Disclaimer from the end of this blog
No connection between different aspects of life (such as AI and psychology) is worthy of writing if it doesn’t lead to any practical utility. It has to provide tangible values; otherwise, it would just be a random observation you read. No worries if you decide not to read the entire thing or the title doesn’t excite you. Even if you completely disagree with my opinions, it doesn’t mean that you can’t extract some valuable tips from the content. So, if that’s the case, check out the highlighted words in this blog post as they lead to a book or a video from which you might get some practical tips. If you prefer books in Turkish, reply to this email with “TR,” and I’ll send you the links to get translated versions.
The problem with most self-help books out there
Two weeks ago, I was walking with one of my dear friends. At the end of our walk, I ranted about the self-help books (although I read most of them with great enjoyment). I was talking about how they don’t go deep into the science of their claims in the references. For instance, one even showed a Jordan Peterson lecture referencing a self-help idea. Another talked about a psychological phenomenon that I was very familiar with, yet to find out the author got it wrong. That’s not how referencing or science should work. That being said, if you write a book, you are artistically free to do whatever you want. You can even cite God and interpret the words he whispered to your ear that one night when you take too much you-know-what in Amsterdam. Nevertheless, as a psychology student, I encounter many ideas that rely on psychology in those self-help books, but in a half-baked manner.
On the other hand, as my friend pointed out, this is the point for those kinds of books. They aim not to give you endurance on your way to an Apollo mission to the Moon. You need years of extensive training for that. The main goal of a good self-help book is to get you started. To introduce you to a variety of ideas. Their value lies in their basket of life lessons that make you realize that there is an entire world of ideas out there that could make your career, mental health, and overall life 10 times better. It’s surprising (at the same time very understandable) how we don’t utilize all the books that have lessons from dead and wise people who were gracious enough to provide us with their failures and how to overcome them. To develop an extraordinary engineering career, it’s a no-brainer that you need lots of reading, math, tutorials, failures, and years of practice. Yet, to develop a great character, we choose to trust our intuitions and “street-smartness” and do no external work for it. We settle with beliefs and character traits left to us as inherited from our parents or past experiences.
Let’s return to the point of this blog post. I want to argue for well-written nonfiction books that discuss broad topics and how we can develop our characters like a deep neural network learns.
What is a deep neural network?
First, I want to define what a deep neural network is. In short, the deep neural network is a unique subset of artificial intelligence algorithms. Inspired by the human brain, it has many interconnected artificial neurons that can learn complex patterns in any data. For instance, it can predict a person’s identity from how they walk or make sense of our half-assed and grammatically chaotic questions to ChatGPT. All you need to know is that these networks try to understand the world around them by mimicking human neuron behavior (to an extent). It’s no wonder the Nobel prize winner Geoffrey Hinton, the founder of deep learning, is also a cognitive psychologist. The human brain inspired him and his buddies before creating the foundations of deep learning (that is, changing the world as we speak).
If we dig deeper, deep neural networks utilize a foundational optimization algorithm called gradient descent. Although there are different deep neural networks, I want to focus on the one that’s more intuitive to understand. Let’s say that a deep neural network is learning how to recognize an angry face from a picture. First, it needs to be trained where you show him many people, some with angry faces and some with non-angry faces. You explicitly say the neural network each time you give a face, whether it’s angry or not.
Then, the algorithm starts its learning process by initially taking a random guess. You give it a brand new picture with a person laughing in it, and it says, “This is an angry person.” It’s like babies shouting at every random object, giving them a “randomer” name each time. They both take a guess first. That’s how you start to learn about things by doing. Then, after the random guess, the algorithm calculates its error (by looking at all the previous angry faces it was thought). Just like you correct your baby when he points out the cookies and says, “Coco!” You need to explicitly show him that a cookie is called a “cookie.” (We don’t want him to be a cute little dealer, although it would raise no suspicion). After this initial guess, the neural network takes another guess, but this time, a little bit less random because it has made a huge error before. The algorithm wants to go to a point where the error is as minimal as possible so that whenever there’s a new picture with a face, it can mainly identify the emotion in that face accurately. It wants that minimum point of the valley, as shown in the figure below:
I want you to notice only one thing in the figure, and then I promise I’ll stop talking about AI. Look at the learning steps. Initial steps are enormous at first because the algorithm doesn’t know which way is right or wrong, so why not just say f-ck it and take a big risk? Then, it starts to correct itself and adjust its steps because it notices more specific patterns that can lead it to the right path, like frowned eyebrows or wrinkled forehead. Eventually, it finds a spot where it no longer makes too many mistakes. That spot is called “local minimum”. Here, local means temporary. Why? We don’t know if there are more valleys like this, with the lowest point being lower than the current ones. The algorithm needs to climb another valley (by making huge mistakes) to see whether any other valleys are worth the effort.
Learning steps in character development
Let’s take a step back and think about our life. This process resembles many career climbs, educational paths, college exams, or the development of our character. Initially, none of us knew what we were doing. We take a guess. Then, after each step, we evaluate our current situation and adjust accordingly.
Carl Jung calls this circumambulation. It means revolving around some object of interest. A famous example is the Islamic ritual in Mecca, where people walk around the sacred Kaaba. Although according to Jung, self-development is not linear. It’s a circular path, similar to finding the center of a circular labyrinth. We get lost and believe we have found our way, then fail again. We adjust ourselves by each iteration (I.e., each learning step that the above-mentioned algorithm takes). That adjustment brings us closer to the maze’s center, which holds our true selves. As Jung points out in his autobiography:
“I began to understand that the goal of psychic development is the self. There is no linear evolution; there is only a circumambulation of the self. Uniform development exists, at most, at the beginning; later, everything points toward the center. This insight gave me stability, and gradually, my inner peace returned.”
So, do we actually take those learning steps in our lives? Most of us, including me, can stay at the highest point of the valley for too long. It’s too scary to take that first big step when the weather is foggy, and you don’t know where you are going. It’s hard to try out new things when you are invested in other things that don’t fulfill you the way you want it. It’s hard to face the fear of failure. Fear of shame or mockery. The worst is the fear of facing our faults due to our ego. The “ I know it better” confidence. Not knowing what, we know nothing, as Socrates would contempt. Although living with our ego is an art that needs to be learned, some might (perhaps rightfully) claim that ego is the enemy.
Ultimately, it’s a very good idea to take those first steps wherever they might be applicable in your life. Whether it’s your constant anxiety of things that might go wrong or that constant guilt you are trying to get rid of due to a mistake you made seven years ago. There is always hope, and I am sure there is always time to read (not to mention professional support, although it might be out-of-reach for many reasons). If you don’t know where to start, it’s always a good idea to take a random guess and read a non-fiction book that interests you remotely. With time, you could focus on the things that matter to you. Perhaps it will lead to more fearlessness, stillness, discipline, meaning in life, confidence, fewer regrets, and a noble stand against life if needed. Or when you need it.
Best regards,
Bugra