After the smartphone

Sep 03, 2016

Over the last decade, the dominant technology platform has been the smartphone. More than any other device, it has re-shaped our daily lives. We’re a few months away from the decade anniversary of the launch of the iPhone and it’s worth reflecting on what the smartphone really is and what might come next.

The smartphone is a compound technology platform. It combines a touchscreen display, lithium-ion battery, low-power/high-performance mobile CPU, flash memory, tiny but incredibly high-resolution camera, GPS, high-bandwidth LTE mobile data, a range of sensors, scratch and shatter-resistant glass, Bluetooth, NFC and hardware encryption. Oh, and you can use it to make phone calls. All of this is bundled with a desktop computer-grade operating system and an application framework and distribution model that enables the device to perform millions of possible functions using third-party application code.

Take away any one of those things, and the smartphone becomes much less useful. Take away three or four, and it becomes nearly worthless. By using low-cost commodity components and moving the user interface to apps, the smartphone was able to replace many devices that would previously have existed as separate products.

The smartphone has replaced many previous categories of electronic goods

The story of the consumer tech economy - and increasingly the enterprise - over the last decade has been the story of software running on personal pocket supercomputers, linking sensors, touch interactions, remote data and location-awareness. The cultural and economic impact of the smartphone came about because it commoditised a set of of previously disparate and expensive technologies and enabled software developers to build higher-order systems on top of them.

To see how powerful this is, consider ride-sharing services like Uber and Lyft. Without smartphones, each car would require expensive custom hardware to manage location tracking, to send instructions to drivers, to handle billing, payments and ratings. Customers would need to request cars by phone, giving imprecise directions to their pickup location, or place their requests on the web via a desktop PC. Combined, these restrictions would kill the business: no smartphone, no ride-sharing.

The same is true of WhatsApp (data, mobile CPU, apps) and Instagram (camera, data, mobile CPU, apps, GPS). Tinder’s swipe interface doesn’t work without a touch screen. Google Maps can’t tell where you are without GPS. Live-tweeting without mobile data and the Twitter app would need to be done by SMS, or require a laptop and wifi. Ten years in and we’re only just starting to see apps like Pokemon Go that would be impossible to conceive of without knowing what a smartphone is. The iPhone might have made Apple the world’s most profitable company, but the ecosystem around it has spawned multiple multi-billion dollar companies, and will continue to do so because of how it enables these higher-order products.

So what’s next?

So, to understand the smartphone and what might follow it, we need to understand how technology gets commoditised and how higher-order systems are built on top of it. The best model for understanding this is Simon Wardley’s mapping technique. This maps out the process by which a technology goes from its chaotic genesis to a linear, predictable commodity or utility phase.

Consider how servers went from being custom-built, to standard product, to rented, to shared and ultimately to becoming a fungible utility sitting in a cloud of undifferentiated machines. As the technology moved across the map, more and more higher-order products became possible. The eventual move to cloud utility computing has lowered the costs of starting new online services dramatically by eliminating any need for up-front investment in server hardware, and the predictability of pricing and performance enables easy scaling.

This process is recursive - the services built on top of the cloud hosting environments are eventually commoditised themselves, and the cycle begins again, as illustrated here:

Higher order components such as elastic map-reduce jobs build on top of cloud servers, and provide a base layer for the next level

At any given moment, every technology sits somewhere on that horizontal axis, with the forces of competition driving it further to the right. Following the earlier stages of “wonder” and “peace” as vendors trial their ideas and experiment with ways to grow the market, Simon refers to the period of commoditisation as “war”, as vendors have brought competing products to the market and are fighting each other for profit, mind and market share. This triggers a corresponding period of “wonder” as some other part of the economy - cloud platforms on commodity hardware, apps on smartphones - takes advantage of the falling price of the commodity to build something new on top of it.

If we want to predict the next few years of tech, we should look for technologies that are about to become commoditised, and that could be tied together with software. Just as WhatsApp and Instagram were valuable because they used several of the smartphone’s core technologies, the next wave will be most valuable if it can leverage several newly-commoditised technologies at once, to get a ‘compound innovation’ effect as the underlying technologies continue to improve through vendor competition.

Figuring out which technologies are approaching commodity status is always hard. My own sense is that we’re seeing, or about to see, several technologies reach this phase. In no particular order, these are VR/AR, AI and machine learning, blockchains/distributed ledgers and certain forms of robotics/“Internet of Things”. (Others may see different trends - try applying this model to your areas of expertise and let me know what you see).

VR and AR products have been trickling on to the market in the last few years - the Oculus Rift and Google Cardboard are classic “wonder” products designed to establish a market rather than compete with each other - but the momentum is now sufficiently well-established that we can expect to see many competing vendors in the same space. Initiatives like OpenAI, and open source projects like TensorFlow are putting powerful machine learning tools in the hands of ordinary software developers as competition for talent hots up - one of the most common Hacker News or Lobsters posts of mid-2016 is the announcement of some new open source machine learning or AI toolkit. Blockchains, too, are approaching commoditisation; Eris and BlockApps offer open source blockchain-as-a-service solutions. Finally robotics is in a similar phase - drones, for instance, are much cheaper and much more capable than they were a few years ago, and open source software and hardware designs are common.

Imagine a monitoring system for physical infrastructure like solar panels or dams that uses smart image recognition to classify pictures taken by commodity drones, with results stored on a blockchain for audit purposes and VR remote control for human inspection of anomalies. Previously, this would have required robotics experts, a team of PhD computer scientists to build the image recognition, hugely expensive VR equipment and even the secure distributed storage of the audit trail would have been a tricky problem to solve in its own right. We’re now at the point where generalist software developers and product designers have the tools to solve these problems at their fingertips. By massively reducing the cost of these innovations, the war of commoditisation will make many new products possible.

Competition at the commodity level almost guarantees that such higher-order products will continue to improve. Image recognition will get better, drones will become faster, more reliable and more energy-efficient, with better cameras and faster data transmission, and VR will improve in visual fidelity and response times. Even scaled blockchains will improve the throughput of our audit logging system. It’s safe to assume that if you can build something with these technologies in 2016 that just about works, kinda, then in a year’s time it’s going to work a lot better and will continue to experience the same improvements just as smartphone apps got longer battery life, higher resolution screens, more storage and RAM, faster CPUs and greater bandwidth to play with.

In this sense, the next consumer tech revolution is going to look a lot like the last one, and the one before that - this is the story of the personal computer as well as the smartphone. Commoditisation of key components enables higher-order software products, built by generalist developers whose breadth of knowledge enables them to knit multiple pieces together. Product designers who can grasp the possibilities first will have a huge opportunity to create unique and transformative products.

What’s different this time is that markets are much bigger and more integrated. Apple became the world’s first trillion dollar company because the market for personal computation in the last decade is much bigger than it was in the 1980s or 1990s (it took iOS devices barely eight years to surpass the sales figures of Windows PCs, h/t Benedict Evans). Many Asian economies are orders of magnitude bigger than in the mid-1980s, and many more people are able to participate in the global economy as a result. As new platforms appear, their reach is vastly greater than anything experienced in previous generations, which means that the potential rewards are bigger.

This has obvious implications for startups and venture capital; when Dave McClure says that the real bubble is in the valuation of legacy companies rather than startups, I think he’s right. Based on entirely predictable advances in technology, it’s easy to see how billions of dollars in value is going to be created in the next few years, and it’s going to be startups leading the way.


If you find these ideas interesting, I strongly recommend that you head over to http://hexayurt.com/capital, watch the video there and fill in the form. We’re kicking off a conversation about the future of venture capital in an age of accelerating technological change; the psychology, the financing, and the technology and how these things are going to change. Please do reach out on Twitter if you’d like to discuss further.