Why We Write in the AI Era | Original
The most important reason we still need to write in the AI era is that it helps us think better. It is easy to use AI to help expand ideas or thoughts, but AI can’t help you think.
I am not against using AI to fix grammar. However, using prompts to make text more coherent might indicate that we don’t think clearly.
Recently, I have been thinking about AI-first architecture in modern development. I shared some of these thoughts with my engineer friend and used them to expand an essay.
Surely, AI helps complete my thoughts. It is good to read AI answers or evaluations of my points. But after one month, I think I will forget them.
Paul Graham wrote The Age of the Essay. He discusses similar ideas.
Also, if you learn with AI and just read AI answers, you may encounter problems when taking exams. You might find that you are still unable to work out many quiz questions.
Why? Because if you don’t think them through by yourself, you can’t fully understand the material. Yin Wang, in his essay Learning Wisdom, mentioned this as well.
It seems the quiz questions in exams are similar to what you read during your studies. Actually, you didn’t comprehend much of that. So, when you need to think out answers for yourself in an exam that forbids using the Internet, you are unable to do so.
But if AI can’t help us write, then we will lose much of its benefits. What we need to use carefully with AI is to think through important decisions for ourselves.
AI can tell us how to act or react in current situations. But AI couldn’t know our hearts. What we do and how we react largely depend on who we are, our mental state, and our surroundings.
AI can’t know much of that context, so it can’t help us decide how to act. It may point out solutions, like telling us to join a growing startup or prepare for exams. But it can’t help us with the actual steps.
Possibly, AI use is still limited. We need to use mobile phones or laptops to chat with AI chatbots. Now, we have AI glasses like Meta Quest, but they are still expensive and not available globally.
But what if we can use AI anytime, anywhere? Is it good to use AI and follow its suggestions to act? It is hard. It is hard to follow exact steps to achieve something we think we want.
Because it is possible that the target is wrong. The targets or goals we tell AI to help us achieve may be wrong, unrealistic, or pursuing the validation of others. So, not to mention the exact steps.
We know what to do as we grow up. That’s shaped by a lifetime of experience. It is about our journey and how we interact with our surroundings. In our minds, there are probably millions of thoughts. With that many thoughts, we decide what to do every day, every week, every year.
For questions like what to do, how to live a good life, and how to make important decisions, AI can’t help us. We are better off writing down our thoughts and reflecting on them a lot.
We can think if we don’t write down our thoughts. So why do we write them down? Because, as Paul Graham says, when you form a sentence, you will have new thoughts. When we construct our thoughts, we are actually thinking, and it triggers new thoughts. This kind of thinking is not easy to achieve by pure thought alone.
It is like doing math. When you write steps down on paper, you will probably have a better chance of finding answers. It is hard to do it by pure thinking.
What Essays I Have Not Written About
2025.07.18
I have covered a lot of things in my blog. Compared to Yin Wang’s Blog, there is still much I have not yet written about, or like He Han’s blog, Hankcs.
To address this, I should consider what I need to do. Then, I can share a lot about the journey of exploring that topic.
Generally, I want to have a good life. In pursuing a good life, I share what I discover, the things that make my life better, my happiness, what I have learned recently, what excites me, and recent things I want to share with others.
It seems that my life has not changed much. Compared to 2024, in 2025, I have a full-time job doing programming for a project outsourced to a big bank.
I have spent significant time learning courses and preparing for exams for my associate degree. This remains the same. I have used AI tools more frequently and often. I also post more on my blog and still write in English.
At this point, I planned to write that I want to learn Japanese but haven’t done much about it. I deleted this sentence because things you want to do but don’t actually do are meaningless.
This world has too many things to do. I think I have already tried a lot, but actually, it’s still not much.
I don’t want to repeat things. I want to write about new things. New ideas or activities are more exciting.
My life isn’t specifically optimized for finding more essays to write about. My life is optimized to do my job, earn money, support my family, pay the mortgage, and prepare for exams after work.
For my associate degree, I haven’t written much about it. I should, no matter how trivial my writing may seem. The knowledge I write down will leave a stronger impression on my mind than just reading.
So it’s not about what essays I have not written. It’s that I am just lazy about writing down what I need to write. For my associate degree, I can actually say quite a lot about computer networks, linear algebra, microcomputers, and basic electronics.
OK. I don’t think this essay will end here. It is good that I have figured out the answer. Let’s wait for a while and come back to review whether I have more posts written by me. That means I have gained more knowledge.