How the Open Source Underdogs Shaped the Internet and Will Transform AI

I remember the early days of the internet, which felt like the Wild West. Big players like AOL, MSN, and France's Minitel were trying to control how people connected online through their own closed networks, separate from the internet's open standards. These commercial networks aimed to dominate the future of communication. At the same time, tech giants like Sun Microsystems, IBM, and Microsoft (in cooperation with Intel, known as Wintel) were battling for dominance in data center servers with their closed operating systems. It was a world of walled-off, super-expensive solutions.

But then the underdog open-standards and open-source rebels showed up with free-to-use solutions running on more affordable hardware.

In 1990, British computer scientist Tim Berners-Lee and Belgian Robert Cailliau, working at Switzerland's CERN, introduced HTTP and HTML, the open protocols that became the foundation of the World Wide Web. Soon after, Flash claimed HTML was a dead end, especially for animated websites. Flash provided rich media experiences, interactive games, and animations in a way HTML at the time couldn't match. Adobe Flash was seen as the future for multimedia on the web, offering an alternative to the more static nature of HTML pages. However, Flash was inherently a closed structure, requiring proprietary software for creation and viewing, which stood in stark contrast to the open standards movement. Ironically, it's Flash that has died today, with its support officially ended by Adobe in December 2020—a testament to the triumph of open standards.

In 1991, Linus Torvalds, a Finnish student, created Linux, an open-source operating system that challenged the dominance of Unix and Windows servers by running on more affordable x86 architecture. Today, Linux dominates with over 75% of internet servers in datacenters globally and is even more prevalent in AI applications. In 1994, Michael Widenius created MySQL, an open-source database that challenged Oracle and IBM DB2. To this day, MySQL remains one of the most widely used database systems for web-related applications. These innovators, who believed in collaboration, transparency, and freedom, were the rebels against large corporate giants and their financially backed closed systems—and their approach worked. Open standards like HTTP and HTML, combined with open-source Linux and MySQL, running on more affordable x86 hardware, reshaped the internet into the decentralized, global platform we know today.

Fast forward to 2025, and we’re seeing the same dynamics in AI.

On one hand, big closed-source players like OpenAI, Stargate (backed by Microsoft), and other corporate giants are building powerful, proprietary AI models. These systems are impressive but come with concerns about control, accessibility, and monopolization—echoing the early internet days when Microsoft, IBM, Sun, AOL, and others tried to dominate through proprietary solutions.

On the other hand, open-source AI projects like DeepSeek, Stable Diffusion, and EleutherAI are driven by small groups of individual researchers, developers, and large communities of enthusiasts worldwide who believe AI should be open, transparent, and accessible to everyone. Even companies like Meta are contributing to open-source AI with projects like LLaMA, bridging the gap between corporate and community-driven innovation.

Another observation: Real innovation is not primarily coming from the USA and its large, well-capitalized companies, contrary to what their PR and marketing machines often claim. In the nineties, we saw global innovators, mainly from Europe, pushing back against the large, capitalized U.S.-based companies that sought to control how people communicated. Today, we see the same pattern, with significant contributions from around the world, including Chinese innovations that challenge the dominance of established powerhouses.

So let’s celebrate the underdogs, the rebels, and the dreamers—and support open-source AI and the communities working to make AI accessible to everyone. Because, just like in the 90s, the future belongs to those who believe in openness, collaboration, and freedom. Just as HTML eventually evolved to meet the challenges posed by Flash, we can expect AI to follow a similar path, driven by the spirit of open innovation.