The laggard does what he’s told with the technology du jour

Writing the term “ChatGPT” here hurts me. Are those quotation marks giving me some distance? We are in the realm of wincing rather than suffering; it’s high-school pain, born of a self-concept that started young and didn’t allow for commenting on whatever everyone else was losing their minds over.

It was the news of the day, maybe, or the zeitgeist. I am ready to jump in only when enough years have passed to gather the requisite context. By then, of course, the public’s attention is elsewhere, and I’m glad I kept my mouth shut. When Facebook and Twitter hit the mainstream, I had all sorts of nothing to say about them. I didn’t join, I didn’t cast aspersions; I just rooooooode it out.

With the authority of the armchair psychologist—and because I couldn’t persuade my parents to provide the toys, much less the cool technology, that an eleven-year-old would be into—I’m declaring that the death knell for my material yearning rang outside The Sharper Image store in the underground shopping arcade between Yonge Street and Bay Street subway stations in Toronto. I stood outside the store, looking in, and performed a mental accounting: I didn’t have an allowance, or any savings, and all requests outside Christmas and birthdays were denied, so there were no means of getting my hands on one of those scales that verbalized your weight that day, or a complicated massage chair, or even just a chrome Newton’s cradle like the ones that adorned the offices of lawyers and bankers (clack … clack … clack). The bottomless pit of childish energy that goes into wanting things that you can get with money hit bottom, and then sealed, and I imagine I didn’t ask my family for anything many more times after that. Which was just as well; nothing was on offer.?

Still, I was safe and warm and fed (for a little while longer), and I’d have to figure out what the real value of money was on my own, the slow way, which was fine.

I came to study money formally, in a manner of speaking. For a commerce student in a city where the capital markets had thrived for decades, I wasn’t excited about business, but it was practical. Of the four components of that degree, namely economics, accounting, finance and marketing, the latter left me flat. To complete the assignments, I couldn’t rely on math or cogent writing, the two things I knew how to do. Instead, I was forced to concoct plans, out of nothing, to convince people to buy chattels, of questionable quality and of temporary utility, which I neither cared about nor wished them to spend their money on, and which they may or may not have been able to afford.

It was an untoward mess, lent credence by a demanding university classroom and erudite professors. This many years on, I’m not sure I feel much differently, and still companies send oodles of carbon into the air in order to move single-use plastic from here to there? My question isn’t rhetorical: why are you encouraging consumers to buy trivial objects during a climate crisis?

Ugh. I need this discussion to come back down to Earth. Let’s just say that marketing embarrassed me, I wasn’t good at it, and I can’t enthuse about the ostensible success of moving goods faster and in larger quantities.

Things were moving more slowly in 1982, a somewhat arbitrary year I’m picking because I’d been born by then, medical technology was advancing at speed, we had more time to temper the warming of the planet and the irreversible species loss, and we had the means to make lives better in ways that weren’t restricted to ensuring people bought the right stocks at the right time. We had enough to work with in 1982, the year that the first personal-computer virus was likely distributed, to make a fairer, safer and sustainably farmed world.

And so, ChatGPT, now without quotation marks, reminds me of marketing, of shunting matter around the commercial wheel faster, of believing things are going to be better in the future. Just wait, they say.

And yet, all this blather of mine, all these eye rolls and misgivings about the hollow promises of new technology, are superseded by less lofty fare, namely an email from a friend. Ivan, the chap who wondered why I wasn’t writing about the beer we drank, just asked me to review ChatGPT, specifically its ability to “improve” English text written by authors for whom English is not their first language by boosting clarity, correcting defective grammar and syntax, and adhering to US English (which this article is rendered in). Ivan also expects ChatGPT to maintain the author’s style.

“If the rewrite is okay,” he wrote me, “it will save us a ton of time.” Despite all my wordy theorizing above, once parties that I trust enter the room, I get much less choosy, so I’m not certain what exactly is being recast?by ChatGPT, who the “us” is he’s referring to, nor what these text blocks are going to be about. I can simply ask him, of course, but you get a little fast and loose when your friends are involved, so I’ll just get started.

With the confidence of an early adopter rather than the indifference of the laggard I really am, I describe for you here how ChatGPT manages, in November 2023, as it smooths out a certain type of English. Please, you clever people who wrote about this already (I didn’t read it), and you long-suffering readers who have already taken in your share of similar exercises (again, nada), come and tell me to my face how ill-timed this exercise is; we can laugh together.

Right, so the text is guiding designers in technology and digital spaces to prepare for interviews. If I pre-judge based on 41 words plus four bullets, the text is characterized by a straightforward tone and unexciting, understandable sentences.

As for the bullets, did the software apply them or were they included in the original? Ivan isn’t answering my phone call, and life moves too fast around here to wait for an answer—Move those chattels already! At this early stage, we’ll prefer a parallel structure for the bullets, and so we’ll reserve them for the three instructions—compile, consider, and allocate—and we’ll reformat the explanation about the score so that it comes as body text that follows the bullets.

But wait. The text is talking to us about a score as if we knew there were points. At this stage, you’ve seen just as much of this text as I have, so you can see some preliminary explanation is lacking. Also, following these bullets in the manuscript, there’s no obvious list of questions that we should be considering thoroughly, as instructed. Instead, there are 14 categories such as user experience and pride in one’s work, with each category including some guidance and an accompanying question.

We need to introduce what follows and add some clarity. And let’s get rid of those extraneous horizontal lines. Again, I wonder whether they were inserted by the writer or the software, but I don’t have the original text. Here’s a proposed edit of the bullets:

You’ll find that when writing about consecutive steps, numbering them fits better than bulleting them. Is that something ChatGPT understands? Don’t ask this guy! But tell him if you know the answer. Okay, now to the first three of the 14 subsequent categories that designers looking for jobs have to consider.

Right, because I have an editor’s brain, and you have a brain that likes putting raisins in baked goods, or driving an SUV in the city, or thinking the 45th US president is actually funny, there are aspects to this I don’t want to bore you with, like this: now I’m wondering whether the software capitalized every word of the headings (except for articles, conjunctions and prepositions, it seems), or sometimes it does and sometimes it doesn’t, or whether you can get as granular as that when you’re enlisting AI to format your writing.

In any event, the language seems clear enough at first. It’s geared toward technical types who make apps and other software-based tools I half-understand. Ideally the client (i.e. Ivan) would be using an editor who comes from that world and who knows instinctually whether wording rings true or not, but using your friends is drastically cheaper and Ivan knows I’m tightly wound enough to tell him if something needs a second pair of eyes.

There’s no need for those bullets in the first category. Abridged listicles have their place, but it’s not usually clarity they’re boosting. (From here I’ll redline the edits—in green.)

The second category uses sentences that adopt a nice business diction, but the meaning isn’t clear. Is “reflective learning” design-industry jargon, or will other readers be nonplussed by it like I am? (I’ll ask Ivan.)

I’m not sure that the heading “User-Release Engagement” is clear, either, but let’s move on to the the third category, hmm … and here I don’t know what a “resolution choice” is, but I’ve made edits that reduce room for confusion if I’ve got the meaning right.

Right, so I have 29 pages to review, editing for clarity and correctness but more importantly doing what my client/buddy asked for, which is accounting for how well the software has performed on these metrics: clarity, grammar and syntax, and an adherence to both US English and the author’s voice. At this terribly early stage, it seems like the software is doing a good job of sounding professional and it has a handle on grammar and syntax, which is helpful for a business audience. Still, I needed to jump in at quite a few places, and in this small sample there’s plenty of language that seems all right only if you’re skimming.

Did you read the original text and think it was good enough? You’re not wrong. There will be plenty of forums where text that is error-free and doesn’t signal problems to the audience is sufficient, especially if it’s generated using a large language model. It’s not the type of text that I look to edit because I’m after the really good stuff, which my clients write themselves, unfurling as it does from their progression of thought, which may or may not move at the speed of commerce. Their work is more valuable than fast and cheap, of course. It’s more valuable than money.

But here, back down on Earth, ChatGPT may be doing its job, so I’ll go confer with the client and see whether this effort is warranted, or if, instead, these 1,900 words of mine become a very 2023 artifact of one laggard’s late, quasi adoption of a technology that won’t change the world for the better, if his cynicism holds.