but have a a surface where youpull it apart, you visualize it withother people, you give comments on thespecific screen that like actually needsa build up on the idea and you make thereview process not only a lot more funbecause now it's joint, but you make thenext steps very, very clear and youclear the backlog in that way.>> Yeah, I've seen a lot of people buildinglike custom tools to go from prototypeto even just taking out the screens andbringing it back to the canvas. I Idon't know. I'm sure that maybe you'veencountered that in your own research,too, where it's like, "Man, it felt likesurely surely there's a better way to dothat." Like we felt like we had thischoice between interactivity and thebird's-eye view of something and gettingboth of those for free kind of thingfeels like inevitable. Like you ofcourse it has to be that way.>> That's exactly what we're doing andthat's why we think an infinite canvasis the is the answer as a thinking spacebecause you don't just want to look atone mode, you don't want to be in linearflow, and you want the opportunity for adiscontinuous innovation, right?

00:28:50As anexample, we're building we're oftenbuilding Figma in Figma, but we're alsobuilding all sorts of apps that are, youknow, not products that we want torelease in the world, but where we putourselves in the shoes of somebodydeveloping a mobile app for, you know,finance as an example. And so, when youstart looking at it looking at code inthe same way that you might look at adesign screen, then your instinct is to,you know, what if we could do it this adifferent way? So, you duplicate thatcode layer on the canvas, and all of thesudden you could put in like a differenttype of interactivity on like, you know,the keyboard buttons of that calculatoror what whatever it is that you arebuilding. And then somebody else comesin and says, well, actually we tweak itin a in a different direction. It mightfeel even more playful. And now you havethat same process. The medium haschanged underneath, and that'sabsolutely perfect. That's not a problemat all, but the the type of thinking andthe type of collaboration that you couldget to is the one that feels familiar tous and [music] fun.>> 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 Talent[music] Network, and it's working. Likeright now, I'm helping many of the mostexciting startups

00:30:05that 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.>> I remember being at Config, I guess itwould have beentwo or probably three years ago, whenyou released dev mode. And in some waysit felt like, yeah, there are bridgesbetween code and design, but it was alsolike the deepening of the trenchesbetween roles a lot of times. And noweverything that you're saying is likecode is just a medium that is availableto designers and everything's gettingreally blurry and I guess I'm kind ofwondering like do you have any thoughtson this sentiment that you know we'reall just going to become builders andthe lines drawn between designers andengineers in the long run just kind offade away and everybody's working withcode and who is what role anymore. Doyou have any thoughts there?>> I do. I think it's happening and what'sreally interesting is that everyone iselevating uh up the stack in some way.Design is having a an identity crisisand a renaissance at the same time.>> So well said.>> We'll release a um study that we didwhere we asked about the importance ofdesign. This is something that we'vebeen doing for a while

00:31:21and what's reallyinteresting is that developers are themost likely ones to say that design ismore important and that they are doingmore design. So we are sometimes myopicfrom from our identity as designers aswe're like doing more coding about likehow much more design is happening. Likedevelopers are already like you know,you work with a great developer. Theyhave a design mind. And like what'sreally important for them right now islike not only to check the code. I thinkit's important check the code but alsolike participate in what we're buildingand why. The roles are blurring.Expertise still very very much matters.Of course I would I would say that adesigner should feel empowered to taketheir idea all the way to like front endproduction and like make sure that theinteraction that they had in mind is nottranslated through a document into youknow, back into the product if they'reable to do that themselves. Like we shipPRs all the time.>> I ship PRs.>> Really?>> It's quite interesting cuz you get intothe who should stamp my PRs question. Uh>> Oh, that's funny. What kind of PRs doyou ship? What what what Give me anexample.>> Eval tools. I I specialize in eval toolsthese days.Um and that's where I work with a teamthat you know, is in our eval We couldtalk about evals more. I I it's reallyreally important

00:32:36for designers toembrace eval's because that is how youcreate the system. That's yourdefinition of good. And then um bugs,like, you know, simple things on ourproperties. So, like, if I see somethingmisaligned or, you know, there there'san instance where a hover state actuallywas instantiated with um the kind of newbadge, but the new badge was hidden cuzthere was nothing new there. So, like,it was just misaligned and it took verylittle time to actually fix that, butthen I ran into who should style my PR.And so, I've decided to um ask the VPengineer responsible for for that areato do it because uh it's only fair. Andyeah, it was it was a pretty smoothissue.>> [laughter]>> Like, one of the variables that is in myhead when I think about the future andwhere this is all going and how fastit's changing is like, what percentageof people in an org are actively in thecode base, you know? Does everybody haveit on their machine? Are we allcollaborating code? Is everybodyshipping to production? Or are designersjust working in the medium because it'smore realistic and we have, you know,playground repos like crazy andeverything's this like 80/20, 90/10.Like, how what are you all seeing?>> Both are important and we are doingboth. Prototyping playgrounds can speedyou up because you don't have some ofthe code base dependencies

00:33:51um and youyou might just like not run into as manyissues just like running the code andlike making sure that you could and ifyour intent is not necessarily to shipthat, you're still exploring, like,speed is really, really important atthat point. When you want precision, youwant to be in the code base. There's nopoint like, you know, trying to refinean interaction in your prototypingplayground and then like just gotranslate it over. So, you need to knowin what what stage you are. It's justkind of like uh we tend to bereductionist and say, you should onlywork this way, but in reality, anythingfrom a sketch on your desk all the wayto like production is a tool and mediumthat you have need to have the nuance toknow how to use.>> Okay, let's talk about eval's. I'm notthinking about that near enough,apparently. So, why is that a big dealfor designers?>> Increasingly, we're working withnon-deterministic systems, right? So, alot of what the experience is isproduced by an LLM. Previously, it wouldbe could be produced by an algorithm. Iworked at Meta and can tell you we'vehad a lot of conversations about likethe buttons that would exist on a feedstory. Those buttons are are important,but when you think about what peoplereally experience when they open one ofthose products, it's the content. Andso, all of the conversations that we hadabout buttons as as designers were farfar

00:35:06less important than theconversations that we should have hadabout the algorithm itself, right? Andwhen it comes to these non-deterministicsystems, how do you ensure that likethey they meet your quality bar? There'sthis thing called the e-vows. It's onthe one hand the least sexy thing in theworld because you are now in a placewhere you're trying to specify what goodlooks like and have systems that couldeither recognize that themselves or, youknow, be augmented with with humans thatcould do that. And it's fascinating totry to do this on an unverifiable domainlike design because design is not, youknow, there's not a correct answer todesign, but there are many incorrectanswers. And so, we certainly dive deepin there as we consider how all of ouragents should build and design. Andhonestly, if we care about theexperience, that's where we should spendour time.>> It feels like a lot of the new areasthat we as designers are feelingI'll say pressure, honestly, to to likegrow in and explore are all prettytechnical. And I'm just wondering likeis that just the shift? Like is thatjust where this is heading? We're allgoing to become a little bit engineerywhere even brand is like I'm creatingthesesemi-technical workflows

00:36:21and that, youknow, is that just where this is going?>> I think a lot of it is going in thatdirection, but also there will still beplenty of moments for craft and like forsweating the details in the way that youmight want to. I welcome designersentering some of these technical areasbecause we have a bar for usability thatI just can't wait to bring to all of thetechnical systems that maybe did not umcrave it out way as much. And this isone of the things that we're doing withcode as an example. We wanted to feellike a medium that that's under yourcontrol, not like a technical thing thatyou're just banging your head againstthe wall and like fighting. It's reallyimportant for good work to be in theflow. So, what what does it look likewhen the product that you're using,whether it's like technical, it's aboute-vows and, you know, all of the otherstuff, but it's actually a joy to use. Ibuilt a a little e-vow game. Uh we e-vowon the canvas, of course, where youcould try to beat an auto judge at likefinding design issues. There are so manyways to like all of these things likefun.>> Let's zoom out of the app for a second.So, you been at Figma I think around 10months, which is probably my assumptionlike a really short time, but also 10months is an eternity nowadays. And

00:37:37you're brought in externally, you havethis this long background at Meta.And my assumption is you're havingconversations with existing leadershiparound, you know, what you want toinstill in the team, why me kind ofthing. So, maybe we could talk at thatlevel for a little bit. Like, what aresome of the core ideas around how designorgs can and should operate that you'rehoping to see more of at Figma?>> Honestly, we have incredible talent atFigma, and I think as a leader a lot ofwhat I want to spend my day on isfiguring out how to pull all of that upand like, you know, remove the barriersfrom them literally building what's intheir heads. So, um some some of thethings that AI is creating as challengesbecome entirely virtuosical. Like thiswhole way we're talking about, you know,the review cycles and like how do um youun- you un- clamp that pipeline, etc.That's one thing. The second thing islike you want to make these toolsavailable, you want to inspire everyone,you want to like make sure thatcollaboration happens. And then maybefrom a Figma perspective, the thing thatI consistently say is findintersections. I really, really believein, you know, the magic that happens atthe intersection of things that used tobe disconnected, whether it's design andcode, sound and visual, you know, PM andand design. You've seen me

00:38:52like hoveracross all of those. And then withinFigma itself, FigJam and slides, slidesand design, you know, code layers andthe canvas, there is so much more thatyou could do when these things areinteroperable. And that requires likelooking at teams and like projects in indifferent ways.>> So you you have almost a decade at Meta,now almost a year at Figma. I wouldimagine there's certain things thatyou're noticing that are just like alittle bit different, right?Different shape of an org, differenttype of people. So what are some of theunique ways that the design org operatesthat have stood out to you so far?>> The joy of building and how much we nerdout at Figma over like the details, theinteractions, like you know, the joy ofbuilding a tool is uh is very, veryapparent at Figma, but not just acrossdesign. Like all of the disciplines haveit. Like literally, you know, down toour like HR and sales departments. Likeall of us are makers. And that just likecreates a culture of most heartwarminginner nerds. You know, like coming towork together. That's the thing thatstands out. Everyone is makingsomething. It's constantly creating,constantly thinking about how the toolsshould evolve. Where at Meta, uh therewas a lot of

00:40:07thought about, you know,the people on a global scale and like,you know, what content is important andlike, you know, elements of culture thatare truly key to developing the next setof products. Uh so, the apertures are abit different.>> You talked about how designers arehaving to think a little bit moreholistically, like product strategy iskind of a something that everybody holdson to. And maybe we can pull on thatstrand a little bit because you guys arein an interesting place where you have aheck of a foundation,and yet you're adding so many differenttypes of features, some of them comingfrom acquisitions. I think there'sprobably multiple different ways thatyou could have sliced up the what liveswhere, and what's the relationshipbetween all these tools, and howinterconnected are they, and how doesthat fit into like the broader platformstrategy? Put us in the room on some ofthe core product strategy debates thatwere leading up to Config. What weresome of the things that you guys werereally wrestling with?>> A number of things, like some very smallthat you'd expect us to talk about, andsome that are broader around like howthe products, exactly as you said, areare coming together, and like what isthe the right way to to build. Forexample, with code layers on on canvas,that is a big decision. It changes whatthe design canvas can do.

00:41:22With animationon canvas, you always have to thinkabout how these things compose with eachother, and you know, like should youhave an animation, would you be able tolike build it straight on the canvas? Asa result of us introducing the codelayers on canvas, and like yes, theanswer is like you need to be able to doall of these things. We're just removingthe distance between all of theseactions that were already happening, butin a very sequenced way before, right?And so, it used to be that you'd go toone place to like design what you wantedto do, and that resulted in an artifact,and then that artifact would become abrief row for a design that happened ina tool, and the design would become anartifact that again would gettranslated, and like each of these stepshad a translation cost, and moreimportantly, [clears throat]it's you you don't have the quick loopback to try a different idea, right? Andso, what we're looking at is like how doyou bring it all together? But, thatobviously creates a lot of uh strategicquestions about, you know, what eachproduct is about, like how they shouldinter-operably talk about that um quitea bit, and there's a lot more thatyou'll see from us in in that direction.And then we have the small big decisionsaround, you know, like agents on goingcanvas, and like how should they berepresented,

00:42:38and like, you know, shouldthey feel like people on your team, orshould they be, you know, more in thebackground like working and not takingtoo much attention? Those are the thesole questions that that come up thenwhere, you know, there's no rightanswer, and like there's a lot of let'sbuild something, see see what it feelslike, let's pull back, let's buildagain. Like it takes a little bit oftime to get to something that we feeleven internally. Thankfully, we're bothbuilding Figma in Figma, so we're alwaysin the tool. And we feel that it's itrepresents like the the right balancebetween bringing AI into the foregroundas it builds with you, and allowing youto really stay focused on the on thetask and the humans around you, too.>> And so, let's imagine that I'm a designlead on one of these features that has ahorizontal element to it becauseeverything is kind of touchingeverything when you're putting a bunchof functionality inside of the canvas,and there's this constant iteration,exploring, and just making and buildingthings. Andwhat are the behaviors that you wouldhope to see as the design leader interms of how I'm not only bringingpeople along for the ride, but likehelping move the actual design anddecision-making forward given thesecompressed timelines?>> On the one hand,

00:43:53we used to say designis so important because it makes thingstangible, right? You could talk aboutsomething for a long time, and like youhave the risk of talking over each otherbecause what you imagine in your head isvery different than another person'slike imagination. So, a is worth athousand words exactly for that reason,and a prototype or like, you know,appearing staging around a direction isis worth even more words. And so, a lotof time I think the what I love fordesigners to do is like try to makethings concrete, be very clear with likethe the trade-off space, that their jobis to explore directions, to considerall of the possibilities as much aspossible. This is actually where Figmais a great tool to actually understandwhat variation could be in place thatyou might not have considered. So,elevate all of the tension points. Uh,that's one of the important things sothat we could have the most productiveconversations about what matters andmove the the solution forward. So, thoseare the, you know, understand the systemreally well, understand the trade-offs,expose the trade-offs, show ways inwhich we could we could work throughthose as an example, and bring everyoneon the same page. And like, with that, Ithink at least the direction part of it

00:45:08is taken care of. The The second thingthat I expect is, you know, just theirpoint of view, like what do they theythink is the right thing to do? Where istheir conviction? And thankfully, we wehave that in in space in Sigma. The as Isaid, in Siza is just like absolutelycrushing.>> Points back to the value systemsthinking again for me, where again, cuzlike everything's kind of touchingeverything else. And so, being able tosee that full picture and then how thatreflects in the set of micro decisionsthat you're making on a Tuesday, that'sa is a whole skill just in that fordesigners.>> I think you're right. Absolutely. And ifyou think about it, like we live in anincreasingly complex world, and softwarehas gotten incredibly complex itself.We're building global products thatpeople experience in like many differentincarnations. Some of them read left toright, some of them read right to left.That's not even like the most mostcomplex thing that we need to thinkabout when we build software. So, whenyou're building the tool that needs toexpress software for like the entirepopulation, you really need to figureout how to read between the lines, likefigure out what thoughts connect, andsimplify things that allow you to takethings forward because complexity iskind of the enemy of action. And thisway design like is is increasinglyimportant here, too.

00:46:23>> Before I let you go, I want to tap intoyour perspective as somebody who ispretty frequently hiring and thinkingabout the different types and shapes ofdesigners. And so, we've talked abouthow design gives you a lot out of thebox, right? So, like let's imagine thatsomebody who's listening is in a stackof portfolios and applications that arejust full of eight out of 10s, nine outof 10s, you know? So, what is somethingthat you could see in either somebody'sportfolio or how they're presentingthemselves that would get you to thepoint where it's like, "Wow, that is aheck of a signal. Like they I'm nowpaying attention and this is probablysomebody that I'd want to bring onto theteam.">> It's interesting because you can havelike incredible design thinkers thatmaybe did not work on a domain thatimmediately applies to what you've done.And so, often times you actually have topull yourself back from Oh, have theyworked on a design tool specificallybecause we need somebody who understandslike just the design tool space inparticular. I think that you're actuallylooking more for a way of thinking.Again, like at the at the system level,you look for their craft. And like forme, craft and passion are so united thatoften times I look for passion more thanI look for for anything

00:47:38else becausethat gives you a number of things. Itgives you the drive to like overcomesomething that maybe doesn't go go yourway. It gives you the taste, back toyour you know, where we started. Like itgives you that way of seeing that that'sunique to you that people come toappreciate. It gives you the the sharedlanguage with the rest of the team thatis already like very much, you know,bought into into the church of craft.From that perspective, I think the thepassion like ends up carrying things.You certainly look for signals thatthey're experimenting, that they'reexpansive in their their thinking. Youlook for signals that they'recomfortable exploring technology in thestate that we're in right now. Sounddesigners are quite interested in divingdeep into the model space. We actuallyhire AI model whose job is a little bitdifferent. It's less about figuring outthe pixels and more about figuring outhow to mold that medium. So, it dependsa little bit. Like there's somespecializations that are coming aboutright now in in design where everyone isstill a product designer, but might befocused on on a different area as well.>> Everything that you're saying keepscoming back to what makes you unique asa human, both in terms of how you'reusing the tools, what you're capable ofimagining, why you're caring

00:48:54aboutsomething. I appreciate that as like athrough line through this wholeconversation.>> And isn't that beautiful though? Isn'tthat why we're here?>> I have a annoyingly hard question foryou before we leave and we can ride offinto the sunset on this, but if you had24 hours where you could create abillboard that everybody in the designcommunity would see, what would you puton that billboard and why? And I thinkgiven your position as somebody who'sshaping the future of this industry asmuch as anybody,I I don't reach for this question often,but I'm going to reach for it with youand you can take your time to thinkabout it.>> I think it'll be a version of like putyourself in in what you're making. Theidea of like putting the soul in the inthe music, putting yourself into yourdesign is the message that I'd it giveeveryone right now. That's what theworld needs.>> It's beautiful. Well, I think that's afantastic place to end this conversationand Radana, I really appreciate youcoming on today and sharing a little bitabout behind the scenes of what you allare working on, how you're thinking,what you're seeing in the industry, andalso just reinforcing this human messageat the heart of everything you'resaying. Really goes a long way, and it'sinspiring for me. I hope it's inspiring[music] for others. So, appreciate you.>> Thank you for having me, and thank youfor running this podcast. It's such aninspiration for all of us.