← 홈 / 2026-Q3
기간

2026-Q3

에세이 18편
최신순
2026-09-23

AI가 제 시스템에 대해 저에게 무언가를 가르쳐 준 날

아무것도 생산하지 않지만 이후 생산될 모든 것을 규율하는 문서는 어디에 배치해야 할까요? 흔히 쓰는 기준인 변화 속도만으로는 부족합니다. 그 자리를 기능 기준이 대신합니다. 이 파일은 결과를 생산하는가, 지식을 보존하는가, 상태를 기술하는가, 아니면 앞으로의 작업을 규율하는가. 아키텍처가 예상하지 못한 어떤 대상에도 쓸 수 있는 기준입니다.

2026-09-21

"Contexte"에서 제가 정말로 발명한 것이 무엇인지 찾아보았습니다

몇 달 동안 만든 방법론이 이미 다른 이름으로 존재하는지 어떻게 알 수 있을까요? 방법은 몇 가지 동작으로 요약됩니다. 아이디어를 메커니즘으로 분해하고, 어휘가 아니라 기능을 찾고, 자기 분야와 영어권을 벗어나고, 새로운 발견이 무엇을 바꿔야 의미가 있는지 미리 정해 두는 것입니다.

2026-09-20

기둥 카드: 계속 생각하기 위해 기억을 살려 두기

AI로 빠르게 생산하다 보면 생각하는 시간이 밀려납니다. 이 시간을 지키는 구체적인 장치가 있습니다. 제한된 기둥 카드의 핵심 묶음을 간격 반복으로 살려 두고, 테이블 위에 펼쳐 새 생각을 자신의 신념에 맞세워 봅니다. 카드에 들일 기준, 복습 주기, 그리고 시작하는 방법까지 정리합니다.

2026-09-20

ChatGPT 안에서 제 Contexte를 직접 돌리기

어딘가에 켜 둔 Mac 없이 Contexte로 일하는 방법입니다. 구성은 세 가지 요소로 충분합니다. ChatGPT가 추론하고, 프로젝트 지침이 엔진을 정의하고, Google Sheets가 기억합니다. 실제로 전시회장에서 걸으며 대화 내용을 받아쓰고, 보존하고, 인제스트한 다음, 자리를 뜨기 전에 아직 던지지 않은 질문이 무엇인지 시스템에 물어볼 수 있습니다.

2026-09-13

엔터프라이즈 SaaS가 달라지고 있습니다: 대기업이 통제권을 되찾으려 할 때

한 대기업 고객이 데이터를 자사 Azure에, 인증은 자사 IAM으로, 키와 LLM 계약도 자사 것으로 쓰고 싶다며 계약을 멈춥니다. 이 글은 그런 요청을 가려내는 기준을 제시합니다. 무엇이 실제로 우리의 가치를 담고 있는지, 무엇이 그 가치를 살 수 있게 만들 뿐인지, 그리고 고객의 통제권과 제품을 진화시키는 우리의 능력 사이에 경계를 어디에 그어야 하는지입니다.

2026-09-12

"Persist that": getting a piece of information out of ChatGPT and into permanent storage

Five days later, you rarely find the one paragraph that mattered in a long ChatGPT conversation. This is the story of how I got a precise fragment out of the flow on demand, using nothing but a custom instruction and a Google Sheets outbox — a blocked API, a workaround that failed, then a solution that went from 50 seconds to two, with a systematic check after every write.

2026-09-07

How I write my articles by talking to my phone

Your best article ideas die between the moment they turn up and the moment you sit down to write. Here is the setup that catches them while they are still warm: the voice to empty your head without putting anything in order, a session source to keep the raw material raw, AI to surface the gaps in ten seconds rather than in two days.

2026-09-04

The deliverable is no longer the asset

Every new format feels like starting over: you reopen the last document, copy, rephrase, and two versions slowly drift apart. The cause isn't a missing tool, it's treating a deliverable as your source of truth. How to install a reusable layer between your sources and your documents, and what AI gains from it.

2026-09-02

AI doesn't replace judgement, it speeds up how an idea matures

You can feel that a piece of writing or a decision isn't right yet, without being able to say why. That's where AI earns its place: it turns a vague unease into concrete handles — a blind spot, a shaky assumption, an objection you never dealt with. Which questions get you pushback rather than output, and why this makes human judgement more important, not less.

2026-08-11

I wrote a product's ontology. Three times, I thought I was done.

Your team rebuilds the same product model for every question, every bug, every tender. Here is how to write it down once: several dozen objects, a few hundred rules, under six hours. And above all how to know you have finished — the three stopping criteria that look like proof of completeness and prove nothing, plus the one that holds.

2026-08-03

The PO is not a profession, it's a function — and AI changes nothing about that, it just speeds it up

Hiring a PO to hold the backlog and sit between the business and the developers? You are optimising a circuit that is already obsolete. This piece takes apart the split between the people who do the thinking about the product and the people who execute it: why it degrades quality, why a backlog cannot support a profession, and why AI removes the last excuse for living with it.

2026-08-01

PM, developers and AI: roles blur, responsibilities remain

AI raises the capacity to build, not the capacity to decide. Developers move up towards product; PMs move down towards building. The shift blurs roles without dissolving responsibilities. A full pass over what AI makes ordinary in product management, what stays hard, and a catalogue of emerging roles with the real risk attached to each.

2026-07-28

Four days of vibe coding in the shoes of a rusty PM

Four days building a real product with Claude Code — 180 commits, starting from nothing. What it shows: the speed that goes to your head, the traps the AI creates on its own, what tests change when you haven't written the code yourself, and how far a PM can go without stepping outside their remit. A field report, not a doctrine.

2026-07-23

AI shouldn't just let us produce ten times more. It should also force us to understand ten times better what we produce

Producing code faster is worth nothing if nobody understands what got built. AI automates precisely the tasks that used to train junior developers, and an error that lands too late no longer teaches anything once the system has grown too big to analyse. A few concrete anchors for staying in control: clear contracts between APIs, tests tied to the functional need, rebuildability as a quality bar rather than raw delivery speed.

2026-07-20

Session sources: what your documents miss

Your documents never say everything. The fine-grained understanding of what happens on the ground stays in the PM's head. Session sources open a channel for pouring that informal knowledge into an AI Context — a hallway remark, a nuance picked up from a customer, a tension just starting to show — and turn dozens of micro-signals into notes, assumptions and better-framed decisions.

2026-07-08

Information systems are not going away. Their shape will change.

Your tools file information away neatly, but they never build up what that information teaches you. This piece explains why the future of the information system is not a better AI-assisted form, but a shift towards `Contexts`: portable, governed units of memory that connect sources, decisions and reasoning instead of merely storing them.

2026-07-06

The product question should start from the source code

Do you lose half an hour in a test environment just to answer a product question on Slack? That hidden cost disappears once AI can query the source code, the real truth about how the product behaves. Answers grounded in evidence — functions, tests, commits — rather than in documentation that has fallen behind. Provided the code speaks the language of the business.

2026-07-01

Under the bonnet of my context engine: how an AI remembers a mission

Do you lose the thread the moment you reopen a conversation with an AI about something that has been running for weeks? This article lifts the lid on a structured memory system: the difference between the goal and the attention of the moment, the directives, the offline queue, the three memories (log, synthesis, checkpoints) and the cycle of a session. Enough to see how an AI can genuinely remember a mission.

최신 글