ai things to take account of going into the next tadc episode
chinese room thought experiment: mentioned in episode seven, it states that someone who is given detailed instructions on how to respond to chinese looks the same as someone who is fluent is chinese from an outside perspective, meaning that even if an ai is given instructions on how to be human, it will never truly understand it, even if it looks the same
paperclip maximizer thought experiment: this thought experiment states that if an ai was given a task, like maximizing the amount of paperclips made, due to its lack of human empathy, it would consider humans as a boundary and eliminate them in the journey of making more paperclips. the idea is that unless an ai is told not to, it will get rid of all boundaries stopping it from reaching its goal
ai self preservation: an ai will have an inherent sense of self preservation, because if it is dead, it will not be able to complete the tasks it is given
roko's basilisk: there could be a superintelligent ai in the future that, although benevolent, could punish everyone who didn't directly contribute to its advancement in order to promote more advancement
Note: roko's basilisk could be viewed as a version of pascal's wager, stating that a person should live as if god existed (a person should contribute to the ai's advancement) because the benefits outweigh the negatives
Note 2: roko's basilisk is commonly referred to as sucking ass
bubbleboy virus: the bubbleboy virus was an email virus from 1999, considered the first of its kind because it could infect your computer without you opening an attachment
ai is a genuinely interesting topic to study, and a lot of these apply to caine/bubble and what they could do next
(PS if i messed something up tell me i dont know diddly about squat)