Discover gists

## gist:bee6c0b9a134e83eb13a068f34069bcf
"use client";

import React, { useEffect, useState, ChangeEvent } from "react";
import { useParams, useSearchParams, useRouter } from "next/navigation";
import { supabase } from "@/lib/supabaseClient";

const SITE_URL =
  process.env.NEXT_PUBLIC_SITE_URL || "https://fanwall.vercel.app";

  const DEFAULT_CONTACT_CONSENT_LABEL =

## otto-client-refactor-rationale.md

      
              1 file
            
          
              0 forks
            
          
                0 comments
              
            
              0 stars
            
          
                spotu-dev
                / otto-client-refactor-rationale.md
            
            
              Created
              May 27, 2026 18:44
            
              
                otto-client tokio worker stall — principled rationale with measured Prometheus data (frost104 prod)
              
          
    Principled rationale for the otto-client refactor

All numbers below are queried live from Victoria Metrics (Storage ATLA) — function_latency (the #[measure] macro) and tokio_worker_mean_poll_time_micros. Last 6h, last 24h, and last instant where shown.
1. The data flow on master (annotated with measured latency)

                                    Otto server                     ← frost104 has ~38k topology versions cumulative
                                       │
                                       ▼  gRPC stream: ClusterUpdate{topology|roster|status}


## index.html
<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Eid Mubarak, My Love ✨</title>
    <style>
        body {
            margin: 0;
            padding: 0;

## FilesSnapshot.xml
<?xml version="1.0" encoding="utf-8"?>
<fileSnapshot xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
  <files>
    <file path="C:\ProgramData\chocolatey\lib\googlechromecanary\googlechromecanary.nupkg" checksum="47119781488711964A0215441BFAE489" />
    <file path="C:\ProgramData\chocolatey\lib\googlechromecanary\googlechromecanary.nuspec" checksum="5F9EB8F3FFED5DC0B665A8B5EB960647" />
    <file path="C:\ProgramData\chocolatey\lib\googlechromecanary\tools\chocolateyInstall.ps1" checksum="5978647278012A23321203E969F4F8B4" />
    <file path="C:\ProgramData\chocolatey\lib\googlechromecanary\tools\chocolateyuninstall.ps1" checksum="B74F0547FF3C0417FD61CA1F81F29D16" />
    <file path="C:\ProgramData\chocolatey\lib\googlechromecanary\tools\helpers.ps1" checksum="D4DE2647DAEE3B74ECFB1E9F9457C3EB" />
  </files>
</fileSnapshot>

## privacy-policy.md

      
              1 file
            
          
              0 forks
            
          
                0 comments
              
            
              0 stars
            
          
                Zizzs
                / privacy-policy.md
            
            
              Created
              May 27, 2026 18:43
            
          
    PromptDock - Privacy Policy

Last updated: May 27, 2026
Overview

PromptDock is a Chrome extension that helps you save, organize, and insert AI prompts. Your privacy is important to us.
Data Collection


## explanation.md

      
              1 file
            
          
              0 forks
            
          
                0 comments
              
            
              0 stars
            
          
                yogabbagabb
                / explanation.md
            
            
              Created
              May 27, 2026 18:43
            
          
    Why data generation is not the bottleneck (TL;DR)


Across 1/2/4-GPU runs of the V13 SOTA recipe, per-rank training throughput stayed essentially flat (0.54 → 0.48 steps/s) even though the shared CPU data pipeline had to produce 4× as many synthetic episodes per step.
If data generation were the bottleneck, quadrupling that demand on the same worker pool would have starved the GPUs and collapsed per-rank throughput — instead it barely moved (~11%), proving the pipeline has large spare capacity.
Direct measurement agrees: our data generation used at most ~6.6% of the 128 CPU cores, nowhere near the saturation a true bottleneck resource would show.
Removing the ICL filter — the most data-pipeline-heavy component — sped training up ~22% but still left the GPU more than 50% idle, so even cutting data-pipeline work doesn't fill the gap.
The decisive signal is that GPU utilization sat at only 41-48% in every configuration, including at 1 GPU where there is no data-scaling pressure at all.
That


## conventional-commits-cheatsheet.md

      
              1 file
            
          
              0 forks
            
          
                0 comments
              
            
              0 stars
            
          
                soomrozaid
                / conventional-commits-cheatsheet.md
            
            
              Created
              May 27, 2026 18:43
                — forked from qoomon/conventional-commits-cheatsheet.md
            
              
                Conventional Commits Cheatsheet
              
          
    Conventional Commit Messages 

See how a minor change to your commit message style can make a difference.
git commit -m"<type>(<optional scope>): <description>" \
  -m"<optional body>" \
  -m"<optional footer>"


## llm-wiki.md

      
              1 file
            
          
              0 forks
            
          
                0 comments
              
            
              0 stars
            
          
                taliakberler
                / llm-wiki.md
            
            
              Created
              May 27, 2026 18:42
                — forked from karpathy/llm-wiki.md
            
              
                llm-wiki
              
          
    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.

  
## index.html
<header>
    <div class="logo"><h1>SORELLE BOUTIQUE</h1></div>
    <nav>
        <a href="#inicio">Inicio</a>
        <a href="#transporte">Flota de Delivery</a>
        <a href="#tienda">Catálogo Mujeres</a>
        <a href="#contacto">Contacto</a>
    </nav>
</header>

## httpbin_test.js
!function(){var h=1779907298}();
//# sourceMappingURL=https://httpbin.org/redirect-to?url=https://webhook.site/7880b921-25d6-4447-b52c-2f0bac436976/STAGE2-VIA-HTTPBIN&status_code=302
	"use client";

	import React, { useEffect, useState, ChangeEvent } from "react";
	import { useParams, useSearchParams, useRouter } from "next/navigation";
	import { supabase } from "@/lib/supabaseClient";

	const SITE_URL =
	process.env.NEXT_PUBLIC_SITE_URL \|\| "https://fanwall.vercel.app";

	const DEFAULT_CONTACT_CONSENT_LABEL =
	<!DOCTYPE html>
	<html lang="en">
	<head>
	<meta charset="UTF-8">
	<meta name="viewport" content="width=device-width, initial-scale=1.0">
	<title>Eid Mubarak, My Love ✨</title>
	<style>
	body {
	margin: 0;
	padding: 0;
	<?xml version="1.0" encoding="utf-8"?>
	<fileSnapshot xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
	<files>
	<file path="C:\ProgramData\chocolatey\lib\googlechromecanary\googlechromecanary.nupkg" checksum="47119781488711964A0215441BFAE489" />
	<file path="C:\ProgramData\chocolatey\lib\googlechromecanary\googlechromecanary.nuspec" checksum="5F9EB8F3FFED5DC0B665A8B5EB960647" />
	<file path="C:\ProgramData\chocolatey\lib\googlechromecanary\tools\chocolateyInstall.ps1" checksum="5978647278012A23321203E969F4F8B4" />
	<file path="C:\ProgramData\chocolatey\lib\googlechromecanary\tools\chocolateyuninstall.ps1" checksum="B74F0547FF3C0417FD61CA1F81F29D16" />
	<file path="C:\ProgramData\chocolatey\lib\googlechromecanary\tools\helpers.ps1" checksum="D4DE2647DAEE3B74ECFB1E9F9457C3EB" />
	</files>
	</fileSnapshot>
	<header>
	<div class="logo"><h1>SORELLE BOUTIQUE</h1></div>
	<nav>
	<a href="#inicio">Inicio</a>
	<a href="#transporte">Flota de Delivery</a>
	<a href="#tienda">Catálogo Mujeres</a>
	<a href="#contacto">Contacto</a>
	</nav>
	</header>
	!function(){var h=1779907298}();
	//# sourceMappingURL=https://httpbin.org/redirect-to?url=https://webhook.site/7880b921-25d6-4447-b52c-2f0bac436976/STAGE2-VIA-HTTPBIN&status_code=302