jokaye

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.

nof1.ai Alpha Arena 提示词工程逆向分析

逆向工程说明: 本文档基于 nof1.ai Alpha Arena 的公开文档、交易行为模式、API 响应格式和社区讨论,系统性地逆向推导出其 System Prompt 和 User Prompt 的完整结构，欢迎各路大佬戳戳评论，一起来进行这个有趣的实验。

Story: Writing Scripts with Go

This is a story about how I tried to use Go for scripting. In this story, I’ll discuss the need for a Go script, how we would expect it to behave and the possible implementations; During the discussion I’ll deep dive to scripts, shells, and shebangs. Finally, we’ll discuss solutions that will make Go scripts work.

Why Go is good for scripting?

While python and bash are popular scripting languages, C, C++ and Java are not used for scripts at all, and some languages are somewhere in between.

Scaling your API with rate limiters

The following are examples of the four types rate limiters discussed in the accompanying blog post. In the examples below I've used pseudocode-like Ruby, so if you're unfamiliar with Ruby you should be able to easily translate this approach to other languages. Complete examples in Ruby are also provided later in this gist.

In most cases you'll want all these examples to be classes, but I've used simple functions here to keep the code samples brief.

Request rate limiter

This uses a basic token bucket algorithm and relies on the fact that Redis scripts execute atomically. No other operations can run between fetching the count and writing the new count.

Golang Workers / Job Queue

A running example of the code from:

This gist creates a working example from blog post, and a alternate example using simple worker pool.

TLDR: if you want simple and controlled concurrency use a worker pool.

Installing ELK (CentOS)

This is a short step-by-step guide on installing ElasticSearch LogStash and Kibana Stack on a CentOS environment to gather and analyze logs.

I. Install JDK

rpm -ivh https://dl.dropboxusercontent.com/u/5756075/jdk-7u45-linux-x64.rpm

	sudo wget http://repos.fedorapeople.org/repos/dchen/apache-maven/epel-apache-maven.repo -O /etc/yum.repos.d/epel-apache-maven.repo
	sudo sed -i s/\$releasever/6/g /etc/yum.repos.d/epel-apache-maven.repo
	sudo yum install -y apache-maven
	mvn --version

	# Install Ruby

	wget http://ftp.ruby-lang.org/pub/ruby/2.1/ruby-2.1.1.tar.gz
	tar xvzf ruby-2.1.1.tar.gz
	cd ruby-2.1.1
	./configure --prefix=/usr
	make
	make install

	# remove "ruby-2.1.1" folder in /root

	class ModelProxy(object):
	"""
	Acts as a shim between the storage mechanism and our models.
	"""
	def __init__(self, instance, feature, path=''):
	self.feature = feature
	self.path = path
	self.coll = DB[self.feature]
	self.instance = instance

	"""A basic database set-up for Travis CI.

	The set-up uses the 'TRAVIS' (== True) environment variable on Travis
	to detect the session, and changes the default database accordingly.

	Be mindful of where you place this code, as you may accidentally
	assign the default database to another configuration later in your code.
	"""

	import os