BLOG: About Chess, Tesla’s and Moore’s Law

In the news this week an item about the AlphaGo software beating South Korean player Lee Sedol 4 out of 5 times in a game named GO. That same week I read about the nearing end of Moore’s Law in The Economist. So how are these two related? And what will eventually be the impact on our business?

Let’s start with Intel co-founder Gordon Moore, who claimed in November 1971 that the

processing power of a computer chip would double roughly every two years. And how right he was. For over 44 years his Law stood firm and dictated the rhythm of the industry. The first ever microprocessor in 1971 held 2,300 transistors (electrical switches that can be turned on and off, representing the 0’s and 1’s of digital language). The 2014 Intel Xeon Haswell chip holds 5 billion.

It is difficult to grasp the full impact of Moore's Law. But think of this. In 1997 the US government needed a super computer to maintain its nuclear arsenal. It was named ASCI Red of which one customer made copy was built by IBM. It had the size of half a tennis court and was capable of 1.3 trillion calculations per second. In 2006 Sony presented the PlayStation 3. This machine doubled the speed of the ASCI Red, weighed 5 kg, could fit in a bag and was sold over 65 million times worldwide. Now that is the impact of Moore’s Law!

About Chess, Rice and Tesla's

Us human beings have difficulty understanding exponential growth. But an old parable might be the best story telling option to get to grips with it. Centuries ago an advisor created the game of Chess for his king. When asked what he wanted as a reward, the advisor asked for one grain of rice on the first section of the Chess-board, two on the second, four on the third and so on, doubling the amount every next section. The king laughed and said that he must be joking, and if he didn’t want anything else?
When the counting started only halfway the chess-board, the numbers where very large, but on the second half they were just vastly larger.

Only halfway, on the 32nd section, the king had to deliver 2.147 Billion grains of rice. And

remember, the 33rd section, this amount would double. To cut a long story short: more rice had to be delivered than was produced all over the world for years. The King learned the power of exponential growth (I wonder what happened to the advisor).

This is exactly what happened with computing-power since 1971. Imagine that the computing-power that allowed all Tesla Model S cars to drive themselves since October 2015, will have doubled in October 2017. And all robots that already are replacing work formerly done by humans, will have doubled their processing-power in two years time. And again two years later.

Through exponential growth, what sounded like science fiction a couple of years ago, now has become reality.

The End of Moore's Law

But, according to The Economist, we are approaching the end of Moore’s Law, as the
shrinking of transistors on a microchip is nearing it’s physical limits (the size of an atom). Intel admitted end of 2015 that the investments in R&D, needed to make chips quicker, have started to outpace the increase in speed that resulted from those investments. So chips will still continue to get better, but at a gradually slowing pace.

Two things are happening in parallel and/or in response to that. The first is an increased power to the big cloud-computing companies like Amazon, Google, Alibaba and Tencent. They are improving their cloud infrastructure and will be able to increase their offered computing-power the coming years by combining every single processor of all servers in their global datacenters. With ‘cloud’ we mostly think about storage, like Google Drive or Dropbox. But Cloud-computing offers especially computing-power to customers. Power that is far larger than one could get of a stand alone PC or server.


Artificial Intelligence and Deep Learning

The second movement is not about speed but about intelligence. This brings us to the
other news story last week: the AlphaGo software, built by Google-acquired startup ‘DeepMind', beat the best human player in the game of Go. GO is an ancient game that looks deceptively simple, yet is the most complex game to be solved by a computer, as it has an unpronounceable number of scenario's: 10 to the power of 80. Plus: it is rather a game of intuition than rational strategies.

This game can only be won by a computer when Artificial Intelligence (AI) is involved. Computer-learning, supported by hundreds of algorithms that learn and can copy instinctive moves from humans. This is called ‘Deep Learning’.
So if we have come to the point that computing-power (processing lots of data) combined with machine-learning algorithms (AI), the whole 44 years of Moore’s Law might turn out to be just a lightweight introduction to the real digital age.

The (business) game is on

These two examples, small news articles that you could easily miss, paint a picture of the business environment we have to operate in. The era where not only repeatable patterns are being automated and scaled in the cloud, but where intelligent predictive and prescriptive decision making and AI based consultancy will become part of our business life.

So what will this eventually mean for our rapidly changing world of work? What happens if even the Lee Sedols of this world can be automated? I have a hard time gathering my thoughts around that to be honest.