Of Thoughts and Robots
WARNING - THIS HAS ABSOLUTELY NOTHING TO DO WITH MY LIFE!!
then why i'm putting it up here? cause... I CAN!! ^^
Basic theory, which has no reference to any relevant available theory at the moment…
First step to building a Persocom (yeah, you read right…)
I’m writing this cause… I have absolutely nothing else to do, for the moment. Or maybe it’s to put my thoughts into words so that I don’t start contradicting myself in my head… (happens too many times…) Anyway, let’s start with the basic model:
Persocom - An android / gynoid (most probably later) capable of interacting with humans at a human-to-human level while able to utilize machine level capabilities. (or something like that…)
Simple as it seems, it is a daunting task to even come close to building one. In the process of its creation, I have identified 4 levels which have to be achieved before the creation of one is possible.
Level 1 - Finance
Why finance? Cause money runs the world, that’s why. Say what you want, but that is the truth laid bare to you. R&D cost, payments to researchers, equipment cost (COSTLY stuff, if you are going to be toying with experimental materials). Yeah, unless a major company (think Honda… or maybe a joint, since I doubt only one company will be able to finance it ^^) is willing to carry the cost, then we may begin.
Level 2 - Body
To put it simply, only think of the best, even if it isn’t sensible. Artifical muscles, reactive skin, solar-cell hair (yes, I’m not joking.). The reconstruction of a human body, minus some functions (eating, sleeping…). Power to run it? Batteries ought to suffice. Recharging them? Having a dedicated (large) recharging unit which sucks kilowatts isn’t going to be economical. So we can have them solar powered. Still, there is the problem of the solar-cells; they’re shiny and don’t look good if you patch the Persocom with square meters of it. Simply to put, unless development of more efficient, material looking solar-cells becomes a reality, powering the unit will be a huge problem.
Second part is the electronic system that will be the brain of the unit. Processing power, along with memory and storage capacity will become issues during the design and creating process. With quantum computing at the horizons, creation of small, compact super computers may be a reality, so lets put a check on that… for now.
Next is durability and COST. Yeah, making one prototype will cost, but if that cost translates to the cost per unit, I doubt anyone can afford it. Also, as with all machines, with the accelerated return theorem, the life span of the unit must be taken into consideration. Hopefully, manufacturing process can be simplified, such as modular assembly and cheap China labor (hopefully still as cheep). Of course, my original plans were to have them assembled in third world countries and have them imported. A win win situation.
Level 3 - AI
This is where it gets dirty; an AI that must be able to interact with humans effectively, make its own decisions, and generally, keep itself from ending up in the junkyard. Not an easy task. Current AI developments are gearing a single AI to a specific task, like a machine. Not so with the Persocom. My goal is to create a kind of being, separate from humans but respected as one. And unless the Persocom can act like a human, it will be a difficult task.
My plan; 2 AI’s within a single unit. One Ai known as the Sub-AI will control the unit’s functions e.g motor functions, temperature regulations, sensors. This AI will also limit the units functions to within acceptable levels; don’t want the unit to crush a persons hand now do you? Also, it will relay the sensor data is information and not pure machine code. This is vital for the second AI, the Core-AI
The Core-AI will be a Cognitively-inspired AI, which is essentially an AI but without any algorithms and no function. It will, however have information feed to it from its sensors; images, sound, smell, touch, and possibly, pain. By applying basic language recognition software, it will become possible to teach the AI to make sense of the information it is getting, thus in a way creating an AI that doesn’t ‘think’ in machine code, but about the same way humans do. As the Core-AI learns, the Sub-AI will create algorithms that will allow the Core-AI to further understand the information it is getting. This symbiotic link between the two AI’s will be the pride of the Persocom, essentially making the Persocom seem more human-like then a machine.
Of course, there is the problem of memory and the greatest problem, the ‘Three Law’ problem. As with memory, with triniry computing making its debut with the quantum computer, it will be possible to store vast amount of data within a small shell. However, if the need arises, the Sub-AI will began deleting older data, creating smaller, reference copies of them. Thus, in a way, the Persocom will forget, but will be able to recall the moment, even if the sensory information is erased.
The ‘Three Law’ problem is, in other words, the safety of us human when dealing with AI controlled machines. While Issac Asimov three laws (four if you want it…) seem perfect, in reality, it is useless. Thus to properly create a truly Asimovrion machine, it must thus learn two important concepts; death and life. Only when a machine can comprehend the preciousness of life and the inevitable death which follows, can a machine truly be considered ‘safe’. I would put it as ‘alive’ but some people won’t like it.
Level 4 - Marketability
Face it, if you are going to make something like that, you’ll want to make money out of it. To put it simply, HOW MUCH MONEY CAN I MAKE OUT OF SELLING IT?
In the face answer; probably nothing.
First of all, it isn’t sensible nor is it economical to make a life size, human interacting / imitating machine. Even if it is done, the amount put into its development would be astronomical, and I doubt the sales would be able to cover everything. Finally, how much it would cost? Made and delivered within half a year of its completion, about the price of a luxury car, I think.
Now notice I haven’t even talked about what it can do, but then what it can do is not the point here; it’s simply stupid to invest in the development of a Persocom. Besides, if they want nursing robots, it’s only a few years before they are on sale.
The verdict; I just got too much time on my hand.
then why i'm putting it up here? cause... I CAN!! ^^
Basic theory, which has no reference to any relevant available theory at the moment…
First step to building a Persocom (yeah, you read right…)
I’m writing this cause… I have absolutely nothing else to do, for the moment. Or maybe it’s to put my thoughts into words so that I don’t start contradicting myself in my head… (happens too many times…) Anyway, let’s start with the basic model:
Persocom - An android / gynoid (most probably later) capable of interacting with humans at a human-to-human level while able to utilize machine level capabilities. (or something like that…)
Simple as it seems, it is a daunting task to even come close to building one. In the process of its creation, I have identified 4 levels which have to be achieved before the creation of one is possible.
Level 1 - Finance
Why finance? Cause money runs the world, that’s why. Say what you want, but that is the truth laid bare to you. R&D cost, payments to researchers, equipment cost (COSTLY stuff, if you are going to be toying with experimental materials). Yeah, unless a major company (think Honda… or maybe a joint, since I doubt only one company will be able to finance it ^^) is willing to carry the cost, then we may begin.
Level 2 - Body
To put it simply, only think of the best, even if it isn’t sensible. Artifical muscles, reactive skin, solar-cell hair (yes, I’m not joking.). The reconstruction of a human body, minus some functions (eating, sleeping…). Power to run it? Batteries ought to suffice. Recharging them? Having a dedicated (large) recharging unit which sucks kilowatts isn’t going to be economical. So we can have them solar powered. Still, there is the problem of the solar-cells; they’re shiny and don’t look good if you patch the Persocom with square meters of it. Simply to put, unless development of more efficient, material looking solar-cells becomes a reality, powering the unit will be a huge problem.
Second part is the electronic system that will be the brain of the unit. Processing power, along with memory and storage capacity will become issues during the design and creating process. With quantum computing at the horizons, creation of small, compact super computers may be a reality, so lets put a check on that… for now.
Next is durability and COST. Yeah, making one prototype will cost, but if that cost translates to the cost per unit, I doubt anyone can afford it. Also, as with all machines, with the accelerated return theorem, the life span of the unit must be taken into consideration. Hopefully, manufacturing process can be simplified, such as modular assembly and cheap China labor (hopefully still as cheep). Of course, my original plans were to have them assembled in third world countries and have them imported. A win win situation.
Level 3 - AI
This is where it gets dirty; an AI that must be able to interact with humans effectively, make its own decisions, and generally, keep itself from ending up in the junkyard. Not an easy task. Current AI developments are gearing a single AI to a specific task, like a machine. Not so with the Persocom. My goal is to create a kind of being, separate from humans but respected as one. And unless the Persocom can act like a human, it will be a difficult task.
My plan; 2 AI’s within a single unit. One Ai known as the Sub-AI will control the unit’s functions e.g motor functions, temperature regulations, sensors. This AI will also limit the units functions to within acceptable levels; don’t want the unit to crush a persons hand now do you? Also, it will relay the sensor data is information and not pure machine code. This is vital for the second AI, the Core-AI
The Core-AI will be a Cognitively-inspired AI, which is essentially an AI but without any algorithms and no function. It will, however have information feed to it from its sensors; images, sound, smell, touch, and possibly, pain. By applying basic language recognition software, it will become possible to teach the AI to make sense of the information it is getting, thus in a way creating an AI that doesn’t ‘think’ in machine code, but about the same way humans do. As the Core-AI learns, the Sub-AI will create algorithms that will allow the Core-AI to further understand the information it is getting. This symbiotic link between the two AI’s will be the pride of the Persocom, essentially making the Persocom seem more human-like then a machine.
Of course, there is the problem of memory and the greatest problem, the ‘Three Law’ problem. As with memory, with triniry computing making its debut with the quantum computer, it will be possible to store vast amount of data within a small shell. However, if the need arises, the Sub-AI will began deleting older data, creating smaller, reference copies of them. Thus, in a way, the Persocom will forget, but will be able to recall the moment, even if the sensory information is erased.
The ‘Three Law’ problem is, in other words, the safety of us human when dealing with AI controlled machines. While Issac Asimov three laws (four if you want it…) seem perfect, in reality, it is useless. Thus to properly create a truly Asimovrion machine, it must thus learn two important concepts; death and life. Only when a machine can comprehend the preciousness of life and the inevitable death which follows, can a machine truly be considered ‘safe’. I would put it as ‘alive’ but some people won’t like it.
Level 4 - Marketability
Face it, if you are going to make something like that, you’ll want to make money out of it. To put it simply, HOW MUCH MONEY CAN I MAKE OUT OF SELLING IT?
In the face answer; probably nothing.
First of all, it isn’t sensible nor is it economical to make a life size, human interacting / imitating machine. Even if it is done, the amount put into its development would be astronomical, and I doubt the sales would be able to cover everything. Finally, how much it would cost? Made and delivered within half a year of its completion, about the price of a luxury car, I think.
Now notice I haven’t even talked about what it can do, but then what it can do is not the point here; it’s simply stupid to invest in the development of a Persocom. Besides, if they want nursing robots, it’s only a few years before they are on sale.
The verdict; I just got too much time on my hand.