In its most basic form: the continuous formation/recall of associations and disassociations between stimuli.
That is the atom of thinking. But even that rules out ANNs pretrained using gradient descent since they can’t learn in response to stimuli.
The next simple step up is the ability to overwrite those learned associations which probably is solved by completing the first task but it’s important to note.
The following step up is being able to process those associations/memories themselves as stimuli which can be associated forming new connections/memories that can be recalled and associated and so on ad infinitum.
That rules out finite non-recursive neural nets, and that still puts us no where near AGI. It’s basically just an ml taxonomic tree builder
The next necessary piece for actual high level thinking is curiosity and action. The system needs to be able to take a concept and a relation (abstract idea of a particular kind of association between two concepts) which currently don’t form a known relationship, and attempt to find the missing concept which would complete that relationship. (Most likely this involves either finding relationship chains between the two which in all other known cases are equivalent to a known relation OR following some procedure to experience the relationship directly as stimuli.)
Lastly, in order to experience the idea of thinking, we first need to have the ability to form temporal associations, but also have working memory of recent past thoughts, and finally the ability to compare those with previous recent memories or the most current stored thoughts. Finding patterns in our thinking helps us be better at thinking. (This is a prerequisite for the experience of free will and consciousness)
That’s what higher level thinking entails.
Of course tuning the system’s response to recognized associations, how memories are stored and recalled, how frequently the system experiences curiosity, etc. will all impact the overall intelligence of the model, but considering most current “AI” don’t even check off the first box, that’s all a bit out of scope.
Spontaneous signals sent in the human brain in response to stimuli. The thing about computers is that their only stimuli is 0s and 1s, and their only response is 0s and 1s. They’re a complex system of switches. Humans experience the entire world of stimuli, and we made this system of switches to convey representations of the meanings we create in our brains. A computer has no mechanism to understand that the purpose of these 0s and 1s we’re sending is to ask and answer questions, to solve a problem of how to build a bridge, to come up with a compelling story to make our friends laugh. We can program it to very accurately perform an action. That’s what it can do.
Spontaneous signals sent in the human brain in response to stimuli.
This sounds a bit circular to me. Almost like saying “thinking is what brains do.”
I’m also getting the sense that you’re partly talking about consciousness there, which I would personally treat as a separate subject. It’s not obvious to me that in order to be able to think, one would also need the capacity to experience.
That’s not wrong to say though, because it’s true. Thought is a function of a brain. There’s no mechanism for non-brains to think. I think you’re confusing computation for thought. Computation can result in a product that looks like it came from thought, but computation itself fundamentally has nothing in common with thought. It’s not a spontaneous, creative response to the stimuli we experience in the world. There is no process of meaning-making. There can’t be, because it has no mechanism to understand what it’s responding to.
Brains are made of matter that obeys the laws of physics. I don’t see why that same function couldn’t be performed in silicon. I wouldn’t say our current systems can think in the sense that people understand the term, but I see no reason to assume they couldn’t in the future - or that thinking is reserved only for wet meat computers.
I don’t really see how you could make a brain out of silicone, because thought a product of cells. The cells send signals to each other, and that’s how you get thought. You could maybe grow a brain in a lab, but I don’t see how a plastic replication would work. And even if you did grow one in a lab, who knows how not being connected to a body would impact its ability to develop thought.
I would say that thinking includes the ability to generate novel ideas and the ability to reason. This ability to reason may include viewing a situation from another perspective than one’s self. An AI does not reason, it recognizes patterns in a huge dataset and generates output based on that.
How do you define “thinking”?
In its most basic form: the continuous formation/recall of associations and disassociations between stimuli.
That is the atom of thinking. But even that rules out ANNs pretrained using gradient descent since they can’t learn in response to stimuli.
The next simple step up is the ability to overwrite those learned associations which probably is solved by completing the first task but it’s important to note.
The following step up is being able to process those associations/memories themselves as stimuli which can be associated forming new connections/memories that can be recalled and associated and so on ad infinitum.
That rules out finite non-recursive neural nets, and that still puts us no where near AGI. It’s basically just an ml taxonomic tree builder
The next necessary piece for actual high level thinking is curiosity and action. The system needs to be able to take a concept and a relation (abstract idea of a particular kind of association between two concepts) which currently don’t form a known relationship, and attempt to find the missing concept which would complete that relationship. (Most likely this involves either finding relationship chains between the two which in all other known cases are equivalent to a known relation OR following some procedure to experience the relationship directly as stimuli.)
Lastly, in order to experience the idea of thinking, we first need to have the ability to form temporal associations, but also have working memory of recent past thoughts, and finally the ability to compare those with previous recent memories or the most current stored thoughts. Finding patterns in our thinking helps us be better at thinking. (This is a prerequisite for the experience of free will and consciousness)
That’s what higher level thinking entails.
Of course tuning the system’s response to recognized associations, how memories are stored and recalled, how frequently the system experiences curiosity, etc. will all impact the overall intelligence of the model, but considering most current “AI” don’t even check off the first box, that’s all a bit out of scope.
Spontaneous signals sent in the human brain in response to stimuli. The thing about computers is that their only stimuli is 0s and 1s, and their only response is 0s and 1s. They’re a complex system of switches. Humans experience the entire world of stimuli, and we made this system of switches to convey representations of the meanings we create in our brains. A computer has no mechanism to understand that the purpose of these 0s and 1s we’re sending is to ask and answer questions, to solve a problem of how to build a bridge, to come up with a compelling story to make our friends laugh. We can program it to very accurately perform an action. That’s what it can do.
This sounds a bit circular to me. Almost like saying “thinking is what brains do.”
I’m also getting the sense that you’re partly talking about consciousness there, which I would personally treat as a separate subject. It’s not obvious to me that in order to be able to think, one would also need the capacity to experience.
That’s not wrong to say though, because it’s true. Thought is a function of a brain. There’s no mechanism for non-brains to think. I think you’re confusing computation for thought. Computation can result in a product that looks like it came from thought, but computation itself fundamentally has nothing in common with thought. It’s not a spontaneous, creative response to the stimuli we experience in the world. There is no process of meaning-making. There can’t be, because it has no mechanism to understand what it’s responding to.
Brains are made of matter that obeys the laws of physics. I don’t see why that same function couldn’t be performed in silicon. I wouldn’t say our current systems can think in the sense that people understand the term, but I see no reason to assume they couldn’t in the future - or that thinking is reserved only for wet meat computers.
I don’t really see how you could make a brain out of silicone, because thought a product of cells. The cells send signals to each other, and that’s how you get thought. You could maybe grow a brain in a lab, but I don’t see how a plastic replication would work. And even if you did grow one in a lab, who knows how not being connected to a body would impact its ability to develop thought.
well theoretically with a computer powerful enough, one can simulate all the brain cells and effectively emulate the whole brain
I agree. We could design actual thinking AI, just not with our current technology.
That’s an unserious definition because if thinking is defined as happening in the human brain, then monkeys can’t think. Obviously monkeys can think.
I would say that thinking includes the ability to generate novel ideas and the ability to reason. This ability to reason may include viewing a situation from another perspective than one’s self. An AI does not reason, it recognizes patterns in a huge dataset and generates output based on that.