Openness

AI coding agents and a rerun of the operating-system wars

Photo source: Rod Long (@rodlong) | Unsplash

Attending the Agentic AI for Science workshop at MPCDF in Garching got me thinking about the platforms we are starting to build our scientific workflows around. The workshop programme brings together hands-on work with agents, examples from scientific software development and materials simulations, and perspectives from industry. For someone interested in free and open-source software, this raises some familiar questions: who controls those platforms, what can we change ourselves, and how much of our work can we take elsewhere?

What interests me here is the software around the model. An agent harness gives the model tools, supplies context, and controls what it can do. The flagship GPT, Claude, and Gemini models are proprietary, but the software using them does not have to be (and often isn't). We can use the same model through different harnesses, with different licenses, tools, and restrictions. Choosing the model, deciding where it runs, and choosing the harness are separate decisions, even when a product offering bundles them together.

A few harnesses and development environments for programming, scientific workflows, and other tasks illustrate these differences. This is only part of the picture: agents embedded in enterprise applications bring their own integrations and dependencies.

This led me to a somewhat loose comparison: Claude Code reminds me of Windows, Codex CLI of BSD, OpenCode and the like of Linux, and Antigravity of the Macintosh. Each comparison emphasizes something different: compatibility, permissive reuse, independence from suppliers, or the appeal of an integrated experience.

If you have used Claude Code in a terminal, the first part probably sounds wrong already. How does a program that reads files, invokes shell tools, and fits into a Unix workflow resemble Windows? Bear with me. The resemblance is in how people build tools and workflows around the software, rather than how its interface looks.

Should I do a Ph.D.?

Photo source: Santeri Liukkonen (@iamsanteri) | Unsplash

Tough question, and the one that has been asked and answered over and over. The simplest answer is, of course, it depends on many factors.

As I started blogging at the end of my journey as a doctoral student, the topic of how I selected the field and ultimately decided to enroll in the postgraduate studies never really came up. In the following paragraphs, I will give a personal perspective on my Ph.D. endeavor. Just like other perspectives from doctors of not that kind, it is specific to the person in the situation, but parts of it might apply more broadly.

Don't use RAR

Photo source: Tim Mossholder (@ctimmossholder) | Unsplash

I sometimes joke with my TA Milan Petrović that his usage of RAR does not imply that he will be driving a rari. After all, he is not Devito rapping^Wsinging Uh 😤. Jokes aside, if you search for "should I use RAR" or a similar phrase on your favorite search engine, you'll see articles like 2007 Don't Use ZIP, Use RAR and 2011 Why RAR Is Better Than ZIP & The Best RAR Software Available.

Mirroring free and open-source software matters

Photo source: Tuva Mathilde Løland (@tuvaloland) | Unsplash

Post theme song: Mirror mirror by Blind Guardian

A mirror is a local copy of a website that's used to speed up access for the users residing in the area geographically close to it and reduce the load on the original website. Content distribution networks (CDNs), which are a newer concept and perhaps more familiar to younger readers, serve the same purpose, but do it in a way that's transparent to the user; when using a mirror, the user will see explicitly which mirror is being used because the domain will be different from the original website, while, in case of CDNs, the domain will remain the same, and the DNS resolution (which is invisible to the user) will select a different server.

Free and open-source software was distributed via (FTP) mirrors, usually residing in the universities, basically since its inception. The story of Linux mentions a directory on ftp.funet.fi (FUNET is the Finnish University and Research Network) where Linus Torvalds uploaded the sources, which was soon after mirrored by Ted Ts'o on MIT's FTP server. The GNU Project's history contains an analogous process of making local copies of the software for faster downloading, which was especially important in the times of pre-broadband Internet, and it continues today.

Fly away, little bird

Photo source: Vincent van Zalinge (@vincentvanzalinge) | Unsplash

The last day of July happened to be the day that Domagoj Margan, a former student teaching assistant and a great friend of mine, set up his own DigitalOcean droplet running a web server and serving his professional website on his own domain domargan.net. For a few years, I was helping him by providing space on the server I owned and maintained, and I was always glad to do so. Let me explain why.

Celebrating Graphics and Compute Freedom Day

Photo source: Elena Mozhvilo (@miracleday) | Unsplash

Hobbyists, activists, geeks, designers, engineers, etc have always tinkered with technologies for their purposes (in early personal computing, for example). And social activists have long advocated the power of giving tools to people. An open hardware movement driven by these restless innovators is creating ingenious versions of all sorts of technologies, and freely sharing the know-how through the Internet and more recently through social media. Open-source software and more recently hardware is also encroaching upon centers of manufacturing and can empower serious business opportunities and projects.

The free software movement is cited as both an inspiration and a model for open hardware. Free software practices have transformed our culture by making it easier for people to become involved in producing things from magazines to music, movies to games, communities to services. With advances in digital fabrication making it easier to manipulate materials, some now anticipate an analogous opening up of manufacturing to mass participation.

The academic and the free software community ideals

Photo source: Giammarco Boscaro (@giamboscaro) | Unsplash

Today I vaguely remembered there was one occasion in 2006 or 2007 when some guy from the academia doing something with Java and Unicode posted on some mailing list related to the free and open-source software about a tool he was developing. What made it interesting was that the tool was open source, and he filed a patent on the algorithm.

Free to know: Open access and open source

Photo source: Álvaro Serrano (@alvaroserrano) | Unsplash

Q&A with Vedran Miletić

In June 2014, Elon Musk opened up all Tesla patents. In a blog post announcing this, he wrote that patents "serve merely to stifle progress, entrench the positions of giant corporations and enrich those in the legal profession, rather than the actual inventors." In other words, he joined those who believe that free knowledge is the prerequisite for a great society -- that it is the vibrancy of the educated masses that can make us capable of handling the strange problems our world is made of.

The movements that promote and cultivate this vibrancy are probably most frequently associated with terms "Open access" and "open source". In order to learn more about them, we Q&A-ed Vedran Miletić, the Rocker of Science -- researcher, developer and teacher, currently working in computational chemistry, and a free and open source software contributor and activist. You can read more of his thoughts on free software and related themes on his great blog, Nudged Elastic Band. We hope you will join him, us, and Elon Musk in promoting free knowledge, cooperation and education.

I am still not buying the new-open-source-friendly-Microsoft narrative

Photo source: Patrick Bellot (@pbellot) | Unsplash

This week Microsoft released Computational Network Toolkit (CNTK) on GitHub, after open sourcing Edge's JavaScript engine last month and a whole bunch of projects before that.

Even though the open sourcing of a bunch of their software is a very nice move from Microsoft, I am still not convinced that they have changed to the core. I am sure there are parts of the company who believe that free and open source is the way to go, but it still looks like a change just on the periphery.

All the projects they have open-sourced so far are not the core of their business. Their latest version of Windows is no more friendly to alternative operating systems than any version of Windows before it, and one could argue it is even less friendly due to more Secure Boot restrictions. Using Office still basically requires you to use Microsoft's formats and, in turn, accept their vendor lock-in.

Put simply, I think all the projects Microsoft has opened up so far are a nice start, but they still have a long way to go to gain respect from the open-source community. What follows are three steps Microsoft could take in that direction.

AMD and the open-source community are writing history

Photo source: Andrew Dawes (@andrewdawes) | Unsplash

Over the last few years, AMD has slowly been walking the path towards having fully open source drivers on Linux. AMD did not walk alone, they got help from Red Hat, SUSE, and probably others. Phoronix also mentions PathScale, but I have been told on Freenode channel #radeon this is not the case and found no trace of their involvement.

AMD finally publically unveiled the GPUOpen initiative on the 15th of December 2015. The story was covered on AnandTech, Maximum PC, Ars Technica, Softpedia, and others. For the open-source community that follows the development of Linux graphics and computing stack, this announcement comes as hardly surprising: Alex Deucher and Jammy Zhou presented plans regarding amdgpu on XDC2015 in September 2015. Regardless, public announcement in mainstream media proves that AMD is serious about GPUOpen.

I believe GPUOpen is the best chance we will get in this decade to open up the driver and software stacks in the graphics and computing industry. I will outline the reasons for my optimism below. As for the history behind open-source drivers for ATi/AMD GPUs, I suggest the well-written reminiscence on Phoronix.