Dossier · 01 // LAST UPDATED MAY 2026
What I believe.
The long form of how I think about technology, abundance and the next decade. Updated whenever I learn something I'd want my younger self to know.
This page is the place I keep my actual operating system. Not the marketing version, not the LinkedIn version — the version I'd defend with friends after the second coffee.
It's a working draft. I update it when I notice I've been arguing something for long enough that it deserves to be on the wall.
Effective accelerationism, the kind that ships
I align with e/acc in the simplest sense: I think AI's potential to shake up inefficient systems, automate the mundane, and free humans for meaningful work is one of the best stories of our time, and I want to be inside it, not commenting on it.
The version of e/acc I care about is the one with hands. It picks a problem, ships a tool, watches a human get an hour back, and then asks who else needs it.
Abundance is built, not waited for
The default state of the world is friction. Forms that don't connect. Compliance rules nobody reads. Tools that cost more than they earn. Talent stuck because the right tool wasn't translated.
Most of these are solvable, and the cost of solving them keeps falling. Abundance — actual abundance, not the marketing word — is what happens when you remove enough friction in a row that ordinary people can do extraordinary things.
The right tool, given to the right person, is the highest leverage asset on Earth.
Tools over takes
When I see something inefficient I try to ship a better thing instead of a longer thread about it. Threads have their place; tools change the conditions.
Human-AI collaboration is the headline
Agents amplify operators. The future I bet on has more capable humans, not fewer, and the ladder up is built deliberately — by people choosing to share the keys instead of hoarding them.
The way I work right now is honest about this. I keep Claude, ChatGPT, and Gemini open in parallel; I write code with a harness that argues back; I let the tools propose and I make the calls.
From the global south, on purpose
I build from Barranquilla, on the Caribbean coast of Colombia, because the gap is largest here and the leverage is highest. The default tools were not designed for the people I want to serve — first-time founders, small clinics, family businesses, regional governments — so we get to design them.
What I won't do
- I won't build attention traps. The world has enough.
- I won't ship tools that make their users feel smaller or stupider than they are.
- I won't optimize for metrics that don't survive a conversation with the humans on the other side of the screen.
- I won't pretend AI is magic. It's the best leverage I've seen, but it's a tool — and tools are only as good as the people holding them.
Things I'm probably wrong about
Honest belief comes with a price tag attached: I keep a list of the things I've changed my mind about, so I remember that the next shift is probably already in the air. I'll add it here when it's long enough to be useful and short enough to be readable.
That's the working draft. If you disagreed with any of it, write to me — that's how this page gets better.