Which world would you most like to design on? One is not better than any other, because they come with tradeoffs. The unexplored world on the left has the most potential, but is least accessible; whereas, the explored world on the right is a bit crowded, but is easiest to build on because of its mature infrastructure.
Luckily, it’s easy to get to whichever world you choose. You don’t need to be an astronaut and getting there doesn’t involve rocket science. The world you design on is determined by the maturity of the technology your product is built on. The newer the technology, the more unexplored the world.
Technologies are like worlds because many things can be built on top of them. Cities are built on worlds, and products are built on technologies. Technologies and worlds also have different stages of maturity. So just as there are different stages of technology – novel, growth, and legacy – there are different stages of worlds. The design space of a novel technology is unexplored just like the surface of a new world. As a technology matures, and more products are built on it, more is known about its design space. In the end, legacy technologies are crowded with products just like how a populated world is crowded with cities.
Novel technologies are the most new and advanced, but also have the least amount of adoption because they are so new. They are orders of magnitude (e.g. 10x-100x) better than what came before them, and are usually the result of scientific discovery or engineering breakthrough.
When a novel technology emerges it’s like discovering a new, unexplored world. There are few inhabitants, little has been built on them, and entirely new things are possible. People are drawn to them because they sense this raw potential. The downside is that novel technologies, and new worlds, are least accessible and most difficult to build on because important infrastructure is missing.
Growth technologies are more mature than novel technologies and more products have been built on them. This is why growth technologies are like partially explored worlds that have proven they can ‘sustain life’. The goal here is to explore new continents and expand civilization beyond its first settlements. As their infrastructure matures, growth technologies become more accessible and easier to build on – like how a planet becomes easier to travel with more roads and bridges.
Legacy technologies are fully developed and many products have been built on them. Nothing is unknown and everyone has a good idea of what can be built with legacy technologies. They are like fully explored worlds that have been settled for some time and are inhabited by many. Legacy technologies are most accessible and easiest to build on because their infrastructure is so mature; however, they are the most crowded with ‘inhabitants’ (i.e. talent) and ‘buildings’ (i.e. products), so the chances of finding open real-estate are lowest.
The technologies you work with affect your day-to-day experience as a designer, and perhaps your career overall. The more novel the technology, the less explored its design space. There are likely entirely new product possibilities that novel technologies make possible, but that no one has yet discovered. Also, mental models and design patterns are less likely to exist for novel technologies, which gives you the chance to establish them and influence how products are designed within your category for years to come. It’s most difficult to build with novel technologies because there are so many gaps in infrastructure and tooling – so you’re likely to be inconvenienced by this in some way. Finally, fewer designers and developers have an understanding of novel technologies, which means it’s easier to differentiate yourself if you have an expertise in them.
On the other hand, the more mature a technology is, the more explored its design space. Mental models and design patterns will be more set in stone, and changing them might actually hurt UX because users have become familiar with them over the years. More mature technologies are easier to build on because they’ve been around for longer and have had time to fill their infrastructure gaps. Many designers and developers have already entered the field, and there is an abundance of talent. Legacy technologies are the most proven and support the most amount of products, but this makes them somewhat crowded and there is no longer low-hanging fruit for entrepreneurial opportunities.
Product Design at Generative AI’s Frontier
Novel technologies are the latest and greatest, yet have the least amount of adoption because they are so new. They are usually the result of an engineering breakthrough or scientific discovery, and are orders of magnitude (e.g. 10x-100x) better than what came before them. GPT-4 is an example of a novel technology. It’s the most advanced conversational assistant in the world right now. Assistants like GPT-4 have been around for decades, but recent advances in algorithms and hardware enabled OpenAI to engineer a model 10x better than all the rest.
This makes all sorts of new things possible. For one, conversational assistants will never be the same. chatGPT was the fastest adopted product ever, and recently “Voice” was added where users can speak back and forth with their assistant. There is no limit to the conversations you can have and the assistant also can grab information from the web, making it 10x more dynamic and insightful than Siri (and others).
This is just what OpenAI is doing with GPT-4, but others are building on it too. Just in the past year, GPT-4 has attracted thousands of entrepreneurs who are integrating it into their own products via a paid API. Similar to how many different cities are built on Earth, many different products are built on GPT-4.
So what are all these startups building with GPT-4? You probably won’t be surprised to hear that many are building conversational assistants for specific domains. There are coding copilots for programmers, medical copilots for doctors, and legal copilots for lawyers. It makes sense that advances in conversational intelligence lead to better chatbots, but this is not all they will lead to. “Better chatbots” are most obvious to us, because chatbots are what we’re most familiar with when it comes to language models; however, novel technologies unlock entirely new domains that couldn’t have existed before, and these new domains take longer to discover.
As Chris Dixon points out, when a new technology comes along we first try to carry over use-cases from existing technologies. Over time our design thinking evolves from this skeuomorphic design to native design. By “native design” I mean we consider the use-cases that novel technologies uniquely enable with their 10x capabilities. An example of this is how the internet started as static, read-only websites with content similar to pre-internet magazines and newspapers. It was years later that we discovered user-generated content and social media, which changed the internet paradigm from read-only to read-write.
Here are some early examples of native design thinking in AI.
- Humane’s AI Pin is a paradigm shift in consumer devices. Right now we mostly use GUIs to interact with our smartphones, but the AI Pin has the power to shift us into a voice-first paradigm.
- chatGPT browse is a paradigm shift in internet search. Now AI agents search the web for us, and distill the exact information we want into any format of our choosing.
- LLM OS is a paradigm shift in operating systems. Computer operating systems might start to be designed with large language models at their center.
The point is AI’s design space is mostly unexplored, and it will take decades for us to fully explore. Fortunately, if AI isn’t your thing, but unexplored design space is, there are many other novel technologies for you to choose from. A couple other examples include CRISPR Cas-9 and metal 3D printing. CRISPR is 10x cheaper, faster, and more accurate than previous gene-editing technologies, which is starting to enable personalized medicine at scale. Similarly, metal 3D printing is revolutionizing manufacturing and making it 10x faster for engineers to go from digital designs to physical parts. Printing reusable rockets is one of the initial use-cases of this technology.
As a product designer, your experience building with novel technologies will be like that of a pioneer discovering a new world with little development and few inhabitants. Overall, few products will have been built on the technology, which means established design patterns won’t exist. Also, novel technologies have 10x capabilities than what came before, so it will be like landing on a world with different laws of physics that support new types of buildings and new ways of life. Mental models need to be developed to fit the new paradigms brought about by novel technologies. Designing with novel technologies is more time-consuming compared to technologies that come with off-the-shelf components and mental models; however, you will have the chance to influence how products are designed in your category for years and decades to come.
Also, the tooling and infrastructure that support novel technologies are most immature, and it will take some time for this ecosystem to work through its growing pains. You will often find that the infrastructure you need is missing or under-developed. The main infrastructure component holding AI back right now is a GPU shortage. GPUs are what run AI models, and since the surge in AI products last year, demand for GPUs has far exceeded supply. This has made compute costs prohibitively expensive for some.
There is also the risk that you build your product on infrastructure that becomes obsolete since novel technologies change so rapidly. Right now, OpenAIs models are superior so almost everyone has chosen to build with them, but what if OpenAI gets permanently surpassed by another model provider like Anthropic or Google? Also, a lot can change in 3 to 6 months like technology trends and the dominant products in your category. The fast-paced nature of novel technologies requires that you continuously update your knowledge and refresh your skill sets.
Finally, talent is scarce within novel technologies. Since novel technologies are still relatively unproven, designers and builders are hesitant to spend the time developing an expertise in them. While this makes it harder to form a team, the fact that you are one of a few designers in the field is advantageous to you. It’s easier to differentiate yourself as a designer when your portfolio is aligned with a novel technology.
This should make job interviews easier for you. If product designers are needed badly enough, and your fit with a company is good enough, you might be able to negotiate an above-average salary relative to your experience level. If the technology goes on to flourish, you can ride this wave and define your career by it. Hiring managers will still be impressed by your experience even if the technology eventually fizzles out and you’re forced into a new field. They will rightly sense that you can see further into the future than others, at least in some domains.
Product Design in Bitcoin’s Golden Age
Bitcoin has reached its tipping point. At the time of writing, 100M people own Bitcoin but soon many more will. Institutional investors now want a piece of Bitcoin and they will have their chance with the coming launch of Bitcoin ETFs. What changed for Bitcoin that led to this institutional demand?
First of all, Bitcoin is the oldest, and longest-standing cryptocurrency. It was launched in 2009, and thousands of other cryptocurrencies have come and gone since then. The longer Bitcoin has been around, and the more market cycles it survives, the more people trust that it won’t just disappear. So its age is a significant factor, because it reassures institutions who fear public embarrassment and monetary loss.
Second, there is now regulatory clarity for Bitcoin in the US. Bitcoin is considered a digital commodity by the US government, and this clarity makes it less risky for institutions to own Bitcoin and for businesses to build products on Bitcoin.
Third, the narrative for why people should buy Bitcoin has become better understood. More people are beginning to see that Bitcoin is a scarce digital asset, just like Gold is a scarce physical asset. In other words, more bitcoins can never be created, which is why more and more people regard it as an asset that will hold its value – or increase its value – over the long term.
Not only are more people buying Bitcoin, but more products are integrating Bitcoin. Payment apps like PayPal and CashApp allow users to trade Bitcoin. Some online retailers accept Bitcoin payments. Several nations have adopted Bitcoin as a legal currency within their borders. And most recently, X added Bitcoin tipping. This means that the ecosystem around Bitcoin – like its developer talent and tooling – has matured to the point to have made these integrations possible. This is an example of how infrastructure gaps start to close for growth technologies, and become easier to build on compared to novel technologies.
So long story short, institutions are now willing to adopt Bitcoin because it’s a more proven technology than before. The adoption of Bitcoin will continue to gain speed as it becomes easier to purchase and easier for developers to integrate. This positive feedback loop is true for all growth technologies. As a technology’s ecosystem matures and becomes more accessible, then more people build with it. And with more products being built, the technology’s ecosystem is forced to improve, which brings about even more adoption.
Bitcoin’s design space is more explored than it was 10 years ago. As a novel technology, Bitcoin birthed entirely new product categories like crypto exchanges, wallets, block explorers, node services, and many more. Now, as a growth technology, Bitcoin will need to be adapted to new users and use-cases. For example, institutions that invest billions of dollars in Bitcoin will have different security and customer service needs than Bitcoin’s retail early-adopters. Designers working on these institutional products – like Coinbase Custody – will be able to reuse most of the mental models and design patterns that have come before; however, a significant amount of UX will need to be rethought like safer workflows for transferring large amounts of Bitcoin, managing who has the authority to transfer Bitcoin and how much, etc.
Even though more people understand what Bitcoin is and what it can be used for, Bitcoin’s design space has not been fully explored. Like a partially explored world, we may still discover entirely new ‘continents’ within Bitcoin that we didn’t know existed. Entrepreneurs are extending Bitcoin into new application areas like decentralized finance (DeFi), artwork with Bitcoin Ordinals, and micropayment rewards programs. We don’t know what will be seen in the future, but Bitcoin is still in its earliest stages as a growth technology so I imagine things will change quite a bit from here.
As more products integrate Bitcoin the demand for those with Bitcoin experience grows. There will be an influx of designers and developers for years to come. Right now, there are still relatively few with an understanding of Bitcoin, so it’s still something you can differentiate yourself with; but, in 5 years, many will have “digital currency and collectibles” listed on their resume.
The beginning of the end of smartphones
No one questions the utility of smartphones, and everyone has a general sense of what they can do. This is because 6 billion people now own and use smartphones on a day-to-day basis. But don’t let the fact that Apple releases a “new” iPhone every year fool you. Each new version just incrementally improves on the previous. Every year the iPhone gets a slightly upgraded chip, screen, camera, and form-factor – but there is no 10x leap in capabilities that fundamentally change how we interact with our devices. This is why the smartphone is a legacy technology.
The smartphone paradigm is set, and so are the mental models and design patterns within it. Everyone knows what it means to “unlock your device”, “download an app”, and “silence notifications”. Significant changes to mental models and design patterns are not permitted since users have become accustomed to them over the years. This limits your freedom as a product designer.
Also, the design space has been thoroughly explored by Apple – designing iOS and its native apps – and the thousands of third-party developers who have launched apps on the App Store. This is not to say that new products cannot be built on the iPhone. I imagine many new apps launch daily and some might generate significant revenue. My point is that since the smartphone paradigm is so well understood, and we’re so used to the products built on it, new mobile apps won’t fundamentally surprise us in what they can do.
Building mobile apps is easier than building products on more novel technologies, because smartphone infrastructure is fully mature. There are plenty of resources for learning about mobile app design, and developer frameworks for building mobile apps are stable. Also, think of all the designers and developers with mobile app experience. The talent pool is deep, which makes it easier to form a team. This also means that mobile app design is no longer a skill you can differentiate yourself with. Back in 2010 it was, but now a knowledge of mobile app design is required just to stay relevant.
Legacy technologies eventually get phased out by growth and novel technologies. The smartphone will soon be completely reimagined, catalyzed by the rise in AI Pins and spatial devices. These novel technologies will be 10x more seamless and immersive than current smartphones, and they will fundamentally change how we interact with devices. Right now we mostly use GUIs to interact with our smartphones, but AI Pins will shift us into a voice-first paradigm and spatial devices can also leverage gestures and eye movement. Designers working on next-gen devices will get to develop mental models and design patterns from scratch to fit this new paradigm.