[visarga]

I would complement your knowledge management platform with a passive way to reference previous content. Save every paragraph of text that is being read/written and implement semantic similarity search. For example you could use a Hugging Face model and an Approximate Nearest Neighbor index.

Then integrate this search with an extension on top of your regular search engine, such as Google. Bonus if it logs across all apps, not just the web browser, and if it implements semantic and source based blacklisting to control what is being logged, for privacy.

This is complementary to your approach because manual logging is hard and slow while passive logging + semantic search can follow everything. It will be like an external memory.