惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

J
Java Code Geeks
腾讯CDC
M
MIT News - Artificial intelligence
Y
Y Combinator Blog
L
LangChain Blog
Vercel News
Vercel News
云风的 BLOG
云风的 BLOG
GbyAI
GbyAI
Stack Overflow Blog
Stack Overflow Blog
Microsoft Azure Blog
Microsoft Azure Blog
B
Blog RSS Feed
The GitHub Blog
The GitHub Blog
酷 壳 – CoolShell
酷 壳 – CoolShell
B
Blog
P
Proofpoint News Feed
H
Hackread – Cybersecurity News, Data Breaches, AI and More
博客园_首页
Google DeepMind News
Google DeepMind News
WordPress大学
WordPress大学
aimingoo的专栏
aimingoo的专栏
小众软件
小众软件
IT之家
IT之家
A
About on SuperTechFans
H
Help Net Security

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
I Shipped a Database in 4 Days a Week. Here's What Git Sa...
contour · 2026-06-02 · via DEV Community

contour

I'm not going to tell you if 4-day weeks work.

I'm going to show you my git log for the last 90 days. You decide.

Context: I write C++ search engines. Solo. No meetings. No manager. Just me and a codebase called glyph-engine.

For 3 months, I worked Mon-Thu. Fri-Sun I wrote zero code. No laptop. No "quick fixes".

Here's what happened to my output, my bugs, and my brain.

The Data: My GitHub Doesn't Lie

First, proof I'm not making this up. This is my May 2026: 363 commits across 6 repos.

Notice the pattern? Heavy Mon-Thu. Dead Fri-Sun. That's intentional.

187 of those commits went to glyph-engine, a disk-based FM-index I built:

The repo got traction too. 1,043 clones and 414 unique cloners in 14 days:

Now, the actual experiment. I compared 3 months of 5-day weeks vs 3 months of 4-day weeks on the same codebase.

Results:

Metric 5-Day Weeks 4-Day Weeks Change
Commits / Week 22 14 -36%
Lines Added / Week 1,240 780 -37%
Lines Deleted / Week 410 1,190 +190%
Bugs per 1k LOC 4.2 1.7 -59%
Days to Merge Feature 9.2 4.8 -48%

Translation: I wrote 37% less code. But I deleted 190% more garbage. Net result: features shipped 2x faster with 2.5x fewer bugs.

Why It Worked: The 3-Day Defrag

My brain has cache. Just like a CPU.

Day 1 off: L1 cache flush. I'm still thinking about yesterday's segfault. Useless.

Day 2 off: L2 cache flush. I start forgetting variable names. Good.

Day 3 off: L3 cache flush. This is where magic happens. I wake up and realize my suffix array layout thrashes the TLB. The fix is obvious.

With 2-day weekends, I never hit Day 3. I was stuck debugging symptoms. With 3 days, I fixed root causes before I wrote them.

My CPU does 37 GB/s. My brain does ~37 thoughts/s. Both need idle cycles to defrag. 2 days isn't enough.

When This Fails: The Data Is Brutal

I'm not selling 4-day weeks. They broke my workflow 3 times:

  1. On-call week. Server panic on Friday. I was "off". Worked 6 hours anyway. Result: 5-day week + guilt + broken family time. Net negative.
  2. Upstream dependency. My colleague ships on Fridays. I was blocked every Monday until he replied. Lost 20% of Mon to context-switching.
  3. Crunch time. 2 weeks before a demo, I switched back to 7 days. 4-day weeks are for marathons. Sprints need all hands.

External data backs this up:

  • Cambridge University 2022: 61 companies, 6 month trial. Burnout -71%, Revenue +1.4%. You feel better. The company doesn't print money.
  • Buffer 2020: Tried 4-day for engineers. Rolled back. Reason: collaboration overhead killed flow.
  • Microsoft Japan 2019: +40% sales productivity. But they were sales, not engineers building databases.

The Actual Takeaway: Maker Schedule, Not 4-Day Week

Don't ask your boss for a 4-day workweek. Ask for this:

  1. Block 4 days. No meetings. No Slack. No code review. Just creation.
  2. Take 3 days off. Really off. No laptop. Your L3 cache needs it.
  3. Rotate on-call. Pay them 2x. A 4-day week with on-call is just a 5-day week with lies.

If you can't get #1, a 4-day week will make you slower.
If you can, your git log might look like mine.

My boss didn't give me a 4-day week. I gave it to myself. I tracked the data.
And for this project, alone, writing C++, the data says it worked.

Your codebase is different. Your team is different.

Don't trust me. Trust your git log.
Repo: github.com/yasha1971-coder/glyph-engine
Traffic: 1,043 clones in 14 days during this experiment.