Source: Deep Learning on Medium
Software that’s eating the world < Software that’s creating a new world
“It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.”
– Mark Twain
This simple logic is echoed throughout economic, investment and technological history. The LTCM debacle, Lehman Brothers’ end in 2008, the demise of Sears and Blockbuster, the railroad trusts of the 1800s — all behemoths in their day, now nonexistent or almost completely diminished. In a world dominated by tail events and power laws, financial and economic ruin is most often and easily achieved by overconfidence in sure things. At the same time, the unpredictable, unthinkable events capture almost all the upside from technological and economic change. The idea that most of the world’s information would be freely available and searched online, or that movies and TV shows would be streamed online at home, or that open-source software would support much of our financial system can seem unlikely ideas just prior to their ubiquity. The best bets are often seemingly illogical because they are based on the premise that the way the world works now will not be the way it works in the future.
Ultimately, in the words of venture capital’s godfather, Don Valentine; “to make a great investment you must assume the past is wrong.”
Software that is eating the world:
The disruption of established industries by more nimble, powerful and younger software companies has been an exceedingly lucrative industry in the last couple decades. It has delivered unicorns and mythical 100x multiples on invested capital, a core ideal of VC funds as an asset class.
Incumbent corporations, the dinosaurs, whose profits are built on highly efficient distribution systems in traditional industries such as retail, entertainment, stock trading, taxicabs, hotels, etc. are upended somehow or another. Their dependency on the efficiency that once created their margin leaves their business exposed to new ideas. This allows entrants to eat their lunch over time by using new technology, creating tremendous fortunes in the process. In Forbes’ most recent Billionaire List, 5 of the top 10 richest people in the world attribute their wealth to software companies. The largest IPO in history was the 2014 debut of Alibaba, a software company. Software eats the world as technologist and VC Marc Andreessen observed in 2011. Investors in the stock market loathed technology at the time and felt the lofty private valuations of Facebook and Twitter were just part of another bubble. But Andreessen was right: from 2009–2019, companies that have embraced software solutions have gobbled their competitors. The rise of Airbnb, Stripe, Netflix and Amazon has redefined industries through software. The timely and effective adoption of software by established companies like Walmart, Goldman Sachs and Domino’s has allowed them to capitalize on new technologies at the expense of their competitors. It happens slowly at first, and then all of a sudden, like actual dinosaurs, many of the incumbent companies are gone.
This has been the narrative for the last decade and is evidenced by several software-driven startups that have managed to upend entire industries. By shifting consumers and enterprises into new paradigms, these few who bet the world could and would look different have garnered the lion’s share of returns. It is a vicious, subtle process and usually winner-take all. This story is echoed throughout history; the end of the coal and energy trusts at the start of the 1900s through cheap and reliable electricity, or the Industrial Revolution over feudal economies in the 1700–1800s, or the end of the Bronze Age civilizations in 1200BC as the Iron Age emerged. The common thread is that investments which maximize the return on once scarce things, be it natural, intellectual or talent resources, deliver plenty of upside to society as a whole.
However, software that is eating the world will also one day yield less and less return, just as investors “overdosed” on new technologies like railroads and dotcoms, leading to commodification. To make great investments today, we must assume the past is wrong and look elsewhere for opportunity.
Software that is creating a new world:
Fortunately, such opportunities and technological breakthroughs are not far from sight. Just as the New World of the Americas was discovered by using the best navigation technology available to do something most people believed to be a waste of time and money, new worlds are now being created by software platforms that are solving some of the biggest economic problems in transportation, healthcare, currencies and much more. For example, the automobile is moving from hardware-driven to software-defined as we create autonomous vehicles. This new product aims to deliver nearly 0 emissions, 0 fatalities, and 0 traffic delays (each of which may seem presently unthinkable).
The future of mobility is also much more than rethinking the automotive industry. It’s creating new systems and systems-of-systems, from vehicle-to-IoT technology, advanced communication networks, deep learning-powered computer vision and causal inference algorithms for autonomous decision-making. To achieve this interdisciplinary progress, we must build new foundational levels of software to enable real-world benefit for consumers. At Deeplite, we are fortunate to be part of mobilityXlab in Gothenburg, Sweden, as well as the LSPARK accelerator, working with fellow Canadian company BlackBerry on real-time operating system QNX, currently in 150 million vehicles worldwide. Autonomous cars are highly complicated computers with wheels. Similar to when computers first emerged in people’s homes, the foundational software makers, the operating systems of personal and professional life, yielded incredible returns to the relatively few people building them. We must now create operating systems for self-driving cars, surveillance cameras, IoT networks, drones and more to be successful in creating the new world.
We cannot be naïve nor grandiose in what this new world will bring. Just as petroleum brought great comfort to many by way of efficient transportation, energy and cheap plastics, it also brought pollution and conflict in creating the petro-economy. This new “oil” of data and new world of software will also create new contests and unintended consequences, especially if not implemented with caution, diligence, and foresight. The Pentagon recently noted that it must leverage the billions of dollars of work in the AI software consumer space to adequately improve national defense, namely with respect to its edge computing capabilities for AI algorithms like facial recognition and natural language processing. Applications of deep learning and AI that emerged from a handful of academic labs are now leading examples of applications that lie at the extremes of harm and benefit.
Software that’s creating this new world will be the big bet of the next decade. In the perennial push for new ways of doing old things, consumer electronics giants, existing technology companies and governments will jockey for the best of the best. The exponential improvement in computing power enjoyed for the last fifty years is slowing as we hit the limits of computer architecture’s ability to pack more transistors into smaller areas. Furthermore, the increasing sophistication and rapidly changing algorithmic landscape for deep learning model architectures outpaces hardware adaptations. In a post-Moore’s Law world, where we cannot depend on exponential performance gains, we will instead rely on ingenious new hardware designs coupled with creative software applications. Inspirations from highly efficient systems such as the human brain, and their software implementations may deliver unfathomable new results both in benefits and inequalities. It will define the world, not just disrupt it.
However, caution in such investments is required. Patagonia founder and chairman, Yvon Chouinard, pointed out a fundamental paradox in a recent interview: businesses that do extremely well do not often live extremely long. Most enterprises that are hundreds of years old are family offices in Japan and Europe. They have reinvented themselves over the years, from being a blacksmith shop to radio makers to motorcycle manufacturers. You cannot survive over 200 years and grow at 150% year-over-year while doing the same thing, let alone grow 15% annually for 100 years.
So, we must choose, not just in our bets but in the version of ourselves we wish to be in the future. Regardless, new software technology will play a defining role. Our choices in this, will define the world.