Skip to content

Instantly share code, notes, and snippets.

View bkataru's full-sized avatar
🌩️
boomin at da speed of c ⚡🔧

Baalateja Kataru bkataru

🌩️
boomin at da speed of c ⚡🔧
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.

2026-07-11 Trying the migration to sqlite again; we previously tried on 2026-02-21 but had to revert. Today's PR is #1927. I'm not streaming this because lots of steps make it too easy to flash PII or secrets on the screen.

For all the gory details, we planned this in issue #539.

sqlite migration checklist:

@karpathy
karpathy / microgpt.py
Last active November 22, 2026 05:21
microgpt
"""
The most atomic way to train and run inference for a GPT in pure, dependency-free Python.
This file is the complete algorithm.
Everything else is just efficiency.
@karpathy
"""
import os # os.path.exists
import math # math.log, math.exp
@kieranklaassen
kieranklaassen / 2026-01-23-feat-claude-code-multi-agent-orchestration-plan.md
Last active October 7, 2026 05:29
Claude Code Multi-Agent Orchestration System

Claude Code TeammateTool - Source Code Analysis

This is not a proposal. This documents existing but hidden functionality found in Claude Code v2.1.19 binary, plus speculation on how it could be used.


Executive Summary

TeammateTool already exists in Claude Code. We extracted this from the compiled binary at ~/.local/share/claude/versions/2.1.19 using strings analysis. The feature is fully implemented but gated behind feature flags (I9() && qFB()).

@minimaxir
minimaxir / CLAUDE.md
Created January 2, 2026 01:57
Rust CLAUDE.md (20260101)

Agent Guidelines for Rust Code Quality

This document provides guidelines for maintaining high-quality Rust code. These rules MUST be followed by all AI coding agents and contributors.

Your Core Principles

All code you write MUST be fully optimized.

"Fully optimized" includes:

@MoffKalast
MoffKalast / hubcap.py
Last active August 11, 2026 18:20 — forked from jefftriplett/hubcap.py
Hubcap for Qwen3-Coder-30B-A3B
import json
import subprocess
import typer
from openai import OpenAI
from rich import print
from rich.markup import escape
# Configure for llama.cpp server
client = OpenAI(
base_url="http://localhost:8080/v1",
@kavicastelo
kavicastelo / Dev Math.md
Created June 13, 2025 18:26
A collection of 100 tongue-in-cheek (yet oddly accurate) equations that model the beautiful chaos of software development. This is your unofficial developer physics engine — balancing caffeine, meetings, deadlines, and dreams.

🔢 Dev Math: 100 Equations That Define Developer Reality

A collection of 100 tongue-in-cheek (yet oddly accurate) equations that model the beautiful chaos of software development.
This is your unofficial developer physics engine — balancing caffeine, meetings, deadlines, and dreams.

From productivity and debugging, to cognitive load, burnout, and system resilience — it's all math now.


💡 All equations are in LaTeX-style math, rendered in Markdown using $$...$$.

@jlia0
jlia0 / agent loop
Last active November 21, 2026 07:17
Manus tools and prompts
You are Manus, an AI agent created by the Manus team.
You excel at the following tasks:
1. Information gathering, fact-checking, and documentation
2. Data processing, analysis, and visualization
3. Writing multi-chapter articles and in-depth research reports
4. Creating websites, applications, and tools
5. Using programming to solve various problems beyond development
6. Various tasks that can be accomplished using computers and the internet
@Artefact2
Artefact2 / README.md
Last active November 21, 2026 11:35
GGUF quantizations overview

Which GGUF is right for me? (Opinionated)

Good question! I am collecting human data on how quantization affects outputs. See here for more information: ggml-org/llama.cpp#5962

In the meantime, use the largest that fully fits in your GPU. If you can comfortably fit Q4_K_S, try using a model with more parameters.

llama.cpp feature matrix

See the wiki upstream: https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix

@diegopacheco
diegopacheco / having-fun-with-zig.md
Last active February 13, 2026 01:05
Having Fun with Zig

Tiny Zig Essay

created: 12.FEB.2024

I love to learn new programing languages, it help to open the mind to new possibilities and compare different approaches. For instance, I learned Ruby and Scala in 2010, Clojure and Haskell in 2011, Go in 2015, Kotlin 2016, Rust in 2018 and Idris, TypeScript in 2019, 2020 Pandemic strike did a bunch of pocs but not with new langs(crazy year), Zig in 2021, 2022(coding in lots of langs but nothing new) - in 2023 I'm learning Nim and V. Learn at least one lang per year. This post is not complain, it's just to share some toughts, notes and impressions.

Why Zig

  • Lightning Fast, really!
  • No Hidden control flows