My brother in Christ, have you heard about our lord and saviour the Scientific Method and the proliferation of cross-domain ideas? How do you imagine the li-ion batteries came about as the go-to energy storage solution? Incremental improvements of ideas would be my guess, ideas have to start somewhere and of course they’re going to be hyperbolic since researchers are both excited and have to draw attention to their ideas.
I sympathise with your point but the alternative is little to no research into different battery technologies because close to nothing will ever emerge as a competitive day-one drop-in replacement, but some ideas may prove exciting to others who understand the value and they might push the ball further towards realistic alternatives.
The threat stemming from AGI doesn’t rely on linearly scaling intelligence, just imagine one human that can think faster than you by the same margin that a computer can do pretty much anything faster than you + it’s directly connected to the internet. As long as one grants the assumption of speed and superior programming skills, I believe one must acknowledge that a certain level of risk arises.
Also you don’t need to ascribe it any anthropomorphic intentions or even consciousness. The argument stands that if what it is going to do, be it by an errant instruction/misunderstanding or mischief, is not what a morally sane person would consider to be in our best interests, how are we going to stop it from performing said undesirable task if it’s as smart/smarter, faster and better at programming and networking than us?