Computer science affects the potential abilities of weapons just as much as it does talking refrigerators and email routing and everything else. One key area of effect will be to obsolete vehicles slower than 500 mph like tanks and attack helicopters. This is because these vehicles which cost over $1 million each can be fairly easily destroyed with missiles costing less than $1000 each. Currently these vehicles are still operational because the production of such missiles has been artificially restricted. The United States allows only a few favored contractors to build such missiles using old and inferior designs and in limited numbers. The main such missiles are the "Red Eye" and "Stinger". Curiously enough other large countries are just as jealous of military production as the US and do exactly the same thing with one twist: instead of making their own inferior missile they just copy the US's inferior one. For example China has the Qianwei-1 and Anza missiles.

I think the US pays about $50,000 for each Stinger and in the open market in Afghanistan or Pakistan they sell for $200,000. However this will change dramatically in the future. The real cost of producing these devices is less than $1000 per item and only crumbling governmental restrictions prevent them from being more widespread. These devices are are so easy to make that even Third World countries could easily make them if they so desired. When this happens helicopters will be unable to operate in hostile areas.

The story is similar for tanks. A whole bunch of different missiles like the "Copperhead" are making tanks very unhealthy places to be. If you are interested in the full details of tank obsolescence you can read this paper by Major Ralph L. Schutte, a Canadian Army analyst. I am not a military equipment specialist so I cannot evaluate the kind of minutia that Schutte can, but what I can tell you is that the if the production of missiles was not restricted they would be significantly cheaper and more effective than the ones Schutte analyzes in his paper.

An important thing to be aware of is that the critical part of the missile is its brains--the part the computer scientist makes. Everything else is secondary. The common view of the "Stinger" is that it is a single device. Yet this is very wrong. In fact the Stinger (made by General Dynamics) has many "variants". Often the difference between these variants is just a different program, but the difference is huge. It's like the difference between Microsoft Paint and Adobe Illustrator. Currently these programs are typically written by a small, incestuous group of dopey engineers who are trained in electrical engineering, not computer science, and don't know linear optimization from a hole in the ground. If you made production of this software open to the free market and allowed real computer scientists to write it the results would scary: missiles that would not miss.