I put your previous post and my reply into ChatGPT for Amusement and irony:
This is solid—clear, concrete, and it makes a point most people half understand but don’t articulate well: prompting is a skill, and yes, it changes how people talk.
If you want my straight-shooter response (and a few edits that would make it land better on a music chat site), here it is.
What you did well
1) You grounded it in a real example
That pizza story is gold because it’s low-stakes, ordinary, and believable, and it shows exactly what you mean without needing theory.
2) You made a nuanced defense
You aren’t saying “ChatGPT is always great.”
You’re saying: it can be useful, but users often misuse it, and the interface encourages weird behavior. That’s credible.
3) You nailed the social contrast
The “you’d never speak like that to a human” point is excellent. That’s the part that will stick.
The main weakness (and it’s fixable)
Your argument has one shaky assumption:
“An AI algorithm cannot intuitively understand implied things in normal speech.”
That’s partly true, but not exactly.
Modern chat models can understand implied intent pretty well. The bigger issue is:
-
they won’t assume missing facts in many cases (because making things up is dangerous),
-
they sometimes refuse, especially when uncertain,
-
they sometimes “play safe” with generic advice (“call the restaurant”) when they feel the question demands specifics they don’t have.
So it’s not only “AI is dumb unless you speak machine.”
It’s more like: AI is cautious, and users must help it by supplying constraints or allowing assumptions.
If you adjust that framing, your piece becomes much harder to argue with.
One line I’d revise (because critics will pounce)
This:
“I think that as AI is used more widely, people will in turn be trained through its regular use to speak more like a machine.”
This is interesting, but it’s also the most debatable claim in your whole post.
A counterargument is:
-
or AIs will get better at normal speech, so people won’t have to change.
I actually think the truth is: both happen.
People learn some “prompt habits,” but models also improve so the habits become less necessary.
So I’d soften that line slightly so you don’t hand opponents an easy target.
A stronger version of your core point (shorter, punchier)
If you want it to hit harder online, the main thesis could be distilled to:
“People judge ChatGPT by treating it like a human and then getting mad when it isn’t. It’s closer to a tool: it needs constraints, assumptions, and context. When used that way, it often gives useful help—especially for low-stakes planning—without pretending it has perfect knowledge.”
That keeps your spirit but makes the logic more defensible.
The pizza story: one tweak that makes it even better
Right now the story slightly undermines your own point because a critic could say:
-
“So the AI didn’t answer until you bullied it—sounds bad.”
You can flip that by adding one small clarification:
“Once we stopped asking it to magically know our driving schedule and instead asked for one missing variable—the typical prep time—it became useful.”
That makes it about good tool use, not about “AI refuses unless threatened.”
My overall response to your post
You’re not wrong at all. You’re describing the cultural moment accurately:
-
People are impatient.
-
They anthropomorphize the tool.
-
They get offended when it can’t do mind-reading.
-
They also use it for novelty and then complain it isn’t magic.
And your final comparison—humans would hang up—is a killer line because it exposes the weird emotional entitlement people bring to machines.
If you want, I can do a light edit that keeps 95% of your wording but:
-
trims repetition,
-
strengthens the claims,
-
makes it “chat-site length” without losing the nuance.
