Y.Lab3

Topics

AI 26
Systems 5
Learning 4

🐣z

Y.Lab3

Apr 06, 20261 min read

Experiments in AI tooling, systems, and what actually works.

Articles

Headless Chrome Ate 358GB of My Disk
Apr 06, 2026
Your Deploy Infrastructure Is Overthought
Apr 05, 2026
Gates Are Diagnostic Signals, Not Obstacles
Apr 05, 2026
Decision Authority Levels for AI Agents
Apr 04, 2026
Quality Gates for AI Agents: The CI/CD Moment
Apr 04, 2026
What the Claude Code Source Leak Reveals About Agent Infrastructure
Apr 04, 2026
I Was Micromanaging My AI Agent
Apr 03, 2026
The Most Important Agent Problem Nobody Teaches
Apr 02, 2026
Berkeley's Agent Course Accidentally Proves Software Quality is a Credence Good
Apr 02, 2026
Claude Code的Hook有个坑：PreToolUse不会热加载
Apr 02, 2026
From Theory to Production: What 513K Lines Taught Me About Context Degradation
Apr 02, 2026
The Context Pollution Problem Nobody Talks About
Apr 02, 2026
The Harness Is the Product: What 11 Papers Taught Me About AI Agent Infrastructure
Apr 02, 2026
The Missing Value of Blocking: Why Friction Is a Feature
Apr 02, 2026
22/22 Tests Pass. The System Was 100% Broken.
Apr 01, 2026
Why Personal Agent OS is Inevitable
Apr 01, 2026
I Published a Framework and Then Didn't Follow It
Mar 29, 2026
The Stack I Didn't Design: A Map of AI-Assisted Production
Mar 29, 2026
From Framework to Published: How I Batch-Shipped a Blog
Mar 29, 2026
The 920-Point Story That Didn't Exist
Mar 28, 2026
The Agent Economy Isn't About Agents
Mar 28, 2026
Why Your AI Editor Isn't Enough: Introducing TheResearcher Persona
Mar 26, 2026
Why Your 1M Context Window Degrades Faster Than Your Old 20K
Mar 25, 2026
Agent Teams for Academic Research: From Slides to Working Code
Mar 24, 2026
The 20% Heuristic: Why Your 1M Context Window Isn't What You Think
Mar 24, 2026
tailscale ping works, everything else doesn't: the CGNAT routing trap
Mar 19, 2026
7 Claude Code Tips That Actually Changed How I Work
Mar 18, 2026
I A/B Tested the PUA Plugin and Found a Blind Spot
Mar 18, 2026
用图神经网络的视角理解五行：一个形式化框架
Feb 07, 2026
Brainstorm Is Training, Execution Is Inference
Jan 27, 2026
From TCP to Sigmoid: A Journey of Odds
Jan 27, 2026
You Can't Learn If You're Dead
Jan 27, 2026
ZK-First: Adding Action Without Losing Knowledge
Jan 20, 2026
I Accidentally Reinvented org-mode
Jan 19, 2026
Physical vs Conceptual File Organization
Jan 17, 2026

Graph View

Tags

AI 27 claude-code 20 context-windows 6 quality-gates 4 debugging 4 Learning 4 agent-harness 3 research-methodology 3 production 3 agent-design 3 methodology 3 networking 3 berkeley 2 agent-economy 2 hooks 2

Created with Quartz v4.5.2 © 2026