One thing I noticed from my quarterly Claude vault analysis: the 30-minute review was useful, but the specification note I'd built with Claude beforehand was doing most of the work.
'That specification note is carrying a lot of the weight — it provides all the context the model uses.'
The AI is only as good as the context you give it. Full post: https://www.ctnet.co.uk/ai-obsidian-vault-analysis/
When Claude analysed my Obsidian vault, it mapped disconnected note clusters — groups that don't connect to the rest.
My clusters: AI and tech (unsurprising), gaming, personal finance, health and pressure.
What surprised me: it didn't flag philosophy, which I'd have expected.
Has anyone else run an AI analysis of their PKM? What did it find about your knowledge clusters?
#PKM #Zettelkasten #Obsidian #AI #KnowledgeManagement #SecondBrain
Title: P8: I have been reading about Knowledge Graphs and GraphGL. [2024-08-21 Wed]
+ One endpoint require caching and requiests monitoring
+ Not recommended to use for:
- unlimited binary data
- multi-user writing with versioning
- authentification services: cookies and redirect is problematic
Sources:
- https://neo4j.com/blog/what-is-knowledge-graph/
- https://memgraph.com/blog/introduction-to-node-embedding
😶 #dailyreport #graphs #retrival #knowledgemanagement
Title: P7: I have been reading about Knowledge Graphs and GraphGL. [2024-08-21 Wed]
| No contract | There is contract by schema |
| Errors returned by separate code | Errors returned in body |
| Versioning support | No versioning |
GraphGL drawbacks:
+ no limits: big query may cause overload of server
+ idempotency of request
- harder errors handlings #dailyreport #graphs #retrival #knowledgemanagement
Title: P5: I have been reading about Knowledge Graphs and GraphGL. [2024-08-21 Wed]
GraphGL is powerful query language and runtime that offers
several advantages over traditional RESTful APIs.
GraphGL solve REST problems:
+ advanced queries - good for frontend, good when you have many data sources
- reducing the amount of data transferred over the network.
- nested queries - *fetch data from multiple sources* #dailyreport #graphs #retrival #knowledgemanagement
Title: P2: P4: I have been reading about Knowledge Graphs and GraphGL. [2024-08-21 Wed]
sufficient for a product recommendation use case.
Ontologies are available in the OBO and OWL formats.
---------------------------------------------- #dailyreport #graphs #retrival #knowledgemanagement
Title: P3: I have been reading about Knowledge Graphs and GraphGL. [2024-08-21 Wed]
- if complexity will increase and it will require to store the properties of each relationship, it will be a
problem.
levels of KG:
- data - rows
- relationships - rows (node relationship node) - graph
database
- Organizing principles - additional nodes which is rules
or categories around the data that provide a flexible, #dailyreport #graphs #retrival #knowledgemanagement
CTNET
anoncheg