History of Open Source: From GNU to GitHub

Open source did not start at GitHub. It started in a world where software came attached to hardware and the people who wrote it were expected to share. In the 1960s and 70s, at places like MIT and Bell Labs, code circulated freely between researchers. Unix, written at Bell Labs in 1969, was passed around with its source code, and a generation of programmers grew up reading and improving it. That culture, often called hacker culture, assumed the source was part of the package.

Then computing went commercial. By the 1980s, vendors sold binary-only software signed with nondisclosure agreements, and sharing became illegal. Richard Stallman, a researcher at MIT's AI Lab, watched the shift and refused to accept it. In 1983 he announced GNU, a recursive acronym for "GNU's Not Unix", a plan to build a complete Unix-compatible operating system that everyone could use freely. He also wrote a license for it, the GNU General Public License, which used copyright itself to guarantee freedom: anyone could use, modify, and redistribute the software, but modified versions had to be shared under the same terms. That trick became known as copyleft.

In 1985 Stallman founded the Free Software Foundation to fund and protect the project. GNU grew steadily, a compiler here, a text editor there, but a crucial piece was missing: the kernel, the core that talks to the hardware.

That piece arrived in 1991 from a 21-year-old Finnish student, Linus Torvalds, who built a Unix-like kernel for his own PC and posted it to a newsgroup. He released it under the GPL, and the combination of GNU tools and the Linux kernel gave the world a complete free operating system. Linux spread because it was open: thousands of people could fix bugs, add drivers, and build distributions. The internet had finally made that kind of collaboration possible at scale.

By the late 1990s, "free software" had a marketing problem. The word "free" was read as price, and the movement's moral framing made businesses nervous. In 1998 a group of advocates, including Eric Raymond and Bruce Perens, coined the term "open source" to describe the same approach with the emphasis on practical benefits: better code through many eyeballs, faster bug fixes, no vendor lock-in. The meeting where the term was picked was itself a response to a shock. Netscape, the leading browser company, had just released its browser's source code to fight off Microsoft. That event, in January 1998, is what made companies pay attention.

Around the same time, the web had quietly built its own open source success story. The Apache web server, started in 1995 from a series of patches to the NCSA server, grew to dominate the early web. The name itself is a joke, "a patchy server", which tells you everything about how the project felt then. By 1999 Apache served more sites than all proprietary competitors combined. Open source had proven it could run the infrastructure of the new internet.

The next decade industrialized the model. Red Hat went public in 1999, and enterprises began trusting Linux and open source databases. The bigger shift was tooling. Git, released by Torvalds in 2005 to manage the Linux kernel, changed how version control worked. Then GitHub launched in 2008 and made Git social. Suddenly contributing to an open source project meant fork, edit, pull request, review, merge, a workflow so clean it became the default way software teams work everywhere, not just in open source. GitHub did not invent open source, but it turned it from a philosophy into an everyday practice for millions of developers.

That is where the corporate takeover happened, and quietly. The companies that once feared open source are now its biggest funders. Google, Microsoft, and Meta run enormous open source programs, and much of the modern internet's infrastructure, Kubernetes, React, PostgreSQL, is open source. Open source became the default for new software in a way Stallman in 1983 could not have imagined.

The newest frontier is AI. The live argument is about open models and open weights: Meta released Llama openly while others keep models closed. In 2026 that debate, over whether the inner workings of an AI model should be public, echoes the 1980s fight over source code almost exactly. The players have changed, and the question is the same, whether the people who use a technology are allowed to understand it. If you want to be part of it, my guide to finding open source projects for beginners is where to start today.