I’ve got a new thought experiment, along the lines of the Lottery Question, only this one is aimed at computing and artificial intelligence.
What if you were given a computer with infinite computing power? Or, to be a bit more realistic, you suddenly had all of the world’s computing power at your disposal. What software would you run? What new software would you write? What problem could you solve, and how?
Let’s imagine there is a way of measuring the importance or quality of the results of a computer program. Does the importance or quality of the results increase linearly with the increase of computing power?
It seems clear that in many cases, doubling the speed of your laptop’s CPU does not double the quality of a term paper, or cut the time it takes to write it in half. Instead, it might decrease the chance of a spelling or grammar error slightly, or make it possible to get the layout correct in less time. Difficult to quantify but most people would probably say it is less than a 100% increase in quality.
My NaNoWriMo novel from last year covers this topic a bit. A couple of the characters study AI in college and, after a few years working as programmers, they start a project to integrate various AI mechanisms to solve some difficult problems. Eventually they enhance the system until it can teach itself, creating a feedback loop of improvement and learning. The idea is that, at this point, any increase in computing resources does result in a corresponding increase in the quality of the results. They use the software to create new hardware for itself to run on and soon they have unimaginable computing power.
Note: the system in the novel does NOT suddenly wish to exterminate humanity. Much like we are bound to our bodies and our instincts (to eat, sleep, stay alive, have sex) the software is bound to its hardware and its “instincts” (solve the given problem) which have been coded in the lower levels of the system.
On a related note, this book looks interesting: The Myths of Innovation
W. and I often talk about creativity and innovation, generally agreeing that most things are a product of their environment. It is extremely difficult to make any large jumps from the well traveled paths. Not because we are brainwashed by modern society, but because it is just difficult. How can we be sure there is an answer out there in the infinite number of possibilities? How do we know it will be better than current solutions? How do we search through the infinite possibilities in finite time?
Anyways, I find all of those issues fascinating and need to read more.
When the kinks are worked out of this type of netbook, and the price drops even further, it is going to have a huge impact on the industry. Apparently it already is.