A chat with Yamilah Atallah - a staff designer at Instagram
What it really means to be a designer?
Yamilah Atallah is a staff designer at Instagram, currently working on Stories (previously, Instagram Direct, the messaging service within Instagram). Previously, she worked at Twitter, where she worked on a few different products: Twitter Blue (Twitter’s consumer subscription service), Twitter Communities (”group”-like functionality for Twitter), and also an enterprise offering for Twitter’s wide array of enterprise customers. Before that, she worked at numerous startups ranging in size from eight to several thousand employees.
She writes about design stuff, ranging from technical matters to advice for pre-career designers, on her Medium, and offers tips on her LinkedIn, as well.
Being a designer myself, this conversation stems from my concerns regarding AI. Initially, I was apprehensive when I started researching AI's impact on design. But now, after having interacted with Yamilah, I have concluded that my concerns were irrational. If you are in a similar state, I am confident this conversation will settle your dilemma.
Yamilah and I cover a number of topics—from how she got into design to the future of AI’s impact on design. My hope is that you will enjoy the breadth of topics covered, as well as a personal and detailed look behind the scenes at how designers work at companies of all sizes.
How did you get into product design?
Yamilah - I didn’t even know what product design was (it was called “UX/UI design” at the time!) when I got into the industry. I was in a really bad financial situation—I’d stupidly decided to borrow money for living costs to attend my university, though I’d received a full tuition scholarship. My family was dirt poor (I couldn’t even afford a one-way plane ticket to the school that I accepted the offer from). But I really didn’t need to borrow money after that first year once I’d gotten established—I’d figured out living, and was working full time while studying. But I was really uneducated about finances and money, and I allowed myself to accrue a ton of debt—way higher than the US national average. It was totally my fault.
At some point during my second year, I’d realised that I may have destroyed my life by making the decision I’d made to take on this amount of debt, and all for a university degree. I looked seriously at the job options (for the money those universities charge, I’d expect those universities would provide this service for students without asking, but okay) based on architecture, and I knew paying back my loans wouldn’t be possible on those salaries—even assuming the best case scenario (that I could become a top architect). I knew it was going to be a very hard life, even in the best situation. I was scared. I switched to graphic design, as I always liked it more anyway (but simply refused to accept it, as I felt it “wasn’t respectable” enough—hilarious, given their compensation is actually higher than that of architects’, on average). But I still knew that the debt to income ratio would not pencil out.
I started looking for any jobs. I put together a little static site using HTML, CSS, and a little JS. I put up all my school work. At some point, I started applying to hundreds of jobs. I wanted anything, had no idea what career I would even have, and figured that I could just apply for film roles (I had been doing film work on the side), graphic design roles, as well as computer related stuff (I could put together websites and, with my graphic design experience, I could make it look better than a Wordpress template, I guess), and etc. I found no opportunities. I had no idea what to even do.
Eventually, after hundreds of applications at a ton of different jobs, I got two responses, and one offered me a job. It was a 5 hour daily commute to some random place far out in Massachusetts at $25/hour, which was more than anyone probably in the history of my entire lineage has ever made. Obviously, I took the job. This was around 2015, so they asked me to work in Sketch. But had no idea what that even was. It was really similar to Illustrator, but less cluttered, so it was easy to pick up using a 20 minute YouTube tutorial. Off to the races I went.
It was my first “industry” job. I’m extremely grateful but, if I’m honest, it was not a good job. It was a dated software company that had me work on stuff that didn’t really add a lot to my portfolio, in retrospect, but it helped me learn about what product design was at all, and realise that it was a viable career—not just a job, like all the jobs I’d already had. It was there that I discovered that “UX/UI design” even was at all, just from hearing that word from my colleagues.
People ask me how I went from being a painter to becoming someone who designs and makes software—and it’s that I’ve always like computer stuff as a kid, so I naturally gravitated towards it. That, and the sheer grit required to apply to hundreds of places, knowing you’ll have no chance of ever hearing back.
I knew that I had to find a career, as I had loans I needed to start paying—and fast. I knew I had to be good enough to deserve a career, and fight for every job I’d ever get. I knew I was not entitled to anything. I knew I couldn’t show up to an employer with not much else to offer other than a $55,000 diploma I’d made a bad decision to purchase. I find it confusing when I see mentees do this, and I can’t help but think that it’s not a winning strategy. You have to take accountability for the decisions you make—and you have to give employers a reason to hire you. A bad investment is your personal problem—your customers (e.g., companies who want to pay you to do some work for them) have bills to pay, too. Give them a reason to spend on you. Show them incredible skills.
Above all, what has influenced you, your career, and worldview?
Yamilah - I think most people would agree that what makes a person who they are goes so far beyond just one thing, and I’m certainly no different! But one pivotal period in my life was the years I spent at a design high school. It was a public school,—which are notoriously extremely low quality in the States—but this one was unusually excellent. This was about ten years ago, so this was way before designing software became a “thing.” I didn’t study anything to do with software. I knew how to make paintings and could design buildings creatively. Specifically, I studied fine arts (which every student is required to study for all four years in that program), and picked the architecture major (you have to pick one in that program—the options were architectural design, fashion design, film, industrial design, and graphic design).
Nothing but learning how to make software can prepare someone for a career in building software, if I’m honest, so the skills I learned there don’t really factor in to my day to day, as a designer of software interfaces. But the values that I learned there are what really changed me: that school’s curriculum was aggressive, and the expectations were high. Good craft was not a nice-to-have there—it was a requirement. When you showed up to class critique, everything you made had to be excellent—in every respect. “Good enough” was not good enough. As I went on to find out, this is a very common cultural value at top art schools—competition between students is palpable, and professors are extremely demanding. Hard work is required—you really feel the pressure.
One particular experience at that school that shaped me, I think, was when I was in a Civics class (required by the state). I was probably 16 at the time. Mr. Cao was telling us all about some historical movement in American history, probably back in the 1800s—I barely remember. But one stated value of that movement was: “hard work is a virtue.” I don’t think he intended any of that to be profound to teens, but it really hit me, as a young person. I realised how much hard work mattered to me, too, and that I always wanted to distill that as much as I could. I wanted people to feel they could rely on me to get incredible work done—no excuses. It happens that this kind of mindset is encouraged at art schools (for anyone who’s gone to a good one, they’ll tell you that it’s psychologically brutal for this reason).
Relatedly, a colleague here at Instagram and I were once shooting the breeze. I told her that I’ve never really met another product designer that went to art school (this is extremely uncommon, even in the States—I’ve met maybe two prior to joining Instagram). Meanwhile, at IG, there seems to be a higher percentage of designers who went to art school (still not the majority, but higher than I’ve seen anywhere else). She remarked that she’s had the same experience, but that “you can always tell when a designer went to art school.” I asked her to elaborate, and she said: “You can just tell. Their work tends to be playful.” I think there’s some truth to that—even though we study completely unrelated things in art school, the values you carry from that very difficult experience stay with you forever. Excellence never feels optional—there’s a lasting desire to push the envelope, to work hard, to overdeliver.
What are your thoughts on bootcamps, UX university programs, etc.?
Yamilah - As it stands, I would not recommend any design program (e.g., bootcamp, university, etc.). I do not think they are worthwhile investments, most especially for people who want to design software interfaces. I wish I could change my opinion but, until I see higher quality programs, I would recommend even to my dearest loved to stay far, far away.
I think bootcamps and most universities teaching software design mean well, but they completely misrepresent what software services are, and how to make them. They do not create an environment of excellence at all costs which, in my view, is not a nice-to-have in intellectual roles (like the ones you see in tech). For my part, it’s a value that has underpinned my entire career, and what is likely the “secret sauce” behind what (speaking generally) makes the work of people who come from art schools distinct (as mentioned above by my colleague).
I wish someone—who knew about money, who understood current, viable industries and the available options within them—had pulled me aside to tell me how hard this was going to be, and how bad my decision-making really was until that point. If I’m completely honest, there weren’t any competent adults in my life who were in my corner.
I no longer mentor but, when I did, I absolutely make sure to be that adult in someone’s life that I wish I’d had. Something I emphasised to my mentees was the reality of how difficult a career in software is, especially when building it at the start. I make no equivocations about what your customers will expect of you, and how to compete aggressively on the market for their business.
As it stands, programs (like bootcamps and universities) de-emphasise core skills that are actually needed, emphasise ones that never will be needed, and don’t inform people much about the industries that these diplomas are good for. Especially in tech, we do not care about formal education. This is a very merit-driven industry,—perhaps more than most others—so your $17,000 JPG certification from Design Bootcamp #785698654395 or “Master’s of MSC.e and Human Factor Something or Another” at College of Super Cool Liberal Arts #67464560079 means less than nothing to those of us on your hiring panel—but these programs will not tell you that. None of us are reading your cover letters and, if we’re honest, we’re not even reading your resumes. We don’t care about what you say you can do—we care about what you can do. We’re looking at your portfolio and deciding within a few seconds if we think you’re worth our while. In the words of a recruiter friend (we worked at Twitter together, but go way back): “if you can’t make software, I can’t hire you.”
Bootcamps and universities leave prospective students thinking they can “checklist” (e.g., problem statement, Google Forms survey, persona, site map, affinity map, etc.) their way into the industry. This is not going to happen. I think, secretly, most students think “there’s no way someone would pay $80,000 for a persona and fake market research”—you’re right, we absolutely would not, and do not. Tech companies are not interested in your doubled diamonds or dearth of “mappings”—everyone from bootcamps and universities do this, and it all looks identical. We want to see you make software that makes sense for some demand on the market, and software that looks professional. We want to see independent thoughts about software that users will reasonably want, to the best of your estimation. As it stands, bootcamps and universities that claim to teach interface design are misleading and misinforming students about what it takes to make it, which is why so many of them spend months (and sometimes years) banging their heads against job applications trying to find jobs after graduation and why, even worse, even after finding that first job, struggle to find even the next.
Especially in this market, the industry is hard to get into—but not impossible. If you’re not getting bites, your portfolio you were directed to make by your program is nearly 100% likely to be the problem—you very likely need to actually redo it completely, because the market is not interested in your services, and you don’t understand what your customers are looking for. You need to double down on your core competencies. If you want to be paid more than you ever have in your life, you’re going to have to work harder than you ever have in your life—there are no cheat codes in life, though design programs routinely misrepresent this (by telling students a designer “checklists” all day at work—I’ve never made a persona in my 7 years, and neither have any of my colleagues).
The design program of your choice should help you realise all this, but they don’t know any better. They do not know how to compete for the customers you want, so it’s not really their fault. They don’t know what it takes, so it’s hard to blame them. It’s incumbent upon you to understand what you’re buying, and check their services against what your target customers (e.g., tech companies) want. The reality is that design programs do not prepare you to get a job at any tech company. I’ve mentored over eighty people at this point (from CCA to UMich to Designlab, and way more) and, while it deeply pains me to say this, I have yet to see one designer who is hireable at a tech company. This is not a reflection of the eighty individuals and who they are as people—so many of them are lovely, and I am grateful to call some of them great friends now. But this is a reflection of the quality of the services these programs sell to people who put all their hope in them. It also touches on the relative opaqueness of the industry, and how easy it is to make very costly, bad decisions in order to get into it.
I have no issue with the concept of bootcamps or universities but, as we all know, the States just proposed to bail out 40 million people (to the tune of $300 billion) for spending tens of thousands (each) on degrees that have been otherwise entirely bad investments (which is why the nation’s taxpayers are bailing them out in the first place—it’s gone very poorly for all of us). At this point, it should be clear to consumers that educational programs have fatal issues, and these programs have little incentive to correct them.
So just focus on the core competencies that your customers (e.g., tech companies) are looking for, not what your bootcamp or university tells you to focus on—unfortunately, your program is extremely unlikely to be well informed on what those core competencies truly are. We are fortunate to work in an industry that doesn’t require legal certification, so you don’t need to pay anyone a ton of money to come out with an unhireable portfolio. If you wanted that, you could do that for free. Go make amazing software of your own, using your own mind to reason about what users will realistically want (for example, users do not want “a food delivery app for your favourite bakery”, nor a “flower catalogue app for a florist”—these are already solved by Uber Eats and Shopify, respectively), wrap it up in a beautiful online brochure (a website), and then go apply to places. There’s no way around how hard this is going to be, no matter how many thousands you throw at it. I learned this the hard way, but you don’t have to. You get what I paid $55,000 for free!
Subscribe for free to receive new posts and support my work.
You mentioned core competencies. What are those core competencies for designers? What should a pre-career designer focus on right now to get a job?
Yamilah - I talk about this in some of my articles, but here are the two, main competencies designers need:
The ability to make software that makes sense, and reason about full features (or entire products, if you’re in an MVP stage) you’ll build in a product, from start to finish. Some people call this “UX design,” or “interaction design,”
The ability to make software that looks professional. Some people call this “UI design,” or “visual design,” or “aesthetics.”
At some point in your career, there will be a third expectation around competency. You’ll be expected to contribute to the product’s strategy, and get features put on your team’s roadmap. To put this another way:
The ability to understand the context in which your company, product, and team operate in, take in multiple inputs about the current state of the company’s business, understand what your company’s customers generally want and expect, and output a strategy to steward and expand on the company’s product(s).
For people just getting into the industry, you need the first two, or you’re going to struggle enormously to compete on the market. But, in order to make a portfolio of well-reasoned software to show to prospective customers, you will necessarily brush up against the third. But you won’t need to demonstrate a full product strategy, so I would consider this a nice-to-have in one’s pre and early career stages.
The best way to demonstrate all of these is to do the following:
Build your own software. By “build,” I don’t mean go and implement it, like as a developer. I mean go and design it. If you use a product often, what are some things that annoy you about it, or that you think it really ought to have? For instance, I think Discord should have voice notes, just like WhatsApp, iMessage, and Instagram do. Lots of chat apps have this, and I think it’s a really cool and helpful feature when you just have something that’s difficult to express in written word to a friend. Go build that. I also think Discord should have a “recurring events” feature in its servers. Lots of people run servers and have weekly, recurring events—I would love to be able to just make a recurring event, just like you can in Google Calendar. Another idea, much larger in scope than the two thus far: what if LinkedIn had a mentor and mentee matching service? The professionals are all on LinkedIn, so why couldn’t there be a way to match mentors and mentees on there? Go build these, or anything else you think makes sense,
Whatever prompts you decide to choose for features to add or improve into your chosen product, build it and make sure it looks (1) professional and (2) indistinguishable from the product it’s being built into. For most of your career, you’re likely to be building features on to an already established product, so getting practice now is a good investment.
Once you have three of these, start building a beautiful (1) website and (2) slide deck. These should fully explain how you came up with the idea, why you think it is a good one, why you think users would want it, show a recording of you tapping through the full flow(s), and then your thoughts around how you could measure this product’s success.
For example, if Discord had voice notes, how do you think this would affect user metrics on Discord? Is it possible it might increase the rate of engagement in messages? Could it decrease them in some way? If Discord had recurring events, would this increase the rate of creation of events in servers, and therefore increase the rate of engagement in Discord servers? What are your hypotheses around how these fixes and additions could affect metrics? Go hypothesise. Write about this at whatever length you need to fully explain the software, how it works, what it does, and why you think it would be a successful feature for the company to invest in. Don’t explain it in a complicated way. Explain it like you’re explaining it to your mom or dad. If your mom or dad won’t get it, no one on your panel will understand, either. We don’t get a lot of time with you, so there’s only so much time you have to explain yourself. Make sure the average person understands, and that increases the probability that the average interview panelist will, too.
As mentioned, the website and slide deck should both be beautiful—and by beautiful, I do not mean generic. Do not use a template. For those of us who review portfolios, we review 30-60 templates a day. We’ve seen yours before. If you don’t have good visual design, you have to practice this skill every day to get better at it. It is a very difficult skill to acquire. The best way to do this is to go on Pinterest and look up traditional graphic design (e.g., zines, magazines, posters, etc.) and copy them exactly as they are—every day. Visual design for user interfaces is not difficult enough that you’ll get enough practice to push you to your limits and so, if all you do is learn visual design in the context of user interfaces, you’ll make slower gains than if you practice visual design for other purposes—interfaces are too simple and uncomplicated to make significant progress on your visual skills. To practice your visual skills and make the fastest gains, you’ll need to practice traditional graphic design skills. For an easy, starter challenge, reproduce this, this, or this—exactly as you see them. Pixel perfect copies. Copy posters every day, and graduate to making your own. Websites are just very long posters, and slide decks are just 16:9 ratio posters that you can tap through, so being good at making posters (a common thing for graphic designers to make) will help in your career by way of pushing you to become a great visual designer that can take on any surface (interfaces, slide decks, websites, etc.).
A lot of people ask me if they should be networking or not, and if I have tips. Here’s what I would advise: until you have three solid pieces of software you think make sense, presented in a beautiful website, you’re wasting your time, and theirs. Network once you have those two things, but not a moment before. Help others help you: put in the work before asking for help.
How do tech companies hire?
Yamilah - Every company is different, but for a general breakdown, I have a guide here. First, recruiters are often the ones looking at your site—a company’s small enough, they might not have one, and it might be their higher designer (e.g., design manager, or founding designer), or even the CEO, him/herself. They check through a lot of them every day, so make sure yours looks beautiful—especially important if you want attention from FAANG companies.
Most tech companies are checking for (1) good interaction skills (if you know how to make features that make sense and seem usable to most customers), (2) good visual design skills (you can make software that looks professional), and (3) good product thinking (you can reason about why some software should exist, and have opinions about what a business should be working on—some people call this business acumen, or some variation). For pre/early career designers, the first two are not negotiable, but the third one doesn’t have to be as apparent (this often comes as a function of the first two skills, as well as experience with building software at companies you gain through your career—it often just comes with time spent working in businesses and seeing how they work). Some companies have higher standards than others. At companies like Instagram, the standards are extremely high, and so apprentices and junior hires are almost nonexistent, as far as I can tell. You want to do everything you can to create your own luck, though, so push really hard on all of these areas—as hard as you can, and as soon as you can.
When someone decides they want to move forward with you based on your application, you can throw your site out the window—it’s all about your 1:1s and slide deck now. You typically move on to the following stages, but it varies:
A chat with a hiring manager, or the CEO (if the company is extremely small). They may or may not want you to present your slide deck at this point.
If they like what they hear, they might:
Set you up with a panel of peers to present your slide deck. These range from anywhere from 30 to 50 minutes long of talking time, so this is why you need at minimum two and, at maximum, three, portfolio pieces. You can generally fit two pieces in to 30 minutes. With three, you can accommodate 50 minutes of talking time.
A white boarding session.
An app critique.
If the panelists are generally positive about you and your work, they might slightly or strongly lean to “hire.” You want most people to give you a “strong yes,” not a “yes.” At most tech companies, if even one panelist recommends a no hire at any point in the process, your candidacy ends. You want most (if not all) “strong yes”s, because that means you’ll be at the top of the pile when they consider all the applicants they might have in their pipeline. You stand out as their top choice.
At FAANG companies, team matching occurs as well, which is a separate process, and they all do them slightly differently, and at different points in the hiring pipeline. It can be a super long process. For reference, my full interview loop at Meta took 4 months to complete—to the day.
What does a day in your life look like? It'd be really helpful if you could describe it in as much detail as possible.
Yamilah - I talk about this in another piece, but I am mainly gaining alignment with my team about what we think we should be working on for the next two quarters. My team’s been in what we call “half planning” mode—I talk about this in the article, but this is basically the period wherein we, as a feature team (designer + PM + engineers + data scientists + researcher + product writer), come together to decide what we are going to commit to building over the next two quarters.
At Twitter, we did this quarterly, and it wasn’t extremely formal. At Instagram, it’s a very formalised process, and the result is a “visual roadmap” where we present our strategy for the team’s product area. The roadmap consists of high level concepts for features, and it’s connected to our team’s charter. This is mainly led by designers where, at most companies, PMs generally organise this and ask designers and other stakeholders to contribute. At IG, designers drive this, however.
So that’s what I’ve been up to right now—getting my team together to do exercises to tease out our team’s charter (our team is brand new—it came together as a result of a reorg a month ago), agree upon what we think the future of our team is, and how it relates to the other teams in our area. We want to develop opinions about what we want to build, and what metrics we want to move at the company, and what kind of impact we want to make. It’s a lot of open discussion and debate. I’m working with my researcher right now to develop open questions the team has about our target users on the platform, and what we think they hate about IG Stories right now, and how we can fix it in a way that will truly have impact for the company.
I don’t know what’s going to come of it all but, once we have some notion for the direction we want to take this team in, we have to put together that visual roadmap, present it to our leads, and then prove to them that this direction and strategy makes sense and is plausible. This involves a lot of evidentiary work—our researcher can pull studies, or make his own, and then our two data scientists can pull information about our target users on the platform, mainly analytics that track their behaviours over time. We can use all this information to argue for our strategy, and evidence our identified user problems, as a way to support the features we want to build that solve these issues, and how much impact we think it’ll have (e.g., “this could increase DAU in our target cohort by 2%,” which could be tens of millions of dollars in ad revenue, for example).
When we’re able to commit to a roadmap and strategy, we go into execution mode and start building the software. Designers at Twitter and IG spend a lot of design time in Figma but, at IG, Origami dominates—this is unusual and will likely never be the case at any other company. IG is a product that’s got a lot of complicated gestures (pulling and dragging and pinching and zooming and resizing and so on and so forth), in a way that almost no other products have (some notable examples that are like it are TikTok, FigJam, etc.). We are expected to prototype stuff at very high levels of detail. At Twitter, it was more relaxed, and everything happened in Figma.
Design’s expected to be far enough ahead of engineering such that I have a ton of vetted and done designs, ready for engineers to pick up at any time. Priorities do change as time passes and things come up, so giving engineers lots of runway is very important. It’s easier to design stuff than it is to build stuff, so being way ahead of engineers is required.
There’s only one designer per feature team (as well as a singular PM). There might be multiple designers on Stories, or Reels, or etc., but none of them work on the same part of Stories or Reels. I work on Stories, and there are four other designers who work on four distinct aspects of Stories. We all have different focuses, and we don’t share PMs or engineers—these are embedded on our teams. Researchers, product writers, data scientists, and other roles often support multiple teams at once, however, so my researcher might be shared by one or two other teams.
At Twitter, planning was quarterly and it was super informal. You could just do it with your PM pretty casually and it was no big deal. We didn’t present our strategies to any leads or debate about it. We just worked together, as a team (PM + designer + engineers) to agree on what we wanted to work on, and we just went and did it. We didn’t have too many researchers at the company, nor data scientists nor product writers, but we could go to the research, data, and product writer orgs to request help on products we were working on on an ad-hoc basis. I thought this worked just as well, and returned the same results. I don’t have a strong opinion either way, but I might slightly prefer the approach that requires less presentation work (i.e., Twitter’s way).
Previously, you worked at Abstract, a git-like software for Sketch that gives designers access to Version Control. What are your thoughts on extending this approach to all design tools? Do you see something like "Github for designers" emerging?
Yamilah - I had a lot of fun working on Abstract (I worked Abstract’s second product, which isn’t its Git-like version controlling software for Sketch), but I don’t think the future for design collaboration looks like Git or Github, personally. Figma has already added branching, and it has some notion of visual diffing, but it has a lot of problems. I don’t know that I’m convinced visual diffing will ever be an easy problem to solve in design, nor provide much utility to designers, even if it were perfected.
I think most people don’t like managing versions and, considering that designers generally work solo,—even on large teams—there’s not a lot of opportunity there. What’s more, code gets written and rewritten on the same exact lines, and exact line changes are required when trying to understand what’s happening at every step in some code’s execution. This is not how design works. Design software is modelling software—it just models how something should act, and is evaluated entirely visually. Code is not modelling software—it is the software, and what it says is what the software does. Line-by-line evaluation is, therefore, required.
I don’t think there’s much sense in having visual diffing—I just don’t think people who work visually need or want this. One of Figma’s greatest value adds for teams is that they no longer need to deal with merge conflicts (and other problems introduced with version controlling), or final_for_real_copy_v2_v8_forrealthistime.sketch because it’s all handled for them on someone else’s computer. I don’t see the clock ever turning back on this enormous convenience.
What’s the future of AI, especially as it pertains to design? Do you think we should be worried that the recent developments will impact our jobs?
Yamilah - I don’t know if anyone has the answer to that right now. As it stands, we’re in a position where machines are able to synthesise things that are currently known—it cannot be used to generate abstract, new thoughts and answers like never before seen (e.g., new products that respond to market conditions, proofs for thousand year old, unsolved math problems, etc.). As we’ve seen with text and image AI generators, they have to synthesise material using things that are currently available—this is super cool, though. There are a lot of ways this can help many people in many industries right now. But the content that these tools generate are interesting and fun because they require human source material (imagery, text, etc.)—they take what already exists—and blends all these things together in really interesting, sometimes heretofore combinations.
The catch here, though, is that humans can generate completely new ideas without any source material to blend with, irrespective of what they are given. Machines can’t be used to do this right now—it’s all about synthesising things that currently exist, in order to create unusual and interesting combinations of these things. So, to put this in another way, these tools can summarise articles and whitepapers really nicely, or blend images together in ways that look super interesting, or even synthesise from articles available online to write individually on a topic, but it wouldn’t be able to onboard on to a product team, understand the product, its business, develop opinions about current markets and user demand, and then start producing business and product strategies for these product teams.
It just doesn’t chunk this level of information yet, and doesn’t have the intuition needed to do something like develop an abstract opinion on something like product direction. It can only synthesise knowns to create interesting combinations—it can’t delve into complete unknowns, like humans can. It cannot create original thoughts, or have a way to gain access to knowledge that isn’t online (”feelings” about culture and society that are simply deeply held and hard to describe, but have enormous impact on human decision-making), and it has to know how all that information is connected (cultural direction, level of interest on the market, demand from users, recent happenings that might impact public opinion, etc., and then sum all those together). Until we can make machines that can create novel concepts, I don’t think it can be said that product design (or engineering, or product management) will be able to be automated away.
It’s possible some machines can be used to make products that help us do our work more efficiently (e.g., make us faster, happier, etc.), as mentioned—some already come to mind, like Github’s Copilot, for example. For writers, it might be possible that these markets may further contract, and timelines to producing even something fantastical—like fiction novels—may decrease. It might help the writers that do stick around produce more quickly, and more accurately.
So it’s possible that these tools can generally make us so faster and happier. It might be that we wouldn’t need as many people at our workplaces, so perhaps companies might get more productive with fewer people on staff (on average). But who can really say? Until machines can create new ideas like humans can, though, I don’t think we’re going anywhere.
For any pre-career designers who have yet to get into the industry and are worried about tech still being a viable industry, I would not worry about this for now, personally. As cheesy as it sounds, if you’re really interested in design, just go and make it happen! Don’t be deterred by innovations in machine learning. It’s got a ways to go yet.
I do think it is possible, however, for machines to be able to do more than synthesise in some future. The question is: on what time horizon? I don’t know, and I don’t think anyone really does. My personal prediction is that we can revisit this topic in 10 years and see. In that time, someone can build a great design career, though, so don’t sell your future out for something that might happen so far down the line! As the industry changes, we can change with it, too. And, honestly, if we reach a point where machines can create novel ideas without human source material, I think every single one of us will have bigger fish to fry than “is design/engineering/product management still a viable career?” The entire world will be drastically changed forever.
Anyway, big market shifts don’t happen over night, either—even the change from horses to vehicles as the primary mode of transportation (as huge and world-changing as that was, globally) took a long time, and so you have a lot of time to adjust and determine for yourself how to respond to a changing landscape. For this reason, I think getting into tech is a great choice and will probably remain unchanged for a while to come—provided one is sincerely interested in it!
Subscribe for free to receive new posts and support my work.
If you're comfortable with it, can you share your Spotify Playlist? Or your top 5 songs?
Yamilah - I don’t really make playlists! I’m so boring! Honestly, I even forget music exists. But I’ve been enjoying the following tunes lately -
Lastly, what are your favourite desserts?
Yamilah - I have a huuuuge reputation among my loved ones for having the biggest sweet tooth! So, obviously, this is my favourite question!
Cuatro leches cake (what could be better than tres leches? Cuatro leches, obviously!),
Quesillo (the Venezuelan take on flan),
Pastelitos de guayaba (buttery, flakey bread filled with guava paste and often cheese),
Fried ice cream (vanilla ice cream, breaded and fried),
Tiramisu (undeniable classic, obviously).
That’s it, folks; I hope this resonated. It’s my first time writing long-form interviews, and I would love feedback from all of you. Feel free to comment reply to this email, comment on Substack or DM me on Twitter (@ksvchn) about what you liked and, more importantly, what you disliked.
On that note, wishing you a delighted new year!
I think this interview could possibly change the entire vector of my career (like a Graphic designer shifting for UX in the middle of Coursera Google Certification). Thank you so much for this deep and honest talk, soo impressed!
a wonderful and real insightful read. thanks so much for sharing this with us.