This is the sixth and final post in a series summarising the key arguments of Ray Kurzweil’s The Singularity is near: When humans transcend biology. The previous posts were:
- Ray Kurzweil, The Singularity and the accelerating pace of progress,
- Kurzweil predicts personal computers with the power of the human brain by 2025, and
- Kurzwil predicts we will create software that emulates the human mind by reverse engineering the human brain
- Kurzweil on coming revolutions in genetics, nanotech and robotics (strong AI)
- Kurzweil part 5 – the impact
Most of the disagreement with Kurzweil’s thesis stems from what he describes as ‘incredulity’ or ‘simple disbelief that such profound changes could possibly occur’. Many people think he is simply talking about too much change too quickly. The key counter arguments are variations of ‘the world is more complex than Kurzweil allows for and these changes are far more difficult/will take longer/may never happen’.
These arguments are applied in the specific to the different strands of Kurzweil’s Singularity thesis – e.g. hardware development, re-engineering the human brain in software and the nanotech revolution. For examples look in the comments on the posts in this series, or check out this recent conversation on Hacker News where people debate the depth of Kurzweil’s understanding of the brain and one contributor suggests it will take 75 years for us to reverse engineer the brain’s functionality, rather than the 20 years Kurzweil is predicting.
Kurzweil dedicates a chapter of The Singularity to countering the arguments of his detractors. His broad response to the ‘incredulity’ criticism is to point out the long history of exponential rates of increases in all key technologies (covered in the first post in this series) and to repeat the point that our brains aren’t wired to notice when we are on exponential curves, but instead interpret the progress as linear (i.e. the tangent of the exponential) and hence routinely underestimate the pace of change. He also notes (perhaps displaying a little hubris…) that humans have a long history of resisting notions that threaten the accepted view that our species is special – be it Copernicus’s insight that the earth was not at the centre of the universe or Darwin’s idea that we are only slightly evolved from other primates.
My view, having soaked up a lot of Kurzweil related material over the last few weeks, is that (rather boringly) the truth lies somewhere in the middle. Kurzweil has done a lot of good thinking and makes a lot of predictions across a lot of areas. Inevitably some of them will be right, and some will be wrong, particularly when it comes to timing. For me the important thing is not exactly when the brain will be reverse engineered or we will have nano-bots in our blood stream, but rather that we are headed in that direction. This last point is one that not too many people seem to take issue with. I also think that Kurzweil suffers from being a generalist, and as such he attracts criticism from specialists wanting to defend their turf.
Beyond incredulity there is one other objection to Kurzweil’s theories that I’m going to cover off in a little detail here, which is the argument that exponential trends don’t last forever. Proponents of this criticism point out that most exponential trends we observe hit a wall, usually driven by the environment. E.g. human population growth tails off when population density approaches critical levels. Kurzweil’s response is to me convincing. He recognises that the tailing off in computing power will come, but argues it is a long, long way off. To prove the point he describes at some length how we will be able to use more and more of the matter in the universe for computational purposes, and shows how this will provide all the computing power we need to get us a long way past the singularity.
For completeness I’m going to close by noting that there are a number of other criticisms that are beyond the scope of this post and which in my opinion Kurzweil adequately covers off in his book – e.g. brains are too complex to model, quantum computing will be required and is impossible, lock in to legacy technologies will prevent progress, and various philosophical and religious arguments which hold machines can’t be conscious. In most of these cases my (relatively uneducated) view is that Kurzweil’s arguments are much the stronger, although none of these issues are black and white.
So that is it for the ‘Kurzweil series’ – I haven’t enjoyed writing it as much as I expected I would, largely because the subject is much heavier than my normal posts. That said, I have learned a lot :-). Thanks to everyone who commented and retweeted.