I have no problem with a personal note that helps a tool understand the way I work.
I do have a problem with the idea that artificial intelligence should get my whole life at once: finances, health, relationships, fears, goals, writing style, past failures and a five-year plan. That is not neutral background. It is a frame through which the model will start to interpret the questions that follow.
If I describe myself mainly through limitations, the tool may start to justify them very convincingly. If the first assumption is wrong, the following answers may only build that error out further. Excess information also means a higher cost and a wider range of private data entrusted to an external service.
I prefer to pass on context in layers.
I start with the current problem and the result I want to get. I check the first answer. If the direction is good, I add the facts that are needed. If it is bad, I correct the assumptions or start the conversation again.
When working for a company I apply the same rule. A model does not need the whole history of the business to put one procedure in order. It should get the data needed for the task, in as small a scope as possible and without information that changes nothing.
The point is not for AI to know everything about me. The point is for me to know what I am passing on to it, why I am doing so and which part of the answer I still have to judge myself.
The best context is not the biggest file.
The best context is the smallest set of information that lets the task be done well.