Shane Xie xh4n3

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.

Helm 3 is storing description of it's releases in secrets. You can simply find them via

$ kubectl get secrets
NAME                                                TYPE                                  DATA   AGE
sh.helm.release.v1.wordpress.v1                     helm.sh/release.v1                    1      1h

If you want to get more info about the secret, you can try to describe the secret

$ kubectl describe secret sh.helm.release.v1.wordpress.v1

How many times shouldn't it happen...

-- https://news.ycombinator.com/item?id=11396045

SELECT count(*)
FROM (SELECT id, repo_name, path
        FROM [bigquery-public-data:github_repos.sample_files]
 ) AS F

Download videos from egghead

Install the React Developer Tools Chrome Extension.
Go to the egghead website, i.e. Getting Started with Redux
Click View -> Developer -> Javascript Console, then the React tab, then the <NextUpLessonList ...> tag.
Click back to the Console tab, then run:

	#!/usr/bin/python
	bpf_text = """
	#include <linux/ptrace.h>
	#include <linux/sched.h> /* For TASK_COMM_LEN */

	#include <linux/icmp.h>
	#include <linux/netdevice.h>

	struct probe_icmp_data_t
	{

	#!/bin/bash

	name=${1%.*c} # detatch .c from c file name
	clang-9 -O0 -Xclang -disable-O0-optnone -fno-discard-value-names -emit-llvm -c $name.c # convert source code(.c) to bit code(.bc)
	llvm-dis-9 $name.bc -o $name.ll # generate llvm ir(.ll) from bit code(.bc) obtained above without any optimization

	opt-9 -mem2reg $name.bc -o $name.mem2reg.bc # optimize bit code using `mem2reg` optimization pass before generating llvm ir
	llvm-dis-9 $name.mem2reg.bc -o $name.mem2reg.ll # generate llvm ir(.ll) from optimized bit code(.bc)

	#include <bcc/proto.h>
	#include <uapi/linux/ip.h>
	#include <uapi/linux/ipv6.h>
	#include <uapi/linux/icmp.h>
	#include <uapi/linux/tcp.h>
	#include <uapi/linux/udp.h>
	#include <uapi/linux/icmpv6.h>
	#include <net/inet_sock.h>
	#include <linux/netfilter/x_tables.h>

	#include <sstream>
	#include <iostream>
	#include <iomanip>

	std::string uint8_to_hex_string(const uint8_t *v, const size_t s) {
	std::stringstream ss;

	ss << std::hex << std::setfill('0');

	for (int i = 0; i < s; i++) {

	#!/bin/bash

	# Script to determine what named images and containers are dependent on the specified image (or a containers image) from stdin or in the args
	# An unmatchable input is ignored
	# All of the following will be found:
	# Any image name / tag / name:tag / id that has a prefix in it that matches the input
	# Any other image that relies on the images specified
	# Any containers (running or not) that use any of these images

	set -euf -o pipefail

	package main

	// @lizrice, mostly copied from @doctor_julz: https://gist.github.com/julz/c0017fa7a40de0543001

	import (
	"fmt"
	"os"
	"os/exec"
	"syscall"
	)