-
What to Decide Before Running a Software Factory
Define success before agents start optimizing for it
Software Factory를 돌리기 전에 정해야 할 것들
Define success before agents start optimizing for it
-
An Agent Application Evaluation Playbook
Building useful evaluations from the first working prototype
에이전트 애플리케이션 평가 플레이북
Building useful evaluations from the first working prototype
-
As Agents Get Stronger, Where Will Business Moats Move?
Business moats are shifting from applications to data
에이전트가 강해질수록, 서비스의 해자는 어디로 옮겨갈까
Business moats are shifting from applications to data
-
My Thoughts on AWS AI-DLC — The Direction I Believe Agentic Development Should Take
Where AI-DLC diverges from my view of agentic development
AWS AI-DLC에 대한 내 생각 — 내가 믿는 에이전틱 개발의 방향
Where AI-DLC diverges from my view of agentic development
-
As Agents Automate Development, Where Do Developers Go Next?
Exploring FDE as an evolution of traditional developers
Agent로 개발을 자동화한 다음, 개발자는 어디로 갈까
Exploring FDE as an evolution of traditional developers
-
Why Did Some Teams Get Up to 10x Faster with the Same AI Tools? — Five Habits of Frontier Development
Five habits behind Amazon's frontier development teams
같은 AI 도구를 썼는데 왜 어떤 팀은 10배까지 빨라졌을까 — Frontier Development의 다섯 습관
Five habits behind Amazon's frontier development teams
-
When Should AI Adoption Be Measured by Business Metrics? — Revisiting AHEAD and LEVER 2/2
Measuring harness learning before business value
AI 도입 성과는 언제 비즈니스 지표로 봐야 할까 — AHEAD·LEVER 다시 보기 2/2
Measuring harness learning before business value
-
Why AI Code Review Feels Harder — Context Familiarity and Cognitive Debt
Why compressed understanding makes review the bottleneck
AI 코드 리뷰가 버거운 이유 — 컨텍스트에 익숙해질 시간과 인지부채
Why compressed understanding makes review the bottleneck
-
Did AI Coding Tools Actually Cut Development Cost? — Understanding CTS-SW
Measuring AI by software delivery cost
AI 코딩 도구가 정말 개발 비용을 줄였을까 — CTS-SW 이해하기
Measuring AI by software delivery cost
-
Why Separate PRDs, ADRs, and Code? — Reading One Abstraction Level at a Time
Keep intent independent of implementation details
PRD·ADR·코드를 왜 나눠야 할까 — 추상화 계층 하나만 읽고 판단하기
Keep intent independent of implementation details