Computers have never been fast enough. From their very beginnings in antiquity as abaci to the building-sized supercomputers of today, the cry has gone up "Why is this taking so long?"

This is not an idle complaint. Humanity's ability to control the world depends on its ability to model it and to simulate different courses of action within that model. A medieval trader, before embarking on a trading mission, would pull out his map (his model of the world) and plot a course (a simulation of his journey). To do otherwise was to invite disaster. It took a long period of time and a specialized skill set to use these tools. A good navigator was an important team member. To go where no maps existed was a perilous journey.

The same is true today, except that the models have become larger and the simulations more intricate. Testing a new nuclear missile by actually launching it is ill-advised. Instead, a model of the missile is built in software and a simulation of its launching is run on a computer. Design flaws can be exposed in the computer (where they are harmless), and not in reality.

Modeling a missile is much more complex than modeling the course of a ship. There are more moving parts, the relevant laws of physics are more complicated, the tolerance for error is lower, and so on and so forth. This would not be possible without employing more sophisticated tools than the medieval navigator had access to. In the end, it is our tools' abilities that limit what we can do.

It is the nature of problems to expand to fill the limits of our capability to solve them. When computers were first invented, they seemed like the answer to all our problems. It did not take long before new problems arose.