I don’t have a grand AI strategy. I don’t follow launch threads or obsess over benchmarks.
Instead, I pay attention. To friction. To repetition. To the quiet inefficiencies in my day.
And when I notice them, I ask a simple question:
“Is there a way AI can make this more effective, or more insightful?”
That’s my strategy: discovery through reflection.
No hype. No dashboards. Just one tool at a time, layered together like LEGO bricks — until they start to feel like an operating system for how I think and work.
Here are the tools I’m using today.
I use Superhuman as my main email client, not just because it’s fast, but because it plays nicely with my AI workflow. One of my favorite moves: copy an email (Cmd+C) and paste it into ChatGPT. This simple action powers one of my most valuable routines — feeding newsletters and updates into a custom GPT I call Executive Summary Pro.
(If you’re curious, here’s my referral link — it helps support this work.)
Link: Executive Summary Pro GPT →
This is a custom GPT I built to act as a research companion. It reads long-form content (emails, Substacks, academic papers, whitepapers) and distills them into structured, executive-level summaries — without losing nuance.
Every day, I paste in dense materials: market analysis, research memos, deep Substack posts. The GPT acts like a smart analyst: extracting insights, surfacing tensions, preserving tone. I can scan what matters and dive deeper only when I want to.
It’s made my reading stack dramatically more actionable — and less overwhelming.
While I use Vivino to scout wines online, this GPT has become a go-to for deeper questions:
What pairs well with what I’m cooking tonight?
What’s the best bottle in my cellar for a specific mood or dish?
How should I serve, decant, or age this quirky bottle I picked up last year?
What’s the story behind this region or vintage?
It turns wine into a richer experience — part ritual, part learning, part pleasure.
This one’s more niche — but incredibly important for how I think about markets.
I’ve followed John P. Hussman, Ph.D. for years. He’s a valuation-driven investor and economist, with a deep archive of commentaries that blend historical insight, macro discipline, and behavioral awareness.
So I built a GPT that mimics his worldview. It incorporates:
Valuation metrics like MarketCap/GVA and margin-adjusted P/E
Cyclical analysis of market internals (breadth, credit, sentiment)
Skepticism toward speculative manias, but also awareness of past blindspots
I don’t use it for signals or trade ideas. I use it as a bias-check. A lens. A historian who speaks in models and probability. It helps me step outside the present and think in cycles — not headlines.
This post itself was outlined, structured, and refined with a custom GPT I call my Writing Partner.
It’s not a ghostwriter. It’s not a prompt machine.
It’s a calm, reflective collaborator — trained on the themes I care about:
Platforms. Long-term thinking. Brazil. Private markets. Memory. Reinforcement.
Most days I start with a raw idea or phrase — “platforms that host ourselves”, “memory as infrastructure” — and build from there. The GPT helps me test phrasing, suggest structures, rewrite in different tones. It spots recurring themes in my past writing. It proposes threads from essays and essays from threads.
Over time, it’s become a kind of thinking mirror. A reinforcement loop.
Not what to say — but how to say it, and more importantly, how to see it more clearly.
This is the key insight: AI is no longer just a productivity tool. It’s a system. One that you build gradually, like scaffolding around your own way of thinking.
You start with simple tools.
Then you link them.
Then they start linking you — across habits, across decisions, across time.
Each GPT becomes a module: one for reading, one for investing, one for writing, one for pleasure. Together, they form a kind of cognitive OS — modular, personalized, extendable.
And the difference is compounding.
Using AI in this way is like switching from a typewriter to a computer.
One is static, linear, forgetful. The other is dynamic, editable, connected.
With a computer, you can cut, paste, search, remix — and build on everything you’ve done before.
That’s what this moment feels like.
Not science fiction.
Not some far-off dream.
But now — quietly, practically — changing the way we think.
— Art
Writing to think. Sharing to learn.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.