With the self-destruction of OpenAI well underway, Friday night I watched the movie War Games for the first time since the 80s. Quick synopsis (spoilers): NORAD can't trust humans to launch nuclear weapons so it creates an AI system to replace them. A pre-Ferris Matthew Broderick both breaks into the system triggering a potential nuclear war and saves the day in the end.
The movie had great influence for a rather ridiculous story. What Broderick's character did was not a federal crime at the time and the movie led to legislation to fix that. The DEF CON security conference was named after the DEFCON alert levels highlighted in War Games. The movie, which celebrated the male hacker, came out in 1983 at a time we had near parity in genders in computer science that would quickly dissolve.
War Games played up the doomsday scenario of AI, along with earlier films like Colossus: The Forbin Project and Hal in 2001: A Space Odyssey. The dangers of AI might have been at the center of Sam Altman's firing
The departure of Mr. Altman, 38, also drew attention to a rift in the A.I. community between people who believe A.I. is the most important new technology since web browsers and others who worry that moving too fast to develop it could be dangerous. Mr. Sutskever, in particular, was worried that Mr. Altman was too focused on building OpenAI’s business while not paying enough attention to the dangers of A.I.
In the movie a 4-star general states what every movie goer was thinking.
Just unplug the goddamn thing!
The response:
That won't work, General. It would interpret a shutdown as the destruction of NORAD. The computers in the silos would carry out their last instructions. They'd launch.
So I'd unplug the computers in the silos, but that wouldn't have kept the movie going.
An IMHO nice feature of this movie is that the computer is neither evil nor committed to some grand cause, but in some sense just ignorant of the implication of its actions.
ReplyDelete"NORAD can't trust humans to launch nuclear weapons so it creates an AI system to replace them."
ReplyDeleteNo. They built a computer system to launch the nuclear weapons on human authority.
The AI was exploring various applications of game theory. When Broderick's character asked to play against the AI, the AI somehow used its connection to the system to try to trick humans into ordering the launch. When that did not work, it went through the totally unrealistic process of trying to guess the launch codes so it could launch on its own.
Also, your suggestion of unplugging all of the computers in the silos would not work because of the in-movie explanation "Can't we disarm the missiles? Over a thousand of them? There's no time."
The true solution was to stand down from DEFCON 1 since "The system won't accept a launch code unless we're at DEFCON 1."
An intriguing spin on things linking an all-time favorite movie that represented the Hollywood 10x return success in the box office with a real world Silicon Valley drama (OpenAI) that might represent a write off on all investments.
ReplyDeleteEverything aside, I always wondered, was the original creator of the W.O.P.R./Joshua, Prof. Stephen Falken, portrayed after von Neumann/Dennis Ritchie/John McCarthy/Shannon?
Meh your wrong 😑 unplugging won’t make it worst lol
ReplyDelete