The FOSS Academic Lifestyle Dream 2025 Review

Robert W. Gehl, 21 December 2025

Happy solstice! The shortest day of the year is a fine day to think back on another time 'round the sun. What's gone on with me and the FOSS Academic dream this past year?

A new book

Well, the big thing for me was my fourth book, Move Slowly and Build Bridges: Mastodon, the Fediverse, and the Struggle for Democratic Social Media was published. Learn more about it at its website. Move Slowly was the culmination of years of effort, and frankly, I'm exhausted!

Since this is a blog about the FOSS Academic Lifestyle Dream, Move Slowly was done entirely with FOSS tools, with note-taking in Zotero and Zettlr, interviews done with Nextcloud Talk, and files handled by Nextcloud. At some point, however, I had to turn files over to the press, which uses Microsoft Office tools for copyediting. Often that introduces compatibility issues with LibreOffice, but I still found a way to get things done with FOSS. It is possible!

In addition to Move Slowly, I published several articles this year using a similar workflow (and, unsurprisingly, on a similar topic). Along with Nate Tkazc, I published "After Twitter" in Social Media + Society. And I published a paper on fediverse blocklists in Communication, Capitalism, and Critique, a journal I've long wanted to contribute to. Both of these are open access.

So, more success using FOSS to do my academic work! Yay!

You Got Markdown in My Python!

Speaking of Zettlr, this year has really cemented the place of Markdown in what I do. It's so simple to fire up a text editor (such as Kate, which I'm using now) or Zettlr to jot down ideas or even make to-do lists.

But now I'm doing Markdown for more than notes. I'm making presentations with it. Marp, a tool to convert Markdown files into presentation slides, has been a revelation. I started using it late in 2024 and have since given many presentations with it, both at conferences and in my teaching.

Being based on Markdown, Marp makes creating slideshows so much easier. I write out notes and boom! I can make a presentation. I have found that I can turn one around in a few hours before teaching a class -- this is so much faster than the way I used to do it with LibreOffice Impress.

I built a whole custom workflow that not only automates the generation of my presentation scripts, but also pushes both my slides and their notes to my website. This makes my slideshows much more accessible: all the images have alt text and my notes are readily readable by a screenreader. To see it in action, here's a book talk I gave at LSE and here are the corresponding notes.

I start all of this with Zettlr, using it to build out the notes, then using Marp to generate the slides, and then using Python to convert those slides into the various HTML files I push to my website.

You Got Python in My Markdown!

And that brings me to another major part of my 2025. Those custom workflows are thanks to my no-doubt-clumsy-yet-effective abilities with Python, a language I've spent a lot of time learning this year.

A big push came when I adopted a Codeberg-hosted Python project to post little snippets of my Mastodon book to the fediverse. Now, every day at 9:00 am my time, a wee bit of my book is made public on the fediverse.

While I really didn't know what I was doing, I did a lot of searches for solutions to little problems, like how to randomize lines from a .CSV file, or how to trim strings so that they stay under a character limit.

Solving those problems gave me a taste for what Python can do with text. And once I get a taste... it led to a bigger hunger.

I've been running this blog on Jekyll since the outset, and while it's worked, I've found Jekyll to be brittle, especially when I install its required Gems on a new computer. I started to wonder if I could just reimplement this blog in Python, and after a few days of coding, I was able to achieve it.

Now I find myself wondering what else I can do with Python and text files, such as Markdown or CSV. I have a project in mind for Zotero Reports -- more on that in 2026, maybe.

GenAI, GenAI, Who Can I Turn To?

While I'm living the FOSS Academic Lifestyle Dream, it's not all gumdrops and pythons. 2025 has also been marked by a growing sense of dread due to generative AI.

While I am an academic by profession, my core identity is that I am a writer. Writing is a beautiful, horrible mess. It is such a powerful way to work though thoughts, a way to structure thoughts that would otherwise flit and flow like a school of fish, to make ideas alien but not alienating.

I believe strongly in writing-as-thinking, and not only do I put it into practice in my life, I also put at the center of my teaching. All of my classes have writing (and the related practice of reading) at the center.

And here comes generative AI to utterly decimate that.

I won't spend too much time despairing about generative AI here. Instead, I will focus on how generative AI is an exploitaiton of the FOSS Academic Lifestyle Dream.

FOSS stands for Free and Open Source, a means by which people can build software together. The "open" bit has splashed onto the academic career -- people now expect a degree of openness about research. This blog was started precisely for this reason: I was taking a chance, trying to put my work out there a bit more, make it more open, show my process. I hadn't really done that before.

Now my opeenness has been exploited by generative AI.

As I write, a bunch of scrapers are taking what I compose to train models. I know this because traffic to my websites is clearly coming from bots. I get bursts of traffic if I share a new blog post, and then a steady stream of visits from cloud centers around the world.

My books have been pirated by Anthropic to train their models.

My academic articles have been sold off to the highest bidder.

So basically, when Gemini and Claude and ChatGPT pop up and tell my students, "hey, you don't have to read this -- our machine will summarize it," or "hey, you don't need to write! Let the machine do it!" -- when those machines short-circuit my students' thinking, my work is part of the machine. It's utterly demoralizing.

I am wrestling with what to do about this. I may just put this blog, and my other websites, behind a login. I hate to do it to my fellow humans, but I think they'd understand.

Plus ça change, plus c'est la même chose

While generative AI feels like a grenade rolled into the dugout, looking back, I also see that so much of my workflow -- and the broader FALD -- hasn't changed: I'm still using:

• Wireguard to link all my devices
• Jellyfin to stream media to all those devices
• KDE as my desktop of choice
• Ubuntu servers
• Nextcloud for managing files
• Syncthing for syncing files
• LibreWolf for browsing
• LibreOffice for writing
• Zettlr for note-taking
• And the mighty, mighty Zotero for managing bibliographies

Looking back on this, it's actually quite astounding. While so many tech companies have jammed generative AI into their software -- with an egregious example being Mozilla -- what's notable among the tools I use is how little of it I'm seeing. (And what little I see, such as Nextcloud's AI agent, I just don't enable.)

And so even with the terror of the transformers, I have much to be thankful for in terms of FOSS. It's been... boring, in a way. But I kinda like it that way.

Post Tags

• Goal 1
• Goal 2
• workflow