Innovation for Jobs

Innovation for Jobs

Innovating a human
future of work

The People Centered Economy

Timing the Singularity, v2.0

Author: Kartik Gada

Exactly 10 years ago, I wrote an article presenting my own proprietary method for estimating the timeframe of the Technological Singularity. Since that time, the article has been cited widely as one of the important contributions to the field, and a primary source of rebuttal to those who think the event will be far sooner.

Exactly 10 years ago, I wrote an article presenting my own proprietary method for estimating the timeframe of the Technological Singularity. Since that time, the article has been cited widely as one of the important contributions to the field, and a primary source of rebuttal to those who think the event will be far sooner.  What was, and still is, a challenge is that the mainstream continues to scoff at the very concept, whereas the most famous proponent of this concept persists with a prediction that was too soon, which will inevitable court blowback when his prediction does not come to pass.  Now, the elapsed 10-year period represents 18-20% of the timeline since the publication of the original article, albeit only ~3% of the total technological progress expected within the period, on account of the accelerating rate of change.  Now that we are considerably nearer to the predicted date, perhaps we can narrow the range of estimation somewhat, and provide other attributes of precision.  

In order to see if I have to update my prediction, let us go through updates on each of the four methodologies one by one, of which mine is the final entry of the four.  

1) Ray Kurzweil, the most famous evangelist for this concept, has estimated the Technological Singularity for 2045, and, as far as I know, is sticking with this date.  Refer to the original article for reasons why this appeared incorrect in 2009, and what his biases leading to a selection of this date may be.  As of 2019, it is increasingly obvious that 2045 is far too soon of a prediction date for a Technological Singularity (which is distinct from the ‘pre-singularity’ period I will define later).  In reality, by 2045, while many aspects of technology and society will be vastly more advanced than today, there will still be several aspects that remain relatively unchanged and underwhelming to technology enthusiasts.  Mr. Kurzweil is currently writing a new book, so we shall see if he changes the date or introduces other details around his prediction.  

2) John Smart’s prediction of 2060 ± 20 years from 2003 is consistent with mine.  John is a brilliant, conscientious person and is less prone to let biases creep into his predictions than almost any other futurist.  Hence, his 2003 assessment appears to be standing the test of time.  See his 2003 publication here for details.  

3) The 2063 date in the 1996 film Star Trek : First Contact portrays a form of technological singularity triggered from the effect that first contact with a benign, more advanced extraterrestrial civilization had on changing the direction of human society within the canon of the Star Trek franchise.  For some reason, they chose 2063 rather than a date earlier or later, answering what was the biggest open question in the Star Trek timeline up to that point.  This franchise, incidentally, does have a good track record of predictions for events 20-60 years after a particular Star Trek film or television episode is released.  Interestingly, there has been exactly zero evidence of extraterrestrial intelligence in the last 10 years despite an 11x increase in the number of confirmed exoplanets.  This happens to be consistent with my separate prediction on that topic and its relation to the Technological Singularity.  

4) My own methodology, which also gave rise to the entire ‘ATOM’ set of ideas, is due for an evaluation and update.  Refer back to the concept of the ‘prediction wall’, and how in the 1860s the horizon limit of visible trends was a century away, whereas in 2009 it was in perhaps 2040, or 31 years away.  This ‘wall’ is the strongest evidence of accelerating change, and in 2019, it appears that the prediction wall has not moved 10 years further out in the elapsed interval.  It is still no further than 2045, or just 26 years away.  So in the last 10 years, the prediction wall has shrunk from 31 years to 26 years, or approximately 16%.  As we get to 2045 itself, the prediction wall at that time might be just 10 years, and by 2050, perhaps just 5 years.  As the definition of a Technological Singularity is when the prediction wall is almost zero, this provides another metric through which to arrive at a range of dates.  These are estimations, but the prediction wall’s distance has never risen or stayed the same.  The period during which the prediction wall is under 10 years, particularly when Artificial Intelligence has an increasing role in prediction, might be termed as the ‘pre-Singularity’, which many people will mistake for the actual Technological Singularity.  

Singularity

Through my old article, The Impact of Computing, which was the precursor of the entire ATOM set of ideas, we can estimate the progress made since original publication.  In 2009, I estimated that exponentially advancing (and deflation-causing) technologies were about 1.5% of World GDP, allowing for a range between 1% and 2%.  10 years later, I estimate that number to be somewhere between 2% and 3.5%.  If we allow a newly updated range of 2.0-3.5% in the same table, and an estimate of the net growth of this diffusion in relation to the growth of the entire economy (Nominal GDP) as the same range between 6% and 8% (the revenue growth of the technology sector above NGDP), we get an updated table of when 50% of the World economy comprises of technologies advancing at Moore’s Law-type rates.  

We once again see these parameters deliver a series of years, with the median values arriving at around the same dates as aforementioned estimates.  Taking all of these points in combination, we can predict the timing of the Singularity.  I hereby predict that the Technological Singularity will occur in :

2062 ± 8 years

This is a much tighter range than we had estimated in the original article 10 years ago, even as the median value is almost exactly the same.  We have effectively narrowed the previous 25-year window to just 16 years.  It is also apparent that by Mr. Kurzweil’s 2045 date, only 14-17% of World GDP will be infused with exponential technologies, which is nothing close to a true Technological Singularity.     

So now we know the ‘when’ of the Singularity.  We just don’t know what happens immediately after it, nor can anyone with any certainty.

Tanya