If you think Linux dating all the way back to 1991 is old, wait'll you hear about AT&T UNIX. This talk is a whirlwind tour of ancient UNIX quirks that existed long before Linus ever started kernel hacking; quirks we've all accepted in the community and have become so inured to working around that we don't even think of them as problems anymore, if we think of them at all. Come learn some history, learn some real-life gotchas and how to avoid them, and some words of wisdom all the greybeard old-timers forgot to give you.

As a hobby, I started building a personal web app to solve my needs as a Python community organizer and a public speaker. I call these my own “secret codes”, but there’s nothing secret about it. These are all open source Python web apps built on Django.

Why build my own when there are existing SaaS products out there?
Using SaaS products are free but they have hidden costs like: needing to accept third party cookies, having your data used/trained, having to pay for more powerful features.

In this talk I'll walk through what I built, why I built it instead of reaching for free tools, and what it would take to let others use it too. Convenience, privacy, and trust are worth paying for.

Digital badges promise portable, verifiable recognition, but in practice, they’re locked behind proprietary platforms, expensive subscriptions, and rigid workflows. For many nonprofits and open communities, that makes them unusable.

BadgeFed rethinks digital credentials for the open web. Built on the Open Badges standard and the ActivityPub protocol, it brings badge issuance, sharing, and verification into the Fediverse, fully open source, self-hostable, and designed to run on Linux.

This talk explores how federated social infrastructure can replace centralized credential platform and build sustainable recognition systems without vendor lock-in.

Even though finishers will nearly always get the credit, starters are the secret power that changes the world. From the US civil rights movement to the push for community-controlled social media, finishers have always relied on dozens or hundreds of starters to lay the groundwork for them. We should do more to reward starters in open source.

As duopolies in technology continue to flex their power for purposes contrary to what users want to do with their machines, and governments insist upon pushing age-verification requirements into software and hardware, interest in Linux as an alternative (or the foundation of an alternative) continues to grow. However, Linux still holds a reputation as a technical and scary operating system for nerds. It needs a better "elevator pitch," a short, accurate, and interest-piquing description. Let's build one together!

A talk on how modern rockets use software (often times Open Source) to accomplish nigh-impossible tasks, such as taking off from Earth or very accurately getting to a certain orbit. Goes into a simulation I have made, which includes naive flight software capable of takeoff and orbit.

To non-engineers, geeks often sound like a chorus of doom. When we say a codebase is "complete garbage" we're usually describing a 0.x% deviation from technical excellence - but the sales and markting teams quickly update their resumes. This talk explores the "Engineerspeak" dialect: why our brains are wired to prioritize edge cases, minutia, and fatalistic honesty. We'll break down how to bridge the gap between technical rigor and organizational communication without losing our integrity. Expect humour, horror stories, and a healthy dose of self deprecation. Let's optimize the most buggy stack interface: the one between teams.

This is a presentation about how to pick a Linux distribution. For new users, it can be a guide. For intermediate and experienced users, it can be a refresher as to what's out there.

A small effort to keep old laptops out of landfills grew into The Computer Upcycle Project, a community upcycling program that's refurbished 3,000+ computers for people in need. This talk shares how a local donation program in Olympia, WA turned into a full-scale computer reuse effort — breathing new life into old hardware with a mix of Linux systems, including Debian, Bazzite, Fedora, and our own Nixbook OS, a simple, stable, auto-updating NixOS-based build. Learn how thoughtful design, the right OS for the right machine, and community action turn e-waste into opportunity.

AI assistants like Claude and ChatGPT are powerful, convenient, and a trap. The cost, in addition to the price itself, is that your data and your code are being handed off to a corporation and who knows who else. Breaking your dependency on cloud AI takes effort and sacrifices convenience, but the gap between local LLMs and frontier models is narrowing every day. The hidden challenge is the harness, everything that wraps around the LLM to make it useful. This talk shows how to run an agentic AI stack using open source tools and whatever hardware you have to make useful tools without giving up your privacy and your sovereignty.

There is a vast treasure trove of free Latin language pedagogical content available on the Internet, but shockingly little of it is in plaintext -- the greppable, diffable, interoperable, and eternal format that is the fundamental substrate of GNU Emacs.

In this talk I'll demonstrate how I use both local and cloud-hosted ML models to transform recorded Latin audio into glorious structured plaintext that not only encodes the text itself, but also its morphological and semantic context. Thus transformed, we can now render the text with attached metadata that enables powerful word-at-point features like linguistic details in the echo area, accurate dictionary lookups, and audio playback with precise word alignment.

On March 27th, 2026, a coalition of European companies including IONOS and Nextcloud announced "Euro-Office," an AGPLv3-licensed fork of OnlyOffice. While public discourse has rightly focused on the legal nuances of the AGPLv3 license, this talk will rather focus on one of the project's other key objectives: liberating the codebase from a vendor accused of "open-washing."

Learn how to use the man command on Linux to read documentation!

Man pages have a reputation for being cryptic, intimidating, and written for people who already know the answer. However, they're available offline, pre-installed in most places, and consistently structured. Once you can "RTFM" (read the friendly manual), you'll never go back!

This talk is for people who have never heard of man pages, and those who've never opened one without fear and trepidation. I will:

• Use Ubuntu and basic command line tools
• Explain how to navigate the document structure of man pages
• Solve some simple problems using man pages
• Install some tools to make your man prettier and friendlier

Have you ever wanted to achieve website perfection?

Ever wanted to get (and maintain) four 100/100 scores in Lighthouse? Make automated accessibility tests like Axe DevTools pass completely? ... Even tame the most verbose and difficult of checkers, the W3C Validator?

I am here to show you how to get there and show you its not as hard as it seems. Even under such rigid constraints, a lot of creativity is possible. In this talk I will demonstrate a handful of sites that are simple and pass tests with flying colors, and also show how they manage to be incredibly useful and communicate information. And best of all, you can make one too!

I have high demands for my private computing! Data privacy, control over my CPU usage, fast performance on old hardware, and organized retrieval of uncountable local research notes for fiction and non-fiction writing. I am not a developer, and rarely use the command line, but I know what I am doing, and know my limits. Here is one way to be a user of a Linux desktop environment.

Open hardware and open source run on shared knowledge, but we mostly share the things that worked. The dead ends and half-broken attempts stay private, even though they often save the next person the most time. I want to make the case for sharing those too. My example is a flashlight I tried to 3D print from filament that conducts electricity. The first version barely lit up, and I posted it anyway as a "limited success." The second one worked. I will walk through both, with real measurements and a straight answer on what the filament is good and bad at. Cheap, printable, open hardware makes it low-risk to share the messy parts. And yes, I will bring flashlights to hand out.

It's larger than you can comfortably fit in your head.

Resist Tech Monopolies (RTM) runs a game called Mapbattle, at which participants use public transportation to go around the city and make Open Street Map (OSM) contributions to gain territory.
The presentation will go over the idea of the game and what inspired it (spoiler: SeaGL 2025 is part of that story), the technology behind the game's website, and the observed community benefits of running this game as well as other benefits we hope the game achieves as it keeps running.

Functional programming is often seen as difficult or only relevant to languages like Haskell, but its core ideas can make anyone a better programmer. In this beginner-friendly talk, we'll explore concepts like pure functions, immutability, higher-order functions, and type-driven design using familiar examples from Python and Java before seeing how Haskell brings these ideas together. You'll learn how functional thinking leads to code that's easier to understand, test, and maintain, and leave with practical techniques you can apply in your everyday projects, regardless of the language you use.

a simple guide on how to get started building a following or community on the fediverse

it will include: an intro to the mastodon user interface, the common cultural norms and practices, how to administrate a single-user or multi-user instance, a brief bit of history about the fediverse, what the fediverse is and isn't, how moderation works on fedi, how to build and maintain a following, how to handle interpersonal and inter-instance drama, and more!

AsciiDoc is a great text tool for documentation from simple notes to books. Asci
iDoc is human readable plain text markup.

Since it's plain text, AsciiDoc lends itself to revision control systems. Even WebUI systems like GitHub wiki can support it.

AsciiDoc was created in 2002 as a human readable format for publishing DocBook XML. Via different backend processors, it can also be converted to many other formats. Plain text, revision control, and wikifiable is a trifecta of sysadmin documentation awesomeness.

Attendees will become familiar with use cases, document conversion, and basic syntax.

Bring your questions and get local, offline answers on the spot! Or, just watch a parade of sketches. A chance to laugh at, and ponder, the state of AI and how it impacts us. Part fun psychosis, part serious presentation built from S5 format. A standalone companion to Linux Prepper podcast's current season on local tooling and offline workflows. If you are curious as a FOSS enthusiast, hate AI, or love plain text... this show is expressly for this conference.

Running thousands of snapshotted VMs pushes Linux networking to its limits. This talk covers battle-tested strategies balancing performance, isolation, and efficiency across large-scale topologies. We explore Geneve tunneling for scalable overlay networks, eBPF-accelerated networking for line-rate performance, and stateless NAT without connection tracking overhead. Prefix-based iptables/nftables rules scale to thousands of instances, while namespace pooling minimizes snapshot operation costs. We also address RTNL lock contention and its mitigation. Attendees will learn to architect network stacks for snapshotted workloads where VMs are cloned, restored, and migrated while maintaining security and predictable performance.

This talk will discuss the importance of environmentally-minded technologists, how you can become one, & what technologies there are for us to be excited for. Additionally it will discuss why "AI" isn't amongst them.

It was originally presented at the Wakaatu Arts & Action! Club in Nelson, New Zealand.

The world is rich in old computers that work just fine and unprotected free Wi-Fi. Every time I stay in a hotel room, I wonder "how long could you hide a computer here?" The first volley came in the form of https://pretty.dubious.computer/, an internet radio station coming live from an old cell phone I plan to hide in a bike locker somewhere in Bellevue. There's some fun technical challenges and my own personal sense of whimsy. You like whimsy, right?

Oh yeah, come ask for goodies after the show, too.

There are a lot of convenient & cheap ways to learn about AI, but most of them come with strings attached. Usually, free AI services use your data to train their AI models - a privacy nightmare.

Instead of relying on proprietary AI providers, you can take advantage of open-source software and open weight models to learn AI concepts while maintaining your privacy.

In this talk, I'll share the setup I use, giving a lightning-fast tour of many open source projects that can be used to learn about AI engineering at home, with a focus on inference with small models, image generation and recognition, and basic observability.

Kalpa Desktop is a transactional and atomic desktop, based on openSUSE and KDE Plasma

One of it's aims is to provide a more flexible system that is a middle ground between a traditional read-write Linux installation, and more modern read-only "immutable" systems

This talk will focus on introducing the Concepts and Architecture of the Kalpa Desktop, and some of the difference with other similar products

What if we developed a Federated network that was intentionally... not connected to the internet?

An off-line first, off-grid network...

Where limited range and high degrees of locality were a feature to help build connections between neighbors?

Where slow and possibly infrequent, sporadic connectivity between neighborhoods was a design intention, maybe even a special occasion?

Where information could be delivered by bicycle, or seriously, why not carrier pidgeon?

There is free software to make this possible already, time to put the pieces together!

For decades, OSS communities grew through visible friction. Newcomers filed issues, asked questions, and in doing so announced their arrival. Maintainers noticed. Contributors became maintainers. Now AI is removing that friction.

Through an NSF-backed program, stdlib conducted 100+ stakeholder interviews and found the same pattern: newcomers increasingly rely on AI instead of engaging with communities. Maintainers never see them arrive, and communities lose opportunities to grow contributors.

This session is intended for maintainers, OSPO leads, and researchers interested in the future of FOSS, Attendees will leave with evidence-based frameworks for growing contributors & community as the old, accidental on-ramps quietly disappear.

Most Linux distributions ship precompiled software or compile at install time. This talk explains the architecture and design of a distribution that does not do this.

ccix, an experimental Linux distribution, explores deferred compilation (a.k.a. compilation-on-demand), a technique where applications are compiled when they are invoked. It uses a custom 'deferred compiler' based on libtcc that can resolve symbols and cache outputs. It is performant, compiling bzip2 in 0.07 seconds on an i5-6500.

I find that it's difficult for me to transform the words in my head into the audio that people can hear. Sometimes, even the right words sound wrong due to the auditory properties of how I say them.

To be able to get the words out with more ease, in ways that feel more comfy, I've created a portable text to speech software, built on top of open-source hardware and existing FLOSS.

Introducing Painted Wolf Code. This Apache-2.0, local-first AI coding agent is built and tuned for open source models. Agent personas and coordinator guidance use composable templates, multi-step work is declarative YAML workflows you extend, compose from templates, or override per project. An evidence ledger grounds agent output. Security scanning is built in, including supply-chain, secrets, and SAST, wired into workflow gates. Local first, open source AI coding has arrived.

We’re building a full source bootstrap of the Sparrow Bitcoin wallet for the open-firmware POWER9 platform. This goes beyond an ordinary reproducible build in that Sparrow and its full dependency graph, including build tools, are built reproducibly from source, minimizing the number of parties you need to trust in order to trust the program. In the process, we’re bringing the modern Java ecosystem to Guix, and solving problems from Rust bootstrap compiler bugs to a circular dependency between Gradle and Kotlin. I’ll cover the problems we’ve encountered and the tools and techniques we’re using to address them. I’ll also discuss future directions for the project, such as using Guix to package Sparrow for Debian and other distributions.

So you've got a great open source project. How do you get people to try it? One senior technical writer shares how to write tech docs that cater to all skill levels, get users started quickly, and represent your project to the world via SEO and AEO. It's easier than you think!

It's time once again for the 2026 FLOSS Research Roundup!

Hundreds if not thousands of research papers are written every year about open source software! How should we make sense of it all and what does it mean? In this talk, I carefully harvest the latest research from around the world, filter it for real world relevance, shake it up with the latest trends, and then serve it up with a slight dash of attitude.

Whether you want to explore launching a new open-source project, dip your toe into freelance work, or identify a good startup idea, getting direct input from many potential users or customers will ensure that your venture is received with enthusiasm and receives the support necessary to be sustainable over the long term. Unlike user experience research that asks about specific features of specific software already in development, customer discovery is a process of learning what the customer most needs and wants. In this session, the presenter will conduct and debrief several abbreviated interviews to demonstrate effective techniques for uncovering valuable unexpected data to inform new ventures and pivots.

Transactional systems offer a middle ground between Fedora Atomics, and their image based updating, and read-only filesystems and Traditional Fedora, with it's read/write filesystems, as they can be configured to do either, offering great flexibility.

Transactional systems offer Atomic updating, offering the possibility of more reliable updating, more stable systems over time, and the ability to "rollback" snapshots, in case of issues. or for troubleshooting purposes.

This talk will discuss the basic ideas behind Transactional Updates, how they're used in openSUSE, and what it takes to bring them to Fedora.

"Why isn't my distribution shipping the latest version yet?"
It's a question almost every Linux user, developer, or platform engineer has asked. Behind that seemingly simple question lies one of the most important—and least understood—engineering tradeoffs.
Every package update must balance security, compatibility, regression risk, reproducibility, dependency management, architecture support, release criteria, and long-term maintenance. What appears to be "slow" is often deliberate engineering designed to reduce risk for millions of users.

Modern services depend on far more than compute- databases, messaging, observability, CI/CD, service meshes, IAM, and cloud resources must all work together reliably. Managing these consistently across production, staging, and developer environments remains a persistent challenge.This talk explores how open-source Kubernetes operators can be combined to build complete, isolated, self-service environments. Using Config Connector, Istio, Argo CD, observability operators, and Crossplane, we'll see how infrastructure, networking, deployment automation, and cloud resources compose into higher-level platform APIs.

The Signet Hardware Password Manager is an example of just how unstoppable open source hardware and software can be. This talk will include a payload that demonstrates how software password managers can be no better than a text file in the event of even just a momentary non-root compromise. I'll also explain how using hardware limits the damage. As you'd expect, all the hardware is open source and you can build your own at home, inspect compile the firmware and software, or extend it to do more.

What would a Linux text interface designed for mobile devices look like?

At the beginning of 2026, I started building a Nix-based distribution with a TUI shell, called Huckleberry. It's intended for small laptops and handheld devices with keyboards. We'll follow the development, starting with a search for existing, well-supported console software, and continuing to developing a bespoke menu-based mobile shell, and integrating the various selected and developed software into a portable Nix configuration available for public testing. We'll look at some of the issues that have come up from early Huckleberry use, and review the process of designing and redesigning huckleberry to patch them.

The security characteristics of LTS vs regular release models haven't fundamentally changed over the last 30 years, but the general community sentiment toward them has. As adversaries develop and adopt better tools, it becomes increasingly important to understand the difference between LTS and regular release models.
This talk will explore the basic challenges that exist in keeping LTS systems secure, especially as they continue to ship code that is no longer maintained upstream.

Tired of relying on big Tech or klonky stacks? If so, come learn how to build self-hosted and scalable privacy-centered network infrastructure using mainstream FLOSS. This presentation explores complete DNS sovereignty with both LAN and WAN solutions. On the LAN side, it covers deploying OpenWRT on GL.iNet hardware with virtualized Pi-hole instances paired with Unbound for fast cached recursive DNS resolution and powerful network-level ad and tracker blocking. On the WAN side, it details securely hardening Unbound for public-facing use, with clustered Webmin for centralized management and Bind9 delivering authoritative DNS records. Geared toward home lab hobbyists, dedicated hosting, and data center colocation professionals.

This talk explores how systemd-nspawn, a little-known but extremely versatile native process encapsulation subsystem for Linux, could be used as a resource-efficient and secure alternative to containers or virtualization as part of a privacy-conscious self-hoster's toolbox.

OpenHospital is a free and open-source health information management system used in hospitals across Uganda, Ethiopia, and other resource-limited regions. Built by Informatici Senza Frontiere, it manages patient records, laboratory results, pharmacy stock, and statistical reporting for healthcare workers and patients who depend on it daily.

As a team of four undergraduate students contributing to OpenHospital for the first time, we each tackled separate “good first issues” across reporting infrastructure, interface consistency, localization cleanup, and GUI data accessibility. This talk shares what we worked on and what we learned contributing to production healthcare software as first-time open source contributors.

Programming sucks. Code sucks. It's hard to read, hard to test, and hard to maintain. Only a handful of people can understand any particular software project. All users deserve to be able to control their own software, regardless of their capacity to code. Because of this, we should develop tools for visual, literate, and natural language programming that lets anyone instruct their computers what to do without relying on LLMs and Generative "AI".

Placeholder

Not since the 1980s has technology changed as quickly as in recent months. Reprising the early PC era, capitalism dictates the pace and control of Large-Language-Model-backed generative artificial intelligence (LLM-gen-AI).

Understandably, the FOSS community is outraged; LLM-gen-AI ballyhoo violates our core principles of sharing, cooperation, and mutual assistance. Fear abounds while LLM-gen-AI impacts us; it divides our community.

Of course, we shouldn't accept useless slop; yet, LLM-gen-AI does not represent an existential threat to FOSS. Disciplined usage of LLM-gen-AI provides us an opportunity to improve FOSS, and undo the bad policies promulgated by the purveyors of LLM-gen-AI.

Come learn my plan for how we can do it!

Enjoy some refreshments and great conversation at Big Time Brewery located close to the University of Washington on the Ave.

4133 University Wy NE, Seattle, WA 98105

https://www.openstreetmap.org/?#map=19/47.657852/-122.313575