It's baffling to me that anyone would do this in the age of LLMs. All of the author's concerns could've been solved or greatly mitigated by loading her PKM into a model as context. The article doesn't mention that she even considered this as an option. I hope the files can be recovered when she realizes this possibility.
>All of the author's concerns could've been solved or greatly mitigated by loading her PKM into a model as context
No, because the correctly identified concern by the author is much deeper. Knowledge isn't some repository of data, or some solipsistic LLM simulacrum of it, it's practiced, social, contextual. A real brain, unlike the "second brain", intentionally forgets.
To have agency is to purposefully erase, cut through BS, and start from a clean slate. Any person who has done anything meaningful starts with an empty sheet of paper. There's a reason all these note taking gurus have exactly zero actual work to their name, it's just productivity LARPing. Every single time you see one of these "productivity" advocates you try to look up if they've build something that has helped even one person with all their productivity, nothing. They just keep yapping about their notes.
