AI superintelligence, should it obtain the propensity to end humanity, could kill humans in inconceivable ways, says The Times's science writer Tom Whipple. “You cannot conceive of what a superintelligence will do to kill you, just as ants cannot conceive of the technologies we will deploy to kill ants.”

"inconceivable ways"? “you cannot conceive of what a superintelligence will do” is an epistemological nightmare. how do you examine this claim whose missing mechanism is supposedly beyond your capacity to understand? ask for any pathway and the answer is that you’re not intelligent enough to imagine it. that's literally just unfalsifiable by construction (Karl Popper would have questions). also: ants don’t lose to us because we score higher on reasoning tests. humans have the physical means to act on that advantage. meanwhile, the octopus remains impressively intelligent and regularly served as calamari*. intelligence doesn’t make the hardware, energy, manufacturing, and logistics etc. suddenly appear, i.e. intelligence isn’t the kill chain. *(yes, ok, I admit it, that’s technically squid. but the octopus isn’t running the restaurant either).

This BS narrative is being pushed so hard by all the Left affiliated media at the same time.

Despite its forecast ability to, is there anything in there that would actually motivate it to?

Absolute nonsense We can defeat this superintelligence with a bucket of water

Thought you had to be smart to be a science writer?

I feel safe in the knowledge that if all else fails the wife will give it one of her looks.