Log #11

A note for the record.

In terms of both my physical and mental health, I have now almost completely made it through what I could probably call the hardest period of my life. I had slowly begun noticing the deterioration in my mental health toward the end of 2024, but for various reasons and because of certain misconceptions, I did not take any precautions. It manifested itself particularly violently during the first few months of this year. The reason I did nothing despite noticing the decline at the time was entirely my belief that the problems I was experiencing were environmental rather than structural or physiological. Once my surroundings changed, everything would be solved. Immediately after finishing my bachelor’s degree, I would start a new life and suddenly find myself in a situation where none of these problems existed anymore. Of course, that was not what happened. In general, I am someone who is prone to depression. I say this all the time, and I know myself quite well. Yes, but what I went through recently was incomparably worse than anything in the past, and in order to get myself out of it I basically started pressing every button I could find. I can safely say that I tried pretty much everything you could think of. Everything from vitamin and iron supplements to exercise. I started therapy but quit not long afterwards. I tried breathing exercises, picked up a ridiculous number of new hobbies, read a ton of books on mental health and psychology, and changed my diet quite significantly.

More importantly than any of that, I caused quite a lot of inconvenience to the people around me. Especially my close friends, and Tommaso, Alaettin, Efe, and Evren hoca, who deserve an honorary mention here—the last three being from Hacettepe. Without the patience of all these people, I honestly think there is a chance I would no longer be continuing university right now, or perhaps I would have done some considerably more stupid things that I would rather not write about here. Around June 2026, I had started getting the impression that I was improving. At this point, there is almost nothing left of that period, and on top of that I am going through an extremely productive phase. My creativity is through the roof, and I have started being able to enjoy simple activities such as films and games again. For example, after a break of several years, Onur and I have started playing Europa Universalis IV together again. I am playing as Russia and he is playing as Austria, and even though we are only in the 1600s, we are already very close to conquering the world :)

Yesterday, we organized a small barbecue as the Geochemistry Department, and as someone who generally takes absolutely no pleasure whatsoever in socializing, I found myself wishing it would never end. For normal people these may be completely ordinary things, but I personally consider them an achievement. No offense to anyone.

As of recently, ChatGPT and Claude have released new AI models. I am genuinely astonished. Are there still idiots out there who think artificial intelligence is a bubble at this point? What we should be looking at and evaluating by now should absolutely not be how well these models perform on various benchmark tests. We passed that threshold years ago. At this point, we simply have to accept that artificial intelligence is more capable than we are at any given subject. This is extremely difficult for many people to accept because, by nature, we are egotistical creatures, and on top of that, most people are not intelligent enough to understand just how intelligent these models are. I want to make a reference to an example Prof. Celal Şengör uses constantly involving Einstein and monkeys. What would happen if Albert Einstein walked into an environment full of monkeys and started explaining that he had discovered the theory of relativity? How would the monkeys react? I see our own ordeal with artificial intelligence as something very similar. There are various arguments being thrown around—for example, that OpenAI is supposedly drowning in debt and is close to collapsing, and so on and so forth. Is the revolution currently taking place an artificial intelligence and automation revolution, or an OpenAI revolution? Whether one particular company rises or collapses has absolutely no bearing on the importance of these models. Like the atomic bomb, we have already crossed an irreversible threshold. Humanity may or may not be ready, but we have reached this point, and what we should be discussing is no longer whether artificial intelligence should be used, but how it should be used. What we are discussing here is, at the end of the day, nothing more than mathematics, probability, and statistics. LLMs had already existed for a very long time. These companies simply commercialized them. That is the only difference. Research in this field actually goes back at least fifty years. So we have to confront the fact that what society is currently denying is, in reality, pure science. The interesting part is that some of the strongest resistance to artificial intelligence is coming from the supposedly “progressive,” supposedly forward-thinking parts of society. The period we are currently living through feels almost like a direct copy of the Industrial Revolution. It is incredibly interesting. I am following all these developments with enormous excitement. Apparently, we happened to be born into one of the less boring periods of history.

Speaking of artificial intelligence, among these newly released models, Astra in particular has left me completely astonished. When I was in high school, I was developing a small game of my own. After spending months messing around with GameMaker Studio 2, the only thing I managed to produce was an aesthetically barren 2D platformer-like game. Today, in less than an hour, Astra can create something that could almost be described as a triple-A-quality game. What this means is that something which previously required multimillion- or multibillion-dollar companies, dozens of different studios, and thousands of engineers, designers, illustrators, various kinds of artists, and writers could potentially be done by one highly motivated and creative person alone within a few weeks. I am following these developments with my jaw on the floor. And that is only one dimension of it. OpenAI is now placing greater strategic emphasis on end-user optimization. I think that is exactly what needs to happen. I find the constant comparison of different AI models with one another increasingly pointless, because even if Model A performs only 1.5% better than Model B, that difference is already far beyond what the human mind can meaningfully perceive. End-user optimizations, however, directly target the applications we actually use in everyday life—for example, Blender or Unreal Engine—and therefore have a much greater impact. On top of that, artificial intelligence is now very clearly beginning to establish itself in scientific research as well. In the past, if you wanted to start a venture, you needed a certain amount of capital and a certain amount of manpower. Today, at least for now, a relatively cheap AI subscription is enough to let you build a simple start-up and bring great ideas to life.