What are the crucial points in this article or video that make it iconic, ideas I want to remember for the rest of my life? 1. 2. 3.
How was this video or article relevant to my current life? Did it answer a specific question, enlighten me on a topic, etc.
this is extremely directly relevant — I’m already running the same stack (Obsidian + Claude Code) and this vault is the proof of concept. greg’s approach of using claude.md files as compounding team knowledge is what this CLAUDE.md already does. the interesting question is whether I’m getting greg’s intended outcome: a vault that genuinely reduces cognitive overhead, or am I just adding complexity? the “AI as life operating system” framing is aspirational but the test is whether it saves time in practice.
- vault as context for AI — the key principle: the vault works as an AI extension when it has clean, queryable context. my vault is usable but not clean. the order of operations matters.
- CLAUDE.md as compounding knowledge — every session should leave the vault slightly smarter about my preferences, ongoing projects, and constraints. this is already happening via the CLAUDE.md update discipline.
- skills as repeatable workflows — the skill system in this vault is exactly greg’s pattern: encoding repeatable processes into prompts that can be invoked by name. the gap is having more skills built and tested.
- life OS, not just notes — greg uses the vault not just for capture but for decision-support: morning startup, daily reviews, project planning. this is already the direction the skills are pushing toward.
the most useful delta from watching this is recognising that I’ve been building the right system but not fully using it. the vault has morning-startup, daily-review, weekly-shutdown, and shutdown skills — but I use them inconsistently. the compounding only works if the input (daily use) is consistent. the gap isn’t the architecture, it’s the habit.
useful as validation for the current vault direction, and as a calibration check for whether greg’s implementation gives me any new structural ideas. most of what he describes is already here. ★★★☆☆
- am I using the vault consistently enough for the AI integration to compound meaningfully — or is the inconsistency the actual problem to solve?
- what’s one vault skill I should build or refine in the next month that would reduce daily cognitive overhead?
- how does greg handle knowledge decay in the vault — old notes that are no longer accurate or relevant?
- is there anything in greg’s setup that I’m not doing that would make a meaningful difference to output quality?
- greg isenberg – i got a private lesson on claude cowork & claude code — companion video in this vault
- any model. any app. build your ai os to work everywhere. — parallel AI OS setup video
- CLAUDE.md (this file is the live implementation of what greg describes)
- daily vault close — before ending any work session, make one update to the vault that reflects what was decided or learned. even one sentence. this is the compounding mechanism.
- skills-first for recurring tasks — before doing any recurring content or admin task from scratch, check if there’s a vault skill for it first. if not, write one after doing it the first time.
- check the vault inbox/scratch before starting the day — has anything accumulated that needs processing?
- audit which skills are actively used vs which were built but never triggered — delete or archive unused ones
- build one new skill this month for a recurring task that currently gets done from scratch each time
- check: is the CLAUDE.md current? does it reflect the current projects, tools, and constraints accurately?
- use the morning-startup skill for 5 consecutive days and note whether it meaningfully changes the day’s quality