Home

AntFarm Systems

19 Sep 2026

Two pens, four sentences, over and over again.

I started this blog in 2013 when the long TLDs were announced. That was many years ago, and I never did much with it, since work and life took priority. I always imagined it to be a place to share my experience on building systems, the invisible infrastructure of our digital world. I liked the idea of an antfarm because it’s a massive system hidden just below the surface. Most people don’t even realize it’s there. This feels like the internet. (btw, have you tried dark mode yet?) That never took off, and I never put much effort into it. I shared the idea with peers, friends and even collaborated with a friend, a former peer and all around good person, who died at home alone of a heart attack. Life is short and precious. Take care of yourselves, people.

Today in the age of AI, creating blogs on tech content, how-tos and build guides is so easy, they have exploded in number. Claude is very busy helping people write tech content. Claude even helped me build the easter eggs and some scaffolding of this site. (seriously, try the dark mode) What appears to be happening though is that all the sites read the same, have the same content for the most part, written by the same models. Future models will train on these blogs and feed the flywheel. Will it be better? I don’t know. I do know that AI is an amazing technology, and accelerating our build and delivery pipelines faster than we’ve seen. It’s enabling people with no coding experience to build amazing tech solutions solely built on their non-tech domain expertise. Last year (this year?) The top three finishers in Anthropic’s code hackathon were NOT coders. They were people with depth in a non-tech domain.

In Stephen Brozovich’s great talk, he highlights the skills that are valuable in the agentic world. He tells us that we must move from builders to orchestrators, and gives us a list of features in a team that have high value in the agentic era. Among these are curiosity, collaborativeness, customer focus, sympathy for related domains, and a depth of understanding in first principles patterns. These strike me a deeply human traits.

As humans, we can’t write as fast as the models. We can’t research as fast as the models. We can’t even see the patterns in complex systems as fast as the models, and we definitely can’t write code as fast as the models. So if our tech skills are no longer the differentiator, what is? it must be our humanity, that connection between the human and the digital agent, and each other. Our humanity is what we need to lean into.

About a year ago I took up journaling, doodling and all things pen and paper. There’s more research now that supports the idea that writing out notes (in analog) binds our brain better. So I’ve leaned into this. I take meeting notes on paper, todos, and more. I’ve learned a bit more of how my focus works (poorly, but I’m working on it) and how I think. I’ve also started to develop stronger opinions on pens and stationery, with such hot takes as ‘moleskine is mid-tier on it’s best day’ (COME AT ME BRUH!) and ‘Zento is smooooth.

I’ve written so much I emptied a pen.

Two uniball Zento UBR-Z-07 refills lying on a blank grid-ruled page,  one new and full, one empty with only a trace of ink left above the tip.
UBR-Z-07, one new with the wax cap on the tip, one empty.Google Pixel 10 Pro · f/1.7 · 1/120 · ISO 250 · September 18, 2026© Luis E. Cerezo

Last week we took some time off. We rented a nice place facing the meadow and Mt Crested Butte. It fo’ shizzle did mostly drizzle, but it let me slow down and think on paper. I slowed down enough to do a writing meditation. It started with an interview from Rick Rubin, who was talking about how anyone can create, and nobody is special. I was sharing the video with my wife, and he spoke about his metta meditation practice and repeated the mantras. It struck me that I can do this as a writing meditation, the speaking outloud practice always made me think of Stuart Smalley, and the moment gets ruined. So I sat there, writing slowly.

An open notebook spread filled edge to edge with the same four sentences in block capitals — MAY I BE HAPPY. MAY I BE WELL. MAY I BE PEACEFUL AND AT EASE. MAY I BE FILLED WITH LOVE AND KINDNESS — repeated down every line.
Google Pixel 10 Pro · f/1.7 · 1/120 · ISO 209 · September 18, 2026© Luis E. Cerezo

very slowly. I’m not sure if this is human, but I’m trying to be more human every day, and this is part of the journey. I do believe this will make me a better technologist in the world of AI, and possibly a better person. maybe.

btw- the most human thing about this post is I wrote A LOT OF FUCKING PAGES of A5 meditations, that I mis-heard. I like my version fine.

Published on September 19, 2026


Luis at 00:00

scribble