#1 - Understanding AI: the Alchemist

#1 - Understanding AI: the Alchemist.

Every week, I will post a sequence of fundamental concepts for those who, even without a technical background, want to understand how artificial intelligence works. Here we go.

Artificial Intelligence isn't code, it isn't programming language, and therefore, not software either. AI is, above all, math.

It's a math where the size of equations and the amount of results to be processed is not limited by human capacity to solve them. It's math with yeast, whey, steroids, vitamins, like an equation capable of producing millions and even billions of x, y, z. And what we call "learning" in AI is nothing more than finding a balanced value where for every x value, we instantly find the value of y. But let's start from the beginning.

If artificial intelligence is purely math, what kind of math are we talking about? All of them, but mainly... algebra. The word "algebra" has its origin in Arabic. The original term in Arabic is "al-jabr", which means "reduction" or "reunion of parts." The word was used by the Persian mathematician Muhammad ibn Musa al-Khwarizmi in the 9th century in his treatise "Kitab al-Jabr wa-l-Muqabala" (Book of Restoration and Reduction). In this work, al-Khwarizmi developed systematic techniques to solve linear and quadratic equations and is considered one of the founders of algebra.

And if you've always thought that doing calculations with letters is something out of this world, you're not alone. During the Middle Ages, algebra was often considered a mystical or obscure discipline. This was mainly due to the extensive use of abstract symbols. At that time, algebra was associated with advanced mathematical practices and was often taught only in restricted academic circles, such as universities and math schools. Mathematical texts and manuscripts were written in Latin or Arabic, which further limited access to algebra for most people. Additionally, algebra involved the use of techniques and abstract symbols, such as x, y, z, to represent unknown quantities or variables, which was a new and bewildering concept for many students and laypeople.

But why use letters instead of numbers? Isn't math about numbers, right? Hmm, not exactly. This is where the foundations of artificial intelligence start to make sense. Algebra was born to generalize a mathematical structure. Let's simplify. Imagine a kingdom where all inhabitants are numbers. A king decides to write a penal code but realizes it will be a lot of work because he has to write about each number and then the law of number combinations.

For example, if the number four steals from number five and gives it to number three, both numbers four and three will be guilty, and five will be the victim. And now? Should I write the same law for infinitely all numbers? Or should I keep four and three, and they will be stigmatized for it? How to write something that is understood to be for everyone?

Well, since in that kingdom there were only numbers, how about replacing the concept with letters to represent those numbers?

In this way, the king found an algebraic equation and wrote: if y steals from z and gives it to w, y and w will be condemned. Now the judges only needed to substitute the letters and use the equation that generalized the problem. According to Wikipedia: A generalization is a form of abstraction whereby common properties of specific instances are formulated as general concepts or claims. Generalizations posit the existence of a domain or set of elements, as well as one or more common characteristics shared by those elements.

It's no wonder that this was seen with mystical eyes in the Middle Ages. Perhaps we are currently viewing AI with the same medieval mysticism, and now we can understand why.

You will understand it better in the next post.

See you.