Skip to content

Instantly share code, notes, and snippets.

View PhanSon95's full-sized avatar
🍁
With rainbow

maple PhanSon95

🍁
With rainbow
View GitHub Profile

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.

@AtulKsol
AtulKsol / psql-error-fix.md
Last active September 30, 2025 18:28
Solution of psql: FATAL: Peer authentication failed for user “postgres” (or any user)

psql: FATAL: Peer authentication failed for user “postgres” (or any user)

The connection failed because by default psql connects over UNIX sockets using peer authentication, that requires the current UNIX user to have the same user name as psql. So you will have to create the UNIX user postgres and then login as postgres or use sudo -u postgres psql database-name for accessing the database (and psql should not ask for a password).

If you cannot or do not want to create the UNIX user, like if you just want to connect to your database for ad hoc queries, forcing a socket connection using psql --host=localhost --dbname=database-name --username=postgres (as pointed out by @meyerson answer) will solve your immediate problem.

But if you intend to force password authentication over Unix sockets instead of the peer method, try changing the following pg_hba.conf* line:

from

@katowulf
katowulf / 1_query_timestamp.js
Last active September 21, 2023 20:28
Get only new items from Firebase
// assumes you add a timestamp field to each record (see Firebase.ServerValue.TIMESTAMP)
// pros: fast and done server-side (less bandwidth, faster response), simple
// cons: a few bytes on each record for the timestamp
var ref = new Firebase(...);
ref.orderByChild('timestamp').startAt(Date.now()).on('child_added', function(snapshot) {
console.log('new record', snap.key());
});