Skip to content

Contributor(s)

Fichero is 100% AI coded, under the creative direction of its contributors. Right now, that’s just me, Daniel Tubb.

Daniel Tubb

Daniel Tubb

I am an anthropologist and ethnographer. My work is about Colombia and New Brunswick. Since 2016, I have worked in the Department of Anthropology at the University of New Brunswick in Fredericton, Canada. I am the author of Shifting Livelihoods: Gold Mining and Subsistence in the Chocó, Colombia (University of Washington Press, 2020), about the impact of gold mining on communities in Colombia’s Chocó department. In addition, I co-edited Letters from the Future: How New Brunswickers Confronted Climate Change and Redefined Progress (Chapel Street Editions, 2020), a book of speculative non-fiction that offers a unique New Brunswick perspective on addressing climate change. Currently, my research focuses on agrarian change and agrarian histories of lowland crops in the Colombian Pacific, among other things.

From this work, I have been increasingly drawn into public anthropology and history work, working with archives. For that, I wanted a better way to read and work with archives, and find the things I am interested in. Fichero is the result of a three-year effort, which began in the summer of 2023 as part of a SSHRC Connection Grant on Who Owns the Archives. It has emerged out of ongoing conversations with Ann Farnsworth-Alvear (History, UPenn), Andrew Janco, and others in the Chocó, Colombia. Many good ideas have emerged from those conversations, but the misguided endeavour to attempt to make a fully featured Mac app is my own. Andy has provided lots of crucial pointers and advice, because unlike me, he is a server-side programmer and technical lead.

I am not. The last time I took a computer programming course was in high school, in the late 1990s, when I wrote a windowing system for a drawing application in Turbo Pascal. At university, I pursued Spanish literature as a joint major, rather than computer science. However, I have long used code in my work. As an undergraduate, I taught myself Spanish conjugation by writing an app in REALbasic to quiz me on conjugations. By having to code a Bescherelle, I learned how to conjugate. I have long used snippets of code in my work, in other ways.

As well as a writer and anthropologist, I have been a long-time Mac user. I think of computers as tools to do work. But, there are many other tools. I love pencils and pens. I am enamoured by what my colleague Noah Pleshet calls undigital approaches. I believe the method is the message: how we do our work (our research, and our writing) is crucial to the work we ultimately produce.

I love AI for fixing my typos, but like many I have become alarmed by the way it is used in the classroom to prevent critical thinking. Yet I am also convinced that AI tools can allow us to do better research. Research that would otherwise have been impossible.

Why then use AI to write a Mac app? It’s a tool that I began to vibe code as a Python CLI app in 2024 while on sabbatical, and for a while it was a Toga/Briefcase Mac, Windows, and Android app. But, in December 2025, I decided to start over from scratch, to make a SwiftUI app for Mac, iOS, and iPad, with a Python server on the backend. Without AI, Fichero would not exist. So, Fichero has at once been a project that is vibe coded, and increasingly, vibe directed. If my purpose was to write the best Mac app, would this be the way to proceed? I have no idea. But, my purpose is to make a good enough tool to work with archives and research materials. Fichero does that.

Fichero is the tool I want for reading old archives and historical materials, as well as my own research notes and field notes. These are about a million words. The former will allow me to work on a large synthesis project on the origins of agricultural systems in Colombia, and the latter will allow me to finish a project on makeshift writing.

So, Fichero is at once a tool to help me do this work and a tool for others, I hope.

What is it? It is a Mac shoebox app, inspired by and in many ways complementary to other existing apps — some of which I imagine by now also have AI features: like Yojimbo, DEVONthink, and EagleFiler — I use the latter two — alongside citation-management software like Bookends, Zotero, and EndNote; and tools to think with and write notes in, like Obsidian and Tinderbox. It also draws inspiration from photo editors like GraphicConverter, Photos, and Lightroom. Fichero then is an app to store, organize, keep track of, cite, find, take notes on, navigate, and edit images of archives to make them easier to read.

I hope Fichero will make working with many kinds of documents easier — large archival collections, but also interviews and audio recordings — with a document preview with tools that make reading old documents easier (a magnification bar at the bottom). At the same time, because it is built on open-source software, it is fast (I hope), has semantic search, can run AI workflows that do useful things like extract text, and keeps track of all of the data produced by such AI workflows.

If you’re a prospective graduate student interested in studying with me, I invite you to read my letter to students and send a brief inquiry to dtubb@unb.ca. If you’re interested in learning more about my research and publications, feel free to explore my website and the Human Environments Workshop, which I co-direct.

Contact