Things AI can do

[Homer Potvin]

One thing that’s less clear at present is, once working people – comprising 99% of the world’s population – are further impoverished by a technology that is plainly being deployed as a weapon against them, how will they afford to buy all the crap that AI will be producing in their stead?

With AI security robots, they will not need to pay any of the halves, only the fabrication costs.

Let's limit the discussion to the writing implications of AI, please. LLMs have nothing to do with the Terminator.

 

[Naomasa298]

All I will say is, just try it. Give it a short piece to critique and see what it says, with an open mind. Compare it with what a human might give you. Tell it not to suggest rewritten text. Then, by all means, discard it, if you don't find what it gives useful, or you disagree - and disagreeing, doesn't make it useless. You might disagree with what a human critiquer or beta reader says too.

 

[Mogador]

I will embrace AI, maybe even joint the fight to recognise personhood, as soon as it can show delight at a well-chosen phrase or surprise at a revelation or absorption into the world of an author. Until then, I'll do my best to avoid it.

In the mean time can we all agree to formally deny the personhood of the business acquaintance who explained to me in our last meeting how he was able to fire his communications team because now an AI will write his heart-wrenching yet engagement-boosting LinkedIn posts about his latest charity cycle ride?

Edit: Last deviation from me, per previous two messages from OP and the modfather. Will now stay on topic or off thread.

 

[Stuart Dren]

All I will say is, just try it. Give it a short piece to critique and see what it says, with an open mind. Compare it with what a human might give you. Tell it not to suggest rewritten text. Then, by all means, discard it, if you don't find what it gives useful, or you disagree - and disagreeing, doesn't make it useless. You might disagree with what a human critiquer or beta reader says too.

I'm game. I gave it the first three sentences of my WIP.

I'm putting this in a spoiler tag because there's lots of text:

Spoiler: Chat GPT Conversation

​

can you evaluate the opening of my novel?

ChatGPT said:​

I’d love to.

Please paste the opening (as much as you’d like feedback on), and let me know:

• What genre it is
• Who your target audience is
• What kind of feedback you want (big-picture story, prose style, pacing, hook strength, marketability, line edits, etc.)

Once I have it, I’ll give you a detailed evaluation—what’s working, what’s not yet landing, and specific suggestions for strengthening it.

You said:​

It fell into each carbon model through an undiscovered gradient, something we suspected but never measured. If only we had taken the time to first solve the unknown variable, inevitable wildcard that made us, the world might not have ended.

—Momma

A shiftwell could be predicted on short notice, so long as one knew to watch the muffin trees, or rather, how to watch them.

ChatGPT said:​

This is a strong, intriguing start. It signals genre, voice, and mystery very quickly. I’ll break this down into hook, clarity, voice, worldbuilding, and line-level suggestions, then offer a tightened example revision.

---

1. The Hook​

Your opening line:

It fell into each carbon model through an undiscovered gradient, something we suspected but never measured.

This immediately signals science-driven speculative fiction. Terms like “carbon model,” “undiscovered gradient,” and “unknown variable” create a tone that feels analytical, maybe post-apocalyptic, possibly hard sci-fi.

The second line sharpens the emotional and thematic stakes:

If only we had taken the time to first solve the unknown variable, inevitable wildcard that made us, the world might not have ended.

This is the strongest hook element.
You’ve given us:

• Regret
• A mysterious “unknown variable”
• The suggestion of world-ending consequences
• A personal voice (“we,” and signed “—Momma”)

That last element—“—Momma”—is doing a lot of work. It implies:

• A generational perspective
• Intimacy
• Possibly a diary, recorded transmission, or found document

That’s compelling.

---

2. Clarity vs. Intrigue​

Right now, you’re walking a fine line between intriguing ambiguity and cognitive overload.

Phrases like:

• “carbon model”
• “undiscovered gradient”
• “unknown variable”
• “inevitable wildcard that made us”

are all abstract and conceptual. That’s not inherently bad, but stacking too many abstract scientific phrases in one breath risks distancing the reader before they’re grounded.

You’re asking the reader to process:

• A scientific anomaly
• A modeling failure
• A philosophical statement about creation
• The end of the world

All in two sentences.

That’s ambitious. It almost works—but it’s dense.

---

3. The “Momma” Signature​

The em-dash attribution:

—Momma

This is excellent if:

• The story continues in the POV of her child
• Or this is a quoted epigraph
• Or her voice matters structurally

If it’s just a stylistic flourish and we don’t return to her soon, it may feel ornamental rather than purposeful.

Right now, it reads like a diary excerpt or an epigraph. That’s good—but clarity of intent matters. Even a line break that visually sets it apart more could help.

---

4. The Transition to “Shiftwell”​

Then we move to:

A shiftwell could be predicted on short notice, so long as one knew to watch the muffin trees, or rather, how to watch them.

This is where the piece becomes really interesting.

“Shiftwell” and “muffin trees” are specific, concrete, and strange. That’s good. They anchor us in a world.

However:

• “Shiftwell” sounds technical.
• “Muffin trees” sounds whimsical.
• The tonal collision is intriguing but abrupt.

Right now, the reader has:

1. Abstract scientific regret
2. The end of the world
3. A mother figure
4. A mysterious phenomenon
5. Muffin trees

That’s a lot in ~70 words.

---

5. What’s Working Very Well​

• Strong speculative tone
• Immediate stakes (world ending)
• Thematic depth (unknown variable “that made us”)
• Curiosity hooks (“shiftwell,” “muffin trees”)
• Voice with philosophical weight

You clearly know your world.

---

6. What Might Be Strengthened​

A. Reduce Abstraction Density​

Consider making just one element more concrete.

For example:
Instead of:

It fell into each carbon model through an undiscovered gradient

You might hint at what “it” is emotionally, even if not scientifically.

Mystery works best when it’s specific-but-incomplete, not abstract-and-layered.

---

B. Tighten the Second Sentence​

This part:

inevitable wildcard that made us, the world might not have ended.

The grammar makes the reader work hard.

You might clarify the relationship between:

• The wildcard
• Humanity
• The world ending

Right now it’s slightly syntactically tangled.

---

C. The Shiftwell Line​

This line is promising, but slightly long and indirect.

so long as one knew to watch the muffin trees, or rather, how to watch them.

“Or rather, how to watch them” is conceptually interesting but weak rhythmically.

You might make that sharper or more surprising.

---

7. Example Revision (Not Prescriptive, Just Illustrative)​

Here’s a version that keeps your intent but reduces abstraction density and sharpens rhythm:

It slipped into every carbon model along a gradient we suspected but never measured. If we had solved that variable first—the wildcard that made us—the world might have survived.
—Momma

A shiftwell could be predicted, if you knew how to read the muffin trees.

Notice what changed:

• “Fell” → “slipped” (quieter, more insidious)
• Slight simplification of structure
• Cleaner rhythm
• More direct “read the muffin trees” instead of repetition

Again—this is just one possible direction.

---

8. Overall Assessment​

This feels like:

• Literary sci-fi
• Possibly post-apocalyptic
• Philosophical, voice-driven
• Conceptually ambitious

Your biggest risk isn’t lack of intrigue—it’s overloading the reader too quickly with abstraction before emotional grounding.

But you absolutely have something compelling here.

This made me laugh:

Right now, the reader has:

1. Abstract scientific regret
2. The end of the world
3. A mother figure
4. A mysterious phenomenon
5. Muffin trees

That’s a lot in ~70 words.

To give it credit, it gave me similar advice I received from a human on another thread:

You might hint at what “it” is emotionally, even if not scientifically.

Mystery works best when it’s specific-but-incomplete, not abstract-and-layered.

And of course its actual line revisions were shit, but no one's disputing that.

I didn't feel that it told me anything especially useful, and the nonsense is there beneath surface level:

Right now, it reads like a diary excerpt or an epigraph. That’s good—but clarity of intent matters. Even a line break that visually sets it apart more could help.

What does that mean? Clarity of intent matters? There was a break on my end, but not when I copy-pasted it to the forum, so whatever.

 

[Naomasa298]

It's a bit short for it to really tell all that much about the story. As you noted - it didn't really do worse than your human reader.

Clarity of intent, I would read that it's clear what the story is meant to be. Kinda hard for it to tell from three sentences. It tries to give you quite a lot, compared to what you gave it, so I don't think it's a surprise that it doesn't do that well. If you give it a chapter, it'll probably do a lot better.

 

[Louanne Learning]

You might hint at what “it” is emotionally, even if not scientifically.

Hey, that's kind of what I suggested - I'm as good as AI!