By Andrew J Scott, Founding Partner of 7percent Ventures
As an investor in early stage technology start-ups I often find myself repeating Justin Kan’s adage, that “First time founders are obsessed with the product. Second time founders are obsessed with distribution.”
Faced with limited customer growth a first time entrepreneur will often try to build their way out of the problem by adding more features to their product instead of focusing on fixing their marketing challenges: how to make people aware of your product and tempt them to use it or buy it?
Steve Jobs said form follows function and that emphasis on good design will continue to grow in importance, as no-code platforms democratise the ability to produce new products and service, creating ever more SaaS software and apps, giving customers ever greater choice. The winners, when technology itself is not the differentiating factor, will be those who marry good design with great marketing.
Good marketing aside, what are the technologies that will underpin the billion dollar businesses of the next decade? Artificial intelligence, quantum computing and the blockchain are all already impacting our lives but are actually still very early in their development.
AI is already disrupting many laggard industries like insurance, healthcare and finance. We are constantly exposed to AI without even realising it. Much of this technology is better described as “machine learning” (ML) rather than artificial intelligence which has become a nebulous term covering multiple different ways to make systems smarter.
The theory and maths behind today’s AI goes back a long way. Arthur Samuel coined the term machine learning in 1952 when he developed a computer program at IBM to play Checkers (Draughts). Progress spluttered until the 1990’s when a resurgence of interest and the availability of more powerful computers and software languages enabled concepts to be applied at scale and in real world settings. Since then improvements have come fast. By 2006 facial recognition algorithms were 100 times more accurate than those from 10 years earlier. By 2014, Facebook had developed DeepFace, an algorithm capable of recognizing individuals in photographs with the same accuracy as humans!
The original concept of neural networks came from copying, in a simplistic way, how our brain cells interact. Biomimicry research today has widened to emulate a wider swathe of nature’s genius.
Within years not decades, better natural language processing will mean easy voice control of many devices from your smartphone to your washing machine, become standard, like when Captain Kirk talks to the computer in Star Trek. Driverless cars really will become a norm on our roads. Detection of most diseases including cancer will be done by computers not doctors, because computers in many cases can already do this better. Kheiron Medical, a 7percent portfolio company, can already detect breast cancer more reliably than a human. The new drugs to treat those diseases will also have been discovered and conceived by a computer.
ML usually works best when fed with a huge amount of data. That’s how it recognises patterns. This opens up existential questions around privacy, yet the benefits of sharing our medical information are patently obvious. We must find a way to navigate the legitimate privacy concerns.
Recent performance improvements have often been at the cost of interpretability: the degree to which a human can understand the cause of a decision. An urgent challenge is to produce neural networks or equivalents where we can understand the reasoning and it’s not just a ‘black box’.
For an ML powered self-driving car to be trusted it’s important to understand how it decides what is a road sign or a child. Researchers at MIT (Massecuestics Institute of Technology) have been working on the Transparency by Design Network (TbD-net), which breaks the ML decisions down into smaller modules and visually renders the thought process as it solves each part of the problem, allowing human analysts to understand its decision-making process.
Outside research labs, the companies who win will be those who find the best way to tackle these problems while navigating a regulatory and legal environment which is currently ill-equipped to cope with the questions raised by this brave new world. These technologies are not going away and legislation needs to catch up fast.
A technology about to transform our world is quantum computing (QC). If ML and AI are approaching adolescence, QC is still a toddler.
Conceived in 1980 when physicist Paul Benioff proposed a quantum mechanical model of the Turing machine QC’s have the potential to do calculations which would take today’s computers thousands of years.
Today’s computers (referred to now rather quaintly as classical computers) at their base level of operation use a “bit” which can be 0 or 1, a QC uses a “qubit” which can be either of those or a superposition of both 0 and 1. A clumsy analogy would be to say that classical computers boil everything down to a yes or no answer. A QC can make calculations with answers which include yes and no, maybe, or both, etcetera.
A powerful practical QC will open up the possibility to simulate fluids accurately, molecule behaviours, protein folding – in fact, do many things we’ve never been able to before. This will result in a torrent of technological breakthroughs from new materials to the invention of new drugs, and a host of other things we’ve yet to even imagine.
QC’s today have less than 200 qubits. To be really useful a QC needs many thousands of qubits. The race is on to build one and the field is wide open for someone to get there first. IBM and Google have major projects, China has invested billions in their government-backed programs, and there are now a handful of credible startups chasing this goal. One of the leaders is Universal Quantum based in Brighton, England. In 2019 the team published a conceptual blueprint on how they will build a QC that can eventually scale to millions of qubits. With 15 years of research and IP behind them, they are a genuine contender.
This is the start of a new computing revolution worthy of any science fiction.
Blockchain could be as disruptive as the internet itself. Distributed finance (DeFi) which is built on top of blockchain technology really will transform the financial sector. From changing the way financial entities are created and governed, to accelerating services for the unbanked (1.6 billion adults do not have access to banking and other financial services) the financial industry is the world’s largest worth $20 trillion annually.
Decentralized applications (dApps) are being built mostly on Ethereum (ETH). A clumsy analogy would be to think of Etherum for dApps in the same way HTTP powers a web browser. dApps are already changing the way organisations are run. Decentralized autonomous organizations (DAO’s) are attempting to make decision making decentralised and automated, using rules encoded as a computer program, or smart contract in blockchain speak, which are controlled by the organization members, which could be many thousands of people. These are very early in their development but could have far reaching implications for future commercial organisations, NGOs and even governments.
All three of these technologies will revolutionise the world around us, in our lifetime. Marketing is still what gets a product or service into the hands of customers, but technological leaps are the real engines of change in society. If the 1990’s were the decade of the World Wide Web the 2020’s will be remembered as the decade of blockchain and AI.