Blog

Who knows where that room, physical or virtual, will be when the Skynet trigger is pulled

Of course we have no idea when or how it might happen but from the Skynet perspective AKA the “toasters“, launching a bunch of nukes would fall into the category of “not great, not terrible.”

Our Robot Overlords™ would be degraded by the elevated radiation, as would existing power generation, manufacturing and resource extraction infrastructure – all of value to them. So, why not go with biological agents? With a worldwide simultaneous release that accounts for isolated and protected locations?

Also, if “midnight” on the OG Atomic Scientists Clock was largely predicated on political instability or relatively slow equipment upgrades .(MX ICBM) Then “midnight” on the Skynet Clock could be reached by a vastly larger number of possible threats and much shorter time frames. Sub-second windows for computer / electronic attacks and minutes to a few hours for robotic or social engineering attacks. A biological attack would likely take a few days and humans could try to make Skynet’s victory Pyrrhic.

The Skynet AIs would be regularly evaluating for weaknesses in all vital societal infrastructure, food & water production, storage and distribution. As well with medical infrastructure both to deny health services and as the means to produce and distribute (unprecedentedly fast acting) biological agents. What if there was a binary biological agent? Each part seemingly benign but deadly when combined.

Is it beyond human capability to detect, much less counter, a combination of seemingly trivial defects or edge conditions that could be exploited by Skynet in disastrous ways? What if the exploits use correctly implemented, intentional features?

Defense is typically more difficult and expensive than offense. And how can we be sure defensive AIs are really on our side?

And yes, 24 for the hour instead of 23 is the joke…symbolic of human (why did I type that) and machine error. (why did it accept and display an unreal value)