For the past couple of days, I have been getting reacquainted with all of the writing and notes and ideas I was engaged with during fall quarter for my Artificial Intelligence project. Involved in this process is the reading of the massive quantity of journaling that I did, and watching and reading some new research material. Examples of this material include reading new posts on the Accelerating Future blog, watching some lectures from the Singularity Summit at Stanford, featuring a number of speakers from the Singularity Institute for Artificial Intelligence, and other places.
The next two weeks are not going to be about doing a ton more research however. They are going to be about answering the concrete questions necessary to actually start visual and auditory work on this project. What does the world portrayed in this animation actually look like? How will the presence of the Artificial Intelligence be manifested or embodied? How will the AIs experience of reality be embodied? How should a viewer experience this work of animation? What should they get from watching it? What are the larger issues that are important to be addressed with this animation?
Below are a smattering of notes from early portions of the aforementioned research.
what type of embodiment (that is, manifestation of experience of reality… — the perceived structure that one’s reality contains) would the AI’s experience of reality be mediated by?
how would knowledge be encoded? through language? through symbols?
Notes on “The Human Importance of the Intelligence Explosion” – a lecture at the Singularity SUmmit at Stanford.
intelligence is a ‘super power’ that all humans share. it is the foundation of all our creations. technological progress is a byproduct of intelligence.
the technologies of the future that will really matter, are those that act on the amplification of cognitive powers: nuerotechnology, brain-computer interfaces, and artificial intelligence.
intelligence is the most powerful force in the universe, and also the most non-understood. We know a great deal about the mind, but there are still things we don’t understand about it yet. We have found many “mysterious questions”, but no “mysterious phenomenon”.
the brain is the weak side of the brain-computer interface. human brains are not
molecular nanotechnology controlled by a super-human AI is a possibility to advance human intelligence, but perhaps it is not the best strategy for an intelligence explosion.
a better environment would the predictable and malleable one inside of a computer, where changes have known consequences.
The power of intelligence isn’t about new technology and gadgets. The power of intelligence is about ingenuity and surprise, going outside of the rules of the game.
rule number 1: do not trust your sense of how ridiculous things sound, when considering things in the context of an intelligence explosion.
There are different types of intelligence that might result.
Cheesecake fallacy: whenever an argument leaps directly from possibility to actuality, without considering the manufacturative properties necessary for the latter. ERROR: motive, not manufactive. why do something if there is no reason to? What would an AI want?
Ai’s are commonly depicted as an “ethnic stereotype” — all the same. In reality, there are a large variety of possible artificial intelligences. When we talk about AIs, we are really talking about mind in general. Is the space of possible mind-designs smaller than the space of the design of the human mind? no. and yet futurists commonly speak of how AIs will be in the future as confidently or more so than they would predict the behavior of an actual human being.
a self improving mind can have stable motivational architecture. (it wouldn’t modify itself in a way that would alter its current motivations).
The real problem is not a predictive problem of determining what superhuman AIs might do. It is a question of engineering. One must know enough about the problem in order to “reach” into the possible mind-design space outside of a human mind and pull one out that is a desired type… carefully, in order to achieve some measure of the desired effect.
then, making the right choice in proposing a solution to the problem of AI creation, is the most challenging problem.
(a group that is trying to solve a very difficult problem that holds out on suggesting possible solutions until the problem is discussed so much that there is nothing left to say but possible solutions, have better solutions than those that suggest solutions right away)
Ben Goertzel: “Artificial Intelligence: Now Is The Time”
most types of technological advancement have positive possibilities and negative possibilities. how we choose to use this technology is a result of the wielding of our own powers of judgement, linked to our sense of human wisdom: what would be better, as chosen by our own sense of motivation and morality?
“Victim of the Brain,” based on The Mind’s I
Computers are a medium in which we can Model anything we want to. An approach to artificial intelligence is the modeling of the functioning of human thought – the human mind. This has the effect of allowing us to understand our own functioning better. …
is the algorithmic modeling of consciousness / intelligence a valid way of creating AI?
Central attribute of human intelligence: to take an unfamiliar situation and to immediately get at what the essential elements of that situation are, and even though you might have never been in that situation before, and to ignore the silly superficial aspects of it and get right to the deeper significance, = ability.
Continuing Thoughts About Artificial Intelligence (SOS: Media – Week 2 Update)
For the past couple of days, I have been getting reacquainted with all of the writing and notes and ideas I was engaged with during fall quarter for my Artificial Intelligence project. Involved in this process is the reading of the massive quantity of journaling that I did, and watching and reading some new research material. Examples of this material include reading new posts on the Accelerating Future blog, watching some lectures from the Singularity Summit at Stanford, featuring a number of speakers from the Singularity Institute for Artificial Intelligence, and other places.
The next two weeks are not going to be about doing a ton more research however. They are going to be about answering the concrete questions necessary to actually start visual and auditory work on this project. What does the world portrayed in this animation actually look like? How will the presence of the Artificial Intelligence be manifested or embodied? How will the AIs experience of reality be embodied? How should a viewer experience this work of animation? What should they get from watching it? What are the larger issues that are important to be addressed with this animation?
Below are a smattering of notes from early portions of the aforementioned research.
what type of embodiment (that is, manifestation of experience of reality… — the perceived structure that one’s reality contains) would the AI’s experience of reality be mediated by?
how would knowledge be encoded? through language? through symbols?
Notes on “The Human Importance of the Intelligence Explosion” – a lecture at the Singularity SUmmit at Stanford.
intelligence is a ‘super power’ that all humans share. it is the foundation of all our creations. technological progress is a byproduct of intelligence.
the technologies of the future that will really matter, are those that act on the amplification of cognitive powers: nuerotechnology, brain-computer interfaces, and artificial intelligence.
intelligence is the most powerful force in the universe, and also the most non-understood. We know a great deal about the mind, but there are still things we don’t understand about it yet. We have found many “mysterious questions”, but no “mysterious phenomenon”.
the brain is the weak side of the brain-computer interface. human brains are not
molecular nanotechnology controlled by a super-human AI is a possibility to advance human intelligence, but perhaps it is not the best strategy for an intelligence explosion.
a better environment would the predictable and malleable one inside of a computer, where changes have known consequences.
The power of intelligence isn’t about new technology and gadgets. The power of intelligence is about ingenuity and surprise, going outside of the rules of the game.
rule number 1: do not trust your sense of how ridiculous things sound, when considering things in the context of an intelligence explosion.
There are different types of intelligence that might result.
Cheesecake fallacy: whenever an argument leaps directly from possibility to actuality, without considering the manufacturative properties necessary for the latter. ERROR: motive, not manufactive. why do something if there is no reason to? What would an AI want?
Ai’s are commonly depicted as an “ethnic stereotype” — all the same. In reality, there are a large variety of possible artificial intelligences. When we talk about AIs, we are really talking about mind in general. Is the space of possible mind-designs smaller than the space of the design of the human mind? no. and yet futurists commonly speak of how AIs will be in the future as confidently or more so than they would predict the behavior of an actual human being.
a self improving mind can have stable motivational architecture. (it wouldn’t modify itself in a way that would alter its current motivations).
The real problem is not a predictive problem of determining what superhuman AIs might do. It is a question of engineering. One must know enough about the problem in order to “reach” into the possible mind-design space outside of a human mind and pull one out that is a desired type… carefully, in order to achieve some measure of the desired effect.
then, making the right choice in proposing a solution to the problem of AI creation, is the most challenging problem.
(a group that is trying to solve a very difficult problem that holds out on suggesting possible solutions until the problem is discussed so much that there is nothing left to say but possible solutions, have better solutions than those that suggest solutions right away)
Ben Goertzel: “Artificial Intelligence: Now Is The Time”
most types of technological advancement have positive possibilities and negative possibilities. how we choose to use this technology is a result of the wielding of our own powers of judgement, linked to our sense of human wisdom: what would be better, as chosen by our own sense of motivation and morality?
“Victim of the Brain,” based on The Mind’s I
Computers are a medium in which we can Model anything we want to. An approach to artificial intelligence is the modeling of the functioning of human thought – the human mind. This has the effect of allowing us to understand our own functioning better. …
is the algorithmic modeling of consciousness / intelligence a valid way of creating AI?
Central attribute of human intelligence: to take an unfamiliar situation and to immediately get at what the essential elements of that situation are, and even though you might have never been in that situation before, and to ignore the silly superficial aspects of it and get right to the deeper significance, = ability.