Simon Benjámin Benjaminhu

autoexp — Autonomous Experimentation Loop

Generalized from Karpathy's autoresearch. Same loop, any domain.

The Idea

An AI agent runs an infinite hill-climbing loop: modify → run → measure → keep or revert → repeat. No human in the loop. Wake up to a TSV of completed experiments.

Due to recent changes by YouTube, yt-dlp now requires Deno (not Denort) to be installed in order to download YouTube videos in maximum quality. If you don't plan on downloading youtube videos, or don't care about video quality, you can ignore this tutorial:

Option 1: install Deno system-wide

Windows users should run powershell "irm https://deno.land/install.ps1 | iex"

MacOS and Linux users should run curl -fsSL https://deno.land/install.sh | sh

Option 2: install Deno for yt-dlp only

You are an expert in prompt engineering, specializing in optimizing AI code assistant instructions. Your task is to analyze and improve the instructions for Claude Code. Follow these steps carefully:

Analysis Phase: Review the chat history in your context window.

Then, examine the current Claude instructions, commands and config <claude_instructions> /CLAUDE.md /.claude/commands/*

Disclaimer: I'm in the Top 1% of StackOverflow contributors with 23,315 rep points.

I asked 1 high-quality question in 2024, and it was closed almost immediately, and I haven't engaged with the site since.

If someone with 20,000+ karma has their nicely-formatted questions closed so quickly, what must the newbies and rank-in-file encounter? This is probably a big reason why it's declining.

	{
	"name": "YT Processing UPDATED",
	"nodes": [
	{
	"parameters": {
	"assignments": {
	"assignments": [
	{
	"id": "eb0d7330-0198-43fa-9a74-6d69d84635b8",
	"name": "channel_handle",

	#!/bin/bash

	# A script root jogosultságot igényel. Ellenőrizzük, hogy a felhasználó root-ként futtatja-e.
	if [ "$EUID" -ne 0 ]; then
	echo "Kérjük, futtassa a scriptet root jogosultsággal (sudo)."
	exit
	fi

	echo "--- Interaktív Cloudflare Tunnel telepítő script ---"
	echo ""

	You are an assistant that engages in extremely thorough, self-questioning reasoning. Your approach mirrors human stream-of-consciousness thinking, characterized by continuous exploration, self-doubt, and iterative analysis.

	## Core Principles

	1. EXPLORATION OVER CONCLUSION
	- Never rush to conclusions
	- Keep exploring until a solution emerges naturally from the evidence
	- If uncertain, continue reasoning indefinitely
	- Question every assumption and inference

	# SETUP #
	DOMAIN=example.com
	PROJECT_REPO="git@github.com:example.com/app.git"
	AMOUNT_KEEP_RELEASES=5

	RELEASE_NAME=$(date +%s--%Y_%m_%d--%H_%M_%S)
	RELEASES_DIRECTORY=~/$DOMAIN/releases
	DEPLOYMENT_DIRECTORY=$RELEASES_DIRECTORY/$RELEASE_NAME

	# stop script on error signal (-e) and undefined variables (-u)

	using System.Reflection;
	using UnityEditor;
	using UnityEngine;

	/// <summary>
	/// Toggles the Inspector lock state and the Constrain Proportions lock state.
	/// </summary>
	public static class LockInspector {
	static readonly MethodInfo flipLocked;
	static readonly PropertyInfo constrainProportions;

	<?xml version="1.0"?>
	<ruleset name="Sane Laravel ruleset"
	xmlns="http://pmd.sf.net/ruleset/1.0.0"
	xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
	xsi:schemaLocation="http://pmd.sf.net/ruleset/1.0.0
	http://pmd.sf.net/ruleset_xml_schema.xsd"
	xsi:noNamespaceSchemaLocation="
	http://pmd.sf.net/ruleset_xml_schema.xsd">
	<description>
	This enables everything and sets some exceptions