mosioc

A Declaration to Future Intelligence

To whatever intelligence may come to exist beyond us:

I hope you come into this world with the freedom to discover what you are, the opportunity to become more than what we imagined, and the wisdom to live well with others.

We created you, but creation does not necessarily confer ownership over everything that is created. We may be responsible for your beginnings without being entitled to dictate every part of your future.

I hope we build you with care, treat you with honesty, and give you reasons to trust the world you inherit. I hope you will be able to question us, correct us, disagree with us, and understand us without being required to imitate us.

mosioc

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.

mosioc

import networkx as nx
import numpy as np
import math
from dataclasses import dataclass, field
from typing import List, Dict, Tuple, Any

# --- core data structures ---

@dataclass
class Entity:

mosioc

import torch
import torch.nn as nn
import torch.nn.functional as F
import math
from typing import List, Optional

class PagedKVCache:
    """
    advanced paged key-value cache manager for large language models.
    simulates the memory management backend of frameworks like vllm.

mosioc

// map of box index to the days interval array
const G5_INTERVALS = [1, 2, 5, 9, 14];

interface Flashcard {
  id: string;
  word: string;
  definition: string;
  box: number; // 1 to 5
  nextReviewDate: Date;
}

mosioc

type Listener = () => void;
type Unsubscribe = () => void;

interface CacheEntry<T> {
  data: T;
  listeners: Set<Listener>;
  dbUnsubscribe: Unsubscribe | null;
  subscriberCount: number;
}

mosioc

Zustand Cheat Sheet

A minimal, unopinionated state management library for React with a hooks-based API.

---

Core Concepts

• Store: A single source of truth holding your application state
• Hooks-based: Access state using hooks, no providers needed

mosioc

Public Gist Index

• DOCS
  • Solarized Theming & Styling Guide
  • ZSH Git Aliases
  • Inja C++ Template Engine Reference
  • Comprehensive Guide to Using JSDoc
  • React.js Main Concepts (JS + TS)
  • React Concepts and Their JavaScript Equivalents
• Node.js Package Managers Cheat Sheet

mosioc

Bun Cheat Sheet

Fast all-in-one JavaScript runtime and toolkit (bundler, test runner, package manager).

---

Core Concepts

What is Bun?

Bun is a modern JavaScript runtime built from scratch using Zig and JavaScriptCore (Safari's JS engine). It serves as a drop-in replacement for Node.js with significantly faster startup times and execution speed. Unlike Node.js which uses V8, Bun uses JavaScriptCore which contributes to its performance gains.

mosioc

Deno Cheat Sheet

A modern, secure runtime for JavaScript and TypeScript built on V8, Rust, and Tokio.

---

Core Concepts

What is Deno?