What does ‘open source AI’ mean, anyway?

Paul Sawers 8:00 AM PDT • June 22, 2024

Collected at: https://techcrunch.com/2024/06/22/what-does-open-source-ai-mean-anyway/

The struggle between open source and proprietary software is well understood. But the tensions permeating software circles for decades have shuffled into the artificial intelligence space, in part because no one can agree on what “open source” really means in the context of AI.

The New York Times recently published a gushing appraisal of Meta CEO Mark Zuckerberg, noting how his “open source AI” embrace had made him popular once more in Silicon Valley. By most estimations, however, Meta’s Llama-branded large language models aren’t really open source, which highlights the crux of the debate.

It’s this challenge that the Open Source Initiative (OSI) is trying to address, led by executive director Stefano Maffulli (pictured above), through conferences, workshops, panels, webinars, reports and more.

AI ain’t software code

The OSI has been a steward of the Open Source Definition (OSD) for more than a quarter of a century, setting out how the term “open source” can, or should, be applied to software. A license that meets this definition can legitimately be deemed “open source,” though it recognizes a spectrum of licenses ranging from extremely permissive to not quite so permissive.

But transposing legacy licensing and naming conventions from software onto AI is problematic. Joseph Jacks, open source evangelist and founder of VC firm OSS Capital, goes as far as to say that there is “no such thing as open-source AI,” noting that “open source was invented explicitly for software source code.” Further, “neural network weights” (NNWs) — a term used in the world of artificial intelligence to describe the parameters or coefficients through which the network learns during the training process — aren’t in any meaningful way comparable to software.

“Neural net weights are not software source code; they are unreadable by humans, [and they are not] debuggable,” Jacks notes. “Furthermore, the fundamental rights of open source also don’t translate over to NNWs in any congruent manner.”

These inconsistencies last year led Jacks and OSS Capital colleague Heather Meeker to come up with their own definition of sorts, around the concept of “open weights.” And Maffulli, for what it’s worth, agrees with them. “The point is correct,” he told TechCrunch. “One of the initial debates we had was whether to call it open source AI at all, but everyone was already using the term.”