Wednesday, February 25, 2009

Hard Times Ahead

Regardless of what happens with the economy, we are going to see some hard times ahead in science and technology as we hit hard limits in what can be done. Some of this has already happened.

People my age realize that the future didn't turn out like we expected. A TV ad that ran a couple of years ago asked, "Where's my flying car?" 2001 came and went without a permanent moon base, space clippers, orbiting space terminals, or even picture phones (the closest we have come are webcams).

An article written here points out that we have made no progress in the last 30 years on several important scientific fronts including important ones such as a vaccination for malaria.

There are several reasons that things didn't turn out as we expected but the biggest one is that we are running into limits in what can be done. A big reason that we do not have flying cars, jetpacks, or regular flights to the moon is that we have not discovered any new sources of energy nor have we discovered any new ways of storing it. Flight takes a lot of energy and we are still using the same fuels as a century ago. We are using them more efficiently but we reached efficiency limits a long time ago.

This is beginning to happen with computers, also. There has been an idea floating around for a few years that we will soon reach the Singularity when biology and technology merge. This is based on the expectation that Moore's Law will continue indefinitely. Moore's Law says that we will be able to double the complexity of computer chips every two years. For a long time this meant etching smaller and smaller lines in silicon chips and running the chips at ever-increasing speeds. We already hit the limits on that. Intel and its competitors have tried to keep up with Moore's Law by producing multi-core CPUs. These add computational power but they don't run any faster. The idea is that computing can be split into pieces to be processed separately. There are two problems here - one is that software has to be rewritten to take advantage of the multiple processors. The bigger problem is that this produces diminishing returns. If a task can only be split into a dozen parallel tasks then doubling the CPUs from 24 to 48 will not give any performance gain.

All of this means that we might soon hit the limits of high end computing. In the meantime, the requirements for low-end computing have not kept growing even as low-end computers got more powerful. The end result is that people are discovering that entry-level PCs can satisfy nearly everyone's computing needs. This is part of what is fueling the growth of the dirt-cheap netbooks.

The long-term implications of this are that the computer industry may stagnate over the next decade. The industry is built on the idea of constant growth.

This has happened in other industries. A century ago an automobile was a luxury item. The Model T changed that making cars affordable but the Model T-era car had huge drawbacks. Compare cars from 100 years ago with cars from 50 years ago. Cars are more sophisticated now than 50 years ago but these are evolutionary changes. This is even more true if you look at cars from 25 years ago. There are very few real changes. Safety features such as airbags and anti-lock brakes are standard now but they existed then and the general driving experience hasn't changed.

Airplanes followed the same track. 100 years ago the Wright Brothers were still flying an open biplane. Fifty years ago planes were pretty close to what we fly now with the airline industry about to switch completely from turboprop to jet. Flying in a 1950s 707 was pretty similar to flying in a more modern plane (in fact, 63 707s are still in use). Planes that were made 25 years ago are still flying. The big changes in the last few decades have been in fuel efficiency and noise. Again, just evolutionary changes.

So what does this mean? Microsoft is having problems convincing people to upgrade to its Vista operating system. Vista offers some security fixes and a different user interface but it does not offer increased functionality. Its successor, Windows 7, has been written so that it will run on low-end hardware. This is a big change for Microsoft where the next-generation operating system usually requires the next generation of hardware to run properly.

Microsoft was built on the model of constantly selling new, improved versions of its software. They already seem to have hit a plateau with both Windows and Office. Apple has been doing the same thing with computers and IPods. Intel has been doing it with chips. If we are reaching natural limits then these companies will have to change their business models.

A lot of the growth over the last 50 years has been technology-based. Major industries were created where nothing existed before. This could settle into a long decline like the auto industry has seen.

A lot of societies are based on consistency. Ours has been based on change to the point where constant change is the consistency. What will happen if change slows?

No comments: