What the nerds do on weekends, everyone else will do on weekdays

Chris Dixon, an investor at Andreessen Horowitz, wrote an idea back in 2010 that I come back to every so often: what the most curious people do on weekends is what everyone will be doing on weekdays ten years later.

I've watched it happen several times. The internet started as a hobby for university students. Bitcoin spent years as a topic for odd forums, and when I bought in 2022 almost nobody believed in it. Language models were researchers' toys until suddenly they were on everyone's phone.

The same thing is happening today with AI agents and MCP. A few people are building them on a Saturday night, and in a few years they'll be the normal way to work.

My job is watching the weekend

Since September 14 I've been an AI Analyst at the Municipality of Vicente López, in Buenos Aires, on the AI and Innovation team.

On our team we're constantly testing whatever just showed up: new models, new ways to automate, tools that didn't exist six months ago. What others look at out of curiosity on a Sunday is part of our Monday work.

But the quote has a catch

The nerds' weekend is also full of things that never made it to the weekday. For every internet there were a hundred fads that faded on their own. Exploring without judgment is just scattered focus dressed up as creativity.

In a municipality, the filter is a need. There's health, traffic, public works, environment and resident services, each area with its own urgencies. On the other side there are no customers who chose your product: there are residents who need to get something done and have no reason to know how government works inside.

Watching the weekend is useful. What decides what makes it to the weekday is a real need.

That's why the way we work is simple: start small, test fast and cheap, and measure whether it actually solves something. What works grows. What doesn't gets dropped without drama and leaves a lesson for next time.

Government doesn't have to arrive last

The usual assumption is that the public sector adopts technology once it's already old. It doesn't have to be that way. If it waits for something to become "weekday", it arrives late and pays more. If it dares to watch the weekend with focus, it arrives on time. And when something works in a municipality, it changes the day for a lot of people.

What's next

What's internal stays internal: you won't see data or project details here. What I will share is what I learn watching that weekend: which technology is worth exploring and how that gets decided.

My bet: AI agents are weekend material today. In a few years they'll be the normal way to get a government errand done. The municipalities that start testing them now will be ready when Monday comes.