Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
There isn't any AI. Yet. Because the hardware isn't sufficiently advanced.
That is how humans learn from infancy, and that is how computers will learn. The hardware and basic algorithms that allow trial and error in pursuit of a goal. When success is achieved store the code, then alter the parameters and do it again. Only now it will be from a greater experience base. Eventually the robot will have successful responses to a vast range of situations. New robots that are similar but not precisely the same can be loaded with the learned responses from the start. Same with robots that are put in novel situations or presented with new goals.
There will be lots of coding from humans in the beginning to create the rules and basic system and optimize this process, but even then most of the code will write itself. And once the bots get as smart as humans, there will be no need for human intervention.
That's not how software works.
All software has bugs and inherently flawed. That's because human beings are incapable of producing bug-free software. That's iron-clad, proven by decades of programming - humans can't write bug-free software. It's just human nature to transfer the imperfection to the software, just as we transfer imperfection to everything that we built.
A buggy software is going to transfer the bugs down the chain to whatever software it creates - resulting in a chain of buggy and flawed software. Without human intervention, these software will remain flawed and ultimately unuseable.
Furthermore, humans can always create newer and better platforms that render the old platform obsolete. Even if we create software that writes other software as well as human (a scenario that is impossible to being with), it is impossible for the software to keep up with the newer platforms that humans create.
That's why a software-driven economy absolutely needs humans. The humans are the brains behind all the software, and the later will not exist without the former.
Quote:
Originally Posted by rruff
And once the bots get as smart as humans, there will be no need for human intervention.
You tend to skip over statements that are have no possibility of happening. How can the bots get as smart as humans?
.
There isn't any AI. Yet. Because the hardware isn't sufficiently advanced.
That is how humans learn from infancy, and that is how computers will learn. The hardware and basic algorithms that allow trial and error in pursuit of a goal. When success is achieved store the code, then alter the parameters and do it again. Only now it will be from a greater experience base. Eventually the robot will have successful responses to a vast range of situations. New robots that are similar but not precisely the same can be loaded with the learned responses from the start. Same with robots that are put in novel situations or presented with new goals.
There will be lots of coding from humans in the beginning to create the rules and basic system and optimize this process, but even then most of the code will write itself. And once the bots get as smart as humans, there will be no need for human intervention.
Hardware only determines the speed of computation.
You tend to skip over statements that are have no possibility of happening. How can the bots get as smart as humans?
How did you get "smart"? When you were born you didn't know how to do anything but eat and poop. You learned by interacting with your environment and using your likes and dislikes as feedback. Robots will be self-programmed in a similar fashion. And when one robot gets it, the rest get a perfect copy of its experience.
You keep going on about how humans can't write that code. They don't need to write specific codes, just the rules and goals. Which won't be an "easy task", but it'll be a hell of a lot easier than trying to write specific code and contingencies for every situation that might arise.
And unlike humans who have a bunch of dysfunctional responses, these can be easily indentified and removed in robots. For instance bebOp has a tendency to wet his pants when a pretty girl says hello. That's probably not a response that fits with your goals. But damn, how do you get rid of it?
How did you get "smart"? When you were born you didn't know how to do anything but eat and poop. You learned by interacting with your environment and using your likes and dislikes as feedback. Robots will be self-programmed in a similar fashion. And when one robot gets it, the rest get a perfect copy of its experience.
Software is not humans. Robot is just a piece of equipment. Your vacuum cleaner doesn't get smarter over the years, does it?
You need to think in a deeper level - "Robots will be self-programmed in a similar fashion"......... HOW??? We don't know how to program robots that way, we don't even know where to start, and the technologies we do have are heavily reliant on the human brains to drive it. Like I said, you tend to skim over impossible ideas and then treat it as gospel.
Quote:
Originally Posted by rruff
You keep going on about how humans can't write that code. They don't need to write specific codes, just the rules and goals. Which won't be an "easy task", but it'll be a hell of a lot easier than trying to write specific code and contingencies for every situation that might arise.
And unlike humans who have a bunch of dysfunctional responses, these can be easily indentified and removed in robots. For instance bebOp has a tendency to wet his pants when a pretty girl says hello. That's probably not a response that fits with your goals. But damn, how do you get rid of it?
Dysfunctional responses can be easily identified and removed.... by whom?? Humans? Robots? How'd robots identify these dysfunctional responses. No one has ever figured out how to built a software that can identify bugs without humans pre-programming what a bug looks like. Again, you're imagining an impossible task and treating it as possible - robots can't identify flaws in the software without human help. They simply cannot. A software has no concept of what a bug is. None whatsoever.
Now humans... yes, human can remove bugs from the machine, we do it all the time. That's another reason why software needs humans. A bug (or a dysfunctional response) is a human perception. Furthermore, most bugs are emotional responses.
Your example beb0p wets his pant when a pretty girl says hi..... is an emotional cognition. Let's break it down - why is it bad? Well, because it's embarrassing, awkward, and normal people don't do that! But you only know it is a bad thing because you have emotion and this brings up a lot of negative emotion in you. Without human emotional cognition, beb0p wetting his pants is really no big deal. That's why software can't see bugs - because it does not see the world as we do, it has no emotion, and it simply treat a bug as something that is normal.
A software doesn't understand that when a application repeating crashes, it drives the user crazy. Or that a save action that takes 2 seconds is good but 30 minutes is bad. All of that needs to be pre-programmed into the software by humans. I know you like to say, "Let's pre-programm ALL POSSIBLE bugs identification mechanism into the software!!!"
Well, here's the thing... we can't. No humans can possibly pre-program everything into one software. That's impossible.
Thus, when bug arises, a software needs humans to fix it because it won't know what a bug is in the first place.
.
You need to think in a deeper level - "Robots will be self-programmed in a similar fashion"......... HOW???
Thus, when bug arises, a software needs humans to fix it because it won't know what a bug is in the first place.
It didn't skim over it, I told you. What part did you not understand?
You weren't born with very much initial programming beyond the capacity to learn and grow. Basic desires and the hardware capable of human function. This isn't that esoteric, even conceptually. Computers currently lack the capacity to emulate human-type function, which is why it hasn't been done. We are at mouse level, and the abilities of a mouse are simply not that interesting.
The robot knows if it does or doesn't acheive it's goals. The goals and the ability to evaluate trial and error success and failure are the initial programming. When it achieves success it may try to speed up the task until you are satisfied. Then you alter parameters so it needs alter its response. Repeat.
It didn't skim over it, I told you. What part did you not understand?
You weren't born with very much initial programming beyond the capacity to learn and grow. Basic desires and the hardware capable of human function. This isn't that esoteric, even conceptually. Computers currently lack the capacity to emulate human-type function, which is why it hasn't been done. We are at mouse level, and the abilities of a mouse are simply not that interesting.
You keep referring to humans when trying to explain robots. That's why you are incoherent, because you are comparing apples to oranges and trying to use apples to explain oranges.
You skim over parts that is impossible with a robot. Several posts later, you still have not even touched on your favorite subject.... how robots can take over humans. You keep saying it will happen, but you repeatedly avoided explaining how. When pressed, you would say, "Well humans can do it, robot will too." Sorry, robots are not humans. You can't use humans to explain robots.
If you want to talk about robots, then stick to robots.
Quote:
Originally Posted by rruff
The robot knows if it does or doesn't acheive it's goals. The goals and the ability to evaluate trial and error success and failure are the initial programming. When it achieves success it may try to speed up the task until you are satisfied. Then you alter parameters so it needs alter its response. Repeat.
Bugs are not goals. A software can achieve goals while still being very buggy. Case in point - the SSL bug - it more than achieves its goal but it had a major flaw that was easily hackable.
Again, humans are the masterminds of software. The later cannot function well without the former.
.
You keep referring to humans when trying to explain robots. That's why you are incoherent, because you are comparing apples to oranges and trying to use apples to explain oranges.
I just came into the tail end of this conversation and I found the argument exceedingly coherent, and I'm not a programer or a computer scientist (although I hung out with a bunch in college). The whole idea or artificial intelligence is to build computers that can learn just as a child learns. A child's brain at birth is essentially a piece of hardware with a basic operating system. Over time the brain is fed more data and the child's ability to function increases in complexity over time. What is so hard to understand about the ability to develop both the hardware and software that gathers data and learns how to use that data to ever increasing degrees of functionality.
Quote:
You skim over parts that is impossible with a robot.
Just looking at my iPhone, I have to recognize that this device has a "rudimentary" artificial intelligence component called Siri that depends on learning my individual speech patterns in order to achieve greater functionality. Using the word impossible even in the context of even things like the the grail of the impossible, exceeding the speed of light, should be used judiciously.
I just came into the tail end of this conversation and I found the argument exceedingly coherent, and I'm not a programer or a computer scientist (although I hung out with a bunch in college). The whole idea or artificial intelligence is to build computers that can learn just as a child learns. A child's brain at birth is essentially a piece of hardware with a basic operating system. Over time the brain is fed more data and the child's ability to function increases in complexity over time. What is so hard to understand about the ability to develop both the hardware and software that gathers data and learns how to use that data to ever increasing degrees of functionality.
You're making the same mistake - trying to use humans to explain robots.
"What is so hard to understand about the ability to develop both the hardware and software that gathers data and learns how to use that data to ever increasing degrees of functionality. "
What you described is not hard to understand, and it is also not real AI. Yes, we can built software like that, but we still need humans to built it. Yes, a software can theoretically keep increasing the degree of functionality, but to the point of writing itself and functioning independent of humans?? Not possible. A software is just an extension of the human brains. As long as software is being used, humans are needed. This should not be a difficult concept.
If you want to talk about robots, then stick to robots. If you ever find yourself needing to use humans to justify your idea, you're on the wrong path.
Quote:
Originally Posted by TheWiseWino
Just looking at my iPhone, I have to recognize that this device has a "rudimentary" artificial intelligence component called Siri that depends on learning my individual speech patterns in order to achieve greater functionality. Using the word impossible even in the context of even things like the the grail of the impossible, exceeding the speed of light, should be used judiciously.
It is a far cry learning your speech to learning to think for itself. The word impossible is used because it is absolutely, positively not possible to implement given our knowledge. This is where the term, "Talk is cheap" comes in. It's easy to say, "Hey, let's make computer learns to think for itself!" Just like it's easy to say, "Let's solve world hunger!" "Let's eliminate poverty!"
If it is possible, then why do you guys have such a hard time explaining how without using humans as a fallback?
.
Hopefully, people would come up with something better than simple extrapolation.
A few MONTHS is about the end point for most foreseeable horizons. After that, things quickly begin to deteriorate into mish-mash and poppycock.
To the easily amazed, I'm sure. The same is said by some of horoscopes, you know.
So crop usage prediction is no better then a horoscope? Moores law? Etc etc. The predictions being made by futurologists are based on facts, and science, not just wild guesses. Even better is some of the best people in the field do in fact have high rates of making successful estimates on progress.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
