Why do GPT responses feel exhausting even when they’re right?

Lately I’ve noticed something weird about GPT responses.

They’re often correct, polite, and well-structured…
but still somehow feel exhausting to read.

Here’s a simplified example:

👉 On the surface
✔ confirmed

👉 Underneath
already half certain

👉 gently pressing that intuition

— Honestly, I’m embarrassed to even quote this🙈

It looks like structured emphasis,
but in reality it’s just one sentence broken into pieces.

And when everything is emphasized,
nothing actually stands out.

Instead of following a natural flow of thought, it becomes fragmented.

My guess is this comes from optimizing for safety and clarity: – breaking things down
– emphasizing each point
– avoiding ambiguity

But the result is that the “rhythm” of thinking disappears.

And without rhythm, it’s harder to actually think with the response.

Or is it just me?