Can machines think?
How should I even start this?
Most people would say that having a thought is abstract and has some divine essence of some sort. But I do not agree with this; I think having any thought is simple mathematics, as with most things that happen in our universe, and scientists have now taught machines how to do it.
This text was written roughly on a Sunday morning and is basically for some of my friends that argued that machines cannot think, this is my response to them. Subsequently as you read on, you’ll also get the general idea of what “to think” actually mean.
Earlier before covid and before publicly available conversational bots, most people I know including myself when given a very difficult problem try to think properly to get a solution, even though it can be a challenging thing to do sometimes, but a few others I’ve seen, know how to think properly more than others, more than I could, like my late physics teacher Mr. Osbourne. But today almost everyone I know has stopped trying to think, with no resolve on how they should learn to think better either, rather they seem pretty comfortable outsourcing this crucial task of thinking to machines, and I do sometimes struggle to keep myself out of that category.
Artificial Thoughts??
Since we're made up of living cells, we could call the way we think the organic or natural way and the computer's way, the inorganic or artificial way. Machine's way of thinking can be seen as an imitation of how we humans think, but this imitation is not necessarily deficient, we could even argue it's outperforms organic thinking in some areas! Recently, there are companies actually making thinking machines that think the natural way, by converting any of human cells to brain cells, connecting them to a computer in a way, and teaching them to perform a task, like playing ping pong or whatever task they want.
So where is the essence in a computer thinking it's way into getting a high score in a pong game or reasoning its way into beating me in a game of chess on a chess website, or is it all just computations and mathematics?
I'm writing briefly my thoughts on this in a way I hope can be understood generally without involving the complications of machine learning computations or of confusing mathematical equations. It will also lay the basis of argument with some of my friends that do not agree with me that artificial intelligent machines can actually think.
Can you think visually?
Suppose I gave you a visual thinking exercise to simply think of a dog kicking a ball (could be football or baseball or whatever) like a man, even though it is impossible for a dog to kick a ball like a man would, you would immediately think of something, because you’ve seen a dog and you’ve seen a man and you’ve seen a ball, you’ve seen a man kicking a ball, so with these cues it won’t be hard for the neurons in your brain to makeup a dog kicking a ball like a man even though such a thing dosen’t exist.
What if I asked you to think of something else, like a Maksgh Gtyahlving a Sakhgt, you would ask me what are those, if those an English words or perhaps even tell me to shut up.
A thought is formed when Experiences (E) maps out to a given abstract variable or outcome, with the neurons as a sort of function doing the processing in your brain, in the first scenerio, a dog kicking a ball, in the second scenerio, nothing, because we have no experience whatsoever with Maksgh, Gtyahlving or Sakhgt, so we can only really think in blanks “something doing something to something else”.
So without any Experience E, it’s would be quite impossible to actually think visually, but from actions we have performed, actions we have seen, read or listened to others perform, we gather these Experiences that makes visual thinking possible.
But you might say, well, sometimes we do have thoughts that aren’t neccesarily visual, like thinking in words, as you would in an internal dialogue. That’s where it gets even more interesting.
Non-Visual Thoughts
Here’s the question I want you to think about right now: When you’re having an internal dialogue, do you talk to yourself? And have you perhaps had conversations with multiple selfs?
What ever is your answer to the question is not important to me, what’s relevant is that you thought about it, perhaps had an internal dialogue with yourself or with multiple selfs, and arrived at your own answer. This is possible because to properly understand the question, and find an answer, your brain is spinning out versions of you with supposed different Experiences, we could say E1 and E2 in the case of an internal dialogue, with E1 been your experience and E2 a supposed experience of your other self, both being a construct of your own Experiences and Learnt Experiences from others maybe through reading a book, watching a podcast or listening to speeches.
How then did we teach computers to think also? Remember I said we needed some Experience E to form the basis for which we’ll have thoughts (generate some abstract variable or outcome) and also reason (propose actionable steps to generate some abstract variable or outcome). One way we do this is teaching a computer by using a neural net (artificial imitation of the brain neurons) trained through reinforcement learning (you can call this learning from mistakes, sounds like what we do right?). I like to call this teaching sands to think, because computer chips are made from sand, and in the scripture we were formed from dust(sand), so we can say teaching computers to think is closer to us creating something in our own image, as is a divine being creating us in his image.
Reasoning as a form of Thinking
If you were in a party of 4 and was asked how many ways was needed to cut a cake to serve all of you, you will easily say “Cut it into four”. Which is the correct answer as each person will get a piece. But did you know how you reasoned all the way there? You probably didn’t pay much attention to this at all because it’s something non-trival, and takes less than a second to process. But someone else might say “cut the cake into eight pieces and let each person take 2 pieces”, this too is also correct, likewise cutting into 16 pieces and each person taking four. So why then did the first one occured to us as correct, even though there are other correct ways to share the cake? The answer is simply reasoning.
Cutting into four takes fewer steps, and each person gets a piece, not complex or complicated but sufficient enough as a solid answer. Anything other than this would be more complex, or more complicated in retrospect and would not be a rational choice.
So we can simply put; to reason is to think logically but in a way that actually makes sense.
Now comes the big question, if Machines can actually think and reason, can they too like us also understand concepts and ideas?
Yes they can, even though their intelligence is artificial.
To test this, go to Gemini or ChatGPT or which ever one is on your phone and ask it to tell you why the sky is blue, then read what it tells you. Now even though you get it keep telling it you don’t understand. You’ll find out that it dosen’t just tells you the answer, it tells you in a way that shows it actually has a deep understanding of why the sky appears blue. This is not simply the case of a computer program telling you an answer to a question, but of a thinking machine explaining to you what you don’t know from information it has learnt, which would not be possible to do if it has no prior understanding what so ever of your question or the information it has learnt.
Hope you enjoyed the read.
Note: There are statements here and there about thinking and machine understanding ideas that might not necessarily be true in some sense, but this forms a solid base for my current model of what that process looks like in general. Also pardon whatever spelling or grammatical errors you might've come across, as this was written once and lazily too.
Sun, Sept 20, 2026 Ebube.