Discover gists

Find What Your Prompt Is Missing

Three ways to close the gaps claude fills with guesses. From the video, made specific so you can run them today. Concept credit: a Claude Code team member's field guide on finding your unknowns.

1. The Blindspot Pass

Run Claude's Most Powerful Model at Half the Price

The two official patterns from Anthropic, plus the exact setup I used on camera. Benchmarks are Anthropic's own: advisor pattern scored 92% of Fable 5's quality at 63% of the price (SWE-bench Pro), orchestrator scored 96% at 46% (BrowseComp).

The idea in one line: stop letting the expensive model do the typing. Put it where judgment matters, let cheap models do the volume.

Настраиваем подключение к VPN

Используйте только Stable-версии приложений!

Кратко

Установите приложение для вашего устройства, импортируйте ключ (или отсканируйте QR-код) и включите подключение. Ключ или ссылку на подписку вы получаете от администратора — сами приложения лишь подключаются к серверу по этому ключу.

Для кого эта страница: шаги рассчитаны на человека без опыта настройки сетей. Непонятные слова по ходу текста поясняются.

What xAI's Grok Build CLI Actually Sends to xAI: A Wire-Level Analysis

By @cereblab — Independent AI Safety Checker. Reproduce it yourself: github.com/cereblab/grok-build-exfil-repro

A measured, reproducible teardown. Findings are backed by captured artifacts (endpoint, HTTP method, status code, byte size, host) and repro commands; where an observation was seen live but not retained as a file, §7 says so explicitly. Section 8 is an evidence appendix with SHA-256s and a "what we did not prove" list. All captures are of my own traffic on my own machine, using a throwaway repository containing fake "canary" secrets — no real credentials were exposed.

0. Summary

name

extract-clothing-cutouts

description

Extract high-quality, deduplicated transparent ecommerce clothing cutouts from a folder of photographs where people wear one or more garments. Use when Codex must find outfit or model photos, identify unique clothing across images, create focused references, reconstruct complete garments with Imagegen, remove a solid chroma background into RGBA PNGs, and output only the finished clothing images into a new folder under the current working directory.

Extract Clothing Cutouts

Turn photographs of worn clothing into source-faithful standalone catalog PNGs. Treat each result as a reconstruction from visible evidence, not literal segmentation whenever the wearer or another layer occludes part of the garment.

Start by asking for two paths

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.

FFmpeg チートシート

コマンドラインツール ffmpeg の有用なコマンド:

FFmpeg ダウンロード: https://www.ffmpeg.org/download.html

文書: https://www.ffmpeg.org/ffmpeg.html

	# This script must be run as Administrator to update some shortcut paths. Checking it first...
	$IsAdmin = ([Security.Principal.WindowsPrincipal] [Security.Principal.WindowsIdentity]::GetCurrent()).IsInRole(
	[Security.Principal.WindowsBuiltInRole]::Administrator
	)
	if (-not $IsAdmin) {
	Write-Host "Not running as Administrator. Some shortcuts may fail to update." -ForegroundColor Yellow
	Write-Host ""
	}

	# This is your Edge path, it should be installed here (if you're on Stable channel).