the economy is about to grow by trillions of dollars. none of it will reach your bank account.
citrini calls it "ghost GDP" — output that shows up in the national accounts but never circulates through the real economy. productivity is skyrocketing. companies are posting record efficiency. and the people who used to do that work? they're updating their resumes.
this isn't a doomer take. this is math. when agents can transact, build, research, and execute on behalf of companies — the question isn't whether the economy grows. the question is whether YOU grow with it.
so where does that leave us? what do you actually do when the tools are smarter than the average worker and getting smarter every quarter?
i've spent the last year rebuilding how i work from the ground up. not because i wanted to — because i had to. here's what i've learned.
1. build a system that thinks with you
it doesn't matter if you're a type-A planner or a chaotic creative. what matters is that you document.
not for organization's sake. for *compounding's* sake.
i set up a system on obsidian, claude code, and custom commands that lets me do something most people can't: trace how my ideas evolve over time. i type `/ideas` and it pulls patterns from everything i've written in the last 60 days. i type `/connect` and it finds relationships between two topics i never would have linked myself. i type `/trace` and it shows me how my thinking on a subject has shifted over months.
this isn't about having the most information. everyone has access to the same information now.
**the way to stand out is by seeing how one idea impacts another — how two completely unrelated concepts might actually be connected.**
that's a human skill. that's taste. that's the thing agents can't do yet — and it's the thing that makes your work irreplaceable.
practical steps:
pick a note-taking system and commit to it. obsidian, notion, even apple notes. the tool matters less than the habit.
document what you learn, not just what you consume. write down your reactions, your questions, your half-formed thoughts.
set up automations for the parts you dread. if you're doing the same task every day, that's a system failure, not a discipline problem. i automated my daily review, my content pipeline, my research synthesis — and it freed up hours i now spend on actual thinking.
internetvin's podcast on the obsidian + claude code setup is a great starting point if you want to go deep on this.
2. understand the layer beneath the tool
everyone is talking about claude code, cursor, chatgpt, the latest agent framework. fine. use them.
but if all you know is how to prompt, you're a tourist.
the people who will thrive are the ones who understand *why* the tools work — not just *that* they work. what's an API? what's a context window? what does it mean when an agent "calls a tool"? how does retrieval actually function under the hood?
you don't need to become an engineer. you need to develop technical taste. the same way a great film producer doesn't operate the camera but understands lenses, lighting, and what makes a shot work — you need to understand the architecture well enough to direct it.
this is what separates someone who uses AI from someone who *builds with* AI.
practical steps:
read the documentation, not just the tutorials. anthropic's docs, openai's docs — they're written in plain english and they'll give you intuition no youtube video can.
build one small thing end to end. a simple automation. a custom command. a workflow that connects two tools. the act of building teaches you more than a hundred hours of watching someone else build.
when something breaks, don't just re-prompt. figure out *why* it broke. that's where the real learning lives.
3. train the muscle, not the tool
the tools will change. they always do.
six months from now there will be a new model, a new framework, a new "this changes everything" launch. if your entire skill is knowing how to use one specific tool, you're on a treadmill.
but muscle memory transfers.
the more you build, the faster you pattern-match. the more you debug, the sharper your instinct for what went wrong. the more you ship, the better you get at scoping what's actually possible. these are physical skills disguised as technical ones. they compound the same way gym reps do.
without this, you hit what i call *states of entropy* — where every new tool feels like starting over, where you're always a beginner, where the pace of change outstrips your ability to adapt. that's an exhausting way to live.
the antidote is reps. not mastery of one tool. reps across many.
practical steps:
commit to building something every week. doesn't have to be big. a script, an automation, a prototype. volume beats perfection.
when a new tool drops, don't just read about it. open it. break it. make something ugly with it. you'll learn more in 30 minutes of fumbling than in 3 hours of tutorials.
keep a "build log" — a running list of what you made, what worked, what didn't. this becomes your personal curriculum. over time, you'll see your own acceleration.
4. build your hedge in the physical world
here's the part no one in tech is talking about.
as everything accelerates toward the digital — as agents handle more knowledge work, as software eats more jobs, as screens mediate more of life — there will be a massive premium on things that are physical, human, and irreplaceable.
this is my thesis: **the more digital the world becomes, the more valuable the physical world gets.**
elevated experiences. wellness. physical retail that feels like something. spaces that make you feel alive. food made by someone who cares. communities you can touch.
these aren't nostalgic luxuries. they're going to be the growth sectors of the next decade. because when AI can write, code, analyze, and create — the things it *can't* do become the scarce resource. and scarcity is where value lives.
so while you're building your digital skills, don't forget to build something in the real world too. a practice. a space. a product. a community. something that requires your physical presence and human judgment in a way no agent can replicate.
practical steps:
identify one area of your life where you feel the "digital fatigue" — that's a signal, not just a feeling. it's pointing you toward where physical value is underserved.
look at the businesses around you that are thriving despite (or because of) the digital shift. wellness studios, specialty retail, experience-based dining, community spaces. study what they're doing.
start small. host a dinner. teach a workshop. make something with your hands. the gap between digital and physical is where the next wave of entrepreneurs will build.
the real point
this piece isn't really about tools. it's about who you become when the tools can do everything.
ghost GDP is real. the economy will grow and leave people behind. but not everyone. not the ones who built systems that compound their thinking. not the ones who understood the infrastructure well enough to direct it. not the ones who trained their instincts through relentless reps. and definitely not the ones who bet on the thing AI can't touch — the physical, the human, the irreplaceable.
the agents are coming. they're already here.
the question isn't whether they'll change how you work. they will. the question is whether you'll be the one directing them — or the one they replace.
*the way to stand out is not by having the most information, but by seeing what no one else sees.*
build the system. learn the layer beneath. train the muscle. hedge in the real world.
that's how you work in the agentic world.