The Coming Singularity: An Epiphany
After seeing Ray Kurzweil's TED talk about the blurring line between humans and machines I was inspired to pick up his book The Singularity Is Near: When Humans Transcend Biology.
The "Singularity" is a concept which will be familiar to many science fiction readers. The pace of technological change has been accelerating throughout human history. If that trend continues there will come a time where the rate of innovation becomes so great that we will have the equivalent of 100 years of scientific discovery and technological achievement every day. Such a world would be so fundamentally different from that which we live in today that we cannot even imagine what life would be like after that point. In this way it is similar to the singularity at the heart of a black hole, into which we cannot see and from which no information can escape.
By looking at a very wide range of technological indicators - memory density, computing power per dollar, genome sequencing, patents granted, labour productivity, etc - Kurzweil shows that the Singularity is not very far off: by 2018 we'll be able to buy computer memory equivalent to human memory capacity for about $1,000 and by 2045 the total "non-biological" computing capacity on the planet will be one billion times that of the entire human population. If you think that this is all wild-eyed prognostication, think again: that 2018 date comes not from Kurzweil but from the International Technology Roadmap for Semiconductors, which is used by Intel and much of the semiconductor industry for future planning.
I'm only about a quarter of the way through the book at the moment, so I don't want to give a review of it. Rather, I want to describe an epiphany that I had last night.
One of the problems that I've had while reading this book has been what I perceived as a disconnect between the accumulation of ever greater knowledge and its practical application. For example, we already know how to create synthetic genes and proteins and, like so many other technologies, the speed at which we can do it is increasing at an exponential rate. So I am completely comfortable with the idea that in the next 10 to 20 years machines will be capable of producing millions, if not billions, of synthetic proteins every day. Once we reach that speed it becomes cost effective to simply try every possible combination in a family of proteins to see which ones are most effective against a given disease1. We could easily end up with a cure for a new disease every week, then every day, and eventually every hour. So far so good.
But we're not going to just take the drugs coming out of these machines and start prescribing them to people. There will still need to be lengthy human trials. Regardless of how fast the machines find cures it will still be years before they will be declared fit for human use. Furthermore, with new drugs being created on an hourly basis, there simply won't be enough people to man all those millions of drug trials: the whole human population would soon end up doing nothing but testing drugs.
Then I had my epiphany.
One of the many technologies which is progressing at an exponential rate is the modelling of human biology. By the time we get to the point where new drugs are being discovered every hour, we will also have achieved a working model of human biology which, if not perfect, will certainly be good enough for testing drugs. There will no longer be human trials: new drugs will be tested in a few seconds inside a computer model before being made available to people. We might start out only allowing terminal patients access to such drugs, but as our confidence in the model grew we would eventually ease the restrictions and new cures would become available to disease sufferers within days of their discovery.
But what about tooling up the factories to produce the new drugs? Well, remember that by this time we have machines capable of producing millions of automatically designed drugs per hour. A cheaper version of such a machine could be used to create your prescription on demand at the local pharmacy, if not in your own home.
I'm starting to believe that the Singularity may indeed occur within my lifetime. Wouldn't that be cool?
1It's more likely that our improved understanding of human biology will allow us to directly design drugs for each disease. My point was that even without such a paradigm shift, simply increasing the speed with which we create synthetic proteins will still get us to the cure-an-hour point.
The "Singularity" is a concept which will be familiar to many science fiction readers. The pace of technological change has been accelerating throughout human history. If that trend continues there will come a time where the rate of innovation becomes so great that we will have the equivalent of 100 years of scientific discovery and technological achievement every day. Such a world would be so fundamentally different from that which we live in today that we cannot even imagine what life would be like after that point. In this way it is similar to the singularity at the heart of a black hole, into which we cannot see and from which no information can escape.
By looking at a very wide range of technological indicators - memory density, computing power per dollar, genome sequencing, patents granted, labour productivity, etc - Kurzweil shows that the Singularity is not very far off: by 2018 we'll be able to buy computer memory equivalent to human memory capacity for about $1,000 and by 2045 the total "non-biological" computing capacity on the planet will be one billion times that of the entire human population. If you think that this is all wild-eyed prognostication, think again: that 2018 date comes not from Kurzweil but from the International Technology Roadmap for Semiconductors, which is used by Intel and much of the semiconductor industry for future planning.
I'm only about a quarter of the way through the book at the moment, so I don't want to give a review of it. Rather, I want to describe an epiphany that I had last night.
One of the problems that I've had while reading this book has been what I perceived as a disconnect between the accumulation of ever greater knowledge and its practical application. For example, we already know how to create synthetic genes and proteins and, like so many other technologies, the speed at which we can do it is increasing at an exponential rate. So I am completely comfortable with the idea that in the next 10 to 20 years machines will be capable of producing millions, if not billions, of synthetic proteins every day. Once we reach that speed it becomes cost effective to simply try every possible combination in a family of proteins to see which ones are most effective against a given disease1. We could easily end up with a cure for a new disease every week, then every day, and eventually every hour. So far so good.
But we're not going to just take the drugs coming out of these machines and start prescribing them to people. There will still need to be lengthy human trials. Regardless of how fast the machines find cures it will still be years before they will be declared fit for human use. Furthermore, with new drugs being created on an hourly basis, there simply won't be enough people to man all those millions of drug trials: the whole human population would soon end up doing nothing but testing drugs.
Then I had my epiphany.
One of the many technologies which is progressing at an exponential rate is the modelling of human biology. By the time we get to the point where new drugs are being discovered every hour, we will also have achieved a working model of human biology which, if not perfect, will certainly be good enough for testing drugs. There will no longer be human trials: new drugs will be tested in a few seconds inside a computer model before being made available to people. We might start out only allowing terminal patients access to such drugs, but as our confidence in the model grew we would eventually ease the restrictions and new cures would become available to disease sufferers within days of their discovery.
But what about tooling up the factories to produce the new drugs? Well, remember that by this time we have machines capable of producing millions of automatically designed drugs per hour. A cheaper version of such a machine could be used to create your prescription on demand at the local pharmacy, if not in your own home.
I'm starting to believe that the Singularity may indeed occur within my lifetime. Wouldn't that be cool?
1It's more likely that our improved understanding of human biology will allow us to directly design drugs for each disease. My point was that even without such a paradigm shift, simply increasing the speed with which we create synthetic proteins will still get us to the cure-an-hour point.