from Diary 2024-09-24 Nodal point of thought 2024-09-24
- I think External brain GitHub work and 2024-09 Improve vector search and Extract questions at various levels of abstraction are related.
Utility gained by accumulating information
- If this goes down slowly, eventually the utility gained will be less than the cost of adding it, and it will not be added.
- This is a three-day system. Scrapbox enhances the ability to discover and reuse past relevant articles compared to blogs.
- It will improve âprofit per additionâ for a while. They donât keep improving.
- For example, the KJ method already has over 100 links, and the utility of adding another link is low.
- For example, there are various pages on the subject of âintellectual productionâ such as Keicho, Kozaneba, and its predecessor Regroup, the KJ method, etc., but they are not well connected by âkeyword string matchingâ.
-
Nodal point of thought 2024-02-21 ăèȘăă§ăăè ă«èœăĄăăæ°æăĄăćąăă
- I wouldnât have known this, either, if this layer hadnât happened to be written on this untitled page.
- I really think itâs right that itâs presented as an abstraction from a bunch of concrete pages, and then I look at it, and if I think itâs different, then I revise it, and then the higher-order pages are created over and over again.
- Raising the layers
- I tend to write more concrete descriptions because I dislike thinking only in the abstract layer without connecting it to the concrete layer because I consider it ânot being groundedâ.
- KJ Method Ideas
- Each of the 20,000 pages of this Scrapbox is like a fragment of the KJ method
- Higher order structures are discovered as they are bundled together.
- The discovery process is not all done by the machine, but the interaction with humans.
These areas are improved by TTTC extraction and Extract questions at various levels of abstraction in ways that were not feasible before the LLM.
This page is auto-translated from /nishio/æèăźç”çŻçč2024-09-24 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. Iâm very happy to spread my thought to non-Japanese readers.