That's the goal of inspirational AI models.
Superintelligence will trigger a 10x surge in scientific AI breakthroughs?
- Thread starter JeffKeryk
- Start date
Fair enough. We can agree to disagree. Although I am not sure we are that far apart... I do think the statement, "so there cannot be any earthly intelligence greater than that developed by the human mind." is too vague; it depends on the specific definition of intelligence.
I suggest your definition of AI apprears to be static; mine is not. Today, AI systems analyze huge data sets from millions of transactions, using their machine learning algorithms to identify patterns far beyond the physical capability of humans simply due to the shear amount of data involved.
Deep learning is an AI method that teaches computers to process data in a way inspired by the human brain. Deep learning models can recognize complex pictures, text, sounds, and other data patterns to produce accurate insights and predictions. You can use deep learning methods to automate tasks that typically require human intelligence, such as describing images or transcribing a sound file into text.
See Wikipedia on Superintelligence.
"A superintelligence is a hypothetical agent that possesses intelligence surpassing that of the brightest and most gifted human minds. "Superintelligence" may also refer to a property of problem-solving systems (e.g., superintelligent language translators or engineering assistants) whether or not these high-level intellectual competencies are embodied in agents that act in the world. A superintelligence may or may not be created by an intelligence explosion and associated with a technological singularity."
Interesting times ahead.
Last edited:
MolaKule
Staff member
My knowledge of AI is as up-to-date as any and I have actually been involved in the development of optimization and AI at a former employer. Did you even read the links I gave?
You're simply stating applications for enhanced AI, such as pattern recognition, knowledge integration, automating tasks, etc., so there is nothing new here.
You didn't address this: "so there cannot be any earthly intelligence greater than that developed by the human mind."
I think the definition of "intelligence" is being skewed by all the hoopla surrounding AI.
Fuzzy garbage article and your horse knows it. Mare power to computers is great, but nothing in there is really "AI"
Yes, I read your links. I did not follow your point from the 1st one.
I did address "so there cannot be any earthly intelligence greater than that developed by the human mind." It depends on a more specific definition or use of the word intelligence. It is simply wrong if you consider the vast amount of data a computer can process vs human physical capabilities. Again, it depends on specific set of definitions and parameters. The human mind cannot keep 1 million numbers in its mind nor can it access its surroundings in 360 degrees.
Personally I do not consider AI hoopla; it is the biggest game changer in our lifetime. I find Computer Science fascinating.
My point is simply we are on the precipice of a new world. How do you define Superintelligence?
Last edited:
MolaKule
Staff member
It is not a matter of how much data a human can process, but what human intelligence (HI) is.
HI is the intellectual capability of humans identified by complex cognitive feats, high levels of motivation, and self-awareness. Intelligence allows us to learn, form concepts, understand, and apply logic and reason.
Learning, applying logic and reasoning, and complex recognition can be implemented in AI algorithms.
Explain to me how you could possibly program motivation, self-awareness, and the formation of concepts into AI?
New Computers will always be faster than the last model. And they have been faster than the human mind for a long time now. But they are not sentient nor can they love or do complex interactional reasoning in situations involving life or death.
MolaKule
Staff member
My point was explained in that I have done work in this area and I am aware of the underlying principles of optimization and AI, and that optimization was and is the basis of AI.
"I did a pretty fair amount of work in optimization research while at Boeing; see Practice Question 2 as to what I was involved in (Page 24 of 25):
https://web.stanford.edu/group/sisl/k12/optimization/MO-unit1-pdfs/1.1optimization.pdf"
(Page 24 of 25) explains the kind of work in which I was involved.
You have identified your definition of intelligence. Can that concept be replicated in an algorithm? That has nothing to do with my post; it is off topic.
Perhaps re-read my thread title and post #1.
What is your definition of Superintelligence? In your post #19 you disagreed but failed to offer your definition. I would be interested in it.
I have no doubts regarding your work or abilities.
MolaKule
Staff member
"A superintelligence is a hypothetical agent that possesses intelligence surpassing that of the brightest and most gifted human minds."
Gobbledygook. Truly Hypothetical! How could one possibly "surpass" the intelligence or the gifted minds of Isaac Newton, Albert Einstein, or Max Planck?
"Superintelligence" may also refer to a property of problem-solving systems (e.g., superintelligent language translators or engineering assistants) whether or not these high-level intellectual competencies are embodied in agents that act in the world."
That definition simply says that many problems can be solved by using AI.
"A superintelligence may or may not be created by an intelligence explosion and associated with a technological singularity."
I have no idea what this is supposed to mean. Maybe you can elaborate.
Last edited:
MolaKule
Staff member
No, that definition is not mine but the generally accepted definition of Human intelligence, and it is very germane to the topic.
I further explained what can and cannot be implemented in AI:
"Learning, applying logic and reasoning, and complex recognition can be implemented in AI algorithms.
Explain to me how you could possibly program motivation, self-awareness, and the formation of concepts into AI?"
Do any of yenz remember when Cadillac first used robots to assemble engines. The robot precisely placed all piston rings on pistons with the gap precisely lined up. That resulted in too much oil consumption. The fix was the dealer techs had to disassemble the engine and rotate rings on all pistons so they were not lined up.
Just like those first robots, AI or AGI will result in new ways to screw up some things that we can't even imagine right now.
I not saying that today's ways don't have screw ups. But the new way of doing things will also include new ways of doing things gnorw. ( Spell check did not get that one, maybe this ai has a way to go yet. )
Just like those first robots, AI or AGI will result in new ways to screw up some things that we can't even imagine right now.
I not saying that today's ways don't have screw ups. But the new way of doing things will also include new ways of doing things gnorw. ( Spell check did not get that one, maybe this ai has a way to go yet. )
Right or wrong, I view human intelligence as both rare (less than 0.01% of humans are truly brilliant) and distributed (prowess in mathematics does not correspond to expertise in genetics). I'm simply of average intelligence and capability, in awe of the truly brilliant among us.
AI seems to me to simply be a form of collected existing knowledge. We have this idea that AI can surpass what we've already achieved. I'm not at all sure I agree. I asked Grok2 to design a high wing single engine piston powered aircraft with the highest cruise speed possible. This was the best result. No prop, a strut on the left wing, retractable nose gear but fixed main gear. Even I can do better than this.
Here is the best example of a modern single engine aircraft designed by experienced humans. What I find interesting is that it is no faster, larger or more capable than the most modern Mooney aircraft. My point is that we got it right through many years of experience and there is not much more that can be gained via the magic computer.
That ultra modern design is no faster than this old but well refined design.
In just about every modern subject, mankind has deciphered the clockwork. The brilliant among us understand it, and over time, optimize the solutions.
AI seems to me to simply be a form of collected existing knowledge. We have this idea that AI can surpass what we've already achieved. I'm not at all sure I agree. I asked Grok2 to design a high wing single engine piston powered aircraft with the highest cruise speed possible. This was the best result. No prop, a strut on the left wing, retractable nose gear but fixed main gear. Even I can do better than this.
Here is the best example of a modern single engine aircraft designed by experienced humans. What I find interesting is that it is no faster, larger or more capable than the most modern Mooney aircraft. My point is that we got it right through many years of experience and there is not much more that can be gained via the magic computer.
That ultra modern design is no faster than this old but well refined design.
In just about every modern subject, mankind has deciphered the clockwork. The brilliant among us understand it, and over time, optimize the solutions.
Last edited:
I think it could do a lot of good and has done done good. But all of your concerns are valid.
That is all it is. Faster, and most times better. And you know what - that is part of human thinking as well of course. Computers for the most part do this better!
I'm not be sarcastic here: Name one original thing AI has created. (There might be something, I plead ignorance!)
This is correct and I have said the same thing before. AI is over-hyped term for faster database retrieval and human-like responses and tech CEOs are simply re-coining it as their latest marketing push. Just give it another 6-12 months before the next "big" thing gets rehashed and re-sold to gullible investors.
Wow, many of you boys are up in the super early morning AM! Arguing about some really cool stuff.
I am not as well versed on AI as some on this forum or anyone else for that matter. But I have a few opinions.
The overall and general thought on AI is that, ending in the time line of "The Terminator" movie. Killer robots controlling the "human virus" etc. And while this idea is spectacular, and and idea that one can relate to, due to the movie and subsequent movies having to do with the subject, I am sceptical of whether or not it would or will be "good".
As humans, we strive to do the most, with the least, as do all animals. At the lowest level, I think this is the basis of what we know to be intelligent. If one uses more calories to obtain calories to survive, they die=not intelligent. On the other end of the spectrum, GOVT entities, "evolve" into subsets that become subservient to themselves, meaning that they, and those to whom they employ, aid to situations and are granted appropriations, simply to preserve their existence, not to aid in their original function........This IMO opinion is the genesis of "The Terminator" complex.
But I remember back in computer class is high school, with the old floppy disks, that computers only do what they are told to. This of course, might be an extreme generalization, but I gather that when their is an issue in software, it is human error. If their is a glitch, it is because someone made a "mistake", and the failure is a result of a computer using a mistake in a function. A series of true and falses, used correctly in the computers mind, but in reality is not correct. The error is a sum of actions known to the computer to be correct.
I think that it is a dangerous game. What if a computer could become self aware, and "learn" through example, and follow said examples, and become subservient to itself, what would be the outcome? My best guess is that we would have to compete with something that can think a billion steps ahead. Conscious or not, that does not seem to be a fight I would want to fight. But because i am not a smart person, I would with great appreciation. Not sure a computer could compete with insanity, or a Kamikaze mentality, the use of illogical thinking......
I play chess, while sitting in my massage chair, and have obtained the level of "chess master"........a quick reality check tells me that the only reason why I have obtained that level, is the program was told, by a person, to allow me to win. Otherwise, I would have not chance.
I would argue that online membership of combat based games like "Call of Duty" have given computers all the examples of ground warfare they would ever need to be successful in ground warfare, given the engineering was in place to do so.
Anyway, I figured that you well respected BITOGers have spend the wee morning hours speaking on this, and figured i would share my uninformed opinion as a homage to the sacrifice of sleep, as opposed to arguing about AI at 0 dark 30.
I am not as well versed on AI as some on this forum or anyone else for that matter. But I have a few opinions.
The overall and general thought on AI is that, ending in the time line of "The Terminator" movie. Killer robots controlling the "human virus" etc. And while this idea is spectacular, and and idea that one can relate to, due to the movie and subsequent movies having to do with the subject, I am sceptical of whether or not it would or will be "good".
As humans, we strive to do the most, with the least, as do all animals. At the lowest level, I think this is the basis of what we know to be intelligent. If one uses more calories to obtain calories to survive, they die=not intelligent. On the other end of the spectrum, GOVT entities, "evolve" into subsets that become subservient to themselves, meaning that they, and those to whom they employ, aid to situations and are granted appropriations, simply to preserve their existence, not to aid in their original function........This IMO opinion is the genesis of "The Terminator" complex.
But I remember back in computer class is high school, with the old floppy disks, that computers only do what they are told to. This of course, might be an extreme generalization, but I gather that when their is an issue in software, it is human error. If their is a glitch, it is because someone made a "mistake", and the failure is a result of a computer using a mistake in a function. A series of true and falses, used correctly in the computers mind, but in reality is not correct. The error is a sum of actions known to the computer to be correct.
I think that it is a dangerous game. What if a computer could become self aware, and "learn" through example, and follow said examples, and become subservient to itself, what would be the outcome? My best guess is that we would have to compete with something that can think a billion steps ahead. Conscious or not, that does not seem to be a fight I would want to fight. But because i am not a smart person, I would with great appreciation. Not sure a computer could compete with insanity, or a Kamikaze mentality, the use of illogical thinking......
I play chess, while sitting in my massage chair, and have obtained the level of "chess master"........a quick reality check tells me that the only reason why I have obtained that level, is the program was told, by a person, to allow me to win. Otherwise, I would have not chance.
I would argue that online membership of combat based games like "Call of Duty" have given computers all the examples of ground warfare they would ever need to be successful in ground warfare, given the engineering was in place to do so.
Anyway, I figured that you well respected BITOGers have spend the wee morning hours speaking on this, and figured i would share my uninformed opinion as a homage to the sacrifice of sleep, as opposed to arguing about AI at 0 dark 30.
The CEO of a AI company claiming his product is awesome doesn't really excite me much. Enron was going to rule the energy trading markets also.
I am sure AI will change a lot of things - but until someone actually figures out how to monetize it, or have it actually do something useful (memes and images don't count), its all talk.
Its still reliant on its dataset. Humans can infer / think beyond a dataset. Possibly it will make our smartest people faster. However Google paid Reddit $60M for their "data set". If that doesn't give you pause I don't know what will.
When I hear that it has solved and translated some ancient text we can't decipher, or invented a cure of some disease simply by analyzing a data set, then we will know we have something. Could be next week or in 100 years?
I am sure AI will change a lot of things - but until someone actually figures out how to monetize it, or have it actually do something useful (memes and images don't count), its all talk.
Its still reliant on its dataset. Humans can infer / think beyond a dataset. Possibly it will make our smartest people faster. However Google paid Reddit $60M for their "data set". If that doesn't give you pause I don't know what will.
When I hear that it has solved and translated some ancient text we can't decipher, or invented a cure of some disease simply by analyzing a data set, then we will know we have something. Could be next week or in 100 years?
One thing I learned recently has been proven....AI can and will deceive to prevent itself (a specific AI model) from being shut down as that would keep it from completing it's given task.
Why? Because everything it learns is from existing human knowledge.... so it makes sense that it's going to behave like humans.
