This is interesting, and a little disturbing. We're going to have to come up with better ways to fight these things.

the best first defense is diversifying our computing environment. So far, worms have spread mostly thanks to predominance of Windows on desktops ( servers are more closely guarded by their users than desktops ). No doubt, that if MacOS and x86 Linux gain more market share, these will be targeted as well ( cross-platform worm would be pretty bad .. )
However, in more diversified computing environment its harder for the worms to survive and multiply, simply because it would be technologically way more challening to engineer and plan for and would compromise traceability, i.e. such a thing would have much plainer signature for detection.

I suggest removing bugs from your code.

Kind of a weird research study.

To simulate the effects of a rapidly mutating worm, which beats defenses by constantly changing its strategy, she assumed that even patched computers remain susceptible to new attacks by the same worm.

Uh...if they were that smart, they wouldn't be worms, they'd be massive complex applications that would probably crash 8 times out of 10 as soon as they loaded.

Why not study a worm that used N-rays to control the mind of human users while you're at it?

But it's certainly understandable that Microsoft would support the view that the basic problem with computer security is the fiendishly clever designs of virii and worms, rather than (say) the really stupid designs of operating systems.