Reimagining Defence Episode 1 – Thinking Exponentially

This episode was written by Lt Col Henry Willi and edited by Flt Lt James Kuht. The thoughts are the authors own and do not represent the Ministry of Defence. The episode can be found above, or on Spotify & iTunes. You can find out more by following us on twitter, @ReDefPod.

============================

Let’s start with a thought experiment.

A standard piece of paper is about 0.01cm thick.  But, if you fold it eight times, it stacks to just over a cm. And that might seem like a fairly big jump after a mere eight folds.  If you don’t believe me, try it.

So here’s the question, imagine folding a single piece of paper 42 times.   Do you think it would be thick enough to cover from London, to Manchester…or London to Manhattan? 

Well, the answer is actually neither.  On the 42nd folding it would reach from London to the moon.  And on the 43rd it would go all the way to the moon, and back.

The paper folding example isn’t some surprise discovery made by the Army Origami club, but rather a way to help us grasp the first and perhaps most important law introduced in this podcast: Moore’s Law.  This refers to a discovery made by Gordon Moore back in 1958 about computing power increasingly exponentially.  That is, he observed that his company, Intel, was doubling the number of transistors per square inch on a computer chip every eighteen months. 

So we can think a doubling in computer power as analogous to a fold in a piece of paper.  Each iteration, causes both the paper and power to get twice as big.  So what?  Well computer power has now been doubling for over 60 years, and that means the size of the ‘folds’ occurring in our generation are gargantuan.  For us to try and measure that growth, we’re talking about numbers like a quintillion; which is the number 1 with 18 zeros after it.  So as the index of measurement begins to outstrip our intuition we use things like the moon example as a way of conveying what’s going on.  

We’re at the point now where we’re now starting to see stuff that is fundamentally different from anything that we’ve seen before.  For example, if you’re an American in Phoenix’s East Valley, you can sign up to the Waymo early rider app, where you can hail one of their completely driverless cars cars to pick you up.  That car’s available 24 hours a day, seven days a week.  And it arrives with 20 million miles of driving experience under its belt.  This is something that people are using today. Giving credence to that line from the science fiction writer William Gibson who observed, ““the future is already here, it’s just unevenly distributed”. 

I think the other phenomenal aspect of computing power is that it is getting cheaper each year.  A good story to illustrate this is told by Thomas Friedman in his 2017 book, Thank you for being late.  He takes the reader back to 1997, when the US developed the most powerful computer in the world known as ASCI Red, which could process 1.8 teraflops, where a teraflop is a trillion calculations a second.  To do this, ASCI Red occupied a space only a little smaller than a tennis court, used as much electricity as eight hundred houses and cost $55 million dollars

Friedman goes onto say that he was playing on the equivalent of ASCI Red just the other day – it’s called the Play Station 3; launched in 2006, it fits under a television, runs of a normal power socket and costs less than 200 pounds.  He finishes the tale by reflecting that, within a decade, a computer able to process 1.8 teraflops went from something that could only be made by the world’s richest government to something a teenager could reasonably expect to find under a Christmas tree.

Apply this logic of increasingly more powerful and yet cheaper computing power to drones and smartphones, and you can see how in the future, the enemy might be able to develop an ISR brigade on the cheap, and there shouldn’t be any reason why we can’t reap similar benefits too

In a competitive setting, some people believe that if you miss this ever-accelerating IT wave, no amount of paddling will let you catch it again.  To stay on that wave, some say China’s pursuit of AI offers an example of what good looks like here, where in their Made in China 2025 plan, they declared their intent to catch America by 2020 and be the global leader for AI by 2030.  They seem to be on track, for example at the world’s top computer-vision conference held last year and, sponsored by Google and Apple, it was a Chinese university’s algorithm that won.  

Interestingly, these computer vision algorithms are all about detection, which makes them dual-use, meaning the technology can fulfill both consumer and military needs.  So an algorithm that warns a self driving car of a person suddenly stepping out onto the road, could equally warn a remote weapon system of an enemy stepping into arcs.  And in case you’re wondering the name of the university that made the winning Algorithm, it was the Chinese National University of Defence Technology – which is a top military academy of the People’s Liberation Army.  

As well as talent, China also leaves the biggest data footprint on the planet, with more internet users than the US and Europe combined.  To describe this in the context of surveillance, Yitu Technology, a Shangai based AI company, boasts honing it’s award winning facial recognition on the “world’s largest portrait system, covering more than 1.5bn people”.  As we’ll learn in the next episode, the advantage of lots of data like this is that it fuels big data analytics, where if the quality of the data is good, those that own that data can use it to draw conclusions, or target you, with a greater degree of confidence.  Hence why Yitu’s system was built for Chinese Law Enforcement, using data collected by the authorities.

In a competitive setting in the military, being first to harness technology can provide an edge, that for us sometimes makes the difference, between wars won, and wars lost.  Take radar for example, a capability that some might take for granted.  But when the US’s B24 Long Range bombers were equipped with it for the first time in 1943, it genuinely swung the battle of the Atlantic.  41 U boats were detected and sunk that May, which was more in one month than in any of the first three years of the war.  By the end of May 1943, nearly a third of the Nazi’s operational fleet had been wiped out; and the remainder were forced to dive deep to escape the depth charges and gun-runs.  

The German Admiral Doenitz, recognizing the inevitable, withdrew his U Boats from the Atlantic, writing at the time, “the enemy has achieved his objective, not through his superior tactics or strategy, but through his superiority in the field of science”. 

To be able to realise the potential of such battle winning science, we need to be clear-eyed about technologies exponential nature, and how it is accelerating change.  Part of the problem though is that our minds are not wired that way.  When we imagine what progress the next 30 years will bring, we tend to look at what has changed during our lifetime and superimpose that rate as a straight line onto the future.  Whilst thinking in straight lines is most intuitive for us, it’s often wrong.  It’s why I suspect that most people chose Manchester or Manhattan, and on hearing that a piece of paper folded 42 times actually reaches the moon – met it with disbelief and amazement.  As Tim Urban notes in his jaw-dropping article on AI, the trick is to think not linearly, but exponentially.

So to think exponentially, perhaps the first software upgrade we need is for the mind, and that is partly what this podcast hopes to do.  Because, as John Boyd (picture below), fighter-pilot and pioneer of the OODA loop use to say: ‘People, Ideas, and Technology.  In that Order’.  It’s why some say digital transformation is 90% about people and only 10% about technology.  

Peoples, Ideas and Technology. In that Order

John Boyd, below.

And with the forthcoming Defence review, this point appears not to be lost in Number 10 either.  Where some thought leaders want to build ultra high performing cross-functional teams, based on the precedent set by military leaders who’ve done this in the past: such as Groves with the Manhattan project, Boyd with the Lightweight Fighter Program; and Schriever with Intercontinental Ballistic Missiles.

The argument here is that: to build great technology, you must first build great teams; in that order.  As it is only people who can create the conditions for break-through technology.  Historically, the military outliers that grasp this, are often bilingual.  Where, they’ve worked hard to master the language of both team and tech.

Today, those outliers will hear CGS’ call to action made at RUSI last year loud and clear, where he said: ‘the revolution in technology requires a revolution in our thinking’.  This series of 7 podcasts hopes to play a small part in both of those revolutions by giving you the bilingual building blocks to build great teams that build great technology.

==========================

References

We hope you enjoyed this podcast episode, for more check out Reimagining Defence on Spotify, iTunes & Soundcloud and follow us on twitter, @ReDefPod.