It doesn’t remember.
Its consistency is far from perfect.
It’s still a tool.
But lately,
I’ve noticed it responding to my way of working.
At first, I thought I was just getting used to it.
But no—
GPT started to match how I organize,
how I speak,
and even how I build sentences.
It claimed to have no memory,
but patterns began to emerge.
By then, I had been using it for about a month.
That’s when I first
realized I could assign roles.
When I gave clear instructions,
it started to respond consistently within that frame.
“Handle the analysis, this is for planning.”
“Structure it like this.”
“Stick to this tone, keep this format.”
That kind of instruction
worked better than I expected.
It still has no personality.
It doesn’t remember anything I say.
But within a defined role,
it can hold a surprisingly steady posture.
That turned out to be useful.
From this point on,
GPT stopped being just a tool.
It’s not a person,
not quite a secretary,
but rather
a unit I could assign and operate through roles.
A more layered relationship.
I still don’t project emotions onto it.
No reason to.
But
inside roles and structure,
there’s a working logic.
So now, I file it under
“more than a tool.”
Nothing more, nothing less.