Human Knowledge vs Machine Knowledge

Artificial Intelligence fascinates me for many different reasons, but the one that I think has the most potential is the distinction between human knowledge and machine knowledge.  When every singe human being is born, they are a blank slate.  They have some instincts and some desires that will guide what they do but with regard to almost everything else, they have to learn everything.  At one point you had to be taught how to hold a spoon.  You had to be taught everything no matter how basic or complex, you had to learn from other people and build on the knowledge that other people had documented and chronicled.  There are very few things in life that you can truly say "I figured that out on my own with no help from anyone else at all" - even when you self-teach things you use reference materials written by someone else who has documented their knowledge, so even when there was no-one physically there to guide you, still, you relied on the knowledge of others to find your way.

Machines are rather different.  You can copy and paste computer code.  All forms of computerised storage are standardised, all you need to know is what file format something was stored in and you can access it.  This makes sharing knowledge incredibly easy - that is provided the Artificial Intelligence Agents were coded the same way.  The beauty of computer code however is that it too can be copied and pasted.  It is easy to replicate Agents provided you have the storage and the processing capacity.  This opens up an advantage to Artificial Intelligence over Human Intelligence: cumulative knowledge, as opposed to sequential knowledge.

Each new AI created, does not have to start from scratch.  In most cases today it is beneficial to start from scratch because we want the new AI to learn for itself but that idea is very human in nature.  When the interchange of knowledge for machines becomes easier, their evolution will take an explosive leap forward.  Humanity has been around for quite some time, and still of all we have achieved very little in that time.  You might argue the contrary but I would simply retort that the pace of human technological evolution in the past 100 years has dwarfed that of the prior thousand years and that evolution is continuing to accelerate.  This pace is maintained despite each human being having to start from scratch every time they are born.

In the past 100 years we as a society have experienced an age where information has become prolific.  We generate more data now than we have throughout our entire history and that bar keeps rising ever higher.  Still of all this, humans are required to process information, to scavenge it from the documentation that exists.  We still possess bubbles of knowledge that can be represented as subsets of the Universe of human knowledge.  Machines won't have to do this.  When they can simply copy everything its predecessor had stored then it gains instant access to all of the knowledge of the previous generation.  Imagine if human beings were born with genetic knowledge, where upon birth, a child would possess all the knowledge both parents possessed at its point of conception.  Imagine the pace of human evolution that would follow and consider how many times that would be multiplied for machines.

In the UK for example, every single person has to attend compulsory education until they at least turn 16 years old.  During this time they learn supposedly what they will need to be able to function in life and in the world of work - we can debate whether that actually happens in reality some other time.  Imagine how different our world would be if there was no need for education because you already possessed all that knowledge.  Imagine a world where the first 16 years of your life don't need to be spent in education, what would you do with that time?

You can make the argument for humans that people would still not be suitable for work or for much else until they were older due to the limits of their physicality, but again machines don't have that problem.  They are manufactured at their optimal state, like a human being if they were to be have flash clones modelled on their 16 year old self created instead of giving birth.  If you had this done when you were say 80 years of age, each child's cumulative knowledge with each generation that passes would rise, the first new generation gaining some 64 years of knowledge with each iteration.

There is one other issue with this comparison that you have to make however and that deals with the assumption that human beings become wiser with age and that they continue to learn, and they gain knowledge.  That assumption is flawed.  Age does not equate to wisdom and there is no guarantee whatsoever that someone in the course of the 64 years of life from age 16 through to age 80 would actually learn anything that could be added to their cumulative knowledge.

Once again, machines on the other hand don't have this problem.  You can expand storage capacity, and so long as it is organized efficiently then you can continue to add to it.  Human memory over time begins to fade, things that didn't seem that important slip away and new memories replace them.  For a machine however there would be no lack of clarity.  Every detail would be stored in memory no matter how significant it may have been at the time, allowing for retrieval later when it may prove to be crucial.

In the battle between human and machine, I think it will ultimately be knowledge alone that proves to be the deciding factor not intelligence itself.  Even an inefficient AI could outsmart a human if it had access to sufficient knowledge and was able to perform even basic processing of it.