Category: Writing

Categories
Software Work Writing

Reblog: good takes on writing with LLMs

I read these two pieces a few weeks ago and they were still kicking around in my head so I re-found them to share. They are nice complements to my 2023 post about LLMs being good coders and useless writers. They argue that, in fact, LLM writing is often worse than useless.

Link 1: Using LLMs at Oxide. This is the best guide I’ve seen for expectations related to LLM usage at a particular workplace. It acknowledges LLMs as valuable tools while focusing on their ultimate purpose, serving humans. It’s good throughout, but the can’t-miss section is 2.4, LLMs as Writers. Here’s an excerpt:

To those who can recognize an LLM’s reveals (an expanding demographic!), it’s just embarrassing — it’s as if the writer is walking around with their intellectual fly open. But there are deeper problems: LLM-generated writing undermines the authenticity of not just one’s writing but of the thinking behind it as well. If the prose is automatically generated, might the ideas be too? The reader can’t be sure — and increasingly, the hallmarks of LLM generation cause readers to turn off (or worse).

Finally, LLM-generated prose undermines a social contract of sorts: absent LLMs, it is presumed that of the reader and the writer, it is the writer that has undertaken the greater intellectual exertion. (That is, it is more work to write than to read!) For the reader, this is important: should they struggle with an idea, they can reasonably assume that the writer themselves understands it — and it is the least a reader can do to labor to make sense of it.

If, however, prose is LLM-generated, this social contract becomes ripped up: a reader cannot assume that the writer understands their ideas because they might not so much have read the product of the LLM that they tasked to write it. If one is lucky, these are LLM hallucinations: obviously wrong and quickly discarded. If one is unlucky, however, it will be a kind of LLM-induced cognitive dissonance: a puzzle in which pieces don’t fit because there is in fact no puzzle at all. This can leave a reader frustrated: why should they spend more time reading prose than the writer spent writing it?

Link 2: Your Intellectual Fly Is Open, linked in the above quote. It’s a short post. My favorite chunk:

When you use an LLM to author a [LinkedIn] post, you may think you are generating plausible writing, but you aren’t: to anyone who has seen even a modicum of LLM-generated content (a rapidly expanding demographic!), the LLM tells are impossible to ignore. Bluntly, your intellectual fly is open: lots of people notice — but no one is pointing it out. And the problem isn’t merely embarrassment: when you — person whose perspective I want to hear! — are obviously using an LLM to write posts for you, I don’t know what’s real and what is in fact generated fanfic. You definitely don’t sound like you, so…​ is the actual content real? I mean, maybe? But also maybe not. Regardless, I stop reading — and so do lots of others.

I see this from a few people in my professional network. It’s brutal.

“Your intellectual fly is open” is a good comparison to say “we see something embarrassing, we’re just not saying it” but it’s not strong enough in terms of the impact. Once I see someone I know writing through AI without disclosing it, I permanently distrust what they say from then on.

I was prompted to write this post when at a friend’s recommendation I listened to a podcast episode, AI and I: Why Opus 4.5 Just Became the Most Influential AI Model. The episode was okay, and I didn’t like the second episode of that show I tried. But I was struck by something the guest, Paul Ford, said. He spends much of the show discussing how he uses LLMs all day for coding and research. He’s building an AI-based product. But when it comes to writing, he said the bottom-line limitation of using AI is simple: “it’s not me*.”

It’s 2026 and I stand by my 2023 take. I double down on it, in fact: current LLM coding tools are leaps and bounds better than they were in 2023. When I wrote that post, Claude 3 had not yet been released, to say nothing of Claude Code, Github Copilot, Agent mode, etc.

But generating code is writing for machines. And LLMs still aren’t useful for writing to humans.

*I’m quoting that line from memory. I’m not going to re-listen to fact-check myself but please correct me if I got it wrong.

Categories
ruminations Software Work Writing

LLMs are good coders, useless writers

My writer friends say Large Language Models (LLMs) like ChatGPT and Bard are overhyped and useless. Software developer friends say they’re a valuable tool, so much so that some pay out-of-pocket for ChatGPT Plus. They’re both correct: the writing they spew is pointless at best, pernicious at worst. … and coding with them has become an exciting part of my job as a data analyst.

Here I share a few concrete examples where they’ve shined for me at work and ruminate on why they’re good at coding but of limited use in writing. Compared to the general public, computer programmers are much more convinced of the potential of so-called Generative AI models. Perhaps these examples will help explain that difference.

Example 1: Finding a typo in my code

I was getting a generic error message from running this command, something whose Google results were not helpful. My prompt to Bard:

What’s the matter with this Python dict definition?

details = {
‘electrification_lvl’: decoded_vin.get(‘ElectrificationLevel’, None),
‘engine_kw’ = decoded_vin.get(‘EngineKW’, None),
‘fuel_type_primary’ = decoded_vin.get(‘FuelTypePrimary’, None),
‘fuel_type_secondary’ = decoded_vin.get(‘FuelTypeSecondary’, None),
‘battery_kwh’ = decoded_vin.get(‘BatteryKWh’, None)}

Bard told me I had a “significant issue”:

“You used an assignment operator (=) instead of a colon (:) to define the key-value pairs. This will cause a syntax error because assignment statements are not allowed within dictionary definitions.”

Yep! So trivial, but I wasn’t seeing it. It also suggested a styling change and, conveniently, gave me back the fixed code so that I could copy-paste it instead of correcting my typos. Here the LLM was able to work with my unique situation when StackOverflow and web searches were not helping. I like that the LLM can audit my code.

Example 2: Writing a SQL query

Today I started writing a query to check an assumption about my data. I could see that in translating my thoughts directly to code, I was getting long-winded, already on my third CTE (common table expression). There had to be a simpler way. I described my problem to Bard and it delivered.

My prompt:

I have a table in SQL Server called EV_VINs_Presence. It contains columns “vin”, “snapshot_date”, and “locale”. Every record that has a locale value of “Ann Arbor” should have a corresponding record with the same “vin” and “snapshot_date” but with the locale value of “State”.

Write a SQL Server query to return any records that do not have such a corresponding record.

Bard replied:

Continue reading “LLMs are good coders, useless writers”
Categories
ruminations Writing

Finish The Book, Sam

Another installment in what I’m realizing is a series of book-update blog posts.

In November 2021 I wrote half of a novel. In another six months I’d finished the rough draft. Then in November 2022, I resolved to edit it so that someone could read it. Then I said I would get it done by the end of June. I missed that deadline, but finally “finished” this fall! Two finishes so far.

In October I shared it with a few beta readers. There was good news and bad. The good news: it is not terrible! Most of the readers – my friends, granted – finished and enjoyed it.

It is, of course, rife with problems. Most of them I can chalk up to the ignorance of a beginner. For instance:

  • The book opens with my protagonist, Dani, waking up. I now realize that’s a deadly cliche.
  • Dani’s emotional growth arc needs work. I failed to plan that aspect of the novel before writing it.
  • Beta readers universally disliked my handling of a small romance subplot. They are right.

This is, in one sense, the worst news. Were the manuscript irredeemably bad, I would be done. Actually finished. But I believe there’s a chance it could become a good, solid book … with a lot more work. And it would be a shame to waste that chance. Argh. I was so glad to be “done” and finally hand it to people!

It has been satisfying to discuss the world that only lived in my head with others who have now visited it. They had good questions and ideas for making it better. Now I need to steel myself and commit to re-entering that world, not leaving until I’ve shored up its weaknesses as best I can.

If I buckle down, could I make the changes in … two or three months? The rewrite list isn’t enormous. For instance, I’m not changing the perspective from 1st to 3rd person, a task I would not accept.

Discussing this at Workantile, my friend Anthony reassured me that I do finish things. And I have finished things. Now I need to keep finishing. I hope I can find the focus and willpower to finish until this is actually Finished.

I keep coming back to this quote from an old interview with Andre 3000 about releasing imperfect art:

As an artist you can sit and tinker with stuff forever. You can add and take away but I think that’s kind of the importance of having someone over you saying, “We need this, this is a deadline.” Sometimes those oppositions or those who push and pull are needed because we’ll just sit and tinker forever. There are actually songs on The Love Below that were not finished, but that’s how they are, that’s how it came out.
…
I just knew I wanted to put that [imperfect song] on there, but it wasn’t done, but it was enough.

The GQ&A: Andre 3000

The worst would be to sit on this book and not keep improving it. It’s not timeless material and I need to wrap it up and move on and stop having it paralyze me. So I think I need to say, February is the deadline. I hope I work hard on it until then and fall back in love with the story and the process. But even if I don’t, or I only complete some of what I hoped to, it might be time to say, it’s enough.

In the meantime, feel free to ask me about the project and encourage and/or shame me depending on how it’s going!

Categories
Gardening ruminations Writing

This thing is still on

I miss writing this blog. Things have been busy. I draft posts in my head but nothing has gotten onto the virtual page. I’ve meant to blog some recent happenings: a nice win at work, my beloved bike commute that is about to change, getting a heat pump, plants I’m growing. I hope I still will.

I stopped using my last regular social media outlet. Mastodon was a nice improvement on Twitter but it was still sucking up my attention. That leaves me without a place to write and share shorter posts. Maybe I can get comfortable blogging faster and more briefly.

On the plus side, I have been back in the groove of working on my novel manuscript. I am more than halfway through line editing and made a pact with a friend to finish this edit by June 23rd (somewhat arbitrary, but I need a deadline). Perhaps when that’s done I’ll write more here.

Here’s a micro-update: I am enamored with Silphium terebinthinaceum, aka Prairie Dock. Gangly, deep-rooted, whimsical flowers, leaves so ugly they’re pretty. I thought about writing an ode to the plant but someone else already did the job nicely. My two Prairie Docks came back this spring and there’s a new one that might survive to join them. Around Ann Arbor there are some nice specimens in the YMCA’s wildflower garden and along the Stadium Blvd bridge, between the bridge and Graydon Park.

See you soon, I hope!

Categories
Writing

This year it’s NaNoEdMo

A year ago I was working on my outline for National Novel Writing Month, better known as NaNoWriMo. And the following month I completed the challenge, writing 50k words. Progress slowed after that, but I finished the first draft of my book around May of this year. It currently clocks in at 98,354 words, longer than I’d expected.

And many of those words have got to go. No one has read my draft yet because it needs a thorough edit. At this stage the big to-dos are to fill in placeholders (“it’s in [PLACE]”, says one character to another) and slice out crud that makes the book drag.

I haven’t been prioritizing that editing. Which then bums me out because if you write a book and no one reads it does it make a sound? And while I seek to finish the thing, it’s not a chore. I enjoy returning to that fictional world.

The maples have mostly shed their leaves, while the oaks remain mostly clad. It’s NaNoWriMo time once more. I’m thinking I will participate on my own terms. This year’s challenge will be:

  • Work on the book at least a little every day
  • Get the draft to a point where it’s ready to share with an alpha reader on Dec. 1. Fill in all the placeholders and clean up as much of the rest as I can.

I have been developing an outline for another story. This one is a science-based thriller, Jurassic Park vibes but part of a tech billionaire’s sinister plot. All while celebrating one of evolution’s most incredible feats. I think it would be significantly easier to write, because (a) I’ve done it once before (b) it’s a little more basic, with more cliffhangers and less character development.

But that’ll have to wait until I make more progress with Book 1. Maybe I’ll take a crack at the thriller for NaNoWriMo 2023, if all works out.

Categories
Writing

I completed NaNoWriMo 2021 – but my story’s not done

The last time I sat down at the blog it was to declare that I was going to attempt to write a 50,000 word novel in November. Since then I’ve written a lot, just not here. To be precise, I met the NaNo word goal a day early and finished the month with 51,553 words in my story, writing substantially on each of November’s thirty days.

It was a blast! The story has tumbled out. At times I feel like I’m reading it as it materializes in front of me. It will definitely need editing, but I think I was right about having an interesting plot, and my prose has not been as wretched as I feared it might be. I type fast and my natural tendency is to be wordy in both my speech and writing, so NaNo let me play to my strengths and pile up the words.

(There is a metaphor that makes the rounds in NaNo circles along the lines of, writing your book is like building a sandcastle. The first draft is digging up the sand to work with. Don’t worry about the quality yet, just get it out so you can shape it as you revise.)

Lessons learned include:

  • The targets and progress tracking were hugely motivating. This, plus talking with people about what I was doing, was the magic of NaNo.
  • I’d thought dialogue would be hard to write. Turns out it flows much better for me than descriptions of scenery.
  • Beginning with an outline that described 25+ chapters was essential. Once a good idea for a chapter was in place I was comfortable telling its story in detail.
  • Many of these ideas and plot points occurred when I was walking my dog and would tell her the story. Now if I get something juicy, I take care to dictate to my phone so there’s no risk of forgetting it.
  • I had success with an old digital typewriter (an AlphaSmart Neo) I’d had lying around. I wrote everything on there, transferring it to a computer later. The featurelessness of the Neo deterred me from editing, which kept my words flowing, and it entirely blocked me from getting distracted by the internet.

Despite having 50,000 words, I’m not done writing my story. I want to finish it, in part because I want to know how it ends! (I know the general ending, but want to know the details I’ll only think of while writing).

I’m guessing I’m three-quarters done with the story and I fear that if I take a day off, I’ll lose steam. So I’m going to continue writing, setting a target of averaging 1,000 words a day for the first half of December. That would take me to 67k, which might be enough.

I guess if I’m not done at that point, I’ll keep going. During NaNo I averaged 1,700 words per day. Sometimes that was difficult, and I relied on a few vacation days where I racked up several thousand. But averaging 1,000 per day feels sustainable.

Then I’ll take a little break before I come back and re-read what I’ve written. Editing will be a whole ‘nother ordeal. But that’s for later. For now, here’s to my story – it ended up drawing on many of my interests, experiences, and dreams, and it’s a weird little story no one else could have written, for better and for worse.

P.S.: I typically edit blog posts for a while without making them better. One lesson I hope I’ve learned from NaNo is to rein that tendency in. So this post gets merely a quick read-through.

Categories
Someday Writing

Let’s write a novel this November

For the longest time, I wanted to write a book. My “bucket list” evolved over my teenage years and adulthood, but this item stayed constant. Eventually, I removed it, for two reasons. I didn’t feel I had material worth writing about and even if I did, my prose would fall short.

This year, I finally had an idea for a story worth telling. The smallest seed for it was planted a few years ago, as a book someone should write. I kept turning the idea over, growing it slowly. Then I had a breakthrough this summer during a chat with my ten-year-old – we settled on the main character’s quest and her path to victory. The plot is genuinely compelling (in my eyes) and while it’s not my personal story, it’s a mix of settings that I have a bit of familiarity with.

Soon I had my eye on National Novel Writing Month, “NaNoWriMo,” which begins in a month. The timing was great: I could sit with my story in September and October and see whether I lose interest or stick with it and keep plotting the story and characters. So far it’s been the latter. So I’m signed up and planning to give it my best shot! The goal is 50,000 words in 30 days. I’m not sure how long my story might get once I unspool it, but my guess is that 50k words could be enough to tell it all. I have read a few books of ~200 pages this year and it’s a nice snappy length, so that’s my current vision.

I would love to talk NaNoWriMo with others as it approaches and gets underway. Anyone out there want to take the plunge with me this year? Everyone else: if I stick with this you might see less from me as I put my extra energy into the book.

I’m still not sure my writing will be any good – I’m a rank amateur. But I’ll have fun telling the story I’ve dreamt up so even if the result is lousy, I’ll have enjoyed myself and the experience of taking on this challenge. And I can check it off my list.

The biggest challenge I foresee right now is not editing. The idea with NaNoWriMo is to pump out a draft as fast as possible and hit the word count. Then you go back and edit it in future months. When I write here and professionally, I spend more time editing (on the fly and afterward) than I do writing. I’m not sure it always improves my writing, and it slows my blogging down considerably. So perhaps if I can embrace the NaNoWriMo mode of write-without-editing, it’ll lead to more blog posts in the future. Time for me to stop re-reading and publish this post!

Categories
Jokes Writing

Hamilton walks in the woods

Alexander Hamilton finished writing his paragraph, drained his water glass, donned his top hat, and stepped out of the door. He checked his watch – he’d be late to the gala, even if he hurried. He walked quickly, resisting the urge to run. He wished to show up composed and if he went too fast, he’d sweat through his finery. And he was wearing his best summer suit.

Halfway there, all the water he’d gulped while immersed in writing caught up with him. He stepped to the side of the road and gingerly waded through the bushes, only going far enough off the path to attain the privacy required for a gentleman to relieve himself. He carefully picked his way back out to the road and brushed himself off, as to leave no trace of this errand.

Much relieved indeed, Hamilton picked up his pace and arrived at the ball only a little later than was fashionable. He had flushed cheeks and rapid breathing, but was pleased not to have made a scene by being any later.

Thomas Jefferson spotted him: “Mr. Hamilton!” Jefferson walked over to shake his hand. Hamilton apologized for being late, admitting that it was hard for him to take a break from writing. Jefferson looked him over and winked, remarking “oh? It looks to me like you were enjoying a walk through the woods on the way here. Perhaps you had more urgent business to attend to?”

Hamilton was aghast. He’d stepped with such care! He blurted out, “whatever gave me away?”

Jefferson chuckled and pointed to Hamilton’s ankle: “Your undoing? Why, an errant burr, sir!“