Do you know about the idea of intelligence explosion? Or do you want to know about this terminology? If yes, then you’re at the right place as here, we are going to get into this term and try to grab all about this one. Let us begin below and dig into the article one.
Pic Credit: Pixabay.com
Why Computers Won’t Make Themselves Smarter?
In the 11th century, Canterbury’s Saint Anselm proposed a unique argument related to the God’s existence and roughly, came up with something like that:
“God is the greatest being that we can ever imagine, a god with no existence is not as great as a God that has an existence and therefore, God must have existence. The entire imagination or statement is called an “Ontological argument”. There are sufficient people who are convinced through this argument as they are still discussing the term number of times even after a period of thousands years. Some critics of this argument affirm that it defines the existence of a God but the statement is not at all indicated like how it has to be worked.
God is the only being here about whom they argued into existence. A mathematician Irving John Good, in the year 1965, wrote that an Ultra intelligent machine can also be defined as the one that has the ability to surpass all sets of intellectual activities of any person, however clever he may be clever.
How did the idea of “Intelligence Explosion” come about?
Since the machine designing is the only intellectual activity but an ultra intelligent machine can be helpful to design a better version, then it would be questionable to be an “intelligence explosion” and thus, the intelligence of a man can also be left behind. Therefore, the first-ever ultra intelligent machine is the best invention ever as made by a man, provided that the machine is the docile one that is sufficient to tell us about how to keep it in control.
The idea “Intelligence Explosion” was derived in the year 1993 by the author cum computer scientist Vernor Vinge, who marked it as “the singularity” and the idea has gained popularity among some prestigious philosophers and technologists. Some books like Superintelligence: Paths, Dangers, Strategies by Nick Bostrom, Life 3.0: Being Human in the age of Artificial Intelligence by Max Tegmark, and Human Compatible: Artificial Intelligence and the Problem of Control by Russell all described some scenarios of self-improvement and Artificial Intelligence program designs improved version.
What’s common in Anselm’s and Good’s argument?
Throughout this, we can find something common in the arguments of Anselm and Good. The common part is all work is done by some initial definitions. Also, these definitions seem too reasonable and because of which all of these are accepted by face value at first but it requires close examination. I think that we scrutinize all suggested assumptions from the arguments of Good and the idea of “Intelligence Explosion” has become less plausible, then.
What recursive self-improvement seems like for human beings?
For convenience, we can explain the intelligence of human beings in terms of I.Q. but not as an endorsement for testing I.Q. as I.Q. indicates an idea that the I.Q. can be easily captured by a single number, the idea being an assumption as made by proponents of intelligence explosion. In such a case, the term “recursive self-improvement” seems like:
“Once there was a person with an I.Q level of 300 and a problem that such a person with an I.Q. of 300 can solve is how to convert the I.Q level of 300 to 350. Further, the person with an I.Q. level of 350 can solve the problem of converting the I.Q. level “350” to 400. And the process goes on.
How does Intelligence work?
Don’t we have even a single reason to say that intelligence works in the same manner? I do not believe that we have even one. To exemplify, there are a number of people with an I.Q. level of 150 and others may have 180. But none of them have the ability to increase an I.Q. level from 60 to 110. If increasing I.Q. level is such an easy task like Mathematics puzzles, then we might see some of the good examples to prove it. But we didn’t see even a single evidence of it.
Is there any Intelligence explosion or recursive self-improvement?
Maybe it happens because we are so far from the actual required threshold. It might be possible that there is a need for at least 300 I.Q. level as a minimum criteria to increase the level of another person. Still, we do not have any good reason left behind to believe in recursive self-improvement.
The Bottom Line
At the end, some proponents of this argument contend that it might be possible to increase the intelligence of a person without even getting a proper understanding of how the system exactly works. They indicate that Intelligent systems such as Artificial Intelligence or the human brain may have one or more Intelligence Knobs and we have to be smart enough to understand where these knobs are.
Currently, we need some right candidates for such knobs and it is so hard to evaluate the idea’s reasonability. The common suggestion to increase the AI is to speed up the hardware on which the program operates. Superhuman Intelligence will be created effectively once we make software that is as intelligent as that of a human being. Therefore, we are so far from creating an AI as equivalent to humans.