 Okay, we're gonna be talking about the future of design and there's a lot that's packed into this talk. So I actually gave everyone out these notes. We'll be getting to them in just a minute. But before we get started, I want you all to do an exercise for me. Actually two exercises. So the first of these, and you can use the back of this if you want, or you can use a notebook. The first exercise is I want you to make a list of all of the things you do in a week. The activities you do. And it could be things like prepping screens for handoff, facilitating meetings, interviewing users, paper prototyping, whatever it may be. I'm gonna give you about 45 seconds just to write down some of the things you do in a typical week. And we'll come back to this exercise at the end of the talk. Okay, and exercise number two, slightly different question. I want you to think about, I want you to think about sci-fi movies, sci-fi books, fiction, maybe nonfiction, and technical problems. All right, there we go. I want you to think about what stories truth or fiction have helped you reflect on the unintended consequences of technology. And so for me, I put up some things here. Ex machina, thinking about AI, the movie Gattaca, thinking about genetics. But think about some books or movies that have helped you reflect on the unintended consequences of tech. I'll give you about 30 seconds to write something down here. And you can thank me later. A lot of us are among strangers here, and a lot of us tend to be very introverted, so I just gave you a conversation starter at the break. So you can ask each other what you wrote down in the answer to this question. All right, so let's get into the talk proper. I feel that design is in the midst of a shift, a pretty big seismic shift. And it's a shift that's gonna require that we develop a new set of skills. And I'll come out and say, I fear it's gonna require that we develop a new set of skills different than what we have today, or we may risk becoming irrelevant in the future. And so what I wanted to talk about today is what are those new skills that we need to be thinking about developing, focusing on right now? So that'll be the bulk of the talk, is what are these new skills. But before we do that, I'm gonna start by spending about 10 minutes talking about the context and the situation we're in, kind of taking a long view of things so you can understand why I'm arguing for this new set of skills. And then we'll finally wrap up at the end with what doesn't change in all this? What is it about being a designer that doesn't change? And by the way, I'll quickly add that I take a very broad view of design, like I think many of us in all professions design in one way or another, but today is really speaking to those of you who practice design, who have designer in your title. That's who I'm really speaking to today. So let's start with what is the shift that design is going through. And there have been a number of things they've got to be thinking in this direction over the years. I think one of the earliest signals of these thoughts goes back to November 2014. And I was at a conference retreat for the weekend in Tessel. It's an island north of Amsterdam about an hour or so. And it was an event I had not been to before. It was a marketing event. So a lot of people who do growth hacking or conversion sciences, things like that. So it was a whole new tribe for me, which leads to interesting, rich conversations. And it was at the end of the event, where a group of us were waiting for the fairies to take us back. Again, it was on an island. And then it was some speakers and attendees. We were chatting about things and there were three conversations that suddenly converged for me. So one was everything we've been talking about all weekend long. So web page conversion metrics, things like that. The second coming more from the tribe I'm used to of designers and web developers and stuff was design systems, pattern libraries, style guides and how those things were becoming more clear and more defined. And you have things like Google material design coming out. And I started thinking about AI and machine learning. Even though this was 2014, I believe a service called the grid had just been launched and the grid claimed to be able to build websites without any humans involved. And so I was thinking about this and we started looking and discussing these three topics and how they might converge in the near future. Something became clear to me. It won't be very long before no one would need to hire a web designer for a website, right? You can put in the algorithms, the goals, you can put in the design patterns and the AI can test out a thousand configurations until it finds out what works best for those metrics. And I said, okay, that's interesting, scary, curious, but I do more web apps and mobile apps and things like that. So I'm safe, that's software. But then I started thinking more about that and I was like, well, wait a second. What makes this all the different from web apps? I mean, web apps are more complicated, but they're not more complex. And give it time and the same thing will happen with applications and software and a lot of the things I do. Same patterns, the same AI. And so suddenly I'm thinking, wow, do I have a future anymore? Where is this all headed? So that's one signal, one thought. There are many, there are probably dozens of such stories I could share that I've collected over the last five years. But I think the best way to characterize the shift is to think of it from one where as designers we craft things, think of the potter at the clay wheel crafting things, to one where design will mean to twist knobs and tune things and fiddle with things. And I'm calling the shift, you can forgive me, I'm calling the shift from design 2.0 to design 3.0. I know you can slap me for this. But to understand design 2.0 and 3.0, I think we really need to look back and talk about what design 1.0 was first and then move through the years. So design 1.0 is what I would consider the very direct hands-on manipulation of things. So making products, making chairs, building houses, making brochures, making pamphlets. And there was a direct correlation between the work of the designer, often singular, and the work that was produced. So then if we shift to design 2.0, which is what many of us have grown up in, things are a little bit different. There's not as much direct control, there's more indirect control. So you have things like different browsers and screen sizes, we've had to wrestle with that for years. So with print you could design for 11 by 17 or A4 or whatever your print size was. With browsers you had to think about different sizes and renderings, classic problem we've had for a couple of decades. I love this photo of Josh Clark and sitting in front of all the different mobile apps, mobile sites, devices, but at some point it's not gonna be dozens or hundreds of devices, it's gonna be millions of touch points and they're not all gonna be screens. So it's not sustainable to continue designing for every break point and touch point. Plugins and browser customizations. So in design 1.0 you could control what people experience to some degree. But with browser based design, I can have a plugin that makes my experience different from anyone else's. So this is what I see when I go to Amazon. I've installed the fake spot plugin which actually will crawl through all of these customer reviews, 81 reviews and it says five stars and actually tell me most of those are fake, right? It uses AI to do that and it's pretty cool. So this is my version of Amazon. Amazon didn't build this, but this is a plugin I downloaded and I was able to tailor the experience. Dynamic content. So the whole reason we have content management systems and things like that is we're working indirectly, we're defining page archetypes that then work together to build these pages. Data management. So thinking about the data itself and I wanna call out a specific example here and think about designing search results. And I call this out because I think how we respond to a problem or how we responded to a problem like designing search results is a good indication of how prepared we are for what's coming next. So if you had a problem like design search results there's kind of two ways to respond to it. One would be the design one-oh sort of mindset which would say let's work on the information design, let's make the typography easier, let's figure out how to drive the focus to the right areas. Maybe let's play with some iconography and create some animation and create an emotional experience there. So that's one way to respond to the experience of designing search. The other is to actually say if we wanna create a great search experience I need to like dive into the behind the scenes working and dive into stuff like search queries and ontologies and metadata and keywords and understand how that works so that the experience that people have is a good one. So that's a lot of what I would call design two-oh and then more recently I might even say we have design 2.5 so that's where as the touch points have exploded things like service design are a necessity and nice to have because you have to orchestrate how these things fit together. So we have things like customer journeys and blueprints and the like. Cross-functional teamwork is a requirement now. We can't be siloed as designers and then hand stuff over the wall to engineers like we have to work together because things are more complex, they're more interrelated, they're more complicated, all these things. So cross-functional teams are a requirement. Then you have rules-based systems and what I mean by that some of you may have worked in a system where you were filling in the madlib statement sort of like this. So listen to data when an event happens in data trigger event with content. This is very common with tools like intercom when you're configuring reports and reporting but I'm seeing it show up in other areas like when people are configuring how the page will render. This is what some design teams are working with. They're working on rules to define when certain patterns or when certain components are gonna be shown to the page. And this I would say by the way is the first step towards machine learning, artificial intelligence, those things on the right, rule-based decision-making. If this happens then do this. And it's no longer the realm of engineers like designers are doing this or programming these things, determining what will happen when. And then we have things like AI-generated websites. I mentioned the grid that was launched or hyped up in 2014 but it's actually starting to become real and I think five of the leading website systems like Wix and Squarespace and others say they're all working on this and analysts say it'll be about two years before that technology reaches prime time but basically people will be able to go on non-designers, non-developers, anyone could build a website with AI and have it rendered the page and have it look great and convert well and do whatever it needs to do. So that's design 2.0 that we're living in right now and I don't know the exact words to describe what 3.0 will be but some of the things that come to mind are it's about thinking about the system behind the things. It's about how things relate so it's less about the things and more about the connections between things. It's more focused on outcomes and less on how we get there and it's more focused on facilitating structures. This is a phrase I've been using a lot lately. How do we create facilitating structures that create the conditions for certain things to happen? So going back to this analogy there's a lot of themes I could call out. We could talk about the individual versus the team things like that but there's one big theme I want to focus on which is a loss of control. That's what it can feel like is we're moving from this place of having a lot of control to having very little and if anything we can tweak knobs, tweak rules and hope that something changes. And I love how my friend Rural says it. He works at Google and has been navigating between engineering and design for the last decade. He says instead of pushing pixels you learn how to orchestrate them. They push you. I was like wow that's a brilliant way to describe like where we're headed and what's happening. So when I start to talk to people about this topic a curious thing has happened. So I would say imagine a future where we don't need someone to design the interface and I'm thinking all about this more AI driven things all these big ideas, these systems and I would get a curious response from folks. Oh do you mean voice and VR and those types of things instead? And I would scratch my head and say well not exactly that wasn't what I was thinking. And so I want to show a video clip that I think sets up kind of where I'm thinking when we talk about the future design and where things are going. This clip is from a movie that came out several years ago called Elysium. Some of you may have seen it had Matt Damon and this is one of my responses to this question that I gave you earlier. But a little bit of setup for the scene before we go to it is set in the future dystopian future as we've come to expect in movies and media. And Matt Damon's character is checking in with his parole officer. It happens to be a machine you'll see. And just watch what happens. Hello before we start I'd just like to explain. Matt DeCosta violation of penal code 2219 today at bus stop 34V. Yes that's exactly what I wanted to talk to you about. I believe there's been a misunderstanding. Immediate extension of parole by a further eight months. Wait what? No no no no no. I can explain what happened. I just made a joke. And you know. Stop talking. Elevation and heart rate detected. Would you like a bill? No. Thank you. What I'd like to do is explain what happened. Personality matrix suggests a 78.3% chance of regression to old behavior patterns. Grand death bottle. Assault with a deadly weapon. Resisting arrest. Would you like to talk to a human? No I am okay. Thank you. Are you being sarcastic and or abusive? Negative. And if you've seen the film it gets worse and escalates. I'm sharing this for a couple reasons. One, we tend to talk about sci-fi predicting the future. I think it's more accurate to say that science fiction is often a comment on the present. And so the reason a lot of us feel something and feel an emotional resonance with this scene is it feels like something we may have experienced already. So if you've ever made a call to customer support you may have had a similar experience where things were very automated and routed and there was a script. I know in the U.S. at least we're feeling a prescription of the pharmacy can be very frustrating. Dealing with the voice assistant. So we joke about how Siri and Cortana and these other things get stuff wrong but imagine when it's for something critical, right? If we look at the present capabilities of sensors then you see that this stuff is possible. We can sense people's elevated heart rates and we can calculate algorithms and probabilities. We're already doing that now and we can go on. But I feel like sci-fi is a comment on the present more than it is a prediction of the future. The other reason I shared this and this is going back to the point I was making earlier is as designers we can watch this and there's one of two ways to respond when I ask this question. What do you wanna fix about this experience? And I think there's a good number of us that when we see this we're like wow, I would love to fix that interface. So I'd love to change that creepy face, right? Or I'd love to change that robotic voice so it's more human and that's great. In fact I've dedicated 20 years to making those touch points and those interfaces more human, more emotional. I think that's critical. And you do those types of things and it might keep the situation more calm. It might not lead to the negative outcomes. But more often I see experiences like this and I feel like the real design challenge is to focus on what happens and where it leads. The entire experience, the entire exchange and the outcome. Like him getting frustrated, him getting hauled away by people. Like these are bad outcomes. Increasingly I find myself focusing more on that than the experience and the interaction along the way. And so if we go back to this I wanna reframe this in a slightly different way using someone else's model. I'm gonna change design 1.0 to product. Design 2.0 to experiences. And design 3.0 to outcomes. And let's think of it this way. For years, for many years we talked about products which are designing things that are appealing and easy to use. We've been living in this era of user experience design where we create touch points that create a desired experience. And I believe we're moving into out of necessity a period where we'll need to focus on outcomes. Designing experiences that contribute to positive societal outcomes. We're talking designing for societies in many ways and what happens. This credit for this version, this illustration comes to, goes to Cheryl Cababa. I picked this up in February at Interaction 19 in Seattle. So the big question I'm left with though is, okay, what does it mean to design for outcomes? And this is what I've been wrestling with for the past year or so. What does this actually mean? And I've formed some conclusions. One is this, that if we want to design experiences that contribute to positive societal outcomes, then I think we need to get comfortable with these two things. Designing with machine intelligence and designing for systems at scale. And there's a lot of hype around this thing on the left, the machine learning and AI and all that. I think it's the combination of both of these things that's critical. And the reason being, I think it's summed up well in this quote, never before has technology allowed individuals to do more harm or good with such low effort. And I think that's what we're seeing happen and play out with these systems that are built at scale that affect literally millions of users, sometimes billions. So if we roll the clock back a year ago, these are the questions I was wrestling with that I did not have answers to. How should designers be thinking about machine learning? How will the design profession evolve to meet 21st century demands? How do we develop our ability to think in systems and prototype possibilities? So these are the types of questions I've been musing on. And what I wanna share next is where I've landed and where I've arrived. I don't have answers to all these questions, but what I think I can offer is a frame or a model to help us all start thinking about these types of questions and what specific questions might we ask. So you have a page at your, all of you should have had a page or a handout at your chairs when you sat down. This is what we'll be walking through very quickly given the time we'll be going through this. This is an outline of the skills that I believe we need moving forward. I wanna start with the most important part and it's the thing that doesn't change in all this. And it's the thing that intentionally is at the center of this whole model. And that is the human experience. The humans at the center of this all. That is key. That doesn't change. That's always been important. That always will be important. And I think just to go back to our design 2.0 paradigm, I think we've had a nice time where we've focused on designing interfaces and interactions. And so for many years, it was a web top desktop screens. I think there's a lot of maturity here. When we have new things like voice UI which is still relatively new or VR there's a lot still yet to be done, particularly with how we use space in a virtual world. So there's still lots of work to be done at these interfaces and touch points. But increasingly, I think as designers we need to think beyond the interfaces themselves and think about all the stuff outside of it. All the stuff, the data, the algorithms that fuel this. And it's kind of scary. That's why I put in a dark background. It's unknown, right? What does this mean? What do I need to learn? And so hopefully what I'm gonna share will shed some light on that. At a high level there's really four themes that I'm gonna go through. You can see them called out here. And then there are nine skills within these four themes. So let's go through some of these. We'll start with the stuff that's got the most hype that we hear about the most often. It's the stuff on the left which is about training the engine. And let's start with some of this stuff. I wanna start by playing this clip from Breakout. This is one of the early examples. I think going from three or four years ago from DeepMind, they were acquired by Google, but they built a machine learning algorithm that they didn't program it to do anything other than learn just like humans or toddlers do. And so it's learning to play the game. And so in the beginning it's just learning, it's training itself. And again, no one programmed it how to play the game. They just said look for things, look for patterns and respond. So like the first several plays that just sat there and did nothing. It's getting better. Here's what's really interesting. I wanna go to 600 training episodes and look what the machine has taught itself how to do. It's figured out that the most efficient way to beat the game is to work on the edges and get the ball up there and then just sit back and watch. And this is a move that human players who are expert arrive at for some time or if you see someone else do it, then you pick this up. But the machine figured this out on its own. So again, no one programmed it to do this. So that's pretty amazing technology. But when I first saw this, there was one thought that came to mind. One, I was amazed, but two, I was like there was one clear goal, win, right? One clear goal. And yet in the work we do, what happens when there isn't a single clear goal? Cause I don't know about you, but oftentimes there are many competing goals on the work we do. And so what do you do there? And then I started thinking of another thing. What are we already doing? So forget machine learning and programming. What are we already doing on the teams that we're a part of to actually help define the objective or objectives and the goals for the work? And so this isn't even a tech question. Like obviously we're gonna have to program and tell a machine what the learning goals are or the training goals are, but what are we doing just within our teams to set clear objectives and key results or KPIs or whatever metrics we do? That's an opportunity right now, something we can lean into and learn to do better so that when it's critical for this machine programming, we have an edge. Let's talk about algorithms. So this is the thing that's kind of scary. I don't know, I saw a lot of math formulas like this when I would dig into different kinds of computer programming algorithms. And to be honest, even though I'm really good at math, I felt a bit like this when I was staring at these things. But I feel like if this is the material that we're working with, we need to understand the pros and cons, the weaknesses, the blind spots of different algorithms. And so I started looking for some resources to help me and Stanford D-School has identified the six types of algorithms that every designer should be aware of. So this is a project they came up with. I started reading some books. Books are really hard because the bulk of them are targeted at engineers. So it was hard to find some books that actually speak to the non-expert or the non-technical expert, but I found a few. I created a Pinterest board, as you do, with all sorts of pins and things related to machine learning and AI. And that was good for exposure to terminology, but it wasn't really good for understanding. And so I actually created a project for myself to make this a little more fun, where I started digging into all of the essays and the reports that come out about every six months about machine learning, beating some video game. And for a little context, video games are a great Petri dish for engineers to test these machine learning algorithms. And again, it's all test and priming for a larger societal application. But the first 50 or so early 80s arcade games, just simply by throwing reinforcement learning at it, these algorithms have been able to beat the games. But then you wonder, okay, well, why were there five or six games that reinforcement learning couldn't beat? Things, games like MonoZoom is Revenge or Pitfall. And the reason it couldn't beat those is there were few reliable reward signals. So if games like Pac-Man gave you an instant feedback loop, games like MonoZoom is Revenge would require that you recall and remember that that key that did nothing a few screens ago is actually important to this door or this box now. And so you have to have memory. And reinforcement learning doesn't have memory of something that happened a few milliseconds ago. And so the solutions there were imitation learning where let's let the machine watch humans and learn, which is a bit of cheating, right? Or let's do reinforcement learning but teach the machine to be curious and remember information and be intrigued by stuff that doesn't do something right away. And then of course you have things like StarCraft which is incredibly insane. It's been called one of the biggest challenges for machine learning and earlier this year there were some groups that finally solved this with supervised learning and reinforcement learning. So anyway, I'd read all these and then I'd go to my friends and know better and say, okay, explain to me the difference between reinforcement learning as it was used here versus here and what's the difference between imitation learning and supervised learning? But at least I knew sort of the terminology and the questions I needed to ask to start up leveling my understanding. So, simple challenge here. How are you learning to speak the language of data science? Again, we don't have to become programmers and understand this stuff. There are a lot of engineers who can't do what data scientists do but we need to know generally about how algorithms and work and how they fail. So finally data, I'll just keep this really brief. We have to engage with data. It's a material unto itself and we have to start with a project. Well, give me the data that's gonna be fed into these screens or this experience. And we have to do more than that. We have to ask ourselves hard questions like where did the data come from? How was it collected? What data was not collected? What data was omitted? Is this raw data or is this aggregate data? There's all sorts of stuff we need to learn there. And I'll just quote from an academic, a teacher I was interviewing. She said data is reductive and political. So by the very nature that you don't have visibility in how it was collected and you don't see what was not collected, it's a political act, the collection of data. Which I think is interesting because normally we think of data as very black and white and binary. That's the data, that's truth. Well, the collection process would say otherwise. So, simple question. What are you doing to engage with data? All right, let's shift over to the far, we'll just go through this section briefly. Here's what normally happens. You have these learning goals, this data, these algorithms that feed into an AI engine and then something happens, right? We hear about the news often, it makes the headlines because it's very bizarre, it's very biased. And it's actually hard right now for anyone to actually see what happened or understand why. And so there's an incredible opportunity moving forward to help close the loop on actually seeing what's happening. So, simple question I would ask. What are people right now? What are people doing right now with your product, with your site, like could you pull up your mobile app and tell me exactly right now what people are doing? How would you know? And the challenge there, I think most of us don't. In fact, a lot of us work in environments where we work on screens and mock-ups and maybe we hand them off to engineers to get built and some months later we'll actually look at what's live and be like, wait, that wasn't what I put in the mock, right? Can you relate to that? Like we don't often know what's actually made it to production. If you work for a smaller company, you may have better visibility into that. But my challenge would be is how do we close the loop? So at any moment you know exactly what's going on and what's happening on your website or with your web app at this very moment. What might you do to improve visibility into these real-time activities? So that's monitoring results. And again, incredible opportunity for that. I think last year was the first UI-based way to actually track this. So a group out of Austin called Argo Design built that. So it's still a wide-open space and a much-needed need. Let's talk about designing in real-time. And this is a mind-shift thing. I could talk about getting faster and I could talk about a team that, the design team that unless your idea can be thought of, coded and released in the afternoon that the company won't even entertain it. But it's not about speed necessarily, it's about the mind-shift. And so I think all of our notions of speed are rooted in the manufacturing mindset. How do we move things through a process until we hit the release button? And I think the new mindset we need to have is more like surfing. How are we in the moment responding to changes real-time as they happen? It's a completely different mindset. Think of surfing, think of being a sports game. It's different from manufacturing. And I think that's the real-time mindset that we need to identify. Finally, and this one I think is really, really critical to us, our role as designers, is we need to figure out ways to quantify the things that are intangible. So things like trust, things like joy. How do we turn those into things that a machine can quantify and factor in alongside bottom-line metrics like profit? I think that's a key thing. And as I speak to companies, I'm finding that there are teams they're working on. How do we quantify trust and make it something that's calculated in the algorithm? Because this is critical. How do we take those principles that are on the wall and actually turn them into actionable things? So how do we quantify things that are fundamentally intangible? Trust, understanding, joy? I think that's a key challenge that we can double down on. All right, let's shift down to modeling possibilities. Because so far everything I've talked about is happening in the wild, in the real world, seeing what's happening. But I think that can be incredibly dangerous and risky. And so we need safe ways to explore possibilities and see what's happening. And I think there's a whole new breed of tools still to come that don't exist today. But I think we see hints of what will come in things like this. Generative design, which I'll talk about, computer modeling, simulations, explorable explanations, which I'll talk about in video games. So let me just pick a few of these. Generative design, a quick show of hands. How many of you've heard this phrase, generative design? Okay, a good number of you. So this is a part that was built by a human, designed by a human engineer. This is the same part designed by a machine with human sort of governance or intervention. It actually supports the same workload and same stress and everything, but the two options on the right here use half as much materials and weigh half as much. So far better in terms of efficiency and use of materials designed by a computer. It's not just about physical manufacturing though. This is, from a presentation I attended, also at Interactions, it's an architecture space and urban design. So here's the traditional way that things were done where you had to focus on all the variables and all the considerations in order to make the recommendation, the solution right or the three options. Well, the new way we're working with or some architects are working with is the computer does all the modeling, looks at all the possibilities, far more than a human could, and then spits out tens of thousands of options. And so the new role of a human in this case is to actually curate and sift through possible solutions rather than try to figure out, do the synthesis and come up with the solutions. So the way the speaker here characterized it was we're just pushing complexity from upstream problem solving to this downstream sort of curation selection process. So question there, something to think about is how might machine learning with generative design tools alter or improve the work you do? So think about machine learning embedded in sketch or Figma or these things and how might that change what you're doing as a designer? Explorable explanations, this is something else I am super excited about and thrilled about. Nicky Case is leading the way with a lot of these and what Nicky is doing is taking things like papers written in the 1970s on desegregation and turning them into playful sort of simulations very visual, very interactive where you can engage with these really difficult concepts very easily. So in this case, this is based on a game in a prize winning paper on segregation, what happens? And Nicky introduces just three rules from the paper but what you can do is you can adjust those three rules as you see on the right, play it out over time and see if you end up with a segregated society or not. Incredibly complex ideas but he's making them incredibly approachable, something you can play with. So imagine then not just something with three variables but tens of thousands of variables. Imagine what we might do with modeling to figure out ways to solve problems like climate crisis or ending poverty. What kinds of modeling are needed as problems get more complex and tricky? So let's shift to the top then. This is where we get into complex adaptive systems, systems thinking, these types of things. I've been talking about systems thinking as a phrase generally for years but it's only in the last year that I'm starting to double down and learn some of the tools and techniques that come from formal systems thinking training and I'll share a bit of those but just I'll start by saying one thing I've learned over the years is problems are rarely straightforward. Like you may be assigned to this team or this team, this project team, right? And you're working and before long you meet someone from this team and realize that you overlap, right? You're both working on similar things and you're also overlapping with this other team and by the way, you're all part of a bigger project. Things naturally are complex and there's lots of overlap. So how this has affected my thinking. I've started shifting from talking about human-centered design to humanity-centered design. I've started shifting my thinking from how might we types of things and what if to why should we and at what cost given the scale that we're working at. I've shifted from happy paths to edges and exceptions. These edges and exceptions also can be, you might say that's only like half a percent but half a percent of millions of people could be a lot of people affected. From big changes, I used to be all about big changes to small tweaks, let's make tweaks so we know what's actually catching hold, what's changing, what's not. So these are some of the results of looking at things as a system, some of the shifts in my thinking. Things I've learned, simple rules give rise to complex behaviors. So this has seen a nature in the flocking patterns of birds. There's really only three rules that guide all this complex flying flocking behaviors. Simple things like the personal Kanban. Two rules, visualize your work, limit work and progress. But I think it's precisely because there are only two rules that this has become the most popular personal management system in the world. So simple rules, double loop learning is another thing I've learned about. So very simply and very quickly, double loop learning might say, or single loop learning might say we observe our current customers, we assess possible corrections, we develop new strategies and we implement new actions, rinse and repeat. And that's all rooted in the idea that we can change within the existing framework. What double loop learning says is well, maybe we should go back and assess the current structure, the mental model, the vision, go to these deeper things, develop new structures, mental models and then implement new actions based on that. And so essentially you're changing the framework itself. And this is what we've done for years as designers when we do problem reframing or we do five wise, but this is a more structured and analytical way to talk about those very same ideas. And when we talk about porting this to machines and machine learning, this kind of structured way to say are we solving the right problem? I think translates better. So we have all sorts of complex problems getting teams of teams to work together, business models that are anything but straightforward. And we really need tools that are suited for the complexities of the 21st century. So that's another thing I'm looking for. So tools I've been collecting are things like polarity maps and polarity mapping to facilitate conversations where there shouldn't be a decision, but there should be understanding at the end. Things like consolidated flow models where you have lots of actors involved and you can't double down on one. Things like wordly maps and understanding the system or what we can build on. Things like back casting where you can look at possible futures and work backwards and ask the question, how did we get here? Very different from forecasting where you ask how will we get there. So I've been looking at tools like this that are more suited for 21st century problems. So there's lots there. We can also talk about designing for long-term effects. There's tools like the tarot cards of tech to ask us questions like what happens or who or what disappears if the product's successful. Other good questions like that. What would using your product too much look like? All right, and these are asking us questions about where things might go. So went through that very quickly. All of this is online if you wanna reach out or look for this stuff. So you can go to links for everything I referenced. I want to end by focusing on what doesn't change in all this. So with so much focus and much changing, it's good to remind ourselves of what doesn't change. If we go back to that first exercise, there's kind of something I've been doing with this. I've been taking the list that groups come up with and asking them to run their list through two meat grinders. Poor analogy I know. And the first meat grinders remove anything at all related to screen interfaces. And here's the scenario. The demand for screen design has disappeared. Design systems and decent machine learning algorithms combined with robust metrics driven business and marketing tools have automated most of the work currently done by hand. So I would remove anything related to screens. And that usually catches a bunch of things. And then the next one catches even more scenario too. Remove anything that could be done reasonably well. Not great, but reasonably well by any other member of your team. Non-designers, non-researchers. So here's the scenario. Research, customer journeys, copywriting. These are all things that can now be done reasonably well by most members of a cross-functional team. Set aside any concerns, however legitimate about expertise and quality. If good enough from a non-trained expert is sufficient, then cut it from your list. So when I've been doing this, it's been interesting to focus on what's left. But I think that's the point of the exercise is to look at what's left and start to ask, what does it mean to be a designer? And you may have heard me mention some of these things throughout this talk. I think what doesn't change in all this is a mindset. And so things I've been doubling down on are things like framing and reframing problems, working from principles and values, thinking in systems and context, focusing on human needs and motivations, seeing possible futures where other people see present realities, thriving on ambiguity and complexity. I think these types of things, and there are more, are what we do as designers, are the value we bring to teams. I think this is the stuff that we need to double down on. So I wanna close with one final example, something we can all relate to, let's typography, raise your hand if you love good type, good typography, all right, good, I got most of everyone. I love great typeface, but I started asking myself the five whys. Why is this? Why is typography so important? And my questioning went something like this, well, one, I hate this kind of stuff, right, bad kerning? Yeah, it just, it drives me nuts. So I say, I like good type because it just, it feels better, right? It's more aesthetically pleasing. Well, why is it important? Why is aesthetically pleasing more important? Well, I like things that have balance and harmony and beauty. Well, why is harmony and beauty important to you? Well, I like a world that's nice to live in. I care about the world, I care about the world we live in, we create. And when I boil it down to that, caring, I think that gets at the heart of what's really important in all this. We care about making the world a better place. And so whether it's fixing bad typography, whether it's fixing creepy interfaces, or whether it's engaging with really scary algorithms, I think the core is the same, which is we care about making the world a better place. And that's what doesn't change in all this. That, thank you very much.