Getting to the End of the Thought; or, Why Write in the Age of AI?

• June 6, 2025
• essays  
• education  
• learning  
• teaching  
• AI  
• artificial intelligence  
• writing  

I’ve been thinking a lot about the value of writing recently, especially in regard to how it’s been influenced by the advent of generative AI. In the past few months, I’ve had a lot of conversations in which people claimed that generative AI is just as good at writing as humans now, or better, in many cases. While I agree that AI generation tools are, at this point, better at the mechanics of writing than the average person¹, I have always found their outputs to be shallow and devoid of interesting surprises. There is rarely any depth to the ideas in the generated text, and whatever personality is apparent feels like a disastrous photocopy of a sketch of something real. I don’t think these are problems that will go away, because they fundamentally stem from how the texts were created. You can’t approximate a real human’s perspective by using a machine that consumes content in aggregate. AI enthusiasts tend to talk a lot about outputs and content and products, but if we want to get to the bottom of why I’m disturbed by the proliferation of texts generated by AI, we need to start talking about process, and start thinking of it as something that matters².

I’m extremely concerned by the fact that kids in elementary and high schools are using generative AI to do their writing assignments. I’m horrified by the quickly approaching reality in which the average person might have no idea how to come up with their own ideas and organize them coherently in textual form. I am aware that as a writer, I’m biased and likely to attach disproportionate value to the skills required for my discipline, so this is my obligatory disclaimer. However, I think there is something universally important about being able to express your thoughts using words. I’m a lot more coherent in writing than I am verbally, and I suspect this is true for most people. Writing things down helps me slow down enough to actually think.

I find thinking linearly and precisely to be difficult without external aid. When I’m trying to deal with new ideas or concepts, the thoughts floating around in my brain tend to be loose and unspecific. A thought might be an inkling of a starting point for exploring something, or a general idea of a structure for exploring something, but the specific details of how to do so are usually blurry at best. Even when some of the details are clear in the moment the thought hits me, they’re fleeting, and I’m only left with a vague impression afterwards. Sometimes I try to write down my thoughts as they happen, but come on, I have a life. It’s really hard for me to drop everything and start writing whenever I have an urge to say something. I have other stuff to do.

Conversations, I find, can help push ideas a step beyond free-form thinking. They’re still fluid and unstructured, which makes them good for exploring and generating new ideas, but they force you to explain enough of your thought that the person you’re talking to can both understand and build upon what you’re saying. If my unaided brain often spends quite a bit of time floating just below the surface, maybe getting to shallow depths on a good day, a conversation can get my understanding of something to maybe medium-depth level on a really good day. Sometimes talking to someone can help you refine and go deeper with an idea, but because conversations never unfold the way you expect them to, I find that they generally tend to suggest new directions or adjacent things to consider or help unveil another perspective. It’s easier to talk about details as they come to you than to hold them in your brain, and hopefully the person you’re talking to will have details of their own to share that will have an interesting synergy with your ideas. You still might not remember all of the details, but you’ll remember more of them, and you’ll remember some of the alternative directions for sure.

While writing can also function in the exploratory modes I just described above, as far as I’m concerned, what the written mode uniquely excels at is focused investigation of a single idea, or what I’m going to call “getting to the end of the thought”. It can be useful to wander around the edges of an idea or opinion to see its contours and different ways to approach it, but that’s not always enough. When clarity and precision are at stake, when I want to figure out exactly what it is that I think and why it is that I think it, it is imperative for me to write things down. Getting to the end of a thought involves following an idea as far as it’ll take me and seeing what emerges, and I need somewhere to put all of the thoughts I’ll gradually dig out of my brain on the way there. Only then can I sift through the materials and figure out what the core of my actual argument is; only then can I construct a conclusion. This takes time, focus, and trial and error, and I find it’s almost impossible to do without writing. At the end of the process, it feels like I’ve unknotted some pathways in my brain, like I’ve untangled some deep knot of confusion and come out of the other side with understanding.

This, more than any other reason, is why I bother to write anything at all. A lot of corporate communications feel pointless, and a lot of the academic writing done in courses feels stilted, forced, and transactional to me, more about proving that one can go through the motions of explaining and arguing something than about actually presenting any ideas of value. I think lots of people have gotten the idea that writing texts is always like this, that it’s always some kind of bullshit generation exercise where you make a product no one will read or care about. It can be, and if your objective is to create bland corporate texts, I can understand why you may want to delegate your tasks to a machine. But I think it’s dangerous to conflate learning to write with learning to generate bullshit content. The skills you gain by learning to communicate cannot be replaced by a machine.

The problem, of course, is that a lot of the basic training in how to use words to communicate ideas is acquired in schools, which are notoriously terrible at assigning meaningful work. The school system places undue importance on writing as a product, which is why kids learning to write see no functional difference between them writing a bullshit essay and a large language model generating one for them. It’s hard to find value in the process of doing schoolwork when it can functionally be treated as a box-checking exercise. But despite the general blandness of primary school writing assignments, they are a big part of how I first cut my teeth as a writer. They are a big part of how I learned to trust my own ideas and figured out how to make others understand them. What I’m really worried about is that kids relying on AI to do all of their writing for them won’t develop the ability to “get to the end of the thought”, which is precisely the thing I think writing is best for.

Generative AI is essentially glorified autocomplete. No matter how sophisticated an autocomplete engine gets, the text it generates, even if it’s based on ideas or structures you provide, is unlikely to do anything to help you understand what is going on in the crevices of your brain. Through writing (or more likely, through failing to write), you untangle the different threads that are causing the knot in your brain and have the opportunity to examine what they’re made of. That’s why we think of writing as a process; it’s as much about what happens in the brain as it is about what gets on the page. I think it’s really dangerous to view creating products as the goal of having a writing practice. The finished product, in a lot of cases, is simply a representation of the complex process that happened in the brain of the author. For the reader, it offers the chance to start integrating the author’s ideas into their brain, which might kick off a similar process where they create and untangle new knots in their own brain.

Learning to write well isn’t just learning to communicate – it’s also learning to think more clearly. Being a writer has made me more in touch with my thought process; I feel like the constant practice of having to distill my ideas in such a way that they’re understandable to others has made them clearer to understand for me as well. I’m more in touch with where my ideas come from and am better at watching them form in my brain. I think learning to write has also made me a more coherent speaker, able to speak off the cuff for longer in a somewhat linear way. I guess what I’m saying here is that there is so much more I get out of engaging in the writing process than simply having a completed written piece at the end.

There’s an old quote, traditionally attributed to Dorothy Parker, that says, “I hate to write, but I love having written”. A lot of writers, myself included, have historically liked to identify with this statement. Writing is hard. Like any creation process, it’s often painful, grueling, and uncertain. Working on a personal writing project can feel a lot less like a fun hobby than like an act of self-torture. But in recent times, my thoughts on this quote have been changing. Sometimes pithy statements feel quite different when their reality is staring you in the face. Generative AI quite literally gives us the tools to completely skip the writing process and get straight to “having written”, and it terrifies me, because it’s an illusion based on lies. You haven’t really written anything if your neural pathways haven’t been altered in the process. I may hate the writing process occasionally, but not enough to put my name next to the words of a probabilistic prediction machine.

I’m concerned that my generation is turning into an army of AI-augmented bullshit machines, and even more concerned for the next generation, which hasn’t lived in a world where relying on AI isn’t an option. I’m concerned, because we don’t have the skills to deal with this yet. I’m worried that we might collectively lose our ability to think and analyze and create at the time when people are most clamouring for something real.