Strunk and White on Steroids

Strunk and White on Steroids

Hypocrite lecteur,—mon semblable, —mon frère!

Baudelaire, “Les Fleurs du mal” 

Last night I did something mildly perverse: I fed a commercial Facebook post into my ChatGPT—the same one I use for research, copy editing, and tech triage—and asked it if the copy had been generated by AI.

BTW, I refer to my ChatGPT as “Chad,” not because I think it’s human, but because it’s easier to say (and type) than ChatGPT.

Anyway, Chad conjectured that the Facebook post had been produced by AI and then provided these telltale signs of AI authorship (as if Chad were human, and didn’t engage in the same techniques itself).

            The first indicator is that AI produces balanced sentences. [1]

Three-part structures. Nicely paced. Very composed. Humans do this too, of course—but we’re sloppier about it. We interrupt ourselves. We go long when we shouldn’t, or bail out early.[2]

e.g., “Stay alert, stay ready, and don’t talk your way out of the moment.”

The second sign of AI-generated prose Chad calls “hinge sentence moves”—phrases like

“Because it does happen” 

“And here’s the part that matters”

Those little turns feel less like thoughts arriving and more like pre-installed signposts.

The third characteristic of AI writing Chad cites is its tendency to engage in “emotional generalities instead of specifics.”

            “move your life in a direction that feels right”

That’s technically fine, but it floats. A human writer—especially you— tends to ground emotion in something tactile or slightly odd.[3]

Phrases like “move your life forward in a direction that feels right.”

No one ever missed a bus in these sentences. No one spills anything. No one says the wrong thing.

            Number four is “controlled charm.”

“Your dog, who is clearly the decision-maker.”

That line has been focus-grouped by the internet into harmlessness. It lands, but it doesn’t leave a mark.

And, finally, number five: evenness of tone.

No spikes. No awkward sentence that makes you pause and reread. No moment where the writer sounds just a little off—and therefore real.

Humans leave fingerprints. AI wipes the glass.

And here’s the twist: The more people read AI copy, the more they’ll start to distrust fluency.

So, there you have it. ladies and germs, the AI Style Sheet: Strunk and White on steroids. 

Stay tuned. My next project is plugging some Faulkner into AI to see how it would rephrase the opening paragraph of Absalom, Absalom.


[1] I’ve italicized Chad’s direct quotes.

[2] Note the self-referential pronoun “we.”

[3] Chad doesn’t mention this, but AI-prose loves them em-dashes. I’m assuming the “you” in the sentence refers to Wesley Moore III.

Elon Musk as Bond Villain

Oscar Wilde’s counterintuitive observation that “life imitates art” certainly seems a valid observation when you consider that Elon Musk obviously modeled his persona on the Bond Villains he identified with as a child watching 007 movies in his South African enclave.

I asked ChatGPT to create a Venn diagram comparing characteristics of Dr. No, Auric Goldfinger, Ernst Blofeld, and Elon Musk, but rather than creating the visual representation I asked for he/she/it/they came up with this:

“Creating a Venn diagram comparing characteristics of Dr. No, Goldfinger, Blofeld, and Elon Musk could be interesting! Here’s how we might categorize some of their traits”:

Dr. No

  • Monologuing: Dr. No is known for his villainous speeches and arrogance.
  • Desire for Global Power: His main goal is to dominate the world.
  • Tech Savvy: Uses technology for his villainous deeds, such as a secret base and advanced machinery.

Goldfinger

  • Love of Wealth: Obsessed with gold and wealth accumulation.
  • Manipulative: Skilled in manipulating people and situations for personal gain.
  • Planning for Destruction: Plans to increase his wealth by threatening global economics.

Blofeld

  • Mastermind: A schemer and manipulator who plans complex, worldwide operations.
  • Philanthropic Facade: Often hides his true intentions behind a charitable or business façade.
  • Desire for Control: Wants to control global events to advance his personal agenda.

Elon Musk (Common Traits)

  • Innovative: Known for pushing technological advancements (SpaceX, Tesla).
  • Wealth: A billionaire who accumulates wealth through various ventures.
  • Ambitious: Has ambitious goals such as colonizing Mars and transforming energy production.
  • Public Persona: Often in the media spotlight and makes bold statements.[2]

Note how ChatGPT avoided portraying any of Musk’s traits as negative, much less as villainous.

Allow me to rectify that failure:

Elon Musk (Common Traits)

  • Monologuing: an asshole known for his villainous speeches and arrogance.
  • Desire for Global Power: a megalomaniac who desires to dominate the world.
  • Tech Savvy: an innovator who builds rocket ships, electric cars, etc.
  • Wealth: a billionaire obsessed with wealth accumulation
  • Mastermind: a schemer and manipulator who plans complex, worldwide operations.
  • Desire for Control: a self-obsessed piece of shit who wants to control global events to advance his personal agenda.

Oh, yeah, what about the autism, the lack of social grace and human empathy, not to mention operating outside the bounds of law?  The stilted speech patterns all four share?

On this assignment, I’d give ChatGPT a big fat D if he/she/it/they were my student and myself a C- on this blog post, which has forsaken hard work for expediency. 

That said, it’s kind of scary that a Bond villain now has access to my social security number.


[1] The name Elon Musk sounds like Ian Fleming himself came up with it.

[2] Not how ChatGPT doesn’t bother with parallel construction. The catalogue he/she/it/they came up with is about as elegant as a Tesla truck.