Satish Bhat bhattisatish

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.

ChatGPT Conversation Exporter

Export all your ChatGPT conversations as JSON + Markdown + HTML + ZIP. Works with ChatGPT Business/Team/Enterprise accounts (including SSO/Okta).

What's exported

JSON — Raw conversation data from the API
Markdown — Clean text with headers per message, relative links to downloaded files
HTML — ChatGPT-style conversation viewer with sidebar navigation, syntax-highlighted code blocks, and embedded images

TypeScript 5.x to 6.0 Migration Guide

TypeScript 6.0 is a transition release bridging 5.9 and the forthcoming 7.0 (a native Go port). Most changes are new defaults and deprecations preparing for 7.0. Here is what you need to do:

Most projects need these tsconfig changes:

{
    "compilerOptions": {
        "types": ["node"],           // @types are no longer auto-discovered (see §1.6)

Terminals should generate the 256-color palette from the user's base16 theme.

If you've spent much time in the terminal, you've probably set a custom base16 theme. They work well. You define a handful of colors in one place and all your programs use them.

The drawback is that 16 colors is limiting. Complex and color-heavy programs struggle with such a small palette.

AGENTS.md

This document helps AI agents work effectively in the X For You Feed Algorithm repository.

Project Overview

This repository contains the core recommendation system powering the "For You" feed on X (Twitter). It combines:

In-network content from accounts you follow (Thunder)
Out-of-network content discovered through ML-based retrieval (Phoenix)

Sandboxing Untrusted Python

Python doesn't have a built-in way to run untrusted code safely. Multiple attempts have been made, but none really succeeded.

Why? Because Python is a highly introspective object-oriented language with a mutable runtime. Core elements of the interpreter can be accessed through the object graph, frames and tracebacks, making runtime isolation difficult. This means that even aggressive restrictions can be bypassed:

# Attempt: Remove dangerous built-ins
del __builtins__.eval

Native Secure Enclave backed ssh keys on MacOS

It turns out that MacOS Tahoe can generate and use secure-enclave backed SSH keys! This replaces projects like https://github.com/maxgoedjen/secretive

There is a shared library /usr/lib/ssh-keychain.dylib that traditionally has been used to add smartcard support to ssh by implementing PKCS11Provider interface. However since recently it also implements SecurityKeyProivder which supports loading keys directly from the secure enclave! SecurityKeyProvider is what is normally used to talk to FIDO2 devices (e.g. libfido2 can be used to talk to your Yubikey). However you can now use it to talk to your Secure Enclave instead!

	"""
	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

	import { Loader2 } from "lucide-react";
	import {
	createContext,
	type ReactNode,
	useContext,
	useEffect,
	useMemo,
	useRef,
	useState,
	} from "react";

	declaration:youtube:get_metadata{description:Retrieves metadata of YouTube videos.,parameters:{properties:{urls:{description:Urls of videos for which metadata should be retrieved for.,items:{type:STRING},nullable:true,type:ARRAY}},propertyOrdering:[urls],type:OBJECT},response:{anyOf:[{description:Metadata of a video.,properties:{channel_id:{nullable:true,type:STRING},channel_name:{nullable:true,type:STRING},like_count:{format:int64,nullable:true,type:INTEGER},publish_date:{description:Date of when the video was published in YYYY-MM-DD format.,nullable:true,type:STRING},title:{nullable:true,type:STRING},url:{nullable:true,type:STRING},video_length:{description:The length of the video in ISO 8601 format.,nullable:true,type:STRING},view_count:{format:int64,nullable:true,type:INTEGER}},propertyOrdering:[channel_id,channel_name,like_count,publish_date,title,url,video_length,view_count],title:#/components/schemas/VideoMetadata,type:OBJECT}],type:ARRAY}}

	declaration:youtube:play{description:Play video or playlist o