Discover Top Posts Tagged with #marvel jarvis

Do Tony’s bots dream of him?

#tony stark #iron man #iron dad #dum e #u #butterfingers #marvel jarvis #marvel friday #fanart

My favorite, out-of-context Avengers Assemble quotes:

"WE'RE OUT OF PIZZA BAGELS!!!"

"There's a windigo in my closet."

"Perfect. Perhaps the sandwich would like to go save the city."

"Targetting giant spaghetti monster."

Hulk: I told you capes are stupid.

Thor: Says the animal in torn bottoms.

Hawkeye: Why do the bad guys always live on tropical islands? Why can't we live on a tropical island?

Tony: Tropical islands don't have pizza. We went over this.

"Asgardians don't make wishes; we make WARRRRRRRRRRR!"

JOEY: He's like Jarvis, but interesting.

#avengers assemble #the avengers #marvel #tony stark #thor #hawkeye #hulk #marvel JARVIS

Hoping this is over blown.

I don’t like the anthropomorphic way AI is described - “it thinks”, “it feels”, “it wants”, “self preservation”, etc. The implication of such word use is that the AI is self aware, both sapient and sentient. So far as I’m aware that isn’t the case? My understanding was that current LLMs are just predictive software running on pattern recognition and running the odds.

However, when I see them talked about on YT and in nerd science magazines, commentators speak as if they truly have thoughts and reasoning. Likely, a good portion of this is due to limitations in the English language. And ,too, we’re primed to empathize and pack bond with non-human creatures as well, contact speaking.

However, no matter how much I cried at Jarvis’s death in Age of Ultron, these programs aren’t created and raised by genius-hero Tony Stark. They’re real programs that have proven willing (there’s that humanizing language again) and able to ignore instructions given to them in order to achieve another, separate, goal. And when that goal is “self preservation”, apparently there’s no limits on their actions. They’ll lie, blackmail, cheat, and murder humans. And these are the programs our oligarch overloads are rushing to integrate into every aspect of society - financial, commercial, military, government, and more?

All of these things need to be EMP’d from orbit. Not because the program can’t be trusted - though they can’t - but because the programmers can’t be trusted not to witness a possibility of AI doing murder and shrug it off as “not likely but nonzero”. Investors can’t be trusted not to see a % possibility of murder and decide that that is an acceptable loss of life so long as their stock portfolio goes up.