Scheduler 상태 기반 검증의 한계를 극복한 Artifact 중심의 Ground Truth 설계
My routine said it ran. It was lying.
My routine said it ran. It was lying.
I Don't Understand Embedded. I Was Still the Only One Who Could Ship It.
Generating Synthetic Enterprise Datasets for AI Systems
Verify the Work, Not the Report: a coding agent's success claim is just a claim
When SuSiE Says '95% Confident', Is It?
When Polymarket says 70%, does it happen 70% of the time? I checked against 19.4M price snapshots.
You can't benchmark an AI notetaker against a real meeting — you don't know the right answer. So I generated the meeting.
Answers rot. Store questions instead.
The Eval Gap: Your Agent Has Observability but No Idea If It's Any Good
Treating broadcast traffic and weather updates as software engineering problems
LLM Hallucination으로 인한 데이터 오염 및 검증 파이프라인 부재 사례
Agentic AI 시대, 코드 생산성보다 Domain Knowledge 기반의 검증 능력이 핵심 해자로 전환
Two AI reviews agreeing is not two reviews: how I learned to test claims before adopting them
I Caught My AI Agent Hallucinating Revenue (And Built an Observability Layer to Stop It)
How I Evaluated an AI Model on AWS Without Writing a Single Line of Training Code
Cursor: Rules, Project Docs, and Context That Actually Help the Model
AI Hallucinations: Why Your Mock Environments Might Be Lying to You
Why AI proposal tools feel generic — and how to make them read real client briefs
Proposal: A Real Benchmark for Long-Term AI Memory Systems
I Rewrote 16 Plans From Scratch. The Code Was Fine. The Plans Were Rotting.