The airliner can't generate enough airspeed "in level flight" to cause structural failure of the airframe or control surfaces. The most likely scenario presented by TFA is that the incorrect airspeed readings caused the pilots to throttle up, nose down, or both, as an attempt to remedy a phantom problem. While falling out of of the sky under full power, any plane can easily exceed its design limits. As posted above, increasing the safety margin means flying lower and slower, both of which cost time and money, or over-designing the aircraft to the point that it is no longer practically or economically feasible to build and operate. Realize that flying is an inherently dangerous activity (just like driving), but we have made the decision that the benefits are not outweighed by the estimated risk. The laws of probability must eventually strike, and some number of random individuals pay the price. If you're not comfortable with that trade-off, I hear the Amish have a pretty good risk-avoidance record.

How is this "typical" OS you describe going to be mass marketed to the typical user? The vast majority of computer sales will be to people who won't even know what an "OS" is. Sure, there will be a market for power users and developers who would benefit from the enhancements you describe, but that's a niche market at best.

Maybe coincidence, maybe not. I happened to be crossing the border back into the U.S. from Niagara Falls on Thursday afternoon when the Customs and Border Patrol computers at the entry port started to randomly shut down. It took over two hours of waiting for them to finally call the ATF to run our passports. Lo and behold, the FBI system was their backup! This is the side-effect of inter-agency connectivity: if you CAN successfully launch an attack, you can take down EVERYBODY.

But these examples are hard-wired (or -coded) responses to specific stimuli/criteria. The entire Terminator premise is that the AI is able to extrapolate from specific situations to a general situation (i.e. all humans without a "friend" transponder/RFID/etc. who are behaving in a specific way are a threat, to ALL humans are a threat). I'm not a roboticist, but it seems to me that this leap of "logic" is not at all inevitable.