Views: 3261
Submissions: 62
Favs: 125
~KitXune
Gender: male.
Orientation: straight.
Religion: none.
Likes: cute things and thought-provoking things, darkness and bright-light, overcast or vivid-blue skies, silence, time, innocence and naiveté, optimism, open-mindedness, observing things and people from a distance and, when I have the courage to do so, experiencing new things and new places.
Orientation: straight.
Religion: none.
Likes: cute things and thought-provoking things, darkness and bright-light, overcast or vivid-blue skies, silence, time, innocence and naiveté, optimism, open-mindedness, observing things and people from a distance and, when I have the courage to do so, experiencing new things and new places.
Featured Submission
Recently Watched
Stats
Comments Earned: 626
Comments Made: 469
Journals: 11
Comments Made: 469
Journals: 11
Recent Journal
Will future robots be hostile?
12 years ago
I wrote this masterpiece in the comments section of a youtube video (in many parts). I felt it deserved better than that, so I'm also posting it here. This was written in response to a user who stated that there's no reason to believe robots of the future will be capable of hostility (we've been conversing for a while).
I'm going to take a computer-programming approach to this. In some programming languages, there's a distinction between a function's interface and its implementation.
The interface defines what the function's inputs and outputs are, and often includes a comment describing what it does and how it should be used. The implementation is the actual compilable code that does the work.
In theory, if you want to build upon source code someone else wrote, you should be able to learn how to use a function by merely referring to its interface, without bothering to understand the implementation.
For now, let's say we completely disregard the implementation. If we look at human emotion from this vantage point, it doesn't matter how it works -- only what its end results are.
Now, a hostility function in an AI might be implemented completely differently from a hostility function in a human, but do the interfaces match? If they do, then they're effectively the same thing. We can call the function, and it will serve the same purpose.
To predict whether AIs in the future will have a hostility function, we need only consider whether a programmer would have cause to include one.
I contend that a sufficiently complex AI created with an objective in mind might possess a hostility function or something similar. I think a hostility function serves a number of purposes, and that programmers will inevitably mimic it.
I'm going to take a computer-programming approach to this. In some programming languages, there's a distinction between a function's interface and its implementation.
The interface defines what the function's inputs and outputs are, and often includes a comment describing what it does and how it should be used. The implementation is the actual compilable code that does the work.
In theory, if you want to build upon source code someone else wrote, you should be able to learn how to use a function by merely referring to its interface, without bothering to understand the implementation.
For now, let's say we completely disregard the implementation. If we look at human emotion from this vantage point, it doesn't matter how it works -- only what its end results are.
Now, a hostility function in an AI might be implemented completely differently from a hostility function in a human, but do the interfaces match? If they do, then they're effectively the same thing. We can call the function, and it will serve the same purpose.
To predict whether AIs in the future will have a hostility function, we need only consider whether a programmer would have cause to include one.
I contend that a sufficiently complex AI created with an objective in mind might possess a hostility function or something similar. I think a hostility function serves a number of purposes, and that programmers will inevitably mimic it.
User Profile
Accepting Trades
No Accepting Commissions
No Character Species
Fox
Favorite Music
Electronic
Favorite TV Shows & Movies
Fantastic Mr. Fox
Favorite Games
Pokemon Black 2, Minecraft, Braid, Super Meat Boy
Favorite Gaming Platforms
DS
Favorite Animals
Foxes
Favorite Site
YouTube
Favorite Foods & Drinks
Mexican, Indian
SleepyHelen