Genuine Creativity Is Your New Moat (2026)

AI makes it fast and easy to copy what’s already out there. If you build products or run a business, your competitive edge is not necessarily features or price, it’s ideas. This means getting in the habit of inventing new things is more valuable than ever. And to explain how, let’s look back at a strange time for the web: the era of Flash. Back in 2006, I worked at a tiny creative agency called Mutant Media, on 35th and 8th in Manhattan. We designed and built websites, motion graphics, print design, and a little something called microsites. If you’re old enough to have lived through the Flash era, you remember a particular species of microsite: you’d be on a major brand’s website and there’d be a link that said something like “Launch Experience >>>” or “Enter Flash Site” and clicking on that would open up a fixed, 640x480 popup window (yeah) containing a Flash app. This Flash app would have some kind of fancy, eye-catching quasi-functionality, like allowing you to spin a sneaker around in “3D,” and there might be a low-fi music loop playing (mute button only if you’re lucky). As another example, one of the many projects I worked on at that time was a Flash game for Bacardi where you mix Mojitos by dragging and dropping ingredients. The game had a bank of sound samples of a voice actor portraying the bartender and saying things like “Nice” and “It’s pronounced mo-hee-to.” Hahaha. Sorry, I don’t have a screenshot of this. I wish. Whether you liked or hated Flash, this period of web development was nothing short of a creative explosion. At a glance: - 1996: FutureSplash becomes Macromedia Flash - 2000: Newgrounds creates its “Flash Portal;” experimental net art site PrayStation launches; the Favourite Website Awards (FWA) launches specifically to recognize innovative Flash websites - 2001: PrayStation wins the Prix Ars Electronica - 2002: Les Manovich writes Generation Flash - 2004: Wired magazine covers a functioning Flash simulation of a Mac System 7 machine - 2006: Flash Player is installed on almost 98% of web-connected PCs. ActionScript 3 is released with Flash Player 9 (which was the first version to be officially called Adobe Flash Player) - 2008: Flash has become the standard delivery mechanism for web video - 2010: Steve Jobs publishes his “Thoughts on Flash”; Three.js is released - Beyond this point, Flash usage steadily plummets until it was officially discontinued in 2020. Today, the Flashpoint archive catalogues 157,509 Flash works—and those are just the ones that happen to be preserved. I’ll remember one other thing that heavily influenced me: a site called Born Magazine that paired writers and artists with designers and programmers, creating sometimes beautiful multimedia experiences that I haven’t forgotten to this day. If anyone knows anyone who has all that data backed up, oh my god, please get in touch with me. I’d be more than happy to revive all of it. But I’m not here to praise Flash. It objectively sucked, and everyone knew it objectively sucked. Accessibility? Yeah, no. Unskippable intro animations? Browser back button? Everyone hated Flash even while using it everywhere for all kinds of new stuff. There are entire books about this. I don’t think it’s a crazy claim that Flash enabled a whole lot of people to make a whole lot of crazy, weird, unexpected shit. It reminds me of when synthesizers became cheaper and more portable in the late 70s, which resulted in entirely new genres, like New Wave and Synthpop and on and on and on. When you give the world a new tool, with which you can make animations, make arbitrary graphics interactive, have stuff move around and be clickable, etc., etc., you get an entire generation of designers and artists and ordinary randos—not “web programmers”—experimenting with new ideas. People came up with autoplayed music, animated menus, all kinds of animated navigation methods, drag and drop interactions, loading screens as a creative medium, product configurators, and a million other things that JavaScript could not do at the time. And yes, I know a lot of that stuff sucked. But man, you don’t get new good ideas without sifting through piles of bad ones. Experimentation is how we discover interesting, valuable, worthwhile things that didn’t exist until some kid learned how to use the weird new tool. The Present Works and is Boring Fast-forwarding to today: the web is far more standardized, far more functional, far less busy, far more accessible, far less annoying. But it’s also boring as hell. I don’t think there’s any way I can convey to you how interesting all that Flash garbage was if you didn’t live through it. It’s trivially easy to complain about it. It’s less easy to recognize that we, as a culture, had to go through that phase in order to figure out a bunch of stuff, like, “Hey, maybe screenreaders should be able to tell what this button is for.” Out of curiosity I went to nike.com to see what it’s like today. And, you know, it’s fine. It’s perfectly cromulent. There’s a little carousel. The pages load pretty quickly. Design-wise, it’s, you know, fine. AI-‘designed’ Sites It’s also extremely banal. It’s background design: the kind that no one notices. And it’s not just this one particular site. Now that LLMs can vibe up a site in minutes, anyone can whip up an equally banal, inoffensive site that is nonetheless technically fine. LLMs, after all, are trained on this vast internet of banality. So why would we be surprised that the output is ordinary, (mostly) legible, more-or-less accessible, and all looks the same? I’m sure you won’t be surprised to hear that there are a lot of LLM websites being generated. An April 2026 study found that, by mid-2025, around 35% of newly-published websites were AI-generated. OK, so now you’re wondering if I’m drawing a comparison, at the cultural level, to the Flash phenomenon. But there’s a pretty gigantic difference: the humans initiating the creation of these LLM-generated sites aren’t making any important creative decisions. There’s no experimentation happening here. If there are new ideas here, they have to do with the subject matter, not the visual or interactive presentation. “But Ted,” you say, “there doesn’t need to be visual or interactive experimentation. In fact, there shouldn’t be. We’ve figured out the correct way to make websites.” And for many intents and purposes, I’d actually agree with you. I’d rather have a boring but clean, legible and accessible site with decent visual hierarchy than, say, a plain text file. (This also appears to set me apart from a lot of programers who appear to think that undifferentiated plain text is literally the only user interface that should exist.) To be clear, experimentation and art on the web never went away entirely. There’s always been communities of artists and programmers and technologists that continue to push boundaries and hopefully there always will be. But if you’re worried that your SaaS is now going to start having a lot more competition, you might be right. Generative AI is lowering the barrier of entry, and at least some companies are struggling with that. Between 2022 and the end of 2024, the number of Y Combinator applications jumped from 20,000 to 27,000—per batch—and then starting in 2025 they moved from 2 batches per year to 4 batches per year. So: how do we differentiate ourselves? How do we stand out from the crowd? It is in Your Best Interests to be Creative The answer is genuine creativity. I’m not talking about trying an edgy color palette. I’m talking about starting over from scratch. Don’t take one step back from your project, take 100 steps back. What are you trying to accomplish? What is the heart of your idea? What do your users want? What would be cool? What would be awesome? What the Gen-AI-enabled world needs now is better taste. It’s easy—trivial, even—to stand up a site or a basic app. But so what? Are you looking at the user experience from end to end? Are you questioning your own premises? Have you considered a completely, fundamentally different approach? And if that sounds hard? Well, yeah, exactly. If it’s hard, really hard, then you might just be on to something. Often this is exactly why something hasn’t been done before: because it’s a hard problem and nobody’s solved it yet—until you come along with a new idea. If you want to differentiate your product, your moat isn’t one novel idea. This can, and will, be copied. Your moat is the ability to keep inventing original solutions to problems while everyone else is just iterating on the same tired answer. Reexamine Your Assumptions “Starting over from scratch” doesn’t mean throwing out accessibility or standard practices that have real value. It means creative problem-solving starts way, way deeper. - Does this product need to work this way? - Is this actually the customer’s problem? - Is this feature a good solution, or merely the customary solution? - What would we build if we hadn’t already seen 20 answers to the same question? (I went deeper into this subject in A Hundred Steps Back.) AI Helps You Experiment Faster Now that we have machines that can generate humanlike outputs, AI can potentially be a great boon to creativity. It can help you try things and iterate much more quickly than ever before. The critical thing is that you don’t let your AI decide important things for you. You need to be the one questioning everything and keeping an eye out at all times for novel solutions. We have an opportunity here, as a culture, to leverage AI’s speed to make more experiments, not more replicas. So next time you’re deep in conversation with your LLM about a product idea, ask it to identify the assumptions you’re making. Pick one of those and consider what it would mean for that assumption to be wrong. Then ask it to critique your approach from the POV of someone who actually has to use the thing in question. Expect some its response to be nonsense, but deciding what is likely true and what is likely irrelevant is actually in itself a helpful exercise. I have a lot more ideas about this subject but those will have to wait for a future article. The Business Argument When competent execution becomes abundant: - “We built an app” is no longer special or impressive. - Professional-looking design is not a differentiator. - Standard feature sets are easy to replicate. - Even a novel output can still be copied quickly. What remains scarce is your capacity to: - Identify problems everyone else has overlooked - Frame familiar things differently, even radically differently - Invent a truly new interaction, model, experience, or category - Make unusual or experimental choices coherent - Continue evolving after your competitors copy your ideas I can’t overemphasize the word “habit” enough. And you may have noticed that the value in this is not new. It’s always been valuable to be innovative; it’s just suddenly much more apparent. Experimentation is Not a Guarantee, but it is a Requirement Let me reiterate one other point: You don’t get new good ideas without sifting through piles of bad ones. Every time I see some internet commenter shitting on an experiment I just think “Well, good luck ever having a new idea with that attitude.” The purpose of the experiment is not to be “good” or “bad.” It is to try. When you were a kid, you probably reached a point where your taste overshot your ability to create, and you started experiencing frustration looking at, say, a drawing you did, and angrily throwing it away. You may have come across Ira Glass’ famous quote on this point. I’m begging you to change that reaction. You must expect that a lot of what you try won’t work. You’ll make something awkward, or confusing, or discover after three hours that your incredibly clever idea has a fairly obvious problem. Congratulations, you’re experimenting! This is what it looks like before there’s something good to show anybody. Creativity is a habit, and part of that habit is letting an idea live long enough to find out whether there’s anything to it. “That probably won’t work” can be the beginning of an interesting afternoon if you let it. Why won’t it work? Could some version of it work? What would have to change? My single favorite source of inspiration is simply: “What if…?” So now we have tools that can help us investigate those questions quickly. Like, really, really quickly. That weird interaction you’re imagining? You might be able to try it out today. The approach you dismissed because it would take too long to prototype? Maybe it doesn’t anymore. This can be—should be—an incredibly exciting time to make things. But we have to actually take advantage of that opportunity. If every experiment starts with a competitor’s website and ends when ours looks approximately as “professional,” we’ve set ourselves a pretty low ceiling. We could be asking what our customers would love that nobody’s bothered to make yet. That’s where the business value lies. Getting comfortable with this process gives you a real value prop when everyone else is quickly and efficiently duplicating past success. Practice questioning the premise, following the odd idea, finding something useful in an attempt that mostly failed. And keep doing it and literally don’t ever stop. A competitor cannot acquire all of that by prompting Claude. I think about those ridiculous Flash sites because people were so visibly excited to find out what they could make. I’d love to see more of that excitement again. We have extraordinary tools at our disposal, and there’s still plenty we haven’t tried. So start giving credence to weird ideas and go find out if any of those weird ideas work. Just, if you include a soundtrack, please let us mute it ;)