(April 26, 2018)
You are about to lose your job. No matter what you do, no matter which industry you work in or the position you hold, it is highly likely to disappear in the next 10 years. It is not a joke. Nor an alarmist propaganda. The skillset that distinguishes you from your peers will become totally redundant within a decade. And that is a very generous estimate. Machines are not likely to wait that long.
The Economist this week published a rather uncharacteristically blasé piece titled ‘The Kamprad test’. For a heartbeat it felt that one of the finest publications on the planet which had been a personal source of guidance for me for decades was mocking the artificial intelligence (AI). A group of AI researchers managed to teach a group of industrial robots how to put together an Ikea chair. The job that humans can accomplish within no time, took the robots 20 minutes to finish. And the publication’s science editors danced with joy. Moravec’s paradox holds true then. It states that while high reasoning requires significantly less computational resources, low level chores requiring sensorimotor dexterity need a lot of them. So finally, it has conclusively been proven that the machines will pose no threat to the human civilisation because they are far too naïve to replace human skills. Right? Dead wrong. The Economist really disappointed us in reaching that hasty and premature conclusion.
It is true that human dexterity at question here has evolved over billions of years and AI hasn’t been around for that long. Except under that hood time doesn’t flow in that simple a fashion. The newer versions of computers are capable of running trillions of subroutines per second. This makes the job of committing mistakes, learning from them and evolving a child’s play for machines. Want to test the hypothesis? Sure.
An exabyte is a million terabytes. But I am getting ahead of myself. This piece you are reading is a little longer than a kilobyte (1,000 bytes). 1,000 kilobytes make a megabyte. 1,000 megabytes a gigabyte. 1,000 gigabytes a terabyte (1TB). That 1TB drive that you’ve got can store your entire life in high definition video. A thousand TBs equal one petabyte. A thousand petabyte make one exabyte. Clear?
According to IDC’s Digital Universe Study sponsored by EMC dated December 2012, humanity produced 130 exabytes of data since dawn of time till 2005. This data included everything like the Dead Sea Scrolls and Beethoven’s symphonies. The figure had reached 1,200 exabytes by 2010. Further acceleration brought it to a 7,900 exabytes by 2015. And by 2020 it will reach a whopping 40,900 exabytes. Now some of this data is undoubtedly white noise generated say by your air conditioner’s interaction with the internet or simply put waste. But even so, just take a step back and appreciate where we are at in our evolution. Where humanity’s evolution generated 130 exabytes in only 15 years it grew to 40,900 exabytes. What brought this difference? Smart machines.
And they are getting smarter every single minute. And now to dexterity. If you want to see how robots are improving their sensorimotor skills all you need to do is open the YouTube app and input two words: Boston Dynamics. Their prototype Atlas gave Elon Musk goosebumps when video of it doing backflips went viral. Since then we have seen robots enjoying classical music and gently dancing with it. And every day new videos keep coming.
Please do not confuse this rise of the machines with the Terminator or Matrix like Armageddon. Far from it. There are strong reasons to believe that even when AI realises its true potential it will remain benign to its creators. It will serve you all right. But in serving you it will take all your jobs. All of them. And since in a capitalist society you have to pay for the services you want to enjoy, you have to ask yourselves how would you plan to pay for them, especially when you are out of job. If you are rich, no problem. You will get richer. But if you belong to the working class you are doomed.
Calum Chace in his book The Economic Singularity makes a convincing case that in striking contrast to the Industrial Revolution which created more jobs by shifting mankind’s focus from manual labour to the service industry, the information revolution led by automation we are witnessing today will deprive you of your jobs by taking over the service industry. If you think your job will be spared because you possess unique skills, well, think again. AI’s forays into my own profession began in 2010 when Associated Press began use of Quill, a software that writes sports and business stories for it. How about that? In warfare, a software named Alpha keeps beating human drone pilots in repeated simulations. Daimler’s 18-wheeler driverless trucks are already being tested in Nevada. Dexterity much?
When futurists talk about the rise of AI they usually have a distant date in mind. Somewhere around 2040. Why this contention then that this rise will be possible within the next 10 years? Two laws come to mind. The first is called Moore’s Law. Named after Intel co-founder Gordon Moore, it states that the number of transistors per square inch on an integrated circuit doubles every year. So, machines double their intelligence every year. Wait. There is more. The second law is actually more of an adage than a law. Murphy’s law maintains “that which can go wrong, will go wrong.” So far we have talked about binary computing. Enter quantum computing. And as if that was not enough, as Ray Kurzweil pointed out in the book The Singularity is Near, humanity now has the access to much faster DNA computing which can adjust far more transistors per square inch than silicon-based flat chip. So, the acceleration that ensures untimely arrival of self-aware, super-intelligent AI is a foregone. All of this is based on what we already know about technology. There is no assurance that a faster mechanism will not enter the race soon.
As Max Tegmark has demonstrated in his book Life 3.0, unlike previous versions of intelligence on earth, AI has the capacity to redesign its software as well as hardware. And here is the kicker. Unlike the cost of human labour, information technology gets cheaper every day. The technological marvel called smartphone that fits the palm of your hand and must have been bought for a few hundred dollars would have been worth millions of dollars if not more only a decade ago if it existed back then.
So within a decade if this technology reaches its critical mass, jobs will disappear, the middle class will rapidly shrink and we will have social problems of all sorts. People like me however can find solace in the fact that annoying little articles mocking technology will stop because AI most likely will supplant The Economist’s science and technology editor as well.