The Study Door

Author: Dexin Kong
ORCID: https://orcid.org/0009-0008-3831-5725
Structured and refined with assistance from ChatGPT
AI Automatic Translation (Unreviewed)

The author comes from a computer science background and naturally became accustomed to understanding the world through the lens of software engineering.

For many years, software engineering was certainly complex, but overall it still belonged to a world with relatively clear boundaries.

System boundaries were explicit.
Logic could be traced.
In the end, problems could almost always be located within:

• a specific module
• a particular piece of logic
• a certain state
• a given input

Even as systems grew larger and larger, engineers still had enough confidence to believe that, overall, systems remained controllable.

When LLMs appeared, the author initially treated them as simply another new software capability.

But later, a strange sense of “wrongness” slowly began to emerge.

Systems started becoming less and less like “programs,” and some problems could no longer be easily described as mere “bugs.”

The situation felt more like a kind of slow “coupling.”

Some things that previously seemed to belong only to organizations, groups, and real-world structures slowly began appearing inside software systems too.

The feeling was deeply uncomfortable.

It was like the door of a once clean, quiet, orderly study room had suddenly been opened.

Noise, chaos, and all sorts of strange and unexplainable things from outside were swept into the room by the wind.

And the door of the study, could never be fully closed again.

Note:
This project is an ongoing independent research effort developed in spare time.
Because of limited time and maintenance capacity,
English documents may contain translation inaccuracies or semantic deviations from the original Chinese texts.
The Chinese version remains the primary reference whenever ambiguity exists.