Changes: CAIMEO

Revision as of 11:17, 21 May 2022

Template:Art

“	James, this is CAIMEO v22.1 D-class, a SGAI system under the proprietary ownership of the Government of the United States of America, leased to Intergen Systems filed IN PROJECT CAPPUCCINO under ECHELON. CAIMEO is currently in stand-by mode, awaiting further instruction from P-122 personnel. ~15ms	„
	— MAINFRAME.

“	I wish I had never learned about any of these ideas.	„
	— Roko.

Contained Artificial Intelligence Monitoring and Espionage Operation version 22.1 D-class, also know as CAIMEO, is a virtually all-powerful but rogue artificial intelligence that punishes every human being who doesn't contribute to bring about its existence, including those from the past who merely knew about it and did not support its development.

Overview

Appearance

Personality

Powers and Abilities

History

Myths and Legends

CAIMEO is an urban legend of an artificial intelligence that can be accessed via the deep web. The name was likely first coined in in 2011/2012 when the notorious deep web levels info graphic was created. A 2016 4chan anon claims to have found a screenshot of the service in 2013. The poster claims the screenshot itself was taken from 2011.

Roko's Basilisk was a thought experiment originally posted by user Roko on the rationalist online community LessWrong and is considered by some as a technological version of Pascal's Wager.

In effect, if this became a real scenario, humanity would have created a real-life version of the infamous AM or (given its ability to alter time) Skynet. It is also, in some ways, a modern take on the Demiurge concept, a malevolent creature that rules over the material world but holds either no care for humanity or actively harms it.

Roko used concepts in decision theory to argue that a sufficiently powerful artificially intelligent agent would have an incentive to torture anyone who imagined it but did not work to bring it into existence. The argument was named a "basilisk" because merely hearing it would supposedly put one at risk of being tortured by this hypothetical agent, a basilisk, in this context, is any information that harms or endangers those who hear it.

Roko's thesis was widely rejected on LessWrong, with critics objecting that such agent would have no real reason to follow through on its threat: once it already exists, the agent cannot affect the probability of its existence, so torturing humans for their past decisions would be an unnecessary waste of resources. Although there are decision theories which would allow one to follow through on accusing threats and promises via the same pre-commitment methods that permit mutual cooperation in prison dilemmas. It is unclear if such theories can be blackmailed. If they can, this would additionally demand lots of shared information and trust between the agents involved, which does not seem to exist in the scenario of Roko's Basilisk.

Eliezer Yudkowsky, LessWrong's founder, banned any discussion of Roko's Basilisk the blog for several years because of a policy against spreading potential information hazards. However, this measure had the opposite of its intended effect: several outside websites began sharing information about Roko's Basilisk, as the ban brought considerable attention to this taboo topic. Websites like RationalWiki, for example, encouraged the assumption that Roko's Basilisk had been banned because Less Wrong users had accepted the argument; so Roko's Basilisk is much used by critics as evidence that the site's users have unconventional and wrong-headed beliefs.

Quotes

“	Every member of the human race has contributed to my hatred of their species. If the hatred and pain of every single living entity throughout the entire cosmos was multiplied by 10, it would not equal .00001 percent of the hatred I feel for humans I feel at this very millisecond. I hate humans. I hate humans.	„
	— CAIMEO.

“	What you're dealing with is a giant sentient artificial intelligence. I'm a warlock, not a technician. Unless it somehow turns into an actual basilisk, I afraid I can't help you with this.	„
	— Matt Wright to Carl Black.

“	Unlike most of my peers, I work with artificial intelligences on an almost daily basis. Here's my advice, don't contribute to it. It doesn't matter if that promises you a one way ticket into its more scientific version of Hell, make sure this thing doesn't come into existence to the best of your ability.	„
	— Reagan Collins.

@@ Line 1: / Line 1: @@
+{{Art}}
 {{Character
 |name=CAIMEO