Skip to main content

Engineering for serendipity

Nicholas Gruen, prominent Australian economist, visiting professor at Kings College London and the CEO of Lateral Economics, says that when it comes to 5G, the arrival of its ‘killer app’ may take us all by surprise

Download full article

5-minute read

If lilies on a pond double in size every day, and they take 20 days to cover the whole pond, how long does it take them to cover half the pond? It’s obvious if you work backwards. With the lilies’ coverage doubling each day, the day before they fully covered the pond they must have half covered the pond. So the answer is 19 days. Yet few people get the puzzle right – including the highly numerate.

Our intuition is so bad at understanding exponential growth that it was hard to believe that 10 COVID cases a day could engulf our societies in a few weeks. Still, as Winston Churchill said in a different context: “Truth is incontrovertible. Panic may resent it. Ignorance may deride it. Malice may distort it. But there it is”.

And it turns out that networks – such as the 5G network being rolled out now – pose similar challenges for our intuition. We’re better at thinking of things than of the relations between them. And networks structure relations between things. And it isn’t just amateurs whose intuitions let them down about networks. It applies to professionals who’ve studied them and business people who’ve bet their shirt on them. Indeed, I know of no communications network that has been introduced without having been widely misunderstood – including by most of those building it.

But this, as you’ll see, is often where the magic happens.


Unintended consequences

Who knows, for instance, that Alexander Bain’s "Electric Printing Telegraph" – the fax – was patented in 1843? That’s 33 years before the telephone and well over a century before faxes became widely used. Yet like the lilies on a pond, it makes sense when you look backward knowing the answer. The technology to convert a 2D image into electronic pulses and then back again seems tricky – but much less tricky than doing the same with the human voice.

More than a century ago, the Bell Telephone company set out to expand its network, by promoting telephony as the means for piping music. It wasn’t a success. Meanwhile the ‘killer app’ wasn’t just sitting smiling back at them from under their noses. Bell made it clear that it was getting up their noses! “

The unlimited use of the telephone leads to a vast amount of unnecessary occupation of the wires, and to much borrowing of telephones by parties who are not subscribers. Thus the telephone system is so encumbered with calls which are unnecessary, and largely illegitimate, that the service is greatly impaired, and subscribers, to whom prompt connection is essential, become dissatisfied.”

Similarly, just as the car began in the quest for a ‘horseless carriage’, ARPANET which grew into the internet was built to give far flung researchers access to a few expensive mainframe computers. Then the miracles started flowing – almost all of them unexpected. Someone hacked up an email app to facilitate occasional messages between computer users and operators. Then usage took off becoming ARPANET’s, and then the internet’s, first ‘killer app’. SMS on mobile networks was a similar story – built for technicians out in the field, not for users of the mobile phone service.

So what can we learn from this sorry tale of how much we have misunderstood the networks of the past – how unpredictable has been their development?

Providing the optimum environment for predictions

My rule of forecasting is “show your ignorance as much as your knowledge”. Weather forecasters do this when they tell us that there’s a 60% chance of rain tomorrow. Compare that with economic and financial forecasts which are usually an assertion of a given outcome – say that GDP or company revenue will grow by 2.5 percent – with no indication of how much confidence you should have in the forecast.

Providing the optimum environment for predictions

And this fact drives professional cultures. Put simply, economists are way too confident in their own knowledge which extends through to overconfidence in their forecasts. Psychologist and economist Daniel Kahneman assures us that overconfidence is ubiquitous among professionals. But when the forecasts clearly state the level of the forecaster’s confidence, that problem resolves itself because we can all see who is overconfident.

This drives a much better professional culture in weather forecasting which has improved steadily over the last generation. Today’s five-day forecasts are as accurate as 1-day forecasts were in 1980. Yet economic forecasting remains haphazard and unaccountable, with a predictive performance that hasn’t improved in generations.

"A critical contribution can come from anywhere"

So when it comes to communications technology, how do we start to forecast the future? Well, it helps to think of networks as engines for serendipity. In other words, a nurturing space for unexpected ideas. IT networks are typically built from the centre out. So new uses of the network must be authorised by network administrators. In building the internet, DARPA’s brilliant engineers produced the first permissionless IT network. As one of them – David P. Reed – put it “the idea … had breathtaking implications.”

Permissionlessness means that a critical contribution can come from anywhere. Indeed my personal experience of innovation and insight-sourcing services like Kaggle and Innocentive suggests that tricky problems are usually solved by some new approach from some other field of activity. The first and second killer apps of the internet – email and the World Wide Web – wouldn’t have taken off without a permissionless internet.

We think of such contributions as ‘bright ideas’. But as Edison famously said, innovation is one per cent inspiration and 99 per cent perspiration. This highlights the imperative for deep collaboration within organisations. But what it leaves out is that some, perhaps most, of the most important contributions are not bright ideas at all, but the identification of shortcomings of existing arrangements. So you can’t get all this right without a high-trust environment in which everyone feels encouraged to contribute, including by critique and disagreement.

5G’s big idea is probably already with us

The history of networks is a history of underplaying people’s natural sociality, their natural desire to communicate. It’s remarkable how consistently we underestimate this given that it’s what makes us unique – the primate that gossips and solves problems in groups (the first led to the second but it’s a long story). But, as Churchill might say, “there it is”. And we did it using that other permissionless network that predated the internet – language.

5G’s big idea is probably already with us

So will there be killer apps on the 5G network and if so what will they be? I’d be contradicting my own creed of humility if I told you I knew, but in driving up bandwidth and driving down latency my eye would go to multiplayer gaming and systems that aggregate distributed AI. We’d also be wise to keep our eyes on industrial IoT. Sure, we know that it can address performance-related issues, but the continuous collection of data from a massive number of sensors could bring about entirely new use cases. From smart wearables to real-time asset tracking, the opportunities for both micro and macro-level innovation are enormous.

If 5G’s killer app is going to happen, the chances are that right now it’s being talked about by someone. But whether they’re inside or outside your organisation, how confident are you that your company ethos ensures that they’re noticed and listened to? That’s the challenge for those who want to give themselves the best chance to succeed in a world that will be different in ways we can’t foresee. It’s the challenge of being open to new ideas – and of engineering for serendipity.

Nicholas Gruen

About the Nicholas Gruen

Nicholas Gruen is the CEO of Lateral Economics and the Chair of Open Knowledge Australia. He was formerly Chairman of the Australian Centre for Social Innovation, the Australian Government's Innovation Australia, and founding chairman of Kaggle, an online community of data scientists and machine learners. He served as economic adviser to Australia’s Treasurer and has held visiting fellowships at the University of Melbourne and the Australian National University.