Everyone's using AI wrong. Including me, until last month.
We ask AI to write emails, generate reports, create content. But that's like using a supercomputer as a typewriter. The real breakthrough happened when I flipped my entire approach.
AI's superpower isn't creation. It's consumption.
The creation trap
Here's how most people use AI:
- "Write a blog post about engineering leadership"
- "Generate code for this feature"
- "Create a summary of this meeting"
Makes sense. These tasks save time. But they're thinking too small.
My Obsidian vault contains: → 3 years of daily engineering notes → 500+ meeting reflections → Thousands of fleeting observations about building software → Every book highlight and conference insight I've captured
No human could read all of this in a lifetime. AI consumes it in seconds.
The consumption breakthrough
Last month I connected my Obsidian vault to AI. The questions changed completely:
Instead of "Write me something new" I ask "What have I already discovered?"
Real examples from this week:
"What patterns emerge from my last 50 one-on-ones?" AI found that performance issues always preceded tool complaints by 2-3 weeks. I'd never connected those dots.
"How has my thinking about technical debt evolved?" Turns out I went from seeing it as "things to fix" to "information about system evolution" around March 2023. Forgotten paradigm shift.
"Find connections between Buffer's API design and my carpeta.app architecture" Surfaced 12 design decisions I'm unconsciously repeating. Some good. Some I need to rethink.
Your knowledge compounds, but only if accessible
Every meeting, every shower thought, every debugging session teaches you something. But that knowledge is worthless if you can't retrieve it.
Traditional search fails because you need to remember exact words. Your brain fails because it wasn't designed to store everything.
AI changes the retrieval game: → Query by concept, not keywords → Find patterns across years, not just documents → Connect ideas that were separated by time and context
The constraint was never writing. Humans are already good at creating when they have the right inputs.
The constraint was always consumption. Reading everything. Remembering everything. Connecting everything.
Building your consumption system
My setup is deceptively simple:
- Everything goes into Obsidian (meetings, thoughts, reflections)
- AI has access to the entire vault
- I query my past self like a research assistant
But the magic isn't in the tools. It's in the mindset shift.
Stop thinking of AI as a creator. Start thinking of it as the ultimate reader of your experience.
Every note becomes a future insight. Every reflection becomes searchable wisdom. Every random observation might be the missing piece for tomorrow's problem.
The compound effect
After two months of this approach:
→ I solve problems faster by finding similar past situations → I make better decisions by accessing forgotten context → I see patterns that were invisible when scattered across time
Your experience is your competitive advantage. But only if you can access it.
Most people are sitting on goldmines of insight, locked away in notebooks, random files, and fading memories. AI turns that locked vault into a queryable database of your own expertise.
The real revolution
We're still thinking about AI like it's 2023. Writing assistants. Code generators. Content creators.
The real revolution is AI as the reader of everything you've ever thought.
And that changes everything about how we should capture knowledge today.
Start documenting. Not for others. For your future self and the AI that will help you remember what you've forgotten you know.
This piece originally appeared in my weekly newsletter. Subscribe for insights on thinking differently about work, technology, and what's actually possible.