The university lecturer I laughed at became the one prediction that predicted everything
In 1990, a lecturer told me and a room full of Cambridge computer science students that there was no money in hardware. In the future, he said, all the money would be in software.
We laughed.
It was an absurd thing to say. A high-end PC cost close to four thousand pounds – a fortune, more than my PhD student grant. You paid for the machine; the programs were thrown in for free to make the expensive thing useful. The value was obviously in the silicon you could hold in your hands. Everyone in that room knew that. So, we laughed.
The lecturer was not just any lecturer. It was Sir Andy Hopper, one of Britain’s most accomplished computer scientists and entrepreneurs. Looking back, it is perhaps no surprise that the man teaching us to reason forward about software would later become a founding partner of Cambridge AI Venture Partners and Chairman of CommonAI. Seeing where technology is heading has been a recurring theme throughout his career.
He was, of course, completely right. Within a decade the hardware had become a commodity, while software ate the world and minted the largest fortunes in human history.
I have thought about that lecture more than any other moment in my life in technology. Because Andy explained his reasoning. He wasn’t guessing. He was reasoning forward from where technology was heading and valuing the future rather than the present. The rest of us were staring at the cost of a PC in 1990. Andy was already living in 2000.
That, I have come to believe, is the rarest skill in the tech industry. Not invention. Not even engineering brilliance. The rare thing is seeing where the world is going clearly enough to plan your life around it before anyone else will agree with you - and being willing to look foolish in the meantime.
So, when I finished my PhD, I went to work for Andy at the Olivetti Research Laboratory (ORL). I reasoned, why not work for the guy who can see the future?
ORL was an extraordinary place. We built things that made little sense at the time, and complete sense the moment you saw the trend they were riding. ATM networking, which went on to carry the world’s telecoms traffic. Virtual Network Computing (VNC) which still underpins how we operate one machine’s screen from another. Piconet, a low-power wireless network for everyday devices, years before Bluetooth made the same idea ubiquitous. The Broadband Phone, an early attempt at carrying voice over the internet, long before anyone said VoIP. None of it looked inevitable from where we sat. All of it looks obvious now. That was the whole lesson, and I learned it on the job rather than in a lecture: reason forward from where the technology is heading and build for the world it will create.
Today I run Efekta Education, where we build artificial intelligence that teaches.
People ask whether AI in the classroom is hype. It is the wrong question. The right one is what Andy would have asked: where is this heading, and who is building for it already? The cost of teaching a child one to one is collapsing, the way the cost of computing once did.
We believe every academic institution will need to rethink how it teaches and supports students in an AI-enabled world. Partly because it is cheaper. But mostly because it can tailor teaching to each individual student, and we have known since Benjamin Bloom’s work forty years ago that a child taught one to one learns dramatically faster than one lost in a class of thirty. For most of history that has been a privilege of the wealthy few. We intend to make it universal. Give every child the personal tutor that only the rich could once afford, and you do not simply raise a test score. You change what a country can become.
We are betting big on that future. Exactly as Andy taught us.
We are working with governments to transform public education at scale, focusing first on the countries with the most to gain. Those with massive teacher shortages. Roughly four million students now learn with Efekta. And it is working. In the Brazilian state of Paraná, we have seen learning outcomes improve by 32% across 750,000 students. Behind that figure are hundreds of thousands of children who might otherwise have been written off.
I can’t imagine more meaningful work.
The digital revolution that Computer Weekly has chronicled for sixty years has, at bottom, been one long story of value moving, from the mainframe to the PC, from the PC to software, from software to the network, from the network to data, and now to intelligence itself. Each migration looked absurd at the start. Each was obvious at the end. The whole game is learning to lean in on the direction of travel early.
I owe Andy a great deal, most of all the habit he drilled into us without our noticing: look past the price tag in front of you to the curve behind it. He taught a lecture theatre of sceptics to reason forward, and I know that for many of us it was the most valuable lesson we ever received.
Written for Computer Weekly.