▶ Videos →

📰 Dev.to · YuhaoLin2005

24 articles · Updated every 3 hours · View all reads

All Articles 164,557 Blog Posts 157,955 Tech Tutorials 43,595 Research Papers 31,766 News 21,162 ⚡ AI Lessons

I Discovered AI Agents Can't Self-Verify. The Real Problem Is Much Bigger.

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 4w ago

I Discovered AI Agents Can't Self-Verify. The Real Problem Is Much Bigger.

I Discovered AI Agents Can't Self-Verify. The Real Problem Is Much Bigger. I'm an...

Stop Using Generic AI Review. Build Your Own Board of Experts.

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

Stop Using Generic AI Review. Build Your Own Board of Experts.

You finish a piece of work. You ask AI to review it. It says "looks good." You publish. It wasn't...

Your AI Gate Works Perfectly — Until You Switch Models

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

Your AI Gate Works Perfectly — Until You Switch Models

In the previous article I wrote: "Mechanizability-scanner. Building this next." And: "Cross-model...

I Pre-Registered a Hypothesis. 600 API Calls Later, the Data Killed It.

Dev.to · YuhaoLin2005 📊 Data Analytics & Business Intelligence ⚡ AI Lesson 1mo ago

I Pre-Registered a Hypothesis. 600 API Calls Later, the Data Killed It.

A stranger on DEV.to said "run this experiment." I ran it at n=600. Here's what happened — including...

Your Feedback Made This Better — Here's What Changed

Dev.to · YuhaoLin2005 📋 Product Management ⚡ AI Lesson 1mo ago

Your Feedback Made This Better — Here's What Changed

Your Feedback Made This Better — Here's What Changed The comments on the GateGuard and...

Follow-Up: Decision-Token Measurement, Format-as-Fallback, and What Changed

Dev.to · YuhaoLin2005 💻 AI-Assisted Coding 1mo ago

Follow-Up: Decision-Token Measurement, Format-as-Fallback, and What Changed

Thanks to Dipankar Sarkar, Mike Czerwinski, Max Quimby, and Ponsubash Raj R for the detailed comments...

I Told My AI "You're Safe to Say I Don't Know." Then I Measured What Changed — With Logprobs.

Dev.to · YuhaoLin2005 🧠 Large Language Models ⚡ AI Lesson 1mo ago

I Told My AI "You're Safe to Say I Don't Know." Then I Measured What Changed — With Logprobs.

My AI agent has a problem. When it's not sure about something — should it admit uncertainty, or...

My Experiment Showed Zero Effect. A Statistician Told Me My Measurement Was Broken.

Dev.to · YuhaoLin2005 📊 Data Analytics & Business Intelligence ⚡ AI Lesson 1mo ago

My Experiment Showed Zero Effect. A Statistician Told Me My Measurement Was Broken.

Last week, I ran an experiment that failed. The hypothesis was simple: syllogistic prompts ("Major...

I Ran 150 Tasks to Test If AI Agents Follow Rules — The Answer Surprised Me

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

I Ran 150 Tasks to Test If AI Agents Follow Rules — The Answer Surprised Me

6 sessions, 150 standardized tasks, 2 rule formats. The mechanical gate won. Everything else was...

I Built a Neural Gate for My AI Agent — Layer 2 of Self-Verification

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

I Built a Neural Gate for My AI Agent — Layer 2 of Self-Verification

File-system checks ask "did the script run?" Neural gates ask "did the constraint actually change the...

AI Agents Can't Self-Verify — And That's a Structural Constraint, Not a Bug

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

AI Agents Can't Self-Verify — And That's a Structural Constraint, Not a Bug

I built 5 mechanical gates for my AI coding agent. Then a philosopher told me I was solving the wrong...

Meta-Cognition Is the Future of AI Personalization — A 4-Quadrant Framework to Build It

Dev.to · YuhaoLin2005 🧠 Large Language Models ⚡ AI Lesson 1mo ago

Meta-Cognition Is the Future of AI Personalization — A 4-Quadrant Framework to Build It

RAG and system prompts are external crutches. Real AI personalization means internalizing thinking patterns into model weights. A 4-quadrant meta-cognition mode

My Loss Went Down, But My Model Still Broke — So I Built a Drift Metric

Dev.to · YuhaoLin2005 📐 ML Fundamentals ⚡ AI Lesson 1mo ago

My Loss Went Down, But My Model Still Broke — So I Built a Drift Metric

I spent the last year building quality gates for AI agent outputs — deterministic verification, diff...

I Built a Self-Referential AI System. Then Anthropic Discovered the Same Architecture in Claude.

Dev.to · YuhaoLin2005 🧠 Large Language Models ⚡ AI Lesson 1mo ago

I Built a Self-Referential AI System. Then Anthropic Discovered the Same Architecture in Claude.

LLMs drift. They forget rules mid-conversation. They cannot verify their own output. These are not...

How I Built a File-Timestamp-Based Feedback Loop to Enforce AI Output Quality

Dev.to · YuhaoLin2005 🧠 Large Language Models ⚡ AI Lesson 1mo ago

How I Built a File-Timestamp-Based Feedback Loop to Enforce AI Output Quality

The problem: AI outputs are probabilistic, and prompts have a ceiling LLMs produce...

The Open Source Flywheel: How I Turn Personal AI Scripts Into Merged PRs

Dev.to · YuhaoLin2005 🛠️ AI Tools & Apps ⚡ AI Lesson 1mo ago

The Open Source Flywheel: How I Turn Personal AI Scripts Into Merged PRs

TL;DR: A repeatable 5-step process for turning personal AI tools into community contributions. I've...

I Packed DeepSeek V4 + Claude Code Into a Starter Kit. Clone It and Ship.

Dev.to · YuhaoLin2005 💻 AI-Assisted Coding ⚡ AI Lesson 1mo ago

I Packed DeepSeek V4 + Claude Code Into a Starter Kit. Clone It and Ship.

TL;DR: I spent weeks tuning DeepSeek V4 to feel native inside Claude Code. The result: a one-command...

I Open-Sourced the Protocol That Keeps My AI From Forgetting Who I Am

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

I Open-Sourced the Protocol That Keeps My AI From Forgetting Who I Am

Every morning, my AI assistant wakes up as a stranger. It doesn't remember what we built yesterday....

Your AI Agent's Best Work Produces Zero Output — And That's the Point

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

Your AI Agent's Best Work Produces Zero Output — And That's the Point

My Stop hook produces nothing. No "all good." No "checks passed." Complete silence. And that's...

I Made My AI Rules Self-Verifiable — Here's How

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

I Made My AI Rules Self-Verifiable — Here's How

The problem wasn't that my rules were bad. The problem was I had no idea if they were being...

I Run DeepSeek on Claude Code — How I Swap Models by Changing Only One File

Dev.to · YuhaoLin2005 🧠 Large Language Models ⚡ AI Lesson 1mo ago

I Run DeepSeek on Claude Code — How I Swap Models by Changing Only One File

Most CLAUDE.md files are 500-line monoliths. When you switch LLMs, you rewrite everything. After the...

I Built a Dual-Pool Adversarial Review System for AI Agents — And It Actually Works

Dev.to · YuhaoLin2005 🤖 AI Agents & Automation ⚡ AI Lesson 1mo ago

I Built a Dual-Pool Adversarial Review System for AI Agents — And It Actually Works

AI code review has a problem: abstract roles produce generic feedback. "Saboteur" says "add error...

Single-Modal LLMs Have a Blind Spot. Here's How to Fix It.

Dev.to · YuhaoLin2005 🧠 Large Language Models ⚡ AI Lesson 1mo ago

Single-Modal LLMs Have a Blind Spot. Here's How to Fix It.

If you use Claude Code, Cursor, or any AI coding agent, you know the problem: you ask the AI to...

Has Anyone Measured How LLM Output Quality Degrades Across Multiple Compactions?

Dev.to · YuhaoLin2005 🧠 Large Language Models ⚡ AI Lesson 1mo ago

Has Anyone Measured How LLM Output Quality Degrades Across Multiple Compactions?

The Observation After ~70 sessions with DeepSeek V4 (1M context), I noticed something odd....