Me Myself & AI

It's nice to know a piece of software thinks I can write well

A great big hello to all my readers and subscribers and followers! Thanks for clicking on the latest post from Crone Life, my newsletter about getting to know yourself better as you get older.

We’ve all been reading a lot recently about AI (or LLMs, which I think is the correct term). There was a New Yorker article by Hua Hsu published at the end of June. Hsu used to teach at the same SLAC where I work but now teaches at a similar one nearby. He met with some annoying NYU students and talked with them about how they used AI for college work. One student, Alex, appeared to use it for just about everything. I suppose an LLM is useful for those who don’t like to type or think. Maybe it’s good that something exists for those who have writing or thinking deficits. Although Chat GPT or the equivalent are not intended to be therapeutic, no matter how they’re advertised.

Before I read this article, I didn’t know much about how people used LLMs. I knew that you could give a program a prompt and ask it to create something for you. I knew you could use it for “research” (mostly quick and dirty). I understood that an LLM is not AI in the sense of being intelligent, it’s just a big piece of software that can predict speech patterns based on access to large amounts of digitized material found on the internet. It doesn’t think, it doesn’t write, it doesn’t advise, it echoes. But maybe the echoes can be useful.

Reading about Alex and his use of LLMs, I sensed a deep insecurity underlying everything. He has used various programs to prepare his college application, to write essays for class, to ask if he can go running wearing a specific kind of Nikes. He has switched his major from art to accounting. His friend, whom Hsu also interviews, is a business major. Neither one is an English major nor indeed any type of humanities major. Writing isn’t important to them.

That’s probably OK, right? If they were good at writing, they would be going to school for something that required good writing. Although plenty of people write things and are not very good at it. They write to communicate, perhaps, but not to be good writers. Having spent some frustrating times as a copywriter and a “content creator” for businesses, I am well acquainted with how people who are not good writers think about writing. They think anyone can do it. They don’t know how to judge a piece of writing. If anyone can do it, what’s the point of paying someone? If a machine can do it, there’s no point anyway. Why learn to write at all?

I did learn a lot about how one might use an LLM from this article. I noticed that Alex and some of the other students quoted had employed an LLM to critique their original writing and ask for suggestions to improve it. My own previous use of such programs (usually the free version of ChatGPT) has mainly consisted of asking it to create bullet points and summaries of articles I had already read and summarized while in grad school. I was curious to see how my approach, which admittedly can be idiosyncratic, held up against that of a machine. I’ve also used ChatGPT to make cover letters for jobs based on the written job description but the results were always so terrible that I abandoned the practice. But that was over two years ago. Technology has progressed.

Intrigued by the idea that using an LLM might help me to improve my writing somehow, I took the text of my Crone Life post from June 30, “Self Confidence,” and ran it through ChatGPT. The results were interesting, at least to me. In truth, I didn’t know what to expect. It turns out that ChatGPT likes my writing, responding with sentences like “Your essay is vulnerable, thoughtful, and grounded in personal experience, which makes it compelling. You blend personal narrative, cultural critique, humor, and social commentary in a way that feels natural and relatable.” It then complimented me on the strengths of my writing such as “Honest and Engaging Voice,” “Complex Themes Handled Accessibly,” “Well-Chosen Cultural References,” and “Balance of Humor and Depth.” The “Suggestions for Improvement” section was about as I expected–tighten structure and flow, add transitional phrases (always a challenge, because I think they sound stupid and obvious), “clarify key concepts,” (got me there, probably), “be mindful of repetition”--OK, but sometime I use repetition as a rhetorical device, does ChatGPT know the difference?--and “polish the grammar and sentence clarity”--again, some of the choices it singled out were those I had made deliberately because I wanted the sentence to read a certain way and in a particular voice.

There was also a section called “Optional Enhancements.” Most of these seemed stupid to me. While they might make the tone friendlier or more “evocative,” they didn’t add anything but a (perhaps more) palatable gloss. The “Final Thoughts” conclusion was prefaced with an orange heart emoji (does that mean anything?):

---

🧡 Final Thoughts

This is a deeply personal, culturally aware, and emotionally intelligent piece of writing. With some light editing and stronger structural signposts, it could really shine—whether in a newsletter, magazine, or essay collection.

If you’d like help editing or restructuring the essay section-by-section, I’d be happy to help.

---

I was chuffed that ChatGPT said good things about my writing, but didn’t take it up on its offer, mainly because I thought the "optional enhancements” didn’t enhance. Still, it was nice to know a piece of software thinks I write well. I told my sister about my experiment and she said “Of course you’re a good writer!” I told my son ChatGPT had praised me and he raised both eyebrows in concern and said “You know it’s programmed to flatter you.” Damn, am I not actually culturally aware and emotionally intelligent, then? Is this just one of those machine dream hallucinations? I mean, I think I am both of those things, but I could be badly misled. Who can I trust? I don’t think I should take my twenty-something son as an arbiter of “good writing,” necessarily, since his cultural intake consists mainly of manga, computer games, and things he reads on Reddit or Discord. Also, memes. But why am I feeling defensive about it?

We’ve already established that despite my advanced age, my years of experience and my objectively good education, my self-esteem is not the highest. I don’t need ChatGPT to believe in my ability to write, I need to believe in it myself. But I can understand how, if I were college-student age, I might seek this type of reassurance. I might ask the software to edit my work to make it better. I might run everything I wrote through it, if only for the comfort of being told I did at least some things right.

When is it appropriate to make use of this technology? I don’t know. I recently applied for two positions and while I got good feedback, didn’t make it to the interview round for either. If I had asked ChatGPT to “enhance” my cover letters, would I have done better? Would “tightening the structure and flow, polishing the grammar and sentence quality” have made enough of a difference? Using ChatGPT or another LLM for this is a cheap method of career coaching, I suppose. I’ve never wanted to hire someone for that because I find it difficult to trust they could help me the way I want to be helped. I want to be more clearly myself, not just more digestible for the mainstream. Probably, though, using ChatGPT to improve my job-search attempts is a good idea. Cover letters are just forms, after all, and you might as well write something that conforms to expectations. I don’t need to stand out in this realm as much as I need to fit in.

There are no conclusions here, either way. There are only individual decisions. Is using ChatGPT for writing “cheating?” It depends how you use it, clearly. That you will need to work out for yourself.

The absence of Veronica

I’ve barely seen her this week. I went back to work, so I’m not here during the day. When I come home, I collapse into bed once I’ve eaten something. My husband says she has curled up on my pillow a few times, but I slept through it. It’s hot, so she prefers the basement, unless the A/C is on. We are still trying to use as little of that as possible, for various reasons. I am in charge of deciding if I can endure without it. Mostly, it’s the humidity that’s going to to kill me.

Thanks for reading! I am so tired and stressed out, still. My immediate supervisor retired this past week and starting on Monday I will be on my own until they hire somebody. Naturally, this make me uneasy. I don’t know what to expect. My future seems suddenly unsure. And it’s hot.

Please feel free to comment, forward, restack, share, or what you will. Click the heart to reassure me that I really can write as well as ChatGPT says I can. Stay cool!