Open Source (OSS): Advantages, Challenges, and AI
Open source software refers to software projects shared online on GitHub, Codeberg, Tangled, or other GitHub Alternatives, usually with a git repository. Similarly, modifications and derived works are allowed based on Licensing Models.
The acronym of OSS
In our context it refers to Open Source Software. Wikipedia
# The Hook of Open Source
I’ve been a big fan of open source since very early on, when I worked in Business Intelligence, and all I had to use was closed-source software, and you could only try it after talking to a sales agent. With open-source, this was different. The code wasn’t perfect or running smoothly, but I could try it, get a feel for it, fork it, change it, and expand on it. That’s what really hooked me early on.
Today (2026-08-18), two decades later, everything has changed; see some more thought on How is it Changing with AI below, but I still think OSS is what is best for humanity in general. Why build 100 of the same things individually? Why not work together and make one great product?
# Why Open Source?
Given that a single organization can’t create and maintain every data connector, framework, or component along the data stack, open source is a method to tackle the challenges in a new way together.
Let’s look at a few software companies as examples. Red Hat, Elastic, and Cloudera have implemented successful open-source strategies. Airbnb has more than 30 open-source projects, and Google has more than 200. In data, Databricks open-sourced Apache Spark, so now everyone has access to its internals.
The beauty of open-source is that the more you share, the more feedback, contributions, and, most importantly, trust a product receives. Usually, open-sourcing improves a product, as community contributions can’t compare to those from a small team of in-house engineers.
Even though it may seem counterproductive to share code publicly, as competitors may try to replicate it, doing so may be the most effective way to foster collaboration and creativity. The pressure to turn a profit is ever-present. Still, there is ample opportunity to build AI-powered, business-driven innovations on top of a data stack that strengthens with each new open-source contribution.
# What’s the Advantage of Open Source?
The advantage of OSS is not really that it’s “free” (that is relative in IT), but strategic with interoperability — says Matthias Stürmer in 1 zu 1 mit Matthias Stürmer, BFH - 1 zu 1 - der Bedag Fachtalk | Podcast on Spotify (more).
# Challenges of Open Source
One of the biggest challenges, and a common one, is “how to make money”? We saw software is eating the world with Why Software Is Eating the World by a16z, to open-source strategies in Silicon Valley VC strategies, to give everything for free, build a great product, and figure out the finances later. But that’s not as easy.
Sponsoring would probably be best, or if the repository is big enough, e.g., the one that started it all, the Linux repo, you get status and might get another job or so. But it’s still, to this day, not a solved problem.
# How OSS Makes Money
I like how 37Signals handles it, they produce 100 of repos with Kamal, Fizzy, Omarchy, Shape Up and so much more. But I heard them saying that first you make the money, and only then you share it. And after that, you don’t expect any return. DHH called it a Make Gifts, this is all open source is. Because others can take it or leave it, but can’t demand that you do things. It’s a gift to the world.
I like this framing, too.
What open source also enforces is a focus on Openness and Open Standards. If you become the standard through openness and transparency, you might capitalize down the road with services or just good marketing.
# Related: Opposite of OSS, Real Names and Fun of it
# The Opposite of OSS: Proprietary Software
Technically, “closed-source” is not the right opposite. Proprietary software is more correct, and it’s about licenses, which define what can be used and how.
# Why Open Source Issues Should Be Created with Real Names
Good write-up on Dear anonymous internet user asking for help.. - Bert Hubert’s writings:
First, I just consider it rude. You come at me hiding who you are but still expect me to do free work for you. Try doing that in real life. What were you thinking? Not introducing yourself AND using a fake identity?
Found it linked in the DuckDB issue creation.
# The Fun and Joy of OSS
The joy of contributing to an OSS project. Building something for yourself, but sharing it and getting it distributed to 1000s. Not because of anything, but for the fun and joy of it. Very rewarding. That’s why I believe the OSS-based strategy will never die, despite the dbt/Fivetran merger’s efforts to convince us otherwise. My Bsky post ^1cb4bd
# How is it Changing with AI
AI is the best and worst thing that has happened to open-source at the same time. As a user, you can customize every piece of software by just prompting the cloned git repo. E.g., Omarchy is my OS, and I can just prompt to change everything in my OS; that is amazing.
The worst part is from the maintainer’s perspective—so many PRs and noise to deal with. Also, programming is leading to AI exhaustion, as you typically work more and just less satisfied. You don’t even know what you’ve built anymore.
A great video highlighting all of it is here by ForrestKnight.
# A Great Example: Agentic OS
With Omarchy as the first Agentic OS, you basically prompt your Agent to add a widget for time zones, or a stats panel for your website, or a weather radar warning with satellite images, directly built into your OS. All of these are plugins I’ve built or installed from the Omarchy Marketplace that people quickly prompted.
And the key here is Open Source. Because Linux, the foundation of Omarchy is open source, it’s in the training data of every model, the Omarchy provides skills for how to build plugins with clear instructions to use Quickshell, a Flexible toolkit for making desktop shells that uses something called QtQuick that lets you build easily with C++. Because agents know that language well, it’s super easy and super fast, with C++ being one of the fastest Programming Languages out there.
# COVID: Collaborativeness of the Whole World
Another good example was COVID-19, where a vaccine was developed faster than ever before, with people working together all over the world. Not really open source in the sense of code, but sharing the learnings openly among labs is the same spirit.
When resources are limited, we need to work together and share our findings instead of re-creating the same at every company again and again.
# Further Reads
- How open-source envolved explained in Open Source: From Community to Commercialization by a16z (internal)
- Open source security at Astral (company behind uv)
- Supporting Open Source
- Open-Source Data Engineering Projects
Origin:
References: Open Source at Airbyte